BULLETIN OF

THE BRITISH MUSEUM

(NATURAL HISTORY)

ZOOLOGY

VOL. 9
1962 1963

BRITISH MUSEUM (NATURAL HISTORY)

LONDON: 1963

DATES OF PUBLICATION OF THE PARTS

No. i
No. 2
No. 3
No. 4
No. 5
No. 6
No. 7
No. 8
No. 9

28 September 1962

13 November 1962

13 November 1962

30 November 1962

30 November 1962

30 November 1962

30 November 1962

10 May 1963

14 May 1963

PRINTED IN GREAT BRITAIN
BY THOMAS DE LA RUE &
COMPANY LIMITED LONDON

CONTENTS

ZOOLOGY VOLUME g

PAGE

No. i. Land birds of Guadalcanal and the San Cristoval group, Eastern

Solomon Islands. By IAN C. J. GALBRAITH & EBBA H. GALBRAITH i

No. 2 A review of the Indo-Pacific Gizzard Shad genera Nematalosa,
Clupanodon and Konosirus (Pisces : Dorosomatidae) . By P. J. P.
WHITEHEAD 87

No. 3. The Pantanodontinae, edentulous toothcarps from East Africa.

By P. J. P. WHITEHEAD 103

No. 4. A revision of the Lake Victoria Haplochromis species (Pisces,

Cichlidae) part V. By P. H. GREENWOOD (PI. i) 139

No. 5. Notes on, and descriptions of New Hebridean land snails. By

ALAN SOLEM (Pis. 1-2) 215

No. 6. Observations on the Heteromi, an order of Teleost fishes. By N. B.

MARSHALL 249

No. 7. Darwin's type specimens of varieties of Balanus amphitrite. By

J. P. HARDING (Pis. i-io) 271

No. 8. A revision of the genus Gordiodrilus Beddard (Oligochaeta :

Megascolecidae) . By BARRIE G. M. JAMIESON 297

No. 9. Mites of the genus Macrocheles Latr. (Mesostigmata) associated with
coprid beetles in the collections of the British Museum (Natural
History). By G. OWEN EVANS & K. H. HYATT 325

Index to Volume 9 43

LAND BIRDS OF

GUADALCANAL AND THE

SAN CRISTOVAL GROUP,

EASTERN SOLOMON ISLANDS

IAN C. J. GALBRAITH and EBB A H. GALBRAITH

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY VoL 9 No. i

LONDON: 1962

LAND BIRDS OF GUADALCANAL AND THE

SAN CRISTOVAL GROUP,
EASTERN SOLOMON ISLANDS

BY

IAN C. J. GALBRAITH and EBBA H. GALBRAITH

Pp. 1-86 ; 3 Text-figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 9 No. i

LONDON : 1962

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series, corresponding to the Departments
of the Museum, and an Historical series.

Parts will appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 9, No. i of the Zoological
series.

Trustees of the British Museum. 1962

Issued September 1962 Price Twenty Six Shillings

LAND BIRDS OF GUADALCANAL AND THE

SAN CRISTOVAL GROUP,
EASTERN SOLOMON ISLANDS

IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

SYNOPSIS

A collection from Guadalcanal, San Cristoval and Ugi is reported upon, with systematic and
distributional notes covering also the land and freshwater forms not collected. A new subspecies
of Halcyon Moris is described from the Three Sisters group ; a recently-described race of
Tyto alba is synonymized ; changes of arrangement, affecting the nomenclature of forms
currently placed in Coracina tenuirostris, Monarcha barbatus, Myiagraferrocyanea and Myzomela
nigrita, are suggested ; and the application of the the names Pachycephala pectoralis and
Zosterops rendovae is discussed.

INTRODUCTION

THIS paper is primarily a report on the collection of bird-skins made by the Oxford
University (Department of Zoology) Expedition to the Solomon Islands in 1953 ;
but is expanded to cover islands in the San Cristoval group not visited by the expedi-
tion, and land and freshwater forms unrepresented in the collection. New subspecies
discovered in the mountains of Guadalcanal have already been described (Cain &
Galbraith, 1955), while new distributional information and a few taxonomic notes
have been given incidentally (Cain & Galbraith, 1956). The expedition, financed
mainly by the Percy Sladen Trustees and the Parliamentary Grants Committee of the
Royal Society, was led by Dr. A. J. Cain, supported by I. C. J. G. and by native
assistants from the Malanggo district of Guadalcanal and the Ravo district of San
Cristoval. Almost all the linear measurements presented here were taken by
E. H. G., while I. C. J. G. is responsible for the rest of the paper.

Avifaunal Geography

A study of the Solomons avifauna, with special emphasis on the geographical
variation for which it is remarkable, is in preparation by I. C. J. G., and only the barest
outline will be given here. Grover gives geographical and general accounts (1955 :
3-26 ; 1957 ; 1958 : 3-10), and much detailed geological information with valuable
general references (1955 ; 1958 ; 1960), for the British Protectorate. Nissan, Buka
and Bougainville, with the Bismarck Archipelago, belong politically to the Territory
of Papua and New Guinea (see Australia 1950-1951). The area is shown to a
convenient scale, and all but the smallest islands named, in the new Times Atlas
(1958, pi. 15). Apart from minor differences in spelling, only the following names
used here for islands in the Solomons are not to be found in the Atlas : Buena
Vista = Vatilau ; Gower = Ndai ; Murray = Buraku ; Ranongga = Ganongga ;
San Jorge = St. George ; Sikaiana = Stewart ; Simbo, five miles south of Ranongga ;
Small Malaita = Maramasike.

ZOOL. Q. I. I

4 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

The Solomons form the easternmost extension of a zone of islands round the
western, northern, and eastern coasts of New Guinea, beyond the continental shelf
but accessible by island-hopping across gaps of less than two hundred miles
(mainly less than eighty). The avifaunas of these islands the Tenimber and Kei
groups, the Moluccas, Biak and Numfor, and the Bismarcks and Solomons place
them in the Papuan region, and have clearly been derived mainly from New Guinea
by colonization across the sea. The Moluccas have a considerable Malaysian com-
ponent, and the Solomons a smaller Polynesian one. Enclosed by this " Molucco-
Melanesian " zone is that of the islands on the Papuan continental shelf Aru, the
Western Papuan Islands, Japen, and less certainly the d'Entrecasteaux and Louis-

v

MAP i. The Papuan region, showing islands of the " Molucco-Melanesian " zone (black),
in relation to the Papuan continental shelf (stippled to the 2oo-metre isobath, from B.H.I.
1942), Weber's Line of faunal balance (from Mayr, 1944^), and the wider water-gaps of
the Polynesian region.

iades with richer faunas perhaps largely derived by direct colonization at periods
of lower sea-level, yet lacking many birds characteristic of the mainland. Further
out, Micronesia to the north and Southern Melanesia (including the Santa Cruz
group and New Hebrides) to the southeast, form a zone with the much poorer faunas
characteristic of the Polynesian Region (Mayr, 19416). All three zones are avifaun-
ally characterized not merely quantitatively but qualitatively ; though even the
continental islands can have had land connections with one another only through
mainland New Guinea. The remarkable avifaunal resemblances, in the repre-
sentation or otherwise of Papuan stocks, between (say) Biak and the Solomons,
thirteen hundred miles apart, must be almost wholly due to the different colonizing
potentials of the various stocks, across similar barriers and into ecologically similar
habitats.

Within the middle zone, the Bismarcks and Solomons together form a rather
homogeneous unit, distinguished as Northern Melanesia (Mayr, 19416). As might be

LAND BIRDS OF GAUDALCANAL 5

expected, avifaunas progressively poorer in species are encountered in passing from
the Huon Peninsula of New Guinea, through New Britain and New Ireland to the
Solomons. The geography of the area suggests this as the most important coloniza-
tion route, while some direct evidence for it in the patterns of distribution and varia-
tion of some (apparently rather recent) arrivals is not convincingly paralleled for
the possible route to the Solomons through the Louisiades. However, the faunal
parallelism characteristic of the three zones is shown within Northern Melanesia also.
In many stocks, the representative populations in the Solomons are more distinct
and diverse, and thus probably older, than those in the Bismarcks. The striking
endemism of the Solomons avifauna is exemplified by Haliaeetus. H. leucogaster
occurs, without appreciable geographical variation, from India to Australia, New
Guinea and the Bismarcks ; but is represented in the Solomons by the very distinct
H. sanfordi.

Though a fairly compact and homogeneous archipelago seven hundred and twenty
miles across, with no water gaps of much more than fifty miles, and mostly covered
by tropical rain forest the Solomons proper show considerable avifaunal diversity,
in both the representation and the geographical variation of species and superspecies.
The most striking discontinuity is met in crossing the thirty-five mile strait from
Guadalcanal to San Cristoval : no fewer than 29 of about 95 breeding species are
unrepresented, including such widespread and common birds as Accipiter novaehol-
landiae, Ptilinopus superbus, Coracina papuensis, Aplonis cantoroides, Mino dumontii
and Nectariniajugularis. The internal richness in geographical variation is still more
striking : perhaps in no other archipelago does such a high proportion of the avifauna
break up into so many and so diverse representatives across such narrow barriers.
The most celebrated example is the Zoster ops kulambangrae (=rendovae) superspecies
in the New Georgia or Central Solomons group (see Mayr, 1942 : 227 in which the
narrowness of the straits is exaggerated and Mees, 1961 : 143-151). The discon-
tinuities and affinities are not geographically random but express themselves in a
rather clear-cut pattern, which may be formalized in a system of avifaunal districts
and sections (Mayr, 19450 : 275 expanded from Rothschild & Hartert, 1905 : 244).
This system is reproduced here with only minor emendations, though a modified
form is in preparation. The possibilities of confusion in the avifaunally misleading
term " Central Solomon Islands" have been pointed out (Cain & Galbraith, 1956 : 102),
and it has since been adopted in yet another sense (Grover, 1957 ; 1958). Here it is
replaced by " New Georgia Group ", as used by Grover and the Times Atlas (1958).
Names and spellings in this paper are as far as possible those officially current in the
Protectorate.

A. Main Chain

a. Northern Islands (Buka), Bougainville, Shortland Is., (Fauro), Choiseul,

(Robroy, Wagina, Gagi, Barola, Baroraite), Ysabel, (San Jorge)

b. Guadalcanal Group Florida Is., (Savo), Guadalcanal

c. Russell Is.

d. Malaita Group Malaita, (Small Malaita), ? Ulawa

6 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

B. New Georgia Group

a. Western Section Vella Lavella, (Ranongga), ? Simbo

b. Main Section Gizo, Kolombangara, (Wanawana, Arundel), New Georgia,

(Vangunu, Gatukai)

c. Rendova Section Rendova, (Tetipari)

C. San Cristoval Group San Cristoval, Ugi, (Bio), Three Sisters Is., Santa Ana,
(Santa Catalina)

Parentheses indicate small or moderate-sized islands, lying close to the groups with
which they are associated, for which avifaunal lists are not available. Unpublished
distributional information from the Whitney Expedition was no doubt used by
Mayr (19450) in associating some of them with districts and sections. Smaller offshore
islands probably resemble the more isolated ones Nissan, Treasury, Murray,
Buena Vista, Ramos, Gower, Sikaiana and Ongtong Java ; and perhaps Simbo
(Sibley, 1951 : 83) and Ulawa (Cain & Galbraith, 1956 : 104) in having such depleted
avifaunas, composed almost entirely of widespread forms, that they cannot be associ-
ated faunally with any particular subdivision of the Solomons (nor indeed with one
another, within the unity imposed by the ecologically-limited range of potential
colonists).

Rennell, with Bellona, has a very distinct avifauna, somewhat intermediate in its
affinities between Northern and Southern Melanesia (Mayr, 19310 ; Bradley & Wolff,
1956 ; Braestrup, 1956), while the Santa Cruz group is associated with the Solomons
only politically, and has a largely Southern Melanesian avifauna.

Within the Solomons proper, the most fundamental cleavage is between the Main
Chain and New Georgia Group on the one hand, and the San Cristoval Group on the
other. Failure of almost a third of Guadalcanal species to colonize San Cristoval has
already been mentioned ; while of 67 San Cristoval species whose distribution is
sufficiently known, 17 are represented on Guadalcanal by different (mostly very
distinct) subspecies, 5 by different species, and 7 are not represented. Between
Guadalcanal and San Cristoval there is a major break in the southeastwards spread
of Papuan birds, comparable to that between the Bismarcks and Solomons, and a
lesser one in the northwestwards spread of Polynesian forms. Although evident in
Mayr's (19450) summary of the Solomons avifauna, this striking discontinuity has
not been especially remarked ; largely because most of the San Cristoval endemics
were discovered early and only gradually localized within the Solomons, while the
absence of otherwise characteristic species has also been only slowly recognized. Ugi
and the other small islands have reduced but distinctively San Cristoval avifaunas,
with some specialization. The presence of Aplonis cantor oides and N ectarinia jugularis
on these islands, though not on San Cristoval itself, supports other evidence in sug-
gesting that the latter is not an avifaunal vacuum preserved by physical isolation ;
but that the endemic forms have redeployed ecologically to form a thin but harmo-
nious avifauna which presents an effective barrier to colonization. This in turn suggests
that the island may have had a distinctive history as a terrestrial biotope during the
Caenozoic which the bathygraphy, relief, stratigraphy and tectonics of the Solomons
(Grover, 1955 ; 1958 ; 1960) do not contradict.

LAND BIRDS OF GUADALCANAL 7

Whereas the San Cristoval forms Coracina salomonis, Monarcha vidua and Myiagra
cervinicauda have recently been regarded as conspecific with their representatives
elsewhere in the Solomons (Mayr, 19450 ; 1955), we treat them as full species. Some
of Mayr's species-arrangements of allopatric forms are too much influenced by geo-
graphy rather than characters, and so may be misleading when faunal affinities are to
be considered. Possibly we have fallen into the opposite error through over-preoccu-
pation with the peculiarity of San Cristoval. However, other recent authors have
divided some of Mayr's (19450) polytypic species on similar grounds (e.g. Mayr,
I 955 : 2 3 J Amadon, 1956 : 23 ; Mees, 1961 : 132).

Avifaunal differences between and within the Main Chain and New Georgia Group
are much less profound than those which separate the San Cristoval Group. Bougain-
ville and Guadalcanal, the largest and most massive islands, have the richest and
most typical avifaunas, with surprisingly little differentiation between them. The
long but narrow and rather low islands of Choiseul and Ysabel lack many species
common to Bougainville at one end and Guadalcanal at the other, while the few
differentiating forms which they do possess link them rather more with the former.
The low-lying Russell group has a poorly-differentiated fauna reminiscent of those of
the more isolated islands, but with a surprising number of endemic subspecies. The
avifauna of Malaita can be derived almost entirely from that of Guadalcanal, but has
many very distinct representatives. Still more distinct, and showing some special
affinities both with the Northern Islands and with Guadalcanal, the New Georgia
Group is distinguished by marked internal variation, in five species and superspecies,
upon which its subdivision is largely based. Though New Georgia is the largest island
of the group, the taller volcanic cone of Kolombangara supports the richest avifauna.

Ornithological Literature

The avifauna of the Solomons is admirably summarized (Mayr, 1945 a), in a work
indispensable in the field or study which embodies otherwise unpublished records and
taxonomic judgements. A series of papers on Northern Melanesia as a whole (Mayr,
19456 ; 19496 ; 1955 ; 1957) is unfortunately incomplete ; so that no comprehensive
recent work treats the pigeons, parrots, kingfishers, or several other non-passerine
groups, of the Bismarcks. The Novitates Zoologicae (to 1932) and the American
Museum Novitates (from 1924) are the richest sources for revisions which include
Northern Melanesian birds.

Mayr (1955) lists the most important regional works on Northern Melanesia, dating
from 1899 to 1951. Several subsequent papers, some based largely on field observa-
tions, have to be added (Bradley, 1957 ; 1962 ; Bradley & Wolff, 1956 ; Cain &
Galbraith, 1955 ; 1956 ; 1957 ; French, 1957). The latest development is the explora-
tion of mountains on New Britain, which have already yielded two remarkable new
species (Gilliard, 19600 & 6). Mayr omits several papers based largely on field obser-
vations (Dahl, 1899 ; Donaghho, 1950 ; Hamlin, 1931 ; Meyer, 1927 ; 1930 ; 1933 ;
1937). Useful details may be gleaned from earlier works, and miscellaneous papers
also omitted (Beecher, 1945 ; Danis, 1938 ; Davidson, 1929 ; 1934 ; Hartert, 1908 ;
1929; Mayr & Camras, 1938 ; Ogilvie-Grant, 1887; 1888; Ramsay, 18790 & b '> 1881;
18820, c & e ; 1883 ; Schonwetter, 1935 ; Sharpe, 1888 ; Sibley, 1946 ; Stresemann,

MAP 2. Localities visited on Guadalcanal (from D.C.S. 1955, sheets 2, 3, 6, j, 10 and n,
with modifications from Grover, 1955 and 1960). Names and altitudes obtained by the
expedition are shown in parentheses, and uncertain localities are indicated by queries.

LAND BIRDS OF GUADALCANAL g

1933; Tristram, 1879 ; 1882; 1892; 1894; 1895; White, 1938). However, the
earlier locality records must be treated with reserve : Ramsay's unreliability in this
respect has been repeatedly remarked. Still earlier reports are now of mainly nomen-
clatural importance (see Mayr (19336) on Sclater's (1869) ostensible Solomons
collection) .

Though ecological interactions, and geographical variation in ecology and behaviour,
present some of the most interesting problems in the Solomons avifauna, very little
information is avilable. Field notes and locality records from the Whitney Expedition
must be a rich source of unpublished data, especially on the geographical variation of
altitudinal distribution (see Mayr, 19450; on Rhipidura rufifrons and Pachycephala
pectoralis). Fairly comprehensive field-notes have been published for areas on Bou-
gainville (Virtue, 1947), New Georgia (Sibley, 1951), Guadalcanal (Donaghho,
1950 ; Cain & Galbraith, 1956), San Cristoval and Ugi (Cain & Galbraith, 1956), and
Rennell (Bradley & Wolff, 1957). Comparison of major islands, and of altitudinal
zones, has been attempted only by Cain & Galbraith (1956).

Collecting Localities

At the time of the Oxford expedition, the best available maps of the Solomons were
the American military series at i : 500,000 (A. M.S. X4oi), and larger maps of indi-
vidual islands prepared by the Lands Department ; both useless for inland localities.
Sketch maps at i : 50,000 are now available for Guadalcanal (D.C.S. 1955) and for
San Cristoval, Ugi, Bio, the Three Sisters, Santa Ana, Santa Catalina, and Ulawa
(D.O.S. 1958). Maps 2 and 3 are based on the appropriate sheets of these, with
additions from Geological Survey maps of Guadalcanal (Grover, 1955 figs. 9 & 48 ;
1960 fig. 42), and a sketch-map of eastern San Cristoval provided by Mr. R. B. M.
Thompson.

Only localities at which collections were made are listed here. Native collectors
based at these localities ranged for several miles in search of specimens. Cain &
Galbraith (1956 : 106) briefly describe all the localities visited. They use a more
phonetic spelling than the official form, rendering the Melanesian " b " as " mb "
and the San Cristoval " g " as " gh ".

Guadalcanal : Tenaru. Tenaru Mission School on coastal plain, among grassland,
lowland forest, second-growth jungle, and cultivated land.

Guadalcanal : Tsea. Waterfall on Tenaru River near sea-level (see Grover, 1955
fig. 9 & pi. VIII), in lowland forest with gardens and second growth.

Guadalcanal : Betilonga. Village near floor of Betilonga basin at 1,800 ft (see
Grover, 1955 Ch. XXIV ; Hill in Grover, 1960 Rept. 21). Hill forest on ridges rising
to about 2,500 ft, with clearings and second growth.

Guadalcanal : Nala. Deserted village on spur above Tasi (? Baliai) River at about
i, 600 ft, at foot of Kavo range (see Grover, 1960 fig. 42). Village clearing and extensive
second growth in hill forest.

Guadalcanal : Turipava. Camp on spur of Kavo range above Nala, at about 4,000
ft in lower mist forest. The position of the spur was determined, from the drainage
pattern of the area and compass bearings taken from about 5,000 ft at Vatuvulele,
before publication of the map (Grover, 1960 fig. 42) confirmed the positions of the

io IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

Nunukama/Betibau confluence and Nala. The peak marked as Mt. Tambunugnu at
5,300 ft may well be the high point recorded by Cain & Galbraith as Tambunangu
(" forbidden to speak ") at about 6,000 ft, but the position of the sites along the spur
is conjectural.

San Cristoval : Manewiriwiri. Village on coast of Wainoni Bay, among coconut
plantations, gardens and second growth in coastal forest.

San Cristoval : Goge. Village in Warihito valley at Goge confluence, near sea-level
in lowland forest with extensive clearings and second growth. The name of the village
is correctly Napagiwae ; but this was not used by the natives during I. C. J. G.'s stay,
and is ignored here for ease of reference to Cain & Galbraith (1956) and the specimen
labels.

San Cristoval : Nagasi. Inland ridge rising to about 2,200 ft, with hill forest,
second growth and deserted village clearing. The position of this ridge is very
uncertain, and merely suggested on map 3. Unfortunately, no compass directions
were recorded, and no views of recognizable points were obtained. The following may
later assist in the identification of the Nagasi ridge. Crossing the Goge River from
Goge, the path southwards up the crest of the nearby ridge was followed (through a
deserted village, Nanoni, at about 1,400 ft after 2hr.), without descending appreciably
or crossing water until, after a sharp descent to the left from a ridge at about 1,700 ft,
a small river, Halauma, flowing from right to left, was reached in 3 hours rapid march-
ing. A recently deserted village, Wuranakumau, lay across this and io minutes climb
away, at about 1,400 ft on a platform of the Nagasi ridge. The path to the highest
point, from which the return to Wuranakumau took 40 minutes rapid descent,
approximately followed the crest, past overgrown sites, Huniara at about 1,600 ft,
and Nagasi at about 2,000 ft. Above Huniara it crossed a swampy saddle, with water
trickling to the right. Just above Nagasi the path branched ; the left-hand branch
was said to go towards the south coast, the other back to the Goge path. A view from
the left of the latter showed at least three parallel ridges at about 2,000 ft, separated
by steep valleys. Both paths soon began to descend. Thompson (in litt.}, who had
previously visited Namasi, made enquiries in the Wainoni area, and was convinced
that the Halauma was the Haruta or upper Warihito River, where there was once a
village Kehaha at the crossing (names not shown in map 3), and that the Nagasi ridge
lay between this and the Waiakoko River, while the path down from the ridge prob-
ably reached the south coast at Marauni. The only discordant point is a clear though
restricted view, to the right of the path below Nagasi, sharply down to a steep shore
at an acute angle to the ridge, and an apparently extensive body of water. It seems
impossible, from the distances involved, that this could have been the south coast as
I. C. J. G. supposed at the time ; and hardly possible that it could have been the shore
of a large unknown lake in the area not covered by aerial photographs (D.O.S. 1958
sheet 7).

Ugi. Although the collectors worked from the mission school at Pawa, they
ranged over at least the southern half of the island, and did not collect within the
extensive grounds of the school. The island reaches a height of 570 ft, and supports
a variety of habitats lowland forest, old and new second growth, gardens of native
and European types, fields, coconut plantations and coastal formations.

Rivers & streams
Paths (approximate) ^
Heights in feet

MAP 3. Localities visited on San Cristoval (from D.O.S. 1958, sheets 2, 4, 5, 7 and 8). Names
and altitudes obtained by the expedition, are shown in parentheses, and uncertain
localites are indicated by queries.

12 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

The distinction drawn by Cain & Galbraith (1956), between hill forest on Guadal-
canal (Betilonga) and " ridge forest " at comparable altitudes on San Cristoval
(Nagasi) , is probably not tenable. Intensive botanical study would be necessary to
establish a real difference between primary forest at about 2,000 ft on the two islands,
since forest physiognomy is considerably subject to local influences, while old second
growth (especially common on San Cristoval Walker, 1948 : 49) may be misleading.
Avifaunally, the evidence is at most suggestive. Phylloscopus trivirgatus is the only
species, taken at Nagasi, which was present on Guadalcanal in the montane mist
forest (Turipava) but absent from Betilonga. Edithornis silvestris, Vitia par ens and
Rhipidura fuliginosa, not known to be represented on Guadalcanal, might prove to
be hill forest birds on San Cristoval ; as might Zoothera margaretae on both islands.
Besides Phylloscopus, the only other species known from both islands, and strictly
montane on Guadalcanal, is Petroica multicolor, whose altitudinal distribution on
different islands has not been recorded (Mayr, 19346) though " above 1800 feet "
(Mayr, 19450) may well refer only to San Cristoval. The attempt to delimit zones of
bird life in the Solomons, to study their boundaries, and to compare strictly similar
habitats on different islands, has yet to be made.

THE LAND AND FRESHWATER BIRDS

Although the expedition was concerned only with the land and freshwater birds
(in the sense of Mayr, 19450, the Charadriformes, Procellariiformes, Phaethontidae,
Sulidae and Fregatidae being neglected), a very few waders were taken incidentally :

Actitis hypoleucos (Linnaeus, 1758). Guadalcanal : Betilonga i unsexed 25th
August ; Tenaru i $ i5th September. San Cristoval : Goge i <$ 3oth October, i $
3rd December, 2 unsexed 2nd and 3rd November.

Heteroscelus incanus brevipes (Vieillot, 1816). San Cristoval : Goge i 27th
October.

Calidris acuminata (Horsfield, 1821). San Cristoval : Goge i < i4th November,
i unsexed 29th October. Ugi : Pawa i unsexed igth December.

The Heteroscelus, in winter plumage, was identified subspecifically according to
Stickney (1943). As Bull (1948) found, the white bars on the upper tail-coverts are
poorly developed.

Every species and subspecies of land and freshwater bird known from the islands of
Guadalcanal, San Cristoval, Ugi, the Three Sisters and Santa Ana is listed below, a
total of 154 forms. Those represented in the expedition collection are shown in bold-
face type. The number, prefixed by " M.", associated with each name is for reference
to Mayr (19450 : 212-280) and Cain & Galbraith (1956) : there are several gaps, while
conspecific races, and sometimes species of a single superspecies, bear the same
number. Additional species (marked* by Cain & Galbraith) are distinguished by
letters :

A. Pelecanus conspicillatus (p. 14) F. Poliolimnas cinereus (p. 22)

B. Egretta intermedia (p. 15) G. Cichlornis whitneyi (p. 52)

C. Platalea regia (p. 16) H. Gymnorhina tiUcen (p. 71)

D. Circus approximans (p. 19) J. Myzantha melanocephala (p. 75)

E. Falco peregrinus (p. 19)

LAND BIRDS OF GUADALCANAL 13

The names adopted are those used by Mayr (latest of 19410 ; 19450 ; 19456 ;
19496 ; 1955 ; 1957), except as otherwise discussed under " Notes ". Nomenclatural
references are not given, since these are easily found (especially from Mathews,
1927-30 ; 1931 ; 1932 ; 1933). Reference is made to recent revisions where possible,
otherwise to a standard work (usually the Catalogue of Birds in the British Museum,
London 1874-98, cited by volume as Cat. Birds B.M.), or in a few cases to several
partial discriminations. References to Mayr, 19450 are given for species not there
recorded from the Solomons. More general references are introduced by " See . . .".

In the range indications, " Solomons " includes only the islands from Buka to
Santa Ana, and not Nissan, Ongtong Java, Rennell, Bellona or the Santa Cruz group.
Forms which are not native residents of the Solomons are indicated as introduced,
migrant or straggling ; those which are not normally to be found in or near the major
hill-forest habitat are indicated as coastal, lowland or montane. The available
information on altitudinal and ecological distribution is summarized by Cain &
Galbraith (1956).

In the sections on the specimens collected, sexes and growth-stages are indicated
by : (, male ; $, female ; ?, unsexed ; ad, adult ; subad, subadult ; imm, immature ; juv,
juvenile ; nest, nestling ; ret, retarded phase. A few unsexed specimens (mainly adult
males of strongly dimorphic forms) have been sexed by plumage. Growth-stages were
determined from plumage. Most of the specimens are in the British Museum (Natural
History), referred to as BM(NH), but some are indicated as having been sent to the
American Museum of Natural History (AMNH) or the Oxford University Museum
(OUM). The destination of a few specimens in spirit has not yet been decided. The
BM(NH) specimens are registered as nos. 1959.21.1-1590, 1604-1611.

Specimens were collected at the various stations between the following dates in
1953 (expedition serial numbers in parentheses) :

Betilonga 4th June-24th July (1-27, 31-337)

3rd August-i2th August (356-443)

22nd August-29th August (563-613)
Tsea 9th June-ioth June (28-30)

Turipava 3oth July-3ist July (338-348)

i5th August-2oth August (444-562)
Nala 3ist July-ist August (349~355)

Tenaru 3rd September-24th September (614-834)

Goge 8th October-i4th November (835-1404)

30th November-3rd December (1573-1648)
Nagasi i8th November-26th November (1405-1572)

Manewiriwiri 6th December-8th December (1649-1712)
Pawa I4th December-3oth December (1713-2003)

From 26th October to 3ist October (nos. 1115-1191), the native collectors were
working unsupervised, recording sex and gonadial development pictorially, but not
weights nor colours. These determinations seem wholly reliable : there is a reasonable
proportion of undetermined specimens, and the recorded gonadial states do not

I 4 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

conflict with the plumage, even where external differentiation is inconspicuous.
Labels from this period are marked " nat. coll.". All other specimens were sexed,
weighed and described by A. J. Cain or I. C. J. G.

The freshly collected birds, if not severely damaged, were weighed without allow-
ance for loss of blood and feathers or for lodged shot. Weights less than 185 gm
were taken to the nearest 0-5 gm, with a long-scale spring balance designed by Dr.
J. A. Gibb for weighing small passerines alive (as advertised by the B.T.O. in Bird
Study] . Used with a 50 gm weight for the lighter specimens, this gave a linear response
over its whole range, and would have been reliable to a greater degree of accuracy.
It was calibrated in the field, and checked at intervals. Weights between 185 and
500 gm were taken with a dietary balance to the nearest 5-0 gm, and their means
are given to the nearest i-o gm. Weights of discarded and spirit specimens have been
used where possible, so that a weight series may be longer than the corresponding
series of skins. Linear measurements were almost all taken by E. H. G., while re-
measurement of some series by I. C. J. G. showed good agreement. Except where
otherwise stated, wings were measured flattened down, tails from the common sheath
to the tip of the longer central rectrix, and bills from the angle of skull and culmen
chord wise to the tip of the mandible.

Weights are given in grams, linear measurements in millimetres, to the nearest
0-5 unit, and their means to the nearest o-i unit (except for weights above 185 gm).
Measurements of unsexed specimens are not usually given. Though the standard
deviation should ideally be given as the measure of spread, the undesirability of
combining or accurately comparing measurements made by different workers makes
the labour of computation not worthwhile in reports on collections, and the range is
here given instead. Three or more measurements are given in the form : number of
individuals measured /sex /stage /range /(mean) ; e.g. 5 $ ad 42-5-52 (46-7). For 2
specimens, the measurements are given instead of the range and mean.

Descriptions of the colours of soft parts are compiled from the field-labels for each
series, including those of discarded and spirit specimens. They apply to both sexes
and all represented stages, unless otherwise stated ; stages represented but not
mentioned agree with the next older stage. Other notes on individual specimens refer
to those taken by the expedition, unless otherwise noted. In the colour-comparison
of Ptilinopus solomonensis and Aplonis metallica (pp. 24 and 68), the longest and
most variable series was divided into as many colour-classes as could be re-assembled
with little error, to which the remaining series were then matched, rather than into a
predetermined number of classes.

(A) Pelecanus conspicillatus Temminck, 1824

REFERENCES. Cat. Birds B.M. 26 : 483 ; Alexander, 1954 : 189.

RANGE. Australia ; straggling from Tenimber Isles to Melanesia (coasts and
lowlands) .

NOTES. Recorded from the Solomons, New Hebrides and Rennell after high winds
in 1952 (Anon, 1952 ; Dorst, 1954 ; Laird, 1954 ; Bradley & Wolff, 1956 ; Cain &
Galbraith, 1956).

LAND BIRDS OF GUADALCANAL 15

(M. 2) Phalacrocorax melanoleucos melanoleucos (Vieillot, 1817)

REFERENCES. Amadon, 19420 : 2.

RANGE. Java and Celebes to Australia and Santa Cruz, including Guadalcanal and
Santa Ana (lowlands).

SPECIMENS. Guadalcanal : Betilonga i $ ad. To OUM.

Bill 42.

Iris dark brown, silvery outer ring. Skin round eye yellowish fuscous, chin dirty
yellow. Bill yellow, maxilla mottled with black. Foot black.

NOTES. The known range of this species is irregular. In the Solomons proper,
it is well-known from Guadalcanal (Mayr, 19450 ; Donaghho, 1950 ; Cain & Galbraith,
1956) and Santa Ana (Mayr, 19450 ; French, in litt.), while the endemic race brevi-
cauda occurs on Rennell. A single specimen has been recorded from Bougainville
(White, 1938), though the species is not known from the Bismarcks. Another, hitherto
unrecorded, was taken at Star Harbour, San Cristoval, by Commander J. F. A.
O'Neill of the Melanesian Mission's m.v. " Southern Cross ". It was examined by
I. C. J . G. at Pawa in January 1954, but was later lost when the refrigerator broke
down (Turbott, in litt.}.

Though liable to be neglected by collectors, the species is conspicuous where it
occurs. Virtue (1947) on Bougainville and Sibley (1951) on New Georgia apparently
did not study suitable habitats on lakes or large rivers, so that their failure to record
it is not conclusive. However, it definitely does not seem to maintain itself on San
Cristoval. French (in litt.} doubts its presence there, natives did not recognize its
unmistakable description, and it was not seen during six weeks on the apparently
suitable Warihito River. The Star Harbour specimen must surely have been a
straggler from Santa Ana, across the three-and-a-half mile strait.

The presence of the species on several small or low-lying islands without rivers
(e.g. Santa Ana, Rennell, Ticopia) clearly depends on their possession of freshwater
lagoons. However, its absence from such large islands as San Cristoval is surprising.

(M. 3) Butorides striatus solomonensis Mayr, 1940

REFERENCES. Mayr, 19400 : 6. See Bock, 1956.

RANGE. Solomons (lowlands).

SPECIMENS. Guadalcanal : Bilaha, Tenaru R. i $ ad.

Weight 245. Wing 181. Tail 64. Bill (from feathering) 66-5.

7ns bright yellow. Eyelid dull cherry-red. Skin round eye pale brown. Bill maxilla
black, mandible dirty white. Foot dirty fuscous green, shank bright yellow behind,
sole dirty yellow.

NOTES. We follow the generic arrangement of Bock (1956) for the Ardeidae.

(B) Egretta intermedia plumifera (Gould, 1848)

REFERENCES. Amadon & Woolfenden, 1952 : n. See Bock, 1956.
RANGE. Australia, New Guinea region, southern Moluccas (lowlands) ; San
Cristoval straggler ?

16 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

NOTES. The young bird taken by French at Kira Kira (Cain & Galbraith, 1956 ;
French, 1957) may have been a straggler. However, the extensive marshes of San
Cristoval (Grover, 1958) provide a habitat suitable for the species, which because of
its wariness and inaccessibility was not for a long while recognized as widespread and
locally common in New Guinea (Rand, 1938 : 291).

(M. 4) Egretta sacra sacra (Gmelin, 1789)

REFERENCES. Mayr & Amadon, 1941 : 3. See Bock, 1956.
RANGE. Asia to western Polynesia (coasts and lowlands).

(M. 5) Nycticorax caledonicus mandibularis Ogilvie-Grant, 1888

REFERENCES. Amadon, 19420 : 5. See Bock, 1956.

RANGE. Solomons, Nissan and New Ireland intergrading via Bismarcks with
N.c. hilli (lowlands).

SPECIMENS. Guadalcanal : Betilonga 2 $ ad, i $ ad. i <$ ad to OUM.

Wing 2 <$ ad 270, 272 ; i $ ad 262. Tail i $ ad 95. Bill (from feathering) 2 <$ ad
66-5, 69-5 ; i ? ad 63-5.

Iris yellow. Skin round eye dull yellow or yellow-green. Bill black, $ mandible
yellowish. Foot yellow.

(M. 6) Ixobrychus flavicollis woodfordi (Ogilvie-Grant, 1888)

REFERENCES. Mayr, 19456 : 5 ; Bradley, 1962 : n. See Bock, 1956.

RANGE. Solomons and Rennell (lowlands).

NOTES. This apparently widespread, though rare or retiring, species has not yet
been recorded from the San Cristoval group In addition to the localities already
recorded (Mayr, 19450), there is a specimen in the BM(NH) collected at Bouin,
Bongainville by W. F. H. Rosenberg.

(C) Platalea regia Gould, 1838

REFERENCES. Mayr, 19450 : 280 ; Amadon & Woolfenden, 1952 : 5.

RANGE. Australia, straggling as far as Celebes, Rennell and New Zealand (low-
lands).

NOTES. We follow Amadon & Woolfenden in regarding regia as a distinct
species. It is recorded as a rare non-breeding visitor to the Three Sisters (French,
1957), perhaps from Rennell where it may be resident (Bradley & Wolff, 1956 : 92).

(M. 7) Anas poecilorhyncha pelewensis Hartlaub & Finsch, 1872

REFERENCES. Amadon, 19430 : 3.

RANGE. Micronesia and northern New Guinea to Polynesia (lowlands).
SPECIMENS. Guadalcanal : Betilonga i $ ad. To OUM.
Wing 217. Bill (from feathering) 46-5.
Iris light brown. Bill bluish grey. Foot dull fawn.

NOTES. We follow Delacour & Mayr (1945 : 21) in uniting specifically the superciliosa
group of Australasian mallards with luzonica of the Philippines and poecilorhyncha of

LAND BIRDS OF GUADALCANAL 17

Malaysia and southeast Asia. The species has been recorded from the San Cristoval
Group only by Ramsay (i882c : 41), but was familiar to natives of Goge (Cain &
Galbraith, 1956 : 117).

(M. 8) Aviceda subcristata gurneyi (Ramsay, 1881)

REFERENCES. Mayr, 19456 : 9.

RANGE. Guadalcanal, Malaita and San Cristoval Group (mainly lowlands).
SPECIMENS. Guadalcanal : Betilonga i $ ad. To OUM.
Weight 340. Wing 316. Tail 193. Bill 30.

/n's bright yellow. Bill maxilla black, cere and mandible pale silvery blue. Foot
white.

(M. 9) Haliastur Indus flavirostris Condon & Amadon, 1954

REFERENCES. Condon & Amadon, 1954 : 206.

RANGE. Feni, Duke of York, Solomons (mainly coasts and lowlands).

SPECIMENS. Ugi i ? juv.

Wing 375. Bill 136-5.

7ns dark brown. Eyelid dark slate, visor and skin round eye dirty yellowish-green.
Bill blackish slate streaked with bluish grey, gape dull greenish yellow, cere slaty
black. Foot lemon-yellow, claws black.

The bills of this specimen and two BM(NH) juveniles are conspicuously heavier
than those of girrenera, though not paler in the skin.

NOTES. A BM(NH) subadult specimen from the Duke of York group has the
heavy and wholly yellow bill characteristic of this race ; which with Ptilinopus
viridis lewisi, Ducula pistrinaria pistrinaria, Eos cardinalis, Hemiprocne mystacea
woodfordiana, Eurystomus orientalis solomonensis and Hirundo tahitica subfusca
(and probably Tyto alba cf . crassirostris) is a form characteristic of the Solomons
which has colonized certain small islands in the Bismarcks.

(M. 10) Accipiter meyerianus (Sharpe, 1878)

REFERENCES. Mayr, 19340 : i.

RANGE. Moluccas, Japen, Bismarcks, Solomons Kolombangara, Guadalcanal
(mountains ?).

NOTES. This very rare species is known in the Solomons by two specimens
(Hartert, 1929 : 5).

(M. n) Accipiter novaehollandiae pulchellus (Ramsay, 1882)

REFERENCES. Hartert, 1929 : 4. See Mayr, 19456.
RANGE. Guadalcanal (mainly lowlands).
SPECIMENS. Guadalcanal : Tenaru i $ ad. To OUM.
Weight 280. Wing 221. Tail 173. Bill 29.

7ns very dark reddish brown. Eyelid dull orange, visor dull yellow. Bill black,
cere and base orange. Foot orange, claws black.

This specimen (OUM No. 6/3978) is uncharacteristic, and differs from three

ZOOL 9. I. 2

i8 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

BM(NH) females oipulchellus from Tenaru and Lunga, in having much deeper rufous
thighs and under tail-coverts, not contrasting strongly with the belly, and somewhat
more rufous on the under wing-coverts. However it agrees with pulchellus, and differs
from rufoschistaceus-rubianae-bougainvillei, in having the thighs and under tail-
coverts decidedly paler than the belly, the under wing-coverts primarily grey-and-
white-speckled rather than solid rufous, and extensive white areas at the bases of the
inner webs of the wing quills.

(M. 12) Accipiter albogularis woodfordi (Sharpe,

REFERENCES. Mayr, 1957 : 7.

RANGE. Main Chain of the Solomons, Bougainville to Ulawa.

SPECIMENS. Guadalcanal : Betilonga i $ ad, I $ ad, i $ imm. I ^ ad to OUM.

Weight i <$ ad 265 ; i $ imm 335. Wing i $ ad 210 ; i $ ad 248 ; i $ imm 241.
Tail i $ ad 146 ; i $ ad 123. Bill i $ ad 26 ; i $ ad 30 ; i $ imm 28-5.

Iris orange ; imm yellow. Eyelid yellow, visor fuscous green, skin of chin greenish
yellow. Bill <$ black, base of mandible pale blue-grey ; $ dark grey, base of mandible
white ; cere greenish yellow ; imm bill fuscous-streaked. Foot yellow, claws black.

The immature resembles a juvenile in the normal (i.e. not the " holomelas ") phase,
but has the mantle blacker, the ground-colour of the underparts almost pure white,
the underwing with bold chevrons rather than narrow shaft-streaks, and the thighs
with pale rufous chevrons and a few sub-basal bars rather than shaft-streaks. The
adults are both in the normal phase, with the rufous collar slightly developed.

(M. 12) Accipiter albogularis albogularis Gray, 1870

REFERENCES. Mayr, 1957 : 10.

RANGE. San Cristoval Group.

SPECIMENS. San Cristoval : Goge i $ ad.

Weight 280. Wing 243. Tail 170. Bill 29-5.

Iris deep orange-yellow. Eyelid deep yellow, visor pale fuscous green. Bill dark
grey above, base and mandible pale and bluish, cere dull pale yellow-green. Foot
orange-yellow, shanks paler and greener.

Like all specimens known from San Cristoval or Santa Ana, this is in the normal
phase, without any trace of a rufous collar and with black speckling on the sides of
the breast.

NOTES. Two of the seven specimens in the BM(NH) are from San Cristoval :
an unsexed adult and a male juvenile, both in the normal phase. The other five are
among those recorded by Ramsay (1882^ : 30) as having been collected on Ugi by
Stephens. Formerly in the Gurney collection (Gurney, 1884 : 33-34, footnote 4),
these have been added to the BM(NH) collection since Mayr's (1957) paper. Two are
adults in the normal phase, agreeing well with specimens from San Cristoval. The
remaining adult and two juveniles are in the " holomelas " phase, otherwise unknown
outside the range of woodfordi. It is unfortunate that Ramsay's known unreliability
over locality records makes it impossible to accept Ugi as their provenance without
confirmatory evidence. Doubt of the record is increased by the fact that the locality

LAND BIRDS OF GUADALCANAL 19

" Ugi Island " has been added subsequently to the labels, though in the same hand.
Adults of albogularis cannot certainly be distinguished from woodfordi. However, the
" holomelas " juveniles agree closely with each other, and differ from one such of
woodfordi, in having the spotting of the breast give way to the barring of the belly
much more gradually (in which they agree with a normal juvenile of albogularis,
though in the latter the transition takes place much lower on the breast), the under-
wing with less black, and the thighs with very vague chevrons rather than shaft-
streaks. It would be interesting to find the all-black "holomelas " phase on Ugi but
not on San Cristoval, in view of the tendency towards melanism of flycatchers on the
former island (Rhipidura rufifrons ugiensis, Monarcha castaneiventris ugiensis, M.
vidua squamulatus] .

(M. 14) Haliaeetus sanfordi Mayr, 1935

REFERENCES. Mayr, 1936 : i.

RANGE. Solomons.

NOTES. Definitely known only from the larger islands (including Bougainville
Beecher, 1945 : 35) and the Three Sisters (Cain & Galbraith, 1956 : 106), though
reported from Ulawa and Ugi (op. cit. : 118). Condon & Amadon (1954 : 230) discuss
the status of the representative species H. leucogaster on Nissan.

(D) Circus approximans gouldi Bonaparte, 1850

REFERENCES. Amadon, 1941 : 371 ; Mayr, 1945^ : 54.

RANGE. Australia, southern New Guinea, Solomons, Southern Melanesia, western
Polynesia, New Zealand (lowlands).

NOTES. In the Solomons, known from the lowlands of Guadalcanal (Beecher,
1945 : 35) and as a rare non-breeding visitor to the Three Sisters (French, 1957).

(M. 15) Pandion haliaetus melvillensis Mathews, 1912

REFERENCES. Amadon, 1941 : 376.

RANGE. Malaysia to Micronesia, Solomons and New Caledonia (coasts).

(M. 16) Falco severus papuanus Meyer & Wigglesworth, 1893

REFERENCES. Mayr, 19456 : n.

RANGE. Celebes to New Guinea and Northern Melanesia.

NOTES. In the Solomons, known from Bougainville (Sibley, 1946 : 97) and Gizo
(Mayr, 19456), and doubtfully recorded from Guadalcanal (Cain & Galbraith, 1956 :
119).

(E) Falco peregrinus ernesti Sharpe, 1894

REFERENCES. Mayr, i94ic : i ; 1945^ : 56.

RANGE. Malaysia to Philippines, New Guinea region and Northern Melanesia.
NOTES. In the Solomons, known only as a rare non-breeding visitor to the Three
Sisters (French, 1957 ; see Cain & Galbraith, 1957)-

2 o IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(M. 17) Megapodius freycinet eremita Hartlaub, 1867

REFERENCES. Mayr, 19380 : 12.

RANGE. Northern Melanesia, with irregular minor geographical variation (coasts
and lowlands).

SPECIMENS. Guadalcanal : Betilonga i <$ ad, 2 $ ad. San Cristoval : Goge i <$
ad, i ? juv. T $ ad to OUM.

Weight i ? juv 68. Wing 2 $ ad 221, 227 ; 2 $ ad 225, 227 ; i ? juv 55. Tail i
ad 72 ; 2 $ ad 71, 73. Bill 2 $ ad 28, 29-5 ; 2 $ ad 26-5, 27 ; i ? juv 14.

7ns brown ; juv grey-brown. Skin of throat and sides of face dull purplish-red
to dull pink, crown strawberry ; juv throat pinkish-brown. Bill yellow, greenish
basally ; juv dark fuscous grey. Foot dark olive green to blackish fuscous ; juv dark
grey-brown.

NOTES. Frith (1956) summarizes the incubation methods of megapodes, which
seem to be especially diverse in the Solomons.

(M. 18) Turnix maculosa salomonis Mayr, 1938

REFERENCES. Mayr, 19386 : 3 ; Slitter, 1955 : 114.

RANGE. Guadalcanal (lowlands).

SPECIMENS. Guadalcanal : Tenaru 3 <$ ad, i <$ juv, i $ ad. i ^ ad to AMNH.

Weight 3 3 ad 38-5, 39-5, 41 ; i juv 31 ; 2 $ ad 52-5, 57-5. Wing 3 ad 75, 76,
77 ; i ^ juv 72 ; i $ ad 86. Tail 3 ^ ad 27, 27, 28 ; i < juv 23-5 ; i $ ad 31-5. Tarsus
3 <$ ad 19-1, 19-5, 20-3 ; i <$ juv 19-4 ; i $ ad 22-4. Culmen (dorsal) 3 ad 9-5, 10-0,
10-7 ; i c juv 10-7 ; i $ ad 11-5 ; (lateral) 3 ^ ad n-i, 11-7, 11-7 ; i g juv n-o ;
i $ ad 12-8. Linear measurements taken by Sutter.

7ns yellowish white. Bill blackish, base of mandible yellowish, gape bright yellow ;
whole of juv mandible and edge of maxilla dull yellow. Foot dull yellow, toes greyer.

Four other specimens of this race have been recorded : an adult female at the
AMNH (Mayr, 19386), an adult female at the Museum of Vertebrate Zoology,
Berkeley, California, and an adult male at the Chicago Museum (Beecher, 1945), and
a juvenile female at the U.S. National Museum (Baker, 1948 ; Sutter, 1955). Dr.
Sutter has kindly examined our specimens and provided very helpful notes, though
comparative material of the other eastern races was not available to him at the time.
All the specimens have the black markings well-developed on the underparts, but
none approaches the remarkable condition shown by the juvenile in the USNM
(Sutter, in. litt.}.

The female specimen is in poor condition on rump, back and throat, and adds little
to Mayr's description ; except that the belly is decidedly paler and the back apparent-
ly less grey than those of saturata. It differs from five adult females from the central
highlands of New Guinea (in the type-series of giluwensis Sims, 1954) in having the
plain rufous feathers of the mid-breast darker, contrasting more with the paler black-
barred feathers at the sides ; the belly darker than average, matching the most
rufous-tinged belly of giluwensis ; the head paler, with wider grey edges, and without
a distinct pale stripe ; the rufous collar well-developed (only one giluwensis has
concentrations of rufous near the wing-roots) ; and the upper back more rufous. It

LAND BIRDS OF GUADALCANAL 21

differs from three adult females from New Britain (including the holotype of saturata)
in having the belly and perhaps the throat paler ; the black-barred feathers of the
breast perhaps encroaching more towards the mid-line ; the rufous collar much
better-developed than in the one saturata which shows it (see Mayr, 19386) ; and the
back apparently with more rufous and more black. The head is similar to those of
two saturata which lack crown-stripes. This specimen differs from a female from
south-eastern New Guinea (holotype of horsbrughi) in having the belly much paler
(the holotype being unusual among horsbrughi in having the belly rufous see Mayr,
19386 : 2 ; Sutter, 1955 : in) and the throat possibly so ; more barring on the breast,
with the barred feathers more conspicuously pale ; the edges of the upper wing-
coverts much paler ; and the crown with broader grey edges and no central stripe.
It differs from two adult females from Cape York Peninsular (yorki) in having the
midbreast deeper rufous ; more barring on the sides of the breast ; the crown-stripe
less distinct ; the collar deeper rufous and apparently narrower ; and the mantle
apparently blacker, less grey.

The male has not hitherto been described. Sutter (in litt.} describes the three
adults as "in coloration and pattern somewhat intermediate between horsbrughi
and mayri of the Louisiades. Upper-parts in general appearance rather plain-coloured
(apart from the ochraceous edges on mantle and scapulars), principal colours black
and dark olive-grey with narrow rufous vermiculation or bars. The rufous bars are
distributed regularly over the whole upper back, without indication of a nuchal collar
(as in Australian males). Breast rather deep rufous, middle of belly whitish in strong
contrast to the breast collar colour (as in giluwensis). Sides of the breast and wing-
coverts heavily spotted." These specimens agree with two males of giluwensis in the
coloration of the underparts and in the crown-stripe, and differ in having the pale
edges of the crown-feathers wider and more rufous ; the back variable but averaging
greyer, less black, and more rufous towards the interscapulars ; and the edges of the
scapulars and upper wing-coverts deeper ochraceous. They differ from three adult
males of saturata in having the belly-patch whiter and larger ; the edges of the crown-
feathers more rufous, less grey ; the mantle more black, less grey, with considerably
more rufous anteriorly ; and the edges of the upper wing-coverts and scapulars
deeper ochraceous.

NOTES. Mayr (19496 : 2) combines T. maculosa specifically with T. sylvatica, but
they overlap on Mindanao (Sutter, 1955 : 137). Pendleton (1947) studied this race in
the field.

(M. 19) Rallus philippensis christophori Mayr, 1938

REFERENCES. Mayr, 19386 : 7.
RANGE. Guadalcanal and San Cristoval Groups.
SPECIMENS. San Cristoval : Goge i $ ad.
Weight 240. Wing 134. Bill 34.

Iris reddish-brown. Bill slate, mandible paler, base pinkish-brown. Foot pale
greyish-buff, joints greyer.

The specimen, caught in a native drop-trap, is in poor condition posteriorly.

22 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(F) Poliolimnas cinereus leucophrys (Gould, 1847)

REFERENCES. Mayr, 19496 : 17.

RANGE. Moluccas to northern Australia and Northern Melanesia (lowlands).

NOTES. Known in the Solomons from Bougainville (Baker, 1948) and Guadalcanal
(Donaghho, 1950) ; and from the Three Sisters (Cain & Galbraith, 1956 ; French,
1957) where it is common. No doubt, as French suggests, it will prove to be wide-
spread.

(M. 20) Amaurornis olivaceus ultimus Mayr, 1949

REFERENCES. Mayr, 19496 : 21.

RANGE. San Cristoval, Santa Ana, Gower (lowlands).

(M. 21) Nesoclopeus woodfordi woodfordi (Ogilvie-Grant, 1889)

REFERENCES. Mayr, 19496 : 15.
RANGE. Guadalcanal.

(M. 22) Edithornis silvestns Mayr, 1933

REFERENCES. Mayr, 19330 : i.
RANGE. San Cristoval (mountains).

NOTES. The species is well-known to the natives, and apparently not rare below
Nagasi, yet the holotype remains unique.

(M. 23) Porphyrio porphyrio samoensis Peale, 1848

REFERENCES. Mayr, 19496 : 22.

RANGE. Northern and Southern Melanesia and western Polynesia (individual and
irregular geographical variation) .

SPECIMENS. Guadalcanal : Betilonga 2 $ ad. San Cristoval : Goge i $ ad. i $
ad to OUM.

Weight San Cristoval i $ ad 495. Wing Guadalcanal 2 $ ad 222, 226 ; San Cristoval
i $ ad 203. Tail Guadalcanal i $ ad 86. Bill (from back of shield) Guadalcanal 2
ad 61-5, 64 ; San Cristoval i $ ad 57-5.

Iris orange-brown to red. Bill and shield deep orange-red. Foot flesh colour to deep
pink.

The San Cristoval specimen agrees with Mayr's in being small and black above,
but has a distinct breast-shield.

(M. 24) Ptilinopus superbus superbus (Temminck, 1810)

REFERENCES. Stresemann, 1914 : 45 ; 1923 : 72. See Cain, 19546 ; 1954^ ; 1957.

RANGE. Moluccas to eastern Australia and Solomons, except San Cristoval Group
(mainly lowlands).

SPECIMENS. Guadalcanal : Betilonga 3 ^ ad, 4 $ ad. i <$ ad, 2 $ ad to OUM.

Weight 4 <$ ad 111-5-133-5 (120-4) ; 5 $ ad 104-118 (113-4). Wing 3 <$ ad 124-129
(127-3) ; 4 $ ad 115-127 (121-5). Bill 3 ^ ad 18-5-19 (18-8) ; 3 ? ad 18-19 (18-7). '

LAND BIRDS OF GUADALCANAL 23

7ns yellow, sometimes greenish. Eyelid pale yellow or green, skin round eye
greenish blue. Bill sage green to fuscous green. Foot dull purplish red (" bright " in
i of 12).

A male (not preserved) was taken by collectors from Turipava, probably below
the mist-forest.

NOTES. It has only recently been pointed out (Cain & Galbraith, 1956 : 121)
that this widely-distributed species is apparently absent from the San Cristoval group.
There are no specimens from there in the AMNH (Amadon, in litt.} nor the BM(NH).
and natives of San Cristoval and of Ugi failed to recognize a coloured figure of this
distinctive lowland species (Mayr, 19450; pi. 2, fig. n). Ramsay's (18820 : 39) record
for Ugi cannot be accepted without further evidence.

(M. 25) Ptilinopus solomonensis ocularis Mayr, 1931

REFERENCES. Mayr, 19317 : 6. See Cain, 19546.

RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Betilonga 2 ^ ad, i ^ juv ; Turipava 2 <$ ad, 4 $ ad.
i ^ ad, i $ ad to OUM.

Weight 5 <? ad 112-136 (125-3) ', i c? juv 91 ; 5 $ ad 114-5-141-5 (127-1). Wing
4 $ ad 127-5-131 (129-9) ; I <J J uv I22 '5 i 3 ? ad 123-5-129 (125-8). Bill 2 ^ ad 19,
20 ; i <$ juv 20 ; 4 $ ad 19-5-21 (20-2).

Iris sage green (light sage green to dull green) ; juv pale yellow. Skin round eye
pale green to bluish. Bill sage green to greenish-grey. Foot dull purplish red (" dull
pink " in i) ; juv dull reddish brown and grey.

NOTES. Iris colours were not given for the three high-altitude races in the Solo-
mons (bistictus, ocularis and ambiguus] at their original description (Mayr, 1931).
Labels of the AMNH series of ocularis show some variation in iris colour (P. Vaurie,
in litt.}, but it is not clear whether the iris is ever yellow. We have the labels of ten
adults, all showing it as some shade of dull green. This is very different from the
warm yellow, with a narrow green inner ring, of solomonensis. BM(NH) specimens
of lowland races have the irides indicated as bright yellow to red (solomonensis),
yellow (neumanni], pale greenish yellow to dull cream yellow (meyeri] and pale
greenish yellow (johannis] . Careful study of the AMNH collections will be necessary
to determine the geographical variation of iris colour in this species

(M. 25) Ptilinopus solomonensis solomonensis Gray, 1870

REFERENCES. Mayr, 19317 : 6. See Cain, 19546.

RANGE. San Cristoval Group.

SPECIMENS. San Cristoval : Goge 39 $ ad, 27 ? ad, i ? juv ; Nagasi 5 <$ ad,
i ? ad, i ? juv. Ugi i g ad, i ? ad, i ? juv. 3 $ ad, 3 ? ad to AMNH ; 3 <? ad, 2 ? ad
to OUM.

Weight Goge 30 $ ad 80-101-5 (93-1) ; 22 <j> ad 83-5-105 (93-7) ; Nagasi 6 ad
88-5-106-5 (98-1) ; i $ ad 97-5 ; Ugi i ^ ad 101 ; i $ ad 86-5. Wing Goge 39 ^ ad
114-123 (119-5) ; 27 $ ad 111-5-123 (116-8) ; Nagasi 5 $ ad 116-124 (120-3) ; i $
ad 120 ; Ugi i $ ad 127 ; i <j> ad 119. Bill 40 $ ad 18-21 (19-3) ; 28 ad 16-5-20
(19-1).

24 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

7ns orange-yellow (pale pinkish yellow to dull orange), inner ring pale green ; juv
dark grey-brown to dirty yellow, inner ring greenish. Skin round eye pale green to
pale blue ; juv blackish to blue-green, eyelid paler. Bill sage-green, sometimes
greyer and bluer ; juv dark, fuscous-green to blue-grey. Foot dull mauve, shanks
redder ; juv pale pinkish brown to pale grey.

In this, and the other long series of Ptilinopus, the distribution of weights is
markedly bimodal. However, this is not paralleled in other measurements or characters,
and no first-year or retarded plumages (as described by Mayr, I93i/ : 7, for ambiguus]
can be discerned. The juvenile from Nagasi, presumably a male, has pink forehead
feathers which evidently belong to the juvenile plumage.

The series from Goge is extremely variable in depth of coloration. This is evidently
due to variation in carotenoid pigmentation (see Volker, 1953 ; Auber, 1957), the
yellow, green and lilac areas being equally affected. The variation does not seem to
be due to bleaching, whether in life or post mortem. It is most conspicuous in the green
of the back, but there interacts with variation in structural coloration (blue and
bronzing). Table I shows the distribution of specimens among seven colour-classes
(A palest to G deepest) for the males, judged from the yellow bib ; and among 6
classes (H to N) for the females, judged from the yellow under tail-coverts. BM(NH)
specimens of solomonensis are included in the series for the coast of San Cristoval and
Ugi. Other specimens, of the races neumanni, meyeri, and johannis, are lumped
together, since they are too few to indicate significant differences among themselves
in variability or average depth of pigmentation.

TABLE I. Intensity of yellow carotenoid pigment in adults of Ptilinopus solomonensis.

Classes A-G males, judged on bibs ; classes H-N females, judged on under tail-
coverts : classes A and H palest, G and N deepest.

ABODE FG HJKLMN
P.s. solomonensis:

Nagasi . 14 I

Goge . . .1433 14 56 123 13 71

San Cristoval coast . i i i

Ugi ... 21 i

P. s. ocularis . . 3 i i i

Other races . . 342 122

NOTES. The subspecific assignment of the population on the Three Sisters must
be tentative, since it is represented in the BM(NH) by a single juvenile only.

(M. 25) Ptilinopus solomonensis cf. solomonensis

SPECIMENS. Guadalcanal : Betilonga i $ ad.

Weight 97-5. Wing 121. Bill 19-5.

7m yellowish orange. Skin round eye pale blue-green. Bill greyish green. Foot
dull lilac, shanks dull purplish red.

This specimen agrees with the lowland races, and disagrees with ocularis, in its
smaller dimensions, yellow iris, and complete lilac cap. It is richly-pigmented (group
F of Table I), and has the scapular spots slightly larger and the cap and belly perhaps

LAND BIRDS OF GAUDALCANAL 25

slightly richer lilac than in most males of solomonensis . It could belong to vulcanorum
of the New Georgia Group (P. Vaurie, in litt.}.

NOTES. Although Amadon does not rule out the possibility of this being a mutant
of ocularis (P. Vaurie, in litt.}, the concordance of three characters makes it much more
probable that it is a straggler from another island. Its provenance must remain uncer-
tain until the solom&nensis-like populations on small islands in the Solomons are
better known (see Mayr, I93i/: 8). Geographically, Buena Vista near the Florida
group seems the most probable source.

Errant individuals are almost unknown among the strongly-localized races of
Northern Melanesia. Their occasional occurrence is necessary to the hypothesis that
character-progressions in insular areas may be due to gene-flow (e.g. Galbraith, 1956:
160-165) > but they must remain rare if the geographical variation is not to be swamped.
The possibility should not be overlooked, that striking character-differences between
neighbouring insular populations may sometimes be of selective importance in reducing
introgression from such errant individuals. This would help to explain why very
isolated populations tend to lose their distinctive patterns while relatives in archi-
pelagos, not sympatric with closely-related species, retain them.

(M. 26) Ptilinopus viridis lewisi Ramsay, 1882

REFERENCES. Cat. Birds B.M. 21 : 153. See Cain, 1954^.
RANGE. Lihir, Solomons except San Cristoval Group.
SPECIMENS. Guadalcanal : Betilonga i <$ ad. To OUM.
Weight 112. Wing 116. Bill 17-5.
Iris yellow. Bill yellow. Foot dull purplish red.

(M. 26) Ptilinopus viridis eugeniae (Gould, 1857)

REFERENCES. Cat. Birds B.M. 21 : 153. See Cain, 1954^.

RANGE. San Cristoval, Ugi, Three Sisters.

SPECIMENS. San Cristoval : Goge 10 ^ ad, i $ juv, n $ ad ; Nagasi i <$ ad. Ugi
2 $ ad, 2 $ ad, i $ juv. i ^ ad, i $ ad to OUM.

Weight 13 < ad 93-5-129-5 (113-8) ; i $ juv 66 ; 12 $ ad 101-5-127-5 (113-2) ; i ?
juv 89. Wing 12 < ad 114-128 (122-6) ; i $ juv 113-5 ', 12 $ ad 116-124-5 (120-2) ;
i $ juv 115-5. Bill 13 <J ad 17-19 (18-2) ; i ^ juv 18 ; 12 $ ad 17-18-5 (18-1) ; i $
juv 17-5.

7ns tawny orange to tawny red (" deep scarlet " in 2 of 3 from Ugi), inner ring
(when noted) pale yellow to dull green ; juv dull brownish, inner ring greenish. Skin
round eye bluish grey, eyelid usually yellower. Bill yellowish, usually greenish
(sometimes golden), base and cere dull maroon ; juv dirty yellowish-green, yellower
distally, cere once dull maroon. Foot mulberry-red, sometimes bluer ; juv dull
(buffy pink to greyish purple) .

The maroon-based bill of eugeniae has been overlooked as a further character of
this very distinct race. In the dried skin the maroon turns blackish. Apart from
eugeniae, only geelvinkianus shows slight blackish traces in the dried bill. Of BM(NH)

26 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

specimens, only one of vicinus and six of nominate viridis have the colour in life
indicated. In these the base is indicated as redder (scarlet to orange) than the tip
(rich orange to greenish-yellow).

(M. 27) Ptilinopus richardsi richardsi (Ramsay, 1882)

REFERENCES. Ripley & Birckhead, 1942 : 8. See Cain, 19546 ; 1957.

RANGE. Ugi, Three Sisters, Santa Ana.

SPECIMENS. Ugi 15 ^ ad, 2 ^ juv, 6 $ ad. I $ ad, I $ ad to OUM.

Weight 15 $ ad 94-117 (108-6) ; 2 $ juv 88, 92-5 ; 6 $ ad 94-5-109-5 (99-1).
Wing 15 $ ad 128-5-134 (131-5) ; 2 $ juv 118, 122 ; 6 $ ad 122-129 (125-9). Bill 14 o
ad 19-20 (19-5) ; 2 <$ juv 19, 20 ; 6 $ ad 18-5-20 (19-2).

Iris orange to orange-red, inner ring pale yellow (once golden-yellow) to pale green.
Skin round eye pale blue-grey, eyelid pale yellow. Bill dark sage green basally, paler
yellow-green distally, cere dull purple. Foot deep purple (bluish to reddish).

NOTES. Cain (1957) dismisses the character " malar apex concolorous with crown "
(Ripley & Birckhead, 1942 : 8) as a mere form of words when applied to this grey-
crowned species. However, the malar apex and the crown are pure grey, while the
face and neck are washed with yellow. In individuals whose greys are exceptionally
deep, the malar apex contrasts sharply with the paler and yellower face. It seems
probable that the ancestral population did have the apex brightly coloured, like the
crown though this still further emphasizes the polyphyletic origin of the character.

(M. 29) Ducula pacifica pacifica (Gmelin, 1789)

REFERENCES. Amadon, 19430 : 10. See Goodwin, 1960.

RANGE. Small islands from Louisiades and Solomons to Polynesia.

NOTES. In the Solomons, the species is known from Buena Vista, Ramos, Gower,
Sikaiana, Ongtong Java, Rennell (Amadon, 19430), Bellona (Mayr, 19316) and the
Three Sisters (Cain & Galbraith, 1956 : 106 ; French, 1957). Only on the latter does
it overlap with its close relative, D. rubricem.

(M. 30) Ducula rubricera rufigula (Salvadori, 1878)

REFERENCES. Cat. Birds. B.M. 21 : 179. See Goodwin, 1960.

RANGE. Solomons.

SPECIMENS. Guadalcanal : Betilonga 2 <$ ad. San Cristoval : Goge i ? juv. i
ad to OUM.

Wing 2 $ ad 240, 253. Bill 2 ^ ad 33, 33-5.

Iris deep red or orange-red ; juv dark brown. Skin round eye blue-grey. Bill dark
grey, base and cere cherry-red ; juv slaty-grey distally, grading to pale greyish-fawn
base and cere. Foot purplish red (cherry-red in i of 5 ad), claws dark grey ; juv fawn,
fuscous-tinged, claws grey.

Though small, the juvenile is as brightly-coloured as adults,

LAND BIRDS OF GUADALCANAL 27

(M. 31) Ducula pistrinaria pistrinaria Bonaparte, 1855

REFERENCES. Hartert, 19260 : 34. See Goodwin, 1960.

RANGE. Feni, Nissan, Solomons (coasts).

SPECIMENS. Ugi i <$ imm. To OUM.

Weight 390. Wing 220. Tail 130. Bill 29-5.

7ns dark brown. Skin round eye pale grey, eyelid reddish brown with blackish
wash. Bill dark slate. Foot pinkish brown.

The specimen has faint juvenile barring on the breast, and is somewhat smaller
than adults, with narrower rectrices. Oiliness of the plumage conceals any colour-
difference.

(M. 32) Ducula brenchleyi (Gray, 1870)

REFERENCES. Cat Birds B.M. 21 : 225. See Goodwin, 1960.

RANGE. Eastern Solomons Guadalcanal, Malaita, Ulawa, San Cristoval, Ugi,
Three Sisters (mainly coasts ?).

SPECIMENS. Ugi i $ imm. To OUM.

Weight 380. Wing 210. Tail 105. Bill 31.

Iris reddish brown, inner ring yellowish brown. Skin round eye grey, eyelid
greyish purple. Bill dark fuscous grey, tip black. Foot deep dull purplish-brown.

Like two juveniles from the Three Sisters, the specimen differs from adults in
somewhat smaller size and narrower rectrices ; underparts less purplish (more rufous)
brown, breast greyer and with pale barring, sides of breast grey not purplish, throat
yellowish rather than pink.

NOTES. Though considered by Mayr (19450:) to be a forest bird, mainly of higher
altitudes, this species is common on the coast of San Cristoval, especially in isolated
trees, and also on the small islands of Ulawa, Ugi and the Three Sisters (Cain &
Galbraith, 1956 : 123 ; French, 1957).

(M. 33) Gymnophaps solomonensis Mayr, 1931

REFERENCES. Mayr, i93i/: n.

RANGE. Mountainous islands of the Solomons Bougainville, Kolombangara,
Vangunu, Guadalcanal, Malaita.

SPECIMENS. Guadalcanal : Betilonga i ^ ad, 2 ? ad. i $ ad to OUM.

Weight i (J ad 385 ; 2 $ ad 310, 310. Wing i $ ad 226 ; 2 $ ad 221, 222. Bill
i $ ad 29 ; 2 $ ad 25, 25-5.

7ns orange-red (i <) or pinkish-orange (3 $) ; juv orange-red. Skin round eye,
and eyelid, red (i <) or reddish purple (2 $) ; juv blue-grey. Bill pale yellow, pinkish
terminal or subterminal patch ; juv fuscous, tip pale. Foot lilac, pinkish in dried
skin (i c), or dull purplish-blue to dull lavender-grey, greyish in skins (3 $), claws
fuscous ; juv dark grey.

(M. 34) Columba vitiensis halmaheira (Bonaparte, 1855)
REFERENCES. Amadon, 19430: : 14. See Goodwin, 1959.
RANGE. Moluccas, New Guinea region, Solomons.

28 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

SPECIMENS. Guadalcanal : Turipava 2 <$ ad. San Cristoval : Goge I <$ ad, I $ ad.
i $ ad to OUM.

Weight i $ ad 490 ; i $ ad 395. Wing 3 $ ad 219-229 (224-3) ; i $ ad 212. Bill
3 <? ad 29-5-31 (30-2) ; i $ ad 31.

7m orange (yellowish to pinkish), inner ring brownish or greenish. Skin round eye
maroon, more or less mixed with pale grey, eyelid dull reddish (brownish red to bluish
purple). Bill ivory, base and cere dull reddish (pinkish red to bluish purple). Foot
pinkish red, with back of shank bluer, to dull bluish purple, claws ivory.

(M. 35) Columba pallidiceps (Ramsay, 1877)

REFERENCES. Cat. Birds B.M. 2 1 : 319 (pallidiceps + philippana) ; Mayr, 1934^ : 4.
See Goodwin, 1959.

RANGE. Bismarcks, Solomons.

NOTES. Rare or retiring in the Solomons, but known from Guadalcanal and San
Cristoval, and recorded by Ramsay (18820; : 722) from Ugi.

(M. 36) Macropygia mackinlayi arossi Tristram, 1879

REFERENCES. Cat. Birds B.M. 21 : 363 (rufocastanea] ; Mayr, 19317 : 12.

RANGE. Nissan, Solomons, Rennell.

SPECIMENS. Guadalcanal : Tenaru i $ ad ; Betilonga i $ juv ; Turipava i
juv, i $ ad. San Cristoval : Goge 2 $ ad. i $ ad to OUM.

Weight i <$ ad 79 ; i ^ juv 91-5 ; 6 $ ad 79-88 (84-8) ; i $ juv 75. Wing i <$ juv
138 ; 3 $ ad 137-5-142 (140-5) ; i $ juv 134-5. Bill i 3 juv 18 ; 4 ? ad 17-18 (17-7) ;
i $ juv 19-5.

Iris scarlet, sometimes irregularly mottled with yellow, inner ring yellow ; juv
dull brown or pale brown. Skin round eye dark grey, eyelid grey or red. Bill black.
Foot red (bright or dull, pinkish orange-red to crimson) ; juv fawn, toes fuscous.

The juveniles differ from adults in iris and foot colour (left and right irides seem to
redden independently), and in more pointed rectrices, laxer plumage, and more
black at the bases of the contour-feathers (yielding a scalloped pattern on the mantle).

(M. 37) Reinwardtoena crassirostris (Gould, 1856)

REFERENCES. Cat. Birds B.M. 21 : 368 ; Mayr, 19440 : i.

RANGE. Solomons.

SPECIMENS. Guadalcanal : Betilonga i 3 ad. To OUM.

Weight 2 $ ad 258, 260. Wing 208-5. Bill 26.

Iris yellow. Skin round eye deep red. Bill red basally, tip more orange. Foot red
(purplish in i of 2).

NOTES. In discussing the characters which separate this species from the repre-
sentative R. browni and R. reinwardtsi, Mayr (19440) inadvertently omits the
most trenchant, the crest, which the others totally lack.

The species is known primarily from the larger islands of the Solomons, but also
from Gatukai, Rendova, Ugi (Ramsay, 18820 : 36) and the Three Sisters. On the
latter it rarely breeds, but is sometimes common (French, 1957).

LAND BIRDS OF GUADALCANAL 29

(M. 38) Chalcophaps stephani mortoni Ramsay, 1881

REFERENCES. Rothschild & Hartert, 19016 : 189 ; Mayr, 1934^ : 6.

RANGE. Nissan, Solomons (mainly lowlands).

SPECIMENS. Guadalcanal : Betilonga i ^ juv. San Cristoval : Goge 3 < ad.
i c? ad to OUM.

Weight 3 <$ ad 133-5-150 (143-8) ; * 3 J uv 87 ; i ? ad 132. Wing 3 $ ad 148-154
(150-0) ; i $ juv no. Bill 3 <$ ad 23-5-24 (23-8) ; i <$ juv 22-5.

7ns dark brown ; juv brown. Skin round eye brownish grey to dull purple, eyelid
dull purple. Bill deep purple basally, cere blackish ; distal 2/3 deep orange with
bright yellow tip (3 $}, or dull yellow separated from purple by blackish cross-streak
(i ?) ; juv fuscous-brown, edges dull yellow. Foot cherry-red, toes more purplish ;
juv dull pale purplish.

(M. 39) Gallicolumba beccarii solomonensis (Ogilvie-Grant, 1888)

REFERENCES. Cat. Birds B.M. 21 : 594 (granti) ; Mayr, 19316 : 12. See Hachisuka,
1931 ; Amadon, 1943^.

RANGE. Guadalcanal, San Cristoval Group, Gower, Rennell (lowlands).

SPECIMENS. San Cristoval : Goge 2 <$ ad, 2 ? juv ; Nagasii^ad. i ^ ad to OUM.

Weight 3 $ ad 93-98 (94-8). Wing 3 <$ ad 105-106-5 (105-7). Bill 2 <$ ad 18, 19.

Iris very dark brown. Skin round eye bluish white, eyelid dull yellow ; smaller
juv dark grey round eye and under throat, wing and belly. Bill and cere black ;
larger juv with yellowish tip, smaller fuscous, whitish distally with dark subterminal
band and white tip (flexible in dried skin). Foot purplish red, toes bluer ; larger juv
pinkish brown with somewhat fuscous toes, smaller pinkish fuscous-grey with back
of shank flesh-colour.

(M. 40) Gallicolumba jobiensis cf . chalconota Mayr, 1935

REFERENCES. Mayr, 1936 : 2,

RANGE. Vella Lavella ; Guadalcanal (subspecies ?).

NOTES. Only four specimens are known from the Solomons : the adult male
holotype from Vella Lavella, and three juveniles from Guadalcanal (Cat. Birds B.M.
21 : 599, footnote) one of which is in the BM(NH).

(M. 41) Gallicolumba salamonis (Ramsay, 1882)

REFERENCES. Ramsay, 1882^ : 299.
RANGE. San Cristoval, Ramos.

NOTES. Only two specimens are known : the unsexed adult holotype from San
Cristoval, and one from Ramos (Mayr, 19450).

(M. 42) Caloenas nicobarica nicobarica (Linnaeus, 1758)

REFERENCES. Mayr, 19316 : 12.

RANGE. Malaysia to New Guinea region and Northern Melanesia.

SPECIMENS. Guadalcanal : above Tsea i $ ad. To OUM.

Wing 244. Tail 82. Bill 31.

7ns salmon pink. Bill and cere matt black. Foot purplish red, sole yellow.

30 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(M. 44) Eos cardinalis (Gray, 1849)

REFERENCES. Cat. Birds B.M. 20 : 22 ; Hartert, 19260: : 38 ; Auber, 1938 : 698.
See Peters, 1935.

RANGE. Duke of York, Lavongai, Tabar, Lihir, Tanga, Feni, Nissan, Solomons
(mainly lowlands).

SPECIMENS. Guadalcanal : Tenaru 9 <$ ad, 3 ^ juv, 5 $ ad, 4 $ juv, 4 ? juv ;
Betilonga I $ ad. i ^ ad, I $ ad to OUM.

Weight 9 ad 173-205 (184-9) > 4 3 Juv 124-160 (145-4) ; 5 ? ad 168-5-195 (178-8) ;
4 $ juv 142-5-155 (148-4). Wing 4 <? ad 177-183 (180-7) ; 3 c? juv 166-173 (168-7) ;

4 $ ad 172-5-180 (175-4) i 4 ? Juv 165-171 (169-3). Tail i ad 154 ; i <? juv 116-5 ;
i $ ad 150 ; i $ juv 144. Bill (from cere) 10 $ ad 20-22 (21-2) ; 4 $ juv 19-5-21 (20-1) ;

5 $ ad 20-21 (20-5) ; 4 $ juv 19-5-20 (19-6).

Iris orange-red to pinkish orange, very narrow black and yellow inner rings ;
juv dull yellow. Skin round eye and under throat black, sharply-marked semi-circular
yellow patch at base of mandible ; juv blackish or fuscous, irregularly mottled with
more or less yellow. Bill reddish-orange, tip yellower, cere and base of maxilla black ;
juv dull orange more or less suffused or mottled with blackish, especially towards
cere. Foot black, ad sometimes with irregular patches of orange- to whitish-yellow.

The juveniles differ from adults in the coloration of their soft parts, especially that
of the bare throat ; and in more pointed rectrices, more uniformly-coloured upper-
parts, ventral fringes yellower less white, and vaguer and whiter violet-blue areas at
wing-bends and under wing-coverts. One BM(NH) juvenile from Guadalcanal has a
much shorter tail, very wide yellow fringes ventrally, and conspicuously orange
thighs.

The outer edges of the remiges and rectrices are subject to bleaching during life,
turning from bronzy red to yellowish olive.

NOTES. The generic placement of this species is obscure. Auber's (1938) intensive
study associates it with the atra-scintillata-duivenbodei group of New Guinea, which
Peters (1935) separates as Chalcopsitta, while Verheyen (1956) creates for it a mono-
typic genus Cardeos. Whereas Mayr (1941^) separates Chalcopsitta and Pseudeos from
Eos, he later (1945^) includes cardinalis with the latter, thus giving to the unwary a
false impression of the scope and range of the genus. His intention is clearly to adopt
a broader generic concept for the Loriinae (as by following Amadon, 19426 in uniting
Charmosyna with Vini], and to return all these forms to one genus. This seems the
most acceptable solution until the subfamily has been critically revised.

(M. 45) Trichoglossus haematodus massena Bonaparte, 1854

REFERENCES. Rothschild & Hartert, 19010 : 70 ; Cain, 1955 : 438, 445.

RANGE. Bismarcks to New Hebrides (lowlands).

SPECIMENS. Guadalcanal : Tenaru 14 ^ ad, i ^ juv, 9 $ ad, i $ juv. San Cristo-
val : Goge i ^ ad, 2 $ ad. i ^ ad, i $ ad to OUM.

Weight Guadalcanal 13 <$ ad 83-103-5 (92-3) ; i ^ juv 75-5 ; 9 $ ad 74'5-95'5
(84-9) ; i $ juv 82 ; San Cristoval i ^ ad 117-5 ; 2 $ ad 97-5, 103. Wing Guadalcanal
10 $ ad 125-134 (130-9) ; i <$ juv 130-5 ; 7 $ ad 122-132 (128-8) ; San Cristoval

LAND BIRDS OF GUADALCANAL 31

I $ ad 145 ; 2 $ ad 132, 135-5. Tail Guadalcanal 8 $ ad 96-5-104-5 (100-3) ', * c?
juv 101-5 ; 5 $ ad 93-101 (97-1) ; i $ juv 86-5 ; San Cristoval i ^ ad 101-5. Bill
(from cere) 14 ad 18-20-5 (19-2) ; i $ juv 18 ; 10 $ ad 17-19-5 (18-4) ; i $ juv

17-5-

Iris orange-red (sometimes pinkish or dull, occasionally yellower or more scarlet) ;
juv greyish yellow to pale dull brown. Skin round eye dark fuscous grey to black,
throat whitish to pale flesh-colour with yellow tinge. Bill orange-red, tip yellower,
cere dark fuscous grey to black ; juv dull orange streaked with blackish, especially
distally. Foot fuscous-grey, more or less greenish or occasionally bluish.

The juveniles differ from adults in the colour of iris and bill (which in the dried
skin is whiter basally), more pointed rectrices, and slight differences in plumage
coloration less blue and orange on the crown, no concealed red at the bases of
feathers on the mid-back, more gradual transition from red breast to green belly,
more evenly yellow-green thighs and under tail-coverts.

(M. 46) Lorius chlorocercus Gould, 1856

REFERENCES. Cat. Birds B.M. 20 : 38.

RANGE. Eastern Solomons Guadalcanal, Savo, Malaita, Ulawa, San Cristoval,
Ugi and Rennell.

SPECIMENS. Guadalcanal : Betilonga 5 ^ ad, 2 $ ad ; Turipava i < ad. San
Cristoval : Goge i $ ad ; Nagasi i <$ ad. Ugi 4 <$ ad. i ^ ad, i $ ad to OUM.

Weight Guadalcanal & San Cristoval 7 ad 168-185 ( I 76-i) ; 4 $ ad 153-174-5
(160-6) ; Ugi 4 $ ad 190-225 (206-3). Wing Guadalcanal & San Cristoval 7 $
ad 160-174 (166-7) i 3 ? a d 159-164 (161-0) ; Ugi 3 ad 176-181 (178-7). Tail
Guadalcanal & San Cristoval 4 < ad 89-5 -101-5 (95'4) Ugi 3 ad 95-100 (98-2) ;
Bill (from cere) Guadalcanal & San Cristoval 6 ad 21-5-24 (22-7) ; 3 $ ad
21-22 (21-5) ; Ugi 4 <$ ad 23-24-5 (22-1).

Iris dull orange to orange-red (Guadalcanal), pinkish-orange (San Cristoval and
Ugi), inner ring yellow. Skin round eye black, throat white. Bill orange to reddish-
orange, base of maxilla blackish, cere black. Foot dark grey, scutes black.

(M. 47) Vini meeki (Rothschild & Hartert, 1901)

REFERENCES. Rothschild & Hartert, 19016 : 187 ; Amadon, 19426 : 2.

RANGE. Higher islands of the Solomons Bougainville, Kolombangara, Guadal-
canal, Malaita (mainly mountains).

SPECIMENS. Guadalcanal : Betilonga 3 ^ ad, 3 ? ad ; Turipava i ad, 2 $ ad.
i c? ad to OUM.

Weight 4 ad 21-25-5 (24-1) ; 5 $ ad 21-5-25-5 (23-5). Wing 4 ^ ad 78-81 (81-2) ;
5 $ ad 78-82 (80-4). Tail 3 c? ad 73-78-5 (75-0) ; 4 $ ad 64-5-70-5 (68-7). Bill (from
cere) 4 < ad 11-5-12 (11-7) ; 3 $ ad 10-5-11-5 (11-2).

7ns yellow to reddish orange. Skin round eye yellowish fuscous. Bill reddish
orange, cere sometimes duller, mandible yellower, tip and edges blackish. Foot pale
to reddish orange, claws fuscous.

32 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(M. 49) Vini margarethae (Tristram, 1879)

REFERENCES. Cat. Birds B.M. 20 : 81.

RANGE. Solomons.

SPECIMENS. Guadalcanal : Tenaru i <$ ad ; Betilonga 3 <$ ad, 2 $ ad. San Cristo-
val : Goge i ^ ad ; Nagasi i $ ad. i <$ ad to OUM.

Wwgto 6 c? ad 52-5-59-5 (557) ; 2 $ ad 43, 44-5. Wwg 5 $ ad 106-5-115 (1117) ;
2 $ ad 108-5, II2 - TVwJ 2 c? ad 95> 103 ; 2 $ ad 85-5, 101-5. *'# (from cere) 5 ^ ad
14-5-15 (14-7) ; 2 $ ad 13-5, 14-5.

Iris orange, occasionally paler or duller. Skin round eye pale yellowish fuscous to
black. Bill and cere orange, tip blackish. Foot orange, claws fuscous.

(M. 50) Micropsitta finschii aolae (Ogilvie-Grant, it

REFERENCES. Hartert, 1924^ : 202.

RANGE. Guadalcanal, Florida and Russell Is., Malaita.

SPECIMENS. Guadalcanal : Betilonga 5 $ ad, 2 $ juv, i $ ad, i ? juv. i c^-ad to
OUM.

Weight 5 # ad 15-17-5 (15-9) ; 2 <$ juv 14, 15-5 ; i $ ad 14-5. Wing 3 3 ad 62-66
(63-8) ; 2 <$ juv 62, 63-5 ; i $ ad 59. Tail 4 <? ad 27-5-34 (31-4) ; 2 J juv 30, 32 ;
i $ ad 26. Bill (from cere) 5 ^ ad 7-5-8 (7-8) ; i $ ad 7.

Iris <$ orange ; $ and I juv dull yellow. Skin round eye grey. Bill <$ blackish,
whitish stripe under mandible, cere dull purplish red ; $ mottled horn, tip and base of
maxilla darker, cere whitish ; juv as $ but cere pinkish or grey. Foot pale bluish grey.

(M. 50) Micropsitta finschii finschii (Ramsay, 1881)

REFERENCES. Hartert, 19240; : 202.

RANGE. San Cristoval, Ugi, Rennell.

SPECIMENS. San Cristoval : Manewiriwiri 2 ^ ad ; Goge 3 $ ad ; Nagasi 2 <$ ad,
i $ ad. Ugi 3 3 ad, i $ ad.

Weight 8 $ ad 15-5-17-5 (16-6) ; 5 $ ad 14-18 (16-1). Wing 6 ^ ad 62-67-5 (65-2) ;
5 $ ad 59-65-5 (63-0). Tail 6 ^ ad 3O'5-35'5 (33'3) ', 5 ? ad 27-5-32-5 (30-6). Bitt
(from cere) 5 ^ ad 8-8-5 (8-3) ; 5 ? ad 7-8-5 (7-9).

Iris very dark brown (i dull yellowish brown). Skin round eye blackish to pale
grey. Bill <$ slate, whitish stripe under mandible, cere pinkish grey to dull mulberry-
red ; $ mottled bluish grey to horn, tip and sides of maxilla darker, cere dark grey
(i pinkish flesh) . Foot pale bluish grey.

Though conspicuous in the intact males, the red belly-patch was lost or reduced
during skinning.

(M. 51) Micropsitta bruijnii rosea Mayr, 1940

REFERENCES. Mayr, 19406 : 2.
RANGE. Kolombangara, Guadalcanal (mountains).
SPECIMENS. Guadalcanal : Turipava i ^ ad.
Weight 14-5. Wing 65-5. Tail 28. Bill (from cere) 6-5.

Iris dark brown. Skin round eye fuscous. Bitt grey, tip of maxilla and stripe under
mandible white, tip of mandible yellowish. Foot grey.

LAND BIRDS OF GUADALCANAL 33

(M. 52) Cacatua ducorpsii (Pucheran, 1853)

REFERENCES. Cat. Birds B.M. 20 : 129. See Vane, 1959.

RANGE. Solomons except San Cristoval Group.

SPECIMENS. Guadalcanal : Betilonga 3 < ad. i $ ad to OUM.

Weight 3 $ ad 355-460 (408). Wing 3 <? ad 258-274 (264). Tail 2 ^ ad 120, 121.
Bill (from cere) 3 ^ ad 32'5-34'5 (337)-

Iris dark grey-brown. Skin round eye bright blue, throat grey to dull purplish.
Bill pale horn, bluish at base. Foot black, greyish between scutes.

(M. 53) Larius roratus solomonensis (Rothschild & Hartert, 1901)

REFERENCES. Rothschild & Hartert, 19010 : 82.

RANGE. Solomons.

SPECIMENS. Guadalcanal : Betilonga i $ ad, i J juv, i $ ad, i juv. Ugi i $
ad. i c? ad, i $ ad to OUM.

Weight 2 <$ ad 370, 375 ; i^ juv 425; i ? ad 355 ; i? juv 370. Wing 2 $ ad 245,
248 ; i $ juv 230 ; i ad 229 ; i $ juv 227. Tail 2 $ ad 109, 110-5. #*7/ (from cere)
2 (? ad 39, 40-5 ; i ^ juv 37 ; i $ ad 36 ; i $ juv 34.

7ns cJ orange ; $ yellow ; <$ juv dull olive, $ juv brownish grey. Skin of throat
straw-coloured. Bill <$ maxilla orange, distal 1/3 golden-yellow, mandible black ;
$ black ; juv black, tip of maxilla yellow or horn. Foot fuscous to dull black.

(M. 54) Geoffroyus heteroclitus heteroclitus (Hombron & Jacquinot, 1841)

REFERENCES. Cat. Birds B.M. 20 : 412 ; Mayr, 19316 : 13.
RANGE. Bismarcks, Solomons (lowlands).
SPECIMENS. San Cristoval : Goge 2 ? juv.

Iris very pale yellow, dull green inner ring. Skin round eye pale grey, eyelid dull
yellow. Bill dull greyish horn, cere dull green. Foot dull greyish green.

(M. 55) Cuculus saturatus subspecies

RANGE. Eastern Palaearctic and Oriental regions, wintering in Malaysia and
western Australasia.

NOTES. Six specimens have been recorded from the Solomons, under various
specific names : four from New Georgia (Rothschild & Hartert, 1905 : 258) and one
each from Malaita (Mayr, 19317 : 13) and Ugi (Ramsay, i882c : 21).

(M. 56) Cacomantis variolosus addendus Rothschild & Hartert, 1901

REFERENCES. Amadon, 19426 : 20.

RANGE. Solomons.

SPECIMENS. Guadalcanal : Tenaru 2 ^ ad ; Betilonga 4 ^ ad, 2 J juv, 3 $ ad ;
Turipava 2 ad. San Cristoval : Goge 2 $ ad. i $ ad. to OUM.

Weight Guadalcanal 9 $ ad 38-45*5 (42-4) > 2 3 J uv 3, 39 i 3 $ ad 38-42-5
(40-0) ; San Cristoval 2 <$ ad 35-5, 38. Wing Guadalcanal 7 $ ad 121-131 (124-4) ;
2 <$ juv 113-5, IX 5 '> 3 ? a( i 110-5-112 (111-2) ; San Cristoval 2 J ad 116, 121. Tail

ZOOL 9. I. 3

34 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

Guadalcanal 2 $ ad 139, 139 ; 2 <$ juv 106, 124-5 ; i $ ad 141 ; San Cristoval i <$ ad
141. Bill Guadalcanal 8 <$ ad 19-20 (19-5) ; 2 ^ juv 17, 18-5 ; 2 $ ad 19, 20 ; San
Cristoval 2 <$ ad 19, 20.

Iris dark reddish brown, wholly or partly ringed with white. Eyelid yellow (bright
to dull and greenish). Bill maxilla black, mandible horn-colour more or less mottled
with fuscous and yellowish, gape yellow to pale grey externally, orange-red internally.
Foot yellow (orange-yellow to pale and dull) more or less tinged with greyish or
fuscous, claws fuscous or blackish.

One female, with much enlarged ovarian follicles, has the throat to the upper belly
barred with dark subterminal chevrons. This is clearly a somewhat retarded plumage,
seen also in a short-tailed subadult female in the BM(NH) (recorded by Ogilvie-Grant,
1888 : 191 as tymbonomus), and in a female from Bougainville, probably immature,
in the AMNH (P. Vaurie, in litt.}.

(M. 57) Cacomantis pyrrhophanus pyrrhophanus (Viellot, 1817)

REFERENCES. Amadon, 19426 : 16.

RANGE. New Caledonia and Loyalty Is.; partial migrant, occasionally wintering
in Solomons.

SPECIMENS. Guadalcanal : Tenaru i ^ ad.

Weight 56-5. Wing 140. Tail 148-5. Bill 22-5.

7ns bright red-brown. Eyelid dull yellow. Bill maxilla black, mandible grey
mottled with darker grey and yellow, gape pale flesh-colour externally, pinkish orange
internally. Foot dull yellow tinged with fuscous, claws dark fuscous.

NOTES. Four other specimens are recorded from the Solomons : three from Ysabel
(Rothschild & Hartert, 1902 : 587) and one from Bellona (Mayr, 19316 : 14).

(M. 58) Chrysococcyx lucidus lucidus (Gmelin, 1788)

REFERENCES. Mayr, 19326 : 2.

RANGE. New Zealand, Chatham, Norfolk (and Lord Howe ?) Is.; migrant,
wintering in Northern Melanesia.

SPECIMENS. Guadalcanal : Tenaru 2 ^ ad ; Betilonga i <$ ad. i $ ad to OUM.

Weight 3 cJ ad 21-5-27 (24-4). Wing 3 < ad 103-104-5 (103-7). Ta ^ 3 c? ad 68-5-70
(69-2). Bill 3 <$ ad 18-19-5 (18-5). Bill width (at front of nostril) 3 $ ad 5-5-6 (5-8).

7ns dark brown to grey-brown. Eyelid whitish. Bill black, base of mandible
silvery to dark grey. Foot bluish grey, sole pale yellowish.

NOTES. We follow Berger (1955), who shows that Chalcites cannot be separated
from Chrysococcyx and Lampromorpha.

Since Mayr's (19326) summary of records, White (1938) and Baker (1948) have
recorded this race from Bougainville, in May and September. As Mayr conjectured,
one of the specimens recorded by Tristram (1882) as plagosus actually belongs to
lucidus (Russell Is., 24th September, 1880). Another BM(NH) specimen (Bouin,
Bougainville, 25th November, 1938, W. F. H. Rosenberg) was taken long after the
migrants normally return to New Zealand (Mayr, 19326 ; Fell, 1947).

Three of Mathews' specimens from Norfolk Island support his statement (1918 :

LAND BIRDS OF GUADALCANAL 35

352) that lucidus breeds there : an adult female (2nd December, 1912) is labelled
" this bird would also have laid tomorrow "; and there are two juveniles (20th and
23rd January, 1913), one labelled " was fed by No. 536 " (the identity of which
cannot be traced). Despite Mayr's (19326 : 5) view that this race does not migrate
through Southern Melanesia, a juvenile from New Caledonia (Ausevata, 26th April,
1877, E. L. Layard) belongs to it, and not to the resident layardi.

(M. 58) Chrysococcyx lucidus plagosus (Latham, 1801)

REFERENCES. Mayr, 19326 : 6.

RANGE. Tasmania and southern Australia ; migrant, wintering from northern
Australia to Lesser Sunda Isles and Northern Melanesia.

SPECIMENS. Guadalcanal : Betilonga i ^ ad, 2 ? ad.

Weight i <J ad 21 ; 2 $ ad 21-5, 23. Wing i $ ad 89-5 ; 2 $ ad 95-5, 95-5. Tail
i ^ ad 65-5 ; i $ ad 68. Bill i ^ ad 18-5 ; i $ ad 19-5. Bill width (at front of
nostril) i <J ad 4 ; 2 $ ad 4-5, 5.

Iris dark brown to fawn. Bill black, base of mandible whitish to pale grey. Foot
bluish grey, sole pale yellow.

Two of these specimens agree in all respects with Australian C. I. plagosus, except
that their upper throats and chins are white with only a trace of barring. They have
been compared with the specimens of plagosus at the AMNH by Mrs. Vaurie (in litt.},
and at the BM(NH). The third specimen, a female, has more white and green on the
head than most of those from Australia, but falls within their range of variation. It
is damaged, but has the stump of the bill conspicuously narrow.

NOTES. It is not surprising that this race should prove to winter in the Solomons,
despite Mayr (19326 : 3), since both lucidus and plagosus are known from the Bismarcks
(see Mayr, 19410 : 73). An adult male from Guadalcanal (recorded by Ogilvie-Grant,
1888 as C. basalis) belongs to this race.

(M. 59) Eudynamis scolopacea alberti Rothschild & Hartert, 1907

REFERENCES. Rothschild & Hartert, 1907 : 440 ; 19080 : 356.

RANGE. Solomons (lowlands).

SPECIMENS. Guadalcanal : Tenaru i $ ad. San Cristoval : Goge i $ ad.

Weight i $ ad 190 ; i ? ad 148. Wing i ^ ad 185 ; i $ ad 179. Tail i ^ ad 197 ;
i $ ad 207-5. Bill i ^ ad 30 ; i ? ad 29.

Iris <$ bright crimson ; $ light rufous-brown. Bill <$ greenish silvery grey, base and
tip of maxilla black ; $ dark horn, gape and mandible light greenish-tinged horn.
Foot greenish grey, toes fuscous.

(M. 60) Eudynamis taitensis (Sparrman, 1787)

REFERENCES. Bogert, 1937 : i ; Mayr, I944C : i.

RANGE. New Zealand ; migrant, wintering in Polynesia and more rarely in
Southern and Northern Melanesia and Micronesia.

NOTES. Bogert (1937) records 18 specimens from the Solomons, mainly from small
outlying islands. No further specimens have been recorded, but French (1957) reports
the species as common in the Three Sisters.

36 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(M. 61) Centropus milo milo Gould, 1856

REFERENCES. Cat. Birds BM. 19 : 335 ; Rothschild, 1904 : 59.
RANGE. Guadalcanal, Florida Is.

SPECIMENS. Guadalcanal : Betilonga 2 < ad, i $ imm, i $ ad, i ? juv. i ^ ad to
OUM.

Wing 2 $ ad 270, 272 ; i $ ad 265. Bill 2 $ ad 63-5, 64 ; i <$ imm 57-5 ; i $ ad

63-5.

7ns dark brownish red ; imm and juv grey. Skin round eye and at base of bill dark
grey. Bill black, whitish streaks under mandible ; imm and juv maxilla grey-brown
to dark fuscous, mandible pale horn whitish below. Foot silvery blue-grey.

An immature and a juvenile in the BM(NH) agree with ours in showing that the
dull black plumage of the foreparts, though not loose in texture, belongs to the same
feather-generation as the barred loose juvenile plumage of the hinderparts. It is
replaced by immature feathers which have large subterminal spots and bars, rufous
above and whitish below, with a much paler appearance (cf . Cain & Galbraith, 1956 :
132) . The immature plumage of the hinderparts is similar in coloration, though not
in texture, to the juvenile plumage which it replaces. Variation in the colour of the
quills and their bars, and in the width and definition of the bars, seems to be merely
individual.

(M. 62) Tyto alba cf. crassirostris Mayr, 1935

REFERENCES. Mayr, 1936 : 10.

RANGE. Tanga, Nissan, Solomons and Rennell Group ; geographical variation
not yet understood.

NOTES. Only four specimens are known from the Solomons proper from Vella
Lavella, New Georgia, Malaita and Santa Ana besides the topotypical series from
Boang (Tanga group near New Ireland), two specimens from Nissan, and one from
Bellona (Mayr, 1936 ; Sibley, 1951 ; Bradley, 1962). However, the species is
apparently not uncommon on New Georgia (Sibley, 1951). It is also known from
sight-records on the Three Sisters, as a rare non-breeding visitor (French, 1957),
and on Rennell (Bradley & Wolff, 1956) where, as on Guadalcanal, San Cristoval
and Ugi (Cain & Galbraith, 1956), it is known to the natives.

The inadequate material from Nissan and the Solomons, tentatively assigned to
this race by Mayr (1936), suggests considerable geographical variation without
revealing its pattern. In coloration, these specimens agree generally with topo-
typical crassirostris, with delicatula of Australia, and with lulu of Central Polynesia ;
and are distinct from the pale meeki of south-eastern New Guinea and the ventrally
ochraceous interposita of northern Southern Melanesia. These forms are reviewed
by Mayr (1936) and Amadon (19426). Topotypical crassirostris differs from delicatula
in its much deeper bill, larger size, richer ochraceous and darker grey upper parts,
broader and darker barring on tail and wing, and somewhat bolder spotting above
and below. Four specimens from the Solomons and Nissan are intermediate in
dimensions, and one from Santa Ana agrees rather with delicatula, while they span
the range of coloration between the two races. While lulu averages still smaller and

LAND BIRDS OF GUADALCANAL 37

still less ochraceous above than delicatula, they overlap in both coloration and
dimensions.

Bradley (1962) describes as a new race, bellonae, a specimen recently collected on
Bellona. This she compares with the somewhat inadequate material of delicatula,
lulu, meeki and interposita in the BM(NH). She mentions the size of topotypical
crassirostris, given in Mayr's original and scarcely more than nomenclaturally valid
description (1935), but does not quote Mayr (1936) nor Amadon (19426). I.C.J.G. has
compared the type of bellonae with a paratype of crassirostris, with specimens from
Nissan and Santa Ana, and with the BM(NH) material. Measurements of both sexes
are combined in quoting (within square brackets) those extracted from Mayr and
Amadon, since there is apparently no sexual dimorphism even in mean dimensions ;
though in fact the four critical specimens are all females. In the following discussion
bellonae refers to the type, crassirostris to the topotypical series.

Wing-length. In many of the specimens examined, the metacarpo-phalangeal joint
distal to the wing-bend is flexed, to a degree that more or less seriously reduces the
measured wing-length. This flexure was especially marked in bellonae, so that after
freeing the joint it is possible to obtain a measurement 15 mm greater than that
taken by Bradley. The value given here was taken with the joint somewhat flexed,
for comparison with the other specimens : bellonae 282, Santa Ana 274 [273] ;
Malaita & Vella Lavella [279, 283] ; Nissan, $ 283 [281], $ [288] ; crassirostris 291
[285-290 (287-8)] ; delicatula [273-291 (284)] ; lulu [262-282 (272)].

Tail-length. Bradley clearly measured the tail from a different basal point, giving
a much higher value than those obtained by Mayr and Amadon, and (in combination
with the unduly low wing value) yielding a tail /wing ratio of 45% ; whereas Mayr
and Amadon's mean figures yield ratios between 40% and 41%. The corrected
measurements for bellonae yield a ratio of just under 40%. In order to obtain the
same values as Mayr from the same specimens, the tails were measured from the
common sheath of the central rectrices to the tip of the longer second rectrix, the
central pair being shorter and variably developed : bellonae 112 ; Santa Ana 109
[109] ; Malaita & Vella Lavella [112, 115] ; Nissan, $ 113 [113], $ [113] ; crassirostris
116 [113-116 (114-8)] ; delicatula [109-120 (114)] ; lulu [105-119 (112)].

Maxilla-depth. This character, on which Mayr assigned the Nissan and Solomons
specimens to " crassirostris ", is not mentioned by Bradley. Since bellonae has the
culmen damaged at the cere, it has been necessary to take this measurement slightly
further forward, perpendicular to the tomium just in front of its obtuse angle, across
the front edge of the nostril. This method, probably less reliable, yields measurements
approximately 0-2 mm. less than Mayr and Amadon's, which are here adjusted by
that amount : bellonae 9-3 ; Santa Ana 9-5 [9-4] ; Malaita & Vella Lavella [10-4,
10-5] ; Nissan, $ 10-5 [10-5], <$ [9-9] ; crassirostris n-o [io-8-n-i (10-9)] ; delicatula
[8-3-10-0 (9-1)] ; lulu [8-2-9-8 (9-1)].

Coloration. The female from Nissan and crassirostris differ from the specimens
of delicatula and lulu, and that from Santa Ana, in having the distal white spots
on the mantle feathers reduced and the proximal white bars somewhat widened.
This trend is carried furthest in bellonae, in which the large black spots are margined
by subequal white bars proximally and distally. The feathers denning the facial disc

38 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

have unusually small black tips in bellonae, but this condition is paralleled in a few
specimens of lulu and delicatula, and seems to be due in part to wear of the tips.
The purer and darker greys, especially on the head, of bellonae are certainly due to
foxing of the other specimens, of which the most recent (crassirostris) was taken
twenty-four years earlier ; while the same is probably true of its deeper blacks. In
its rich ochraceous areas on the upper parts, bellonae agrees with crassirostris, while
the bars on its tail and wings are intermediate in width, as in the Nissan and Santa
Ana specimens.

It seems probable that the next revision of these populations will result in a single
variable race (delicatula) in Australia, Northern Melanesia and Central Polynesia
(whether or not a broader subspecies-concept finally results in further combination) .
The data on lulu and delicatula given by Amadon are suggestive, but inadequate to
determine the statistical separation of the populations according to the "75% rule "
or a variant (see Amadon, 1949). Only the differences in wing and tail lengths are
put forward to justify separation, and it seems improbable that the statistics for the
latter (difference between means 2 mm, spreads of samples 14 & n mm, overlap
between samples 10 mm) conceal sufficiently separated distributions. Only the pro-
vision of standard deviations or related statistics could have shown whether the
wing-length distributions (for which the corresponding values are 12, 20 & 18, and
9 mm) were separable. While topotypical crassirostris seems amply distinct from
delicatula according to current standards, the few other Northern Melanesian speci-
mens already reveal inextricable intermediacy. However, it seems best to maintain
the accepted nomenclature pending a full revision, while emphasizing the hetero-
geneity of " crassirostris ". Only the pattern of the dorsal spots could conceivably
justify the separation of bellonae, and much more material from Northern Melanesia
is necessary to determine the constancy and importance of this character.

(M. 63) Ninox jacquinoti granti Sharpe, 1888

REFERENCES. Sharpe, 1888 : 183.

RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Turipava i <$ ad, i $ ad. i ad to OUM.

Weight i <$ ad 153-5 ; i $ ad 158. Wing i <$ ad 173 ; i $ ad 182. Tail i $ ad 95.
Bill i <$ ad 25-5 ; i $ ad 26.

Iris yellow. Bill dirty greenish yellow, more fuscous at base ; cere <$ greenish
yellow, $ greenish fuscous. Foot pale dirty yellow, fuscous-tinged, claws fuscous.

(M. 63) Ninox jacquinoti roseoaxillaris (Hartert, 1929)

REFERENCES. Hartert, 1929 : 6.
RANGE. San Cristoval.
SPECIMENS. San Cristoval : Goge i $ ad.
Weight 147. Wing 150-5. Tail 82-5. Bill 25-5.

Iris brownish grey. Bill greenish straw-colour, whitish blue-grey basally, cere pale
fleshy grey mottled with blackish. Foot dull whitish yellow, claws dark grey.

LAND BIRDS OF GUADALCANAL 39

This is probably the third known specimen, Ramsay's (1883 : 672) " Ninox taeniata
? juv." being the first. Dr. Amadon finds that it agrees with the holotype, except in
being slightly more conspicuously marked with white on wing-coverts and hind-neck,
and in a stronger suggestion of barring on the mid-abdomen (P. Vaurie, in litt.}.

The porphyrin pigment which gives the characteristic pink tinge to the under
wing-coverts and axillaries of this species fluoresces brilliantly under ultra-violet
illumination. The pigment is generally distributed in the dark bases of the body-
feathers in many Ninox and other owls, but forms clear patches only under the wings.
The extent of the fluorescent patches in N. jacquinoti may be of systematic signifi-
cance. In 4 specimens of eichhorni, 5 of jacquinoti and 8 of granti it extends to the
wing-bend ; whereas in this specimen of roseoaxillaris a broad area within the bend is
without porphyrin, showing white under the UV lamp. However, in a ninth specimen
of granti this area shows little fluorescence. The fluorescence is faint in some of the
older skins, and UV examination of longer series of fairly fresh specimens will be
necessary to determine the constancy of this character.

(M. 67) Collocalia whiteheadi orientalis Mayr, 1935

REFERENCES. Mayr, 1936 : 12.
RANGE. Guadalcanal (mountains ?).
NOTES. Only the holotype is known.

(M. 68) Collocalia vanikorensis vanikorensis (Quoy & Gaimard, 1830)

REFERENCES. Mayr, 1937 : 4.

RANGE. New Hebrides to Solomons and ? Bismarcks; minor geographical variation.

SPECIMENS. San Cristoval : Goge i $ ad. Ugi 2 <$ ad. i ^ ad to OUM.

Weight 3 <$ ad 9-11-5 (10-5). Wing 3 ad 113-115-5 (114-7). Tail 2 ^ ad (inner)
44, 45 ; (outer) 54, 54-5. Bill 3 <$ ad 8-5-9 (87)-

Iris very dark brown. Bill black. Foot fuscous grey, shanks purplish, fleshy above,
toes blackish.

NOTES. In addition to the islands in the Solomons listed by Mayr (1937), the
species is known from Malaita (Davidson, 1934 : 193), San Cristoval and the Three
Sisters (Cain & Galbraith, 1956 : 133, 106), and from Rennell (Bradley & Wolff,
1956 : 102).

(M. 69) Collocalia spodiopygia reichenowi Stresemann, 1912

REFERENCES. Streseman, 1912 : 350 ; Hartert, 19246 : 269 ; Mayr, 1937 : 16.

RANGE. Solomons, Rennell.

SPECIMENS. San Cristoval : Goge i <$ ad. Ugi i <$ ad.

Weight 2 c? ad 7, 8-5. Wing 2 $ ad 103, 105. Tail 2 ad (inner) 41, 41 ; (outer)
47, 49-5. Bill i < ad 8.

Iris very dark brown. Bill black. Foot fuscous (i very pale), toes blackish.

NOTES. Formerly known only from Guadalcanal and Kolombangara (Stresemann,
1912 ; Mayr, 1945^ : 239), but now also from San Cristoval and Ugi (Cain & Galbraith,
1956 : 133) and Rennell (Bradley & Wolff, 1956 : 102). In these localities at least, it
is not confined to the mountains.

40 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(M. 70) Collocalia esculenta becki Mayr, 1931

REFERENCES. Mayr, 19316 : 16.

RANGE. Solomons except San Cristoval Group.

SPECIMENS. Guadalcanal : Betilonga i <$ ad. To OUM.

Weight 7-5. Tail inner 39 ; outer 42. Bill 6-5.

Iris dark brown. Bill black. Foot pinkish grey, claws fuscous.

NOTES. Recorded from Malaita by Mayr & Camras (1938).

(M. 70) Collocalia esculenta makirensis Mayr, 1931

REFERENCES. Mayr, 19316 : 15.

RANGE. San Cristoval, Three Sisters, ? Ugi.

SPECIMENS. San Cristoval : Goge i <$ ad ; Nagasi 2 <$ juv, 2 $ ad, i ? juv. i $
ad to OUM.

Weight i <? ad 5 ; 2 ^ juv 5-5, 6-5 ; i $ ad 5-5. Wing i $ ad 87 ; 2 $ ad 93, 93.
Tail (inner) i g ad 39-5 ; 2 $ ad 39, 41 ; (outer) i ^ ad 41 ; 2 $ ad 44, 44-5. Bill
i c? ad 6-5 ; 2 $ ad 6, 6.

Iris very dark brown. Bill black. Foot purplish grey, pale to blackish, claws black ;
juv pale, dull flesh to grey, with purplish tinge.

NOTES. Two BM(NH) specimens recorded from Ugi (Gray, 1873) carry field labels
showing them to have been collected there by Brenchley, and belong to this race.
The species has not recently been taken on Ugi (Mayr, 19450 : 279) and seems not to
be present (Cain & Galbraith, 1956 : 104).

(M. 71) Hemiprocne mystacea woodfordiana (Hartert, 1896)

REFERENCES. Hartert, 1896 : 19 ; Stresemann, 1921 : 38.

RANGE. Feni, Solomons, Rennell (lowlands).

SPECIMENS. Guadalcanal : Nala i $ ad, i $ imm. i $ imm to OUM.

Weight i $ ad 59-5 ; i $ imm 53. Wing i $ ad 202. Tail i $ ad (inner) 58-5 ; (outer)
185. Bill i ? ad 16-5 ; i $ imm 16-5.

Iris dark brown. Skin round eye and gape purplish grey. Bill black. Foot blackish,
legs purplish.

(M. 72) Alcedo atthis salomonensis Rothschild & Hartert, 1905

REFERENCES. Rothschild & Hartert, 1905 : 255 ; 19086 : 361 (ispida subsp.) ;
Stresemann, 1913 : 316.

RANGE. Solomons ; irregular geographical variation, some populations scarcely
distinct from A. a. hispidiodes (mainly lowlands).

SPECIMENS. San Cristoval : Goge i $ imm, i $ imm. i $ imm to OUM.

Weight i $ imm 35 ; i $ imm 35-5. Wing i $ imm 73 ; i $ imm 74. Tail i $ imm
33-5 ; i $ imm 32-5. Bill (from front of nostril) i <$ imm 31 ; i $ imm 31-5.

Iris very dark brown. Bill black ; <$ gape dull pinkish fuscous, pale salmon-pink
internally ; $ gape dull orange, base of mandible dull orange. Foot orange-red with
fuscous wash, especially on toes.

LAND BIRDS OF GUADALCANAL 41

(M. 73) Ceyx pusillus aolae (Ogilvie-Grant, 1914)

REFERENCES. Ogilvie-Grant, 1914 : 13.

RANGE. Guadalcanal, Florida Is. (coasts).

NOTES. This mangrove-living species is known by only a few specimens from any
island in the Solomons, and though it has not yet been recorded from San Cristoval
or Malaita it may still be found there.

(M. 74) Ceyx lepidus nigromaxilla Rothschild & Hartert, 1905

REFERENCES. Rothschild & Hartert, 1905 : 256 ; Mayr, 1936 : 5.

RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Betilonga 3 <$ ad, i $ ad ; Tasinasa, near Nala, i $ ad ;
Turipava i $ ad. I $ ad to OUM.

Weight 3 3 ad 19-19-5 (19-2) ; 2 ? ad 19, 20-5. Wing 4 $ ad 57'5-59'5 (58-8) ; 2 ?
ad 60, 62. Tail 4 < ad 23-5-25-5 (24-8) ; 2 $ ad 25, 25. Bill (from front of nostril)
4 3 ad 29-30-5 (29-8) ; i $ ad 30-5.

Iris dark brown. Bill maxilla black, more or less streaked with dull orange,
mandible orange-red streaked with black at tip and sometimes towards base. Foot
orange to orange-red.

The small weight (16-5) of one male from Betilonga is not listed above, since the
bird was evidently starved : it was caught by hand, and found to have one wing
hampered by sticky fruit.

(M. 74) Ceyx lepidus gentianus Tristram, 1879

REFERENCES. Tristram, 1879 : 438.

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Goge 3 <$ ad, 2 $ ad, i $ juv, i ? juv ; Nagasi 2 $ ad.

1 $ ad to OUM.

Weight 3 $ ad 24-5-28 (26-8) ; 4 $ ad 27-29-5 (28-4) ; i ? juv 25. Wing 3 $ ad
61-5-64 (62-5) ; 4 <j> ad 63-65 (64-4). Tail 3 $ ad 27-5-28-5 (28-0) ; 3 ? ad 29-31
(30-2). Bill 3 3 ad 31-5-33 (32-2) ; 4 ? ad 29-31 (29-9) ; i ? juv 19.

Iris very dark brown ; juv very dark grey -brown. Bill black, gape greyish to
brownish fuscous ; juv tip white. Foot orange ; juv pale pinkish flesh.

The juveniles (which have growing quills, and white overhanging tips to their bills)
were taken from a nest hole, by a native collector who had seen them flying.

(M. 75) Halcyon chloris alberti Rothschild & Hartert, 1905

REFERENCES. Hartert, 19266 : 132 ; Mayr, 1936 : 6.

RANGE. Solomons, except Russell Is., Malaita and San Cristoval Group : very
similar to H. c. tristrami.
SPECIMENS. Guadalcanal : Tenaru 3 < ad, i $ juv, 2 ? ad, i ? juv ; Betilonga

2 ^ ad. i c? ad, i ? ad to OUM.

Weight 5 $ ad 73-5-82-5 (76-8) ; i ^ juv 70 ; 2 ? ad 77, 77-5. Wing 5 $ ad 104-5-
110-5 (107-3) ; i $ juv 106-5 ; 2 $ ad 104-5, 105-5. Tail 4 $ ad 64-68 (65-4) ; i 3

42 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

juv 67 ; 2 $ ad 67, 69-5. Bill (from front of nostril) 5 ^ ad 37-39 (38-2) ; i <$ juv
33 ; 2 $ ad 38-5, 41.

Iris dark brown ; i juv dark grey-brown. Bill black, basal half of mandible ivory,
sometimes pinkish ; juv tip ivory, gape pale grey, base brownish. Foot fuscous black.

(M. 75) Halcyon chloris solomonis Ramsay, 1882

REFERENCES. Hartert, 19266 : 132.

RANGE. San Cristoval Group, except Three Sisters.

SPECIMENS. San Cristoval Manewiriwiri i <$ ad ; Goge 19 <$ ad, 8 9 ad, 5 $ imm.
Ugi 12 c? ad, i juv, 18 $ ad. i <$ ad, i $ ad to OUM.

Weight San Cristoval 17 < ad 47-5-61 (57-4) ; 7 ? ad 61-71-5 (65-0) ; 5 $ imm
57-66 (61-5) ; Ugi 12 $ ad 54-62-5 (56-5) ', i c? juv 58-5 ; 18 $ ad 57~69'5 (62-9)-
Wing San Cristoval 20 <$ ad 89-5-96 (92-7) ; 6 $ ad 93-95-5 (94-4) ; 5 ? imm 90-5-94
(92-2) ; Ugi 9 c ad 92-5-96 (94-7) ; i <$ juv 92 ; 15 $ ad 90-100 (94-8). Tail San
Cristoval 19 <J ad 59-66-5 (63-1) ; 6 $ ad 63-67-5 (64-7) ; 5 $ imm 62-64 (62-8) ;
Ugi 9 <$ ad 61-65 (63-4) ; 12 $ ad 59-5-68-5 (66-0). Bill (from nostril) San Cristoval
20 c? ad 34'5-39 (37' 1 ) ; 8 $ ad 37'5~4o-5 (38-5) ; 5 ? imm 36-5-40-5 (38-0) ; Ugi
12 3 ad 33-5-38-5 (36-7) ; i 3 juv 32-5 ; 18 $ ad 33'5-4'5 (38-5)-

Iris dark to very dark brown (i dark grey-brown). Bill black, base of mandible
white to very pale grey, sometimes pinkish ; juv base of mandible dull grey. Foot
blackish, shank purplish.

As Ramsay (18826 : 834) points out, the pale supraloral spots, which are the rem-
nants of the " ringband " (Mayr, 1931^ : 3) in this race, are sometimes still further
reduced to vanishing-point. This tendency seems to be confined to Ugi, whence n of
our 31 specimens (6 , 5 $) have the spots obsolete or lacking. All 33 specimens from
San Cristoval (and the one from Santa Ana at the AMNH P. Vaurie, in litt.) have
the spots well developed, extending back as more or less concealed superciliary stripes.
Indeed, one male from Goge (BM(NH) reg. no. 1959.21.329) has a well-developed
though concealed ring-band connecting the superciliaries across the nape ; although
the ochraceous area is subterminal on most of the nape feathers, a few are pale to the
tip.

There are also striking differences, in the mean values of several colour characters,
between the San Cristoval and Ugi series, with the former much the more variable.
This is most conspicuous in the coloration of the male underparts, which are pre-
dominantly rufous in many San Cristoval specimens but mainly white in the Ugi
series. However, plotting colour-classes against collection-dates shows a progressive
paling, superimposed on great variation at any one time, during eight weeks on San
Cristoval ; and suggests that the more uniform palor of the Ugi series (collected in
two weeks) may result merely from the continuance of this trend. Like the bleaching
of phaeomelanin, most of the other colour characters which vary in these series are
known to be affected by wear (Mayr, 1931^ : 3). In each the San Cristoval series is
the more variable, while the mean displacement of the Ugi series is towards the worn
condition : head bluer, mantle less blackish green, less dark fringing on sides of
breast, female underwing less ochraceous.

LAND BIRDS OF GUADALCANAL 43

The impression given by the present series, that solomonis is unusual among the
races of H. Moris in its susceptibility to wear and consequent variability, needs
confirmation on fresher specimens. Intergradation from pale to dark ochraceous also
appears unusually abrupt on the underparts of the most deeply-coloured males,
though the effect may be exaggerated by the make-up of the skins. The rufous is most
intense at the sides of the breast, between which the paler mid-breast contrasts
sharply with the large white throat-patch and intergrades with the still paler belly ;
but the field impression is not of an interrupted ochraceous gorget (cf . Rothschild &
Hartert, 19086 : 361 ; Hartert, 19266 : 132 ; Mayr, 19450 : 243).

The 5 female immatures from Goge (collected 2nd-i2th November, ovaries enlarged,
i with oocytes c. 3 mm diam.) do not have the short bills with white overhanging
tips, nor the deeply ochraceous and strongly-fringed underparts, characteristic of
juveniles. However, they show the juvenile characters of buffy under wing-coverts,
buff edges to the upper wing-coverts, a little buff in the forehead, and somewhat more
olivaceous upperparts. No comparable stage is distinguishable in the other series :
3 males from Goge have some buff in the upper wing-coverts, but no other juvenile
characters.

(M. 75) Halcyon Moris sororum new subspecies

HOLOTYPE. British Museum (Natural History) reg. no. 1940.7.4.17. $ adult,
Malau Paina Island, igth December, 1938, coll. W. French.

RANGE. Malaupaina and Malaulalo Islands, Three Sisters or Olo Malau group,
north of San Cristoval, British Solomon Islands.

DIAGNOSIS. Apparently very like topotypical santoensis Mayr (19310) from
Espiritu Santo, New Hebrides, but with wider bill. Bill width in mm, taken with
vernier calipers at front edges of nostrils :

sororum 4 < ad 13-4, 13-7, 13-8, 13-8 ; 3 $ ad 13-7, 14-4, 14-4.
santoensis 3 ad 12-6, 12-7, 12-9 ; 5 ? ad 12-8, 12-9, 13-3, 13-4, 13-4.

MATERIAL. Three Sisters: n specimens, BM(NH) reg. nos. 1934.8.15.3,
1938.11.23.41-43, 1940.7.4.15-21. All but one collected by W. French, though
some marked as collected by W. H. Barrow. Malaupaina : 4 ^ ad (igth December,
1938, 25-27th March, 1939), i $ imm (nth May, 1934, coll. R. A. Lever), i $ ad (igth
December, 1938), i $ imm (25th March, 1934), i ? ad (27th February, 1934), i ? imm
(ist March, 1934). Malaulalo : 2 $ ad (30th March, 1939).

Wing 4 $ ad 96-5-101 (98-2) ; i ^ imm 94 ; 3 $ ad 95-100 (98-0) ; i $ imm 95 ;
i ? ad 99-5 ; i ? imm 95-5. Tail 4 ^ ad 66-5-67-5 (66-9) ; 3 $ ad 63-5-69-5 (67-5) ;
i ? ad 68-5. Bill (from front of nostril) 4 <$ ad 38-39 (38-4) ; i $ imm 39 ; 3 $ ad
39-42 (40-3) ; i $ imm 40 ; i ? ad 37 ; i ? imm 39-5.

7ns " dark brown ". Bill " black " (and " white "). Foot " dark flesh ", " grey "
or " black " (February to May), " dirty white " (December somewhat paler in dried
skins) .

DESCRIPTION. Differs from solomonis, and agrees with the Southern Melanesian
forms (see Mayr, 19310), in having the ringband fully-developed across the nape.

44 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

The series is in more or less worn plumage, so that critical comparisons of the blue
and ochraceous colours are impossible. However, where comparative material is
suitable, reasonable predictions can be made about the coarser differences, and these
are given in square brackets. The series has been compared with specimens of
solomonis, all the Southern Melanesian races, and three from western Polynesia.
Compared with :

solomonis (Solomons San Cristoval and Ugi) : ringband continuous [<$ ringband

deeper rufous relative to underparts ; <$ underparts much less rufous than on

San Cristoval ; wing and tail paler and greener relative to mantle] .
amoena (Rennell) : much larger [< much paler and greener above].
vicina (Sta. Cruz group Duff group) : bill longer and much wider.
brachyura (Sta. Cruz group Reef group) : somewhat larger, bill considerably so

[$ ringband relatively deeper rufous, $ whiter below].
ornata (Sta. Cruz group Santa Cruz) : [^ mainly white below] .
utu-puae (Sta. Cruz group Utupua) : supraloral spots much smaller, ringband

narrower [$ paler above].
melanodera (Sta. Cruz group Vanikoro) : ringband much wider, black nuchal band

narrower [less black fringing on breast, mantle much less blackish].
torresiana (New Hebrides Torres group) : ringband wider, especially above auri-

culars [<$ mantle duller and greener, $ head bluer relative to mantle],
cf. santoensis (New Hebrides Banks group) : ringband wider, especially above

auriculars, bill longer and somewhat wider.
santoensis (New Hebrides Santo) : bill wider [tail relatively paler and greener, head

relatively bluer, <$ under wing-coverts deeper buff].
juliae (New Hebrides Efate) : bill longer and much wider.
erromangae (New Hebrides Erromanga) : bill longer and much wider, collar narrower

[c mainly white below, $ underwing almost pure white].
tannensis (New Hebrides Tanna) : bill longer and much wider, collar narrower [<

mainly white below, 9 underwing almost pure white] .
cf . juliae (New Hebrides Aneiteum) : bill much wider.
eximia (Fiji Kandavu) : bill longer and wider, ringband less mixed with blue

posteriorly [<$ whiter below].
vitiensis (Fiji Viti Levu, Ovalau) : bill longer and much wider, ringband wider,

especially over auriculars, and less mixed with blue [ J much whiter below] .
sacra (Tonga) : bill longer and wider [J underwing with some ochraceous] .

NOTES. The general resemblance between sororum and solomonis, with the one
trenchant difference bridged by the ringbanded individual from Goge mentioned
above, confirms that the major cleavage between the races characteristic of Southern
and of Northern Melanesia lies (despite Mayr, 19310; : 10) not between amoena and
solomonis, but between solomonis and alberti. This affects the faunal affinities of
Rennell (see Mayr, iQ3ic ; Bradley & Wolff, 1956 ; Braestrup, 1956). It is remarkable
that otherwise similar individuals on Ugi, fourteen miles away, differ so markedly
from the Three Sisters population in lacking the ringband altogether.

This subspecies is named in honour of our daughters Angela, Nino and Janet.

LAND BIRDS OF GAUDALCANAL 45

(M. 76) Halcyon saurophaga saurophaga Gould, 1843

REFERENCES. Cat. Birds. B.M. 17 : 249. See Mayr, 19490 : 56.

RANGE. Moluccas, northern New Guinea and nearby islands, Northern Melanesia
except Admiralty and Ninigo groups (coasts).

SPECIMENS. Ugi 3 ^ ad, i ^ subad ?. i $ ad to OUM.

Weight 3 < ad 123-146 (132-2) ; i $ subad 111-5. Wing 3 ^ ad 122-131 (127-8) ;
i <J subad 124. Tail 3 $ ad 80-84-5 (82-2) ; i $ subad 80. Bill (from nostril) 3 $
ad 53-53-5 (53'3) ; i cJ subad 48.

7ns very dark brown. Bill black, base of mandible white to very pale grey, pinkish
towards extreme base. Foot slate-grey, scutes fuscous black.

The " subadult " specimen is here separated solely on measurements, especially
bill-length.

(M. 77) Halcyon australasia sancta Vigors & Horsfield, 1827

REFERENCES. Cat. Birds B.M. 17 : 267 ; Keast, 1957 : 68.

RANGE. Southern Australia ; partial migrant, many individuals wintering from
Lesser Sunda Isles to Solomons, where a few apparently remain all year (lowlands).

SPECIMENS. Guadalcanal : Tenaru 6 <$ ad, 2 $ ad, i ? ad. i J ad to OUM.

Weight 6 $ ad 46-58 (50-0) ; 2 $ ad 54-5, 54-5. Wing 6 $ ad 88-94-5 (92-0) ; 2 $
ad 90, 90. Tail 6 $ ad 57-61 (58-8) ; 2 <j> ad 56, 61. Bill (from nostril) 6 J ad 32-38
(34-5) ; 2 ? ad 32-5, 37-5.

Iris dark to very dark brown. Bill black, base of mandible dirty white to silvery,
sometimes pinkish. Foot fuscous grey, toes blackish.

NOTES. We follow Mayr (19446 : 119) and van Bemmel (1948 : 359) in regarding
sancta as conspecific with H. australasia of the Lesser Sunda Isles, despite the overlap
in that area outside the breeding-season. In the Solomons, some of the individuals
which remain during the breeding season do appear to breed (Cain & Galbraith, 1956 :
263 ; French, 1957). French (in litt.} writes of the Three Sisters : " The Sacred King-
fisher (Halcyon sancta} bred, and its breeding habits and choice of sites were similar
to the White-collared Kingfisher (Halcyon Moris [sororum]). Both utilize termites
nests on trees in fairly open forest and frequently those on coconut trees in the planta-
tions, and seem to prefer those at a height of from 3 ft. to 10 ft. above ground level.
The drilled hole is very like that of our Green Woodpecker [Picus viridis] but not
nearly so deep. The Beach or White-headed Kingfisher (Halcyon saurophaga} always
nested in dead and rotting stumps along the shore, and the hole was invariably drilled
straight in, and in many cases straight through leaving a double entrance. All three
species laid 2 or 3 eggs, the latter being the commonest full clutch."

(M. 78) Halcyon leucopygia (Verreaux, 1858)

REFERENCES. Cat. Birds B.M. 17:252; Hartert, 19266:134.
RANGE. Bougainville, Shortland, Choiseul, Ysabel, Florida Is., Guadalcanal
(lowlands) .

SPECIMENS. Guadalcanal : Tenaru 5 ad, 6 $ ad. i <$ ad, i $ ad to OUM.
Weight 5 <$ ad 42-5-52 (46-7) ; 6 $ ad 43'5-5i'5 (49' J )- Wing 5 ^ ad 84-92 (87-9) ;

46 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

6 ? ad 87-90 (88-3). Tail 5 ad 55-63 (56-1) ; 6 ? ad 57-61 (58-5). Bill (from nostril)
5 <? ad 32-5-34-5 (337) ; 6 ? ad 32-5~34'5 too).

/ns dark brown. Skin round eye white. Bill black, extreme base of mandible dirty
white to blackish. Foot dark grey, toes and scutes black, upper shanks purplish.

(M. 79) Halcyon bougainvillei excelsa Mayr, 1941

REFERENCES. Mayr, 1941^ : 3.
RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Betilonga 2 $ ad. i $ ad to OUM.

Weight 2 $ ad 160, 215. Wing 2 $ ad 135, 137. Tail 2 $ ad 99, 100-5. Bill (from
nostril) 2 $ ad 40-5, 41.

7ns dark brown. Bill orange. Foot orange.

These specimens, whose ovaries were somewhat developed, agree with the hitherto
unique holotype in having softer plumage than H. b. bougainvillei, so that this does
appear to be a racial character (cf. Mayr, 1941^). Differences of these two from the
holotype (apparently larger ; colours of crown, wing and tail and sizes of collar and
rump-patch as in bougainvillei) may reflect the fact that the latter is in very worn
plumage (P. Vaurie, in litt.).

NOTES. This race differs from bougainvillei chiefly in the pure olive upper back and
paler underparts of the female, while the male remains unknown. Despite Mayr
(1941^), the differences are not very marked, considering the separation of the races
and the peculiarity of the species.

(M. 80) Merops ornatus Latham, 1801

REFERENCES. Cat. Birds B.M. 1 7 : 74.

RANGE. Australia ; migrant, wintering from Celebes to Northern Melanesia.
NOTES. The only published records for the Solomons seem to be those of Ramsay
(18820 : 20).

(M. 81) Eurystomus orientalis solomonensis Sharpe, 1890

REFERENCES. Ripley, 1942 : 174.

RANGE. Feni, Solomons.

SPECIMENS. Guadalcanal ; Betilonga i <$ ad, 3 $ ad. San Cristoval : Goge 2 $
ad, i <$ juv ; Nagasi i ad. i <$ ad, i $ ad to OUM.

Weight 4 $ ad 149-5-166 (159-0) ; i <$ juv 120 ; 3 $ ad 151-5-185 (166-8). Wing
4 $ ad 190-5-201 (197-8) ; 3 $ ad 191-200-5 (197-0). Tail 4 <$ ad 125-134 (129-9) > 2 $
ad 132, 136. Bill (from feathering) 4 <$ ad 23-23-5 (23-2) ; i $ juv *4 ; 3 ? ad 21-5-24
(22-5).

Iris dark brown. Eyelid dull orange. Bill orange-red, sometimes pinkish ; juv
maxilla black with flesh-coloured base and edges, mandible flesh-colour. Foot dull
orange-red, claws blackish ; juv pinkish fawn, washed with fuscous on toes, purplish
on shanks.

LAND BIRDS OF GUADALCANAL 47

(M. 82) Rhyticeros plicatus mendanae Hartert, 1924
REFERENCES. Mayr, 19340 : 10 ; Sanft, 1960 : 120.
RANGE. Choiseul, Ysabel, Guadalcanal, Malaita.
SPECIMENS. Guadalcanal : Betilonga i ad, i ? ad. To OUM.
Wing i $ ad 386. Tail i $ ad 227. Bill i ^ ad 185-5 ; i ? ad 169.
Iris <$ orange, brown. Skin round eye blue ; eyelid <$ red, $ grey ; throat white.
Bill dark red basally, tip pale horn. Foot black.

(M. 84) Hirundo tahitica subfusca Gould, 1856
REFERENCES. Mayr, 1934*2 : 13 ; 1955 : 6.

RANGE. Polynesia to Solomons and Lihir, intergrading via New Ireland with
H. t. ambiens (coasts).

(M. 85) Hirundo nigricans nigricans Vieillot, 1817

REFERENCES. White, 1936 : 90.

RANGE. Australia ; migrant, wintering in New Guinea region and Northern
Melanesia (mainly lowlands).

NOTES. Only two specimens have been recorded from the Solomons, both from
Guadalcanal (Hartert, 1929 : 7 ; Mayr, 1955 : 6).

(M. 86) Lalage leucopyga affinis (Tristram, 1879)

REFERENCES. Mayr & Ripley, 1941 : 17.
RANGE. San Cristoval, Ugi.

SPECIMENS. San Cristoval : Manewiriwiri i ^ ad ; Goge i ^ ad, 2 <$ imm, 3 $
ad, 2 $ imm, 2 ? juv ; Nagasi i ^ ad, 2 $ ad. Ugi 8 $ ad, i <$ imm, 4 $ ad. i ^ ad,

1 $ ad to OUM.

Weight San Cristoval 3 ^ ad 24-25 (24-2) ; 2 $ imm 21, 24-5 ; 5 $ ad 21-28
(24-3) ; i $ imm 24 ; Ugi 8 $ ad 23-5-27 (25-6) ; i $ imm 23-5 ; 4 $ ad 24-26 (24-9).
Wing San Cristoval 3 <$ ad 81-5-87 (84-2) ; 2 J imm 80, 82-5 ; 4 $ ad 80-82-5 (81-6) ;

2 $ imm 79, 84 ; Ugi 8 ^ ad 82-5-87-5 (85-1) ; i $ imm 78-5 ; 4 $ ad 80-85 (82-9).
Tail San Cristoval i ^ ad 74 ; i $ imm 68 ; i $ ad 75 ; 2 $ imm 66, 74 ; Ugi 6 $
ad 66-76 (71-5) ; 2 ? ad 69-5, 73. Bill San Cristoval 3 <$ ad 17-5-18 (17-7) ; 2 imm
17, 17-5 ; 3 ? ad 17-5-18 (17-7) ; 2 $ imm 17-5, 18; Ugi 7 <? ad 16-18 (17-3) ; i <J
imm 18 ; 3 ? ad 17-5-18-5 (18-0).

Iris very dark brown Bill black, base of mandible paler ^ more or less grey, $
grey to horn ; imm pale area yellower and more extensive ; juv gape yellow. Foot
dark grey or black, upper shank fuscous or brownish.

One of the immature males from Goge had enlarged testes.

NOTES. Clearly the single very large male from Ugi (Mayr & Ripley, 1941) was
not characteristic of the population.

(M. 87) Coracina holopolia holopolia (Sharpe,
REFERENCES. Sharpe, 1888 : 184 ; Rothschild & Hartert, IQOIC : 374 ; Mayr,

1931/1 17-

RANGE. Buka, Bougainville, Choiseul, Ysabel, Guadalcanal (mainly lowlands ?).

48 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

SPECIMENS. Guadalcanal : Tenaru i $ imm ; Betilonga i $ subad.

Weight i <$ ad 43 ; i $ subad 41 ; i $ imm 47. Wing i $ subad 105 ; i <j> imm 104.
Bill i $ subad 25 ; i $ imm 23-5.

Iris dark brown. Bill black ; subad and imm fuscous to dark grey. Foot black.

Neither the juvenile nor the immature plumages have been described for this
species, and the latter is not represented in the AMNH (P. Vaurie, in litt.}, while
geographical variation has been studied only in the adult male. The subadult is in
adult plumage, except for its juvenile remiges (as retained by immatures), apparently
immature axillaries and under wing-coverts, and a single immature feather on the
lower breast. The immature is moulting from the juvenile to the immature plumage,
but accidental loss and replacement of some feathers has superficially confused its
plumage sequence : some of the secondaries in the right wing are essentially adult,
while an underlying patch of feathers on the right lower breast is more worn and
more juvenile in character (more boldly barred, the white tinged with yellow) than
the surrounding immature plumage. The immature contour-feathers are grey above
as in the adult, but on throat, breast, under tail-coverts, under wing-coverts and
axillaries they are boldly barred in black and white, grading on the flanks and belly
into grey with narrow white bars.

The juvenile plumage, softer than the succeeding ones, is fuscous and fawn in
colour. The rectrices are pointed, the central pair grey and the remainder browner
than in the adult, with a terminal pattern best developed on the outermost feathers.
This pattern consists of a wide fawn tip, not extending far up the margins of the
webs, which bears a dark line parallel to the margin on each side, and indications of a
second line within this. The wing feathers are fuscous, with fawn outer margins
expanding into wider tips. The tips of the coverts and inner secondaries bear a
pattern of two dark bars similar to that on the tail. Both bars run straight across the
tip, while the inner one bends sharply at each end to run for a short distance parallel
to the edges of the feather ; the inner edge of this bar is ill-defined. The contour
feathers are square-cut, fawn with brown cross bars. On head and back there are
two straight bars, while on breast and flanks there is usually a single }-shaped bar,
and a vague shaft-streak basally.

NOTES. The juvenile plumage is very different from that of Coracina lineata, in
which the pale areas are yellowish white and the dark blackish, the contour feathers
rounded, with pale edges (giving a scalloped effect) on the underparts, and the pale
edges on remiges and rectrices simple. The immature plumage, with its ventral
barring and white-flecked auriculars, is reminiscent of the female of C. dohertyi (which
represents the morio-tenuirostris superspecies on Sumba and Flores Stresemann,
1939 : 124 ; Mayr, 19440 : 142), though its barring is less bold.

An adult male, too damaged to be retained, was taken near Tsea. The locality
Tenaru was omitted by Cain & Galbraith (1956 : 266). Thus two of the three speci-
mens were taken in the lowlands. The vertical distribution of the species is not clear :
possibly it is subject to geographical variation.

This species and the morio-tenuirostris superspecies were formerly placed in
Edolisoma, now considered as merely a subgenus of Coracina (Delacour, 1946 : 2 ;
see Mayr, 1955).

LAND BIRDS OF GUADALCANAL 49

(M. 88) Coracina tenuirostris erythropygia (Sharpe, 1888)

REFERENCES. Rothschild & Hartert, IQOIC : 373 ; 1902 : 582 ; Mayr, 19317: 18 ;
1955 : 12.

RANGE. Guadalcanal and Malaita Groups, with slight geographical variation.

SPECIMENS. Guadalcanal : Tenaru 2 ^ ad ; 2 ? imm ; Betilonga u <$ ad, 4 $
ad, i $ imm, 3 ? imm, i ? juv ; Turipava i $ ad. i ^ ad, i $ ad to OUM.

Weight ii (? ad 54-63 (58-1) ; 2 <$ imm 54, 61-5 ; 5 $ ad 52-64-5 (57-4) ; i $ imm
55. Wing 10 <? ad 113-120 (117-5) ; i <J imm 112 ; 5 $ ad 109-113-5 (111-5) ; i ?
imm 112. Bill 9 $ ad 25-5-29 (27-0) ; 2 ^ imm 26-5, 26-5 ; 4 $ ad 26-27-5 (26-6) ;
i $ imm 26.

Iris dark brown. B?7/ black, base of $ mandible sometimes paler ; imm maxilla
fuscous to blackish, mandible pale horn to warm grey with darker tip ; juv bill dull
brown. Foot fuscous black, $ sometimes paler ; imm and juv dark grey.

Apart from their juvenile remiges and rectrices, the immatures differ from adult
females in having the cap brown instead of grey and the mantle more rufous, less
greyish or olivaceous.

In the juvenile plumage, each upper wing-covert and inner secondary has a rufous
tip bearing a white terminal spot and a blackish bar. The head is yet more brown
than that of immatures, with whitish fringes. The underparts bear dark drop-shaped
shaft-streaks.

One of the adult females (BM(NH) reg. no. 1959.21.542) has sparse and irregular
shaft-streaking and barring of the underparts. This seems relevant to the origin of
the barred race nisoria of the Russell Is. (see Mayr, 1955 : 13).

(M. 88) Coracina salomonis (Tristram, 1879)
REFERENCES. Ramsay, 1883 : 667 ; Rothschild & Hartert, 19086 : 362 ; Mayr,

1955 : 12.

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Goge 13 ^ ad, 3 ^ imm, 8 $ ad, i $ imm, i ? imm ;
Nagasi 3 ad, 2 $ imm, 6 $ ad, i $ imm, 2 ? juv. i ^ ad, i $ ad to OUM.

Weight 15 c? ad 55-75 (64-6) ; 5 <$ imm 62-5-65-5 (63-8) ; 12 $ ad 60-5-73 (65-7) ;
i $ imm 64-5. Wing 14 <$ ad 117-122-5 (119-2) ; 5 $ imm 115-5-118 (116-5) '> I2 ?
ad 115-120-5 (117-4) ; 2 $ imm 112, 112. Bill 15 $ ad 27-5-32 (28-9) ; 4 imm
27-5-29-5 (28-4) ; 14 $ ad 28-30-5 (29-4) ; 2 ? imm 28-5, 29.

Iris very dark brown ; juv dark grey-brown. Bill black ; imm gape and base of
mandible sometimes grey ; juv mandible grey or brown with pale subterminal spot.
Foot black, shanks sometimes brownish ; juv dark grey.

The immatures are indistinguishable from adult females, except for their juvenile
wings and tails. In the juvenile plumage, the wing-feathers have the white terminal
spots obsolete and the dark subterminal bars not clearly demarcated. The feathers
of the crown are dark fuscous grey, with black subterminal bars and whitish-buff
tips ; those of the mantle and the upper tail-coverts have white tips, followed by
rufous and then by black subterminal bars. The underparts are paler than those of
immatures and adult females, with vague blackish streaks-shaft and whitish tips.

ZOOL 9. i. 4

50 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

NOTES. In Coracina, two character-series can be traced in the female plumage, in
each of which the first-mentioned term is closest to the male and the last to the
juvenile plumage. The series in distribution of eumelanins runs : all-grey grey
with narrow pale bars black, grey and pale bars pale with narrow black bars
all-pale. That in the intensity of phaeomelanins in the pale areas runs : none (white)
pale (buff) deep (rufous). Total cock- feathering has clearly arisen repeatedly in
the morio-tenuirostris superspecies. However, discordance between the coloration of
dorsal and ventral surfaces is rare, being confined to the females of morio and closely-
related races in Celebes, dohertyi of Flores and Sumba, and salomonis. In the two
former, the underparts lie towards the " male " ends of both character series, whereas
in salomonis they are as " juvenile " as possible. In all three, the upperparts are as in
the adult male. Since morio is sympatric with a characteristic form (edithae) of C.
tenuirostris , while Mayr (19440 : 142) considers dohertyi as a full species, and since
salomonis is also distinct in the more conservative male plumage, it should be con-
sidered as another representative species.

(M. 89) Coracina lineata solomonensis (Ramsay, 1879)

REFERENCES. Ramsay, 18790; : 71 ; Tristram, 1892 : 294 ; Rothschild & Hartert,
1905 : 264 (pusillus) ; Mayr, 19317 : 17 ; 1955 : 14. See Ripley, 1941 ; Voous & van
Marie, 1949.

RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Tenaru i ^ ad ; Tasinasa i $ ad ; Betilonga 5
ad, 6 ? ad ; Turipava i <$ ad, i $ ad. i $ ad, i $ ad to OUM.

Weight 7 ad 56-64 (60-4) ; 8 $ ad 50-65-5 (58-9). Wing 7 <$ ad 129-135 (131-8) ;
8 $ ad 124-5-129-5 (126-3). Bill (from front of nostril) 7 $ ad 10-5-12 (11-2) ; 8 $ ad
11-11-5 (n-i).

7m yellow. Bill black. Foot black.

The males show considerable variation, between individuals and between the
feathers of one individual, in the degree of barring of under wing-coverts and axillaries;
the extremes being blue-grey with faint white bars on the one hand, and boldly
barred in black and white on the other. The barred feathers are probably remnants
of the immature plumage, which (in solomonensis, ombriosa, nigrifrons, and axillaris
at least) is ventrally barred (P. Vaurie, in litt.}.

(M. 89) Coracina lineata makirae Mayr, 1935

REFERENCES. Mayr, 1936 : 13. See Ripley, 1941 ; Voous & van Marie, 1949.

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Goge 8 <$ ad, 10 <$ imm, 8 $ ad, 2 $ imm, i $ juv,

1 ? imm ; Nagasi i ^ ad, i ^ imm, i ? ad. i ^ ad, i $ ad to OUM ; i <J imm to AMNH.
Weight 7 ad 70-78-5 (73-9) ; 9 <J imm 61-75 (71-8) ; 7 ? ad 65-5-84-5 (73-5) ;

2 $ imm 69-5, 70 ; i $ juv 66-5. Wing 8 <$ ad 135-5-144 (139-4) '> IO c? i mm I 3'5- ]C 37
(133-4) ; 9 $ ad 133-140 (137-3) ; 2 $ imm 132, 132 ; i$ juv 129. Bill (from front of
nostril) 8 <$ ad 12-5-14 (13-3) ; 9 <? imm 13-14 (13-3) ; 7 $ ad 12-13-5 (13-1) ; 2 ?
imm 12-5, 13 ; i ? juv 13.

LAND BIRDS OF GUADALCANAL 51

Iris bright lemon-yellow ; imm bright lemon-yellow to pale yellowish grey ; juv
pale grey-brown. Bill black ; juv gape flesh-coloured. Foot black ; juv dark slate.

Immature individuals have juvenile wings and tails, and sometimes dull irides.
They are otherwise indistinguishable from adult females, except that the barring is
occasionally black-and-white, as in adults of the wholly hen-feathered races gracilis
and lineata, instead of having grey inner halves to the dark bars. Two immature
specimens each of gracilis (Bradley & Wolff, 1956 : 104, nos. 125 and 131) and lineata,
and one of sublineata, have the white bars markedly wider than in corresponding
adult females. Before the immaturity of the heavily-barred males was recognized,
makirae was erroneously mentioned as an unusually variable race (Galbraith, 1956 :
162).

NOTES. There seems no reason to suppose that gracilis of Rennell is more closely
related to lineata of Australia than to the Melanesian races, despite Mayr (19316 : 18 ;
1955 : 14). They are linked by their complete hen-feathering, and by the associated
" juvenility " of their barring, without any grey (see p. 50). The races makirae,
malaitae and sublineata are intermediate in degree of dimorphism, while some immatures
of the former show black-and-white barring. In other respects, the small dark gracilis
is more similar to the Solomons races than to the large pale lineata. In fact, Mayr
(1931^ : 7 ; followed by Bradley & Wolff, 1956 : 115) elsewhere lists gracilis as more
closely related to the races in the Solomons.

(M. 90) Coracina papuensis elegans (Ramsay, 1881)

REFERENCES. Rothschild & Hartert, 1916:289; Mayr, 19317 :i6; 1955: 13.
See Ripley, 1941 ; Voous & van Marie, 1949.

RANGE. Guadalcanal, Russell Is., New Georgia Group (mainly lowlands).

SPECIMENS. Guadalcanal : Tenaru 9 ^ ad, 4 $ ad, i ? imm, i ? juv : Betilonga
i <J ad. i < ad, i ? ad to OUM.

Weight 10 $ ad 56-64 (61-2) ; 4 $ ad 58-5-65 (61-1). Wing 9 $ ad 130-137 (132-3) ;
4 $ ad 127-136 (131-8). Bill (from front of nostril) 7 $ ad 16-18 (17-4) ; 3 ? ad
16-17-5 (17-0).

Iris very dark brown. Bill black ; juv maxilla blackish, mandible dark horn with
blackish tip, gape yellow. Foot grey, scutes black ; juv pale fuscous grey, claws
black and dull white, soles very pale yellow.

Apart from its juvenile wing and tail, the immature differs from adults in its duller
black mask and whiter breast.

The Betilonga specimen was taken " in native gardens near village ", not in forest,
despite Cain & Galbraith (1956 : 267).

(M. 91) Coracina caledonica amadonis Cain & Galbraith, 1955

REFERENCES. Cain & Galbraith, 1955 : 90. See Mayr, 1955 : 15. See Ripley,
1941 ; Voous & van Marie, 1949.

RANGE. Guadalcanal (mountains).

SPECIMENS. Guadalcanal : Turipava 2 ^ ad, i $ ad. ig ad to AMNH.

Weight 2 (J ad 133-5, 136 ; i $ ad 165. Wing 2 3 ad 176, 180 ; i $ ad 173. Bill 2 $
ad 35, 36-5 ; i $ ad 38-5.

52 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

7m dark red-brown. Bill black. Foot black.

Theholotype (adult female, expedition no. 459) is BM(NH) reg. no. 1959.21.434.

NOTES. This is the most distinct race in the species.

(M. 92) Coracina novaehollandiae melanops (Latham, 1801)

REFERENCES. Cat. Birds B.M. 4 : 30 ; Keast, 1958 : 253. See Ripley, 1941 ;
Voous & van Marie, 1949.

RANGE. Southern and eastern Australia ; migrant, wintering in eastern New
Guinea region, Bismarcks, and occasionally Solomons (coasts).

NOTES. Recorded from the northern Solomons only, Nissan (Mayr, 1955 : 17)
and Bougainville (Baker, 1948 : 17) both records in August.

(M. 93) Turdus poliocephalus sladeni Cain & Galbraith, 1955

REFERENCES. Cain & Galbraith, 1955 : 92. See Mayr, 19316 : 21 ; 1941^ : 6.

RANGE. Guadalcanal (mountains).

SPECIMENS. Guadalcanal : Turipava i ^ ad, 2 $ subad, i ? subad. i <$ subad to
AMNH.

Weighty <$ ad & subad 57-67-5 (61-9). Wing 4 ^ ad & subad 107-111-5 (109-6).
Tail 4 $ ad & subad 82-87 (84-5). Bill 4 ^ ad & subad 23-24-5 (23-7).

Iris dark brown. Eyelid yellow. Bill yellow. Foot yellow, shins and claws fuscous-
tinged.

The holotype (adult male, expedition no. 544) is BM(NH) reg. no. 1959.21.553.

(M. 95) Zoothera margaretae turipavae Cain & Galbraith, 1955

REFERENCES. Cain & Galbraith, 1955 : 92. See Mayr, 1955 : 17.
RANGE. Guadalcanal (mountains).
SPECIMENS. Guadalcanal : Turipava i $ ad.
Weight 53-5. Wing 90-5. Tail 63-5. Bill 24-5.

Iris very dark brown. Bill black, base greyish, gape dull orange. Foot greyish
fuscous, joints paler and greyer, claws ivory.

The unique holotype (expedition no. 485) is BM(NH) reg. no. 1959.21.557.

(M. 95) Zoothera margaretae margaretae (Mayr, 1935)

REFERENCES. Mayr, 1936 : 14. See Mayr, 1955 : 17.
RANGE. San Cristoval (mountains).

(G) Cichlornis whitneyi turipavae Cain & Galbraith, 1955

REFERENCES. Cain & Galbraith, 1955 : 91. See Mayr, 19450 : 191 ; Gilliard,
19606.

RANGE. Guadalcanal (mountains).

SPECIMENS. Guadalcanal : Turipava i <$ ad.

Weight 36. Wing 65-5.

Iris dark brown. Bill blackish, mandible streaked with whitish, gape dull yellow.
Foot fuscous brown.

LAND BIRDS OF GUADALCANAL 53

The unique holotype (expedition no. 448) is BM(NH) reg. no. 1959.21.558.

NOTES. In the scattered group of babbler-warblers to which Cichlornis appears
to belong (Mayr, 19330 : 3 ; 19440 ' I 5^ ; 1955 : 18 ; Cain & Galbraith, 1955 : 91),
two subgroups (whether phyletic or adaptive) are discernible. On the one hand there
are pale-coloured, relatively long-tailed forms, somewhat resembling Megalurus, in
open and more or less arid habitats : Buetikoferella bivittata of Timor, Eremiornis
carteri of Australia and Megalurulus marei of New Caledonia. On the other, there is
an even more scattered subgroup of dark-coloured, shorter-tailed forms, with less
depressed nasal opercula, occupying the dense wet forest on mountainous islands :
Cichlornis w. whitneyi on Santo, C. w. turipavae on Guadalcanal, C. grosvenori and
Ortygocichla rubiginosa on New Britain, and 0. rufa on Viti Levu. With the recent
discovery of Cichlornis grosvenori Gilliard (19606), the mountains of New Britain
become the only locality known to support two species of this group.

(M. 96) Vitia par ens Mayr, 1935

REFERENCES. Mayr, 1936 : 15.

RANGE. San Cristoval (mountains).

SPECIMENS. San Cristoval : Nagasi i 9 ad, i 9 imm.

Weight i $ ad 14 ; i 9 imm 13. Wing i 9 ad 55. Tail i 9 imm 42. Bill i 9 imm

Iris brown ; imm dark grey-brown. Bill maxilla black, mandible grey with whitish
base, gape pale grey, dull yellow internally ; imm mandible with whitish stripe and
yellowish tip, gape dull yellow. Foot dull yellow, toes and fronts of shanks washed
dark brown.

The immature specimen, whose wings are not fully-grown, has the crown duller
and less rufous than the adult, but retains none of the juvenile plumage described by
Mayr (1936).

(M. 97) Acrocephalus stentoreus cervinus de Vis, 1897

REFERENCES. Mayr, 1948 : 209 ; 1955 : 18.

RANGE. Sumba, northern Queensland, scattered localities in New Guinea, New
Britain, Solomons Bougainville, Gijunabena near Ysabel, Guadalcanal (lowlands).

SPECIMENS. Guadalcanal : Tenaru 3 ^ ad, i 9 ad.

Weight 3 c? ad 17-5-18 (17-8) ; i 9 ad 16. Wing 3 <$ ad 68-5-71-5 (69-8) ; i 9 ad
65-5. Tail 3 (J ad 63-64-5 (63-5) ; i 9 ad 62. Bill 3 $ ad 20-22 (21-2) ; i 9 ad 20-5.

Iris yellowish brown. Bill maxilla blackish with yellowish horn edges, mandible
yellowish horn with greyer tip, gape flesh-colour. Foot brownish grey, shanks browner.

(M. 99) Phylloscopus trivirgatus becki Hartert, 1929

REFERENCES. Hartert, 1929 : 13 ; Mayr, 1936 : 17.
RANGE. Ysabel, Guadalcanal, Malaita (mountains).
SPECIMENS. Guadalcanal : Turipava 2 $ ad.

Weight 2 $ ad 10, 11-5. Wing i $ ad 59. Tail i $ ad 44-5. Bitt i ad 14.
Iris dark brown. Bitt maxilla blackish, mandible horn, paler basally. Foot pale
blue-grey, tinged with fuscous, sole dull yellow,

54 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

NOTES. The remark (Mayr, 1955 : 19) that the species is to be expected from
Guadalcanal is of course a lapse, since this was one of the first birds to be discovered
in the mist-forest of Guadalcanal.

(M. 99) Phylloscopus trivirgatus makirensis Mayr, 1935

REFERENCES. Mayr, 1936 : 18.

RANGE. San Cristoval (mountains).

SPECIMENS. San Cristoval : Nagasi 2 ^ ad, 2 $ ad.

Weight 2 ^ ad 9, 9-5 ; 2 $ ad 9, 9-5. Wing 2 ^ ad 50-5, 53 ; 2 $ ad 48-5, 49. Tail

2 (J ad 37-5, 39-5 ; i $ ad 35. Bill i $ ad 13-5 ; 2 $ ad 13, 13-5.

Iris dark brown. Bill maxilla black ($ with yellowish edge), mandible greyish or
greenish horn with dark brown subterminal patch, gape dull yellow to pale fuscous
with yellow tinge. Foot bluish grey, <$ shank blotched with blackish.

(M. 100) Rhipidura rufifrons rufofronta Ramsay, 1879

REFERENCES. Mayr, 1931^ : 19. See Mayr & Moynihan, 1946.
RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Tsea i <$ ad ; Betilonga 7 J ad, 3 $ ad, 2 ? ad. i J
ad to OUM.

Weight 8 $ ad 11-12 (n-6) ; 4 ? ad 10-5-14-5 (11-9). Wing 8 $ ad 70-74-5 (72-7) ;

3 $ ad 68-72 (69-3). Tail 5 3 ad 80-5-84 (82-4) ; 2 $ ad 80-5, 84-5. Bill 8 ^ ad 13-5-
15-5 (14-3) ; 2 $ ad 14, 14-5.

Iris dark brown. Bill black, base of mandible whitish to pale grey or horn. Foot
grey, brownish to bluish.

(M. 100) Rhipidura rufifrons russata Tristram, 1879

REFERENCES. Mayr, 1931^ : 18. See Mayr & Moynihan, 1946.

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Manewiriwiri i ^ ad, i ^ imm, i ? imm ; Goge 9 <$
ad, 10 <$ imm, 7 $ ad, 2 $ imm, 3 ? imm ; Nagasi 2 <$ ad, i J imm, i $ ad, i ? imm.
i c^ ad to OUM.

Weight 13 c? ad 8-5-10 (9-1) ; 13 ^ imm 8-5-10 (9-2) ; 9 $ ad 8-9-5 (8-5) ; 2 $
imm 8, 8-5. Wing n ^ ad 63-68 (65-2) ; n $ imm 61-68 (63-0) ; 8 $ ad 60-5-65-5
(62-4) ; 2 $ imm 61, 61. Tail 6 ^ ad 71-5-75 (73-3) ; 6 ^ imm 67-74 (77) i 3 ? ad
7 -5-73 (71-8) ; i ? imm 71-5. 5i# n < ad 12-5-14-5 (13-5) ; n < imm 12-5-14
(13-5) ; 8 ? ad 13-14 (13-3) ; 2 $ imm 13-5, 14.

7ns very dark brown. Bill black, base of mandible ivory, sometimes horn-coloured
or pinkish ; imm mandible occasionally purplish with orange gape and base, only
tip blackish. Foot dark grey, usually somewhat fuscous, browner towards upper
shank.

In this worn and badly-prepared series, the immature birds are separated only by
the rufous edges of their wing-feathers, since their gorgets do not seem to be uni-
formly duller nor their plumage softer. The development of the gonads tends to
support this separation, since there is little overlap between the adult and immature
series : but see under ngiensis below,

5.5

(M. 100) Rhipidura rufifrons ugiensis Mayr, 1931

REFERENCES. Mayr, 1931^ : 19. See Mayr & Moynihan, 1946.

RANGE. Ugi.

SPECIMENS. Ugi 3 <$ ad, 4 imm, 4 $ ad, 2 $ imm. i $ ad to OUM.

Weight 3 (J ad 10-5-11-5 (n-o) ; 4 < imm 10-5-12 (10-9) ; 4 $ ad 10-10-5 ( IO '4) '>
2 $ imm 10-5, 10-5. Wing i ^ ad 71 ; 3 <J imm 64-74 (68-3) ; 3 $ ad 65-67 (66-0) ;
2 $ imm 62, 64. Ta7 i $ ad 72-5 ; 3 $ imm 76-80 (77-7) ; 3 $ ad 72-75 (73-5) ;
2 $ imm 73-5, 74-5. JBt# 2 ^ ad 14-5, 14-5 ; 4 <? imm 14-15 (14-2) ; 4 $ ad 13-5-14-5
(14-0) ; 2 $ imm 13-5, 14.

7m very dark brown. Bill black, base of mandible ivory, sometimes grey or horn-
coloured. Foot dark grey, somewhat fuscous, brownish towards upper shank.

In this series, shorter and worse-preserved than that of russata, the separation of
immatures from adults is more dubious : there is more overlap in gonadial develop-
ment, and the distributions of rufous edges to the wings, duller black throats, and
whitish chins (the two last mentioned as characteristic of juveniles in another black-
throated race, melanolaema, by Mayr, 19312) do not coincide.

(M. 100) Rhipidura rufifrons kuperi Mayr, 1931

REFERENCES. Mayr, 19312 : 18. See Mayr & Moynihan, 1946.
RANGE. Santa Ana, Santa Catalina.

(M. 101) Rhipidura drownei ocularis Mayr, 1931

REFERENCES. Mayr, 19312 : 12.

RANGE. Guadalcanal (mountains).

SPECIMENS. Guadalcanal : Turipava 4 ^ ad, i ^ juv, 3 $ ad, 2 ? ad. I ? ad to
OUM.

Weight 4 ad 11-5-12-5 (12-0) ; 3 $ ad 10-10-5 (10-3). Wing 4 < ad 72-5-77-5
(75-1) ; i $ juv 72 ; 3 $ ad 68-5-70 (69-2). Tail 3 $ ad 81-82-5 (81-5) ; i c? juv 72 ;
2 $ ad 73, 75. Bill 4 ad 14-15 (14-2) ; i <$ juv 15 ; 3 $ ad 13-13-5 (13-3).

7ns dark brown, sometimes greyish. Bill maxilla black, mandible white (sometimes
tinged with pink or mauve), tip blackish. Foot dark fuscous or brownish grey, sole
pale yellow.

(M. 102) Rhipidura tenebrosa Ramsay, 1881

REFERENCES. Mayr, 19310 : 12.

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Goge 3 < ad, 2 $ ad, 4 ? ad ; Nagasi 2 ad, 2 $ ad.
i ? ad to OUM.

Weight 5 J ad 18-5-19-5 (18-9) ; 4 $ ad 15-5-16-5 (15-9)- Wing 4 ad 84-90
(85-9) ; 4 $ ad 75-79-5 (777). Tail 4 ad 88-95-5 (92-1) ; 2 $ ad 86, 86-5. Bill 4
ad 16-18 (17-0) ; 4 $ ad 16.

Iris dark or very dark brown. Bill maxilla black, mandible mainly whitish or pale
grey (sometimes mauve-tinged) with blackish tip and edges of variable extent. Foot
fuscous grey with bluish grey patches.

56 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(M. 103) Rhipidura fuliginosa cf. brenchleyi Sharpe, 1879

REFERENCES. Mayr, 19310 : 14.

RANGE. San Cristoval (mountains).

NOTES. It is unfortunate that this is one of the high-altitude species missed by the
expedition on San Cristoval, since the only known series is too worn to show the
expected differences from brenchleyi of the New Hebrides (Mayr, 1931).

Reversion by Mayr (1955 : 20) to the name R. flabellifera (Gmelin, 1789) instead of
R. fuliginosa (Span-man, 1787), is clearly a lapse.

(M. 105) Rhipidura cockerelli cockerelli (Ramsay, 1879)

REFERENCES. Mayr, 19310 : 4.

RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Betilonga 8 ^ ad, 3 ^ imm, 4 $ ad, i 9 imm, 2 ? ad.
i <$ ad to OUM.

Weight 8 $ ad 16-5-18-5 (17-6) ; 3 < imm 15-16-5 (15-7) ; 4 ? ad 17-18 (17-4) ;
i $ imm 15. Wing 8 ^ ad 87-93 (90-1) ; 3 $ imm 84-86 (84-8) ; 4 ? ad 81-86-5 (83-5) ;
i $ imm 78-5. Tail 7 <J ad 80-85 (82-7) ; 3 $ imm 81-84 (82-0) ; 4 $ ad 79-82-5
(80-5) ; i $ imm 75-5. Bill 8 <$ ad 18-20-5 (19-0) ; 3 $ imm 18-5-20 (19-2) ; 3 $
ad 18-5-20 (19-5) ; i $ imm 18.

7ns dark brown. Bill black. Foot black, occasionally dark fuscous or grey.

NOTES. Though the fact has apparently never been remarked (except by Galbraith
1956 : 193), this species is clearly the geographical representative of the widespread
and highly polytypic R. rufiventris. This makes it less surprising (Mayr, 1955 : 21)
that the latter is absent from the Solomons. R. rufiventris shows tendencies (especially
in kordensis of Biak) towards melanism, white spotting of the breast, and loss of all
phaeomelanin, which are carried to extremes in R. cockerelli. Though much the
largest form in the superspecies, R. cockerelli coultasi of Malaita approaches certain
races of R. rufiventris in its white throat, reduced white wing-patch, and grey back.
In behaviour, R. cockerelli is very unlike the smaller species of Rhipidura (Cain &
Galbraith, 1956 : 271), though this is not apparent from other accounts (Mayr,
1945^ ; Virtue, 1947 ; Sibley, 1951) ; while R. rufiventris apparently approaches the
same Myiagra-\ike mode of feeding (Mayr & Rand, 1937 : 159 ; Mayr & Schauensee,
1939 : 32). In Malaysia, the rufiventris superspecies seems to be represented by a
doublet : R, perlata and R. euryura.

(M. 106) Rhipidura leucophrys melaleuca (Quoy & Gaimard, 1830)

REFERENCES. Mayr, 19310 : 2.

RANGE. Moluccas, New Guinea region, Northern Melanesia.

SPECIMENS. Guadalcanal : Tenaru 2 ^ ad, i $ ad, i $ imm ; Betilonga 2 < ad,
i $ ad ; Nala i ^ ad. San Cristoval : Manewiriwiri i 9 imm ; Goge i <$ ad, i <$ imm,
i 9 ad. Ugi i $ ad, 3 <$ imm, 2 $ ad, i $ imm. i ^ ad to OUM.

Weight 7 <? ad 3i'5~37 (34'4) ', 4 c? imm 3-34 (32-4) ; 5 $ ad 27-36 (31-7) ; 3 $
imm 29-30 (29-3). Wing 7 ad 96-102 (99-2) ; 4 $ imm 97-100 (98-6) ; 5 $ ad
96-102 (98-5) ; 2 $ imm 95-5, 99. Tail 6 ^ ad 93-101 (97-4) ; 2 < imm 94, 95-5 ;

LAND BIRDS OF GUADALCANAL 57

4 9 ad 95-100-5 (977) ; 3 9 imm 91-5-98-5 (94'3)- Bill 6 3 ad 20-5-23 (21-5) ; 4
imm 20-5-21-5 (21-1) ; 3 9 ad 21-22 (21-3) ; 3 $ imm 20-21 (20-3).

Iris very dark brown. Bill black ; some imm mandibles partly grey, gape dull
flesh. Foot black, shanks somewhat fuscous.

(M. 108) Monarcha castaneiventris castaneiventris Verreaux, 1858

REFERENCES. Ramsay, 18790 : 79 (rufocastanea) . See Mayr, 1942 : 81.

RANGE. Choiseul, Ysabel, Florida Is., Guadalcanal, Malaita.

SPECIMENS. Guadalcanal : Tenaru 2 < imm, i 9 imm, i ? ad ; Betilonga n $ ad,

2 < imm, 2 $ ad, 3 9 imm, i ? ad ; Turipava i $ ad, i 9 ad. i < ad to OUM.

Weight 12 c? ad 22-5-27 (24-6) ; 4 ^ imm 25-28-5 (26-4) ; 3 $ ad 22-5-25 (23-8) ;

4 9 imm 21-5-24 (22-8). Wing 10 < ad 82-5-88 (84-7) ; 4 <J imm 81-5-83-5 (82-5) ;

3 $ ad 83-84-5 (83-8) ; 4 9 imm 79-82 (80-7). Tail 10 $ ad 66-75-5 (70-8) ; 4 $ imm
67-69 (67-7) ; 3 9 ad 67-5-73-5 (69-8) ; 4 $ imm 67-69-5 (68-0). Bill 12 $ ad 18-5-21
(20-0) ; 4 c imm 20-21 (20-6) ; 3 $ ad 19-5-21 (20-0) ; 4 9 imm 19-5-21 (20-0).

7m dark brown. Bill silvery blue-grey, occasionally streaked with black, some-
times paler towards tip and edges, tip blackish. Foot blue-grey.

Immature and adult birds seem to be much less distinct in this race than in
megarhynchus, and are here separated tentatively. Both in this series and in 5 males
and 3 females in the BM(NH), all the specimens are intermediate, in colour and texture
of wing and tail and shape of rectrices, between adults and immatures of megarhynchus.
Some of the males which appear the most juvenile had enlarged testes. There is some
overlap between the sexes, but the females average much less glossy above, especially
on rump, wings and tail.

(M. 108) Monarcha castaneiventris megarhynchus Rothschild & Hartert, 1908

REFERENCES. Rothschild & Hartert, 19086 : 363. See Mayr, 1942 : 81.

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Goge 20 <$ ad, 5 $ imm, 13 9 ad, 6 9 imm, 3 ? ad,
9 ? imm. i $ ad to OUM.

Weight 20 <$ ad 25-28-5 (26-2) ; 3 <$ imm 25-27-5 (26-2) ; n $ ad 22-33 (25-0) ;
4$ imm 23-25-5 (24-2). Wing 19 $ ad 83-91-5 (87-6) ; 5 <$ imm 80-88-5 (82-9) ;
13 9 ad 81-86-5 (83-8) ; 6 9 imm 78-81-5 (79-6). Tail 18 < ad 73-5-81-5 (77-6) ;

5 < imm 73-5-80 (76-0) ; 13 9 ad 70-78 (74-9) ; 6 9 imm 70-74 (71-7). Bill 17 $ ad
23-25-5 (24-3) ; 4 $ imm 23-5-24 (23-7) ; 12 9 ad 22-24 (23-0) ; 6 9 imm 22-24
(23-0).

Iris dark to very dark brown. Bill pale blue-grey, whitish distally and sometimes
towards edges, tip and sometimes edges blackish. Foot blue-grey, often dark especi-
ally on toes.

Immatures are readily distinguished from adults by their softer and browner wings
and tails, roundedly-pointed rectrices, and blackish brown upper wing-coverts (not
dull black with a slight blue gloss at the edges). Their bills were adult in coloration,
but one male, with enlarged testes (noted in the field as " ! somewhat autolysed, but
recognisable ") and adult plumage, had the bill black with pinkish- white base and
pale yellow gape. There is no appreciable sexual dimorphism in this series.

58 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(M. 108) Monarcha castaneiventris ugiensis (Ramsay, 1881)

REFERENCES. Ramsay, 1882^ : 128. See Mayr, 1942 : 81.

RANGE. Ugi, Three Sisters, Santa Ana.

SPECIMENS. Ugi 9 $ ad, 2 $ imm, 7 $ ad, 2 $ imm, i ? ad, 2 ? imm. i <$ ad to
OUM.

Weight 9 ^ ad 28-5-33-5 (31-6) ; 2 ^ imm 26-5, 27-5 ; 7 $ ad 28-30-5 (28-9) ; 2 $
imm 27-5, 28. Wing 8 ^ ad 89-95-5 (92-2) ; 2 $ imm 85-5, 86 ; 5 $ ad 85-88 (86-1) ;
2 $ imm 81-5, 82. Tail 9 < ad 77-83-5 (80-5) ; 2 ^ imm 78-5, 79 ; 6 ? ad 73-5-76-5
(75-4) ; 2 $ imm 70-5, 74-5. Bill 8 <J ad 25-26-5 (25-9) ; 2 ^ imm 25, 26 ; 6 $ ad
24-26 (24-9) ; 2 ? imm 25, 25-5.

Iris very dark brown. J5z7/ pale bluish grey, whitish distally, tip blackish ; imm
more or less overlaid with black especially distally, gape dull yellow. Foot bluish
grey, shank paler ; imm more blackish.

The belly and under tail-coverts average distinctly duller in the females than the
males, though there is no sexual difference in those areas which are black in other races.
The belly is slightly glossy in all the males, but in females it is sometimes decidedly
brownish with shafts whitish to the tips. Among the immature birds, the two females
seem older than the two males and two unsexed specimens, in having had the bill of
adult coloration (except that the gape of one was " greyish flesh ") rather than black
with dull yellow gape. However, they are less glossy in plumage, and have the white
shafts of the belly-feathers extended and becoming tawny at the tips, with tawny
tips to the feathers themselves, especially near the crissum. These tips are extensive
in the pale-gaped specimen so that the whole lower belly appears dull tawny. It
seems probable that tawny-fringed belly-feathers are characteristic of immature
females. Two rufous-bellied specimens have been reported (Davidson, 1934 : 195 ;
Mayr, 1942 : 81) . In the latter at least, the appearances is as in adults of megarhynchus ,
not at all as in our immature females (P. Vaurie, in litt.}.

(M. 109) Monarcha barbatus barbatus Ramsay, 1879

REFERENCES. Ramsay, 18790 : 80 ; Rothschild & Hartert, 19016 : 182 ; 1905 :
262 ; 19080 : 357 (brodei) ; Mayr, 19317 : 23.

RANGE. Bougainville, Choiseul, Ysabel, Florida Is., Guadalcanal.

SPECIMENS. Guadalcanal : Betilonga 6 J ad, 6 $ ad, 2 $ imm, 2 ? imm. i $
ad, i ? imm to OUM.

Weight 6 $ ad 20-23 (21-5) ; 7 $ ad 19-21-5 (20-5) ; 2 $ imm 19-5, 20. Wing 4 <
ad 79-83 (80-9) ; 6 $ ad 75~79'5 (77'3) ; 2 ? imm 7^, 7 8 '5- Tail 5 J ad 70-73 (71-6) ;
5 $ ad 64-73 (67-4) ; i $ imm 69. Bill 4 ^ ad 17-17-5 (17-1) ; 5 ? ad 16-5-17-5 (16-7) ;
2 $ imm 16, 16-5.

Iris dark brown. Bill silvery blue-grey, whitish distally and along edges, tip and
occasionally edges blackish ; imm more or less overlaid with black. Foot blue-grey,
sole yellowish.

All but one of these specimens, and all those in the BM(NH), are either clearly
adult or in the pale rufous- washed immature plumage. However, BM(NH) no.
1959 . 21 . 799 appears to be an immature female in an advanced plumage. It resembles

LAND BIRDS OF GUADALCANAL 59

immatures of M. vidua in being generally similar to adults, but with wing and tail
somewhat brownish and of juvenile shape and texture, the black throat-patch less
scaly, and the gloss duller. Unfortunately, the puzzling juvenile plumages of this
species have not yet been reviewed (see Mayr, 1931/1 23).

NOTES. At first sight, the black-tipped white outer rectrices of M. b. malaitae,
reversing the common condition in the verticalis superspecies, suggest a marked
genetic difference. However, close inspection, especially of juveniles, shows that the
change is produced by a distalwards shift of the tail pattern, with enlargement of
the normally small white basal area. This confirms the suggestion of its geographical
distribution, that the character is readily acquired : it is found on Buru, Kei, Biak,
the Admiralty group and Malaita (loricatus, leucurus, brehmii, coultasi, infelix and
malaitae). Clearly malaitae is properly associated with barbatus. On the other hand,
it does not seem appropriate to separate the races of the New Georgia Group as a
distinct species, R. browni, despite Mayr (1955 : 27). This arrangement reflects the
geography rather than the affinities of the forms concerned, since they differ from
barbatus in relatively trivial characters and do not form a homogeneous group. They
are large and somewhat melanic, but browni -f meeki is black-breasted, nigrotectus
is black-winged, and ganongae is closest in pattern to barbatus.

(M. 109) Monarcha vidua vidua (Tristram, 1879)

REFERENCES. Ramsay, 1880 : 468 (melanocephalus) .

RANGE. San Cristoval, Santa Ana.

SPECIMENS. San Cristoval : Goge 8 ad, 7 ^ imm, i $ juv, 13 $ ad, 4 $ imm,
8 ? ad, 4 ? imm, i ? juv ; Nagasi 5 ^ ad, 2 ^ imm, 5 $ ad, 2 $ imm, i ? ad, i ? imm.
i c? ad, i 9 imm to OUM ; i ^ imm to AMNH.

Weight ii c? ad 16-19-5 (17-4) ; 9 $ imm 15-5-17-5 (16-7) ; i $ juv 15 ; 19 $ ad
16-18 (16-8) ; 6 9 imm 15-5-18 (16-7). Wing 13 $ ad 74-83 (78-7) ; 7 <J imm 74-79
(76-0) ; i (? juv 73-5 ; 17 $ ad 72-5-79 (75-2) ; 4 9 imm 70-73 (71-6). Tail 12 $ ad
68-76-5 (71-6) ; 7 $ imm 68-78-5 (72-9) ; i ^ juv 69 ; 13 $ ad 67-72 (70-8) ; 5 9
imm 65-5-71 (68-7). Bill 13 $ ad 16-17 (16-2) ; 7 $ imm 15-5-16-5 (16-1) ; i $
juv 16 ; 16 9 ad 15-5-16-5 (16-0) ; 6 9 imm 15-5-16-5 (15-9)-

Iris very dark brown ; juv dark grey. Bill blue-grey, usually pale especially
distally, tip and more or less of edges black, occasional black patches in front of nostrils
(especially 9) ; imm range from ad to juv condition ; juv black, gape flesh-coloured,
yellow internally. Foot blue-grey, sometimes dark especially on toes ; juv paler.

Even the juveniles are generally similar to the adults in plumage coloration, not
pale and rufous as in M. barbatus. They differ from adults in their soft parts, and in
softer plumage, differently-shaped and browner wings and tails, brownish mantle-
feathers with little black at the tips, and unspecialized dull black throat-feathers.
The immatures have juvenile wings and tails, but otherwise resemble adults except
that their throat-feathers are less specialized, their greater upper wing-coverts have
the black outer edges obsolescent, and their gloss is greener and somewhat duller.
There are no specimens of vidua or squamulatus in a pale and rufous plumage, in the
AMNH (P. Vaurie, in litt.) or BM(NH).

60 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

NOTES. It is remarkable that the uncharacteristic non-adult plumages of vidua
and squamulatus have not previously been pointed out. Rothschild & Hartert
(19086 : 363) suggested it in their remark that the " females " of vidua did not differ
from the males ; but their views on the existence of sexual dimorphism in the pied
monarchs fluctuated (19016 : 183 ; igoic : 374 ; Hartert, 1908 : 107).

Even supposing a constant minority of immature barbatus to show an advanced
plumage like that of vidua, it seems likely that juveniles of that and most other pied
monarchs are always pale and rufous. Thus the juvenile plumage presents a striking
difference between barbatus and vidua. In addition, these forms differ in adult pattern
as radically as many of the representative forms currently separated as species. Indeed
they have little in common, and vidua agrees rather with other members of the super-
species in having the nape and rump white and the sides of the throat black. Despite
Mayr (1955 : 27) , it does not seem possible to unite vidua with barbatus while they are
specifically separated from their relatives in the Bismarcks and beyond.

(M. 109) Monarcha vidua squamulatus (Tristram, 1882)

REFERENCES. Tristram, 1882 : 136.

RANGE. Ugi.

SPECIMENS. Ugi 10 $ ad, 2 ^ imm, i <$ juv, 7 $ ad, 2 $ imm, i $ juv, 4 ? ad,
2 ? imm, 2 ? juv. i ? ad, i ? imm to OUM ; i $ imm to AMNH.

Weight 10 $ ad 16-19-5 (17-7) ; 2 <$ imm 18, 19 ; i <$ juv 19-5 ; 7 $ ad 15-5-18
(16-9) ; 2 $ imm 16, 17-5 ; i ? juv 18. Wing 10 <$ ad 73-5-79-5 (76-2) ; 2 <$ imm
73'5. 75 ; i J juv 75 ; 7 ? ad 73-76 (74-2) ; 2 ? imm 75-5, 76; i $ juv 69-5. Tail

8 $ ad 65-5-70 (68-3) ; i ^ imm 69 ; i <$ juv 69 ; 7 $ ad 64-5-68-5 (67-3) ; 2 $ imm
67, 68 ; i $ juv 65-5. Bill 9 $ ad 16-5-17 (16-7) ; 2 < imm 17, 17 ; i ^ juv 17 ; 7 $
ad 16-17 (16-6) ; 2 ? imm 17, 17 ; i $ juv 16-5.

Iris very dark brown (including juv). Bill bluish grey, paler distally, tip and
occasionally distal edges black ; imm between ad and juv conditions ; juv black,
base of mandible grey or horn, gape dull yellow. Foot bluish grey ; juv sometimes
paler.

This race is a melanistic derivative of M. vidua, and shows the same difference
between plumages. The melanization affects the breast, wingpatch and nape, the
rump slightly, but not the tail tips. It is very variable in degree, and the most highly
melanic specimens have black breasts and upper wing-coverts with narrow white
shaft-streaks, no white collars, and somewhat narrowed white rump-patches. This
variation is shown in all plumages.

(M. no) Myiagra ferrocyanea ferrocyanea Ramsay, 1879

REFERENCES. Ramsay, 18790 : 78 & 79 (-{-pallida} ; Rothschild & Hartert,
1905 : 261 ; 19080 : 357 ; Mathews, 1928 : 373 ; Mayr, I93I/ : 24.
RANGE. Choiseul, Ysabel, Florida Is., Guadalcanal.
SPECIMENS. Guadalcanal : Tenaru 3 ^ ad, i <$ imm, i $ ad ; Betilonga 17 <$ ad,

9 $ ad, 2 $ imm, i ? imm. i ^ ad, i $ ad to OUM.

Weight 21 J ad 11-5-14 (12-9) ; i < imm 12 ; 10 ^ ad 10-14-5 (12-6) ; 2 $ imm

LAND BIRDS OF GUADALCANAL 61

12-5, 13. Wing 19 <$ ad 63-71 (68-6) ; i <$ imm 67 ; 10 $ ad 64-68-5 (66-3) ; 2 ?
imm 67, 67. Tail 19 ^ ad 58-5-62-5 (60-6) ; I <$ imm 60-5 ; 8 $ ad 58-61 (597) : 2 $
imm 59-5, 59-5. Bill 18 <$ ad 15-5-17-5 (16-3) ; 8 ? ad 14-5-16-5 (15-9) ; i $ imm 16-5.

Iris dark to very dark brown. Bill silvery grey-blue, tip and edges black, maxilla
sometimes streaked with black, especially along culmen and in front of nostril ; imm
from more or less black with greyish base and yellow gape, to ad condition. Foot
dark grey to black.

The immatures are like adult females in body-plumage, but have softer wings and
tails with more pointed rectrices, bills decidedly brownish in the dried skin, and
wing-feathers with darker and browner centres more sharply marked off from the
broader fawn edges. The adults vary considerably in shape, texture and colour of
wing and tail, and the females in amount of buff on the underparts ; but no distinct
phases are separable. The adult males, in somewhat worn plumage, are less purple
in body-gloss than fresher BM(NH) specimens.

(M. no) Myiagra cervinicauda Tristram, 1879

REFERENCES. Tristram, 1879 : 439 ; Ramsay, i882a : 726 ; Rothschild &
Hartert, 19080 : 357 ; 19086 : 363.

RANGE. San Cristoval, Ugi, Santa Ana.

SPECIMENS. San Cristoval : Manewiriwiri 2 $ ad, i $ ad, i $ imm ; Goge 2 <$
ad, 5 subad, 2 <$ imm, i $ ad, 2 $ imm, i ? imm ; Nagasi i <$ ad. Ugi n ^ ad, i ^
imm, 5 $ ad, i $ imm, 2 ? imm. i ^ ad, i $ ad to OUM.

Weight San Cristoval 5 <$ ad 10-5-12 (10-9) ; 5 ^ subad 9-5-11 (10-6) ; 2 <$ imm
10-5, 10-5 ; i $ ad 13 ; 3 $ imm 10-11 (10-5) ; Ugi 12 <$ ad 11-13 ( II- 7) '> I 3 i mm TI i
5 $ ad 10-12 (n-i) ; i ? imm 10-5. Wing San Cristoval 4 $ ad 62-5-64-5 (63-6) ;
4 <$ subad 64-66-5 (64-9) ; 2 <$ imm 62-5, 63 ; i $ ad 62 ; 3 ? imm 58-62-5 (60-0) ;
Ugi ii (J ad 63-5-67 (65-7) ; i $ imm 64-5 ; 5 $ ad 61-5-64 (62-7) ; i $ imm 61. Tail
San Cristoval 4 $ ad 56-59 (57-7) ; 3 <? subad 56-59 (57-7) ; 3 $ imm 55-60 (57-3) ;
Ugi 7 c? ad 55-58-5 (57'3) ; i 3 imm 58-5 ; 5 $ ad 54*5-57 (55'9) : * ? imm 57'5-
Bill San Cristoval 4 $ ad 16-5-17-5 (17-0) ; 4 $ subad 16-17 (16-6) ; 2 <$ imm 16-5,
17 ; i $ ad 16 ; 3 $ imm 15-16-5 (15-8) ; Ugi 9 <J ad 16-17-5 (16-8) ; i <$ imm 16-5 ;
4 $ ad 16-16-5 (zfi-i) ; i ? imm 17.

Iris very dark brown. Bill dark or dull blue-grey, tip and edges and sometimes
culmen black, maxilla occasionally more or less black ; imm average more black.
Foot blackish, shank dark blue-grey.

The immatures differ from adult females in the texture and shape of wing and tail,
and in the colour of the dried bill. The females and immatures vary considerably in
the intensity of rufous above and below, but without any apparent correlation with
age. The males also vary markedly, in the shape, colour and texture of wing and tail,
and in the colour and intensity of gloss. Five males, separated as subadults, have
soft wings with pale edges to the inner feathers, pointed tails, duller gloss above and
greener gloss on the throats (though this may be due to wear). However, they do not
differ from adults in testis development and bill colour, while several males are inter-
mediate in character.

62 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

NOTES. The genus Myiagra forms a single superspecies distributed from the Lesser
Sunda Isles to Samoa, yet curiously absent as breeding populations from the Bis-
marcks and most of New Guinea. Cain (19560 : 106) gives a brief account based on a
review by I. C . J. G. A satisfactory species arrangement is made difficult by the overlap
between very similar forms (especially rubecula and cyanoleuca in Australia) and the
extreme variation between representative isolated forms, together with striking
convergences made possible by a limited repertoire of variation. Mayr's arrangement
(19410 ; 19450) is not consistent, as is seen especially in the aggregation of very
distinct Micronesian forms into a single " species " M. oceanica, while the Polynesian
forms caledonica, vanikorensis and albiventris are separated. It might be most
satisfactory to combine the majority of forms in a single polytypic species (M. rube-
cula) from the Moluccas to Samoa and Micronesia to Australia, sympatric with three
other species, M. ruftcollis, M. cyanoleuca and M. azureocapilla. However, the specific
separation of distinct representatives is here accepted, though Mayr's arrangement
for the Solomons does not seem satisfactory. The inclusion of cervinicauda in M.ferro-
cyanea emphasizes their geography rather than their characters, since cervinicauda
resembles the Southern Melanesian forms vanikorensis and caledonica at least as
closely, in coloration and bill-form. It seems to be rather distinct from both groups in
ecology, in its confinement to the canopy and avoidance of the substage and second
growth (Cain & Galbraith, 1956 : 275), and is best treated as a distinct species.

(M. in) Petroica multicolor dennisi Cain & Galbraith, 1955

REFERENCES. Cain & Galbraith, 1955 : 93. See Mayr, 19346.

RANGE. Guadalcanal (mountains).

SPECIMENS. Guadalcanal : Turipava 6 <$ ad, 3 <$ ad retarded, 3 ? ad, i $ imm.
i <$ ad, i $ ad to AMNH.

Weight 6 g ad 10-11 (10-3) ; 3 ^ ad ret 10-5-11 (107) ; 3 $ ad 11-14-5 (12-3) ;
i $ imm 10. Wing 6 ^ ad 62-5-64-5 (63-5) ; 3 <$ ad ret 60-5-66 (62-7) ; 3 $ ad 62-62-5
(62-2) ; i $ imm 62-5. Tail 5 $ ad 39-5-44-5 (42-1) ; 2 J ad ret 41-5, 42 ; 3 ? ad
41-42-5 (42-0) ; i ? imm 42-5. Bill 5 ,3 ad 13-5-14-5 (13-9) ; 2 ^ ad ret 14, 14 ; 3 ?
ad 14-14 (14-0) ; i $ imm 14.

Iris dark brown. Bill maxilla black, mandible horn-colour to blackish, paler
basally. Foot dark brown to dull orange-yellow, fuscous-tinged, sole yellow to orange.

The holotype (adult female, expedition no. 520) is BM(NH) reg. no. 1959.21.957.

The three adult males in retarded plumage have enlarged testes. They differ from
those in advanced plumage in having the crown blackish-brown (with white traces
on the forehead), the wing and especially the primary-coverts browner, the outer
coverts brown not white, the white edges of the secondaries washed with ochraceous,
and the throat and mantle somewhat brownish.

(M. in) Petroica multicolor polymorpha Mayr, 1934

REFERENCES. Mayr, 19346 : 11.
RANGE. San Cristoval (mountains).

LAND BIRDS OF GUADALCANAL 63

(M. 112) Pachycephala pectoralis cinnamomea (Ramsay, 1879)

REFERENCES. Mayr, iQ32c : 14. See Galbraith, 1956.

RANGE. Guadalcanal, Beagle.

SPECIMENS. Guadalcanal : Betilonga 18 ^ ad, 2 imm, 8 $ " ad ", i ? imm, i ?
juv. i $ " ad " to AMNH ; i <? ad, i $ " ad " to OUM.

Weight 18 $ ad 48-56 (517) ; 2 J imm 45, 48 ; 8 ? " ad " 42-47-5 (44-9). Wing 16 <
ad 100-5-105-5 (103-1) ; i <$ imm 99 ; 8 $ " ad " 93-5-102 (98-1). Tail 16 $ ad
74-79-5 (76-3) ; 2 cJ imm 71-5, 81 ; 8 $ " ad " 70-5-77-5 (74-5). Bill 15 ^ ad 21-5-24
(22-9) ; 2 c imm 23, 24 ; 8 $ " ad " 21-24 (22-3).

Iris grey more or less suffused with red, yielding brownish grey to rufous brown,
sometimes blotched or veined. Bill ad black ; ^ imm and $ horn-colour, usually
darker and greyer above, sometimes streaked with black ; juv paler and yellower,
gape yellow. Foot grey, more or less tinged with bluish and /or fuscous, $ averages
paler ; juv whitish, tinged with fuscous and blue in patches.

Males in the " II Phase " plumage (Mayr, 19326 & d) are here considered as imma-
tures, which in this species have apparently already passed through a moult of the
wing and tail. They (and corresponding specimens of christophori) are extremely
variable in tail-shape, which suggests that the onset of moult is not well co-ordinated
with the maturation of the gonads and feather-follicles. The series of " adult "
females probably includes some immatures, but these are apparently not recognizable.
Unlike the 7 corresponding specimens described by Mayr (1932^ and 3 in the BM(NH),
5 of the 8 non- juvenile females are strongly washed with yellow pigment. It remains
to be determined, whether yellow coloration is confined to undetected immature
females (as an exception to the general rule that younger birds have less carotenoid) ;
or whether the character is more variable in expression than Mayr's material had
suggested. Since cinnamomea and bougainvillei differ consistently from orioloides
only in this character, and from each other only in presence or absence of a rufous
wash (which is variably developed in orioloides) , proof of the latter would necessitate
the union in a single subspecies, P. p. orioloides, of all the populations along the Main
Chain from Buka to Guadalcanal which are certainly very closely related.

NOTES. The name Aiuscicapa pectoralis Latham, 1801, has been validated by the
International Commission (ICZN 1956 ; 1958) ; and so remains the name for this
polytypic species whether or not the New Caledonian form, to which the name
Muscicapa caledonica Gmelin, 1789 applies, is included (see Galbraith, 1956 : 174).

(M. 112) Pachycephala pectoralis christophori Tristram, 1879

REFERENCES. Mayr, 1932^ : 19. See Galbraith, 1956.

RANGE. San Cristoval, Santa Ana.

SPECIMENS. San Cristoval : Goge 23 J 1 ad, 8 $ imm, i $ juv, 16 $ " ad ", 2 $ juv,
2 ? imm, 4 ? juv, i ? nest ; Nagasi 18 $ ad, 4 <$ imm, 5 $ " ad ", i ? imm, 2 ? juv,
i ? nest. 2 $ ad, i $ " ad " to OUM.

Weight Goge 23 $ ad 28-36 (33-3) ; 9 $ imm 28-33 (30-7) ; i juv 31-5 ; 16 ?
" ad " 28-37-5 (31-8) ; 2 ? juv 31, 32 ; Nagasi 18 <? ad 32~35'5 (33'9) > 4 3 imm 29-33
(317) ; 5 ? " ad " 31-36 (33-6). Wing Goge 22 < ad 81-90 (85-4) ; 7 <J imm 80-85

64 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

(82-8) ; i juv 84 ; 16 ? " ad " 80-89 ( 8 4'2) ; 2 $ juv 80, 80-5 ; Nagasi 18 <$ ad
85-90 (87-0) ; 4 <$ imm 81-5-86 (84-2) ; 5 $ " ad " 83-89 (86-3). Tail Goge 19 <J ad
54-5-63-5 (59-i) J 4 c? imm 56-61-5 (58-6) ; 12 $ " ad " 57'5-64 (59'9) i 2 ? J uv 56,
59-5 ; Nagasi 13 <J ad 57-5-62-5 (59'8) ; 3 3 imm 60-61-5 (60-3) ; 4 $ " ad " 58-62-5
(59-9). Bill Goge 19 $ ad 20-21-5 (20-7) ; 8 $ imm 20-21-5 (20-8) ; i <$ juv 18-5 ;
15 $ " ad " 20-5-22 (20-8) ; 2 $ juv 21-5, 22 ; Nagasi 18 <$ ad 20-22-5 I 21 ' 1 ) ; 4 ?
imm 20-5-21-5 (21-1) ; 5 ? " ad " 20-21-5 (20-8).

Iris grey more or less suffused with red, yielding brownish grey to rufous brown,
sometimes blotched or veined ; juv greyer. Bill <$ ad black, base of mandible often
paler, gape black to whitish or yellow, mouth-lining whitish to bright orange-yellow,
sometimes blotched or rimmed with black ; other specimens maxilla blackish to dark
greyish horn, mandible paler and yellower, often streaked, gape and mouth-lining
never black. Foot grey, more or less tinged with bluish or brownish, sometimes in
irregular patches.

In this series, there is scarcely any overlap in bill-colour and testis-development
between the 41 adult males and the 12 males in " II Phase " plumage, listed above as
immatures. The latter clearly cannot be adults in a retarded plumage, even if this is
the explanation for such specimens in other races (Mayr, I932C & d). The rectrices
of these immatures differ in shape and texture from those of juveniles, and in average
colour (with little overlap) from those of adults. It thus appears that an additional
moult of the tail, and probably of the wing, is involved.

There is an unusual degree of overlap in this race, in the appearance of the wing,
between the various plumage phases. The wing of nestlings and juveniles (" I Phase ")
is firm in texture (only the upper coverts being markedly soft) and much mixed with
olive, which sometimes extends far up the edges of the secondaries. Conversely,
several adult males and specimens of both sexes in " II Phase " have much rufous in
the secondaries and upper coverts.

Presumably, females in " II Phase " include adults and immatures, but it seems
impossible to separate them. The immature males are also indistinguishable except
for one which (like two recorded by Mayr, i-932c) has a distinct gorget of irregularly-
marked black-and-yellow feathers. This approaches the condition in some adults
(with black bills, enlarged testes and blacker tails) which have the gorget much
mixed with yellow. Otherwise, the gorget in " II Phase " is very variable in colour,
from olivaceous or tawny yellow to olive.

The variable degree of melanization of the adult male is correlated, though imper-
fectly, between the head and tail, but is apparently independent of the development
of rufous on the breast. In contrast, the Santa Ana population (Mayr, I932c) is much
more constant in both characters, having the head and tail mostly green and the
breast with much rufous.

(M. 113) Pachycephala implicata implicata Hartert, 1929

REFERENCES. Mayr, I932C : i.
RANGE. Guadalcanal (mountains).

SPECIMENS. Guadalcanal : Turipava 7 ad, i < imm, i <$ juv, 12 $ ad, i $ juv.
i <$ ad, i ad to OUM.

LAND BIRDS OF GAUDALCANAL 65

Weight 7 J ad 33'5~37 (357) ; i $ imm 33 ; i <? juv 29-5 ; 13 ? ad 30-39 (33-3) ;
i $ juv 32. Wwzg 7 c? ad 86-5-91 (89-1) ; i <? imm 88-5 ; 12 $ ad 84-88 (86-8) ; i $
juv 83. Tail 6 $ ad 69-72 (70-2) ; n $ ad 67-72 (68-7) ; i $ juv 67. Bill 4 $ ad
19-20 (19-4) ; i (J imm 19 ; i ^ juv 14 ; 8 $ ad 18-5-20 (19-2).

Iris dark brown, sometimes reddish, $ greyish ; juv dark grey-brown. Bill black ;
juv fuscous (mandible and base paler) with pale yellow gape, or black with yellowish
edges. Foot grey, usually bluish, shank darker and sometimes mottled ; $ average
paler ; juv shanks pale blue-grey, toes ivory with bluish tinge, or as ad.

The male juvenile, with small pale bill and feet and growing wing and tail, agrees
with Mayr's (i932c) description of nestling richardsi. The female juvenile, evidently
older, is much less rufous, though a small patch of the yellow immature plumage on
the breast shows it not to have undergone a complete moult. The adult females also
vary greatly in degree of phaeomelanization ; especially on the rump, wing, tail and
underparts.

(M. 114) Aplonis cantoroides (Gray, 1862)

REFERENCES. Cat. Birds B.M. 13 : 128 ; Heinroth, 1903 : 74.

RANGE. New Guinea region and Northern Melanesia including Ugi, Three
Sisters and Santa Ana, but not San Cristoval (coasts) .

SPECIMENS. Guadalcanal : Tenaru i ^ ad, 3 ^ imm, 2 $ ad, 10 $ imm, i $ juv,
i ? juv. Ugi 7 $ ad, i J imm, 7 $ ad, i $ juv. i $ ad, i $ imm to OUM.

Weight Guadalcanal i $ ad 53-5 ; 3 imm 52 -5-53 (52-7) ; 2 $ ad 53-5, 58 ;
10 $ imm 43-53'S (48-8) ; i $ juv 44 ; Ugi 8 $ ad 51-5-59 (54-9) ; i $ imm 53 ;
7 $ ad 50-5-56 (53-4) ; i ? juv 50-5. Wing Guadalcanal i $ ad 100 ; 3 ^ imm 96-
97-5 (96-8) ; 2 ? ad 98, 98-5 ; 10 ? imm 91-99 (94-1) ; i ? juv 91-5 ; Ugi 7 ^ ad
100-105-5 (102-0) ; i ^imrn 96 ; 7 $ ad 96-101 (98-7) ; i ?juv 96. Tail Guadalcanal
i c? ad 66 ; 3 $ imm 63-64 (637) ; i $ ad 64 ; 9 $ imm 56-5-65 (59' 1 ) > i ? juv 50-5 ;
Ugi 6 <$ ad 65-70 (67-4) ; i $ imm 63 ; 6 $ ad 62-66 (64-1) ; i $ juv 52. Bill
Guadalcanal i ^ ad 24 ; 3 ^ imm 23-24 (23-3) ; 2 $ ad 23-5, 23-5 ; 10 $ imm 21-5-
24 (22-5) ; i $ juv 21 ; Ugi 6 ^ ad 22-24-5 (23-3) ; i $ imm 23-5 ; 6 $ ad 22-5-23-5
(23-2) ; i $ juv 22.

7ns bright orange ; juv greenish yellow to dark brown. Bill black ; juv gape
yellowish. Foot black, occasionally brownish-tinged ; juv dark grey, brownish-
tinged, to black.

NOTES. Amadon (1956 : 22) regards the species as monotypic, and separates
crassa of the Tenimber Isles as a representative species.

The immature and juvenile plumages are described by Heinroth (1903).

Though listed as " widespread " in the Solomons by Mayr (19450 : 277), the species
has been recorded only twice from San Cristoval (Tristram, 1882 : 137 ; Davidson,
1934 : 194). There are no specimens from this island in the AMNH (Amadon in litt.)
or the BM(NH). Natives of San Cristoval, even on the coast, denied knowledge of it ;
whereas the inland people of Betilonga, where neither species occurs, were well aware
of the distinction between A. cantoroides and A. metallica. French (in litt.) never saw
A . cantoroides on his many visits to San Cristoval, nor were its eggs brought to him

ZOOL 9. i. 5

66 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

by the collectors he employed there. He does not believe that it can be resident.
Tristram did not state the number of specimens, nor their precise locality. Davidson's
single bird was taken at Star Harbour, and may well have been a straggler from
Santa Ana (see p. 15).

The species was not taken on Malaita by the Whitney Expedition (Mayr, I93i/),
but Davidson (1934) records a specimen, and there are 4 in the BM(NH) collected by
R. A. Lever. The species is resident on Rennell (Bradley & Wolff, 1956 : no), where
the nesting sites recorded for A. cantoroides and A. insularis need confirmation.

(M. 116) Aplonis grandis macrura Mayr, 1931

REFERENCES. Mayr, 19317 : 21.

RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Tenaru i ^ ad, i $ ad ; Betilonga 13 $ ad, 19 $
ad, i $ imm, i ? ad, I ? imm. i $ ad, I ? imm to OUM.

Weight 15 <$ ad 114-5-151 (136-4) ; 21 $ ad 110-5-150 (133-3) > i ? imm H9- Wing
14 $ ad 132-146 (140-7) ; 19 <j> ad 130-144 (136-4) ; i $ imm 134. Tail 12 $ ad
92-5-103 (98-3) ; 17 $ ad 87-99-5 (93-6) ; i $ imm 89. Bill 14 ad 26-29 (27-5) ;
9 $ ad 26-29 (27-3) ; i $ imm 27-5.

Iris deep red, sometimes orange-red ; imm orange-red. Bill black. Foot black,
sole sometimes yellowish.

The immatures differ relatively little from the adults, in having less gloss on the
wing, the primaries dull brown and without fawn outer webs, the hackling of the head
and neck greatly reduced (despite Rothschild & Hartert, 1905 : 268) and without
green tinges in the purple gloss, and the gloss of the body duller and bluer and
restricted to the tips of the feathers so that the back appears scaly and the belly
and under tail-coverts dull and brownish. Individual adults approach this condition
in separate single characters.

(M. 116) Aplonis dichroa (Tristram, 1895)

REFERENCES. Ramsay, 18820 : 726 (minor).

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Goge 7 <$ ad, 4 <$ imm, 8 $ ad, i $ imm, 3 ? imm ;
Nagasi 3 < ad, 3 <$ imm, 3 $ ad, 5 $ imm, i ? imm. i ad, i $ imm to OUM.

Weight 8 ad 82-5-90 (86-9) ; 7 ^ imm 68-90 (76-6) ; 9 $ ad 73-91-5 (81-2) ; 6 $
imm 68-82 (74-1). Wing 8 ^ ad 112-122 (117-9) ; 7 3 imm 108-5-118 (112-4) Ir - ?
ad 110-119 (113-5) i 6 $ imm 109-116 (in-8). Tail 9 $ ad 71-5-76-5 (73*5) '> 3 3 imm
61-5-62-5 (62-0) ; 9 $ ad 66-5-72-5 (69-1) ; 4$ imm 61-5-68 (64-7). Bitt 10 <J ad
26-28 (26-9) ; 7 ( imm 25-28 (26-0) ; II $ ad 25-27 (26-0) ; 6 $ imm 24-5-26 (25-3).

Iris tawny orange (deep orange-red to pale tawny orange) ; imm pale golden-
brown (dull yellow-brown to golden yellow with orange wash) . Bill black ; imm gape
sometimes fuscous, yellow internally. Foot black.

The immature plumage differs from that of the adult much more than in our series
of A. grandis. The upper wing-coverts are dark brown without gloss, the head and

LAND BIRDS OF GUADALCANAL 67

neck are without hackles, the contour feathers are fuscous with gloss only at the tips,
and the rectrices are paler and narrower. However, the remiges are like those of the
adult.

The proportion of young birds taken near Goge shows a striking change after the
visit to Nagasi, where an intermediate proportion was taken. The numbers of imma-
tures and adults in the collection are : Goge (i5th October-i3th November) 2/15 ;
Nagasi (20-24th November) 9/6 ; Goge (30th November-3rd December) 6/0. This
suggests that the young birds were moving about together apart from the adults, as
in the related A . cantoroides ; though the feeding flock seen was not noticed to be
composed mainly of immatures (Cain & Galbraith, 1956 : 280).

NOTES. We agree with Amadon (1956 : 23) that this form should be kept specific-
ally distinct from A . grandis, its representative in the remainder of the Solomons.

The identity of the third species of Aplonis on the Three Sisters, recorded in error
by French (1957) as A. grandis, remains in doubt (see Cain & Galbraith, 1957,
footnote). There is now no indigenous Melanesian population in these islands, who
alone might be expected to distinguish between A . cantoroides and A . feadensis or
A. insularis. However, through the kindness of the Chief Secretary to the High
Commission and the District Commissioner, workers on the Lever estate were asked
to observe, and after two months in 1957 reported only A . cantoroides and A . metallica.
If able to distinguish these two species, they would immediately have recognized
A . grandis, A . dichroa or A . brunneicapilla. If one of these species was present during
Mr. French's residence, it cannot still be common as he recorded it.

(M. 117) Aplonis metallica nitida (Gray, 1858)

REFERENCES. Stresemann, 1912 : 311 ; 1914 : 151 ; Amadon, 1956 : 22.

RANGE. Bismarcks, Solomons (lowlands).

SPECIMENS. Guadalcanal : Tenaru 15 ^ ad, 9 ^ imm, 10 $ ad, 4 $ imm, i ? ad,

1 ? imm. San Cristoval : Goge 23 $ ad, 16 < imm, 6 $ juv, 14 $ ad, 4 $ imm, 3 ?
imm, 12 ? juv. Ugi 5 $ ad, 2 $ imm, 3 $ juv, 2 $ ad, 7 $ imm. All to OUM.

Weight Guadalcanal 15 ^ ad 51-5-60-5 (55-7) ; 9 $ imm 50-5-60-5 (55-9) ; 10 $
ad 52-66-5 (57-1) ; 4 ? imm 47-5-52-5 (50-4) \ San Cristoval 19 <? ad 49-63-5 (58-7) ;
9 cJ imm 48-5-61-5 (55-2) ; 6 3 juv 47-5-58 (53-1) ; 10 ? ad 47'5~57 (54'2) ; 3 ? imm
50-51 (50-7) ; Ugi 5 c? ad 52-5-60-5 (56-5) ; 2 $ imm 52, 55-5 ; 3 $ juv 44-5-53
(49-8) ; 2 ? ad 54-5, 57-5 ; 7 ? imm 46-57 (52-0). Wing Guadalcanal 15 $ ad 106-
115 (110-4) i 9 c? imm 99-5-108-5 (103-5) ; 10 ? ad 102-108 (105-3) ; 3 ? imm 96-5-
100 (98-5) ; San Cristoval 23 $ ad 101-5-109-5 (106-0) ; 13 < imm 95-109 (101-4) ;
6 $ juv 99-5-104-5 (102-9) > 14 ? ad 98-5-107 (102-2) ; 3 ? imm 97-5-101 (98-8) ;
Ugi 5 <J ad 106-109 (107-7) ; 2 <$ imm 98-5, 100-5 ; 2 juv 98, 99-5 ; i ? ad 105 ;
5 $ imm 94-5-98-5 (96-8). Tail (central rectrices) Guadalcanal 13 < ad 91-111-5
(102-5) ; 8 c? imm 82-5-95-5 (88-4) ; 10 $ ad 88-101-5 (93-2) ; 4 $ imm 79-88 (83-2) ;
San Cristoval 19 <? ad 88-5-105 (97-1) ; 14 $ imm 83-94-5 (87-5) ; 5 ^ juv 62-70
(67-2) ; 9 ? ad 80-93 (87-8) ; 4 $ imm 79-84-5 (81-2) ; Ugi 3 ^ ad 99-106 (102-2) ;

2 $ imm 80-5, 92 ; i ^ juv 68 ; 2 $ ad 89, 93 ; 5 ? imm 78-81-5 (79-6). Tail (2nd
rectrices) Guadalcanal 13 <$ ad 73-5-86 (79-9) ; 9 $ imm 58-75-5 (70-2) ; 10 $ ad

68 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

6 9-79*5 (73*4) ; 3 ? imm 57~ 68 ( 6 3'2) ; San Cristoval 21 <J ad 68-5-82 (75-3) ; 16 <$
imm 62-76-5 (70-1) ; 5 <$ juv 54-5-63 (58-3) ; 8 ? ad 65-75-5 (71-9) ; 4 ? imm 64-70-5
(66-9) ; Ugi 4 c? ad 76-5-81-5 (78-2) ; 2 < imm 58-5, 73 ; 2 juv 54, 55 ; 2 ? ad 67,
71 ; 6 $ imm 63-5-69 (65-7). Bill Guadalcanal 14 < ad 21-5-24-5 (22-8) ; 8 <$ imm
20-5-24 (21-9) ; 10 $ ad 21-5-23-5 (22-3) ; 3 $ imm 21-22-5 (21-7) ; San Cristoval
22 <$ ad 21-23-5 (22-0) ; 16 <$ imm 21-23-5 (22-3) ; 7 < juv 20-5-22-5 (21-4) ; 14 $
ad 20-5-23 (21-7) ; 4 $ imm 20-22 (21-1) ; Ugi 5 ^ ad 21-23 (22-2) ; 2 <$ imm 21-5,
22 ; 3 $ juv 20-5-21 (20-8) ; 2 $ ad 22, 22 ; 7 $ imm 21-23-5 (21-3).

Iris bright orange-red, often orange centrally, sometimes pinkish ; imm average
yellower, juv golden buff through pale greenish brown to dark grey-brown in
apparently youngest birds. Bill black ; juv gape fuscous to flesh, youngest birds have
dull black maxilla, horn mandible with yellowish base and edges, golden-yellow gape.
Foot black, sometimes fuscous.

The immatures have juvenile wings, but their tails (and that of one of the juveniles)
are adult.

The series from San Cristoval is extremely variable in the colour of the gloss.
Some specimens are more purplish than any others seen, with scarcely a trace of
green when viewed with the light, as in individuals of A . mystacea ; while others are as
greenish as any, having traces of purple on the shoulder only, as in A . bmnneicapilla.
Table II shows the distribution of adult specimens of both sexes among eight arbitrary
colour-classes. The 24 specimens in the BM(NH), from 10 islands in the Bismarcks
and Solomons, are grouped together, while all the specimens from Guadalcanal, San
Cristoval and Ugi were collected by the expedition.

TABLE II. Colour of body gloss in adults of Aplonis metallica nitida
Class A purplest, H greenest.

ABCDEFGH

Bismarcks and Solomons . i i 2 i 14 5

Guadalcanal ... 2 3 n 9 i

San Cristoval . . .12666691
Ugi .... 21 21

NOTES. The non-adult plumages are described by Stresemann (1914).

The character distinguishing nitida from metallica is not so much that the inter-
scapular patch is reduced (Amadon, 1956), as that its central blue-green area is either
totally abolished (Stresemann, 1912) or occasionally represented by a bluish tinge.
Also, even very purplish individuals have the green throat less mixed with purple
than have generally greener specimens of metallica.

(M. 118) Aplonis brunneicapilla (Danis, 1938)

REFERENCES. Beecher, 1945 : 31 ; Cain & Galbraith, 1956 : 281.

RANGE. Solomons known from Bougainville and Guadalcanal ; i specimen
from Rendova.

SPECIMENS. Guadalcanal : Betilonga 23 ^ ad, i ^ juv, 17 $ ad. i $ ad, i $ ad
toAMNH; i ad to OUM.

LAND BIRDS OF GUADALCANAL 69

Weight 24 ad 59'5-76'5 (69-2) ; i $ juv 62-5 ; 18 ? ad 59-73'5 (64-2). Wing 22 $
ad 108-5-115 (112-9) ; 16 $ ad 106-5-114 (109-4). Tail (central rectrices) n $ ad
119-5-204-5 (150-5) ; 6 $ ad 98-133-5 (115-7). Tail (2nd rectrices) 22 $ ad 73-5-91
(80-1) ; 15 $ ad 68-81-5 (72-4). Bill 23 $ ad 25-5-28 (26-8) ; i ^ juv 24 ; 17 ? ad
24-5-27 (25-5).

Iris white ; juv dark grey-brown. Bill black ; juv grey, maxilla dark with black
base, mandible paler with ivory base, gape orange-yellow. Foot black, sole whitish or
yellowish.

The juvenile at first sight resembles immatures of other species of Aplonis, in its
firm glossy plumage with wings and tail of almost adult gloss and texture. However,
the latter clearly do not belong to the adult feather-generation yet are still growing ;
and the weak pale bill, dark iris and sparsely-feathered throat confirm that the
specimen is in its first plumage.

Adults usually have little purple in the body gloss, but one adult male is mainly
purple above, and several specimens are bronzy beneath. The distribution of gloss
colours on the head and throat is reminiscent of that of A . metallica purpureiceps of
Manus : an arc over the eye purple ; lores, superciliary and cheeks brilliant green ;
the throat blue-green with purple mainly at the sides and across the chin and breast.
In fresh plumage, the crest of the adult male (about 15 mm long) and the crown are
glossy black with a brownish tinge, those of the female somewhat browner. They
bleach in worn plumage to pale dull brown.

NOTES. This species has been compared and associated with A. mystacea of
New Guinea lowlands (Amadon, 1943^ : 16 ; 1956 : 22 ; Mayr, 1955 : 36). However,
both are clearly related to the widespread A . metallica, and all three should be con-
sidered together. A. brunneicapilla outdoes A. mystacea in specialization of the head-
feathering, pallor of the iris, arching of the bill, and suppression of the distinctive
juvenile pattern. This last is a common trend in Aplonis, especially among long-
established island populations. While A . mystacea is the smallest and purplest of the
three species (retaining traces of the blue-green area in the interscapular patch, and
adding a purple rump to the pattern shown by A. m. metallica), A. brunneicapilla is
the largest and greenest, and its enormously developed tail-streamers are more like
those of most races of A. metallica. On geographical grounds, it seems probable that
A . mystacea and A . brunneicapilla are both derived from A . metallica, and owe their
special resemblances to convergence.

(M. 119) Acridotheres tristis tristis (Linnaeus, 1766)

REFERENCES. Baker, 1926 : 53.

RANGE. Afghanistan through India to Thailand ; introduced to islands in the
Solomons (coasts).

SPECIMENS. Guadalcanal : Tenaru i J ad.

Weight 127-5. Wing 143. Bill 26.

Iris dirty white overlaid with reddish brown. Skin round eye orange yellow, eyelid
blackish. Bill greenish yellow, base of mandible blackish. Foot dull yellow, greenish
on claws and edges of scutes.

ZOOL Q. I. 5

yo IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

NOTES. In the Solomons, known from the Russell Is. (Mayr, 19450; ; 1955 : 36),
Guadalcanal (Davidson, 1929 : 259 ; Donaghho, 1950 : 129 ; Cain & Galbraith,
1956 : 287) and the Three Sisters (French, 1957).

(M. 120) Mino dumontii kreffti (Sclater, 1869)

REFERENCES. Amadon, 1956 : 34.

RANGE. Bismarcks, Solomons except San Cristoval Group irregular geographical
variation in size.

SPECIMENS. Guadalcanal : Betilonga 5 ^ ad, 3 $ ad. i <$ ad to OUM.

Weight 5 $ ad 182-200 (191); 4 $ ad 170-215 (189). Wing 5 $ ad 149-155
(i53'3) i 3 $ ad 146-150 (148-3). Tail 3 <$ ad 97-100 (98-5) ; 3 $ ad 95-98-5 (967)-
Bill 5 J ad 33-36-5 (34'4) > 3 ? ad 31-33 (32-2).

Iris yellow. Skin round eye orange. Bill orange. Foot yellow to orange-yellow.

NOTES. Amadon (1956) shows that the small birds of Guadalcanal and Malaita
(sanfordi Hartert) are best combined with kreffti.

(M. 121) Dicrurus hottentottus meeki Rothschild & Hartert, 1903

REFERENCES. Vaurie, 1949 : 291. See Mayr & Vaurie, 1948.

RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Betilonga 4 ^ ad, 4 $ ad, i $ imm, i $ juv, i ? imm.
i <$ ad to OUM.

Weight 4 $ ad 73-5-81-5 (78-1) ; 4 $ ad 79-86 (81-6) ; i $ imm 71 ; i $ juv 65-5.
Wing 4 ad 146-5-152 (148-6) ; 4 $ ad I39~ I 43 (141-2) ; i $ imm 138 ; i $ juv 135.
Tail (outer rectrices) 4 $ ad 129-135 (131-1) ; 3 $ ad 120-127-5 (124-8) ; i $ imm
121-5; i $ juv 119. Tail (fork depth) 4 <J ad 7-10-5 (8-1) ; 3 $ ad 2-5-7 (47) '>
i $ imm 6 ; i $ juv 0-5. Bitt 4 $ ad 34-37 (36-0) ; 4 $ ad 32~34'5 (33'6) ; i $ imm
33; i $ juv 32-5.

Iris orange to orange-red ; imm red ; juv dull orange. Bill black. Foot black.

The " immatures " are doubtfully separated as birds in the second year plumage,
on the basis of somewhat more rounded and less curled tails (see Vaurie, 1949 : 206) .

(M. 121) Dicrurus hottentottus longirostris (Ramsay, 1882)

REFERENCES. Vaurie, 1949 : 291. See Mayr & Vaurie, 1948.

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Goge 2 ^ ad, 2, <$ imm, i ? imm ; Nagasi i ^ ad,
3 <$ imm, i ? imm. i ^ ad to OUM.

Weight 3 ad 85-5-96-5 (91-5) ; 5 <? imm 78-5-89-5 (83-4). Wing 3 3 ad 144-152
(149-3) ; 5 <$ imm 137-145 (140-1). Tail (outer rectrices) 2 <$ ad 130, 140; 3 <$ imm
120-5-128 (124-2). Tail (fork depth) 2 ^ ad 7, 15 ; 3 J imm 2-4-5 (3-0). Bill 3 <
ad 44-46 (44-8) ; 5 <? imm 39-44 (41-8).

Iris deep dull red to deep red-brown ; imm dark brown to dark greyish brown.
Bill black ; i imm with pale horn tip. Foot black.

The immature specimens are separated on their more rounded and less curled

LAND BIRDS OF GUADALCANAL 71

rectrices. They average less glossy above, and smaller, and must be in their second-
year plumage (see Vaurie, 1949 : 206).

NOTES. The two races in the Solomons span virtually the whole range of variation
shown by the highly polytypic D. hottentottus, in bill-form and the development of
rictal bristles. One might expect a greater difference in diet than Cain & Galbraith's
(1956) small samples showed. With its very long compressed bill and weak bristles,
longirostris seems the better adapted to a diet of lizards and large insects.

(M. 123) Corvus woodfordi woodfordi (Ogilvie-Grant, 1887)

REFERENCES. Vaurie, 1958 : 9.

RANGE. Choiseul, Ysabel, Guadalcanal irregular geographical variation in size.

SPECIMENS. Guadalcanal : Betilonga i $ ad, 3 $ ad. i $ ad to OUM.

Weight I $ ad 425 ; 3 $ ad 375-405 (386-0). Wing 2 $ ad 262, 265. Tail I ^ ad
130 ; 3 ? ad 118-126 (123-0). Bill i ad 64 ; 3 $ ad 60-5-61 (60-8).

Iris grey. Skin of throat purplish red. Bill ivory, shading to pale blue and purple
basally, tip black. Foot black.

NOTES. We follow Vaurie (1958) in regarding the black-billed meeki of Bougain-
ville as conspecific with woodfordi, and the population of larger birds on Ysabel
(vegetus) as belonging subspecifically to the latter.

(H) Gymnorhina tibicen subspecies

REFERENCES. Condon, 1951 : 67 ; Amadon, 1951 : 20. See Amadou, 1953.

RANGE. Southern and eastern Australia, introduced to Guadalcanal.

NOTES. Records for Guadalcanal are all from the grasslands near the Tenaru
River (Beecher, 1945 : 36 ; Baker, 1948 : 17 ; Cain & Galbraith, 1956 : 288), where the
species is probably not well-established.

(M. 124) Nectarinia jugularis flavigaster Gould, 1843

REFERENCES. Rothschild & Hartert, 1914 : 297 ; Mayr & Rand, 1937 : 210.
See Delacour, 1944.

RANGE. Admiraltys, Bismarcks, Solomons except San Cristoval and Santa Ana
(coasts).

SPECIMENS. Guadalcanal : Tenaru 8 $ ad, i <$ juv, 6 $ ad, i $ juv, 2 ? ad. Ugi
2 $ ad. i <? ad to OUM.

Weight 8 $ ad 9-5-11 (10-2) ; i $ juv 9 ; 8 $ ad 8-5-9-5 (8-9) ; i $ juv 8. Wing
7 $ ad 54-58-5 (56-o) ; i 3 Juv 53 ; 8 $ ad 48-5-53 (51-9) ; i $ juv 47-5. Tail 7 $
ad 34-5-38-5 (37-o) i i c? juv 31-5 ; 7 ? ad 31-5-35 (33-9) ', * ? J uv 2 9'5- Bill 6 $ ad
22-24 ( 2 3 >2 ) > i c? J uv 2I i 8 $ ad 21-22 (21-6) ; i $ juv 21-5.

Iris dark to very dark brown. Bill black ; i juv base of mandible blackish horn,
gape yellow. Foot black, shanks sometimes fuscous ; i juv leaden-black.

The juveniles are distinguished from adult females by laxer plumage, paler caro-
tenoid, and lack of blackish subterminal patches on the dorsal feathers.

NOTES. This widely-distributed species has not been recorded from San Cristoval
(see Mayr, 19450 : 279), and there are no specimens from there in the AMNH (Amadon,

72 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

in litt.) or BM(NH). The expedition did not encounter it at Manewiriwiri, where
natives did not recognize a coloured figure (Mayr, 19450 pi. 3, figs. 34-35). French
(in litt.} doubts its presence on the island. Nor is it found on Santa Ana, though
present on Ugi (see Cain & Galbraith, 1956 : 104) and the Three Sisters (Cain &
Galbraith, 1956 ; French, 1957). Its absence from San Cristoval may be related to
the presence of two species of Myzomela, ecologically very similar to Nectarinia and
both abundant along the coast. Ripley (1959) discusses competition between these
two genera in the Moluccas, while Salomonsen (personal communication) finds a
similar geographical relationship in the Bismarcks (see Ripley, 1961). That N. jugu-
laris and M. nigrita have precisely complementary ranges may be fortuitous : M.
cardinalis, which coexists with the former on Ugi and the Three Sisters, might be
expected, as another wholly coastal species, to compete with it the more intensely.

(M. 125) Myzomela cardinalis pulcherrima Ramsay, 1881

REFERENCE. Mayr, 19320 : 24. See Koopman, 1957.

RANGE. San Cristoval, Ugi, Three Sisters (coasts).

SPECIMENS. San Cristoval : Manewiriwiri 6 <$ ad, 12 $ imm, 2 <$ juv, 6 $ ad, 3 $
imm, i $ juv, I ? imm, 3 ? juv. Ugi 33 <$ ad, 13 <$ imm, 3 $ juv, 7 $ ad, 4 $ imm, I ?
imm, i ? juv. i ^ imm, i <$ juv, i $ ad, i $ imm, i ? juv to AMNH ; i $ ad, I <$
imm, i $ ad to OUM.

Weight San Cristoval 6 <$ ad 14-5-16 (15-4) ; 12 <$ imm 14-16 (15-0) ; 2 <$ juv
12-5, 15-5 ; 6 $ ad 11-5-13 (12-3) ; 3 ? imm 11-5-13-5 (12-5) ; i $ juv 13-5 ; Ugi
33 <J ad 14-18 (15-8) ; 13 <? imm 13-5-17-5 (15-5) ; 3 c? juv 14-5-16 (15-3) ; 7 ? ad
12-14 (13-0) ; 4 $ imm 12-13-5 (12-7). Wing San Cristoval 5 <$ ad 68-74 (70-6) ;
12 ^ imm 65-5-70-5 (67-9) ; i < juv 66 ; 5 $ ad 61-5-65 (63-8) ; 3 $ imm 60-64
(62-0) ; i ? juv 61-5 ; Ugi 28 <? ad 67-74 (69-7) ; n $ imm 65-68-5 (67-2) ; 3 <? juv
66-69-5 (67-5) ; 6 $ ad 61-66 (63-3) ; 3 $ imm 61-63 (61-8). Tail San Cristoval
4 ^ ad 46-47-5 (46-7) ; 10 $ imm 38'5-44'5 (4* '5) > * 3 Juv 40'5 ', 5 ? ad 39-41-5
(40-0) ; 3 $ imm 39-40 (39-7) ; i $ juv 37-5 ; Ugi 22 < ad 42-48 (44-9) ; 13 3 imm
39-44 (41-6) ; 3 c? juv 40-43 (41-3) ; 3 ? ad 38-5-40 (39-3) ; 3 ? imm 37-38-5 (377)-
Bill San Cristoval ; 6 $ ad 20-5-22 (21-0) ; n ^ imm 19-5-21 (20-1) 2 <$ juv 16-5,
19-5 ; 5 ? ad 18-5-20 (19-3) ; 3 $ imm 19-19-5 (19-3) ; I $ juv 18 ; Ugi 29 $ ad
20-22 (20-8) ; 10 c imm 19-21-5 (20-3) ; 3 <$ juv 20-21 (20-3) ; 6 $ ad 19-19-5 (19-3) ;
3 $ imm 19-19-5 (19-2).

Iris very dark brown. Bill black, J gape occasionally fuscous or yellowish fuscous,
$ base of mandible sometimes brownish and gape yellowish (dark horn to golden-
yellow) ; imm and juv base of mandible horn, gape fuscous-yellow to golden-yellow,
base of $ maxilla greyish horn, ^ sometimes greyish. Foot dark grey, toes darker and
browner, shanks bluish (often patchily) with silvery sheen, $ ad average darkest.

The adult males are readily separable and do not show much colour variation, but
the other plumages are similar, intergrading and variable. The colour and texture
of wing and tail are much affected by wear, and there is little difference between
those of juveniles, immatures and adult females. The red pigmentation is very variable
in extent, and in whether it forms a general wash or (as in the adult males) is confined
to specialized feather-tips lacking barbules.

LAND BIRDS OF GUADALCANAL 73

Adult females are dorsally like immature males, and differ ventrally in having the
red virtually restricted to throat and breast, with the belly usually sombre grey
scarcely washed with pink, olivaceous or buff. Immature females differ from adults
in having red tips only on the back ; red on the head confined to the forehead and
front of the crown ; the back averaging paler and more olivaceous or brownish, with
much less red ; the ventral red much less intense, but extending further posteriorly ;
and the belly paler and more decidedly washed with red or olive. Immature males
differ from immature females in having much more red ventrally, with specialized
red tips. The brownish hind-crown and nape of a few immature males, resembling
those of immature females, probably belong to the juvenile plumage. There is
scarcely any overlap, in either sex, in the gonadial development of the adult and
immature series.

Juveniles are very like immatures, in both texture and coloration of plumage.
Their identification as such depends on one male BM(NH) reg. no. 1959.21.1374)
which is coloured much like immatures, as far as can be determined from its some-
what disarranged plumage, yet is clearly very young : its bill is small (16-5 mm),
wing and tail growing (52-5 and 20-5 mm), and contour-plumage scanty with large
naked tracts. The plumage considered to be juvenile has the hind-crown and nape
dull brown with a faint reddish wash, instead of bright red as in immatures, and is
much less intense red beneath. Female and unsexed juveniles are like immature
females, but with less red on the crown.

NOTES. Most of the populations attributed to this species (Mayr, 1932 ; Koopman,
1957) are rather uniform in basic colour-pattern, though variable in dimensions,
intensity of pigmentation, and degree of sexual dimorphism. However, the form on
Rotuma (chermesina Gray) is very different, and perhaps belongs to the lafargei
superspecies. It resembles M. malaitae of that group rather than the races of M.
cardinalis. The presence of a putative member of the superspecies (M. jugularis) in
Fiji demonstrates its colonizing potential.

(M. 126) Myzomela melanocephala (Ramsay, 1879)

REFERENCES. Mayr, 19320 : 26. See Koopman, 1957.

RANGE. Guadalcanal, Savo, Florida Is.

SPECIMENS. Guadalcanal : Tenaru 2 $ ad, 2 $ juv ; Betilonga 9 $ ad, i $ imm,

4 $ ad, i ? juv ; Turipava I <$ ad, I $ imm, i $ juv, i $ ad, i ? imm, I ? juv. i $
ad, i $ ad, i ? imm to OUM.

Weight 12 (J ad 11-14 (12-8) ', 5 c? imm 12-15 (13-6) ; 3 $ juv 10-12-5 (n*3) i

5 $ ad 10-12-5 ( II<2 )- Wing n $ ad 63-68-5 (65-6) ; 2 $ imm 64, 65 ; 2 ^ juv 62,
63-5 ; 4 $ ad 57-60 (58-7). Tail 8 ^ ad 44-46 (44-8) ; 2 <$ imm 43, 44 ; 3 $ juv
37-5-42 (40-2) ; 4 $ ad 38-4'5 (39*4)- Biu IZ c? ad 20 ~ 2 3 (21-9) > 2 imm 20-5,
22 ; 3 juv 19-5-20 (19-7) ; 5 $ ad 19-5-20-5 (20-2).

Iris dark brown. Bill black, $ gape sometimes dull yellow ; juv base (+ i tip)
dull yellow, gape bright yellow. Foot grey, bluish or greenish to fuscous, soles
yellowish.

The immature plumage is considerably laxer than the adult, especially where it is

74 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

most ochraceous in colour (rump, belly, upper wing-coverts). The juvenile plumage
is very similar, but somewhat laxer still, duller on nape and interscapulars, and with
the black on the head still duller and more restricted.

(M. 127) Myzomela tristrami Ramsay, 1882

REFERENCES. Mayr, 19320; : 29. See Koopman, 1957.

RANGE. San Cristoval, Santa Ana.

SPECIMENS. San Cristoval : Manewiriwiri 5 <$ ad, i <$ imm, 5 $ ad, 2 ? imm ; Goge
12 $ ad, 4 c? imm, 5 $ ad, i ? juv ; Nagasi 4 $ ad, 3 <$ imm, i $ juv, i ? imm. i < ad,
i $ imm, i $ ad to OUM.

Weight 21 # ad 12-5-16-5 (14-4) ; 8 <J imm 12-5-15-5 (14-1) ; 10 $ ad 11-5-13
(12-3) ; i $ juv 11-5. Wing 18 <$ ad 66-69-5 (67-5) ; 8 <$ imm 64-67 (65-7) ; 8 $ ad
61-5-64-5 (62-9) ; i $ juv 60. Tail 19 < ad 40-46 (44-2) ; 8 $ imm 39-43 (41-6) ;

7 $ ad 40-42-5 (41-1) ; i $ juv 38. Bill 18 $ ad 21-23-5 (22-0) ; 5 <J imm 21-24
(22-1) ; 8 $ ad 18-20-5 (19-7) ; i $ juv 19.

7ns very dark brown. Bill black ; imm and juv basal 2/3 yellow, greyish above,
gape brighter. Foot black or dark slate, sometimes bluish, < ad average darkest.

What appears to be the juvenile plumage is somewhat softer than the immature,
and less blackish below, but not markedly more olivaceous (cf. Mayr, 19320).

NOTES. It seems unwise to associate this form with M. nigrita of the New Guinea
region (Stresemann, 1923 : 51 ; Mayr, 19320 ; Koopman, 1957) on the single all-
obscuring character of total melanism. It may rather be a member of the lafargei
superspecies ; especially since its nearest neighbour, M. melanocephala, also lacks red
and shows little sexual dimorphism. The fact that two specimens (from Sant Ana
Mayr, 19320) have red on the rump (like M. eichhorni and M. malaitae), and not on
the forehead and throat (as in the females of many races of M . nigrita) nor the crown
(as in the male of M. n.forbesi), supports this conclusion, as does intensive study by
Salomonsen (personal communication).

Ramsay's (i882c : 27) record for Ugi refers only to Morton's account, unsupported
by specimens. The species has not since been found there (Mayr, 1932 ; Cain &
Galbraith, 1956), and is not to be expected on a small island in company with both
M. cardinally and Nectarinia jugularis.

(M. 129) Meliphaga inexpectata (Hartert, 1929)

REFERENCES. Hartert, 1929 : 8 ; Mayr, 19320 : 13.

RANGE. Guadalcanal (mountains).

SPECIMENS. Guadalcanal : Turipava 6 $ ad, i $ imm, 10 $ ad, 3 $ imm. i <$ ad,
I $ ad to OUM.

Weight 6 ^ ad 41-5-48-5 (46-8) ; i 3 imm 40 ; 10 $ ad 37'5-44'5 (4 0>I ) '> 3 ? im m
33-5-34 (33*8). Wing 5 $ ad 109-5-115 (112-9) > * 3 imm IO 4 i 10 ? ad 96-102-5
(100-0) ; 3 $ imm 95-97-5 (95-2). Tail 4 ^ ad 95-96 (95-4) ; i <J imm 91-5 ; 7 ? ad
76-86 (82-8) ; 3 $ imm 79-80-5 (79-7). Bill 4 ad 31-5-35 (32-9) ; i <2 imm 32 ;

8 $ ad 28-31 (29-6) ; 3 $ imm 26-5-31 (28-2).

LAND BIRDS OF GUADALCANAL 75

7ns brown, light to dark (once " black "), sometimes reddish ; imm dark grey-
brown. Bill black. Foot blue-grey, sometimes pale or silvery, claws fuscous ; imm
shanks fuscous-tinged.

The rectrices of the adult females are narrower than those of adult males, but not
as soft and pointed as those of immatures.

NOTES. We follow the generic assignment of Salomonsen (in litt,}, who places
" Meliphaga " bougainvillei of the Bougainville mountains in Melilestes. With the
discovery of a new genus and species in the mountains of New Britain (Gilliard,
19600), the distribution of large honeyeaters in Northern Melanesia (except for the
Philemon moluccensis superspecies) becomes even more puzzling : Vosea white-
manensis in the mountains of New Britain, Melilestes bougainvillei in the mountains
of Bougainville, Meliphaga inexpectata in the mountains of Guadalcanal, and Meliar-
chus sclateri on San Cristoval.

(M. 130) Meliarchis sclateri (Gray, 1870)

REFERENCES. Cat. Birds B.M. 9 : 279 ; Mayr, 19320 : 12.
RANGE. San Cristoval.

SPECIMENS. San Cristoval : Goge 13 ^ ad, i ^ imm, 4 $ ad, 2 $ juv ; Nagasi 2 3
ad, 2 <$ imm, 5 $ ad. i # ad, i ? ad to OUM.

Weight 14 $ ad 75-5-88 (80-2) ; 3 3 imm 68-77 (72'5) ', 9 ? ad 51-5-58-5 (55-6) ;

1 $ juv 45. Wing 4 3 ad 122-136 (129-5) ; 3 ^ imm 121-123 -5 (122-5) .' 6 $ ad 107-115
(in-i). Tail 7 c? ad 104-5-110-5 (108-5) '> 3 3 i mm 99~ I0 2'5 (100-5) i 7 $ ad 87-97
(93-1). Bill 12 3 ad 39-42 (40-5) ; 3 3 imm 39-5-41 (40-2) ; 8 2 ad 34'5-36'5 (35 '4) '>

2 $ juv 30, 30.

Iris pale greenish grey or brown, dark brown centrally ; juv dark brown. Skin
round eye pale blue-grey, eyelid whitish tinged with yellow, greenish or fuscous.
Bill silvery greenish or straw distally, shading to silvery blue-grey basally, gape
whitish with yellow inner surface ; imm and juv more or less streaked with black.
Foot pale blue-grey, shank sometimes yellow-washed.

Besides the characters noted by Mayr (19320), immatures differ from adults in
lacking bristly black shafts in the throat plumage, lacking the blackish subterminal
patches which give a scaled effect to the mid-back, and having the back browner and
the rump more rufous. The juveniles differ from immatures in having softer plumage,
especially ventrally and on the crown and nape, the centres of the dorsal feathers less
blackish and their edges less yellowish, and the ventral shaft-streaking still less
distinct.

(J) Myzantha melanocephala subspecies

RANGE. Eastern and southeastern Australia and Tasmania ; introduced to Three
Sisters (French, 1957).

(M. 132) Zosterops ugiensis oblita Hartert, 1929

REFERENCES. Mees, 1961 : 162.
RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Betilonga 15 ^ ad, i ^ juv, 9 $ ad, i ? ad, 2 ? juv ;
Turipava 6 3 ad, 4 ad. i <J ad to OUM.

76 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

Weight Betilonga 16 <$ ad 13-5-16-5 (14-9) ; i <$ juv 14-5 ; 9 ? ad 15-17 (16-4) ;
Turipava 6 ^ ad 13-16 (15-1) ; 4 <j> ad 13-5-18 (15-6). Wing Betilonga 15 <J ad
64-69 (66-0) ; 8 $ ad 63-5-69 (65-4) ; Turipava 5 <$ ad 66^70-5 (68-3) ; 4 $ ad 62-66
(64-0). Tail Betilonga 14 < ad 44-5-48-5 (46-5) ; 7 $ ad 44-47 (45-4) ; Turipava
5 <J ad 46-5-49 (48-1) ; 3 ? ad 44-46 (44-8). Bill Betilonga 13 ad 15-16-5 (16-1) ;
i < juv 14-5 ; 8 $ ad 15-16-5 (16-1) ; Turipava 5 <? ad 15-5-16-5 (16-1) ; 3 $ ad
16-16-5 (16-2).

7ns light brown ; juv grey-brown to light brown. Skin round eye dark grey. Bill
black, base of mandible whitish ; juv maxilla blackish streaked with pale horn,
mandible whitish or pale horn streaked with blackish, gape flesh. Foot bluish grey,
often with yellowish or fuscous tinge.

The juveniles differ from adults in having softer plumage, more narrowly-pointed
rectrices, and primaries with less falcation, and in more dilute carotenoids (under
tail-coverts paler yellow, olive duller and more citrine especially on outer edges of
remiges), olive edges continued to the tips of rectrices, and lower cheeks without
grey wash.

NOTES. The geographical and altitudinal distribution of species of Zosterops in
the Solomons is puzzling (despite Cain, 19566 : 13).

(M. 132) Zosterops ugiensis ugiensis (Ramsay, 1882)

REFERENCES. Mees, 1961 : 158 (rendovae).

RANGE. San Cristoval.

SPECIMENS. San Cristoval : Nagasi 5 $ ad, i <$ juv, 4 $ ad. i ^ ad to OUM.

Weight 5 J ad 18-5-19-5 (18-9) ; i # juv 17 ; 4 $ ad 17-5-21-5 (19-5). Wing 5
ad 68-70 (69-1) ; i $ juv 64 ; 4 $ ad 66-68 (66-6). Tail 5 <$ ad 43-45-5 (44-5) ; 4 $
ad 40-5-44 (42-3). Bill 5 <? ad 16-5-18 (17-3) ; i J juv 17 ; 4 $ ad 16-5-18 (17-1).

Iris light brown. Bill black, extreme base of mandible dull fuscous to ivory ; juv
base of mandible grey. Foot bluish to greenish or fuscous grey, sometimes patchy,
toes average darker and bluer.

The plumage of the juvenile is slightly softer than that of adults, but much firmer
than that of juvenile oblita, and it has growing rectrices of adult shape. It differs in
colour from adults in having the brownish wash on the forehead (very variable in
intensity between adults) much reduced though discernible, and the dorsal olive duller
and more citrine.

NOTES. Mees (1955 ; 1961 : 160-162) has pointed out that the name rendovae,
which had always been applied to the quite different Zosterops on Rendova, is the
valid name for the present species under the International Rules. Arguments put
forward to disprove this (Galbraith, 1957) are incorrect (China in lilt. I.C.Z.N. ref.
Z.N.(S)i223). While the other objections to the change are largely irrelevant, the
cardinal point remains : transfer of a name in use for one taxon to another is highly
undesirable, because of the confusion it is likely to cause, in a way that simple
adoption of an unfamiliar synonym is not. It is quite untrue that name-transfer in
the present case would affect only museum workers (Mees, 1961), even if we accept a
pessimistic view of future field work in the Solomons. The name Zosterops rendovae

LAND BIRDS OF GUADALCANAL 77

has been repeatedly used for the polytypic species or superspecies in the New Georgia
Group, in widely-read works on evolution (e.g. Mayr, 1942 ; Lack, 1947 ; Cain,
19540).

In order to avoid this transfer, I. C. J. G. will apply to the International Commission
for the suppression of the name rendovae under the plenary powers, an action approved
by Mees (1961). Use of the name rendovae in either sense would now be objectionable,
and in order to avoid confusion a favourable decision by the Commission is here
anticipated. Mees (1961 : 143-151) should be followed for the nomenclature of the
New Georgia Group superspecies, and Mayr (19450) for that of the present species and
race.

(M. 136) Dicaeum aeneum becki Hartert, 1929

REFERENCES. Mayr, 1955 : 39 ; Salomonsen, 1960 : 33. See Mayr & Amadon,
1947.

RANGE . Guadalcanal .

SPECIMENS. Guadalcanal : Betilonga 19 $ ad, 5 <$ imm, 10 $ ad, i ? imm ; Nala
i $ ad ; Turipava I ^ ad. i ^ ad, i $ ad to OUM.

Weight 22 $ ad 8-5-10-5 (9-6) ; 5 < imm 9-0-10-5 (9-5) ; 9 $ ad 8-9-5 (8-6). Wing
20 $ ad 50-5-56-5 (52-9) i 5 3 imm 50-5-54 (52-4) ; 8 ? ad 49-53 (49-9). Tail 15 $
ad 26-29 ( 2 7' 2 ) '> 3 $ i mm 2 5~ 2 6 ( 2 5'5) i 5 $ ad 23-5-26 (24-8). Bill 14 ad 12-13-5
(i 2 '5) : 5 3 imm 11-5-12-5 (12-2) ; 7 $ ad 11-5-12-5 (12-2).

Iris dark brown, i <$ ad light brown, 3 $ ad dark grey-brown. Bill black, ^ base of
mandible often grey to whitish, $ bill sometimes greyish or fuscous, and more or less
of mandible always whitish ; imm gape and more or less of mandible orange. Foot
dark grey to black, often fuscous, sometimes bluish.

The females have a more or less conspicuous hoary loral spot, and (where the orbital
area is well-preserved) a narrow white circlet round the eye reminiscent of Zosterops
species. Neither character is shown by any male (except that an adult sexed as a
male with unenlarged testes, and two unsexed adults, agree with the females in all
characters, and have been included with them) .

The plumage of the immatures is somewhat scantier on flanks and rump than that
of adults, but not otherwise different in texture or the shape of remiges or rectrices.
The immature female is distinguished from adult females only by its orange-based
bill ; while immature males also have less red on the breast, more white on the lower
throat, the grey of the lower breast paler and more olivaceous, the olive of the flanks
duller, and the dorsal green gloss somewhat duller and bluer and more sharply
restricted to the feather-tips.

NOTES. The population on Malaita is a distinct race, D. a. malaitae Salomonsen
(1960 : 34), not simply intermediate between aeneum and becki (cf. Mayr, 1955).

(M. 137) Dicaeum tristrami Sharpe, 1883

REFERENCES. Sharpe, 1883 : 579 ; Rothschild & Hartert, 19086 : 364. See Mayr
& Amadon, 1947 ; Salomonsen, 1960.
RANGE. San Cristoval.

78 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

SPECIMENS. San Cristoval : Manewiriwiri i $ ad ; Goge 10 ^ ad, i ^ imm, 4 $
ad, 2 $ imm ; Nagasi 4 ^ ad, 3 $ ad. I ^ ad, I $ ad to OUM.

Weight 14 <$ ad 11-13 (11-9) ; i <$ imm n ; 8 $ ad 10-12-5 (11-4) ; 2 $ imm 10-5,
IT. Wing 12 <J ad 56-5-60-5 (58-9) ; i $ imm 56-5 ; 8 $ ad 54'5-57'5 (55 '9) > % ? imm
55'5. 56-5- Tail 7 ^ ad 27-30 (28-8) ; i imm 27 ; 6 $ ad 25-5-28-5 (27-3) ; 2 $
imm 26-5, 28. *7/ 12 $ ad 12-5-13-5 (13-1) ; i $ imm 13 ; 8 $ ad 11-5-13 (12-5) ;
2 $ imm 12, 12-5.

7m very dark brown. Bill black ; imm pale in dried skin, gape and more or less of
mandible dull orange-yellow or pale yellow. Foot black to dull slate, often brownish
or bluish especially on shanks.

The sexes are not quite alike, but differ in the distribution of white on the head
(Rothschild & Hartert, 19086 ; despite Mayr & Amadon, 1947 ; Salomonsen,
1960 : 35). Males have a white area behind and below the auriculars (which it may
invade to some extent), no white on the chin and upper throat, but whitish scaling
on the lower throat and breast. Females have the auriculars themselves and the
chin and upper throat white, while the brown breast-shield is unsealed but smaller.

The immatures have sparsely-feathered chins and distinctly softer plumage than
adults, and might be thought much younger than those of D. aeneum becki. However,
they do not agree with Rothschild & Hartert's (19086) description, which may refer
to juveniles. They differ in pattern from adults in being more uniform, without the
sharp contrast between dark crown and pale mantle, with the white edges on the
head confined to the forehead, with the breast-shield paler and greyer and less
sharply distinct from the whitish underparts, and (in the female) with little white on
the throat.

NOTES. It must be emphasized that this species can be associated with the super-
species to which D. aeneum belongs (by Mayr & Amadon, 1947 : 21 ; Mayr, 1955 : 39)
solely on the geographical grounds that only the erythrothorax-hirundinaceum complex
occurs in the Australo-Papuan area. Apart from its large size, heavy bill and total
lack of carotenoids, D. tristrami shows a pattern of melanin distribution unique in the
family, though somewhat reminiscent of certain western Dicaeum (e.g. D. rectrocinc-
tum) and even of Oreocharis arfaki. It seems best to regard it as a relict species of
unknown affinities, which by competition excludes D. aeneum from San Cristoval
(cf. Salomonsen, 1960 : 17). It cannot be associated with any particular species ;
certainly not with D. eximium merely because both are somewhat phaeomelanic.

(M. 138) Erythrura trichroa woodfordi Hartert, 1900

REFERENCES. Mayr, 1931^ : 5. See Delacour, 1943.

RANGE. Guadalcanal.

SPECIMENS. Guadalcanal : Betilonga i ^ ad, i $ imm.

Weight i <$ ad 15-5 ; i $ imm 16. Wing i $ imm 61. Tail i $ imm 35. Bill i $
ad 15, i $ imm 14.

Iris dark brown. Bill black ; imm base of mandible whitish. Foot flesh, fuscous-
washed ; imm pale straw.

Both specimens are rather dark green dorsally.

LAND BIRDS OF GUADALCANAL 79

ACKNOWLEDGEMENTS

We record with pleasure our indebtedness to the following for assistance in the
preparation of this paper. Dr. H. W. Parker, C.B.E., Dr. F. C. Fraser, C.B.E. and
Mr. J. D. Macdonald gave exceptional facilities for study at the British Museum
(Natural History). Dr. Dean Amadon lent specimens from the American Museum
of Natural History, and Dr. H. Vols0e lent a type from the Universitets
Zoologiske Museum, K0benhavn. Dr. Amadon, Mrs. Diana Bradley, Dr. A. J. Cain,
Dr. W. E. China, C.B.E., Mr. William French, Mr. Derek Goodwin, Mr. J. C.
Grover, Prof. Ernst Mayr, Mr. S. A. Parker, Mr. P. A. Pudsey-Dawson, Dr. Finn
Salomonsen, Dr. Ernst Sutter, Mr. R. B. M. Thompson, Dr. E. G. Turbott, the
late Mr. E. N. T. Vane, Mrs. Patricia Vaurie, and the Chief Secretary to the
Western Pacific High Commission, provided information. Miss F. E. Duckett
assisted in checking and proof-reading.

The Directorate of Overseas Surveys and the Geological Survey of the British
Solomon Islands gave permission for maps 2 and 3 to be compiled from copyright
sources.

REFERENCES

The typography of titles is made uniform throughout, and the collective title " Birds collected
during the Whitney South Sea Expedition, no. " is abbreviated to " Whitney, ". Issue
numbers are given in parentheses, where each comprises a single paper or group of related papers.

ALEXANDER, W. B. 1954. Birds of the Ocean. A handbook for voyagers. 2nd ed. London &

New York ; Putnam : xiv + 306 pp.

AMADON, D. 1941. Notes on some Australian birds of prey. Emu, 40 : 365-384.
1942^. Whitney, 49. Notes on some non -passerine genera, i. Amer. Mus. Novit. (1175) :

I II.
19426. Whitney, 50. Notes on some non-passerine genera, 2. Amer. Mus. Novit. (1176) :

1-2 1.

I943. Whitney, 52. Notes on some non-passerine genera, 3. Amer. Mus. Novit. (1237) :

1-22.

I943&. The genera of starlings and their relationships. Amer. Mus. Novit. (1247) : 1-16.

1949. The seventy-five per cent rule for subspecies. Condor 51 : 250-258.

1951. Taxonomic notes on the Australian butcher-birds (family Cracticidae). Amer. Mus.

Novit. (1504) : 1-33.

1953- Further remarks on the Cracticidae. S. Aust. Orn. 21 : 6-7.

1956. Remarks on the starlings, family Sturnidae. Amer. Mus. Novit. (1803) : 1-41.

& WOOLFENDEN, G. 1952. Notes on the Mathews' collection of Australian birds. The

order Ciconiiformes. Amer. Mus. Novit (1564) : 1-16.
ANON. 1952. Pelicans blown into the Solomons. Pacific Islands Monthly, 23 (2) : 78. Sydney.

[n.v.]
AUBER, L. 1938. Die Rassen- und Artenkreise des Genus Eos Wagler (Aves). Bin taxonomisch-

phylogenetische Studie. Festschrift zum 60. Geburstage von Professor Dr. Embrik Strand.

4 : 673-782. Riga ; Latvija.
1957- The distribution of structural colours and unusual pigments in the class Aves.

Ibis, 99 : 463-476.
AUSTRALIA 1950-1951. The resources of the Territory of Papua and New Guinea. Canberra;

Dept. Nat. Devel., 1 : 150 pp. + corrigenda ; 2 : 9 maps + vi pp.
B.H.I. 1942. Carte generale bathymetrique des oceans. 3rd ed. Sheet AHI- Monte-Carlo ;

Bureau Hydrographique International.
BAKER, E. C. S. 1926. The fauna of British India. Birds, 3 London ; Taylor & Francis :

xx + 489 pp.

8o IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

BAKER, R. H. 1948. Report on collections of birds made by United States Naval Medical

Research Unit No. 2 in the Pacific war area. Smithson. misc. Coll. 107 (15) : 1-74.
BEECHER, W. J. 1945. A bird collection from the Solomon Islands. Fieldiana, Zool. 31 (4) :

31-37-
BEMMEL, A. C. V. VAN. 1948. A faunal list of the birds of the Moluccan Islands. Treubia, 19 :

323-402.
BERGER, A. J. 1955. On the anatomy and relationships of glossy cuckoos of the genera Chryso-

coccyx, Lampromorpha, and Chalcites. Proc. U.S. nat. Mus. 103 : 585-597.
BOCK, W. J. 1956. A generic review of the family Ardeidae (Aves). Amer. Mus. Novit. (1799) :

1-49.
BOGERT, C. 1937. Whitney, 34. The distribution and the migration of the Long-tailed Cuckoo

(Urodynamis taitensis Sparrman). Amer. Mus. Novit. (933) : 1-12.
BRADLEY, D. 1957. Birds of the Solomon Islands. Ibis, 99 : 352-353.

1962. Additional records of birds from Rennell and Bellona Islands. The Natural

History of Rennell Island, British Solomon Islands, 4 (32) : 11-12. Copenhagen; Danish
Science Press ; ed. Wolff.

& WOLFF, T. 1956. The birds of Rennell Island. The natural history of Rennell Island,

British Solomon Islands, 1 (7) : 85-120. Copenhagen ; Danish Science Press ; ed. Wolff.
BRAESTRUP, F. W. 1956. The significance of the strong "oceanic" affinities of the vertebrate

fauna on Rennell Island. The natural history of Rennell Island, British Solomon Islands,

1 (9) : I 35- I 48- Copenhagen ; Danish Science Press ; ed. Wolff.
BULL, P. C. 1948. Field notes on waders in the south-west Pacific with special reference to the

Russell Islands. Emu, 47 : 165-176.

CAIN, A. J. I954&. Animal species and their evolution. London; Hutchinson : 190 pp.
19546. Subdivisions of the genus Ptilinopus (Aves, Columbae). Bull. Brit. Mus. (nat.

Hist.} Zool. 2 (8) : 265-284.
19546. Affinities of the fruit-pigeon Ptilinopus perousii Peale. Ibis, 96 : 104-110.

1955- A revision of Trichoglossus haematodus and of the Australian platycercine parrots.

Ibis, 97 : 432-479.

1956^. The genus in evolutionary taxonomy. Syst. Zool. 5 : 97-109.

19566. Classification. The ornithologists' guide : 10-14. London ; Brit. Orn. Union ; ed.

Hutson.

1957- Range-changes and differential selection in fruit-pigeons of the Ptilinopus purpuratus

species group. Proc. Pacific Sci. Congr. (8) 3 A : 1393-1412. Quezon ; Nat. Res. Council
Philippines.

& GALBRAITH, I. C. J. 1955. Five new subspecies from the mountains of Guadalcanal

(British Solomon Islands). Bull. Brit. Orn. Cl. 75 : 90-93.

1956. Field notes on birds of the eastern Solomon Islands. Ibis, 98 : 100-134, 262-295.

1957- [Birds of the Solomon Islands]. Ibis, 99 : 128-130.

CONDON, H. T. 1951. Notes on the birds of South Australia : occurrence, distribution and
taxonomy. 5. Aust. Orn. 20 : 26-68.

& AMADON, D. 1954. Taxonomic notes on Australian hawks. Rec. S. Aust. Mus. 11 :

189-246.

D.C.S. 1955. Sketch map Guadalcanal (British Solomon Islands Protectorate] i : 50,000. London ;

Directorate of Colonial Surveys ; D.C.S. (Misc.) 91 : 14 sheets.
D.O.S. 1958. British Solomon Islands San Cristoval sketch map i : 50,000. London ; Directorate

of Overseas Surveys ; D.O.S. 456 (Series 714) : 8 sheets.
DAHL, F. 1899. Das Leben der Vogel auf den Bismarckinseln nach eigenen Beobachtungen

vergleichend dargestellt. Mitt. zool. Mus. Berl. 1 (3) : 107-222.
DANIS, V. 1938. Etude d'une nouvelle collection d'oiseaux de 1'Ile Bougainville. Bull. Mus.

Hist, nat., Paris (2) 10 : 43-47.
DAVIDSON, M. E. M. 1929. On a small collection of birds from Torres Strait islands, and from

Guadalcanal Island, Solomon group. Proc. Calif. Acad. Sci. (4) 18 (10) : 245-260,

LAND BIRDS OF GUADALCANAL 81

DAVIDSON, M. E. M. 1934. The Templeton Crocker Expedition to western Polynesian and

Melanesian islands, 1933. No. 16. Notes on the birds. Proc. Calif. Acad. Sci. (4) 21 (16) :

189-198.
DELACOUR, J. 1943. A revision of the subfamily Estrildinae of the family Ploceidae. Zoologica,

N. Y. 28 : 69-86.

1944. Revision of the family Nectariniidae (sunbirds). Zoologica, N.Y. 29 : 17-38.

1946. Notes on the taxonomy of the birds of Malaysia. Zoologica, N.Y. 31 : 1-8.

& MAYR, E. 1945. The family Anatidae. Wilson Bull. 57 : 3-55.

DONAGHHO, W. R. 1950. Observations on some birds of Guadalcanal and Tulagi. Condor,

52 : 127-132.

DORST, J. 1954. La Pelican a Lunettes aux Nouvelles-Hebrides. Oiseau, Paris (n.s.) 24 : 149-151.
FELL, H. B. 1947. The migration of the New Zealand Bronze Cuckoo Chalcites lucidiis lucidus

(Gmelin). Trans, roy. Soc. N.Z. 76 : 504-515.

FRENCH, W. 1957. Birds of the Solomon Islands. Ibis, 99 : 126-127.
FRITH, H. J. 1956. Breeding habits in the family Megapodiidae. Ibis, 98 : 620-640.
GALBRAITH, I. C. J. 1956. Variation, relationships and evolution in the Pachycephala pectoralis

superspecies (Aves, Muscicapidae). Bull. Brit. Mus. (nat. Hist.} Zool. 4 (4) : 131-222.
1957- On the application of the name Zosterops rendovae Tristram, 1882. Bull. Brit. Orn.

Cl. 77 : 10-16.
GILLIARD, E. T. I96o. Results of the 1958-1959 Gilliard New Britain Expedition, i. A new

genus of honeyeater (Aves). Amer. Mus. Novit. (2001) : 1-5.

19606. Results of the 1958-1959 Gilliard New Britain Expedition. 2. A new species of

thicket warbler (Aves, Cichlornis) from New Britain. Amer. Mus. Novit. (2008) : 1-6.

GOODWIN, D. 1959. Taxonomy of the genus Columba. Bull. Brit. Mus. (nat. Hist.) Zool.

6 (i) : 1-23.

1960. Taxonomy of the genus Ducula. Ibis, 102 : 526-535.

GRAY, G. R. 1873. Birds. In Brenchley, Jottings during the cruise of H.M.S. Curacoa among

the South Sea islands in 1865 : 353-394. London ; Longmans, Green & Co.
GROVER, J. C. 1955. Geology, mineral deposits and prospects of mining development in the British

Solomon Islands Protectorate. Mem. interim geol. Surv. B.S.I, (i). London ; Crown Agents :

ix + 1 08 pp.

1957- Some geographical aspects of the British Solomon Islands in the western Pacific.

Geogr. J. 123 : 298-317.

1958. The Solomon Islands geological exploration and research 1953-1956. Chapters by

Pudsey-Dawson, P. A. & Thompson, R. B. M. Mem. geol. Surv. B.S.I. (2). London ; Crown
Agents : viii + 151 pp.

(ed.) 1960. The British Solomon Islands geological record 1957-1958. Reports on investiga-
tions into the geology and mineral resources of the Protectorate. London ; Crown Agents :
v + 113 pp.

GURNEY, J. H. 1884. A list of the diurnal birds of prey, with references and annotations ; also a

record of specimens preserved in the Norfolk and Norwich Museum. London ; van Voorst :

xv + 187 pp.
HACHISUKA, M. 1931. Notes sur les oiseaux des Philippines, i. Revision des colombes terrestres.

Oiseau, Paris (n.s.) 1 : 23-29.
HAMLIN, H. 1931. Whitney, 14. Notes on the geography of Rennell Island and the ecology of its

bird life. Amer. Mus. Novit. (488) : 1-6.
HARTERT, E. 1896. Contributions to the ornithology of the Papuan islands. 3. On the forms of

Macropteryx mystacea. Novit. Zool. 3 : 19-20.

1908. [Exhibition of birds from the Solomon. Islands.] Bull. Brit. Orn. Cl. 21 : 105-107.

19240. The birds of New Hanover. Novit. Zool. 31 : 194-213.

19246. The birds of St. Matthias Island. Novit. Zool. 31 : 261-275.

1926^. On the birds of Feni and Nissan Islands, east of South New Ireland. Novit. Zool.

33 : 33-48.
19266. On the birds of the district of Talasea in New Britain. Novit. Zool. 33 : 122-145.

82 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

HARTERT, E. 1929. Whitney, 8. Notes on birds from the Solomon Islands. Amer. Mus. Novit.

(364) : 1-19.
HEINROTH, O. 1903. Ornithologische Ergebnisse der "I. Deutschen Siidsee-Expedition von Br.

Mencke " [concluded]. /. Orn. 51 : 65-125.
I.C.Z.N. 1956. Direction 43. Opin. int. Comm. zool. Nom. 1 (D.y.) : 171-210.

1958. Official list of specific names in zoology, (i). London ; Int. Tr. Zool. Nomen ; ed.

Hemming : xvi + 206 pp.

KEAST, A. 1957. Variation in the Australian kingfishers. (Aves : Alcedinidae). Rec. Aust. Mus.

24 (7) : 61-72.
1958. Variation and speciation in the Australian Campephagidae (Passeres). Aust. J. Zool.

6 : 248-267.

KOOPMAN, K. F. 1957. Evolution in the genus Myzomela (Aves ; Meliphagidae) . ^^,74:49-72.
LACK, D. 1947. Darwin's finches. Cambridge; C.U.P. : x + 208 pp.
LAIRD, M. 1954. Australian Pelicans in the Solomon Islands and New Hebrides. Notornis,

6 : 11-13.

MATHEWS, G. M. 1918. The birds of Australia, 7 (4) : 321-384. London ; Witherby.
1927-1930. Sy sterna avium A ustralasianarum. A systematic list of the birds of the A ustralasian

region. London ; B.O.U. 1 : xvii -f 426 pp.; 2 : iv + 4 2 7- IO 47 PP-
1928. Bird notes. Novit. Zool. 34 : 372-373.

1931. Additions and corrections to the " Systema avium Australasianarum ". Ibis (13)

1 : 44-57-

1932. Additions and corrections to the " Systema avium Australasianarum ". 2. Ibis (13)

2 : 132-161.

1933. Additions and corrections to the " Systema avium Australasianarum ". 3. Ibis (13)

3 : 87-97-
MAYR, E. i93ia. Whitney, 12. Notes on Halcyon chloris and some of its subspecies. Amer. Mus.

Novit. (469) : i-io.
19316. Whitney, 13. A systematic list of the birds of Rennell Island with descriptions of

new species and subspecies. Amer. Mus. Novit. (486) : 1-29.

I93ic. Whitney, 14. The relationships and origin of the birds of Rennell Island. Amer.

Mus. Novit. (488) : 7-11.

1 93 id. Whitney, 15. The parrot finches (genus Erythrura) . Amer. Mus. Novit. (489) : i-io.

1931^. Whitney, 16. Notes on fantails of the genus Rhipidura. Amer. Mus. Novit. (502) :

1-2 1.

I93I/. Whitney, 17. The birds of Malaita Island (British Solomon Islands). Amer. Mus.

Novit. (504) : 1-26.

19320. Whitney, 18. Notes on Meliphagidae from Polynesia and the Solomon Islands.

Amer. Mus. Novit. (516) : 1-30.

19326. Whitney, 19. Notes on the Bronze Cuckoo, Chalcites lucidus, and its subspecies.

Amer. Mus. Novit. (520) : 1-9.
19326. Whitney, 20. Notes on thickheads (Pachycephala} from the Solomon Islands. Amer.

Mus. Novit. (522) : 1-22.

1932^. Whitney, 21. Notes on thickheads (Pachycephala} from Polynesia. Amer. Mus.

Novit. (531) : 1-23.

i933. Whitney, 22. Three new genera from Polynesia and Melanesia. Amer. Mus. Novit.

(590) : 1-6.
!9336. On a collection of birds, supposedly from the Solomon Islands. Ibis (13) 3 : 549-552.

i934- Whitney, 28. Notes on some birds from New Britain, Bismarck Archipelago.

Amer. Mus. Novit. (709) : 1-15.

19346. Whitney, 29. Notes on the genus Petroica. Amer. Mus. Novit. (714) : 1-19.

1935. Whitney, 30. Descriptions of twenty-five new species and subspecies. Amer. Mus.

Novit. (820) : 1-6.
1936. Whitney, 31. Descriptions of twenty-five species and subspecies. Amer. Mus. Novit.

(828) : 1-19.

LAND BIRDS OF GUADALCANAL 83

MAYR, E. 1937. Whitney, 33. Notes on New Guinea birds, i. Amev. Mus. Novit. (915) : 1-19.

- 1938^. Whitney, 39. Notes on New Guinea birds. 4. Amer. Mus. Novit. (1006) 1-16.

- 19386. Whitney, 40. Notes on New Guinea birds. 5. Amer. Mus. Novit. (1007) 1-16.

- 19403. Whitney, 41. Notes on New Guinea birds. 6. Amer. Mus. Novit. (1056) 1-12.

- 19406. Whitney, 43. Notes on New Guinea birds. 7. Amer. Mus. Novit. (1091) : 1-3.

- 19413. List of New Guinea birds. A systematic and faunal list of the birds of New Guinea and
adjacent islands. New York ; A.M.N.H. : xi + 260 pp.

-19416. [recte 1940]. Borders and subdivision of the Polynesian region as based on our
knowledge of the distribution of birds. Proc. Pacific Sci. Congr. (6) 4 : 191-195. Berkley &
Los Angeles ; U. Calif. P.

- I94ic. Whitney, 45. Notes on New Guinea birds. 8. Amer. Mus. Novit. (1133) : 1-8.
-19410". Whitney, 47. Notes on the genera Halcyon, Turdus and Eurostopodus. Amer. Mus.

Novit. (1152) : 1-7.

- 1942. Systematics and the origin of species from the viewpoint of a zoologist. Columbia bio-
logical series (13). New York ; Columbia U.P. : xiv + 334 PP-

- 19443. The birds of Timor and Sumba. Bull. Amer. Mus. nat. Hist. 83 (2) : 129-194.

- 19446. Timor and the colonization of Australia by birds. Emu, 44 : 113-130.

- 1944^. Whitney, 54. Notes on some genera from the southwest Pacific. Amer. Mus. Novit.
(1269) : 1-8.

- 19443". Wallace's Line in the light of recent zoogeographic studies. Quart. Rev. Biol.
19 : 1-14.

- 19453. Birds of the southwest Pacific. A field guide to the birds of the area between Samoa, New
Caledonia, and Micronesia. New York ; Macmillan : xix + 316 pp.

- 19456. Whitney, 55. Notes on the birds of Northern Melanesia, i. Amer. Mus. Novit.
(1294) : 1-12.

- 1948. Geographical variation in the Reed-Warbler. Emu, 47 : 205-210.

- 19493. Artbildung und Variation in der Halcyon-chloris-Gruppe. Ornithologie als biologische
Wissenschaft : 55-60. Heidelberg : Universitatsverlag ; eds. Mayr & Schiiz.

- 19496. Whitney, 57. Notes on the birds of Northern Melanesia. 2. Amer. Mus. Novit.
(1417) : 1-38.

- 1955. Whitney, 63. Notes on the birds of Northern Melanesia. 3. Amer. Mus. Novit.
(1707) : 1-46.

- 1957. Whitney, 64. Notes on the birds of Northern Melanesia. 4. Amer. Mus. Novit.
(1823) : 1-14.

- & AMADON, D. 1941. Whitney, 46. Geographical variation in Demigretta sacra (Gmelin).
Amer. Mus. Novit. (1144) : i-n.

1947. A review of the Dicaeidae. Amer. Mus. Novit. (1360) : 1-32.

& CAMRAS, S. 1938. Birds of the Crane Pacific Expedition. Publ. Field Mus. (Zool.)

20 (34) : 453-473-

& MOYNIHAN, M. 1946. Evolution in the Rhipidura rufifrons group. Amer. Mus. Novit.

(1321) : i -2 1.
& RAND, A. L. 1937. Results of the Archbold Expeditions. 14. The birds of the 1933-1934

Papuan Expedition. Bull. Amer. Mus. nat. Hist. 73 (i) : 1-248.
& RIPLEY, S. D. 1941. Whitney, 44. Notes on the genus Lalage Boie. Amer. Mus. Novit.

(1116) : 1-18.
& SCHAUENSEE, R. M. DE. 1939. Zoological results of the Denison-Crockett Expedition to

the South Pacific for the Academy of Natural Sciences of Philadelphia, 1937-1938. i. The

birds of the island of Biak. Proc. Acad. nat. Sci. Philad. 91 : 1-37.

& VAURIE, C. 1948. Evolution in the family Dicruridae (birds). Evolution, 2 : 238-265.

MEES, G. F. 1955. The name of the white-eye from Rendova Island (Solomon Islands). Zool.

Meded. 33 (28) : 299-300.

1961. A systematic review of the Indo- Australian Zosteropidae. 2. Zool. Verh., Leiden

(50) : 1-168.

84 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

MEYER, O. 1927. Zur Ornithologie des Bismarckarchipels. Orn. Mber. 35 : 112-113.

- 1930. Uebersicht iiber die Brutzeiten der Vogel auf der Insel Vuatom (New Britain).
/. Orn. 78 : 19-38.

I 933- Vogeleier und Nesten aus Neubritannien, Siidsee. Beitr. FortPflBiol. Vogel, 9 : 122-
136, 182-185.
1937. Australische Zugvogel im Bismarckarchipel. Orn. Mber. 45 : 48-51.

OGILVIE-GRANT, W. R. 1887. A list of the birds collected by Mr. Charles Morris Woodford in
the Solomon Archipelago. Proc. zool. Soc. Land. 1887 : 328-333.

1888. Second list of the birds collected by Mr. C. M. Woodford in the Solomon Archipelago.
Proc. zool. Soc. Lond. 1888 : 185-204.

- 1914. [Description of two new subspecies of kingfishers from the Solomon Islands.] Bull.

Brit. Orn. Cl. 35 : 13-14.
PENDLETON, R. C. 1947. Field observations on the Spotted Button-Quail on Guadalcanal.

Auk, 64 : 417-421.

PETERS, J. L. 1935. Remarks on the avian genus Eos. Proc. biol. Soc. Wash. 48 : 67-70.
RAMSAY, E. P. 1 879(3. Notes on zoology of the Solomon Islands, i. Aves. Proc. Linn. Soc.

N.S.W. (i) 4: 65-84.

1 8796. Notes on some recently described birds from the Solomon Islands. Proc. Linn. Soc.
N.S.W. (1)4:313-315.

1880. Contributions to the zoology of New Guinea. 6. On some new and rare birds, from

south east coast of New Guinea, &c. Proc. Linn. Soc. N.S.W. (i) 4 : 464-470.
1881. Notes on the zoology of the Solomon Islands, with descriptions of some new birds. 2.

Proc. Linn. Soc. N.S.W. (i) 6 : 176-181.
1882^. Notes on the zoology of the Solomon Islands, with descriptions of some new birds.

Proc. Linn. Soc. N.S.W. (i) 6 : 718-727.
18826. Descriptions of two new birds from the Solomon Islands. Proc. Linn. Soc. N.S.W.

(i) 6 : 833-835.

i882c. Notes on the zoology of the Solomon Islands. 4. Proc. Linn. Soc. N.S.W. (i) 7 :

16-43.

1882^. Descriptions of two new birds from the Solomon Islands. Proc. Linn. Soc. N.S.W.

(i) 7 : 299-301.
18820. Descriptions of some new birds from the Solomon Islands and New Britain. /. Linn.

Soc. (Zool.), 16 : 128-131.

1883. Notes on birds from the Solomon Islands. Proc. Linn. Soc. N.S.W. (i) 7 : 665-673.
RAND, A. L. 1938. Results of the Archbold Expeditions. 19. On some non-passerine New

Guinea birds. Amer. Mus. Novit. (990) : 1-15.
RIPLEY, S. D. 1941. Notes on the genus Coracina. Auk, 58 : 381-395.

1942. The species Eurystomus orientalis. Proc. biol. Soc. Wash. 55 : 169-176.

1959. Competition between sunbird and honeyeater species in the Moluccan islands.

Amer. Nat. 93 : 127-132.

1961. Aggressive neglect as a factor in interspecific competition in birds. Auk, 78 : 366-371.
& BIRCKHEAD, H. 1942. Whitney, 51. On the fruit pigeons of the Ptilinopus purpuratus

group. Amer. Mus. Novit. (1192) : 1-14.
ROTHSCHILD, W. 1904. [On a new species of Centropus from the Solomon Islands.] Bull. Brit.

Orn. Cl. 14 : 59.

& HARTERT, E. 19010. Notes on Papuan birds. Novit. Zool. 8 : 55-88, 102-162.

igoift. List of a collection of birds from Kulambangra and Florida Islands, in the

Solomons group. Novit. Zool. 8 : 179-189.

igoic. List of a collection of birds from Guadalcanar Island, in the Solomon group.

Novit. Zool. 8 : 373-382.

1902. List of a collection of birds made on Ysabel Island in the Solomon group by Mr.

A. S. Meek. Novit. Zool. 9 : 581-594.

1905. Further contributions to our knowledge of the ornis of the Solomon Islands.

Novit Zool. 12 : 243-268.

LAND BIRDS OF GAUDALCANAL 85

ROTHSCHILD, W. 1907. Notes on Papuan birds. Novit. Zool. 14 : 433-446.

igoSa. The birds of Vella Lavella, Solomon Islands. Novit. Zool. 15 : 351-358.

igoSb. On a collection of birds from San Cristoval, Solomon Islands. Novit. Zool. 15 :
359-365.

1914. The birds of the Admiralty Islands, north of German New Guinea. Novit. Zool.

21 : 281-298.

1916. On some forms of Coracina (Graucalus auct.) from the Solomon Islands. Novit.

Zool. 23 : 289-291.
SALOMONSEN, F. 1960. Notes on flowerpeckers (Aves, Dicaeidae). 3. The species group Dicaeum

concolor and the superspecies Dicaeum erythrothorax. Amer. Mus. Novit. (2016) : 1-36.
SANFT, K. 1960. Aves/Upupae. Bucerotidae. Tierreich (76) : [4] -f- 1-174.
SCHONWETTER, M. 1935- Vogeleier aus Neubritannien. Beitr. FortPflBiol. Vogel, 11 : 129-136.
SCLATER, P. L. 1869. On a collection of birds from the Solomon Islands. Proc. zool. Soc. Land.

1869 : 118-126.

SCOTT, W. E. 1946. Birds observed on Espiritu Santo, New Hebrides. Auk, 63 : 362-368.
SHARPE, R. B. 1883. Notes on some species of birds of thefamily Dicaeidae. Proc. zool. Soc. Lond.

1883 : 578-580.

1888. Descriptions of some new species of birds from the island of Guadalcanar in the Solo-
mon Archipelago, discovered by Mr. C. M. Woodford. Proc. zool. Soc. Lond. 1888 : 182-185.
SIBLEY, C. G. 1946. [Four notes on Northern Melanesian birds.] Condor, 48 : 92-93, 97, 182,

281-282.

1951. Notes on the birds of New Georgia, central Solomon Islands. Condor, 53 : 81-92.

SIMS, R. W. 1954. A new race f button-quail (Turnix maculosa) from New Guinea. Bull. Brit.

Orn. Cl. 74 : 37-40.
STICKNEY, E. H. 1943. Whitney, 53. Northern shore birds in the Pacific. Amer. Mus. Novit.

(1248) : 1-9.
STRESEMANN, E. 1912. Ornithologische Miszellen aus dem indo-australischen Gebiet. Novit.

Zool. 19 : 311-351.
1913. Ornithologische Miszellen aus dem indo-australischen Gebiet. 2. Novit. Zool. 20 :

289-324.

1914. Die Vogel von Seran (Ceram). (Aus den zoologischen Ergebnissen der II Freiburger

Molukken-Expedition). Novit. Zool. 21 : 25-153.

1921. [Neuen Formen aus dem Stromgebiet des Sepik (nordliches Neu guinea) und aus

Neupommern.] Am. orn. Ges. Bayern, 1 : 33-38.

1923. Dr. Burgers' Ornithologische Ausbeute im Stromgebeit des Sepik. Ein Beitrag zur

Kenntnis der Vogelwelt Neuguineas [concluded]. Arch. Naturgesch. 89A (8) : 1-92.

1933- Neue Vogelrassen von Lihir (Bismarck- Archipel) . Orn. Mber. 41 : 114-116.

!939- " Heterogynie " im Rassenkreis Edolisoma morio. Orn. Mber. 47 : 124-126.

SUTTER, E. 1955. tiber die Mauser einiger Laufhiihnchen und die Rassen von Turnix maculosa

und sylvatica im indo-australischen Gebiet. Verh. naturf. Ges. Basel, 66 : 85-139.
TIMES ATLAS, 1958. The Times atlas of the world. Mid-century edition. 1. London; Times

Publishing Co. : xxvii + 24 pp, 24 pis.
TRISTRAM, H. B. 1879. On a collection of birds fom the Solomon Islands and New Hebrides.

Ibis (4) 3 : 437-444-

1882. Notes on a collection of birds from the Solomon Islands, with descriptions of new

species. Ibis (4) 6 : 133-146.

1892. On two small collections of birds from Bugotu and Florida, two of the smaller

Solomon Islands. Ibis (6) 4 : 293-297.

1894. On some birds from Bugotu, Solomon Islands, and Santa Cruz. Ibis (6) 6 : 28-31.

1895. Further notes on birds from Bugotu, Solomon Islands, with description of a new

species. Ibis (7) 1 : 373-376.

TUBE, J. A. 1945. Field notes on some New Guinea birds. Emu, 44 : 249-273.
VANE, E. N. T. 1959. Some observations on the Cacatuinae the cockatoos. Avicult. Mag.
1959

86 IAN C. J. GALBRAITH AND EBBA H. GALBRAITH

VAURIE, C. 1949. A revision of the bird family Dicruridae. Bull. Amer. Mus. nat. Hist. 93 (4) :

I99-34 2 -

1958. Remarks on some Corvidae of Indo-Malaya and the Australian region. Amer. Mus.

Novit. (1915) : 1-13.
VERHEYEN, R. 1956. Analyse du potentiel morphologique et projet d'une nouvelle classification

des Psittaciformes. Bull. Inst. Sci. nat. Belg. 32 (55) : 1-54.

VIRTUE, R. M. 1947. Birds observed at Torokina, Bougainville Island. Emu, 46 : 324-331.
VOLKER, O. 1953. Das Farbkleid der Flaumfusstauben (Ptilinopodinae) . /. Orn. 94 : 263-273.
Voous, K. H. & MARLE, J. G. VAN. 1949. The distributional history of Coracina in the Indo-

Australian archipelago. Bijdr. Dierk. 28 : 513-529.
WALKER, F. S. 1948. The forests of the British Solomon Islands Protectorate. London ; Crown

Agents : 186 pp. (+corr. and add. 1950).
WHITE, C. N. M. 1936. [Note on the Australian Tree-Martin Petrochelidon nigricans.] Bull.

Brit. Orn. Cl. 56 : 90-92.
1938. [Notes on some Solomon Island birds.] Bull. Brit. Orn. Cl. 58 : 46-48.

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

0.

A REVIEW OF THE INDO-PACIFIC

GIZZARD SHAD GENERA
NEMATALOSA, CLUPANODON

AND KONOSIRUS
(PISCES : DOROSOMATIDAE)

P. J. P. WHITEHEAD

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vol. 9 No. 2

LONDON : 1962

A REVIEW OF THE INDO-PACIFIC GIZZARD

SHAD GENERA NEMATALOSA, CLUPANODON

AND KONOSIRUS (PISCES : DOROSOMATIDAE)

BY

P. J. P. WHITEHEAD

Department of Zoology, British Museum (Nat. Hist.

Pp. 87-102 ; 4 Text-figures

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 9 No. 2

LONDON: 1962

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts will appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 9, No. 2 of the Zoological series.
The abbreviated titles of periodicals cited follow those
of the World list of Scientific Periodicals.

Trustees of the British Museum, 1962

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued November, 1962 Price Six Shillings

A REVIEW OF THE INDO-PACIFIC GIZZARD
SHAD GENERA NEMATALOSA, CLUPANODON
AND KONOSIRUS (PISCES : DOROSOMATIDAE)

By P. J. P. WHITEHEAD

CONTENTS

Page
ABSTRACT ........... 89

1. INTRODUCTION. .......... 89

2. Nematalosa AND ITS ALLIES IN THE INDO-PACIFIC .... 90

3. DIAGNOSTIC FEATURES ......... 91

(i) Gill filaments ......... 92

(ii) Gillrakers . . . . . . . . .92

(iii) Parietal sculpture ........ 92

(iv) Scutes .......... 92

(v) Suboperculum ......... 93

(vi) Dorsal rays ......... 95

(vii) Upper and lower jaws. ....... 95

(viii) Second sub-orbital ........ 97

4. DISCUSSION ........... 97

5. A REDESCRIPTION OF Nematalosa arabica REGAN .... 98

6. GENERIC DIAGNOSES ......... 101

7. A KEY TO THE DOROSOMATIDAE OF THE INDO-PACIFIC . . . 101

ABSTRACT

The three genera of Indo-Pacific Dorosomatidae with filamentous last dorsal rays have been
examined. Although closely related, Clupanodon and Konosirus are here separated from Nema-
talosa and from each other on a number of characters, including the shape of the suboperculum
and jaw form. There seems to be little justification for placing the Australian species in the genus
Fluvialosa, and all are here referred to Nematalosa. N. arabica is redescribed from specimens now
in the Museum collection.

i. INTRODUCTION

WHILE identifying specimens of Nematalosa arabica Regan from the Somalia coast and
the Gulf of Aden, I noticed that no further description of this species has been made
since Regan's (1917) description of the holotype. But in redescribing N. arabica, it
was necessary to make comparisons not only with other species of Nematalosa, but
also with the monotypic genera Clupanodon and Konosirus, the last in some respects
an annectant form between the other two. The result of this comparison has been to
reinforce the view of Herre & Myers (1931) that K. punctatus (Schlegel) is generically
distinct from Clupanodon thrissa (Linn.), but that it should not be relegated to
Nematalosa. The latter genus shows some variation, especially in lower jaw shape,
but there are too many intergrading forms for any but specific divisions to be con-
sidered.

ZOOL. Q, 2 6

go P. J. P. WHITEHEAD

2. NEMATALOSA AND ITS ALLIES IN THE INDO-PACIFIC

Seven species are included under this heading (eight if Nematalosa elongata (Macleay)
is accepted). All have an elongated last dorsal ray and a single supplemental or supra-
maxilla. Although the importance of an elongated dorsal ray is perhaps over-rated,
I have here omitted consideration of the two Indo-Pacific dorosomatid genera
lacking a filamentous ray, Gonialosa and Anodontostoma. Regan (1917) proposed
Nematalosa for those species (until then included in Chatoessus Cuvier), which in
addition to possessing a dorsal filament and single supra-maxilla, also had a flared or
reflected edge to the dentary ; two related forms without flared dentaries he placed
in Clupanodon Lacepede (C. thrissa (Osbeck) and C. punctatus (Schlegel)), putting
Konosirus Jordan & Snyder in the synonymy of Clupanodon.

However, Herre & Myers (1931) stated that the dentary in C. punctatus is to some
extent reflected as in Nematalosa, but that C. punctatus differs sufficiently in other
respects from Nematalosa, as well as from Clupanodon, to justify a monotypic genus,
Nealosa (suppressed (Myers, 1932) since Konosirus had already been applied to
Punctatus). Konosirus has frequently been placed in the synonymy of Clupanodon
(e.g. Fowler, 1941, Roxas, 1934) or again considered of subgeneric standing only, but
some authors have recognized it (Iwai, 1956).

After examining material in this museum, I conclude that the criteria used by
Herre & Myers to separate Konosirus are valid and should be re-emphasized. To
them can be added also the rather higher vertebral count in K. punctatus, e.g. :

K. punctatus . . 48, 50, 51

C. thrissa . . 43

N. arabica . . 46

N. erebi . . 43*

(Also Anodontostoma 42 ;* Gonialosa 44,* 46.*

Clupanodon thrissa is unique in this group in having the outer demibranchs of the
first two arches very short, about half the length of the corresponding inner demi-
branchs, and this character, together with its normal [non-flared] dentary and its
subopercular shape (see below), clearly distinguish it from Nematalosa. But if such
differences merit generic distinction, then the high vertebral count in Konosirus
punctatus as well as its flared dentary and normal demibranchs must be sufficient
reason to separate it generically from Clupanodon, while its subopercular shape and
proportions of maxilla and premaxilla must distinguish it from Nematalosa. The
characters on which this argument is based are discussed separately below.

Of the species normally placed in Nematalosa, four require little comment. N.
arabica Regan and N. japonica Regan occupy the western and eastern limits of the
range of this genus (i.e. Red Sea to Japan), N. nasus (Bloch) occurs in the central
part of this region (India and Burma) and N. come (Richardson) is a Western Austra-
lian species. The remaining species, all Australian, have been placed in Fluvialosa,
a genus proposed by Whitley (1943) for Chatoessus elongatus Macleay, N. horni (Zeitz)
and N. richardsoni (Castelnau), and later (Whitley, 1948) for F. paracome Whitley

* Figures from Regan (1917).

REVIEW OF THE INDO-PACIFIC GIZZARD SHAD GENERA 91

and F. bulleri Whitley, and again (Whitley, 1956) for N. erebi (Giinther). These
species can all be adequately accommodated in Regan's Nematalosa, but they were
distinguished by Whitley as being " large fluviatile or estuarine Australian herrings "
as opposed to the genotype of Nematalosa " the marine Clupea nasus Bloch " (Whit-
ley, 1943).

As will become apparent in later discussion, N. horni and N. erebi are rather
distinctive in comparison with other species of Nematalosa, especially in their greater
body depth, few ventral scutes and jaw shape. On the other hand the species which
most nearly approaches these Australian forms is none other than the " marine "
N. nasus. Certainly the Australian forms cannot be separated from the rest with the
same confidence as in the case of Konosirus or Clupanodon. I consider it reasonable,
therefore, to place Flwuialosa in the synonymy of Nematalosa.

Whereas N. horni and N. erebi appear to be valid species, N. richardsoni is proble-
matical. It was originally described as a smaller fish than N. erebi, with an average
length of 7 inches (Castelnau, 1873), but Macleay (1880) refers to it as attaining
10-14 inches. Unfortunately the descriptions of both these authors are too vague to
distinguish this species from N. erebi, and its distinctive features (dorsal filament
short, body more slender) may be juvenile characters or characters whose growth
shows positive allometry with standard length. Fowler (1941) places N. richardsoni
in synonymy with N. come, together with N. erebi and N. horni, but this seems
unjustifiable " lumping " since the latter two at least, differ from N. come in numbers
of scutes and dorsal rays and mouth shape. I have not seen specimens of N. elongata,
but descriptions suggest a species very close to, if not identical with N. horni.

The two remaining species, N. paracome and N. bulleri are also unrepresented in
the museum collections. The former, described from a single specimen, is said to
differ from other species in having no humeral blotch (although a " faint duskiness "
is described) ; a shorter dorsal filament (although as a percentage of standard length
it differs by only 1-7 per cent, from that of N. bulleri according to actual measurements
given by Whitley, 1948) ; and fewer predorsal and lateral scales (respectively two
and four fewer than in N. bulleri) .

The Australian species have not been described sufficiently critically for compari-
sons to be made on the basis of descriptions alone. All however, have rather fewer
scale and scute counts than do the other Indo-Pacific species, and (except for
N. elongata) all are rather deep-bodied fishes. It seems probable that all have the flared
dentary and rather inferior mouth characteristic of N. erebi and N. horni. But
although these Australian fishes form a fairly distinctive group, they are none the
less linked by N. come and N. nasus to the remaining Indo-Pacific species ; there is
no definite point at which the genus Fluvialosa could be said to have begun, and until
more studies have been made of the Australian species, all should be placed in
Nematalosa.

3. DIAGNOSTIC FEATURES

Before assessing further the relationships of Nematalosa, Konosirus and Clupanodon,
eight principal characters must be discussed on which specific and generic divisions
have been made.

92 P. J. P. WHITEHEAD

(i) Gill Filaments

In C. thrissa alone the outer demibranchs of the first two arches are only half the
length of the inner demibranchs ; in the remaining species the outer demibranchs
are at least three-quarters (usually more) of the inner series. The difference is quite
striking and appears to be constant, and it seems to be a strong factor against
including K. punctatus in Clupanodon.

(ii) Gillrakers

In C. thrissa and K. punctatus the longest gillrakers on the lower part of the anterior
arch are at least three-quarters the length of the gill filaments opposite. In all other
species the gillrakers are approximately half the length of the gill filaments, often less.

(iii) Parietal Sculpture

In all species in this group there is a flat, longitudinally striated, wedge-shaped
area at the back of the skull formed partly of the parietals, but to which a postero-
lateral edge of the f rentals also contributes. This area is only lightly covered with
skin. The striations conform to two general types, referred to here as the sardinella
pattern and the harengula pattern (in which two genera they are best developed
amongst the clupeids). In the former there are 5-10 striae running almost the entire
length of the exposed bony area, and posteriorly there is usually a well-defined
transverse bony ridge. This type occurs in N. come (see Text-fig. la), N. japonica,
N. nasus, and usually but not invariably in N. horni. The harengula pattern normally
lacks the transverse ridge, or it is less prominent, and the longitudinal striae are much
fewer (about 3-5) and often discontinuous. This occurs in C, thrissa, K. punctatus
and N. arabica (see Text-fig, ib).

It is difficult to assess the importance of these patterns, and especially since even
in Harengula and Sardinella there are a few exceptions (thus H, nymphaea has 8-10
striae). Certainly N. arabica more closely approaches C. thrissa and K. punctatus in
jaw structure than do any of the species with a harengula pattern. N. japonica on
the other hand shows a striation pattern in some respects intermediate between the
harengula and the sardinella types. There is also some variation with age.

(iv) Scutes

The total number of ventral scutes does not vary much between the three genera
(range 28-36) and there is considerable overlap between species. Nonetheless certain
trends are apparent when pre- and post-pelvic scutes are considered separately.

In the pre-pelvic series the Australian species (Whitley's Fluvialosa) have a consist-
ently lower count. In the post-pelvic series, as well as in the total count, K. punctatus,
N. japonica and N. arabica all have a rather high count in comparison with the Austra-
lian species. Rather surprisingly C. thrissa falls within the latter group. Apart from
this exception, scute numbers tend to confirm the rather distinctive character of the
Australian species, although there is no clear division and C. thrissa appears to be
intermediate between the two groups.

REVIEW OF THE INDO-PACIFIC GIZZARD SHAD GENERA

93

FIG. i. Two types of parietal sculpture, right lateral view.

(a) Nematalosa come (the ' sardinella ' type).

(b) Nematalosa arabica (the ' harengula ' type).

For explanation see text.

TABLE I. Ventral Scutes in Certain Indo-Pacific Dorosomatidae

Number

of

Pre-pelvic

Post-pelvic

Total

Specimens

Clupanodon thrissa . (17) 18 (19)*

11-12

(28) 29-30 (31)

5

Konosirus punctatus . (17) 18-19(20) .

15-16

(32-33) 34-35 (36)

I3t

Nematalosa arabica . (18)19

(13) 14 (15)

(32) 33 (34)

9t

N. japonica . . 18

14-16

32-34

3t

N. nasus . . 18

11-13

29-31

5

N. come . . 18

12-13

30-31

5t

N. erebi . . 16-17

11-14

28-31

5

N. horni . . 16-17

12-13

28-30

5t

N. elongata J . 17

II

28

i

* Parentheses indicate single instance.

t Including the type(s).

J Based on Whitley (1943).

(v) Suboperculum

Herre & Myers (1931) first noticed that the shape of the exposed portion of the
suboperculum in C. thrissa and K. punctatus differed from that of related species.
In these two species it is far less rectangular (see Text-figs. 30 and&). In fact, this
difference in shape is not only a difference in posterior outline of the suboperculum,
but results also from the angle made by the lower edge of the operculum with the

94

P. J. P. WHITEHEAD

FIG. 2. A comparison between the short and rather wide jaws of (a) Nematalosa horni
(specimen 140 mm.), the longer and narrower jaws of (c) Clupanodon thrissa (172 mm.),
and the intermediate condition found in (b) Nematalosa arabica (150 mm.). Ventral and
lateral views.

REVIEW OF THE INDO-PACIFIC GIZZARD SHAD GENERA 95

posterior margin of the inter-operculum. In C. thrissa and K. punctatus this angle
is obtuse, but in the others it approaches a right-angle. The difference appears to
be consistent in fishes of all sizes.

(vi) Dorsal rays

In dorsal (and anal) fin ray counts there is some advantage in counting branched
and unbranched rays separately. Thus some species may be distinguished by a small
but constant difference of only one or two rays, but where the first and second
unbranched rays are small or even minute, this difference may be missed.

Regan (1917) separated N. horni and N. erebi from other species because of their
low dorsal count. In the following table this low count is shown to result mainly from
a low branched ray count.

TABLE II. Branched and Unbranched Dorsal Rays in Certain Species of Dorosomatidae

Number of

Unbranched Branched Total Specimens

N. arabica . . 4-5 . (12)* 13-14 . 17-18 . 8f

N. japonica . . 4 .13 (14) . 17 (18) . 3f

N. nasus . . 4-5 . 12-13 ( J 6) 17 (18) . 4

N. come . . 4 .13 (14) . 17 (18) . 8f

N. erebi . . (3) 4 . 10-11 . 14-15 izf

N. horni . . 3-4 . 9-11 (12) . 13-15 iof

* Single instances in parentheses.
f Including the type(s).

Apart from N. horni and N. erebi, the remaining species are remarkably alike in
this character.

(vii) Upper and Lower Jaws

Between Clupanodon thrissa and Nematalosa horni there exists a graduated series
of jaw forms in which the species can be arranged in the following order.

Clupanodon thrissa

Konosirus punctatus

Nematalosa arabica

N. come and N. japonica

N. nasus

N. horni, N. erebi (and if valid, N. elongata]

This series consists of a progressive shortening of the upper jaw elements, a deepen-
ing of the premaxilla, a more inferior mouth, a widening of the dentary and flaring of
its outer edge (especially at the corners of the mouth), and an increasingly wide angle
made by the dentaries at their symphysis. Three stages in this series are shown in
Text-fig. 2a, b and c.

As Herre & Myers (1931) noted, the expanded portion of the maxilla (i.e. the distal
end) is in Clupanodon and Konosirus rather elongated and extends as far beyond the
premaxilla tip as the length of the premaxilla itself ; with the inclusion of the supra-

9 6

P. J. P. WHITEHEAD

maxilla, the expanded portion of the maxilla becomes approximately rectangular,
the length of the expanded portion comprising a half to two thirds of the whole length
of the maxilla. From the specimens available it appears that the maxilla in K. puncta-
tus does not quite attain the length found in C. thrissa, but is almost exactly inter-
mediate in this respect between the latter and N. arabica (in which the expanded
portion is contained 2^-3^ times in the length of the maxilla) . N. come and N. japonica
represent a further reduction in the length of the expanded portion as well as a trend
towards a downwardly directed tip to the maxilla. In N. nasus, and to a greater
extent in the Australian species, this latter trend is further accentuated by a turning

FIG. 3. The shape of the exposed portion of the suboperculum (stippled), determined in
part by the angle formed by the lower edge of the operculum and the posterior edge of
the interoperculum.

(a) Nematalosa arabica (150 mm. specimen).

(b) Clupanodon thrissa (180 mm. specimen).

inwards of the maxilla tip round the flared corners of the lower jaw. The series shows
a reduction of the narrow stem of the supra-maxilla and the premaxilla becomes
shorter and slightly deeper.

The series also shows a shortening of the lower jaw (from dentary symphysis to
quadrate articulating facet), and at the same time the mouth, when viewed ventrally,
becomes wider and more nearly a straight transverse line (see Text-fig. 2). With the
flaring of the dentary, the lower jaw is no longer included in the upper when the mouth
is closed.

With these changes comes a slight shortening of the snout, the head in profile
becoming deeper and less acutely pointed anteriorly.

Although specimens of N. horni or N. erebi seem to differ strikingly in mouth shape
from say C. thrissa, the difference is one of degree only and the remaining species
provide almost perfect intermediate steps. Once again there is no basis for separating
the Australian species.

REVIEW OF THE INDO-PACIFIC GIZZARD SHAD GENERA 97

(viii) Second Sub-orbital

Normally the second sub-orbital extends to about the centre of the orbit and its
anterior border forms a diagonal (Text-fig. 40). In N. nasus, however, this bone
covers the entire cheek, terminating in front of the eye, and its anterior border is
vertical (Text-fig. 4b). This single exception is not related to any of the other trends
found.

FIG. 4. The shape and extent of the second suborbital (stippled).

(a) Nematalosa come (129 mm.).

(b) Nematalosa nasus (115 mm.).

4. DISCUSSION

Of the characters discussed, jaw form shows the most complete transition, with a
perfectly graded series between the "normal" jaw of Clupanodon thrissa and the pro-
nounced inferior jaws of Nematalosa erebi and N. horni. This series may reflect a
progressive trend towards bottom-feeding, for the mouth shape of the Australian
species strongly resembles that of the iliophagous grey mullets ; unfortunately there
are insufficient records of food and feeding habits to confirm this.

To what extent do other characters conform to this series ? Comparing the tables
given for dorsal rays, vertebrae and pre- and post-pelvic scutes, exactly the same
order of species can be maintained with one exception ; C. thrissa has unexpectedly
low vertebral and post-pelvic scute counts if it is to head the series. However, this
does not seem to be a strong objection to the order of the species and may well repre-
sent a variation similar to the unique sub-orbital shape in N. nasus.

As a whole, species in this group show a great number of similarities and common
features, and there is something to be said for placing all in a single genus, perhaps

98 P. J. P. WHITEHEAD

giving C. thrissa subgeneric status. But there are several rather distinctive features
found in one, two, or at the most three species only, which appear to be unrelated to
the general trend found in mouth shape or meristic characters. Thus the differences
in gillraker and gill filament length, and the shape of the suborbital and subopercular
bones serve to fragment the group.

Two main evolutionary trends are apparent. The first, which can be broadly
linked with geographical distribution, involves mouth shape and certain meristic
characters, and shows a gradual transition from one species to the next. The second,
apparently unrelated to distribution, shows no intermediates and appears to represent
merely variations in individual species or species pairs. Thus within the group, there
is a reasonably complete series between N. horni and N. arabica, but there is a distinct
break between these and C. thrissa and K. punctatm, which themselves can be clearly
separated.

In the first group of characters (those which intergrade) it is difficult to decide
whether the non-meristic series (i.e. jaw form) should be accorded the same status as
the meristic series. Thus, there is good reason to suspect that scute, finray and verte-
bral numbers can be correlated with distribution (i.e. with an extrinsic factor such
as temperature), whereas with jaw shape, these may be merely representative stages
in the evolution of the group. That the species can be arranged in the same order in
respect to both these kinds of character may be coincidental.

In the second, or non-intergrading, group of characters, the differences between the
species are often rather greater than one would expect, and it is perhaps surprising
to find the aberrant form so often represented by only one or two species. Taken in
conjunction with the intergrading series, the impression gained is of a group in which
a few well adapted forms have survived, showing little tendency to speciate except
amongst the Australian fishes, which are the most specialized and perhaps the most
recent. K. punctatus (with a slightly higher vertebral count) and C. thrissa are prob-
ably the primitive members of the group.

5. A REDESCRIPTION OF NEMATALOSA ARABICA REGAN 1917

This description is based on the holotype (131 mm. S.L., B.M. (N.H.) No. 1887 . n . n .
312) ; a specimen from Mukalla labelled N. nasus (150 mm. S.L., B.M. (N.H.) No.
1945.12.31.14) ; and six specimens from Jibuti, Somaliland (94-101 mm. S.L.,
B.M. (N.H.) No. 1962.3.13.1-6).

Description

In percentages of standard length (S.L.) : head length* 29-3-31-7, head depth
at occiput 23-4-26-0, body depth 36-0-40-5, snout length 6-1-7-6, eye diameter
7-3-7-8, post-orbitalf 13-9-15-3, premaxillary length 4-5-5-7, maxillary length
6-5-8-2, pectoral length 20-5-23-0, pelvic length 12-4-14-0, length of anal base

* Premaxillary symphysis to posterior edge of suboperculum, i.e. not a horizontal line. This obviates
measuring to inclined operculum edge.

f These last three measurements are taken in a straight line through the eye and therefore do not
in toto equal head length.

REVIEW OF THE INDO-PACIFIC GIZZARD SHAD GENERA 99

I 5'3~ I 7'9> pre-dorsal distance 49-0-51-0, pre-pelvic distance 52-0-54-5, length of
last (filamentous) dorsal ray 35-5-41-0, depth of caudal peduncle (n-o) 12-1-12-9.

Premaxilla 1-14-1-77 times in length of maxilla ; the length of the expanded
portion of the maxilla 2-40-3-40 times in the length of the whole bone, the depth of
the expanded portion 2-57-3-13 times in maxilla length. Maxilla tip reaching to below
anterior pupil border.

Gill rakers about half length of gill filaments on anterior arch ; outer demibranch
equal or subequal to inner demibranchs.

42-45 scales in lateral series, 19 in transverse series. Ventral scutes 18-19 + 13-15
(total 32-34). Dorsal iv-v, 12-14 (total 17-18) ; anal ii-iii, 15-17 (total 18-20).
Vertebrae 46 (i specimen).

Pelvic fin base lies below 2nd or 3rd branched dorsal ray. Pectoral tips reach or
almost reach pelvic fin base.

Angle between posterior margin of inter-operculum and ventral margin of oper-
culum forming a right-angle, the sub-operculum appearing roughly rectangular
(Text-fig. 3). Parietal striae 3-5 (Text-fig. 16). Mouth inferior, the dentary moder-
ately expanded and reflected outwards (but not as strongly as in N. come) as shown
in Text-fig. 26.

6. GENERIC DIAGNOSES

In view of the foregoing comments on the three Indo-Pacific dorosomatid genera
with filamentous last dorsal rays, more precise generic definitions can be given.

Nematalosa Regan

Nematalosa Regan, 1917, Ann. Mag. nat. Hist. (8) 19 : 313. (Type Clupea nasus Bloch, designated

by Jordan, 1920, Genera of Fishes, Pt. 4, p. 560).
Fluvialosa Whitley, 1943, Aust. Zool. 10 (2) : 170.

Indo-Pacific dorosomatid fishes with the last dorsal ray produced into a filament.
Mouth subterminal or inferior ; premaxilla short but deep ; maxilla slender, expanded
distally and curved both downwards and slightly inwards round the dentary ; a
single supramaxilla ; dentaries meeting at their symphysis at an obtuse angle,
forming sometimes an almost transverse cleft, the edge of each dentary flared or
reflected outwards. Second suborbital with diagonal front border leaving an exposed
area above the lower limb of the preoperculum (Text-fig. 40) except in N. nasus, in
which the front border is vertical (Text-fig. 46). Suboperculum rectangular, the
anterior and upper borders of the exposed portion forming an approximate right
angle and the posterior margin angular and not smoothly rounded (Text-fig. 30).
Gillrakers on first arch half or less than half the length of the corresponding gill
filaments ; outer demibranchs on all arches at least threequarters of the length of the
inner demibranchs.

Dorsal rays iii-iv, 9-14, anal rays 18-24 (f which the first two or three are un-
branched), pelvic rays 8. Scales, 42-50 in lateral series, 14-21 transversely. Ventral
scutes, 16-19 pre-pelvic, 11-16 post-pelvic, total 28-34.

Six species of Nematalosa are recognized here, N. arabica, N. japonica, N. come,
N. nasus, N. erebi and N. horni.

ioo P.J.P.WHITEHEAD

Clupanodon Lacepede

Clupanodon Lacepede, 1803, Hist. Nat. Poiss., 5 : 465. (Type Clupea thrissa Linnaeus, designated

by Bleeker, 1866-72, Atlas Ichth. Ind. Neerland, 6 : 112).
Thrissa Rafinesque, 1815, Analyse de la nature, p. 88.

Indo-Pacific dorosomatid fishes with the last dorsal ray prolonged into a filament.
Mouth subterminal ; premaxilla short but less deep than in Nematalosa, maxilla
slender, expanded terminally but not curved downwards or inwards ; a single supra-
maxilla ; edge of dentaries not reflected outwards, dentaries meeting at their sym-
physis at an acute angle (Text-fig. 20) . Second suborbital with diagonal front border,
leaving exposed area above lower limb of preoperculum. Suboperculum not rectangu-
lar, anterior and upper borders of the exposed portion of the bone forming an angle
greater than 90 (Text-fig. 36). Gillrakers on first arch at least threequarters of the
length of the corresponding gill filaments ; outer demibranchs on the first two arches
only half the length of the inner demibranchs.

Dorsal rays iii-iv, 13-14, anal rays ii-iii, 19-23, pelvic 8. Scales, 44-45 in lateral
series, 17 transversely. Ventral scutes, 17-19 (normally 18) pre-pelvic, 11-12 post-
pelvic, total 28-31 (normally 29-30).

A single species recognized, C. thrissa.

Konosirus Jordan & Snyder

Konosirus Jordan & Snyder, 1900, Proc. U. S. nat. Mus., 23 : 349. (Type Chatoessus punctatus

Schlegel.)
Nealosa Herre & Myers, 1931, Ling. Sci. ] ., No. 10 : 236. (Type Chatoessus punctatus Schlegel.)

See also Myers, 1931, Copeia, No. i, p. 30.

Indo-Pacific dorosomatid fishes with the last dorsal ray prolonged into a filament.
Mouth subterminal ; premaxilla short and in depth intermediate between Nematalosa
and Clupanodon ; maxilla slender, expanded distally and similar to that of Clupano-
don ; a single supramaxilla ; dentaries with slightly reflected outer edges and meeting
at their symphysis at a fairly acute angle (again intermediate between the condition
found in Clupanodon and Nematalosa}. Second suborbital with diagonal front border
leaving exposed portion above lower limb of preoperculum. Suboperculum not
rectangular, anterior and upper borders of the exposed portion of the bone forming
an angle greater than 90 (as in Clupanodon). Gillrakers on first arch at least three-
quarters of the length of the corresponding gill filaments ; outer demibranchs of the
first two arches at least three quarters of the length of the inner demibranchs (as in
Nematalosa}.

Dorsal rays iii-iv, 15, anal ii-iii, 17, pelvic 8. Ventral scutes 17-20 (normally 18-19)
pre-pelvic, 15-16 post-pelvic, total 32-36 (normally 34-35). Scales, 42-46 in lateral
series, 17 transversely.

A single species recognized, K. punctatus.

REVIEW OF THE INDO-PACIFIC GIZZARD SHAD GENERA 101

7. A KEY TO THE DOROSOMATIDAE OF THE INDO-PACIFIC*

* Excluding four doubtful Australian species (N. richardsoni, N. elongata, N. paracome and N. bulleri)
for which there is insufficient published data see p. 91.

I. Last dorsal ray not prolonged into a filament.

A. Maxilla slender, distal end slightly expanded and curved downwards . Gonialosa

i. Depth 2-0-2-5 m length ; 45-47 lateral scales .... G. tnodesta

ii. Depth 2-6-3-2 in length ; 55-65 lateral scales . . . . G. manminna

B. Maxilla straight, thin, tapering terminally .... Anodontostotna

i. Dorsal 15 ; 19 transverse scales ; snout little protruded . . A. chanpole

ii. Dorsal 17-18 ; 12-13 transverse scales ; snout well protruded . A. chacunda

II. Last dorsal ray prolonged into a filament.

A. Gillrakers of the first arch at least threequarters of length of corresponding gill
filaments ; suboperculum not rectangular, its anterior and upper margins forming
an obtuse angle, its posterior margin rounded ; outer edge of dentary not or but
very slightly flared.

i. Outer demibranchs on first two gill arches at least threequarters of length of
inner demibranchs ; vertebrae 48-51 ; post-pelvic scutes 15-16

Konosirus punctatus

ii. Outer demibranchs on first two gill arches only half length of inner demibranchs ;

vertebrae 43 ; post-pelvic scutes 11-12 .... Clupanodon thrissa

B. Gillrakers of first arch half or less than half length of corresponding gill filaments ;
suboperculum rectangular, its anterior and upper margins forming an approximate
right angle ; dentary more or less flared ..... Nematalosa

i. 2nd suborbital covering whole cheek, anterior edge vertical, lower edge hori-
zontal and in contact with preoperculum ...... N. nasus

ii. 2nd suborbital with oblique anterior edge, leaving exposed area above lower
limb of preoperculum

a. Post-pelvic scutes 13-16, total ventral scutes 32-34 ; dentary moderately
flared ; depth 2^-3 in length.

a Depth 2^-2f ; anal 18-20 . . . . . . . N. arabica

fi Depth 3 times in length ; anal 21-23 .... N.japonicus

b. Post-pelvic scutes 11-14, total scutes 28-31 ; dentary strongly flared at
edges, mouth becoming inferior ; depth 2-2^ times in length, Australian
species.

a 13-14 branched dorsal rays ....... N. come

ft 9-12 branched dorsal rays

f Depth 2-2 in length ; pelvics below or immediately in advance of

dorsal origin ......... N. erebi

I Depth 2^-2! in length ; pelvics below anterior half of dorsal . N. horni

REFERENCES

BLEGVAD, H. 1944. Fishes of the Iranian Gulf. Danish Scientific Investigations in Iran, Pt. 3,

pp. 1-247.
CASTELNAU, F. L. DE L. 1873. Contributions to the Ichthyology of Australia. Proc. zool. acclim.

Soc. Viet., 2 : 110-158.
FOWLER, H. W. 1941. Contributions to the biology of the Philippine Archipelago and adjacent

regions. Bull. U.S. nat. Mus. 100 (Vol. 13), 1-879.
HERRE, A. W. & MYERS, G. S. 1931. Fishes from South Eastern China and Hainan. Ling.Sci.J.,

10 (2 and 3) : 233-254.

102 P. J. P. WHITEHEAD

IWAI, T. 1956. The anatomy of the pharyngeal pockets of the Japanese gizzard shad Konosirus

punctatus (Tern, and Schl.). Bull. jap. Soc. sci. Fish., 22 : 911.
JORDAN, D. S. & SNYDER, J. O. 1900. A list of fishes collected in Japan by Keinosuke Otaki,

and by the United States Steamer Albatross, with descriptions of fourteen new species.

Proc. U. S. nat. Mus., 23 : 335-380.

MACLEAY, W. 1880. On the Clupeidae of Australia. Proc. Linn. Soc. N.S.W., 4 : 363-385.
MYERS, G. S. 1931. Nealosa Herre & Myers equals Konosirus Jordan and Snyder (Clupeidae).

Copeia, No. i, p. 30.
REGAN, C. T. 1917. A revision of the clupeoid fishes of the genera Pomolobus, Brevoortia and

Dorosoma and their allies. Ann. Mag. nat. Hist. (8) 19 : 297-316.
ROXAS, H. A. 1934. A review of the Philippine Isospondylous fishes. Philipp. J. Sci., 55 :

231-295.
WHITLEY, G. P. 1943. Ichthyological notes and illustrations (Part 2). Aust. Zool., 10 (2) :

168-187.

1948. New sharks and fishes from Western Australia Pt. 4. Aust. Zool., 11 : 259-276.

1956. Ichthyological illustrations. Proc. R. zool. Soc. N.S.W. 1955-56, 56-71.

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

THE PANTANODONTINAE,

EDENTULOUS TOOTHCARPS

FROM EAST AFRICA

P. J. P. WHITEHEAD

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)
ZOOLOGY Vo1 - 9 No. 3

LONDON: 1962

THE PANTANODONTINAE, EDENTULOUS
TOOTHCARPS FROM EAST AFRICA

BY

P. J. P. WHITEHEAD

Department of Zoology, British Museum (Natural History)

Pp. 103-137 ; 19 Text-figures

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 9 No. 3

LONDON : 1962

THE BULLETIN OF THE BRITISH MUSEUM
(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts will appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 9, No. 3 of the Zoological
series. The abbreviated titles of periodicals cited follow
those of the World List of Scientific Periodicals.

Trustees of the British Museum, 1962

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued November, 1962 Price Twelve Shillings and Sixpence

THE PANTANODONTINAE, EDENTULOUS
TOOTHCARPS FROM EAST AFRICA

By P. J. P. WHITEHEAD

CONTENTS

INTRODUCTION ........... 105

THE SUBFAMILY PANTANODONTINAE :

Relationships . . . . . . . . . .106

Diagnosis ........... 108

DESCRIPTION OF Pantanodon podoxys :

General body form . . . . . . . . .109

Teeth in

Upper jaw . . . . . . . . . . .in

Gillrakers . . . . . . . . . . .114

Pharyngeal dentition . . . . . . . . .116

Feeding mechanism . . . . . . . . .116

Head squamation ......... 123

Head canal system . . . . . . . . .126

Skull 126

Branchiostegal rays . . . . . . . . .129

Pectoral arch .......... 129

Pelvic girdle .......... 131

Axial skeleton .......... 131

Caudal skeleton . . . . . . . . . 133

Pseudobranchiae . . . . . . . . . .135

Habitat ........... 135

DISCUSSION . . . . . . . . . . -135

ACKNOWLEDGEMENTS . . . . . . . . . .136

BIBLIOGRAPHY ........... 136

INTRODUCTION

THE Pantanodontinae, a new subfamily within the oviparous Cyprinodontidae or
" killifishes ", was erected by Myers (1955) to accommodate a single genus and species,
Pantanodon podoxys Myers, based on two small specimens from Dar es Salaam in
Tanganyika which entirely lacked jaw teeth. They were briefly described as " Some-
what similar to Procatopus but with no teeth in the jaws and ventral rays spiny." No
fuller description has since been published.

Recently five more edentulous toothcarps were found in a salt evaporating pool
near Gongoni on the Kenya coast, some twenty miles north of Malindi. All are
apparently females, mature at less than 20 mm. standard length, and all have normal
pelvic fins. They have been compared with a paratype of P. podoxys. I must express
my gratitude to Prof. G. S. Myers, of Stanford University, California, for lending me
this specimen.

ZOOL. 9, 3 7

106 P.J.P.WHITEHEAD

The Tanganyika specimen is a male of 30-5 mm. standard length. I can find no
difference in meristic characters between this fish and the Kenya specimens, and such
proportional differences as exist can be accounted for by its larger size and by sexual
dimorphism. However, some male toothcarps are smaller, often considerably smaller,
than females of the same species, and although this dimorphism is more marked
amongst viviparous than oviparous forms, it is nonetheless surprising to find a male
of apparently the same species so much larger. But I do not think this justifies
placing the Kenya specimens in a separate species, at least until some Kenya males
can be found. The Kenya population was confined to a highly saline pool, whereas
the Tanganyika fishes are labelled as found in " swampy land a few miles inland."
Thus the Kenya fishes may belong to a runted population living under adverse condi-
tions. Runting is not rare amongst toothcarps.

The Sub-family Pantanodontinae : Relationships

The need for a full revision of the oviparous toothcarps, and especially the African
Cyprinodontidae, has become obvious in attempting to establish the relationship of
Pantanodon to other forms. Some characters used in the following comparisons cut
right across the present subfamilial (or tribal) boundaries and yet appear to be
equally as " stable " or " important " as characters now used to separate the sub-
families. The two or perhaps three distinct types of maxilla form, the presence of a
forked or simple post-temporal, and the fusion of the median hypurals are examples
which would seem to merit attention. Since a full revision has not been possible,
I have used an existing classification following Myers (1955), and based mainly on
earlier work by that author.

Family Cyprinodontidae oviparous toothcarps
Subfamilies i. Cyprinodontinae

2. Orestiatinae

3. Fundulinae

4. Rivulinae

5. Procatopodinae (Aplocheilichthyini of Myers, 1931)

6. Pantanodontinae

7. Oryziatinae (Aplocheilini of Myers, 1931)

8. Lamprichthyinae (omitted in Myers, 1955)

(Subfamilies 3-8 were tribes in the subfamily Fundulinae of Myers 1931).

In male P. podoxys the anal fin is not modified into an intromittent organ and the
genus must be presumed oviparous. This, and the closer affinity shown by Pantanodon
to African rather than American or Asian genera, seems to reduce the chances that
Pantanodon is an introduced species.

In attempting to establish the relationship of Pantanodon to other oviparous tooth-
carps, tooth form a useful key character in separating both genera and subfamilies
obviously cannot be used. Myers (1938) felt that less reliance should be placed on
teeth than had been given by Ahl (1924), but nonetheless he was obliged to follow Ahl
in separating the genera Platypanchax , Cynopanchax and Plataplochilus solely on the

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 107

basis of oral dentition. Again, Myers (1931) found little to distinguish his subfamilies
Cyprinodontinae and Fundulinae except tooth form (tricuspid teeth in the former).

The absence of jaw teeth in Pantanodon is of course remarkable ; indeed it is unique
in the whole order Microcyprini. But evidence is given below supporting the hypo-
thesis that degeneration of teeth is but one aspect of a series of modifications arising
from specialisation to a purely filter-feeding habit. Such specialisation is suggested
by the barely protractile premaxillaries, the modification of the gillrakers into fan
structures, the form of the pharyngeal toothpads, and the tightly coiled gut. In all
these characters Pantanodon departs, often markedly, from the normal cyprinodont
pattern. The fact that specialisation has apparently been great enough to eliminate
such a stable character as jaw teeth makes it necessary to seek phylogenetic relation-
ships on the basis of characters other than those involved in feeding. Thus the
absence of vomerine teeth in an edentulous fish such as Pantanodon may not neces-
sarily betoken greater affinities with the subfamilies Fundulinae or Procatopodinae
than with the Rivulinae, while a non-protractile premaxilla may well represent an
adaptation paralleled by that found in the Oryziatinae.

The two principal characters which can be used to place Pantanodon are the fairly
high-set pectoral fins and the moderately wide preorbital, and these suggest affinities
with the Procatopodinae. The pectorals are not so high as for example in Aplocheilich-
thyspelagicus, but can certainly be considered " high " in comparison with the rivulin
Pachypanchax. The width of the preorbital region presents a problem since the
lacrimal is in fact very much reduced and is narrower even than is found in many
Rivulinae. However, the preorbital region as a whole is a little wider than in the
Rivulinae. The other characters which would link Pantanodon with the Procato-
podinae are the absence of vomerine teeth (already commented on) and the absence
also of pseudobranchiae. I have been unable to find further characters sufficiently
consistent and trenchant to distinguish the subfamilies Procatopodinae and Rivulinae.
Thus the presence of a forked or simple post-temporal, and the number of branchio-
stegal rays, varies within these two groups, as does also the point at which the gill
membranes unite, the degree to which the maxillary tip is bound to the lower jaw
by skin, and the form of the hypural rays. In cephalic squamation, Pantanodon
differs from both the Rivulinae and the Procatopodinae. In addition, I know of no
other oviparous genus in which the male pelvic fins develop spiny hooks.

Of the procatopodin genera, Pantanodon most closely resembles Hypsopanchax in
the compressed and deep body form. It bears a great resemblance to A . stuhlmanni
Ahl (which should perhaps be referred to Hypsopanchax} . However in Hypsopanchax
the gill membranes are united much further forward. In addition, the predorsal
region in Pantanodon is shorter (much shorter in the male) than in any genus of the
Procatopodinae, the general appearance being much more like the cyprinid than the
cyprinodont habitus. The pelvic fins are united by a membrane in Pantanodon,
similar to the condition in Procatopus, but the fin bases are not so far advanced as in
the latter genus. There is, however, a strong resemblance between these two genera
in body form.

The occurrence of Pantanodon, at least in Kenya, near the coast in pools whose
salinity often exceeds that of the adjacent sea, makes a link with the Oryziatinae of

io8 P. J. P. WHITEHEAD

the Indo-Malayan region and Japan at least a possibility. Oryzias, the only genus,
has high-set pelvic fins, no vomerine teeth or pseudobranchiae, no parietals (see
Ramaswami, 1946), and the premaxillae are non-protractile. In addition, some species
are extremely small, females of 0. minutillus maturing at only 17 mm. (Smith, 1945).
But in Oryzias the abdomen is not trenchant and the predorsal region is relatively
much longer so that dorsal, pelvic and anal fins appear to be set further back.

The fishes of eastern Kenya are part of a general eastern river fauna which is found
to the east of the Rift Valleys from the Zambezi to Somaliland. While certain genera
(e.g. the cyprinids Barbus, Rasbora, Labeo and Garra) occur also in India, the region
as a whole does not show any closer affinities with the Indian freshwater fishes than
do other regions (e.g. to the west of the Rift Valleys). No genera of Cyprinodontidae
are shared with the Indian region, although Pachypanchax and Nothobranchius are
also found in the Seychelles. Pantanodon shows little affinity with either of these
two rivulins. The only procatopodins which may occur in this eastern river fauna are
Aplocheilichthys stuhlmanni, but Ahl (1924) does not mention the locality in Tangan-
yika from which his specimens came, and A. myoposae and A. carlislei of Natal.

A marine route of entry to the East African coast also suggests affinities with the
cyprinodont Aphanius dispar (Riippell) . This species occurs for example in reef pools
in the Red Sea (Marshall, 1952) and its range extends from N.E. Africa to the Arabian
sea coast (Fowler, 1956). But although the body profile is much closer to that of
Pantanodon than is that of Oryzias, the pectoral fins are set much lower, and the
preorbital is narrower, the lacrimal overlapping the maxilla.

Thus Pantanodon seems to be an isolated form which has become highly specialised
in many of those characters of most value in separating the subfamilies of the Cyprino-
dontidae. The erection of a monotypic subfamily on the basis of trophic adaptation
is here felt to be justified because in Pantanodon modification has far exceeded sub-
familial limits (i.e. supralimital specialisation in the sense of Myers, 1960). At the
same time Pantanodon shows other marked differences from the Procatopodinae
which are unconnected with feeding habits (head squamation, pelvic girdle and rays,
narrow lacrimal), but in no case do these serve to link it more closely with the Rivu-
linae or any other subfamily. Pantanodon therefore appears as a highly aberrant form
most closely allied to the Procatopodinae but whose inclusion in that subfamily
would either destroy the value of the present definition of the subfamily, or so burden
it with unique exceptions that as a working description it would become very clumsy.
Thus on practical as well as biological grounds it seems more reasonable to separate
Pantanodon from the Procatopodinae.

Sub-family Pantanodontinae : Diagnosis
Pantanodontinae Myers, 1955, Tropical Fish Mag., March : 7.

Small oviparous toothcarps of the family Cyprinodontidae characterized by the
complete absence of jaw teeth, apparently as a result of specialisation to a filter-
feeding habit. Similarly, premaxillaries barely if at all protractile, and gut tightly
coiled. Gillrakers slender, but able to expand into minute triangular fans.

No vomerine teeth. Pharyngeal bones bearing transverse bars on which rows of
fine teeth are set. Pectoral insertion moderately high. Dorsal short, set behind anal

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS

109

origin. Distal part of the maxilla closely bound by skin. Epiplurals simple. Pre-
orbital wide, about half eye diameter. Gill membranes united well behind posterior
border of eye. In males ; the tips of the second, third and fourth pelvic rays are
spiny and hooked.

Habitus. Body strongly compressed and abdominal region deep and trenchant,
especially in males.

Genus PANTANODON Myers, 1955
Diagnosis as for species (see below).

Pantanodon podoxys Myers, 1955
(Text-figs, i, 2 and 3)

Description based on five females from Kenya 1 (17-21 mm. standard length) and
one paratype, a male, 30-5 mm. standard length from near Dar es Salaam, Tangan-

FIG. i. Pantanodon podoxys, female, 21 mm. S.L.

yika. Skeletal elements are described from two of the Kenya specimens stained in
alizarin. In the proportional measurements (in per cent, of standard length) and
meristic counts, the first figure given refers to the male paratype, the figures in
parenthesis referring to the Kenya specimens.

Depth of body 36-5 (27-5-29-5), head length 29-5 (31-5-31-8) ; eye diameter
8-9 (9-3-9-4), snout length 8-2 (8-6-9-4) ; predorsal distance 64 (65-69), pre-anal
distance 57 (53-54) ; pectoral length, damaged (22-2 and damaged), pelvic length
I 3'5 (9-9-10-6) ; caudal peduncle, length 20-3 (22-5-23-4), depth 18-4 (17-3-17-5),
length/depth ratio i-n (1-29-1-35).

Body deep, especially in males, strongly compressed, the abdomen trenchant.
Snout, in profile acute, equal or just less than eye diameter. Upper lip broad and
probably non-protrusible. No jaw teeth. Gill membranes free from isthmus, united
behind posterior border of eye. Dorsal about equidistant from snout and tip of
caudal rays, its origin over anterior third of anal fin. Pectoral fins set moderately

1 B.M. (N.H.) No. 1962.4.4.1-2.

no P. J. P. WHITEHEAD

high, the first ray lying on the midpoint of body depth at that point, and just below
horizontal through centre of eye ; pectoral base inclined forwards at 30 to the
vertical. Pelvic fins short, normal in females but tips of 2nd~4th rays modified into

FIG. 2. Pantanodon podoxys (male paratype, 30-5 mm. S.L.) Caudal, anal and pectoral fins

damaged.

FIG. 3. Pantanodon podoxys. Dorsal view of head showing broad upper lip and arrangement

of scales.

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS in

spiny hooks in males ; in both sexes just reaching anal origin. Caudal peduncle
deep.

Dorsal i 6-7 ; anal ii-iii 17-18 (total 20-21), probably not produced posteriorly
in either sex ; pectoral in, rounded in one female but damaged in others ; pelvic
i 5, the inner edges membranously joined to body ; caudal 16-18 branched rays,
10-12 unbranched raylets, the posterior margin of the caudal bluntly round in females
(damaged in paratype). About 60 gillrakers on lower part of first arch.

Scales in midlateral series 26-27 ; 9 scales between pelvic and dorsal origins ; 12
round caudal peduncle. Scales sharply bent in ventral midline.

Colour. In life, not recorded ; in preserved specimens, a uniform light brown
with no distinctive colour markings, but scattered melanophores on head and lip.

Fins hyaline.

Jaw Teeth

These are entirely absent in Pantanodon. In sections of the upper and lower jaws
of a specimen of 20 mm. S.L. there is no indication even of incipient tooth buds.
The very small adult size attained by Pantanodon, the rather poor ossification of some
cranial elements, the poor development of the head canals of the lateral line system,
and the absence of jaw teeth, might all be ascribed to a neotenous condition. But
jaw teeth in some toothcarps at least, appear very early indeed in ontogeny. They
are present at birth in Gambusia affinis (Berner, 1947), are apparently fully formed
in Oryzias curvinotus of 23 mm. (Nichols & Pope, 1927) and can be seen quite
clearly in Lebistes reticulatus of only 17 mm.

It is also arguable whether the jaw teeth of Pantanodon degenerated under some
kind of selection pressure, or whether their incidental loss (through mutation) then
stimulated further adaptation to filter-feeding. It is, however, difficult to imagine
disadvantages so great that teeth would become a hindrance to be actively selected
against. On the other hand, the degeneration of teeth in fishes primarily adapted to
seizing and grasping prey would place at great selectional advantage any subsequent
adaptations which would assist in filter-feeding.

Pantanodon could perhaps represent an end-point in the reduction of jaw teeth
shown by Hypsopanchax from the Aplocheilichthys condition.

Upper Jaw (Text-figs. 4 and 5)

In Pantanodon the upper jaw is barely, if at all, protractile, thus resembling
Oryzias alone among the Cyprinodontidae. It is, however, difficult to be certain of the
degree of protractability of the jaw in such small specimens, especially after preserva-
tion in formalin.

Details of the upper jaw mechanism are shown in Text-figs. 4 and 5. The premaxil-
lae are expanded in the midline but do not meet. The descending ramus of the
premaxilla is moderately broad and loosely attached distally to the lower jaw so
that depression of the latter tends to draw not only the lower ramus but the entire
premaxilla forwards very slightly. As a result of this, the upper portions of the pre-
maxilla are raised slightly so that the premaxillae appear to open rather as a flap

112 P. J. P. WHITEHEAD

upwards rather than to be protruded outwards. The maxillae are very slender. The
upper ramus lies beneath the expanded portion of the premaxilla and is curved
inwards. The descending rami overlap the distal tip of the premaxillae but are
curved forwards. Upper and lower rami of the maxilla meet at a weak joint whose
shape suggests that it might accommodate the posterior edge of the premaxilla but,

max.(u) max.(sup)

nas.

p.max

lac.

max.

(i)

FIG. 4. Pantanodon podoxys. Left lateral view of jaw mechanism.

p. max. Premaxilla (upper ramus).

max (u). Upper ramus of maxilla (= inferior lobe see text).

max (sup). Superior lobe of upper maxillary ramus.

nas. Nasal.

lac. Lacrimal or preorbital.

max (I). Lower ramus of maxilla.

owing to the distance between the two, in fact it does not. The lower ramus of the
maxilla is produced slightly beyond its junction with the upper ramus and similarly
the upper ramus projects beyond this junction also. The rami thus meet edgewise
and so a forward movement of the distal tip of the lower ramus can through a
twisting action, impart a forward movement also to the upper ramus. In the Micro-
cyprini the mechanical principle involved in the protrusion of the premaxillae differs
from that found in other groups (except perhaps the Mugiloidea). Eaton (1935)
describes it as the Cyprinodont or twisting type, in which twisting of the lower end of
a bow-shaped maxilla imparts a forward thrust to the upper maxillary ramus and

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 113

this in turn pushes out the premaxilla. This appears to be mechanically possible in
Pantanodon but it is hard to say if it is in fact used.

Gosline (1961) has, however, challenged Eaton's view, believing the resemblance
between say Fundulus and Mugil as far as jaw mechanism is concerned to be super-
ficial and secondary. " The maxilla (i.e. in the toothcarps) apparently has very little
to do with premaxillary protrusion " (Gosline loc. cit.), the distal ends of the pre-
maxilla being membranously attached to the coronoid portion of the lower jaw. He
concludes that the Microcyprini have a basically fixed premaxilla (i.e. no descending
ramus in Chologaster and Typhlichthys) from which has later developed the protru-
sible upper jaw. Loss of protrusibility in Oryzias and presumably in Pantanodon also,
both of which have descending rami, would thus be a reversion to the primitive

dentary

p. max

max.(u)

max. (sup)

FIG. 5. Pantanodon podoxys. Detail of upper jaw structure, dorsal view, from an alizarin stained
specimen of 17 mm. S.L. Symbols as in Text-fig. 4.

condition. But in Pantanodon, which shows so many structural changes apparently
connected with feeding habits, one might expect a more slender lower premaxillary
ramus than in fact occurs, although the maxilla in Pantanodon is by no means
robust.

In Pantanodon the upper ramus of the premaxilla is fairly broad. A broad ramus
is found in many other species, being perhaps best developed in Aplocheilus lineatus
(see Ramaswami, 1946). Some species have a rather narrow ramus, e.g. Micropanchax
loati and Aplocheilichthys pumilus.*

As already noted, at the point at which the upper and lower rami of the maxilla
meet, there is a tendency for each ramus to be produced beyond the junction, especially
in the case of the lower ramus. This latter projection, which may be more or less
developed, is here termed the superior lobe, in contrast to the main upper maxillary

* The precise limits of Aplocheilichthys and Micropanchax still await definition. Here as elsewhere in
the text, comparisons are essentially between species and not genera.

ZOOL. 9, 3 7

u 4

P. J. P. WHITEHEAD

ramus, which is referred to as the inferior lobe (see Text-fig. 5). Of all the specimens
examined, the inferior lobe lies underneath the expanded portion of the premaxilla,
except where the premaxilla is too narrow for the two bones to meet. In all but two
of the species examined, the superior lobe is no more than a small projection. The
exceptions are A . spilauchena and Procatopus nototaenia, and in these two the superior
lobe is greatly developed, reaching and overlying the expanded portion of the pre-
maxilla. This seems to be an intermediate condition, for in those species with narrow
upper premaxillary rami, there is only a single maxillary lobe, and from its position
it appears to be the superior one ; however, in M. loati the superior lobe bears a
small, pointed hook on its ventral surface and this may be a poorly developed inferior
lobe. Thus there appears to be a gradual transition between forms with only a

abed

FIG. 6. Pantanodon podoxys. Gillrakers : probable stages in the expansion of the external
series of gillraker fans. For explanation, see text.

superior lobe, through those with both lobes well developed, to those in which the
superior lobe is small and insignificant. The number of species is, however, too
few to show whether this is a line of development which can be related to the evolution
of the major groups of oviparous toothcarps.

The Gillraker System (Text-figs. 6 and 7)

There are about 50-60 slender gillrakers on the anterior arch. In the specimens
measured (15-20 mm. S.L.) the rakers are up to 0-8 mm. in length in the external
series and about half this length in the inner series. In addition there is a single
series of rakers on the antero-lateral edges of the lower pharyngeal bones ; these are
essentially similar in shape to those of the external series on the other arches. Pharyn-
geal rakers have not been found in other Microcyprini examined (see below), but
occur in some cichlids (e.g. Tilapia), in the cyprinid Labeo forskalii, and also in the
Grey Mullet Mugil cephalus.

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 115

In situ the rakers appear as thin rods, about 100 fi thick and approximately this
distance from each other. Each rod has a distinct globular cap and within the rod
there is an inner core of bony tissue running the length of the raker and projecting
slightly into the globular cap (Text-fig. 6a).

When, however, the gillrakers are removed from the arch, mounted on a slide in
glycerine and subjected to slight pressure (on the coverslip), they show a more com-
plex structure. The rods then expand into minute triangular fans, their apices being
at the point of attachment to the gill arch. The internal and the external fans open
in towards each other (i.e. towards the midline of the gill arch (see Text-fig. 7). Each

Ext.

Int.

FIG. 7. Pantanodon podoxys. Gillrakers : cross section through gill arch showing
reconstruction of probable shape and position of inner and outer gillraker fan series
when fans are fully expanded, ext. external gillrakers. int. internal gillrakers.

fan has two thickened arms radiating from the apex. In the larger, external fans the
subsidiary or smaller arm is distinctly jointed at its junction with the main arm, but
this was not seen in the internal raker fans. At the apex of the fan is a small root. The
distal edge of the fan is membranous and appears to be derived from the globular cap
since the latter disappears in fully expanded fans. Text-fig. 6 shows a series of external
raker fans in the process of expanding. In fully expanded fans there are about twenty
fine parallel lines running at right angles to the membranous edge but not actually
meeting it. Near the apex of the fan these lines tend to converge and form a reticu-
lated pattern. Under very high magnification the lines resemble fine capillaries. If

n6 P. J. P. WHITEHEAD

this is correct, then their distension might well serve to erect the fan. At the same
time their alignment also suggests that they offer support for the membrane of the
fan. Between these " capillaries " the membrane is marked by a series of transverse
lines, extremely fine and close together, which may be lamellae or ridges or even the
structural supports of a very fine sieve. Such a sieve would be capable of retaining
particles of about 2-3 /* in diameter.

The gillrakers and also the fans when erected, are probably capable of close inter-
digitation. A discussion of the probable feeding mechanism in Pantanodon is given
later.

The Pharyngeal bones (Text-fig. 8)

The pharyngeal toothpads also exhibit specialisation towards a microphagous diet.

The lower bones are roughly triangular, pointed anteriorly, rounded posteriorly
(see Text-fig. Sa). Laterally there are three spines for muscle attachment, the anterior
two pointing at right angles to the bone and the posterior spine directed rather
postero-laterally. Along the antero-lateral edges of the bones lies a series of gillrakers.
The two bones are neither suturally united nor even very closely apposed, resembling
the condition found in some cyprinids. The ventral surface of the bones is composed
of anterior radiating and posterior reticulated bony ridges. The dorsal surface is
made up of a series of ten to thirteen transverse bars each bearing a single row of teeth.
Anteriorly the bars are rather weakly joined in the midline, but the posterior two
or three bars are free. The lateral edges of the bones are, however, fairly strongly
ossified and the lateral tips of the bars are firmly attached to this edge.

The anterior teeth of the lower pharyngeal bones are slender, about 400 /* long,
conical and pointed, quite widely spaced (about 300 /* apart) and curved slightly
backwards. In the posterior rows, however, the teeth are a little shorter and become
so densely packed that there is little or no space between their bases (Text-fig. 8&).

The upper pharyngeal bones are essentially similar in structure to the lower bones
but are oval in shape (Text-fig. 8c). The posterior teeth are about 300 /* in length and
about 100 fi apart, tricuspid with expanded crowns, as shown in Text-fig. 8d. The
cusps show a distinct suture line at their base. In the anterior rows the lateral cusps
are progressively lost until the teeth are conical in the anterior two or three rows.
Generally, the teeth of the upper pharyngeals are rather more widely spaced than
in the lower pads.

The Feeding Mechanism in Pantanodon (Text-figs. 9, 10 and n)

Divergence of the Pantanodontinae has been almost entirely concerned with trophic
adaptation to a filter-feeding habit. Specialisation towards a microphagous diet has
apparently involved both the loss and modification of certain structures. Thus jaw
teeth have been entirely lost which may preclude the seizure of prey in the normal
cyprinodont manner. Secondly, the mouth gape is directed upwards while the pre-
maxillae are more or less fixed. This precludes bottom-feeding and implies that
Pantanodon must obtain its food from the surface layers. The third specialisation is
in the form of the gillrakers. In order to understand their mechanism and possible

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS

117

evolution from the more normal cyprinodont type, comparison was made, first with
other Microcyprini, and then with other fishes which use the gillrakers to retain fine
particles of food.

Aplocheilichthys pelagicus Worth., the small pelagic cyprinodont from Lake
Edward, feeds mainly on small planktonic crustaceans and some insect larvae

a

FIG. 8. Pantanodon podoxys. Pharyngeal bones and teeth from a specimen 17 mm. S.L.

a. Lower left pharyngeal, dorsal view.

b. Teeth from the posterior rows of the lower pharyngeal.

c. Upper right pharyngeal, ventral view.

d. Tricuspid teeth from posterior rows on upper pharyngeals.

(Worthington, 1932). In this species the internal gillrakers on the first arch are rod-
like and do not expand, but the external rakers of the first arch are long, narrow
plates set transversely to the line of the gill arch (Text-fig. 9) . Alizarin staining suggests
that, like Pantanodon, there is an inner and outer ossified support to the raker, but

u8

P. J. P. WHITEHEAD

that the lower and middle portions of the plates are also osseous. In the first four or
five fans there is a curious tubercle projecting from the inner edge of the fan. The
tubercle showed no apparent internal structure and its function is obscure. A narrow
membrane runs the length of the inner edge of the fan, and the fans also have one or
two fine lines running across their surface, encircling both the tip of the fan and the

int.

ext.

tub.

FIG. g. Aplocheilichthys pelagicus. First four gillrakers on ist gill arch, showing rod-like
inner rakers (int) and plate-like external rakers (ext) which bear a small tubercle (tub).

base of the tubercle, joining in places and leading to the base of the fan. The structure
of these lines is not clear ; in places they resemble spirally strengthened capillaries,
while elsewhere they could be interpreted as mucus directing or deflecting ridges.
In the succeeding arches the external rakers become pointed, and on the fourth arch
they are toothed. The large and fairly widely-spaced pharyngeal teeth show that the

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 119

diet is composed of much larger particles than is that of Pantanodon, but the plate-
like gillrakers on the first arch suggest that smaller particles may also be collected.

Plate-like gillrakers occur in a number of toothcarps, usually only on the first arch,
occasionally on the second or on the third, but never in the specimens examined did
they extend on to the fourth arch as in Pantanodon. Aplocheilichthys spilauchena,
A. pumilus, Micropanchax loati, Pachypanchax playfairii and an undetermined
species of Mollienesia all have some development of plate-like gillrakers. In
A . spilauchena the gillraker plates are not set at right angles to the gill arch but are
inclined slightly inwards and backwards, and at the same time the tips of the rakers
lean forwards. Such an arrangement may be the case in Pantanodon but this could
not be seen with the rakers in situ.

In many Cyprinodontidae the gillrakers are represented by teeth or tooth-like
rakers on some or all the gill arches. The exact homology of these tooth-like rakers is
not always clear ; some may merely be modified gillrakers, whereas others are
probably teeth which have replaced the rakers. Here, all are referred to as " toothed
rakers ". Toothed rakers are especially common on the fourth arch and parts of the
third, but in some fishes are found also on the second and even the first arch. In
most cases the toothed rakers are shield-shaped and lie flat against the surface of the
gill arch and their distal edges bear numbers of fine teeth. Those of Aplocheilus
panchax are shown in Text-fig. 10. Such toothed rakers occur in Pachypanchax
playfairii, A. spilauchena and Epiplatys sexfasciatus. In Aphanius dispar the rakers
of the external series on the first arch are tree-like and branched, while on the fourth
arch they are identical in shape to the lower pharyngeal teeth.

In all the specimens examined, those with toothed gillrakers also had a series of
minute microgillrakers on the external faces of all but the first arch. A single excep-
tion to this was P. playfairii. Microgillrakers were not found in any of the species
with plate-like gillrakers. The microgillrakers are in many ways homologous with
those of certain cichlid fishes (see Gosse, 1955 and Whitehead, 1959), but they are
larger and are paired structures (Text-fig. 10). In E. sexfasciatus there are no micro-
gillrakers as separate plates, but the base of each raker toothplate has at each corner
a series of four or five fine-pointed teeth which evidently serve the same purpose.
The disposition of the microgillrakers suggests that they are used to actively clean
the raker teeth on the arch immediately in front.

Filter-feeding, or at least the collection and passing backwards of fine particles of
food, presents a number of problems. Various methods have been evolved for solving
these problems, one of which is the plate system of Pantanodon. Perhaps the greatest
problem is the need to retain very fine particles by a device which will not clog.
In the echiuroid worm Urechis this is solved by periodic swallowing of a very fine
mucus sieve. In some fishes the cleaning of the sieve system appears to be achieved
by the passage of a mucus film to which the particles adhere. The mucus stream
must, however, be forced to take a direction different from that of the exhalent
stream of water and it seems likely that plate-like gillrakers are at least partly
responsible for creating the necessary turbulence for this to happen. In part, also,
the shape of the branchial cavity will assist in forcing the mucus back towards the
oesophagous, the principle being similar to that of a plankton net. In such a system

120 P. J. P. WHITEHEAD

it would seem that mucus-coated plates would provide a greater collecting surface
than would a plain sieve.

A plate system is used by species of Labeo (see Text-fig, n), while the grey mullet
Mugil cephalus employs a sieve. Comparing these two methods it would seem that
the plate system can retain fine particles only in fishes as small as Pantanodon (i.e.
about 20-30 mm. in length), whereas the sieve system is the more efficient in fishes of
100 mm. and more. Labeo is a highly specialized algal grazer, but in spite of inter-
digitation of the gill plates, there is probably little need for the rakers to retain more

FIG. 10. Aplocheilus panchax. External face of second gill arch showing tooth-like gill-
rakers (a) of external series, and position of micro-gillrakers (b).

than just the ingested algal fragments. In Mugil cephalus on the other hand the gill-
rakers, which are long and very closely set, are joined at their tips and each gillraker
bears a double series of minute projections along its length. The projections inter-
digitate. In a 76 mm. fish, with gillrakers 2-0-2-5 mm - m length, the small projections
were 50 fi apart. Each projection, however, bore a series of five to six minute spines
along its upper edge, the tips of which were about 5-10 fi apart, and this presumably
represents the size of the particles which can be retained. Like Pantanodon, Mugil
cephalus has a series of pharyngeal gillrakers.

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 121

The form of the pharyngeal toothpads in Pantanocton differs so greatly from the
normal pattern in other cyprinodonts (large, irregularly scattered teeth) that one
must regard it as another manifestation of high specialisation to a microphagous diet.
It is, however, difficult to imagine why a series of transverse rows of single teeth

sept.

int.

ext.

FIG. ii. Labeo forskalii. Plate-like gillrakers on the lower part of the ist gill arch
(semi-diagrammatic). Dorso-lateral view looking along gill arch from anterior end.

int. Internal gillraker series.
ext. External gillraker series.
sept. Median septum between gillraker series.

P. J. P. WHITEHEAD

FIG. 12.

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 123

should have arisen rather than a fine but dense mat of irregularly placed teeth, such
as is found for example in the microphagous species of Tilapia.

In Tilapia nigra, which feeds principally on diatoms and is to some extent a filter-
feeder, the lower pharyngeal bones are, as in other species, firmly united and cannot
move laterally. In Pantanodon, however, the lower pharyngeals are only loosely
connected, recalling the condition in Labeo, in which the bones are ventrally united
by transverse muscles. Thus in Labeo and Pantanodon the lower bones may be able to
move to some extent laterally, although in Labeo the bones are essentially used for
grinding against the horny pad attached to the basi-occipital facet, whereas in
Pantanodon they appear to act rather as combs.

The final adaptation in Pantanodon to a microphagous diet is seen in the very long
and coiled gut. The form of winding of the intestine is highly characteristic. It first
spirals inwards in a vertical plane, then reverses direction and unwinds in the same
plane but on the other side of the peritoneal cavity. Suyehiro (1942) shows an
essentially similar kind of winding in the case of Chaetodon modestits and Siganus
fuscescens, but he notes that this form of intestinal winding is unusual. This type of
coiling probably allows the maximum length of gut for a given volume of body cavity.
The mucosa of the gut is formed into longitudinal zig-zag lines, but it is not known
whether such a mucosal pattern can be related to a particular diet or if it holds any
taxonomic value.

The Pattern of Cephalic Squamation (Text-figs. 12 and 13)

In a short but useful description of the patterns of scales found on the dorsal surface
of the head in many Cyprinodontidae, Hoedeman (1958) showed that the African
Rivulini bore less resemblance to the South American forms than they did to the
Procatopodinae. Hoedeman considered only a single species of procatopodin,
Aplocheilichthys loati (= Micropanchax schoelleri). I have examined also Procatopus
nototaenia, Aplocheilichthys spilauchena, A . pumilus, A . antinorii and A . pelagicus and
confirm Hoedeman's conclusion.

FIG. 12. Cephalic squamation in some Procatopodinae and Rivulinae (semi-diagrammatic).
Topmost scale stippled. Scales lettered as explained in text.

a. Aplocheilichthys spilauchena

b. Aplocheilichthys pelagicus

c. Aplocheilichthys pumilus

d. Micropanchax loati

e. Aplocheilichthys antinorii
/. Procatopus nototaenia

g. Aplocheilichthys cameronensis

h. Epiplatys sexfasciatus

i. Pachypanchax playfairii

j. Nothobranchius guentheri (juvenile)

k. Nothobranchius sjoestedti

I, Nothobranchius taeniopygus (large adult).

124 p - J- p - WHITEHEAD

In order to compare the rather characteristic pattern found in Pantanodon with those
of other toothcarps, reference must first be made to the system of lettering individual
scales employed by Hoedeman (loc. cit.}. Although this system may occasionally
imply a homology which in fact does not exist, nonetheless it is a useful method and
one that is used here. The disposition of the head scales is to some extent governed
by the position of the head canals. In most Rivulinae and Procatopodinae, there is a
prominent pair of canals between the eyes, from the anterior rim of the orbit to about
the middle of the orbit (absent, however, in Pantanodon) ; these represent the posterior
portion of the supraorbital canal, or pores 2b, 3 and 4^ as designated by Gosline
(1949). Immediately posterior to the eye is another part of this canal (pores 46, 5
and 6). From a comparison of my own material with the figures given by Hoedeman
and also by Rosen & Mendelson (1960), the basic dorsal head scale pattern in the
Cyprinodontidae appears to be reducible to three main elements.

A. Two lateral rows of scales between the anterior part of the supraorbital canal
and the orbits themselves, but continuing forwards sometimes to cover the nasals.
This lateral series is interrupted by the posterior part of the supraorbital canal. The
most posterior scale of this series has been lettered/.

B. A median series of scales (single or paired) lying between the anterior parts of
the supraorbital canals, reaching forwards to the lip in some cases, and backwards to
a line joining the posterior orbital border. This series includes scales e, g, and h.

C. A circle of scales (b, c, and d] immediately posterior to the posterior part of the
supraorbital canal. This circle of scales surrounds a single, median scale, a.

The pattern of squamation can also be characterized by the particular scale which
overlies all the others ; in Text-fig. 12, this scale has been stippled, and the fishes
said to have a g, d or e-type pattern.

The Procatopodinae resemble the African Rivulinae in having a basic g-type
pattern, with the e scale typically paired. However, in most Procatopodinae
examined the a scale overlies the two d scales ; occasionally one d may overlap the
a, but rarely both. With one exception (Nothobranchius guentheri] the reverse is true
in the African Rivulinae. Small differences such as this are remarkably stable within
species and probably do hold a taxonomic weight. Modifications of this basic pattern
occur in both subfamilies. In Micropanchax loati (Text-fig, izd) the e scale is single
and not paired as in all other African species examined. Hoedeman notes that
Pachypanchax has an e-type pattern and I have confirmed that this is not merely
a single, individual variant (Text-fig, izi). An e-type pattern is also found in Aplo-
cheilichthys spilauchena, but this appears to be the only other exception to the rule
(Text-fig. 120). In A. antinorii, possibly because of the small size of the specimens,
the normal g-type pattern is upset (Text-fig. 120), in the first place by an h scale
overlying the g scale (this may be fortuitous, and does not seem to justify designa-
tion of an h-type pattern) ; and in the second place by two pairs of small scales
immediately before and after the e scales. In Text-fig. 12 I have not labelled the
anterior of these pairs since they appear to be too small to qualify as e scales. The
posterior of these pairs I have considered to represent a paired a scale ; this inter-
pretation is not entirely satisfactory, but any other would involve an unlikely
grouping of the d, c and b scales.

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 125

The most striking departure from the normal pattern is found in Nothobranchius
taeniopygus (Text-fig. i2/). Here the scales are arranged in rows as on the body and
do not conform to the circular pattern in other genera. A similar situation occurs in
the larger specimens of the aberrant Orestias, the scales of the head being small and
forming a mosaic pattern on the head. Hoedeman's figure for 0. agassiz is probably
derived from a small specimen ; thus small specimens of N. taeniopygus have the

FIG. 13. Pantanodon podoxys. Cephalic squamation : topmost scale stippled. Lettering

of scales as described in text.

normal circular pattern of scales but this is lost in the larger adults. The figure shown
here for N. sjoestedti is based on a juvenile ; adults may well show the linear arrange-
ment of scales found in N. taeniopygus.*

The scale pattern in Pantanodon is shown in Text-fig. 13. This differs from the usual

* However, Myers (1933) places sjoestedti in Aphyosemion, and the rather regular pattern of head
scales would seem to confirm this.

126 P.J.P.WHITEHEAD

pattern in that there are two rows of paired scales anterior to the a scale. Although
this shows some similarity to the A. antinorii pattern, Pantanodon cannot be desig-
nated a g-type genus, and thus it differs from both the Rivulinae and the Procato-
podinae. There are two or three possible interpretations ; the one shown in Text-fig.
13, which gives Pantanodon an -type pattern, places the e scales after and not before
the posterior margins of the orbits. To have placed them before would have required
a very prominent position indeed for the d scales. The text-figure is based on two
small specimens from Kenya ; unfortunately the somewhat larger paratype has few
head scales.

Whatever interpretation is arrived at, Pantanodon evidently differs from both the
Rivulinae and the Procatopodinae, although it seems closer to the latter, having the
two d scales overlain by the a scale, and not the reverse (a further reason for con-
sidering the d scales to be correctly placed in the text-figure).

Clearly, the pattern of cephalic squamation could well prove a useful systematic
tool for the Cyprinodontidae provided due regard is paid to the size of the individuals
compared. At the same time, the true homology of the scales must be derived from
ontogenetic studies.

Head Canal System (Text-fig. 14)

The degree of development of the sensory cephalic canal system has been shown to
have systematic value in some toothcarps, at least where closed canals occur (Gosline,
1949). In Pantanodon the canals are much reduced and are open (see Text-fig. 14).
The infraorbital canal (C) is represented by a short preorbital groove with three pores.
The supra-orbital canal (A) is also much reduced, and is represented by a post-orbital
groove containing pores 6 and 7 (possibly also 5) as designated by Gosline (loc. cit.).
There is an indication of a single pore on the snout, which may represent pore I of
the supraorbital canal. Between this pore and pore 6 the supra-orbital canal is either
discontinued or firmly closed over with bone ; certainly there is no external indication
of the usually quite prominent canals between the orbits found in almost all other
Procatopodinae and Rivulinae examined, the portion which incorporates pores zb, 3
and 4. The preopercular canal is complete (B), and extends forwards to below the
anterior border of the eye. Seven pores are visible. There is no mandibular canal
developed.

The pore system in Pantanodon shows many basic similarities with other genera
examined. The absence of the central portion of the supra-orbital canal may perhaps
be rare amongst the African forms, but the mandibular canal is often absent (present
in Epiplatys sexfasciatus) .

The Skull

The small size of the specimens and in many places the rather weak ossification of
the bones, have made it difficult to compare the shape, extent, or even the presence of
some cranial elements. Thus I cannot be certain whether parietals are present or not
although they appear not to be. Absence of a parietal would argue affinities with the
Oryziatinae since, contrary to Regan (1911), parietals are absent in Oryzias (Ramas-
wami, 1946). But the sutures between the dermal roofing bones of the skull in

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS

127

Pantanodon can rarely be seen. In general the skull shows no marked departure
from the normal procatopodin type as regards shape and proportions, but some small
differences can be noted.

The lacrimal (or pre-orbital) in Pantanodon is much reduced anteriorly, the bone
consisting entirely of the trough-like groove for the pre-orbital canal. In members
of the Procatopodinae the preorbital canal is but a small part of the lacrimal, the
bone expanding anteriorly and to a greater or lesser extent covering the ventral
portions of the descending rami of the maxilla and premaxilla. In Pantanodon there

B

FIG. 14. Pantanodon podoxys. Cephalic canal system. A, supra-orbital canal. B, preopercular

canal. C, infra-orbital canal.

is quite a wide space between the lacrimal and the maxilla and premaxilla tips. In
the Rivulinae the lacrimal is to some extent expanded dorsally, but quickly tapers
ventrally. On the other hand the whole preorbital region in this subfamily seems to
be narrow, the anterior edge of the lacrimal again meeting and sometimes overlapping
the maxilla and premaxilla. Myers (1931) uses the size of the preorbital in relation
to the size of the eye (respectively more or less than half the diameter of the latter)
to distinguish the Rivulinae from the Procatopodinae, and says that despite the
rather larger eye in some fishes, this character is trenchant when use is made of
comparative material. Although the lacrimal is strictly speaking very narrow in
Pantanodon, I consider that nonetheless it represents a variation on the procatopodin
pattern and is not referable to the rivulin type, principally because the preorbital

128

P. J. P. WHITEHEAD

region as a whole (i.e. anterior orbital rim to corner of mouth) is as wide as in the
former subfamily and not the latter.

The nasal in Pantanodon is also a rather narrow bone, in comparison with members
of the Procatopodinae and Rivulinae, in which it often meets and overlaps parts of
the maxilla and even the posterior part of the premaxilla.

The post-temporal is forked in Pantanodon (see Text-fig. 16), but this cannot be
taken as a subfamilial character. Thus it is unforked in Oryzias (Ramaswami, 1946) ;
forked in some Rivulinae (e.g. Epiplatys sexfasciatus) but not in others (Aplocheilus
panchax, A . lineatus and Pachypanchax playfairii) ; unforked in some Procatopodinae

f.pt.

f.sph

p.op

dent.

int. op

r. art.

FIG. 15. Pantanodon podoxys. Hyomandibular and jaw suspension.

a. pal. Autopalatine

dent. Dentary

dent. sym. Cup-shaped symphysis of the

dentary

ent. pt. Entopterygoid
f.pt. Pterotic facet
f.sph. Sphenotic facet

h.m. Hyomandibular

i. hy. Interhyal

int. op. Interoperculum

p. op. Preoperculum

q. Quadrate

r. art. Retro-articular

5. Symplectic

(e.g. Aplocheilichthys spilauchena) but forked in others (A. pelagicus for example).
The distal end of the main stem is flat and lies on the epiotic. The bifurcation of the
post-temporal occurs more proximally in Pantanodon compared with other species in
which this bone is forked.

The elements in the hyomandibular-jaw suspension series are generally rather weak
and resemble those figured for Oryzias melastigma by Ramaswami (1946), but the
articular facets of the hyomandibular are relatively large and there are two cartilagi-
nous facets and not one (see Text-fig. 15). As in Oryzias, the symplectic is narrow and

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 129

there is no metapterygoid or ectopterygoid ; Myers (1931) believed the former entirely
absent from the whole suborder Poecilioidea, but Ramaswami (loc. cit.} finds a
metapterygoid in Aplocheilus lineatus. The ventral arm of the quadrate is relatively
long. There is a small retroarticular. The junction between the angular and the
dentary is not well demarcated. The symphysial surface of the dentary is cup-shaped,
and presumably accommodates a ball-joint of cartilage.

In the opercular series, the preoperculum is reduced (rather like the lacrimal) to a
curved open groove containing the preopercular canal . The groove is not a surface
formation, but is formed in the posterior and ventral edge of the bone, as Tchernavin
(1946) has shown to be the case in all other Cyprinodontidae. Other opercular elements
scarcely differ from their counterparts in members of the Procatopodinae and the
Rivulinae.

Branchiostegal Rays

In Pantanodon there are five branchiostegal rays, three attached to the epiphyal
and two to the ceratohyal. The number of rays varied between four and six in the
other species examined. The number cannot be correlated either with subfamilies or
even genera. Thus in the Procatopodinae, Procatopus nototaenia has six, Aplocheilich-
thys spilauchena has five, while A . schoelleri, A . pumilus and A . pelagicus have only
four ; and in the Rivulinae Epiplatys sexfasciatus , Pachypanchax playfairii and
Aplocheilus panchax have six, while A . lineatus is reported to have only five (Ramas-
wami, 1946).

Pectoral Arch (Text-fig. 16)

The base of the first pectoral ray is set at about the midline of the depth at that
place and on a level with the lower edge of the eye pupil. In comparison with the
low-set pectorals of the Rivulinae (Epiplatys or Pachypanchax for example), that of
Pantanodon must be considered high, although as Myers (1931) pointed out, the
distinction between high and low pectorals is in some cases confusing. The fin base
is inclined at about 30 to the vertical.

The pectoral arch is shown in Text-fig. 16. The bifurcation of the post-temporal
occurs at the junction of this bone with the tip of the cleithrum, so that the expanded
portion of the post-temporal is contained wholly on the cleithrum, and not merely
its distal tip, as for example in Epiplatys. I have been unable to detect a post-
cleithrum in Pantanodon ; if present, this bone must be extremely small and weakly
ossified, as in Pachypanchax playfairii. The pectoral radialia are weak and the fourth
radial is barely apparent. The coracoid is not keeled ventrally, but has a narrow,
more strongly ossified bar running diagonally to its anterior tip. The coracoid
foramen is large. The scapula and the upper part of the coracoid are weakly ossified
compared with the lower part of the coracoid (suggested by stippling in Text-fig. 16).
There are twelve pectoral rays, the first three articulating directly with the edge of
the scapula, the remainder with the radials.

The pectoral arch of Pantanodon shows no marked specialisation or departure
from the general form found in the other genera examined. The distinction between

13

P. J. P. WHITEHEAD

the high-set pectorals of the Procatopodinae and the low-set pectorals of the Rivulinae
appears to be rather one of degree than of any structural difference. In the Rivulinae
the cleithrum is less steeply inclined while, at least in Pachypanchax playfairii, the
upper end of the cleithrum is not extended in compensation, but the post-temporal
is much longer. As a result, the base of the first pectoral ray is brought on a level
with the lower rim of the orbit. In the small Kenya specimens of Pantanodon the

scap.

cor. for. -

cor.

FIG. 1 6. Pantanodon podoxys. Pectoral arch (left, external face). From an alizarin

stained specimen 17 mm. S.L.

p.t. Post-temporal
cl. Cleithrum
cor. Coracoid

cor. for. Coracoid foramen

scap. Scapula

R IV. Fourth radialia

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 131

first ray is level with the centre of the eye ; in the larger paratype, it is a little below
this, but this may be due to the greater body depth of the paratype.

Pelvic Girdle (Text-figs. 17 and 18)

The pelvic support bones are T-shaped, with a very slender median portion (Text-
fig. 18). In five other genera examined (Procatopus, Pachypanchax, Aplocheilus,
Micropanchax and Aplocheilichthys) the median portion is very much broader than
this. Judging from these five genera, there seems to be considerable variation in the
shape of the two lateral processes of the pelvic support, the inner wings overlapping
more or less firmly, and the outer wings in some cases being produced posteriorly,
presumably for added support. There are no " splint " bones in Pantanodon such
as described by Gosline (1961) for Fundulus majalis ; but this author notes their
absence in another species of Fundulus, and the possession of splint bones may not
hold much signficance in the Cyprinodontidae.

FIG. 17. Pantanodon podoxys, male. Left pelvic fin, ventral view, showing hook-shaped tips
to the first three branched rays. Second ray unsegmented, but signs of segmentation in

third.

In the female specimens of Pantanodon, the six pelvic rays are perfectly normal.
In the male paratype of P. podoxys, however, the tips of the rays are modified (Text-
fig. 17). The first short ray is spiny and terminates in a point ; the second and third
rays, which are branched (the second without segmentation), have hooked, pointed
tips directed laterally ; the fourth and longest ray is clearly segmented, and the two
outermost branches have hooked tips (the innermost branch is damaged in all cases) ;
the fifth and sixth rays are normal and branched.

A thickened, spiny first pelvic ray, and the hooked tips of the second and third rays,
appear to be features which are unique amongst the toothcarps. Since they are con-
fined to the male their function may possibly be concerned with some aspect of
reproductive behaviour. The only other case of a thickened spine that I know of is
in Jordanellaflorida which has a spinous first dorsal ; here the spine is clearly a double
unit (i.e. split longitudinally) whereas in P. podoxys it is a single unit.

132 P. J. P. WHITEHEAD

Axial Skeleton

In the two alizarin preparations of P. podoxys there are 29 vertebrae, of which 19
are caudal. Of the precaudal vertebrae, the first is without ribs or epiplurals, the next
nine bear ribs, and seven of these also bear simple epiplurals also. The first caudal
vertebra has an expanded haemal arch, but this is not found in any of the subsequent
vertebrae.

FIG. 1 8. Pantanodon podoxys. Pelvic girdle, right side, ventral view, from a female 17 mm. S.L.

This vertebral formula is little different from that found in other African cyprino-
donts, but there seem to be rather more caudal vertebrae in Pantanodon. In Procatopus
nototaenia there are 11-12 + 17 vertebrae, 12 -f- 16 in Aplocheilichthys spilauchena,
and II + 17 in Micropanchax loati. In all of these the first vertebra bears an epi-
plural (but no rib). In Pantanodon, only in the first caudal vertebra is the haemal
arch expanded, resembling the condition found in Procatopus (one, two or three),

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS

133

Micropanchax (one), and Aplocheilichthys spilauchena (two) ; it should be stated,
however, that it is not always possible to decide what are elongated parapophyses, and
what are short ribs. But Pantanodon in no way approaches the condition found in
Epiplatys sexfasciatus in which up to nine haemal arches are expanded.

The Caudal Skeleton (Text-fig. 19)

The caudal skeleton is shown in Text-fig. 19. The terminology used here is that of
Gosline (1960), who preferred the use of " terminal vertebra " instead of " urostyle "

HY3

HY2

HYI

w

FIG. 19. Pantanodon podoxys. Caudal skeleton, from an alizarin stained specimen 17 mm. S.L.
HY i, 2, and 3. Hypural rays.
EP. Epural.
UR. Uroneural.

a-d. Neural spines directly concerned with caudal support.
w-z. Haemal spines directly concerned with caudal support.

134 p - J- p - WHITEHEAD

in describing the posteriormost vertebra and the one with which the lowermost and
laterally flanged hypural articulates. In Pantanodon the terminal vertebra is broadly
conical with an upturned end, but there is little indication of the line of demarcation
between the posterior border of the terminal vertebra and the base of the median
hypurals. There is a small and perhaps paired uroneural (UR) lying above the termi-
nal vertebra and overlapping the base of the uppermost hypural. There are three
hypurals. The lowest (Hy i in Text-fig. 19) is a narrow bone which has a slight
expansion along its antero-posterior face, and at each side near its base there is a
lateral flange with posterolateral muscle attachments. Hypurals 2 and 3 are triangular
and similar in shape and size. There is no indication that either is composed of several
fused elements. Above hypural 3 there is a single, narrow epural (EP), similar in
shape to hypural I but without an expanded base. The ultimate neural and haemal
spines are expanded along their anterior faces into thin wings, and the blades of the
last four neural and haemal spines are broader than those of the spines anterior to
them (a-d, w-z). The posterior four spines reach the bases of the caudal rays and
raylets, the spines on preterminal vertebra 4 (i.e. spines a and w) being rather elongated
in order to do this.

The separation of hypurals 2 and 3 is well marked in Pantanodon and does not appear
liable to subsequent fusion in larger fishes, although Hollister (1940) has shown that
a tendency for this exists in certain viviparous toothcarps. In a specimen of Epiplatys
sexfasciatus of 45 mm., however, the division is still complete, whereas in Procatopus
nototaenia at 28 mm. there is a small split anteriorly, but posteriorly the median
hypurals are fused. The hypurals are also split in Aplocheilichthys spilauchena and
Apolocheilus panchax, but no separation of these hypurals was found in Pachypanchax
playfairii (30 mm.) or in Aplocheilichthys pelagicus (40 mm.) . Thus this character is not
apparently stable amongst the Rivulinae or Procatopodinae, and it cannot be used for
subfamilial relationships. Hollister (loc. cit.) shows a totally split median hypural
in Mollienesia sphenops, a partial (anterior) median split in Gambusia holbrooki and
juveniles of Lebistes reticulatus, and completely fused bones in adults of the latter and
in Fundulus bermudae. In no case, however, does hypural I appear to either split or
fuse, but in a specimen of Aplocheilus panchax, hypural 3 is split perhaps an indi-
vidual variant.

In three of the genera examined (Epiplatys, Pachypanchax and Aplocheilichthys)
usually only the last three neural and haemal spines are concerned in the support of the
caudal rays and raylets ; Pantanodon differs from these in its elongated fourth from
last haemal and neural spines, which contributes to its deep caudal peduncle. In
Aplocheilichthys spilauchena also, four spines support the caudal, while in Aplocheilus
panchax only two neural and three haemal spines are concerned.

Uroneurals are not easy to distinguish in larger alizarin preparations, but they
appear to be present also in Procatopus, Aplocheilus and Aplocheilichthys, but could
not be seen in Epiplatys.

Apart from the small differences mentioned, the caudal skeleton of Pantanodon
differs little from the other three genera, or indeed from the poeciliids described by
Hollister (loc. cit.}. Although fusion of the hypurals and the fate of the uroneurals
may provide evidence of evolutionary progress at ordinal level (Gosline, 1960), the

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 135

variations in the toothcarps examined probably cannot be placed in phyletic series,
but represent small variations on a rather similar general plan. In certain amblyop-
soids, however, Gosline (1961) shows a rather different caudal pattern, hypural I
having apparently lost all basal attachment, and hypural 3 being fused with a post-
terminal vertebrae. I have found no post-terminal vertebra in any of the specimens
already mentioned, nor any indication that they have become fused during ontogeny.

Pseudobranchiae

In Pantanodon there are no pseudobranchiae. Pseudobranchiae are not present in
the Procatopodinae (or in the Fundulinae) but are found in the Rivulinae, and in
these three subfamilies at least, their presence or absence is coupled with the presence
or absence of vomerine teeth. The absence of vomerine teeth in Pantanodon may,
however, represent a recent specialisation connected with feeding habits.

Habitat

I have no further data for the Tanganyika specimens beyond " in swampy land
a few miles inland from Dar es Salaam." The Kenya fishes were caught in a | acre
salt-evaporating pool, one of a series built in lowlying, swampy land near enough to
the sea to be flooded at high spring tides. Nearby are similar natural pools leading to
a mangrove swamp, while inland there is a freshwater swamp and the large Sabaki
(Athi) river. The lower Sabaki was intensively studied by the author and the only
toothcarps discovered were species which had been introduced for malaria control.
The pool in which Pantanodon was found contained also juveniles of some marine
species, together with a runted population of Tilapia mossambica, a species well-
known to tolerate saline conditions. The salinity of the pool at the time was approxi-
mately 40 parts per mille.

DISCUSSION

All too often in this paper it has been necessary to curtail discussion on the relation-
ship of Pantanodon to other oviparous Cyprinodontidae owing to the lack of compar-
able studies on a particular feature. However, even the limited material on which
comparisons have been made has indicated several relatively unexplored paths which
should be followed up in a full revision of this section of the toothcarps. But since
familial, let alone generic or specific descriptions and definitions, are in many instances
inadequate, it has been possible here only to suggest Pantanodon s possible affinities,
the more so because specialisation in Pantanodon has affected several useful key
characters. It would seem, however, that Pantanodon is most nearly related to the
Procatopodinae, although this subfamily is only barely represented in East Africa.

There is no reason to suppose that Pantanodon is a primitive form, but rather that
it is a highly specialised genus which has evolved beyond the presently defined limits
of the Procatopodinae. Neither is it an annectant form between the procatopodins
and the rivulins. Zoogeographically, Pantanodon may represent a relic of a once
widespread fauna common to both the eastern and the western watersheds of Africa.
It has been suggested elsewhere (Whitehead in press) that the rivers of the eastern
watershed of Africa may contain elements of an archaic fish fauna which became

136 P. J. P. WHITEHEAD

isolated from the fauna of the west prior to the spread (in the west) of the many
genera now associated with the Nile-Chad-Niger and Congo systems. But whereas
the other cyprinodont genera in eastern Africa do not seem to have diverged greatly,
Pantanodon does not have a counterpart in the west. It may perhaps have become
adapted to a niche filled by other filter-feeders in the faunistically richer waters of the
west.

ACKNOWLEDGEMENTS

Much of this work was carried out as part of a research programme supported by a
grant from the Browne Research Committee. I wish to record my gratitude to the
Chairman and members of that committee for the interest they have shown in this
work. I also wish to thank Dr. E. Trewavas for her many helpful suggestions and for
generously allowing me to make use of her own notes on African toothcarps. I am
also grateful to Dr. P. H. Greenwood for much useful criticism, and to Professor
G. S. Myers go my thanks for many useful comments on the text.

REFERENCES

AHL, E. 1924. Zur Systematik der altweltlichen Zahnkarpfen der Unterfamilie Fundulinae.
Zool. Anz. 60 : 4955.

BERNER, L. 1947. Le development dentaire chez Gambusia affinis (Baird et Girard) . Bull. Soc.
zool. Fr. 72 : 22-30.

EATON, T. H. 1935. Evolution of the upper jaw mechanism in teleost fishes. /. Morph. Phila-
delphia, 58 : 157-169.

FOWLER, H. W. 1956. Fishes of the Red Sea and Southern Arabia, Vol. I. Branchiostomida to
Polynemida, Weizmann Science Press of Israel, Jerusalem, 240 pp.

GOSLINE, W. A. 1949. The sensory canals of the head in some cyprinodont fishes, with particular
reference to the genus Fundulus. Occ. Pap. Mus. Zool. Univ. Mich. 519 : 1-17.

1960. Contributions towards a classification of modern isospondylous fishes. Bull. Brit. Mus.

nat. Hist. (Zool.) 6 (No. 6) : 327-365.

- 1961. Some osteological features of modern lower teleostean fishes. Smith Misc. Coll.
142 (3) : 1-42.

GOSSE, J. P. 1955. Disposition speciales de 1'appareil branchial des Tilapia et Citharinus. Ann.

Soc. zool. Belg. 86 : 303-308.
HOEDEMAN, J. J. 1958. The frontal scalation pattern in some groups of toothcarps (Pisces-

Cyprinodontiformes) . Bull. Aq. Biol. 1 (No. 3) : 23-28.
HOLLISTER, G. 1940. The caudal skeleton of Bermuda shallow water fishes. IV Order Cyprino-

dontes : Cyprinodontidae, Poeciliidae. Zoologica, 25 (i) : 97-112.
MYERS, G. S. 1931. The primary groups of oviparous Cyprinodont fishes. Stanford Univ. Pub.

Biol. Ser. 6, 7-14.
1933- The genera of Indo-Malayan and African Cyprinodont fishes related to Panchax and

Nothobranchius. Copeia : 180-185.
1938. Studies on the genera of Cyprinodont fishes, XIV. Aplocheilichthys and its relatives

in Africa. Copeia, 136-143.

- 1955. Notes on the classification and names of Cyprinodont fishes. Tropical Fish Mag.
March : 7.

1960. The endemic fish fauna of Lake Lanao, and the evolution of higher taxonomic

categories. Evolution, 14 (3) : 323-333.
NICHOLS, J. T. & POPE, C. H. 1927. The fishes of Hainan. Bull. Amer. Mus. nat. Hist., New

York, 54 : 321-394.
RAMASWAMI, L. S. 1946. A comparative account of the skull of Gambusia, Oryzias, Aplocheilus

and Xiphophorus (Cyprinodontes : Teleostomi). Spolia zeylan. 24 (3) : 181-192.

THE PANTANODONTINAE, EDENTULOUS TOOTHCARPS 137

REGAN, C. T. 1911. The osteology and classification of the teleostean fishes of the order Micro-

cyprini. Ann. Mag. nat. Hist. (8) 7 : 320-327.
ROSEN, D. E. & MENDELSON, J. R. 1960. The sensory canals of the head in poeciliid fishes

(Cyprinodontif ormes) , with reference to dentitional types. Copeia, No. 3, 203-210.
SMITH, H. 1945. The freshwater fishes of Siam, or Thailand. Bull. U.S. nat. Mus. Washington,

188 : 1-522.
SUYEHIRO, Y. 1942. A study on the digestive system and feeding habits of fish. Jap. J. Zool. 10

(i) : 1-303.
TCHERHAVIN, V. V. 1946. On the lateral line system of some Cyprinodonts (Order Microcyprini) .

Ann. Mag. nat. Hist, (n) 13 : 429-432.

WHITEHEAD, P. J. P. 1959. The feeding mechanism of Tilapia nigra. Nature, 184 : 1509-1510.
1962. A new species of Synodontis (Pisces : Mochocidae) and notes on a mormyrid fish

from the eastern rivers of Kenya. Rev. Zool. Bot. afr. 65 (1-2) : 97-119.
WORTHINGTON, E. B. 1932. Scientific results of the Cambridge Expedition to the East African

Lakes, 1930-31. 2. Fishes other than Cichlidae. /. Linn. Soc. London, 38 : 121-134.

PRINTED IN GREAT BRITAIN BY
ADLARD AND SON, LIMITED,
BARTHOLOMEW PRESS, DORKING

A REVISION OF THE

LAKE VICTORIA HAPLOCHROMIS

SPECIES (PISCES, CICHLIDAE)

PART V

P. H. GREENWOOD

BULLETIN OF
THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 9 No. 4

LONDON: 1962

A REVISION OF THE

LAKE VICTORIA HAPLOCHROMIS

SPECIES (PISCES, CICHLIDAE)

PART V

BY

P. H. GREENWOOD

Department of Zoology, British Museum (Natural History)

Pp. 139-214 ; Plate i ; 25 Text-fogs.

BULLETIN OF

THE BRITISH MUSEUM (NATURAL HISTORY)

ZOOLOGY Vol. 9 No. 4

LONDON: 1962

THE BULLETIN OF THE BRITISH MUSEUM

(NATURAL HISTORY), instituted in 1949, is
issued in five series corresponding to the Departments
of the Museum, and an Historical Series.

Parts will appear at irregular intervals as they become
ready. Volumes will contain about three or four
hundred pages, and will not necessarily be completed
within one calendar year.

This paper is Vol. 9, No. 4 of the Zoology series.

Trustees of the British Museum 1962

PRINTED BY ORDER OF THE TRUSTEES OF
THE BRITISH MUSEUM

Issued November 1962 Price Twenty Six Shillings

A REVISION OF THE

LAKE VICTORIA HAPLOCHROMIS

SPECIES (PISCES, CICHLIDAE)

PART V

By P. H. GREENWOOD

CONTENTS

INTRODUCTION ....
Haplochromis brownae sp. nov.
Haplochromis guiarti (Pellegrin)
Haplochromis bayoni (Blgr.)
Haplochromis serranus (Pfefifer)
Haplochromis victorianus (Pellegrin)
Haplochromis nyanzae sp. nov.
Haplochromis bartoni sp. nov. .
Haplochromis estor Regan
Haplochromis dentex Regan
Haplochromis artaxerxes sp. nov.
Haplochromis longirostris (Hilgen.)
Haplochromis mento Regan
Haplochromis mandibularis sp. nov.
Haplochromis gowersi Trewavas
Haplochromis macrognathus Regan
Haplochromis pellegrini Regan
Haplochromis percoides (Blgr.)
Haplochromis flavipinnis (Blgr.)
Haplochromis cavifrons (Hilgen.)
Haplochromis plagiostoma Regan
Haplochromis michaeli Trewavas

DISCUSSION ....

SUMMARY .....

ACKNOWLEDGEMENTS .

REFERENCES

Page
141
142

145
149

152

156
159
161
164
167
170
171

174
178
1 80

183
1 86
189
192
196
199
203
206

212
213
213

INTRODUCTION

THIS paper is one of two dealing with the piscivorous Haplochromis of Lake Victoria.
In this part, I have tried to consider a representative sample of species which cover
the different morphological types in this trophic group. A species that I consider to
represent a morphological type from which the fish-eating species could have evolved,
is also described.

ZOOL. 9, 4

i 4 2 P. H. GREENWOOD

Haplochromis brownae sp. nov.
Text-figs, i and 25

Haplochromis stanleyi (part) Boulenger, 1915, Cat. Afr. Fish. 3, 295 (one specimen B.M. (Nat.
Hist.) Reg. No. 1909 . 5 . 4 . 28 ; collected by Bayon in the Sesse Islands ; apparently this speci-
men was not considered by Regan in his revision of 1922).

Paratilapia guiarti (part) : Boulenger, 1915, op. cit. 3, 336 (one specimen B.M. (Nat. Hist.)
Reg. No. 1906.5.30.354).

Haplochromis guiarti (part) : Regan, 1922, Proc. zool. Soc. London, 174 (the same specimen as

above) .

Holotype : A specimen 104-0 mm. S.L. from Entebbe ; B.M. Reg. No. 1906.5.30.

354-

Description. Based on forty-nine specimens (including the holotype) 72-104 mm.
standard length.

Depth of body 32-2-39-8 (mean, M, 35-1) per cent of standard length, length of
head 30-2-33-4 (M = 31-6) per cent. Dorsal profile of head straight or slightly concave
above the eyes, sloping at about 45.

Preorbital depth 13-8-19-4 (M = 16-3) per cent of head, least interorbital width
26-0-34-0 (M 29-8) per cent. Snout as long as broad or very slightly broader, its
length 28-0-33-3 (M = 30-8) per cent of head ; eye diameter 26-0-31-3 (M 28-6),
depth of cheek 18-9-25-9 (M = 22-7) per cent.

Caudal peduncle length 14-6-20-4 (M 17-6) per cent of standard length, 1-3-1-8
(mode 1-6) times as long as deep.

Mouth horizontal or slightly oblique, the jaws equal anteriorly, the lips not
thickened. Length of lower jaw 38-0-42-9 (M = 40-3) per cent of head, 1-7-2-1
(mode 2-0) times as long as broad. Posterior tip of the maxilla reaching or almost
reaching the vertical through the anterior orbital margin.

Gill rakers 9-12 (modal numbers 10 and n) on the lower part of the first gill arch,
the lowermost two to five rakers short and stout, the remainder usually slender
(specially in fishes with more than ten rakers) but occasionally stout.

Scales ctenoid ; lateral line with 30 (f.i), 31 (f.5), 32 (32) or 33 (f.n) scales.
Cheek with 2 or 3 imbricating rows. Six or 7 (rarely 8) scales between the lateral
line and the dorsal fin origin, 6-8 (modes 7 and 8) between the pectoral and pelvic
fin bases.

Fins. Dorsal with 22 (f.i) 24 (f-4), 25 (f.37) or 26 (f.7) rays, comprising 13 (f.i), 14
(f.i) 15 (f.i8) or 16 (f.29) spinous and 8 (f.i), 9 (f.24) or 10 (24) branched rays. Anal
with ii (f.4), 12 (f-42) or 13 (f.3) rays, comprising 3 spinous and 8-10 branched rays.
Caudal truncate, scaled on the basal half only. Pectoral fin from slightly shorter than,
to as long as the head, 24-4-33-4 (M = 28-8) per cent of standard length.

Teeth. In most specimens the outer row in both jaws is composed of relatively
stout and bicuspid teeth, but in many fishes (irrespective of size) the posterolateral
teeth in the upper jaw are unicuspid or tricuspid. In a few specimens there is an
admixture of bi- and unicuspid teeth anteriorly in both jaws. There are 50-66
(M = 56) teeth in the upper, outer series.

Teeth in the inner rows are tricuspid and arranged in 2 or 3 (rarely 4) series in the
upper jaw and i or 2 series in the lower jaw.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 143

Osteology. The neurocranium is that of a generalized Haplochromis. The supra-
occipital crest slopes fairly steeply and is moderately deep ; the ethmoidal region
slopes fairly steeply and the vomer is curved ventrally. The preotic part of the skull
is 55-58% of the basal length (as measured from the tip of the vomer to the posterior
rim of the basioccipital) . The lower pharyngeal bone is slender, the teeth cuspidate
and fine, without markedly enlarged teeth in the median rows. The dentigerous
surface is as broad as it is long and there are 34-40 rows (counted antero-posteriorly)
of teeth.

Number of vertebrae : 29 (f.2), 30 (f.g) or 31 (f.3), comprising 12 (f.i), 13 (f.4) 14 (f.8)
or 15 (f.i) precaudal and 16 (f.8) or 17 (f.6) caudal elements.

FIG. i. Haplochromis brownae ; natural size. (Drawn by Lavinia Beard.)

Coloration in life. Adult males. Ground colour grey-green dorsally, shading to
silver-blue ventrally ; snout dark grey-green. In sexually active fishes a carrot-
orange flush develops on the cheek, operculum, flanks and belly ; in such fishes there
is also a pronounced lachrymal stripe. Dorsal, caudal and anal fins sooty-grey, the
dorsal with three to five horizontal rows of ruddy spots between the spines and rays ;
anal ocelli (usually a single row of three or four) orange. Pelvic fins black.

Females have a silver-grey ground coloration (lighter than that of the males)
shading to silver ventrally. All fins are hyaline, except the pelvics which are light
lemon-yellow. In some individuals there are small orange spots in the position of the
anal ocelli in males.

Colour of preserved material. Females. Brownish above, silvery below and on the
cheek ; a faint lachrymal stripe is visible as are five transverse bars on the flanks,
none of which reaches the ventral profile ; in some specimens there is a short dark,
midlateral streak situated above the anal fin. All fins are colourless. In males the
coloration is more variable, some fishes having a female type coloration (see above)

144 p - H - GREENWOOD

but with a more intense lachrymal stripe and a distinct vertical postorbital bar. All
fins, except the pelvics, are hyaline, the dorsal with dusky lappets ; the pelvics are
dark laterally but greyish medially.

Other specimens have a similar coloration but the ground colour is darker, the
pectoral region sooty and the pelvic fins entirely black. These differences may be
associated with the fish's sexual state.

A third varient (undoubtedly attributable to differences in the mode of preserva-
tion) almost resembles the live colours. Here the belly, operculum and cheek still
retain traces of the carroty flush, the transverse bars are very faint, the lachrymal
stripe is intense and the dorsal fin distinctly maculate. The pectoral region is dusky
and the pelvic fins are black.

Distribution. Known only from Lake Victoria.

Ecology. Habitat. The species is confined to sandy or shingle beaches where the
water is less than thirty feet deep ; such areas are always relatively or completely
exposed to wave action. No specimens have been recorded from areas where the
substrate is soft.

Food. Twenty of the fifty specimens examined contained ingested material. Of
these, three had fed on small fishes (two specimens exclusively so and one on adult
termites as well) tentatively identified, in one case, as Engraulicypris argenteus, but
unidentifiable in the others. Since all three fishes are males the possibility of the
fishes having swallowed their own young is eliminated.

The food from the seventeen other fishes is very varied, but it does seem that
H. brownae is predominantly insectivorous. Larval and pupal Diptera are the com-
monest foods, but after heavy termite hatches the food can become exclusively winged
termites. Other common materials from the gut are macerated fragments of plant
tissue and colonial blue-green algae (cf. Rivularia). Neither plant shows any signs of
digestion. Sand grains and bottom debris are infrequent. Since large numbers of
Rivularia colonies are often broken awav from their substrates after heavy swells, it
is possible that the fishes snap up the floating masses as they drift past. Similar
concentrations of colonial blue-greens have been found in the guts of other Haplo-
chromis whose usual feeding habits do not involve feeding from the bottom. Certainly
plants can contribute little to the diet of Haplochromis brownae because plant material
is apparently indigestible.

Breeding. The species is a female mouth brooder ; the smallest specimen available
(72 mm. S.L.) is brooding. Both sexes seem to reach the same maximum adult size.

Affinities. Structurally, and in its diet, H. brownae is a generalized lacustrine
Haplochromis. It differs, however, from the other generalized and insectivorous
species of Lake Victoria in having a high number of gill rakers (9-12 cf. 6-9) ; in fact
very few Victoria species of any structural or trophic group have more than ten
rakers. In most other characters H. brownae is very similar to H. melanopus,
another littoral insectivore with similar ecology. The two species do, however, differ
in the shape of the lower pharyngeal bone and its dentigerous surface. In H. melanopus
the toothed area is clearly broader than long, whereas in H. brownae the area is
equilateral, or almost so. The species also differ markedly in the colours of the breed-
ing male.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 145

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1906.5.30.354 . Entebbe . Degen

(Holotype)

1909.5.4.28 . Sesse Isls. Bayon

1911.3.3.22 . Jinja, Ripon falls . ,,

1962.3.2.96-7 . Grant Bay . E.A.F.R.O.

1962.3.2.109-113 . Beach near Nasu pt. . ,,

1962.3.2.114-9 . Buka Bay . ,,

1962.3.2.120-7 . Entebbe Harbour .

1962.3.2.128 . Beach near Nasu pt. .

Tanganyika

1962.3.2.76-88 . Mwanza, Capri Bay . ,,

1962.3.2.89-95 . Majita Beach .

1962.3.2.98-108 . Beach near Majita .

Haplochromis guiarti (Pellegrin) 1904
Text-figs. 2 and 25

Tilapia guiarti Pellegrin, 1904, Bull. Soc. zool. France, 29, 186 ; Idem., 1905, Mem. Soc. zool.

France, 17, 184, pi. 16, fig. i.

Paratilapia guiarti (part) : Blgr., 1915, op. cit., 3, 334 (not the figured specimen).
Paratilapia victoriana (part) : Boulenger, 1915, op. cit., 341 (one specimen, B.M. (Nat. Hist.)

Reg. No. 1906.5.30.281).

Haplochromis cinereus (part) : Regan, 1922, Proc. zool. Soc. London, 166 (specimen noted above).
Haplochromis guiarti (part) : Regan, 1922, op. cit., 174.
Haplochromis nigroventralis Lohberger, 1929, Akad. Anz. Wien., 66, 207.

Specimens included in Regan's (1922) synonymy as Tilapia perrieri and Paratilapia
longirostris are no longer considered to be specimens of H. guiarti. In his synonymy
Regan also included part of Tilapia pallida (Boulenger, 1915 Cat. Afr. Fish. 3, 232).
Despite an extensive search, I cannot find any specimens referred to this species
which Regan might have examined. For this reason I have not included T. pallida
(part) in my synonymy.

Holotype. A female 114 mm. S.L. (Paris Museum No. 04 x 150) from the Kaviron-
do Gulf, Kenya.

Description. Based on fifty-five specimens (83-177 mm. S.L.) including the holotype
of the species and that of H. nigriventralis. All these fishes came from Lake Victoria
(see below, p. 147).

Depth of body 27-3-36-5 (M = 32-3) per cent of standard length, length of head
29-5-33-8 (M = 31-4) per cent. Dorsal profile of head slightly curved and sloping at
an angle of ca 4o-45.

Preorbital depth 16-3-21-5 (M = 18-3) per cent of head, least interorbital width
23-4-30-2 (M = 27-4) per cent. Snout longer than broad, or, rarely, as long as broad,
its length 31-7-37-5 (M = 34-4) per cent of head. Eye diameter shows slight negative

146

P. H. GREENWOOD

allometry with standard length, being 23-6-29-8 (M = 26-5) per cent in fishes < 115
mm. S.L. (N = 16) and 19-7-25-3 (M = 22-0) in larger individuals (N = 39). Depth
of cheek shows very slight positive allometry, 20-0-29-0 ( M= 25-9) per cent of head
for the whole sample.

Caudal peduncle 16-2-20-8 (M 18-9) of standard length, 1-4-2-0 (mode 1-5)
times as long as deep.

Mouth slightly to moderately oblique, the jaws equal anteriorly. Length of lower
jaw 39-2-48-2 (M = 44-4) per cent of head, 1-5-2-3 (mode 2-0) times as long as broad.
Lips not thickened, the posterior tip of the maxilla not quite reaching the vertical
through the anterior orbital margin in most fishes but reaching this point in a few.
The dentigerous arm of the premaxilla shows no medial antero-posterior lengthening.

FIG. 2. Haplochromis guiarti ; about X N.S. (From Boulenger, Fish. Nile.)

Gill rakers moderately stout but not stubby ; 9-11 (mode 10) on the lower part of
the first arch, the lowermost 1-3 rakers reduced.

Scales ctenoid ; lateral line with 32 (f.2), 33 (f.26), 34 (f.i8), 35 (f.y) or 37 (f.i)
scales. Cheek with 3 (less commonly 4 and rarely 2) imbricating rows. Six or 7
(rarely 5, 5^ or 8) scales between the lateral line and the dorsal fin origin ; 6 or 7
(rarely 8 or 9) between the pectoral and pelvic fin bases.

Fins. Dorsal with 25 (1.35), 26 (f.19) or 27 (f.i) rays, comprising 15 (1.12), 16 (f-4i)
or 17 (f.2) spinous and 9 (f.25), 10 (f.2g) or II (f.i) branched rays. Anal with 11-13
rays, comprising 3 spinous and 8-10 (mode 9) branched rays ; in one exceptional
specimen there are 4 spines and 9 rays. Caudal truncate, pelvics (particularly in
adult males) with the first two branched rays elongated but rarely extending to beyond
the first or second branched anal ray (cf . H. bayoni where these pelvic rays are greatly
produced). Pectoral fin 21-2-27-4 (M = 25-0) per cent of standard length.

Teeth. In fishes <H5 mm. S.L. the anterior teeth of the outer row in both jaws
are usually unicuspid whilst the lateral teeth are slender and bicuspid. The majority
of fishes >H5 mm. S.L. have all the outer teeth unicuspid, relatively slender and
slightly incurved. However, in several specimens the posterolateral teeth (especially

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 147

in the upper jaw) are slender and bicuspid or weakly bicuspid. There are 48-74
(M = 62) teeth in the outer row of the upper jaw, the number showing a weak positive
correlation with standard length.

In most fishes of all sizes the inner teeth are tricuspid but it is not uncommon to
find a mixture of tri- and unicuspid teeth, the former predominating. Fishes with
the entire inner series composed of unicuspids are rare even amongst specimens >i5o
mm. S.L. The inner teeth are implanted obliquely and arranged in 3 or 4 (less
frequently 2 or 5, rarely 7) rows in the upper jaw and usually in 2 (less commonly i,
3 or 4) rows in the lower jaw. It seems likely that the number of inner tooth rows is
positively correlated with size since there is a tendency for fishes over 125 mm. S.L.
to have the higher numbers of rows. Only one fish has seven premaxillary rows (it is
a large specimen [170 mm. S.L.] but not the largest) ; the arrangement of these
teeth is very irregular and suggests some ontogenetic disturbance.

Osteology. The neurocranium of H. guiarti can be considered as a basic type amongst
the piscivorous species. Although it shows several characteristics of the " extreme "
predator skull (see p. 209) in an early phase of development, it also shows affinity
with the generalized Haplochromis skull type as seen, for instance, in H. brownae.
These points are discussed on pages 207-9.

Vertebrae. Thirty or 31, comprising 13 or 14 precaudal and 16-18 caudals. (Based
on six specimens.)

Lower pharyngeal bone fine, its dentigerous surface equilateral or somewhat
broader than long. The teeth are cuspidate and slender, the most posterior two or
three teeth of the two median rows are generally enlarged ; there are 22-28 (mode
26) rows of teeth.

Coloration in life. Adult males. Dorsal surface of head and body intense malachite
green shading ventrally to silver. All fins colourless except the pelvics which are black,
and the caudal which is dark ; anal ocelli (3 or 4 in a single row) bright orange.
Females have similar coloration except that the pelvics are light yellow and the anal
ocelli are absent or represented by small orange spots. After death the colours
change rapidly to dark grey-black above and silver below.

Colour in preserved material. Adult males. Ground colour greenish-grey overlying
silver, chest and belly sooty ; a dark lachrymal stripe is generally present. Dorsal
fin sooty, the lappets darker and with a small hyaline area between each lappet ;
caudal dark on the proximal two-thirds, hyaline distally. Anal fin greyish, the
ocelli darker grey ; pelvics black. Sexually quiescent males have a female type of
coloration but with the transverse bars fainter and the anteroventral part of the
body sooty. The dorsal fin is sooty, with black lappets ; the caudal is maculate, the
anal hyaline and the pelvics black. Adult females are dark grey dorsally, greyish-
silver below and becoming silvery-white on the chest and belly. Often there are
traces of six transverse bars of lengths varying from short to almost the full depth
of the body ; sometimes these bars are connected by a fainter longitudinal stripe.
All the fins are hyaline, the caudal often maculate on its proximal half to two-thirds.

Distribution. Lake Victoria and Lake Edward. I have examined three specimens
from the latter area. The fishes seem to be referable to H. guiarti but I await further
material before confirming the occurrence of the species in Lake Edward.

148 P. H. GREENWOOD

Ecology. Habitat. The species has a lake-wide distribution but is confined to sand
or shingle beaches where the water is less than 20 ft. deep. No specimens have been
taken over a soft substrate. Information on the distribution and feeding habits of
H. guiarti given by Graham (1929) must be discounted because several species were
confused under this name in his report.

Food. The material found in the alimentary tract of H. guiarti may be divided, in
order of abundance, into three groups : (i) Fishes (particularly small cichlids) (ii)
Insects (particularly winged Termites [Isoptera] and chironomid pupae [Diptera], less
frequently larval Povilla adusta [the boring may-fly]) (iii) Fragments of phanerogam
tissue associated with the colonial blue-green alga Rivularia.

Basing the estimate on both frequency of occurrence and on volume, the difference
between the amount of fish and the amount of insect food eaten is not great nor is it
correlated with the size of the fish. It is difficult to assess the nature of insect-eating
habits in this species. At one locality where seine hauls were carried out regularly,
it was possible to associate the habit with any high level of insect activity. In this
particular case insect-eating might be considered a faculative response to a sudden
abundance of a readily available food. The simultaneous occurrence of other species
also gorged with the same insects (particularly when the usual diet of these fishes
was not primarily insects) seems to indicate that many Haplochromis species are
" opportunistic " in their feeding habits.

None of the plant matter showed signs of digestion and cannot thus be considered
as " food ". As in the case of H. brownae (see p. 144) the material may have been
accidentally ingested.

In general, the food and feeding habits of H. guiarti are similar to those of H.
brownae. The principal difference lies in the greater proportion of fish eaten by
H. guiarti.

Breeding. Haplochromis guiarti is a female mouth-brooder ; brooding females are
found in the same habitat as non-breeding individuals. Sexual maturity is reached
at a length of about 100 mm. although some individuals of 110-113 mm. are still
immature. There is no marked sexual dimorphism in adult size.

Affinities. Perhaps the nearest living relative of H. guiarti is H. brownae. The
species differ in many respects but the divergence of H. guiarti lies in those anatomical
(and associated morphometric) characters which we find in a more exaggerated
condition amongst the entirely piscivorous predatory species. The resemblance
between H . guiarti and H. brownae is most apparent in young specimens (80-90 mm.
S.L.) of the former and adults of the latter species (70-90 mm. S.L.), but even at this
size H. guiarti is distinguishable on certain characters (longer jaws and snout) which
are part of the " predatory species" character complex. However, it is possible that
H. guiarti, itself still a relatively generalized species amongst the predators, evolved
from an H. brownae-like stem.

Amongst the larger predatory species, H. guiarti is perhaps related to H . squamu-
latus ; but the overall resemblance is less than that between H. guiarti and H. brownae.
Another species with about the same degree of relationship as H. squamulatus is
H. bayoni (see p. 152).

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 149

Study material and distribution records

Museum and Reg. No. Locality Collector

B.M. (N.H.). 1906.5

,, 1906.5

1906.5

1906.5

1906.5

,, 1906.5

1909-5

i, 1928.1
1962.3
1962.3

,. 1962.3
1962.3
1962.3
1962.3
1962.3
1962.3
,, 1962.3

Paris Museum 04 x 150

(Holotype)
B.M. (N.H.). 1962.3.2.247

,, 1962.3.2.250-1

B.M. (N.H.). 1962.3.2.249

B.M. (N.H.). 1901.6.24.89

I9H-3-3-23
1962.3.2.252

Uganda

30.210-2 . Entebbe
30.213-5

30.220-8 .

30 . 230 . Nsonga

30.281 . Entebbe

30.355-361 Entebbe

4.6-7 . Sesse Isls.

25 . 22 . Entebbe

2 . 246 . Entebbe (Airport Beach)

2.240-2 . Kagera Port

2 . 243-5 . Ramafuta Isl. (Buvuma Channel)

2 . 260-3 . Entebbe Harbour

2 . 253-9 . Bufuka Bay

2.264-270 . Beach near Nasu point

2.248 . Karenia (near Jinja)

2.271-3 . Jinja pier

2.274-6 . Buka Bay

Kenya

Kavirondo Gulf

Near Usoma Lighthouse,
Kavirondo Gulf
Kasingiri Gingo,
Kavirondo Gulf

Tanganyika

Majita Beach

Lake Victoria, Locality Unknown

Degen

Bayon
Pitman

E.A.F.R.O.

Sir H. Johnston

Bayon
E.A.F.R.O.

Haplochromis bayoni (Blgr.) 1909
Text-fig. 3

Paratilapia bayoni (part) Blgr., 1909, Ann. Mus. Genova (3), 4, 304, fig. (the figured specimen
only ; this specimen is now chosen as the lectotype) ; Idem, 1915, Cat. Afr. Fish., 3, 337,
fig. 227 (figured specimen only).

The syntype on which Regan (1922) based his redescription of the species is no
longer considered to be H. bayoni.

Lectotype. A male, 147 mm. S.L. (collected in the Sesse Islands by Bayon) i.e. the
specimen figured by Boulenger (1909), now in the collections of the Museo Civicio
distoria Naturale, Genoa (No. C.E. 12976).

150 P. H. GREENWOOD

Description. Based on twenty-three fishes (including the lectotype) 82-154 mm.
S.L.

Depth of body 27-0-35-3 (M = 32-2) per cent of standard length, length of head
35-0-38-0 (M = 36-3) per cent. Dorsal head profile sloping at an angle of ca. 3O-35,
its otherwise straight outline broken anteriorly by the prominence of the premaxillary
pedicels.

Preorbital depth 17-7-21-6 (M = 19-6) per cent of head length, least interorbital
width 20-8-25-9 (M = 23-2) per cent. Snout longer than broad (1-2-1-3 times), its
length 35-3-39-5 (M = 37-5) per cent of head. Eye diameter 17-5-21-9 (M = 20-0)
per cent, depth of cheek 23-1-28-1 (M = 25-0) per cent.

Caudal peduncle 14-5-17-2 (M = 16-2) per cent of standard length, 1-1-1-6 (modal
range 1-3-1-4) times as long as deep.

FIG. 3. Haplochromis bayoni ; lectotype, .73 x N.S.
(From Boulenger, Ann. Mus. Genova.)

Mouth slightly oblique, the jaws equal anteriorly or the lower projecting slightly ;
lower jaw length 42-5-48-0 (M = 45-6) per cent of head, 1-7-2-5 (no definite mode)
times as long as deep. Lips slightly thickened, the premaxilla noticeably expanded
anteroposteriorly in the midline. Posterior tip of the maxilla not reaching the level
of the anterior orbital margin, but generally reaching a point slightly behind the
vertical through the posterior tip of the premaxillary pedicels.

Gill rakers usually slender or a mixture of slender and stout, 8-10 (mode 9) on
the lower part of the first gill arch, the lowermost 1-3 rakers reduced.

Scales ctenoid ; the lateral line with 29 (f.i), 30 (f.i), 31 (1.9), 32 (f.8), 33 (f.3) or
34 (f.i) scales. Cheek with 4 (less frequently 3) series of imbricating scales. Five or
6 (rarely 6| or 7) scales between the lateral line and the dorsal fin origin, 5 or 6 (less
frequently 7) between the pectoral and pelvic fin bases.

Fins. Dorsal with 23 (f.i), 24 (f.8), 25 (f.13) or 26 (f.i) rays, comprising 14 (f-4),
15 (f.i8) or 16 (f.i) spinous and 9 (f.8), 10 (f.13) or n (1.2) branched rays. Anal with

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 151

12 (mode) or 13 rays, comprising 3 spinous and 9 or 10 branched elements. Caudal
truncate. Length of pectoral fin 21-2-28-0 (M = 25-5) per cent of standard length.
The pelvic fins have the first unbranched ray greatly produced and filamentous, its
tip reaching to at least the third branched anal ray and usually to between the third
and sixth rays. In some specimens it reaches to beyond the anal base. Both Boulen-
ger (1911) and Regan (1922) imply that the ray is longer in males than in females but
from my sample I can find no clear-cut sexual dimorphism. (Boulenger's remarks are
probably attributable to his having another species in his study material). This
lack of marked dimorphism is very unusual since in all other Victoria Haplochromis
the adult male has noticeably longer outer pelvic rays. Indeed, the marked hyper-
trophy of the pelvic rays in both sexes constitutes one of the most reliable diagnostic
characters for H. bayoni.

Teeth. In every specimen examined the outer row of teeth in both jaws is composed
of unicuspid, slightly to moderately curved and relatively stout teeth. There are
34-52 (M = 44) outer teeth in the upper jaw, the number perhaps showing a positive
correlation with standard length.

The inner rows are also composed of unicuspid teeth, and are implanted obliquely;
there are 2 (less commonly 3) series in the upper jaw and 2 (less commonly i) in the
lower.

Osteology. The neurocranium is clearly derived from the H. guiarti type but is
somewhat more advanced towards the " extreme " predator type of H. mento and
H. macrognathus (see p. 209). Compared with H. guiarti the neurocranium has the
ethmoid-vomer region more strongly decurved, the slope of the preorbital face less
steep (ca. 35), its height less (ca. 3 times in basal length cf. 2-5 times) and a lower
supraoccipital crest. The relative length of the preotic portion of the skull is the same
in both species (65% of basal length).

Vertebrae. 28-30, comprising 12-13 precaudal and 16 or 17 caudal elements (4
specimens examined) .

Lower pharyngeal bone triangular, its dentigerous surface equilateral. The teeth
are relatively coarse and are cuspidate ; the teeth of the two median rows slightly
enlarged. There are 18-20 tooth rows.

Coloration. Data on the live colours of H . bayoni are not available. In preserved
fishes females and immature males are greyish-brown above and silver below, some-
times with very faint traces of four short, transverse bars of irregular outline lying
midlaterally ; less frequently these bars are joined by an even fainter longitudinal
band. All fins are hyaline, the soft dorsal maculate and the proximal two-thirds of
the caudal dark with traces of maculae. The only adult males available are sexually
quiescent and have a coloration like that of the females except for a more intense
midlateral stripe and more distinctly maculate dorsal. The pelvics are slightly dusky
and there is a single row of three opaque ocelli on the anal fin.

Distribution. Known only from Lake Victoria. Earlier records (Boulenger, 1915)
from Lake Kyoga were based on a misidentified specimen.

Ecology. Habitat. The species is apparently confined to water less than thirty feet
deep and to hard substrates (sand or shingle). The majority of specimens come from
exposed sandy beaches but some are from fairly exposed bays where the bottom was

152 P. H. GREENWOOD

of sand or sand overlain by a thin slick of mud. Since only a few of Graham's
specimens were preserved, it is impossible to use his records (Graham, 1929) for
distributional purposes. If it be assumed that he identified his specimens correctly,
then his locality records confirm my idea of the species' distribution.

Food. Information on the feeding habits of H. bayoni is scanty. Seven of the fifteen
fishes examined (size range 82-154 rnm. S.L. ; from six different localities) had food
in the stomach and /or intestines. In each case the food consisted entirely of finely
macerated fish remains, identifiable in four specimens as being small cyprinid fishes
(probably Barbus sp.). One fish (153 mm. S.L.) had the remains of two cyprinids
(ca. 35 mm. S.L.) in the stomach and the remains of at least one other fish in the intes-
tines.

Breeding. Nothing is known about the breeding habits of this species. Both sexes
reach maturity at a size between no and 125 mm. S.L., and both reach the same
maximum adult size.

Affinities. The extreme elongation of the first pelvic ray is unique amongst Lake
Victoria Haplochromis and serves as a ready diagnostic character. On more funda-
mental structures, particularly the syncranial architecture, H . bayoni shows affinity
with H. guiarti but the relationship is not especially close. In these same characters
H. bayoni exhibits a further continuation of the trend leading towards the H. mento
H. macrognathus level of syncranial organization. Phyletically, H. bayoni could be
considered as an isolated (but by no means aberrant) offshoot from an H. guiarti
like stem.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda
Genoa Museum, No. C.E. 12976 . Sesse Isls. . Bayon

(Lectotype)
B.M. (N.H.). 1962.3.2.160-3 . Beach near . E.A.F.R.O.

Hannington Bay

,, 1962.3.2.170-1 . Beach near Nasu point .

,, 1962.3.2.172-4 . Entebbe, Airport beach . ,,

,, 1962.3.2.175-6 . Entebbe, Harbour . ,,

1962.3.2.177 . Kagera port . ,,

Tanganyika
1962.3.2.164-8 . Majita Beach .

Kenya
,, 1962.3.2.169 . Kavirondo Gulf . ,,

Lake Victoria, Locality Unknown
1962.3.2.178

Haplochromis serranus (Pfeffer) 1896
Text-figs. 4, 5 and 25

Hemichromis serranus Pfeffer, 1896. Thierw. D. ost. Afr., Fische, 23.

Paratilapia prognatha (part) : Blgr., 1915, Cat. Afr. Fish., 3, 333 (specimen B.M. (N.H.), Reg.

No. 1906, 5.30.263).
Haplochromis acutirostris Regan, 1922, Proc. zool. Soc. Land., 180 (the lectotype [and figured

specimen] Reg. No. as above).

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 153

I do not consider that any of the specimens used by Boulenger (1915) or Regan
(1922) to redescribe the species can be referred to H. serranus. My concept of H.
serranus is based on Pfeffer's original description and on supplementary notes and a
figure made from the presumed type specimen by Dr. E. Trewavas. The additional
material now available agrees more closely with my idea of the type than do Boulen-
ger's and Regan's fishes.

Description. Based on fifty-two specimens, 101-205 mm. S.L. (including the type
of H. acutirostris}.

Depth of body 327-39-2 (M = 36-0) per cent of standard length, length of head
34-8-387 (M = 36-3) per cent. Dorsal head profile straight or slightly curved,
sloping at an angle of 30-4O, the premaxillary pedicels very prominent.

1cm.

FIG.

Haplochromis serranus. (Outline drawing of a specimen in the Berlin Museum
believed to be the holotype ; made by Dr. E. Trewavas.)

Preorbital depth 14-6-20-0 (M = 177) per cent of head, ratio Eye/Preorbital 1-1-1-5,
(M = 1-3.) Least interorbital width 20-4-26-8 (M = 23-3) per cent. Snout as long as
broad in most specimens, but in some fishes <I3O mm. S.L. it is slightly longer than
broad ; snout length 30-8-37-0 (M = 34-0) per cent of head, diameter of eye 20-4-26-0
(M = 23-3) per cent, depth of cheek 22-9-31-5 (M = 27-5).

Caudal peduncle length 13-3-19-6 (M = 15-4) per cent of standard length, 1-1-1-5
(mode 1-2) times as long as deep.

Lower jaw moderately oblique, sloping at an angle of 25-3O, projecting slightly
and with a distinct mental bump ; its length 47-7-60-0 (M = 54-3) per cent of head
and 1-8-2-5 (mode 2-0) times the breadth. Posterior tip of the maxilla extending to
below the eye or to the vertical through the anterior orbital margin in most fishes,
but not quite reaching this point in a few.

Gill rakers. Short and stout, 8 or 9 (rarely 7 or 10) on the lower part of the first gill
arch, the lowermost one or two rakers reduced.

Scales ctenoid ; lateral line with 30 (f.2), 31 (f.25), 32 (f.i8), 33 (.5) or 34 (f.i)
scales, cheek with 4 (less frequently 5, rarely 3 or 6) rows ; 6 or 7 scales between

154 P- H - GREENWOOD

the lateral line and the dorsal fin origin, 7 or 8 (less frequently 9, rarely 6) between
the pectoral and pelvic fin bases.

Fins. Dorsal with 24 (f-3o), 25 (f.2o) or 26 (f.2) rays, comprising 14 (f.2), 15 (f-4o)
or 16 (f.io) spinous and 8 (f.i), 9 (f-36) or 10 (f.15) branched rays. Anal with n (f.6),
12 (f-45) or 13 (f.i) rays, comprising 3 spines and 8-10 branched rays. Pectoral fin
length 23-8-33-0 (M = 27-0) per cent of standard length. Caudal truncate (the lower
posterior angle somewhat obliquely truncate in a few specimens) scaled on its basal
half to two-thirds (rarely). Pelvic fin with the first ray slightly produced and fila-
mentous, proportionately more so in adult males.

Teeth. In all specimens the majority of outer teeth are unicuspid but in some fishes
<I4O mm. S.L. a few bicuspids occur posterolaterally in the upper jaw. The teeth
are moderately to strongly curved, the lateral teeth often more so than the anterior
ones. There are 44-70 (M .= 63) teeth in the upper outer row, the number showing
some positive correlation with length in fishes <I40 mm. S.L.

FIG. 5. Haplochromis serranus ; . 72 x N.S.
(From Regan, Proc. zool. Soc. Lond.; the type of H. acutirostris Regan.)

All the rows of inner teeth (except in some fishes <I4O mm. S.L.) are composed of
unicuspids.

In the smaller fishes the innermost row (especially in the upper jaw) may be com-
posed of tricuspids or the entire inner series of both jaws may be of tricuspids. The
inner teeth are arranged in 2 or 3, rarely 4, rows in the upper jaw and 2, less frequently
i or 3 rows in the lower jaw.

Osteology. The neurocranium closely resembles that of H. victorianus (see p. 157)
although the dorsal profile is slightly more curved. It can be considered as a develop-
ment of the H . guiarti type.

Vertebrae. Twenty-eight or 29 (mode 29), comprising 12 (f.i) or 13 (f.6) precaudal
and 15 (f.i) or 16 (f.6) caudal elements.

Lower pharyngeal bone triangular, its dentigerous surface very slightly broader than
it is long. The teeth are moderately stout and cuspidate, and are arranged in 22-24
rows.

Coloration. The colours of live fishes are unknown.

Preserved material. Adult males. Ground colour light brown, greyish on chest and
belly ; branchiostegal membrane black or dark grey. In some specimens there is a

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 155

broad, dark, but faint midlateral band running from behind the operculum to the
base of the caudal fin or onto the fin itself, and sometimes an even fainter, interrupted
band running slightly below the base of the dorsal fin. The dark lachrymal stripe is
very prominent and runs obliquely backwards and downwards from the lower anterior
margin of the orbit to the angle of the lower jaw. Dorsal fin brown but with dark
streaks between the rays and, in some specimens, with black lappets. Caudal fin
darkly maculate. Anal grey-brown, darkest at the base of the spinous part ; ocelli
2-4 in number, greyish and often ill-defined. If there are more than three ocelli, they
are arranged in two rows of one and three. Pelvic fins black.

Adult females silvery brown, darker dorsally and apparently without longitudinal
bands. All fins are brownish. Juveniles of both sexes : ground colour silvery-brown
with a broad (sometimes faint) dark midlateral band from the operculum to the
basal part of the caudal fin ; lachrymal stripe, if visible, very faint. All fins are
greyish, often with traces of maculae on the soft part of the dorsal.

Ecology. Habitat. Most records of H. serranus are from sheltered bays and gulfs
where the bottom is of soft, organic mud and the depth of water less than 25 feet.
A few specimens came from exposed habitats and were caught over sand or shingle,
but again the water did not exceed 25 feet in depth. It should be noted that all the
latter localities are near muddy areas. No locality is more than half a mile from the
shore.

Food. The food of fishes in the size range 100-205 mm. S.L. is exclusively fish
and predominantly Haplochromis ; no other genus could be identified in the very
macerated gut contents of the thirty specimens examined. Some food was, however,
too finely divided to even hazard an identification.

Breeding. No information is available on the breeding habits of H. serranus.
Most individuals <I40 mm. S.L. are immature, but one female of 118 mm. S.L. shows
early stages of oogenesis. Males and females reach the same adult size.

Distribution. Lake Victoria.

Affinities. Haplochromis serranus is closely related to H. victorianus and H . nyanzae,
the three species apparently forming a species group amongst the larger predatory
Haplochromis of Lake Victoria. In turn, this group is related to H. spekii and the re-
lationship will be discussed in another paper. The differences distinguishing H. serranus
from H. nyanzae are outlined on p. 161 ; from H. victorianus, it differs in its more
oblique and longer lower jaw (47-4-60-0, M 54-3 per cent of head cf. 44-0-51-8,
M =47-1 per cent), shorter pectoral fin (23-8-33-0, M = 27-0 per cent of standard
length cf. 26-2-32-7, M = 30-4 per cent) and in having fewer teeth in the outer
row of the upper jaw (44-70, M = 63, cf. 67-86 M =74).

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1906.5.30.263 . Bunjako . Degen

(Type of H. acutirostris)

1962.3.2.27-36 . Pilkington Bay . E.A.F.R.O.

1962.3.2.37-51 . Ekunu Bay . ,,

ZOOL. 9, 4 g

I 5 6 P. H. GREENWOOD

Museum and Reg. No. Locality Collector

Uganda (continued]

B.M. (N.H.). 1962.3.2.52-3 . Bukafu Bay . E.A.F.R.O.

1 962 . 3 . 2 . 54-5 Jinj a pier

1962.3.2.56-7 . Karenia (near Jinja)

1962.3.2.61-3 . Pilkington Bay

1962.3.2.74-5 . Williams Bay

1962.3.2.58 . Ramafuta Isl.

Kenya

1962.3.2.26 . Rusinga Isl.

1962.3.2.60 . Naia Bay

Tanganyika

1962.3.2.59 . Ihogororo

Haplochromis victorianus (Pellegrin) 1904

Plate I

Paratilapia victoriana Pellegrin, 1904, Bull. Soc. zool. France, 29, 185 ; Idem, 1905, Mem. Soc.

zool. France, 17, 182, pi. 17, fig. 3.
Pelmatochromis spekii (part) Blgr., 1906, Ann. Mag. nat. Hist. (7), 17, 440 ; Idem, 1915, Cat.

Afr. Fish., 3, 416 (one of the types, B.M. (N.H.) Reg. No. 1906.5.30.300).

Regan (1922) tentatively included Paratilapia victoriana in the synonymy of
Haplochromis nubilus. I have examined the types of both species and can find nothing
to substantiate this arrangement ; indeed, the two species are only distantly related.
None of the specimens identified by Boulenger (1915) as P. victoriana can be referred
to Pellegrin's species.

Holotype. A fish 120 mm. S.L. (Paris Museum Reg. No. 04 x 148) from Kavirondo
Bay, Kenya.

Description. Based on twenty specimens 117-166 mm. S.L., including the holotype.

Depth of body 33-4-41-3 (M = 37-3) per cent of standard length, length of head
33-5-36-0 (M = 34-8) per cent. Dorsal profile of head usually straight (but sometimes
with a slight concavity due to more prominent premaxillary pedicels), sloping at

ca. 35-45-

Preorbital depth 17-9-20-5 (M = 19-2) per cent of head, ratio Eye/Preorbital 1-1-1-3
(mean 1-2), least interorbital width 21-5-24-5 (M = 22-6) per cent. Snout length
31-8-36-0. (M = 34-1) per cent of head, equal to its width ; diameter of eye 21-7-25-5
(M 23-6), depth of cheek 22-5-26-2 (M = 24-6) per cent.

Caudal peduncle 13-5-19-6 (M 17-1) per cent of standard length, 1-1-1-8 (mode
1-5) times as long as deep.

Lower jaw horizontal or slightly oblique, anteriorly equal to the upper or projecting
slightly, its length 44-0-51-8 (M = 47-1) per cent of head, 1-6-2-2 (modal range
1-8-2-0) as long as broad ; there is always a well-developed mental prominence at the
symphysis. The posterior tip of the maxilla reaches the vertical through the anterior
orbital margin, or somewhat more posteriorly, in most specimens but does not reach
the orbit in a few fishes.

Gill rakers short and moderately stout, 8 or 9 on the lower part of the first gill arch,
the lowermost 1-3 rakers reduced.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 157

Scales ctenoid, lateral line with 31 (.3), 32 (f.6), 33 (f.g) or 34 (f.2) scales, cheek
with 3 or 4 series. Six to 8 scales between the lateral line and the dorsal fin origin,
7 or 8 (rarely 6) between the pectoral and pelvic fin bases.

Fins. Dorsal with 24 (f.i), 25 (f.i6), or 26 (f.3) rays, comprising 15 (.4), 16 (f.14)
or 17 (f.2) spinous and 9 (f.i6) or 10 (.4) branched rays. Anal with n (f.i), 12 (f.i6)
or 13 (.3) rays comprising 3 spines and 8 (f.i), 9 (f-i6) or 10 (.3) branched rays.
Caudal fin truncate, scaled on its basal half (rarely two-thirds). Pectoral 26-2-32-7
(M = 30-4) per cent of standard length. First pelvic ray slightly produced and
filamentous in both sexes (an unusual feature) .

Teeth. In fishes less than 125 mm. S.L., the outer teeth in both jaws are mostly
bicuspids, but some may be weakly so ; fishes between 125 mm. and 130 mm. S.L.
show an admixture of unicuspids and weakly bicuspids, or the unicuspids may pre-
dominate ; all fishes >I35 mm. S.L. have only unicuspid teeth in this series. In all
specimens the teeth are relatively slender and slightly curved. There are 64-86
(M = 74) in the outer row of the upper jaw.

The inner teeth are arranged in two or three rows in each jaw ; most specimens
<I3O mm. S.L. have either a mixture of uni- and tricuspids in apparently equal
proportions, or one form may predominate. Fishes >I35 mm. S.L. have only unicus-
pids in the inner series.

Osteology. The neurocranium is very similar to that of H. serranus. It is clearly
distinct from that of H. bayoni (which seems to lead to the H. mento type) and is
somewhat more substantial than the neurocranium of H. guiarti (which is nearest
the generalized insectivore type of, say, H. brownae).

Vertebrae. Thirty (in the six specimens examined) comprising 13 precaudal and
17 caudal elements.

Lower pharyngeal bone triangular, the dentigerous surface noticeably broader than
long. The teeth are cuspidate and slender, and are arranged in 22-24 rows ; the teeth
of the two median rows are often enlarged relative to the others.

Coloration. Live fishes. In adult males the ground colour is a silvery-turquoise
dorsally, shading to silver on the flanks and belly. Dorsal fin dark neutral, with black
lappets and deep red maculae between the branched rays. Caudal fin dark neutral
but with blood-red posterior and ventral margins. Anal dark neutral basally, black
between the rays, and blood-red proximally ; the ocelli are orange-yellow and
numerous. The pelvic fins are black. Females, have a similar ground coloration but
lack the red areas on the caudal and anal fins, and the black area over the anal
spines ; the pelvics are neutral and there are no anal ocelli.

Preserved material. Adult males. Ground colour brownish-grey above, becoming
silver-grey on the flanks and belly. The branchiostegal membrane is sooty and there
is a faint blackening on the chest and belly. Dorsal fin hyaline but with black lappets
and dark maculae on the soft part. Caudal fin colourless but with numerous dark
spots and streaks between the rays. Anal fin black between the spines, otherwise
sooty-grey, the basal half being darker ; the greyish-white ocelli vary in number
from three to seven and are arranged in one or two rows, those of the lower row being
much smaller. Pelvic fins mottled black (appearing uniformly black when closed) .

Females are brownish silver above and silver to silvery-yellow on the flanks and

158 P. H. GREENWOOD

belly. All fins are colourless, the soft dorsal and the caudal sometimes weakly
maculate. In some specimens there is a faint but broad and interrupted midlateral
stripe from the posterior margin of the operculum to the base of the caudal fin.

Distribution. Lake Victoria.

Ecology. Habitat. The species has been recorded from only six different localities ;
in five the bottom is of thick organic mud and in the sixth, shingle (but this place is
near an area of mud) ; all localities are sheltered and the depth of water is between
20 and 40 feet.

Breeding. Very little information is available on the breeding habits of H . victori-
anus. With two exceptions (specimens 128 and 131 mm. S.L.) all the fishes examined
were adult and included several individuals smaller than the exceptional juveniles.
The largest fishes (166 mm. S.L.) are a male and a female.

Food. Fourteen of the twenty-seven fishes examined had ingested material in the
gut. Of these, twelve had fed on fishes and two on organic mud. The fish remains
are very finely divided and come from small fishes ca. 10-15 mm. S.L. (and this
irrespective of the size of the predator). In two cases the fishes are identifiable as
post-larval Haplochromis. The preponderance of small fishes in the prey is interesting
because even in such a small sample most other predatory Haplochromis would have
yielded remains of much larger prey fishes. There is, of course, usually some correla-
tion between prey size and predator size, but in a sample covering a comparable
size-range the prey fishes would be from 30-60 mm. S.L. Perhaps H. victorianus has
specialized in feeding on post-larval fishes ?

The occurrence of mud in two specimens is inexplicable, particularly since there are
no indications that the fishes had been feeding on insect larvae, a common alternative
food in many otherwise piscivorous species.

Affinities. Haplochromis victorianus is very closely allied to H. serranus and H.
nyanzae, see p. 161. The three species form a fairly well-defined group of deep-bodied
and anatomically rather generalized species amongst the piscivorous predators of
Lake Victoria.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1906.5.30.300 . Entebbe . Degen
(Paratype P. speki)

1962.3.2.478 . Ramafuta Isl. . E.A.F.R.O.

(Buvuma channel)

,, 1962.3.2.479-82 . Karenia, near Jinja

1962.3.2.483-87 . Pilkington Bay

,, 1962.3.2.488-494 . Ekunu Bay

Kenya

Paris Museum 04 x 148 (Holotype) . Kavirondo Bay

B.M. (N.H.). 1962.3.2.477 . Naia Bay . ,,

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 159

Haplochromis nyanzae sp. nov.

Text-fig. 6

Pelmatochromis spekii (part) : Blgr., 1915, Cat. Afr. Fish., (one specimen B.M. (N.H.) Reg. No.

1909.5.4.17).
H. serranus (part) : Regan, 1922 ; Proc. zool. Soc. Land. 174 (specimen as above).

Holotype. A specimen 154-0 mm. S.L. (B.M. (N.H.) Reg. No. 1962.3.2.495) from
Jinja.

Description. Based on thirteen specimens (including the holotype) 126-171 mm.
S.L.

Depth of body 33-8-38-6 (M = 36-3) per cent of standard length, length of head
33-0-367 (M = 35-4) per cent. Dorsal head profile straight or moderately concave,
sloping at an angle of 30-4O.

FIG. 6. Haplochromis nyanzae ; .75 X N.S. (Drawn by La vinia Beard.)

Preorbital depth 17-0-20-2 (M 18-9) per cent of head, least interorbital width
20-0-24-5 (M = 22-2) per cent. Snout length 33-4-35-8 (M = 34-5) per cent of head,
i-o-i-i times as long as broad. Eye diameter 19-1-24-0 (M = 22-1) per cent, ratio
Eye/Preorbital 1-1-1-3 (M = 1-2), depth of cheek 24-4-27-6 (M = 25-9) per cent.

Caudal peduncle length 14-0-17-6 (M = 15-9) per cent of standard length, 1-1-1-4
(mode 1-4) times as long as deep.

Lower jaw moderately oblique, sloping at an angle of 25-30, anteriorly equal to
the upper jaw or projecting slightly, its length 45-0-51-6 (M = 48-0) per cent of head,
1-5-2-0 (Mode 1-7) times as long as broad. The anterior outline of the dentary is
smoothly curved and lacks a strong mental projection. The posterior tip of the

i6o P. H. GREENWOOD

maxilla usually reaches the vertical through the anterior orbital margin, but in some
fishes it extends beyond this point or does not reach it.

Gill rakers short and stout (rarely short and slender), 8 or 9 (rarely 7 or 10) on the
lower part of the first gill arch, the lowermost one or two rakers reduced.

Scales ctenoid ; lateral line with 31 (f.i), 32 (f.8) or 33 (.4) scales, cheek with 4 or 5
(rarely 3) rows. Seven or 8 (rarely 6 or 9) scales between the dorsal fin origin and the
lateral line, and between the pectoral and pelvic fin bases.

Fins. Dorsal with 24 (f.5) or 25 (f.8) rays, comprising 15 (f.2) or 16 (f.n) spinous
and 8 (f.4), 9 (f.8) or 10 (f.i) branched rays. Anal with 12 rays (3 spinous, 9 branched)
in all except one specimen which has only 2 spines and 9 rays. Pectoral fin 22-8-28-2
(M = 24-9) per cent of standard length. First pelvic ray very slightly produced in
males and less so in females. Caudal truncate, scaled on its basal two-thirds.

Teeth. The outer row in both jaws is composed of unicuspid, moderately stout
teeth, strongly incurved in most specimens, but less strongly in a few others. The
number of teeth in this row shows a very slight positive correlation with standard
length ; for the whole sample there are 50-76 (M = 60) upper teeth.

The inner teeth are uni- and tricuspid in fishes <I35 mm.. S.L. and unicuspid in
larger fishes. There are 2-4 (usually 3) rows in the upper jaw and 2 or 3 in the lower.

Osteology. With so few specimens available, I have not been able to prepare any
skeletal material. However, on comparing radiographs of this species with others of
H. victorianus I can find no great differences in neurocranial form or general syn-
cranial arrangement. If anything, the supraoccipital crest in H. nyanzae is relatively
lower and the slope of the dorsal skull profile is a little less steep.

Vertebrae. 29 (f.2), 30 (f.4) or 31 (f.i), comprising 13 (f.5) or 14 (f.2) abdominal
and 16 (f.3) or 17 (f.4) caudal elements.

Lower pharyngeal bone triangular, its dentigerous surface noticeably broader than
long. The teeth are slender, cuspidate and arranged in 22-24 rows. Except in the
most posterior transverse row or two the teeth of the median rows are not enlarged.

Coloration. Unknown in live fishes. Preserved material. Adult males have a dark
brown ground colour with an overlying greyish tinge ; the belly, ventro-lateral
aspects of the flanks, the ventral part of the preoperculum and the branchiostegal
membrane are sooty. There is a distinct but narrow black lachrymal stripe, a faint
black midlateral band and a fainter dorsolateral band following the upper lateral
line. The lateral bands are crossed by five or six faint transverse bars. The dorsal
fin is dark brown with darker spots and streaks between the soft rays. Caudal fin
dark brown, anal brown with a faint and narrow darker flush along its base. Pelvic
fins mottled black. In other males (whose sexual state could not be determined)
the general ground coloration is much darker, the lower jaw, snout and ventral
aspects of the cheek, preoperculum and operculum are black, the dorsal fin is darker
(almost black) and the spots on the soft part are more obvious. The caudal fin too is
darker, but the anal is similar. The anal ocelli in all males are difficult to distinguish.

Females are brown above, shading to silver-bronze on the flanks and belly. All
fins are yellowish-brown, the dorsal with a grey overtone which is concentrated
basally and outlined in black on the soft dorsal so that there appears to be a ventro-
caudally curved dark stripe passing across it from the tip of the last spine to the middle

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 161

of the last branched ray. Usually there is a dark midlateral band on the body,
ending as a distinct blotch on the caudal fin base ; traces of three or four transverse
bands may be visible on the flanks.

Immature and quiescent males are indistinguishable from females, although quies-
cent males often show a distinct darkening of the pelvic fins and faint indications of
ocelli on the anal fin.

Ecology. Habitat. The available specimens of H. nyanzae are from five localities,
of which four lie within the Napoleon Gulf and the fifth (a small island) in the nearby
Buvuma Channel. In all localities the bottom is hard (shingle or rock) and the water
less than twenty feet deep. The Napoleon Gulf stations are relatively sheltered but
the island is exposed.

Food. Eight of the thirteen fishes examined contained food in the stomach or
intestines. All eight had fed on fishes (identifiable in each case as Haplochromis) but
one had a few insect remains in the intestine. Judging from the size of the scales and
vertebrae in the gut contents, the prey fishes must have been between 30 and 60 mm.
S.L.

Breeding. No data are available. All the specimens are adult.

Affinities. Haplochromis nyanzae is closely related to both H. victorianus and
H. serranus. It differs from both species in the nature of the coloration in preserved
males. From H. victorianus it is distinguished by its more oblique lower jaw (which
also lacks a pronounced mental bump), shorter pectoral fin (22-8-28-2, M = 24-9%
of S.L., cf. 26-2-32-7, M = 30-4%) and fewer teeth in the upper jaw (50-76, M = 60 ;
cf. 64-86, M =74). From H. serranus it is most readily differentiated by its rounded
dentary (i.e. no pronounced mental bump), shorter lower jaw (45-0-51-6 M =48-0,
cf. 47-4-60-0, M = 54-3% of head) and shorter pectoral fin (22-8-28-2, M = 24-9,
cf. 23-8-33-0 M = 27-0% of standard length).

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda
B.M. (N.H.). 1962.3.2.495 . Jinjabay . E.A.F.R.O.

(Holotype)

,, 1962.3.2.500 . J in j a pier

1962.3.2.501-3 . Karenia (near Jinja)

1962.3.2.504-7 . Jinja (below golf course)

,, 1962.3.2.496-8 . Ramafuta Isl.

(Buvuma channel)
1909.5.4.17 . Sesse Isls. . Bayon

Haplochromis bartoni sp. nov.
Text-fig. 7

Note. The trivial name is given because the species resembles H. worthingtoni, a
species named in honour of Dr. E. Barton Worthington.

Holotype. A specimen 145-0 mm. S.L. (B.M. [N.H.] Reg. No. 1962.3.2.277) from
Ekunu Bay.

1 62

P. H. GREENWOOD

Description based on thirty-five specimens (including the hole-type) 135-195 mm.
S.L.

Depth of body 31-4-37-9 (M = 34-0) per cent of standard length, length of head
36-2-39-7 (M = 37-5) per cent. Dorsal head profile straight or very slightly curved,
sloping at about 40 ; premaxillary pedicels moderately prominent, sometimes
giving the profile a noticeable concavity.

Preorbital depth 17-0-22-4 (M = 20-0) per cent of head, least interorbital width
17-0-21-0 (M = 18-6) per cent. Snout 33-4-40-3 (M = 36-0) per cent of head, and
1-1-1-2 times as long as broad; diameter of eye 20-3-24-1 (M = 22-5), depth of
cheek 23-4-30-2 (M = 27-0) per cent.

Caudal peduncle 13-3-17-2 (M = 15-0) per cent of standard length, 1-2-1-5 (mode
1-3) times as long as deep.

FIG. 7. Haplochromis bartoni ; .75 x N.S. (Drawn by Lavinia Beard.)

Mouth slightly oblique ; the lower jaw sloping at ca. io-i5 and projecting beyond
the upper, usually with a slight mental protuberance. Length of lower jaw 50-8-57-0
(M = 52-5) per cent of head, 1-9-3-0 tines as long as broad. The medial dentigerous
part of the premaxilla is not expanded anteroposteriorly. Lips slightly thickened.
Posterior tip of the maxilla reaching the vertical through the anterior orbital margin
in most fishes (56% of sample) and to a point slightly beyond or slightly anterior to
the vertical in the remainder.

Gill rakers short and stout, 8 or 9 (rarely 7) on the lower part of the first arch, the
lowermost 1-3 rakers reduced.

Scales ctenoid, lateral line with 31 (f.6), 32 (f.8), 33 (.19) or 34 (f.2) scales, cheek
with 4 (less frequently 3, rarely 5) rows. Six or 7 scales (rarely 7^) between the lateral
line and the dorsal origin, 7 or 8 (rarely 6) between the pectoral and pelvic fin bases.

REVISION OF THE LAKE VICTORIA HA PLOCHROMIS SPECIES 163

Fins. Dorsal with 23 (f.2), 24 (f.6), 25 (f.i8) or 26 (f.8) rays, comprising 14 (.3),
15 (f.io) or 16 (f.2i) spines and 8 (f.2), 9 (1.19), 10 (f.n) or n (f.2) branched rays.
Anal fin with n (1.4), 12 (f-3o) or 13 (f.i) rays comprising 3 spines and 8-10 branched
rays. Pectoral fin 23-3-27-0 (M = 25-0) per cent of standard length. First pelvic ray
slightly elongate in adult males. Caudal with a slightly oblique distal margin in
many specimens, but vertically truncate in others ; the obliquely truncate type is
less oblique than in H. plagiostoma (see p. 200). The caudal is scaled on its proximal
two- thirds to four-fifths.

Teeth. The outer row in both jaws is composed of slender, curved and unicuspid
teeth, of which there are 50-80 (M = 62) in the upper jaw. The inner rows in most
fishes >i5o mm. S.L. are composed of unicuspid teeth but in most smaller fishes
there is an admixture of uni- and weakly tricuspids in both jaws, or either type of
tooth may predominate. In the latter case it is usually the innermost series of the
upper jaw which is predominantly tri- or weakly tricuspid, the lower jaw containing
mostly unicuspids. There are 2 or 3 (less frequently 4) rows in the upper jaw and i
or 2 (very rarely 3) in the lower jaw; all inner teeth are implanted obliquely, so that
in many specimens the teeth lie horizontally.

Osteology. The neurocranium is roughly intermediate between the H . bayoni-type
(see p. 151) and the H. mento-type (see p. 176). The medial toothed part of the pre-
maxilla is not markedly expanded, but it does show some development in that
direction. This bone is, however, nearer the generalized H. guiarti-type than is the
premaxilla of H. bayoni.

Vertebrae. Twenty-nine or 30 (mode), comprising 13 abdominal and 16 (f.2) or 17
(f .4) caudal elements.

Lower pharyngeal bone triangular, its dentigerous surface broader than long or
equilateral ; only rarely is it slightly longer than broad. The teeth are slender (but
become coarser in larger fishes), cuspidate and arranged in 20-22 somewhat irregular
rows.

Coloration. Live colours are unknown.

Preserved females are brown above the upper lateral line and on the dorsal surface
of the head. A dark midlateral stripe runs from behind the operculum to the base of
the caudal fin and, in some specimens, can also be seen on the basal half of the caudal
fin membrane. Less well-defined, and absent in some fishes, is a dark band from the
snout, through the eye and onto the operculum where it becomes continuous with the
midlateral band. All fins colourless, the soft dorsal weakly maculate. Adult males
are uniformly greyish-brown, except for the branchiostegal membrane, chest and
belly which are sooty. Dorsal fin grey, with black lappets and dark, often coalesced
maculae on the soft part. Anal fin black in the region of the spines and dark along
its proximal half ; distal part colourless. Anal ocelli whitish-grey, five or six in
number and arranged in two rows. Caudal fin brown with dark streaks between the
rays ; pelvic fins black.

Distribution. Lake Victoria.

Ecology. Habitat. The species has been recorded from several different habitats,
whose common features are : a depth of water less than twenty feet and the nearness
of dense plant stands. The habitats include a sheltered gulf with a hard substrate,

164 P. H. GREENWOOD

sheltered bays with deep mud bottoms, sheltered beaches with hard and soft sub-
strates, an exposed sandy beach and a rock shelf extending from a partially exposed
island. Haplochromis bartoni seems to be more abundant in the sheltered habitats
and, from its relative scarcity in seine-net catches, it seems to be most abundant
some distance offshore (between 200-300 yards).

Food. Nineteen of the thirty-one specimens examined contained food in the stomach
or intestines ; all had fed exclusively on fishes and the remains could be identified as
follows : Cichlidae (undetermined) (f-5), Haplochromis spp. (f.n), Cyprinidae
(undertermined) (f.i) ; unidentifiable fish remains (f.2).

Breeding. Little information is available on this species. Sexual maturity is
reached at about the same size (145 mm. S.L.) in both sexes and there is no marked
dimorphism in the maximum size attained.

Affinities. Some specimens of H. bartoni bear a superficial resemblance to the
holotype and only specimen of H. worthingtoni , a Lake Kyoga species. However, the
likeness is purely superficial and the two species differ in several morphological
characters. Within the Victoria species-flock, H . bartoni is somewhat isolated, not
by any outstanding morphological characters but by the sum of several small
characters. The neurocranium is nearest that of the H. mento complex but it still
retains some of the more generalized characters. The premaxilla is of the specialized
" beaked " type found in H. macrognathus. Another superficial resemblance is to the
H. serranus-H. victorianus complex, but here the likeness is confined to general
facies and is not borne out by any deeper-lying details. Haplochromis bartoni could
represent one of the stages in the evolution of a. H. mento-type from a H . guiarti-like
stem, but a stage nearer the " mento " than the " guiarti " level of organization.
Equally, it could link the H. serranus group with the H. mento complex, but again it
would be nearer the " mento " than the " serranus " condition.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1962.3.2.277 . Ekunu Bay . E.A.F.R.O.
(Holotype)

1962.3.2.278-82 . Ekunu Bay

1962.3.2.283-6 . Pilkington Bay

1962.3.2.287-92 . Ramafuta Isl. (Buvuma ch.)

1962.3.2.293-8 . Fielding Bay

1962.3.2.299-300 . Karenia (near Jinja)

1962.3.2.301 . Old Bukakata

1962.3.2.302 . Beach near Nasu point
1962.3.2.303-6 . Jinja pier
1962.3.2.307-13 . Jinja (below golf course)

Haplochromis estor Regan 1929

Text-figs. 8 and 25
Haplochromis estor Regan, 1929, Ann. Mag. nat. Hist., (10), 3, 391.

Holotype. A specimen 153 mm. S.L. from an unknown locality in Lake Victoria,
collected by M. Graham. B.M. (N.H.) Reg. No. 1959.7.2.1.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 165

Description. Based on the holotype and eight other specimens, 141-170 mm. S.L.

Depth of body 29-6-32-4 (M = 30-3) per cent of standard length, length of head
37-2-38-5 (M = 37-8) per cent. Dorsal head profile sloping fairly steeply (ca. 40),
the premaxillary pedicels prominent.

Preorbital depth 19-3-20-6 (M = 19-8) per cent of head, least interorbital width
20-4-23-7 (M = 21-6) per cent. Snout 1-2-1-3 times longer than broad, its length
24.5-37-0 (M = 36-2) per cent of head ; diameter of eye 20-4-23-5 (M = 21-7) per
cent, depth of cheek 22-8-25-4 (M = 24-5) per cent.

Caudal peduncle 14-3-16-1 (M = 15-8) per cent of standard length, 1-3-1-6 times
as long as deep.

Mouth slightly oblique (i5-2o), lower jaw always projecting, sometimes strongly
so ; its length 54-2-57-5 (M = 55-5) per cent of head, 2-1-2-5 times as long as broad.
Lips slightly thickened, the premaxilla with the medial dentigerous surface antero-

FIG. 8. Haplochromis estor ; holotype. (Drawn by Miss M. Fasken.)

posteriorly expanded. Posterior tip of the maxilla usually reaching the vertical
through the anterior orbital margin, occasionally extending somewhat behind this
line.

Gill rakers short and stout, 8 or 9 on the lower part of the first gill arch, the lower-
most one to four rakers reduced.

Scales ctenoid, lateral line with 32 (.3), 33 (f-5) or 34 (f.i) scales, cheek with 4
(rarely 3) series of scales ; 6 or 7 (rarely 5| or 8) between the lateral line and the dorsal
fin origin, 7 or 8 (rarely 6) between the pectoral and pelvic fin bases.

Fins. Dorsal with 24 (f.i), 25 (f-5) or 26 (f.3) rays, comprising 15 (f.i) or 16 (f.8)
spines and 9 (f .6) or 10 (f .3) branched rays. Anal with u or 12 (mode) rays, comprising
3 spines and 8 or 9 branched rays. Pectoral fin 21-8-25-3 (M = 23-6) per cent of
standard length. First pelvic ray produced in males. Caudal truncate or very slightly
emarginate, scaled on its basal third or half.

166 P. H. GREENWOOD

Teeth. The outer row in both jaws is composed of moderately stout, strongly curved
unicuspid teeth, of which there are 52-70 (M = 60) in the upper jaw. The inner teeth
are tricuspid in the smallest specimen examined (141 mm. S.L.) but are unicuspid in
all others. These teeth are implanted obliquely and arranged in 3 or 4 (less frequently
2) rows in the upper jaw and in 2 or an irregular single row in the lower jaw.

Osteology. The neurocranium of H. estor is of the H. mento type and, because of its
more ventrally curved vomer, closely approaches that of H. dentex (see p. 168).

Vertebrae : 29-30 comprising 12 or 13 precaudal and 16 or 17 caudal elements in
the eight specimens examined.

Lower pharyngeal bone triangular, the breadth of the dentigerous surface slightly
greater than its length or, less frequently, equal to its length. The teeth are relatively
slender and bicuspid, and are arranged in 18-22 rows.

Coloration unknown in life. Preserved males (sexually active). Ground colour dark
chocolate-brown above becoming dusky over silver below, especially on the chest
and belly ; faint traces of a dark midlateral stripe from the opercular margin to the
caudal base are sometimes visible. Lachrymal stripe broad. Lips and lower jaw
almost black, the branchiostegal membrane black. Dorsal fin dark grey, the soft
part maculate. Caudal fin dark grey and densely maculate, the spots on the proximal
half often coalesced so that the fin is dark proximally and lighter distally. Anal
greyish with an ill-defined darker band along the distal margin, and a black but
narrow band along the basal part ; 2-5 grey-white ocelli are present, arranged in
two rows if there are more than four ocelli. Pelvics black. Quiescent males have a
similar coloration but are much lighter ; consequently the midlateral stripe is more
obvious. The dorsal fin lacks spots.

Females are brown, shading to silver on the belly and chest ; a faint midlateral
stripe, and in one specimen three very faint and incomplete vertical cross bars are
visible. Dorsal fin neutral, the soft part maculate. Caudal dark and maculate. Anal
and pelvic fins yellowish.

Distribution. Lake Victoria.

Ecology. Habitat. Four of the localities from which H. estor has been obtained are
sheltered bays and gulfs where the substrate is mud and the water between 10 and
20 feet deep. No information is available for the two other localities. It can certainly
be said that the species (at least when adult) does not occur commonly, if at all, in
exposed inshore areas of the lake.

Food. Five of the eight specimens examined had food in the stomach and intestines.
Each had fed exclusively on fishes (identifiable as Haplochromis in four cases and
merely as " fish " in the fifth).

Breeding. No information is available. All the specimens are adult ; the two
smallest are females and the rest males.

Affinities. The affinities of H. estor are discussed in connection with H. dentex
(p. 169) Haplochromis estor is more advanced towards the extreme H. macrognathus
type, but is nevertheless more closely allied to H. dentex than to H. macrognathus or
even H. mento. Regan (1929) compared H. estor with H. Pellegrini. There is a super-
ficial resemblance and probably an overall phyletic relationship between the species,
but it is not, in my opinion, as close as the relationships suggested above.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 167

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1962.3.2.231 . Karenia (near Jinja) . E.A.F.R.O.

1962.3.2.232-4 . Jinja pier

1962.3.2.235-6 . Bugungu (opp. Jinja)

1962.3.2.237 . Entebbe

1962.3.2.238 . Ekunu Bay

1962.3.2.239 . Pilkington Bay
Lake Victoria, Locality Unknown

1959.7.2.1 . . M.Graham

(Holotype)

Haplochrotnis dentex Regan 1922

Text-figs. 9 and 25

Paratilapia longirostris (part) : Blgr., 1915, Cat. Afr. Fish., 3, 332.
Haplochrotnis dentex Regan, 1922, Proc. zool. Soc. Lond., 182, pi. 3, fig. i.

Holotype. A specimen 127-0 mm. S.L. from the Sesse Islands, B.M. (N.H.) Reg. No.
1909.5.4.1.

Description based on fifteen specimens (gi-o-isg-o mm. S.L.) including the holo-
type.

Depth of body 24-6-29-5 (M = 26-7) per cent of standard length, length of head
33-3-36-2 (M = 34-9) per cent. Dorsal head profile gently curved, the premaxillary
pedicels prominent.

Preorbital depth 18-7-24-5 (M = 21-7) per cent of head, least interorbital width
20-0-24-5 (M = 22-5) per cent. Snout longer (1-3-1-5 times) than broad, its length
36-0-41-5 (M 38-8) per cent of head ; diameter of eye 17-3-25-7 (M = 20-4) per
cent, depth of cheek 22-1-27-5 (M = 24-6) per cent.

Caudal peduncle length 17-0-19-8 (M = 18-2) per cent of standard length, 1-6-2-0
times longer than deep.

Angle of mouth variable, from almost horizontal to slightly oblique (i5-20).
Lower jaw projecting slightly, its length 43-8-49-0 (M = 46-0) per cent of head and
2-0-2-5 times its breadth. Lips moderately thickened, the medial dentigerous surface
of the premaxilla expanded anteroposteriorly. Posterior tip of the maxilla not
reaching the vertical through the anterior orbital margin but always behind a
vertical through the posterior tip of the premaxillary pedicels.

Gill rakers moderately slender ; 9 or 10 (mode) on the lower part of the first gill
arch, the lowermost one to three rakers reduced.

Scales ctenoid ; lateral line with 33 (f.io) or 34 (f.5) scales (in one specimen there
are no lateral line pores on one side) . Cheek with 3 or 4 (mode) rows. Five to 7 scales
between the lateral line and the dorsal fin origin, 7 or 8 between the pectoral and
pelvic fin bases.

Fins. Dorsal with 24 (f.i), 25 (f.12) or 26 (f.2) rays, comprising 15 (f.8) or 16 (f.7)
spines and 9 (f.6) or 10 (f.g) branched rays. Anal fin with 11-13 rays, comprising
3 spines and 8-10 (mode 9) branched rays. Pectoral fin 20-7-25-3 (M = 24-6) per cent

1 68

P. H. GREENWOOD

of standard length. Caudal truncate or weakly emarginate, scaled on its basal two-
thirds. First pelvic ray produced in males.

Teeth. The outer row in both jaws is composed of large, well-spaced, unicuspid,
moderately slender and curved teeth. There are 32-48 (M = 36) in the outer row of
the upper jaw.

In all except the two smallest fishes (91 and 100 mm. S.L.) the inner teeth are uni-
cuspid. The two small fishes have a mixture of unicuspids, tricuspids and weakly
tricuspids, in which the unicuspids predominate. The outermost row in both jaws
is composed of teeth only a little smaller than those of the outer row. There are two
rows of inner teeth (sometimes irregularly arranged and giving the impression of
three rows) in the upper jaw and a single (rarely) double row in the lower jaw.

Osteology. The neurocranium of H . dentex clearly belongs to the H. estor-mento-
macrognathus group. The preotic part is long (65-8% of basal length), the skull is
narrow and shallow (neurocranial height 3-4 times in basal length) and the supra-
occipital crest is low. It differs, however, from other members of the group in its

FIG. 9. Haplochromis dentex, holotype, -8 x N.S. (From Regan, Proc. zool. Soc. Land.)

sharply decurved ethmoid- vomer region which slopes at about 40 (the neuro-
cranial roof slopes at ca. 25). In general, the neurocranium of H. dentex is nearest that
of H. estor, differing mainly in having a sharply decurved vomer.

Vertebrae. 30-32 (mode 31 in the ten specimens examined), comprising 13 (f.3) or
14 (f.7) precaudal and 16 (f.i), 17 (f.6) or 18 (f.3) caudal elements.

Lower pharyngeal bone triangular, its dentigerous surface as long as broad or
slightly broader than long. The pharyngeal teeth are slender, fine and bicuspid
except for the coarser and less obviously cuspidate teeth in the two median and last
transverse rows. There are 18-20 rows of teeth.

Coloration. The colours of live fishes are unknown. In preserved material there is
little difference in the coloration of males and females. The ground colour is dark
grey above becoming silvery-grey below (darker in males). In some females there is
a faint and narrow midlateral longitudinal stripe from the hind margin of the oper-
culum to the caudal origin. Dorsal and caudal fins are greyish, the former with black
lappets and the latter sometimes maculate. The anal and pelvic fins are hyaline in
females, whereas in males the anal is greyish with a narrow black basal line running

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 169

above the spines. In males there are two dead-white anal ocelli, and the pelvics are
black.

Distribution. Lake Victoria.

Ecology. Habitat. Since only fifteen specimens are known, it is impossible to
generalize on habitat preferences particularly since about half the specimens are from
sheltered bays with mud substrates, and the others from exposed, sandy beaches,
open off-shore waters with rock and shingle bottoms, and a fairly exposed gulf, also
with a hard substate. The depth range extends to at least 25 feet, but the species
cannot be considered common in any of the habitats investigated.

Food. Only three of the fifteen specimens available had ingested material in the
gut. Two specimens had fed on small Haplochromis ; in the intestine of the third was
a small quantity of fragmentary plant tissue.

Breeding. The breeding habits of H. dentex are unknown. The two smallest fishes
(91 and 100 mm. S.L.) are both immature and the next largest (128 mm.) may also
be a juvenile. All the other specimens (142-159 mm.) are adult.

Affinities. Superficially, H. dentex looks intermediate between the H. guiarti-type
and the more specialized H. mento-estor types. But, closer study shows that it has
greater affinity with the latter group, particularly with regard to its neurocranial
form and the expanded medial dentigerous part of the premaxilla. Haplochromis
dentex is perhaps most closely related to H. estor, although the latter has progressed
further along the H. mento-H. macrognathus path of specialization. The supposed
relationship between H. dentex and H. estor is based both on points of overall similarity
and on likeness in the neurocrania of the two species. Haplochromis dentex is, however,
easily distinguished by its fewer and larger teeth (mean number of teeth in the outer
row of the upper jaw 37 cf. 60) and its shorter lower jaw (43-8-49-0, M = 46-0 per cent
of head cf. 48-0-57-5, M = 54-9 per cent). The ancestry of H. dentex is obscure ;
possibly it was derived from a H. pellegrini-like stem.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1909.5.4.1 . Sesse Ms. . Bayon
(Holotype)

1962.3.2.129 . Sesse Isls. . E.A.F.R.O.

1962.3.2.130-1 . Macdonald Bay

1962.3.2.132 . Grant Bay

1962.3.2.133-4 Jinja

1962.3.2.135-6 . Buka Bay

1962.3.2.137 . Kagera port

1962.3.2.138 . Ramafuta Isl.

(Buvuma channel)

1962.3.2.139 . Ekunu Bay

1962.3.2.140 . Thruston Bay

,; 1962.3.2.142 . Old Bukakata Bay

Kenya

1962.3.2.141 . Kamiriga, Kavirondo Gulf

170 P. H. GREENWOOD

Haplochromis artaxerxes sp. nov.

This peculiar species is represented by a single specimen. Its diagnostic characters
are such, however, that I have little hesitation in basing the description on one fish.
The specific name is derived from Artaxerxes, King of Persia, also known as Longimanus
and alludes to the extremely long pectoral fins of this species.

Holotype a male 147-0 mm. standard length, from the Napoleon Gulf near Jinja ;
B.M. (N.H) Reg. No. 1962.3.2.508.

Description. Depth of body 27-9 per cent of standard length, length of head 32-7
per cent. Dorsal head profile gently curved, the premaxillary pedicels not making a
prominent projection.

Preorbital depth 17-7 per cent of head, least interorbital width 20-8 per cent. Snout
1-25 times as long as broad, its length 33-3 per cent of head ; eye diameter 22-9 per
cent, depth of cheek 22-9 per cent.

Caudal peduncle 19-0 per cent of standard length , 1-7 times as long as deep.

Lower jaw rather flat and closing within the upper, the anterior tip projecting.
Length of lower jaw 48-0 per cent of head, 2-3 times as long as broad. Median denti-
gerous area of the premaxilla not expanded anteroposteriorly. Posterior tip of the
maxilla reaching the vertical through the anterior orbital margin. Mouth very
slightly oblique, sloping at about 10.

Gill rakers short and stout, 9 on the lower part of the first gill arch, the three
lowermost rakers reduced.

Scales ctenoid ; lateral line with 34 scales, cheek with 4 series. Seven scales between
the lateral line and the dorsal origin, 9 between the pectoral and pelvic fin bases.

Fins. The pectoral fins of this species provide the most readily diagnostic character
since they are longer (34-7% of standard length) than in any other Lake Victoria
species. Only the third and fourth pectoral rays are produced so that the shape of the
fin is also characteristic.

Dorsal with 15 spinous and 10 branched rays, anal with 3 spines and 9 branched
rays. The distal margin of the caudal is damaged so its outline cannot be determined.
The first branched pelvic ray is elongate and filamentous, in fact, intermediate
between the extreme condition found in H. bayoni (see p. 151) and that of other
species.

Teeth. Most of the outer teeth in both jaws are missing. The few remaining teeth
are unicuspid, slender and very strongly curved. The inner teeth are unicuspid and
arranged in two series in the upper jaw and a single, irregular row in the lower jaw.

Osteology. Neurocranial shape cannot be determined from a radiograph. There are
31 vertebrae (13 precaudal and 18 caudal). The lower pharyngeal bone is slender and
triangular, the dentigerous surface being broader than long. The pharyngeal teeth
are slender, fine and cuspidate, and are arranged in about twenty rows.

Coloration. Unknown in life. The preserved adult male is dark brown above, rapidly
shading to a light brass colour with an overall greyish tinge. Lips, lower jaw, hori-
zontal limb of the preoperculum and the posterior opercular margin black. There is a
very broad (ca. half diameter of eye) lachrymal band and two, narrow parallel bands
across the snout. Dorsal fin yellowish-brown, the soft part faintly marbled. Caudal

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 171

dark, anal with a narrow, yellow basal band but otherwise sooty, particularly in the
area of the spines, except for a narrow and yellow distal margin.

Ecology. The only ecological information is that the fish was caught in ca. 10 feet
of water over a mud bottom and near a fringing stand of swamp grass. The location
is near Jinja and is in the relatively sheltered Napoleon Gulf.

Diagnosis and affinities. The species is characterized by the following combination
of characters : pectorals long (35% of standard length and longer than the head),
the third and fourth rays greatly produced ; lower jaw flat and closing within the
upper jaw.

Because of these particular characters, it is difficult to suggest possible relationships
for the species. The elongate pectoral is, of course, a character which could easily
and suddenly develop from the pectoral of any Haplochromis. The lower jaw, on
the other hand, requires a more fundamental change and one rarely encountered in
the Victoria species flock. It is otherwise found only in some of the larval fish-eating
species (Greenwood, 1959). Certainly none of these could represent the ancestral or
descendant condition of a species like H. artaxerxes. Without any knowledge of the
skeleton in this species (and especially without more information on its dentition)
little can be guessed about its phyletic position. Superficially, H. artaxerxes does
resemble H. estor but it is immediately distinguished by the diagnostic characters
listed above, and by its shorter head and lower jaw.

Haplochromis longirostris (Hilgen.) 1888
Text-figs. 10 and n

Paratilapia longirostris Hilgendorf, 1888, Sitzb. ges. naturf.-Fr. Berlin, 77 ; Boulenger (part),

1915, Cat. Afr. Fish., 3, 332.
Haplochromis longirostris (part) ; Regan, 1922, Proc. zool. Soc. Lond., 187, PI. 4, fig. 2 (two of the

three specimens described loc. cit.}.

Haplochromis gracilicauda Regan, 1922, op. cit., 188, PI. 4, fig. 2.
Haplochromis tenuis Borodin, 1931. Proc. New Eng. zool. Club., 12, 50.

The characters which Regan used to separate H. gracilicauda from H. longirostris
(snout i^ to if diameter of eye cf. snout if to twice diameter of eye) are really func-
tions of growth, his species H. gracilicauda representing juvenile H. longirostris.
Through the courtesy of Dr. K. Deckert of the Berlin Museum, I have been able to
examine the holotype of Paratilapia longirostris (Z.M. Berlin Reg. No. 12744) an< i
this confirms the identity of Regan's H. longirostris material. Two of these fishes
(BM. [N.H.] Reg. No's. 1911.3.3.13 and 1906.5.30.516) are retained in the species
but the third (B.M. [N.H.] 1906.5.30.274) is referred to H. argenteus.

Description based on twenty-nine specimens (including the holotype, and the syn-
types of H. gracilicauda} 85-145 mm. S.L.

Depth of body 24-6-30-4 (M = 27-2) per cent of standard length, length of head
29-2-36-2 (M 33-0) per cent. Dorsal head profile slightly concave or, less commonly,
straight ; sloping at an angle of ca. 30. Premaxillary pedicels moderately prominent
in large individuals.

ZOOL. 9, 4 10

172 P. H. GREENWOOD

Preorbital depth 18-2-22-5 (M = 20-8) per cent of head, least interorbital width
17-1-24-0 (M = 21-2) per cent. Snout length 32-0-38-6 (M = 36-0) per cent of head,
1-2-1-5 times as long as broad, narrowest in large fishes. Diameter of eye 18-8-24-3
(M = 21-7) per cent, depth of cheek 21-8-28-4 (M = 24-3) per cent.

Caudal peduncle 17-2-22-2 (M = 19*2) per cent of standard length, 1-7-2-3 (modal
range 1-9-2-0) as long as deep.

Lower jaw markedly oblique, sloping at 40-5o, its tip projecting slightly beyond
the upper jaw in some fishes and level with it in others. Length of lower jaw 42-2-51-4
(M 46-0) per cent of head, 2-4-3-3 times as long as broad. Lips slightly thickened,
the median dentigerous part of the premaxilla not noticeably expanded. The posterior
tip of the maxilla usually reaches the vertical from the posterior end of the pre-
maxillary pedicels, or slightly beyond, but never reaches to below the anterior
orbital margin.

FIG. 10. Haplochromis longirostris, x N.S. (From Regan, Proc. zool. Soc. Lond.)

Gill rakers variable, from long and slender to short and stout, sometimes flattened
and broadly branched ; 9-11 (rarely 8) rakers on the lower part of the first gill arch.

Scales ctenoid, lateral line with 32 (f.6), 33 (f.n) or 34 (f.12) scales, cheek with 3 or
4 rows. Five to 7 (mode 6) scales between the lateral line and the dorsal origin, 5-7
(mode 6), rarely 8, between the pectoral and pelvic fin bases.

Fins. Dorsal with 24 (f-4), 25 (f.19) or 26 (.5) rays, comprising 15 (f.3), 16 (f.23)
or 17 (f.2) spinous and 8 (f-4), 9 (f-i8) or 10 (f.6) branched rays. The dorsal fin of the
holotype is badly damaged and a count gives ca. 14, 9 rays. Anal fin with 11-13 ravs
comprising 3 spines and 8 or 9 (rarely 10) rays. Caudal truncate, scaled on its proxi-
mal two-thirds. Pectoral 20-8-25-2 (M = 23-0) per cent of standard length. First
pelvic ray only slightly produced (not filamentous) in adult males.

Teeth. The outer row in both jaws is composed of slender and fine, moderately to
strongly curved teeth, those situated posterolaterally in the upper jaw are almost
hair-like. These outer teeth are unicuspid in most fishes, but in three (85-5 mm. and
two of in mm. S.L.) there is a mixture of bi- and unicuspids or weakly bicuspids and
unicuspids. There are 40-70 (M = 56) teeth in the upper jaw, the number showing
no clear-cut correlation with standard length.

The inner teeth are either all unicuspid (fishes >I25 mm S.L and a few in the 95-

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 173

125 mm. range) or a mixture of uni- and tricuspids, unicuspids and weakly tricuspids
or, as in the smallest fish (a syntype of H. gracilicauda) , all tricuspids. The inner rows
are implanted obliquely so that the teeth lie almost horizontally. There are usually
2 inner rows in the upper jaw (occasionally 3, rarely i) and one or, less frequently,
2 in the lower jaw.

Osteology. The neurocranium barely differs from that of H. mento despite the more
oblique angle of the lower jaw in H. longirostris. This greater jaw angle is apparently
brought about by a slight difference in the articulatory surfaces of the quadrate and
angular.

Vertebrae. 31 (f.y) or 32 (f.2) comprising 13 (f.i) or 14 (f.8) precaudal and 17 (f.6)
or 18 (f-3) caudal elements.

35 mm

FIG. ii. Haplochromis longirostris, neurocranium.

Lower pharyngeal bone triangular, its dentigerous surface broader than long. The
teeth are fine, cuspidate and compressed (those of the two median rows sometimes
slightly enlarged) and arranged in 24-26 irregular rows.

Coloration. The colours of live fishes are unknown.

Preserved females are grey-brown above, silvery below with, on the flanks and belly,
a slight brassy overtone. The lower lip and sometimes the mental area are sooty.
All fins are hyaline, the dorsal with dark lappets and the caudal often densely and
darkly maculate.

Sexually active males have an overall sooty appearance but with silvery areas
showing through, particularly on the anterior flanks and the lateral aspects of the
chest and belly ; the ventral parts of the chest and belly, however, are always darker
than the dorsal parts of the body. The lower jaw, ventral part of the preoperculum
and the branchiostegal membrane are intensely blank. Dorsal fin dark grey with
black lappets and dark streaks between the soft rays. Caudal dark, especially on its
proximal half. Anal dark, with a black area on the spinous part continued posteriorly
as a narrow black band along the fin base ; the ocelli are barely visible as faint grey
blotches. Pelvic fins black.

The amount of silver visible on the flanks may be correlated with the fish's state of
sexual activity. Juvenile males are indistinguishable from females.

i 7 4 p - H - GREENWOOD

Distribution. Lake Victoria.

Ecology. Habitat. The species is known from only a few localities, each of which is
a sandy beach either exposed to the open lake or within a sheltered gulf ; apparently
H. longirostris is nowhere common.

Breeding. One female contains advanced embryos in the buccal cavity ; no other
brooding fishes are known. Most individuals >no mm. S.L. are sexually mature, as
are some smaller fishes. Apparently both sexes reach the same maximum adult size.

Food. From the scanty data available, H. longirostris appears to feed on both insects
(particularly pupal stages) and small fishes. Fourteen of the twenty-two specimens
examined contained food. Eight had fed exclusively on insects, one on insects and
fishes, and five on fishes only. The fish and insect remains were very fragmentary
so identification could not be taken far. Only one fish could be identified (a small
cyprinid, probably Engraulicypris argenteus} ; the insects are mainly pupal Baetids.

Affinities. The slender, elongate body form and very oblique jaws of H. longirostris
are immediate diagnostic characters and ones which isolate the species from all but
one other in Lake Victoria. This other species is H. argenteus. Haplochromis argenteus
differs from H. longirostris in having the jaws less oblique, the premaxilla more
clearly " beaked " (i.e. the median toothed part expanded) and in having the eye
diameter noticeably larger than the interorbital width.

On neurocranial characters, H. longirostris can be referred to the H . mento complex.
But it differs from other species of this group in several character combinations and
also in its mixed insect-fish diet. Phylogenetically, H. longirostris can be considered
a somewhat isolated offshoot from the " mento "-group stem.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda
B.M. (N.H.). 1906.5.30.262 . Bunjako . Degen

(Paratype H. gracilicauda)

,, 1906.5.30.516 . Bunjako . ,,

,, 1906.5.30.268 . Entebbe

(Lectotype H. gracilicauda}

1911.3.3.13 . Jinja, Ripon Falls . Bayon

1962.3.2.10-13 . Jinja . E.A.F.R.O.

,, 1962.3.2.1 . Grant Bay .

,, 1962.3.2.5 . Karenia (near Jinja)

,, 1962.3.2.14-25 . Beach near Nasu Point . ,,

Tanganyika

,, 1962.3.2.2 . Majita Beach

,, 1962.3.2.3-4 . Mwanza, Capri Bay

Lake Victoria, Locality Unknown
1962.3.2.6-9 . . ,,

Haplochromis mento Regan 1922
Text-figs. 12 and 25

Paratilapia longirostris (part) : Blgr., 1915, Cat. Afr. Fish., 3, 332, fig. 223.
Haplochromis mento Regan, 1922, Proc. zool. Soc. Lond., 183.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 175

Holotype. A specimen 174-0 mm. S.L., from Bunjako ; B.M. (N.H.) Reg. No.
1906.5.30.258.

Description, based on twenty-five specimens (113-178 mm. S.L.) including the
holotype.

Depth of body 24-1-31-5 (M = 28-9) per cent of standard length, length of head
21.5-37-1 (M = 34-5) per cent. There is some variation in the lateral outline of the
head ; in most specimens the head gives an impression of attenuation and of being
pointed, but in others this impression is less marked and the head seems coarser and
more bluntly rounded. There are no clear-cut morphometric differences between
specimens belonging to either group, and intergrades exist.

Depth of preorbital 18-4-23-6 (M = 21-2) per cent of head, least interorbital width
20-4-24-4 (M. = 22-5) per cent. Snout 1-5-1-8 times as long as broad, its length
36-6-43-4 (M = 39-3) per cent of head ; diameter of eye 16-8-22-0 (M = 19-2) per
cent, depth of cheek 18-9-27-8 (M = 23-6) per cent.

FIG. 12. Haplochromis mento ; holotype, -6 X N.S. (From Boulenger, Fish. Nile.}

Caudal peduncle length 14-1-20-9 (M = 17-5) per cent of standard length, 1-3-2-2
(modal range 1-7-1-8) times as long as deep.

Lower jaw projecting slightly in fishes <i5o mm. S.L. and more markedly promi-
nent in larger fishes ; its length 41-8-50-0 (M = 46-8) per cent of head and 2-1-2-8
times as long as broad. Mouth slightly oblique or even horizontal, the medial denti-
gerous part of the premaxilla expanded anteroposteriorly. Posterior tip of the
maxilla generally reaching to a vertical midway between the nostril and the anterior
orbital margin, but sometimes extending a little more posteriorly.

Gill rakers short and stout, 8-10 (mode 9) on the lower part of the first arch, the
lowermost 1-3 rakers reduced.

Scales ctenoid ; lateral line with 32 (f.i), 33 (f.7), 34 (f.i4) or 35 (3) scales, cheek
with 3 or 4 (rarely 5) rows. Six or 7 (rarely 5) scales between the lateral line and the
dorsal fin origin, 7 or 8 (rarely 9) between the pectoral and pelvic fin bases.

Fins. Dorsal with 23 (f.i), 24 (f.2), 25 (f.12) or 26 (f.io) rays, comprising 13 (f.i),
15 (.5), 16 (f.i8) or 17 (f.i) spinous and 8 (f.i), 9 (f.n), 10 (f.12) or n (f.i) branched
rays. Anal with n, 12 (mode) or 13 rays, comprising 3 (4 in one fish) spines and 8-10

176 P. H. GREENWOOD

rays. Pectoral fin 19-7-24-7 (M = 23-4) per cent of standard length. First pelvic
ray produced in adult males. Caudal truncate, scaled on its basal half to two-thirds.

Teeth. The outer row in each jaw is composed of unicuspid, very strongly curved
and moderately stout teeth, there being 38-66 (M = 52) in the upper jaw. The teeth
of the inner series are predominantly unicuspids but in a few specimens there is an
admixture of unicuspid and weakly tricuspid teeth. Teeth in the outermost row of
the inner series, especially in the upper jaw, are often enlarged. All inner teeth are
implanted at a very oblique angle. There are 2 or 3 (rarely 4) rows in the upper jaw
and 2 (rarely I or 3) in the lower jaw.

Osteology. The neurocranium of H. mento presents no outstanding specific character-
istics ; it is typical of the long, shallow and narrow skull found also in H. macrognathus ,
H. estor, H. gowersi and H. dentex. The premaxilla shows pronounced medial expan-
sion, a character usually associated with the " mento " skull type but probably more
marked in this species than in the others of the group.

Lower pharyngeal bone triangular, small and fine, its dentigerous surface broader
than long. The teeth are cuspidate and generally fine but some of the median series
may be coarser. The teeth are rather sparsely distributed and are arranged in 16
(rarely) to 20 (most common) rows.

Vertebrae. 30-32 (mode 30) comprising 13 or 14 precaudal and 17 or 18 caudal
elements (7 specimens examined).

Coloration. Live fishes. Adult males have a steely grey-blue ground colour, darker
(almost sooty) on the chest and belly, branchiostegal membrane dark grey. Snout,
cheek and opercular region have an irridescent sheen. The spinous part of the dorsal
is irridescent blue, the soft part dark neutral ; lappets dusky. Caudal fin dark neutral.
Spinous part of the anal dusky, the soft part dark neutral and bearing the dull
orange-red ocelli. Pelvics black on the outer third, the remainder of the fin dusky.
The colours of live females and immature males are unknown but a recently dead
female has the dorsal aspects of the body and head bright green and the flanks silver.
All fins are yellowish neutral.

Preserved material. Sexually active males. Ground colour brownish with a sooty
overlay on the belly, the branchiostegal membrane blackish-grey and a broad but
indistinct lachrymal blotch. The dorsal fin is dark brown with black lappets. Caudal
dark, densely maculate on its proximal two-thirds. Anal brownish to sooty-grey,
the ocelli indistinct and greyish-brown. Pelvics black on the outer third, the remain-
ing rays black but the intervening membrane light grey. Quiescent males have a
female-type coloration (see below) except that the pelvics are sooty on the outer
three rays, the branchiostegal membrane is dark and there are faint traces of grey-
white ocelli on the anal fin. The vertical bars on the flank are not always visible.
Females are brownish-grey above, becoming golden-silver below the level of the lower
lateral line. There are about nine faint vertical bars on the flank and caudal peduncle ;
each bar extends from slightly above the upper lateral line to a point some one to three
scales below the level of the lower lateral line. The first five bars may be joined by a
midlateral stripe of about the same width as the bars. The dorsal fin is colourless or
light brown, the lappets dark. The anal, caudal and pelvic fins are yellowish brown
to greyish.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 177

Distribution. Lake Victoria.

Ecology. Habitat. Haplochromis mento is apparently confined to areas where the
substrate is sand or sand and rocks, and where the water is not more than twenty
feet deep. Most localities from which specimens were obtained are exposed but one
is a sheltered gulf ; all are within two-hundred yards of the shore, the majority
within one hundred yards.

Food. Eleven of the thirty-one specimens examined contained food in the stomach
or intestines ; of these specimens, one had only the remains of a large dragonfly
larva and the others only fish remains. It is regrettable that so few specimens contained
food because the prey species are predominantly cyprinid fishes ; most other pisci-
vorous Haplochromis seem to concentrate on cichlids. Of the ten specimens with fish
in the guts, three had fed on Cichlidae (determined as Haplochromis in two cases) and
seven on Cyprinidae (identified as Engraulicypris argenteus in two cases) . Admittedly
the contents of seven out of eleven stomachs is far too small a sample on which to base
generalizations. However, it should be remembered that in samples of a like size
from many other species, the identifiable food is entirely of cichlid origin and pre-
dominantly Haplochromis. Also, the eleven H . mento came from five different localities.
Possibly, then, H. mento has specialized in preying on cyprinids, a group not heavily
tapped by other piscivorous fish-predators (see Corbet, 1961).

Breeding. The species is a female mouth brooder. Fishes less than 135 mm. S.L.
are immature. Possible sexual dimorphism in the maximum size attained cannot be
determined from the sample available since few females are represented ; the three
fishes of 170 mm. S.L. and over are all males.

Affinities. Structurally, H. mento is a specialized predator and, therefore, shows at
least group affinity with H. estor, H. gowersi and H. longirostris. At this level there
is also some affinity between H. mento and H. macrognathus, the latter being a
structurally more extreme form of H. mento. The affinities of these two species are
discussed on p. 174. Haplochromis estor and H. gowersi both differ from H. mento in
several ways, but primarily in the shape of the head and the more oblique jaws of the
former species. Thus, it is impossible to consider H. mento as having any close morpho-
relatives, although the species does have phyletically close affinities with several
others. Since H. mento shows the general characters of its specialized predatory line
(particularly syncranial characters) it may represent a fairly basic anatomical state
in that line.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1906.5.30.258 . Bunjako . Degen
(Holotype)

1962.3.2.179-191 . Beach near Nasu Point . E.A.F.R.O.

1962.3.2.192 . Old Bukakata Bay

1962.3.2.193 . Between Yempita and Busiri

Isls. (Buvuma Channel)

1962.3.2.194 . J in j a (below golf course)

1962.3.2.195 . Grant Bay
1962.3.2.196-7 . Kagera Port

i 7 8 P. H. GREENWOOD

Museum and Reg. No. Locality Collector

Uganda (continued)

B.M. (N.H.). 1962.3.2.198-200 . Entebbe (airport beach) . E.A.F.R.O.

1962.3.2.206-7 . Entebbe Harbour . ,,

1962.3.2.201-5 . Bukafu Bay

1962.3.2.209 . Thruston Bay .

1962.3.2.210-1 . Ramafuta Isl. . ,,

(Buvuma Channel)
1962.3.2.212-7 . Buka Bay . ,,

Tanganyika
1962.3.2.208 . Beach near Majita . ,,

Haplochromis mandibularis sp. nov.

Holotype. A specimen 140-0 mm. S.L., from Jinja ; B.M. (N.H.) Reg. No.
1962.3.2.222.

Description based on ten specimens (131-174 mm. S.L.) including the holotype.
Only one of these specimens is a female.

Depth of body 31-6-34-3 (M = 33-1) per cent of standard length, length of head
38-0-39-3 (M = 38-6) per cent ; breadth of head immeditely anterior to orbits 22-4-
28-3 (M = 25-9) per cent of head length. Dorsal profile of head sloping at an angle
of 3O-4O, slightly curved and becoming concave in large fishes ; premaxillary
pedicels slightly prominent.

Preorbital depth 17-7-20-7 (M = 19-3) per cent of head, least interorbital width
19-6-22-2 (M = 21-2) per cent. Snout 1-5-1-8 times as long as broad, its length
36-2-39-7 (M = 38-6) per cent of head ; diameter of eye 17-2-22-2 (M = 19-4) per
cent, depth of cheek 24-5-29-2 (M = 26-1) per cent.

Caudal peduncle 12-2-15-2 (M = 14-2) per cent of standard length, 1-0-1-3 (mode)
times as long as deep.

Mouth slightly oblique, lower jaw extension variable, from projecting markedly to
no extension beyond the upper jaw. Lips somewhat thickened, the median toothed
portion of the premaxilla only slightly expanded anteroposteriorly. The posterior
tip of the maxilla reaches a point near the anterior orbital margin in most fishes, but
it extends to the level of the orbit (or even slightly beyond) in a few specimens.
Lower jaw 47-3-56-8 (M = 51-5) per cent of head, 2-0-2-9 (mode 2-5) times as long as
broad.

Gill rakers short and stout, 8 (mode) or 9 on the lower part of the first arch, the
lowermost I or 2 rakers reduced.

Scales ctenoid, lateral line with 32 (f.4) or 33 (f.6) scales, cheek with 4 (rarely 5)
rows. Six or 7 scales between the lateral line and the dorsal fin origin, 6 or 7 (rarely 5)
between the pectoral and pelvic bases.

Fins. Dorsal with 24 (f.i), 25 (f.8) or 26 (f.i) rays, comprising 15 (f.i) or 16 (f.g)
spinous and 8 (f.i), 9 (1.7) or 10 (f.2J branched rays. Anal with 12 rays, comprising
3 spines and 9 branched rays. Pectoral fin 21-4-24-3 (M 22-2) per cent of standard
length. First pelvic ray somewhat produced and filamentous in adult males. Caudal
truncate, the posterior margin running somewhat obliquely forwards and downwards ;
scaled on the proximal half to two-thirds.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 179

Teeth. The outer row in both jaws is composed of fairly stout unicuspid teeth,
recurved anteriorly and strongly incurved laterally ; there are 72-94 (M = 82) teeth
in the upper jaw. The inner teeth are all unicuspid and obliquely implanted, being
arranged in 4-6 rows in the upper jaw and 2 or 3 (rarely 4) in the lower jaw.

Osteology. The neurocranium closely approaches the H. mento-macrognathus type
but is somewhat less extreme, particularly with regard to its anterior profile which
is steeper. This gives the skull a stouter and more compact appearance. The pre-
maxilla is as beaked as that of H. macrognathus.

Lower pharyngeal bone triangular, its dentigerous surface as long as it is broad, but
with the posterior margin deeply indented in all specimens except the two largest.
The teeth are fine, cuspidate and somewhat irregularly arranged in 20-24 rows.

Coloration. The live colours are unknown. A preserved female is brownish-grey
above, shading to silvery-brown below. On the flanks are traces of about seven dark
transverse bars and there is a very weak spot on the caudal fin base. The snout and
upper jaw are dark grey, the lower jaw is paler. Dorsal fin hyaline but with sooty
lappets and an oblique, ventrally directed dark bar on the soft part ; the base of the
soft dorsal is dark. Caudal greyish-black, the basal half densely and darkly maculate.
Pelvic and anal fins hyaline, the latter with one, faint, dead-white ocellus.

Quiescent males are brownish-grey above and on the snout and upper jaw, becoming
silvery below. There is an interrupted midlateral black band from the posterior
opercular margin to the caudal base, where it ends as a faint spot ; the spot may
extend onto the caudal fin. There are also very faint traces of an interrupted dark
band running immediately above the upper lateral line. The dorsal fin is hyaline but
with dusky lappets and is densely maculate on the soft part. Anal hyaline, as is the
caudal although the latter is densely maculate. Pelvics faintly dusky.

Sexually active males are dark brown above, brassy on the flanks, and black vent-
rally, particularly on the belly, chest and branchiostegal membrane. The lower
part of the preoperculum and the lower third of the operculum are also black. The
dorsal fin is dusky with black lappets ; the soft part is hyaline on the distal half and
densely maculate proximally. The basal third of the caudal is dark brown or black,
the distal part yellowish. The anal is yellowish with a narrow black basal band and
a faint black area over the last two rays ; one or two dark ocelli are present. Pelvic
fins are black.

Distribution. Lake Victoria.

Ecology. Habitat. The five localities from which H. mandibularis were obtained
are all close inshore and have a hard sand substrate ; four are in sheltered areas and
one is partly exposed. The maximum depth at which the specimens could have been
living is between 20 and 25 feet

Food. Five of the ten specimens examined contained food. In each case the gut
contents were fishes, identifiable as Haplochromis in two specimens and as cichlids in
the other three.

Breeding. All the specimens are adult and only one (135 mm. S.L.) is a female. No
other data are available.

Affinities and diagnosis. Superficially and in many morphometric characters,
H. mandibularis resembles H . macrognathus. It is distinguishable, however, by the

i8o

P. H. GREENWOOD

lack of a large, well-defined black spot at the base of the caudal fin, the lack of a black
mental spot and by its deeper body. An obvious difference, but one which cannot be
quantified, is the shape of the head which is less compressed and less acute in H.
mandibularis ; also in this species, at least in the size-range available, the eye diameter
is equal to or is slightly larger than the preorbital depth. In H. macrognathus of the
same size (131-174 mm.) the preorbital is deeper than the eye. Anatomically and
superficially H. mandibularis could represent the ancestral condition from which
H. macrognathus evolved. Haplochromis mandibularis also shows relationship with
H. mento but this is less intimate than that with H. macrognathus ; again the relation-
ship is of a generalized to a more specialized species.

Study material and distribution records
Museum and Reg. No.

B.M. (N.H.). 1962.3.2.222
(Holotype)

1962.3.2.223-4
,, 1962.3.2.225

,, 1962.3.2.226

,, 1962.3.2

Locality
Uganda
Jinja (below golf course)

1962

Jinja (below golf course)

Jinja Pier

Bugungu (opp. Jinja)
Beach near Nasu Point
Lake Victoria, Locality Unknown
3.2.229-230

.227-?

Collector
E.A.F.R.O.

Haplochromis gowersi Trewavas, 1928

Text-figs. 13 and 14
Haplochromis gowersi Trewavas, 1928, Ann. Mag. nat. Hist. (10), 2, 94.

Lectotype. A specimen 154-0 mm. S.L.; B.M. (N.H.) Reg. No. 1928.5.24.478.
Description based on twenty-two specimens (145-224 mm. S.L.), including the
lecto- and paratypes.

FIG. 13. Haplochromis gowersi ; type, -68 x N.S. (Drawn by Miss M. Fasken.)

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 181

Depth of body 26-5-33-5 (M = 29-4) per cent of standard length, length of head
35-8-38-4 (M = 37-0) per cent. Dorsal head profile slightly curved, sloping at an
angle of 30-35 ; premaxillary pedicels slightly prominent, but not sufficiently
prominent to give a pronounced interorbital concavity to the profile.

Preorbital depth 19-3-24-0 (M = 22-0) per cent of head, least interorbital width
16-1-23-3 (M = 19-7) per cent. Snout 1-3-1-5 times as long as broad, its length 36-8-
42-2 (M = 39-6) per cent of head ; diameter of eye 15-5-19-3 (M = 17-5) per cent,
depth of cheek 27-8-33-3 (M = 29-5) per cent.

Caudal peduncle 13-3-17-6 (M = 14-8) per cent of standard length, 1-3-1-6 (mode
1-4) times as long as deep.

Mouth oblique (sloping at ca. 30 ~35 ) , lips slightly thickened, the median dentigerous
surface of the premaxilla expanded anteroposteriorly ; lower jaw always projecting
but to a variable extent. Length of lower jaw 49-1-55-1 (M = 52-0) per cent of head,

39mm

FIG. 14. Haplochromis gowersi, neurocranium.

2-1-2-7 times as long as broad. Posterior tip of the maxilla not reaching the vertical
through the anterior orbital margin, usually reaching a point about midway between
the nostril and the orbit, sometimes a little more posteriorly.

Gill rakers short and stout, 8-10 (mode 9) on the lower part of the first gill arch.

Scales ctenoid ; lateral line with 31 (1.3), 32 (f.2), 33 (f.8) or 34 (f.2) scales, cheek
with 4 or 5 (rarely 6) rows. Six or 7 (less frequently 8) scales between the lateral line
and the dorsal origin, 7 or 8 (rarely 5 or 6) between the pectoral and pelvic fin bases.

Fins. Dorsal with 24 (f.i), 25 (f.io) or 26 (1.9) rays, comprising 15 (f.7) or 16 (f.13)
spinous and 9 (f.7), 10 (f.n) or n (f.2) branched rays. Anal with 12 or 13 rays,
comprising 3 spines and 9 or 10 branched rays. Pectoral fin length very variable,
17-6-27-3 (M = 20-5) per cent of standard length. First pelvic ray slightly produced
in adult males. Caudal truncate, scaled on its proximal third to half.

Teeth. The outer row in both jaws is composed of unicuspid, moderately stout and
slightly to strongly curved teeth. There are 38-52 teeth in the upper jaw, the number
showing a very slight positive correlation with standard length. The inner teeth are
unicuspid and implanted obliquely ; there are 3 or 4 (mode), rarely 5 rows in the
upper jaw and 2 or 3 (rarely 4 or 5) in the lower jaw.

182 P. H. GREENWOOD

Osteology. The neurocranium of H. gowersi is virtually identical with that of
H. mento and is thus similar to the neurocranium of H. estor.

Vertebrae. Twenty-nine or 30 (mode) in the six specimens examined, comprising
13 or 14 precaudal and 16 or 17 caudal elements.

Lower pharyngeal bone triangular, the dentigerous surface longer than broad, or,
less frequently, as long as broad. The teeth are moderately stout and cuspidate,
and are arranged in 18-20 slightly irregular rows.

Coloration is unknown in life. Preserved females have a silvery ground colour,
darker dorsally and on the head. A prominent and fairly broad midlateral stripe
runs from the opercular margin to the caudal origin ; a fainter, narrower and some-
times partly interrupted stripe runs above the upper lateral line. In some specimens
there is a faint lachrymal stripe. All fins are hyaline, the soft dorsal and the caudal
are densely maculate and the former has faintly sooty lappets. Immature males are
coloured like females except that the pelvics are darker. Sexually active males have
a dark brown ground colour, becoming sooty from below the lateral line, with a
faint golden sheen on the flanks and lower part of the operculum. Midlateral and
dorsolateral stripes are developed as in females. The branchiostegal membrane is
black and there is a faint, dark lachrymal stripe. The dorsal and caudal fins are dark,
the anal is yellowish, sometimes with a dusky overlay which does not extend to the
distal margin. There are one to three large, greyish anal ocelli. The pelvic fins are
black.

Distribution. Lake Victoria.

Ecology. Habitat. The species does not appear to be confined to any particular
substrate, since it is caught over both soft (mud) and hard bottoms (sand, rock,
shingle) ; it is found in exposed habitats (open beaches, off-shore islands) as well as
in sheltered bays and gulfs. In no locality was the water more than 20 feet deep.

Breeding. One specimen (a female 153 mm. S.L.), from a sandy beach near the
Kagera river mouth, had larvae in the buccal cavity. Sexual maturity is reached at
a length of about 150 mm.

Food. All the fifteen specimens with food in the stomach or intestines had fed on
fishes, but in two there was also a quantity of macerated plant tissue. The fish
remains were identified as : Haplochromis (f-9), Cichlidae (f.4) ; unidentifiable (f.2).

Affinities. As Trewavas (1928) first suggested, H. gowersi has affinities with H.
mento. It differs from H. mento in having a deeper cheek (mean depth 29-5% of head
cf. 23-6%), a longer lower jaw (M = 52% of head cf. 46-8%) and a more oblique
mouth.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1928.5.24.478 . Entebbe . Graham

(Lectotype)

B.M. (N.H.). 1928.5.24.479 . Entebbe . Graham

(Paratype)

,, 1962.3.2.392 . Williams Bay . E.A.F.R.O.

1962.3.2.393-5 . Beach near Nasu Point

,, 1962.3.2.396-7 . Karenia (near Jinja)

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 183

Museum and Reg. No. Locality Collector

Uganda (continued)

B.M. (N.H.). 1962.3.2.398 . ThrustonBay . E.A.F.R.O.

1 962 . 3 . 2 . 399-400 . Kagera Port

1962.3.2.401-3 . Buka Bay

1962 .3.2. 4045 . Bukassa

1 962 . 3 . 2 . 406 . Pilkington Bay

1962.3.2.407 . Ramafuta Isl.

(Buvuma Channel)
1962.3.2.409-415 . J in j a Pier

Lake Victoria, Locality Unknown
1962.3.2.408

Haplochromis macrognathus Regan 1922
Text-figs. 15 and 25

Paratilapia longirostris (part) : Blgr., 1915, Cat. Afr. Fish., 3, 332.

Haplochromis macrognathus Regan, 1922, Proc. zool. Soc. London, 182 ; PI. 2, fig. 2.

Holotype. A specimen 160-0 mm. S.L.; from Bunjako. B.M. (N.H.) Reg. No.
1906.5.30.260.

Description based on twenty-three specimens (including the holotype) 80-174 mm.
S.L.

Depth of body 26-6-33-3 (M = 30-7) per cent of standard length, length of head
33-8-41-4 (M = 38-2) per cent. Head noticeably compressed, its breadth (as measured
immediately anterior to the orbital margin) 20-4-26-0 (M = 22-9) per cent of its
length. Dorsal head profile straight or slightly curved, sloping at 25-30 ; pre-
maxillary pedicels moderately prominent and breaking the slope of the head as seen
in profile.

Depth of preorbital 18-2-23-3 (M 21-3) per cent of head, least interorbital width
16-5-22-2 (M = 18-6) per cent. Snout 1-5-2-2 times as long as broad (broadest in
smaller fishes ; 1-5-1-6 times), its length 33-3-44-2 (M = 39-0) per cent of head, but
with indications of weak positive allometry. Diameter of eye 16-1-24-1 (M = 18-5)
per cent of head, slightly greater than, or equal to the preorbital depth in fishes
80-104 mm. S.L., less than the preorbital in larger individuals. Depth of cheek
24-2-31-2 (M = 27-0) per cent.

Caudal peduncle length 13-0-20-0 (M = 15-0) per cent of standard length, 1-1-1-8
(mode 1-4) times as long as deep.

Lower jaw projecting beyond the upper, variable in extent but generally more
prominent in larger fishes, its length 48-1-58-8 (M = 54-0) per cent of head and 2-7-3-5
times as long as broad. Mouth moderately oblique, the lower jaw sloping at 25-3O ;
lips slightly thickened. The median dentigerous part of the premaxilla is greatly
expanded anteroposteriorly so that in some specimens the premaxilla has a beak-like
appearance. The posterior tip of the maxilla rarely reaches the vertical through the
anterior orbital margin ; usually it extends to a vertical through the posterior tip of
the premaxillary pedicels.

Gill rakers short and stout, 8 (mode) or 9, rarely 10 on the lower part of the first
arch, the lowermost 1-3 rakers reduced.

184

P. H. GREENWOOD

Scales ctenoid ; lateral line with 30 (f.2), 31 (2), 32 (f.i3) or 33 (f.6) scales, cheek
with 3-5 (mode 4) rows. Six or 7 (rarely 8) scales between the lateral line and the
dorsal origin, 6-8 between the pectoral and pelvic fin bases.

Fins. Dorsal with 24 (f.8), 25 (.14) or 26 (f.i) rays, comprising 14 (f.i), 15 (f.ig)
or 16 (f.i3) spines and 8 (f.3), 9 (f.i3) or 10 (.7) branched rays. Anal with n (.3),
12 (f.i8) or 13 (f.2) rays, comprising 3 spines and 8-10 branched rays. Pectoral
20-7-25-0 (M = 22-7) per cent of standard length. Pelvics with the first ray moderately
produced in adult males. Caudal truncate, scaled on its proximal two-thirds, or,
occasionally, as much as four-fifths.

Teeth. The outer row of teeth in fishes >ioo mm. S.L. is composed entirely of stout,
strongly curved unicuspids, but smaller fishes have an admixture of weakly bicuspid
and unicuspid teeth. There are 50-100 (M = 80) teeth in the upper outer row. Teeth

FIG. 15. Haplochromis macrognathus , holotype, .66 x N.S. (Drawn by Lavinia Beard.)

of the inner series in most fishes >no mm. S.L. (but also in some between 95 and
100 mm.) are unicuspid ; in smaller fishes there is usually a mixture of uni- and tri-
cuspid teeth, sometimes with one type predominating in one jaw but not in the other.
All the inner teeth are implanted obliquely so that in the upper jaw their crowns lie
almost horizontally. The number of inner tooth rows in the upper jaw shows some
positive correlation with the fish's size, there being 3-6 (mode 5) for the size range
available. No such correlation exists in the lower jaw where there are 2 or 3 (less
commonly i) rows.

Osteology. The neurocranium is almost identical with that of H. mento, differing
only in being slightly shallower and narrower, and in having the dorsal profile some-
what flatter. The premaxilla shows an extension of the trend seen in H. mento in
that the median dentigerous section is greatly expanded anteroposteriorly and is
almost beak-like.

Vertebrae. Twenty-nine or 30, comprising 13 or 14 precaudal and 16 or 17 caudal
elements (5 specimens examined).

Lower pharyngeal bone. The outline is somewhat variable and is correlated with
the head-shape of the fish. The dentigerous surface is triangular and usually equi-
lateral but sometimes it is longer than broad, or, less frequently, broader than long.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 185

The teeth are relatively fine (the two median rows somewhat coarser), are cuspidate
and are densely crowded in 22 to 24 rows.

Coloration. The colour pattern of preserved females is a salient feature of the species,
and is rare amongst Victoria Haplochromis. Nevertheless, it only represents the
complete manifestation of the basic patterns shown in part by many species.

The colours of live males are unknown ; females are silvery-grey with a distinct
midlateral black stripe on the flank and caudal peduncle, ending in a black blotch
on the caudal fin near its base ; a very faint dark stripe runs slightly above the upper
lateral line and there is a dark mental spot on the lower jaw. The chest is yellowish
but the branchiostegal membrane is a dark saffron yellow. The dorsal fin is neutral
but with dark spots and streaks between the rays. The upper half of the caudal is
neutral and darkly maculate, the lower half is yellow. The anal fin is faintly yellow
and the pelvics are saffron yellow.

Preserved material. Females. The ground colour is grey-brown dorsally, silver on
the flanks and belly. A dark midlateral stripe runs from the eye to the caudal fin
origin where it expands into a large blotch on the basal part of the fin ; the line is
narrow on the head but widens shortly after it passes the posterior margin of the
operculum. A second dark stripe runs slightly above the upper lateral line, from a
point above the posterior opercular margin to the origin, or slightly beyond the
origin of the caudal peduncle. At the base of the dorsal fin there is a series of dark
spots which extend slightly upwards onto the fin membrane. There is usually a dark
lachrymal spot, and always a dark blotch at the tip of the mandible. All fins, except
the pelvics, are colourless although the dorsal and caudal may be darkly maculate.
The pelvics are greyish, being darkest along their anterior margins.

Adult males. The ground colour is dark orange-brown. The stripes and other
markings are as in females, with the addition of five or six rather faint transverse bars
linking the two lateral stripes and extending a little below the lower stripe. The snout,
lower jaw, branchiostegal membrane, lower limb of the preoperculum, lower part of
the operculum, the chest and belly are black. The dorsal fin is sooty, the caudal
dark and maculate. The anal has a black basal stripe which spreads diffusely out
over most of the fin except for its distal margin which is yellowish-brown ; there are
as many as six, irregularly arranged, blackish ocelli. The pelvics are black. Less
sexually active males have a similar coloration but the pattern is not so intense,
especially with regard to the black antero-ventral parts of the head and body.

Distribution. Lake Victoria.

Ecology. Habitat. The species has been recorded from three different habitats,
viz.: sheltered gulfs with a hard substrate ; sheltered bays with a mud substrate ;
and exposed, sandy beaches. From catch records it seems that H. macrognathus is
commoner in areas where the bottom is hard than in places where it is muddy, and
that the species does not occur at depths greater than twenty-five feet.

Food. Only five of the twenty specimens examined had ingested material in the
stomach or intestines. Four of the five had fed exclusively on fish, and the fifth
contained only fragments of plant tissue with attached colonies of blue-green algae
(see also p. 144). The fish remains were identifiable in two cases as Haplochromis, in
one as Cichlidae and in the fourth were too fragmentary for identification.

i86 P. H. GREENWOOD

Breeding. Haplochromis macrognathus is a female mouth brooder. Most fishes
<I25 mm. S.L. are immature (one male 121 mm. S.L. was classified as " starting ") ;
there is no sexual dimorphism in the maximum size attained.

Affinities. Anatomically, Haplochromis macrognathus belongs to the H. mento-
gowersi-estor complex; but it is immediately distinguished from these species by the
marked compression of the head and its almost beaked premaxillaries. The nearest
living relative is, perhaps, H. mandibularis which although anatomically somewhat
less specialized than H. mento shows certain characters (head compression, shape
of mouth, coloration) strongly suggestive of the H. macrognathus condition.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1906.5.30.260 . Bunjako . Degen

(Holotype)

1962.3.2.420 . Ekunu Bay . E.A.F.R.O.

1962.3.2.421-5 . J in j a Pier

1962 .3.2. 431-3 . Jinja, below golf course

1962 .3.2. 434 . Fielding Bay

1962.3.2.435 . Pilkington Bay

1962.3.2.436 . Entebbe, near Buganga

1962.3.2.437 . Between Vempita and Busiri Isls. .

(Buvuma Channel)
Kenya
1962.3.2.418-9 . Trawl S. of Port

Southby
Tanganyika

1962 .3.2. 416-7 . Beach near Majita

1962.3.2.426-430 . Majita Beach

Haplochromis pellegrini Regan 1922
Text-fig. 16

Paratilapia prognatha (part) : Blgr., 1915, Cat. Afr. Fish., 3, 333.
Haplochromis pellegrini Regan, 1922, Proc. zool. Soc. Lond., 185, fig. u.

Lectotype. A specimen 104-0 mm. S.L., from Entebbe, B.M. (N.H.) Reg. No.
1906.5.30.253.

Description based on twenty-five specimens (71-104 mm. S.L.) including the lecto-
and paratypes.

Depth of body 29-0-33-6 (M = 31-1) per cent of standard length, length of head
34-6-37-9 (M = 36-3) per cent. Dorsal head profile sloping at ca. 30, somewhat
concave above the eyes, the premaxillary pedicels fairly prominent.

Depth of preorbital 15-4-22-2 (M = 17-7) per cent of head, least interorbital width
18-2-24-0 (M = 21-0) per cent. Snout longer than broad, its length 30-8-36-0
(M = 34-0) per cent of head ; diameter of eye 20-5-27-0 (M = 23-6) per cent, depth
of cheek 21-4-27-0 (M = 24-3).

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 187

Caudal peduncle length 13-2-17-8 (M = 15-4) per cent of standard length, 1-1-1-5
(mode 1-3) times as long as deep.

Mouth moderately oblique ; lower jaw sometimes projecting, its length 42-3-51-5
(M 46-8) per cent of head, 1-7-2-4 (mode 2-0) times as long as broad. Posterior
tip of the maxilla not reaching the vertical through the anterior orbital margin, but
approaching this point in a few specimens. The median dentigerous part of the
premaxilla slightly expanded anteroposteriorly.

Gill rakers moderately coarse, 7 (f.3), 8 (f.8), 9 (f.12) or 10 (f.i) on the lower part of
the first arch, the lowermost 1-3 rakers reduced.

Scales ctenoid ; lateral line with 30 (f.3), 31 (f.3), 32 (f.12) or 33 (f-7) scales, cheek
with 3 or 4 (mode) rows ; 6-8 scales between the lateral line and the dorsal fin origin,
6 or 7 (rarely 8) between the pectoral and pelvic fin bases.

FIG. 16. Haplochromis pellegrini, lectotype, .7 X N.S. (From Regan, Proc. zool. Soc. Lond.)

Fins. Dorsal with 23 (f.i), 24 (f.b), 25 (f.i6) or 26 (f.2) rays, comprising 14 (f.5),
15 (f.i3) or 16 (f.7) spinous and 9 (1.9), 10 (15) or n (f.i) branched rays. Anal with
3 spines and 8-10 (mode 9) branched rays. Pectoral fin 19-4-25-3 (M = 21-3) per cent
of standard length. First ray of pelvic fin slightly produced in both sexes. Caudal
truncate or subtruncate, scaled on its basal two-thirds.

Teeth. In most fishes <85 mm. S.L. the outer row is composed of weakly bicuspid
and unicuspid teeth or only weakly bicuspids ; in a few specimens this row may
contain only unicuspids. Fishes >85 mm. have only unicuspid teeth in the outer
row. All the outer teeth are slightly curved. The number of teeth in the upper jaw
(44-62) shows slight positive correlation with standard length.

In the majority of specimens, all the inner teeth are tricuspid, but in some there is
an admixture of tricuspids and weakly tricuspids. The inner teeth are arranged in 2
or 3 (rarely 4) rows in the upper jaw and 2 (rarely i) in the lower.

Osteology. The neurocranium is similar to that of H. bartoni and H. percoides and
does not closely approach the H. guiarti type. In other words it is of a type fairly
advanced along the " extreme " predator line of evolution.

Lower pharyngeal bone triangular, its dentigerous surface broader than long. The
teeth are fine and cuspidate, and are arranged in 22-24 rows.

ZOOL. 9, 4 n

i88 P. H. GREENWOOD

Vertebrae. Twenty-eight or 29 (mode 28) comprising 13 precaudal and 15 (1.5) or 16
(.3) caudal elements.

Coloration. Live females have a dark chocolate-brown ground colour, lighter on the
chest and belly. All the fins are blackish-brown. The live colours of males are
unknown.

Preserved material. Females. The ground colour is brownish-silver, grey on the
upper surface of the head and above the upper lateral line. There is a distinct (if
sometimes faint) lachrymal stripe and, in some specimens, traces of four incomplete
and very faint dark vertical bars on the flanks ; the two anterior bars are crossed
by a midlateral stripe originating behind the operculum but not extending beyond
the first pair of bars. Dorsal fin greyish-black, as is the greater part of the caudal
except for a narrow chevron of yellow-grey which extends forwards from the posterior
margin. Anal greyish, pelvics yellow-brown with a sooty area at their base.

Males are uniformly dark brown, almost black, but the branchiostegal membrane
is lighter ; the lachrymal stripe is very intense. The dorsal, caudal and anal fins
are dark grey to sooty, the anal with two small elongate and dark ocelli. The pelvics
are uniformly black except for a lighter innermost ray.

These notes are from specimens fixed in formol and preserved in alcohol. The two
type specimens (both females) were fixed in alcohol and are light brown above the
upper lateral line and bright silver on the flanks and belly. The dorsal fin is hyaline,
with brown maculae on the basal third of the soft part. The caudal is also hyaline
but is uniformly maculate. The anal and pelvic fins are colourless.

Distribution. Lake Victoria.

Ecology. Habitat. The material examined came from five localities all of which are
relatively exposed sandy beaches near dense stands of submerged and emergent
plants. The types came from Entebbe but no further details are known. If it is
assumed that they were caught in native fishing gear (and, with regard to the date
I think it reasonable to make this assumption) then the chances are greatly in favour
of their having come from a similar habitat. Because of its small adult size, I cannot
be certain that H. Pellegrini does not occur in other habitats where fishing was carried
out with large meshed gear (e.g. in many areas with a soft mud substrate) . However,
no specimens were caught in small-mesh trawls fished in these places.

Food. Of the thirty-five specimens examined, twenty-five had ingested material
in the gut. In three fishes, this material was an unidentifiable sludge, in one there
was sludge plus insect remains ; five had fed only on insects (principally Ephemerop-
teran pupae and larval Diptera) one on insects and small fishes, and fifteen had fed
exclusively on fishes. The fish remains were very fragmentary but in most cases were
identifiable as cichlids. The size range of the prey is 10-15 mm., that is immediately
post-larval fishes. From these records it seems that H. Pellegrini takes over from
those predators which feed on embryo and larval cichlids (Greenwood, 1959). There
appears to be some correlation between the diet of H. Pellegrini and its habitat
preference, since it occurs near areas where brooding cichlids congregate (e.g. reed
beds).

Breeding. The species is a female mouth brooder. All the specimens seen are adults
and both sexes reach the same maximum size.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 189

Affinities. The body-form of H. pellegrini is similar to that of the H. mento-H.
gowersi group and also H. percoides. The species is predominantly piscivorous but its
small adult size is unusual for this trophic group. Anatomically it belongs to the
group of moderately specialized fish-eaters, its neurocranium is fairly well advanced
towards the grade of specialization shown by H. estor and H. mento, and again there
is a strong resemblance to H. percoides. On skull and body form, H. pellegrini must
be allocated to the group of moderately specialized predators, but there are no
characters which allow one to assign it unequivocally to the H. estor complex or to the
H. percoides-flavipinnis group.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.).- 1906.5.30.253 . Entebbe . Degen

(Lectotype)

,, 1906.5.30.254 . Entebbe . Degen

(Paratype)

,, 1962.3.2.510-2 . Grant Bay . E.A.F.R.O.

,, 1962.3.2.513-5 . Beach in Buvuma Channel .

,, 1962.3.2.509 . Beach near Nasu Point . ,,

,, 1962.3.2.516-44 . Karenia (near Jinja) . ,,

Haplochromis percoides Blgr. 1915
Text-figs. 17 and 25

Haplochromis percoides (part) Blgr., 1915, Cat. Afr. Fish., 3, 296, fig. 201 (the two syntypes only).
Haplochromis flavipinnis (part) : Regan, 1922, Proc. zool. Soc. Lond., 172 (the two syntypes of
H. percoides only).

Lectotype. A male 79-0 mm. S.L., from Entebbe, B.M. (N.H.) Reg. No. 1906.5.30.

313.

Description based on twenty specimens 67-93 mm. S.L., including the lecto- and
paratype.

Depth of body 29-5-33-0 (M 30-9) per cent of standard length, length of head
34'3~38'i (M = 35-8) per cent. Dorsal head profile concave, the premaxillary pedicels
prominent.

Preorbital depth 16-7-19-5 (M = 17-7) per cent of head, least interorbital width
19-2-22-8 (M = 20-7). Snout slightly longer than broad or as long as broad, its length
31-4-35-0 (M = 34-7) per cent of head ; diameter of eye 21-9-26-6 (M = 24-2) per
cent, depth of cheek 21-4-27-4 (M = 25-0).

Caudal peduncle length 12-9-16-4 (M 15-3) per cent of standard length, 1-1-1-5
(modal range 1-4-1-5) times as long as deep.

Mouth slightly oblique, the jaws equal anteriorly or, more commonly, the lower
projecting slightly. Length of lower jaw 42-3-49-2 (M = 45-2) per cent of head,
1-7-2-2 times as long as broad. Posterior tip of the maxilla close to or, occasionally,
reaching the vertical through the anterior orbital margin.

igo P. H. GREENWOOD

Gill rakers usually slender, but in some specimens rather coarse and flat ; 7-9
(mode 8) on the lower part of the first arch.

Scales ctenoid ; lateral line with 30 (f.2), 31 (f.y), 32 (f.y), 33 (1.2) or 34 (f.i) scales,
cheek with 3 or 4 (mode) usually imbricating but occasionally irregular rows. Six to
8 (mode 7) scales between the lateral line and the dorsal origin, 7 or 8 (rarely 6 or 9)
between the pectoral and pelvic fin bases.

Fins. Dorsal with 22 (f.i), 23 (f.i), 24 (f.io), 25 (f.7) or 26 (f.i) rays, comprising
14 (f.2), 15 (f.12) or 16 (f.6) spines and 8 (1.4), 9 (f.io) or 10 (f.6) branched rays. Anal
with 12 or 13 rays comprising 3 spines and 9 or 10 branched rays. Pectoral fin
22-4-26-6 (M = 24-4) per cent of standard length. First pelvic ray but slightly
produced and then only in males. Caudal truncate, about the proximal half covered
with small scales (cf. H . flavipinnis) .

FIG. 17. Haplochromis percoides, lectotype -94 X N.S. (From Boulenger, Fish. Nile).

Teeth in the outer series are predominantly unicuspid and slightly curved, but in
some specimens of all sizes the lateral teeth are bi- or weakly bicuspid ; there are
44-56 (mean 50) outer teeth in the upper jaw. Except in one fish (91 mm. S.L.) the
inner series of both jaws are composed of tricuspid teeth ; in the exceptional specimen
all the inner teeth are unicuspid. There are 2 or 3 rows in the upper and I or 2 in the
lower jaw.

Osteology. In its general form, the neurocranium of H. percoides is near that of
H. pellegrini. The preotic face is long (67% of basal length) and its dorsal surface
slopes at a gentle angle (ca 25). The most noticeable difference is in the more depressed
orbital region of the H. percoides neurocranium (reflected in the gentler slope of the
dorsal surface). There is also a close resemblance between the neurocranium of
H. percoides and that of H. flavipinnis. Again, the orbital region in H. percoides is
more depressed.

The lower pharyngeal bone is triangular, its dentigerous surface broader than long.
The teeth are slender, compressed and cuspidate, albeit weakly so in some of the
anterior teeth. There are 18-22 rows of teeth.

Vertebrae. Twenty-nine, comprising 13 and 16 (f.7) or 12 and 17 (f.i) precaudal and
caudal elements.

Coloration. The live colours of H. percoides are unknown, but judging from the
well-marked vertical bands in all preserved specimens these must be conspicuous

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 191

features in life. No marked sexual dimorphism is shown in the coloration of pre-
served specimens, except that the pelvic fins in males are black. The ground colour
varies from silvery to dark brown ; there are five or six dark and broad transverse
bands (extending from the dorsal fin base almost to the ventral midline) on the flank
and caudal peduncle. Each of the two first and the two last bands are often linked
by a narrow midlateral bar ; the last band frequently has a midlateral tongue extend-
ing onto the caudal fin base. Usually there are two transverse and parallel bars
across the snout ; a lachrymal stripe of variable intensity is always present. The
dorsal, caudal and anal fins are yellowish-orange and without other markings. The
pectorals and pelvics are yellowish the latter variously dusky in males. Generally in
males there is only one, ill-defined greyish blotch in the position of an anal ocellus.

Distribution. Lake Victoria.

Ecology. Habitat. The species is apparently restricted to shallow water over
exposed or partly exposed sandy beaches. It is nowhere common but seems to be
widely distributed around the lakeshore.

Food. Fourteen of the twenty specimens examined contained food, and all the
identifiable remains were of fishes. In eight specimens the fish were small Haplochromis
(ca 12 mm. S.L.), and in three the identification could not be carried beyond that of
Cichlidae. In all the remaining specimens, the fish remains were too fragmentary
for further identification. It was noticed that in five specimens the stomach contained
four or five post-larval fishes of a similar size, thus suggesting that H. percoides may
prey on shoals of young cichlids.

Breeding. No information is available on the breeding habits of this species. The
sample studied consisted mainly of males and the only female showing signs of
gonadial activity was a fish 78 mm. S.L.

Affinities. Haplochromis percoides occupies a rather isolated position amongst the
Lake Victoria species, both with regard to its gross morphology and its colour pattern.
The most closely related species is H . flavipinnis . The colour pattern of both species
is virtually identical and both show a very concave upper head profile with pronounced
nuchal hump. In H . flavipinnis , however, the jaws are inclined more steeply and the
body is deeper. In details of neurocranial architecture, H. percoides shows a strong
resemblance to H. Pellegrini and it is conceivable that the latter species represents
a stem type from which H. percoides evolved. There is little doubt that H. flavipinnis
arose from an H. percoides-like ancestor.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda
B.M. (N.H.). 1906.5.30.313 . Entebbe Degen

(Holotype)

,, 1962.3.2.345 . Beach near Nasu Point . E.A.F.R.O.

,, 1962.3.2.346-53 . Entebbe (airport beach) . ,,

,, 1962.3.2.337-41 . Entebbe Harbour . ,,

Tanganyika

1962.3.2.342-4 . Beach near Ma jita . ,,

1962.3.2.332 . Majita Beach . ,,

iQ2 P. H. GREENWOOD

Haplochromis flavipinnis (Blgr.) 1906
Text-figs. 18 and 19

Pelmatochromis flavipinnis Blgr., 1906, Ann. Mag. nat. Hist., (7), 17, 441 ; Idem, 1907, Fish.

Nile, 448, pi. 89, fig. 3 ; Idem, 1915, Cat. Afr. Fish., 3, 418, fig. 286.
Haplochromis flavipinnis (part) : Regan, 1922, Proc. zool. Soc. Londn., 172 (the type specimen

only).

Holotype. A specimen 118-0 mm. S.L., from Bugonga ; B.M. (N.H.) Reg. No.
1906.5.30.308.

Description based on nineteen specimens (including the holotype), 69-156 mm.
S.L.

Depth of body increasing allometrically with standard length, 30-8-39-2 per cent
of the former. Length of head 33-6-38-0 (M = 35-8) per cent of standard length,
dorsal profile concave, the concavity becoming more marked with increasing size.

Preorbital depth 16-7-21-8 (M = 19-5) per cent of head, least interorbital width
19-4-23-6 (M = 21-1) per cent. Snout as long as broad in most specimens but occasion-
ally somewhat broader than long in fishes >I3O mm. S.L. ; its length 29-2-36-4

FIG. 18. Haplochromis flavipinnis, holotype, .9 X N.S. (From Boulenger, Fish. Nile.)

(M = 33-3) per cent of head. Diameter of eye 19-0-25-0 (M = 21-8) per cent, depth
of cheek (showing slight positive allometry) 25-0-32-7 per cent.

Caudal peduncle length 13-8-18-6 (M = 16-0) per cent of standard length, 1-2-1-6
(modal range 1-2-1-4) times as long as deep.

Mouth oblique (sloping at 4O-5o), the lower jaw projecting slightly or, infre-
quently, the jaws equal anteriorly. Length of lower jaw 43-8-51-0 (M = 47-8) per
cent of head, 1-6-2-0 times as long as broad. Posterior tip of the maxilla reaching or
nearly reaching the vertical through the anterior orbital margin.

Gill rakers short and stout in fishes <go mm. S.L., stout and flattened in larger
individuals, 8 (mode) or 9, rarely 7 or 10, on the lower part of the first arch.

Scales ctenoid, the lateral line with 30 (f.2), 31 (f-5) 32 (f.io) or 33 (f.2) scales, cheek
with 4 (mode) or 5, rarely 6 rows ; 7 or 7^ (less frequently 8) scales between the lateral

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 193

line and the dorsal origin, 7 or 8 (less frequently 6) between the pectoral and pelvic
fin bases.

Fins. Dorsal with 23 (f.i), 24 (f.12) or 25 (f.6) rays, comprising 15 (f.n) or 16 (f.8)
spines and 8 (.3) or 9 (f.i6) branched rays. Anal with n (f.8) or 12 (f.n) rays,
comprising 3 spines and 8 or 9 rays. Pectoral fin 23-2-29-6 (M = 25-6) per cent of
standard length. Caudal fin obliquely truncate, the posterior margin sloping down-
wards and forwards to meet the somewhat curved ventral margin. In outline this
fin resembles that of H. plagiostoma (see p. 200) but the lower rays do not have the
appearance of regenerated structures. The caudal is truncate in the two smallest
fishes examined (69 and 70 mm. S.L.) but is slightly oblique in a specimen 86 mm.
S.L. At all sizes, the caudal is scaled over about two-thirds to four-fifths of its length,
an unusually large area for a Lake Victoria species. The first pelvic ray is produced
in adult males.

Teeth. The outer row of teeth in both jaws is composed of relatively slender,
moderately curved unicuspids ; there are 44-66 (M = 55) teeth in the upper jaw,
the number showing a very slight positive correlation with size in fishes <I20 mm.
S.L.

31 mm

FIG. 19. Haplochromis flavipinnis, neurocranium.

The inner rows of teeth, in most individuals, are composed of unicuspids but in some
fishes <ioo mm. S.L. (and even in one of 150 mm.) these rows may contain a mixture
of tricuspids, weakly tricuspids and unicuspids. There are 2 or 3 (less commonly 4)
rows of teeth in the upper jaw and I or 2 rows in the lower.

Osteology. The neurocranium of H. flavipinnis resembles that of H. percoides in
outline except for the shape of the supraoccipital and the points of divergence (slope
of preorbital face, height in the interorbital region) are slight ; as in H. percoides, the
neurocranium has a long preotic part (67% of basal length).

Vertebrae. Twenty-nine or 30, comprising 13 and 16 (f.6), 12 and 17 (1.2) or 14 and
16 (f.i) precaudal and caudal elements.

Lower pharyngeal bone triangular, the dentigerous surface slightly broader than
long. The teeth are bicuspid and arranged in 18-20 rows, those of the two median
rows are somewhat enlarged. In small fishes the teeth are relatively slender but are
coarser in large individuals.

194 p - H - GREENWOOD

Coloration. The colours, and particularly the colour pattern, of H. flavipinnis are
characteristic, even in preserved material.

Live colours. Adult males. Ground colour light silvery-orange dorsally, becoming
greyish-black ventrally ; flanks and caudal peduncle crossed by four, broad, and dark
bars, the two first becoming continuous with the dark ventral coloration. The snout
and interorbital region are olivaceous, the cheek sooty ; anteriorly there is a broad-
based, triangular lachrymal stripe. The dorsal fin is light olive-yellow, with dark,
red-brown blotches between the soft rays. The caudal and anal fins are yellowish
with an overall reddish brown tinge ; the anal ocelli are yellowish-red. Pelvic fins
are black. Adult females have a light silver-yellow ground colour, but are dark, almost
olive on the dorsal surface of the head. The body is crossed by four broad and
irregular bars, the two anterior bars reaching further ventrally than the posterior
pair. Dorsal fin dark neutral with darker spots on the soft part. The distal two-thirds
of the caudal are dark, the proximal part is lighter. Anal fin yellow, the pelvics very
light yellow except along the anterior margin where there is a brownish tinge.

Preserved material. Sexual dimorphism is less marked in fixed material, except that
the chest and belly of sexually active males are dusky to black, and the entire pelvic
fin is black. The ground colour is variable (probably dependent both on the fish's
sexual condition if it is a male, and on preservation), from dark pinkish-brown to
faintly orange-silver ; four, broad, dark and slightly irregular bars cross the flank
and caudal peduncle, the two posterior bars do not extend much below the level of the
lower lateral line, the two anterior bars extend (usually) almost to the ventral surface.
The first and second bars are joined by a narrow midlateral stripe, as are the third
and fourth bars ; from the latter there is often a posteriorly directed tongue which
extends onto the base of the caudal fin. In many specimens the first three vertical
bars extend upward onto the base of the dorsal fin and even part way up the fin mem-
brane. The lachrymal stripe is a very characteristic feature, being a broad-based
triangle which extends over almost half the cheek ; the remainder of the cheek is
sometimes very dusky (perhaps a reflection of the fish's emotional state ?). The
posterior margin of the preoperculum is outlined by a narrow vertical bar which
merges dorsally with an anterior prolongation of the horizontal bar connecting the
first two transverse flank bars. Often an obliquely directed, chevron-shaped inter-
ocular band is visible on the nape. The dorsal, caudal and anal fins are orange-pink
or neutral, the soft dorsal maculate. In males there are three to five dead-white ocelli
on the anal ; if there are more than three ocelli, they are usually arranged in two
rows, the upper containing the greater number. In both sexes the pelvic spine and
the first to third rays are blackish ; in sexually active males the entire fin is black.

Distribution. Lake Victoria.

Ecology. Habitat. It is impossible to generalize on the habitat preferences of H.
flavipinnis from the data available. Specimens have been obtained from exposed
sandy beaches, sheltered bays and gulfs where the substrate is of organic mud, and
from rocky shelves running out from exposed islands. The only common factor in
each locality has been the depth of water : less than twenty-five feet. Since H. flavi-
pinnis is not abundant, nothing further could be determined.

Food. Ten of the eighteen specimens examined had food in the gut. Of these

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 195

specimens, eight had fed exclusively on fishes and two contained fragments of
Ephemeroptera in the stomach. The fish remains were too macerated for accurate
identification beyond familial or generic levels ; in five guts the fish were identified
as cichlids (and, since the scales were ctenoid, probably Haplochromis}. No specimen
had the remains of more than one fish in its stomach. The length of the prey species
is estimated as ranging from 25-35 mm. S.L.

Breeding. No data are available on breeding habits. The sexes reach maturity
at ca. 115 mm. S.L. and there is an indication that males may grow to a larger size
than females.

Affinities. The coloration, and particularly the colour pattern, of H . flavipinnis
serves as an immediate diagnostic character. It is not repeated in other Victoria
Haplochromis. However, the transverse baring of H. flavipinnis is related to that of
H. percoides (see p. 190). These two species are also related anatomically, although
the jaws of H. flavipinnis slope more steeply and the caudal fin is obliquely truncated.
Osteologically, the species show great similarity, particularly with regard to neuro-
cranial architecture. The few differences (apart from coloration and caudal fin shape)
which separate H. flavipinnis from H. percoides could be attributed to the larger
adult size of H. flavipinnis, especially since the differences involve characters which
often show differential growth rates.

Haplochromis flavipinnis also resembles H . cavifrons, at least anatomically, but
again there are differences in coloration, in this case more marked than those between
H. percoides and H. flavipinnis. The morphometric differences (e.g. the longer lower
jaw and wider interorbital of H. cavifrons) are not correlated with size.

Phylogenetically, it seems probable that H . flavipinnis was derived from a small,
inshore species like the present H. percoides, and that besides differences in coloration
the evolution involved a differential growth of certain mouth-parts and associated
neurocranial areas. That H. flavipinnis is less rigidly confined to a definite habitat
may suggest that it is in the process of becoming an off-shore, mud-bottom species.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1906.5.30.308 . Bugonga . Degen

(Holotype)

1962.3.2.143-4 . Ramafuta Isl. . E.A.F.R.O.

(Buvuma Channel)

1962.3.2.145-6 . Jinj a (below golf course)

1962.3.2.149-50 . Jinj a Pier

1962.3.2.154-5 . Napoleon Gulf, near Jinj a

1962.3.2.147 . Ekunu Bay

1 962 .3.2.151-3 . Entebbe Harbour

1962.3.2.156-8 . Pilkington Bay

Tanganyika

1962.3.2.148 . Majita Beach
Lake Victoria, Locality Unknown

1962.3.2.159

ig6

P. H. GREENWOOD

Haplochromis cavifrons (Hilgend.) 1888
Text-figs. 20 and 21

Paratilapia cavifrons Hilgendorf, 1888, Sitzb. Ges. not. Fr. Berlin, 77.
Pelmatochromis cavifrons: Blgr., 1915, Cat. Afr. Fish., 3, 419, fig. 287.
Haplochromis cavifrons : Regan, 1922, Proc. zool. Soc. Lond., 183.

Description based on forty-one specimens, 108-195 mm. S.L.; I have not examined
the type.

Depth of body 31-0-41-5 (M = 35-6) per cent of standard length, length of head
35-4-39-6 (M = 37-3) per cent. Dorsal head profile sloping at an angle of ca 30, its
outline sharply broken by the prominent premaxillary pedicels and thus appearing
concave.

FIG. 20. Haplochromis cavifrons, - 75 X N.S. (From Boulenger, Fish. Nile.}

Preorbital depth 18-3-21-5 (M = 20-2) per cent of head, least interorbital width
21-5-27-5 (M = 25-4) per cent. Snout as long as broad, or slightly broader than long,
occasionally somewhat longer than broad (the latter relationship is found most
frequently in fishes <I3O mm. S.L.) ; its length 34-0-40-0 (M = 36-5) per cent of
head ; diameter of eye 18-8-23-4 (M = 20-3) per cent, depth of cheek 27-2-35-4
(M =30-2) per cent.

Caudal peduncle length 12-7-17-5 (M = 15-0) per cent of standard length, 1-0-1-4
(mode 1-3) times as long as deep.

Mouth oblique (35-40), the jaws equal anteriorly, or the lower projecting slightly.
Lower jaw 49-3-60-5 (M = 55-5) per cent of head, its breadth contained 1-8-2-4
(mode 2-0) times in its length in fishes <I75 mm. S.L., and 1-4-2-0 (mode 1-6) times
in larger fishes. There is usually a broad mental projection developed at the mandi-
bular symphysis. Posterior tip of the maxilla not reaching the vertical through the
anterior orbital margin, and usually not covered by the preorbital.

Gill rakers short and stout 7 (rare) 10, mode 8, on the lower part of the first gill
arch, the lowermost one or two rakers reduced.

Scales ctenoid ; lateral line with 31 (f.6), 32 (f.i8), 33 (14) or 34 (f.i) scales (in one
specimen these scales are very irregularly arranged and give a count of 36) ; cheek

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 197

with 4-6 (rarely 3) rows ; 8-n scales between the lateral line and the dorsal origin,
8 or 9 (less frequently 7 or 10) between the pectoral and pelvic fin bases.

Fins. Dorsal with 23 (f.i), 24 (f.14) or 25 (f.25) rays, comprising 15 (f.22) or 16
(f.i8) spinous and 8 (1.4), 9 (f.27) or 10 (f.g) branched rays. Anal with n or 12 (rarely
13) rays, comprising 3 spines and 8, 9 (mode) or 10 branched rays. Pectoral 21-7-29-3
(M = 24-7) per cent of standard length. First pelvic ray slightly produced in adult
males. Caudal truncate, scaled except for the distal quarter.

Teeth. The outer row in both jaws is composed of unicuspid, fairly stout and slightly
curved teeth. There are 56-74 (M = 63) in this row of the upper jaw.

The inner rows in two of the five smallest specimens (108-129 rnm. S.L.) are made
up of uni- and tricuspid teeth, whilst in the three others and in fishes >I3O mm. S.L.
the inner teeth are all unicuspids. The inner rows are implanted obliquely, and there
are 2-4 (rarely i) rows in the upper jaw and I or 2 (occasionally 3) in the lower.

33 mm

FIG. 21. Haplochromis cavifrons, neurocranium.

Osteology. The neurocranium is very similar to that of H. flavipinnis ; that is, a
moderately advanced type with the preorbital part sloping less steeply and the
vomer not so markedly curved below the level of the parasphenoid as in the generalized
forms.

Vertebrae. Twenty-nine (f.4) or 30 (f.i), comprising 13 precaudal and 16 or 17
caudal elements.

Lower pharyngeal bone triangular, the dentigerous surface as long as broad or
slightly broader than long. The pharyngeal teeth are all basically bicuspid, but in
larger fishes the minor cusp is vestigial except in the teeth which occupy the dorso-
lateral corners of the bone. The teeth are somewhat irregularly arranged in about
twenty rows.

Coloration. The densely freckled appearance of H. cavifrons, both alive and dead,
is very characteristic. Live colours. Adult males. Ground colour olive to yellow-brown,
shading to silvery below, the head dark brown. Both the head and body are densely
speckled with brown spots and blotches. The dorsal fin is muddy yellow, sometimes
(? in sexually active fishes) with traces of deep red mottling over its entire length.
Caudal muddy-yellow with traces of deep red between the rays. Anal dark on its
proximal third to half, muddy-red to pink distally ; ocelli yellow. Pelvics are mottled

ig8 P. H. GREENWOOD

black and pink, becoming blacker in sexually active males. Females have a similar
coloration except that the pelvic and caudal fins are muddy yellow with dark blotches ;
the dorsal is like that of the male. Two aberrantly coloured females were caught near
Mwanza (Tanganyika) ; in these fishes the ground colour is a dark olive green shading
to light grey-green, the flecks darker green. The dorsal fin is mottled yellow-green,
as is the caudal. The anal and pelvic fins are dark yellow with a faint green tinge.

Preserved material. Males. The ground colour is brown, varying from light bronze
to a dark blackish-brown (in sexually active adults). The entire body and head are
densely peppered with darker spots and blotches. The dorsal fin is dusky brown,
marbled with darker pigment, especially on the spinous part. The caudal is brownish
and faintly maculate. Anal fin light orange-brown, darker basally ; ocelli faint,
represented by one to three whitish spots. The pelvics are black in sexually active
fishes, brownish in quiescent individuals.

Females have a similar coloration, but the ground colour is lighter and there are
often traces of six, irregular and rather narrow transverse bars on the flanks and
caudal peduncle. All fins, except the greyish pelvics, are as in males. One large
female (190 mm. S.L.), still sexually active, has five white ocelli arranged in two
rows on the anal fin.

Distribution. H. cavifrons is known definitely only from Lake Victoria ; one of
the specimens listed by Boulenger (1915) is from " Ripon Falls, Jinja " and may thus
be from the Victoria Nile if it was caught below the falls.

Ecology. Habitat. It seems that H. cavifrons is essentially a " hard substrate "
species since only two of the localities in which it was found had a mud substrate.
Because the species is rare in catches from beach-operated seines, but relatively
common in gill-nets set some hundred yards off-shore, it may be inferred that
H. cavifrons is not a member of the inshore community as are, for example, H . guiarti
and H. bayoni. No specimens were obtained from water more than 40 feet deep.

Food. Of the twenty-eight fishes (110-190 mm. S.L.) with ingested material in the
gut, twenty-two had fed on fishes alone, two on fishes and insects and three on insects
only. The insects eaten were : dragonfly larvae, chironomid larvae, and Povilla
(Ephemeroptera) egg masses and larvae. In the individuals that had fed on fishes,
the prey could be identified as follows : Haplochromis (f.i5) ; Cichlidae of indeter-
minable genus (f.5) ; Clarias (f.i) ; indeterminable (f.i).

Breeding. No information was obtained on the breeding habits of H. cavifrons.
Sexual maturity is reached in both sexes at lengths between 135 and 155 mm. S.L.,
and there is no dimorphism in the maximum size attained.

Affinities. No single character or combination of characters gives a clear-cut indica-
tion of the phyletic relationship of H. cavifrons. The coloration is unique and not
easily derived from any other type now existing in the lake. Anatomically, the general
suggestion is of relationships with H. flavipinnis, particularly with regard to head
shape and mouth form. This suggestion is supported by the form of the neurocranium
which is nearly identical in the two species. In turn, the same neurocranial characters
associate H. cavifrons with H. percoides, a more generalized species than H. flavipinnis.
But whereas H. percoides and H. flavipinnis have a related colour pattern, that of
H. cavifrons is completely distinct.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 199

It could be argued that similarity in neurocranial architecture is due to functional
convergence. Against this it can be shown that similar oblique mouths and large jaws
have been evolved in other species (e.g. H. plagiostoma} without producing a similar
neurocranial shape. In fact, the neurocranium of H. plagiostoma is readily distin-
guished from that of H . cavifrons. If on these grounds convergence can be overruled,
then H. cavifrons is possibly an isolated derivative from the H. percoides and H.flavi-
pinnis stem.

Study material and distribution records

Museum and Reg. No. Locality Collector

Uganda

B.M. (N.H.). 1906.5.30.203-7 . Bunjako . Degen

,, 1906.5.30.209 . Buganga . ,,

,, 1911.3.3.35 . Jinja, Ripon Falls . Degen

1962.3.2.358-65 . Jinja . E.A.F.R.O.

,, 1962.3.2.354-5 . Entebbe Harbour . ,,

,, 1962.3.2.366-75 . Ramafuta Isl. . ,,

and 380-6 (Buvuma Channel)

,, 1962.3.2.376-9 . Pilkington Bay . ,,

,, 1962.3.2.391 . Bukarra (Sesse Isls.) . ,,

Kenya

,, 1962.3.2.356 . Naia Bay, Kavirondo Gulf . ,,

Tanganyika

1928.5.24.449 . Mazinga Isl. . Graham

1962.3.2.357 . Mwanza .

Lake Victoria, Locality Unknown

1962.3.2.388-90 . . E.A.F.R.O.

Haplochromis plagiostoma Regan 1922
Text-figs. 22 and 25

Paratilapia longirostris (part) : Boulenger, 1915, Cat. Afr. Fish., 3, 332.
Haplochromis plagiostoma Regan, 1922, Proc. zool. Soc. Londn., 181, Text-fig. 8.

Holotype : a male 113-0 mm. S.L. from Bunjako, Uganda B.M. (N.H.) reg. no.
1906.5.30.261.

Description based on thirty specimens (including the holotype) 69-147 mm. S.L.

Depth of body 32-6-39-0 (M = 36-4) per cent of standard length, length of head
34-0-37-5 (M = 36-0) per cent. Dorsal profile of head gently concave or straight,
sloping at an angle of 2o-25.

Preorbital depth 18-0-21-5 (M = 19-8) per cent of head, least interorbital width
20-6-25-0 (M = 23-4) per cent. Snout slightly broader than long or as long as it is
broad, its length 28-2-35-5 (M = 32-5) per cent of the head ; depth of cheek 28-0-36-8
(M = 33-0) per cent.

Caudal peduncle 13-2-18-5 (M = 15-8) per cent of standard length and 1-0-1-5
(modal range 1-2-1-3) times as long as deep.

Mouth markedly oblique, directed upwards at an angle of 35-5o from the hori-
zontal : jaws equal anteriorly or, occasionally, the lower projecting slightly ; length

200 P. H. GREENWOOD

of lower jaw 44-0-54-5 (M = 49-2) per cent of head, 1-7-2-4 (mode 2-0) times as long
as broad.

Gill rakers stout but occasionally rather slender, 8 or 9 (mode), less frequently 10,
on the lower part of the first arch, the lowermost 1-3 rakers reduced.

Scales ctenoid, lateral line with 30 (f.8), 31 (f.5) 32 (f.12) or 33 (f.5) scales, cheek
with 4 (mode) or 5, rarely 3, rows of scales ; 5-6^ scales between the lateral line and
the dorsal origin, 6 or 7 (rarely 5 or 8) between the pectoral and pelvic fin bases.

Fins. Dorsal with 23 (f.4), 24 (f.i8) or 25 (f.8) rays, comprising 14 (f.4), 15 (1.19)
or 16 (f.7) spinous and 8 (f.6), 9 (f.i8) or 10 (f.6) branched rays. Anal with n, 12 or,
rarely, 13 rays comprising 3 spines and 8 (mode), 9 or 10 branched rays. Pectoral
fin shorter than the head, 24-6-31-0 (M = 28-1) per cent of standard length. Pelvics
with the first branched ray produced, proportionately longer in adult males. The

FIG. 22. Haplochromis plagiostoma, holotype, -63 X N.S. (From Regan, Proc. zool. Soc. Land.)

caudal fin of H. plagiostoma is obliquely truncate, the posterior margin sloping forward
and downwards to meet the upwardly curved ventral margin. The fin shape is
identical in both sexes and throughout the size range studied. The lower four or
five principal rays in most specimens have the appearance of rays regenerated after
damage ; commonly only one of the distal branches reaches the fin margin and
sometimes neither branch extends much beyond the dichotomy. At present I cannot
explain this peculiar fin shape, neither can I suggest whether it is due to some onto-
genetic disturbance or to some behavioural trait of the species resulting in damage to
the lower fin margin. Occasional specimens of other Haplochromis have an obliquely
truncate caudal but H. plagiostoma and H. flavipinnis are the only species in which
this peculiarity can be regarded as one of the specific characters.

Teeth. Throughout the size range studied, the outer row in both jaws is composed
of unicuspid, relatively stout and curved teeth, those of the upper jaw numbering
44-68 (M = 57) and somewhat more curved than the teeth of the lower jaw.

The inner rows are composed of unicuspids in all fishes >I20 mm. S.L. and in some
specimens in the range 70-120 mm.; other fishes in the latter size group have either all
tricuspids or a mixture of uni- and tricuspid teeth. The inner teeth are arranged in 2
(less frequently i) series in the upper jaw and I (less frequently 2) in the lower jaw.
All these teeth are implanted obliquely, the angle sometimes approaching the hori-
zontal.

REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 201

Osteology. The neurocranium of H. plagiostoma is of the generalized H. guiarti
type. Specifically, it differs from H. guiarti in having a slightly narrower interorbital
distance and a shorter preotic region. Despite the marked difference in jaw angle of
the two species (almost horizontal in H. guiarti and very oblique in H. plagiostoma}
the interspecific differences in syncranial proportions and morphology are slight. The
steep jaw angle is brought about by several minor changes. For example, the gentler
slope of the ethmoid region (over which the premaxillary pedicels slide) allows the
dentigerous area of the premaxilla to lie at a greater angle from the horizontal, thus
matching the increased slope of the dentary. The latter gains its slope from slight
changes in the articular surfaces of the quadrate and angular which allow the dentary
to move dorsally through a wider arc than is the case in H. guiarti. There is also a
compensatory change in neurocranial proportions. The preotic part of the skull in
H. plagiostoma is relatively shorter than in H. guiarti. This shortening of the anterior
part of the skull, correlated with a relatively longer dentary and more mobile hinge,
allows the dorsal (i.e. most anterior part of the gape) to be moved from a position
near the horizontal to one nearer the vertical. The lengthening of the lower jaw
permits this change without causing any alteration in the near vertical angle of the
hyomandibula.

The lower pharyngeal bone is triangular, its dentigerous area as long as broad or
slightly longer than broad. The teeth are coarse and cuspidate, those of the two
median rows being slightly enlarged ; there are 16-20 rows of teeth.

Vertebrae. Twenty-eight (f.i) or 29 (9), comprising 12 (f.2) or 13 (f.8) precaudal
and 15 (f.i), 16 (f.y) or 17 (f.2) caudal elements.

Coloration in life. Sexually active males are light blue-grey on the flanks and postero-
ventral surfaces, smokey grey dorsally and on the chest (which is darker than the
back). The head is smokey-grey with lighter blue flecks on the cheek and operculum ;
the branchiostegal membrane is sooty with a blue overlay. The dorsal fin is dark
grey with deep crimson spots between the posterior spines and over the entire soft
part ; the lappets and the margin of the soft dorsal are deep crimson. The caudal is
dark with a crimson margin and spots on the proximal half. The anal is deep red, the
ocelli (one to four arranged in a single row) yolk-yellow. The pelvics are black. In
quiescent males the ground colour is more silvery, the dorsal fin less intensely coloured
and the anal is hyaline except for a red flush which is most intense along the margin
of the fin.

Females are silver with a faint powder-blue sheen posteriorly on the flanks and
faint midlateral and dorsolateral stripes, the latter much shorter than the former.

Colour in preserved material. Females are brownish above, silvery below. There is
a broad midlateral band, sometimes of irregular thickness, running from behind the
operculum to the caudal origin ; a distinct but often faint lachrymal stripe is generally
present. All the fins are hyaline but the caudal is densely maculate. Sexually active
males are brownish above shading to a greyish copper on the flanks, and becoming
sooty on the chest and belly ; the branchiostegal membrane is black. There is often
a midlateral stripe but rarely are there any transverse bars ; a lachrymal stripe is
usually visible. The dorsal, caudal and anal fins are brownish and sometimes faintly
dusky. In some specimens there is a narrow black band on the dorsal fin, running

202 P. H. GREENWOOD

obliquely backwards and downwards from the lappet of the last spine to about the
middle of the last branched ray. The pelvics are black. The anal ocelli (1-4, usually 2)
are opaque and are arranged in a single row.

Distribution. Lake Victoria.

Ecology. Habitat. The species has been caught over both soft (organic mud) and
hard (sand and shingle) substrates. Common features of all localities are their shel-
tered nature (gulfs and bays) and relatively shallow depths (never more than 30 feet).

Food. Although H. plagiostoma is one of the commoner species, little is known
about its feeding habits. The reason for this is the high frequency of fishes without
food in the gut on capture. Those specimens with food (twenty-two in all) are
exclusively piscivorous, the prey species being predominantly Haplochromis and,
less often, small Cyprinidae (probably Engraulicypris argenteus). The two smallest
fishes examined (81 and 91 mm. S.L.) had fed on fishes ; the stomach of the smaller
individual contained two post-larval Haplochromis (S.L., ca. 8 mm.).

Breeding. A single female (107 mm. S.L.) has been recorded with eggs in the mouth ;
this specimen is also the smallest sexually mature individual in the collection. One
larger fish (115 mm. S.L.) is immature, but all other specimens 112 mm. S.L. and above
are adult. Females may reach a larg