esp, 


Agulhas Cape, 13 
Algoa Bay, 20 
Amanzimtoti, 37 
Bashee, 29 
Bazaruto Island, 51 
Beira, 52 
Bredasdorp Coast, 14 
Buffalo River, 26 
Bushmans River, 22 
Camps Bay, 8 

The Cape, 9 

Cape Agulhas, 13 
Cape of Good Hope, 9 
Cape Padrone, 21 
Cape Point, 9 

Cape Town, 7 
Chalumna River, 25 
Dassen Island, 6 
Delagoa Bay, 49 
Durban, 39 

East London, 26 
False Bay, 12 

Fish Point, 24 
Great Fish Point, 24 
Great Kei River, 27 
Inhaca, 47 
Inhambane, 50 
Isipingo, 38 

Kalk Bay, 10 

Kei Mouth, 27 

Kei River, 27 
Knysna, 17 

Kowie River, 23 
Kosi Bay, 45 


— 


OONOahWHh 


Swakopmund 
Walfish Bay 
Port Nolloth 
Lamberts Bay 
St. Helena Bay 
Saldanha Bay 
Dassen Island 
Cape Town 
Camps Bay 

The Cape 

Cape of Good Hope 
Cape Point 

Kalk Bay 
Simons Bay 
False Bay 

Cape Agulhas 
Bredasdorp Coast 
St. Sebastian Bay 
Mossel Bay 
Knysna 
Plettenberg Bay 
Tsitsikama 
Algoa Bay 

Port Elizabeth 
Cape Padrone 
Bushmans River 
Kowie River 
Port Alfred 
Great Fish Point 
Chalumna River 
Buffalo River 
East London 


SOUTHERN AFRICAN 


Kei Mouth 
Great Kei River 
Transkei 
Bashee 

Xora River 
Umtata River 
Umgazi River 
Port St. Johns 
Pondoland 
Port Shepstone 
Umkomaas River 
Amanzimtoti 
Isipingo 
Durban 
Umgeni River 
Umhlanga 
Tugela River 
Zululand 
Richards Bay 
St. Lucia Bay 
Kosi Bay 


Maputoland 49) 
Inhaca cceencee 


L 
Ponte Mahone M 
Ponte Maone 
Delagoa Bay 
Lourenco Marques 
Polana 
Inhambane 
Bazaruto Island 
Beira 
Zambezi River 


ALOR I, ACO R yy, — iglEteaey.. OOO Gy, 


Lamberts Bay, 3 
Lourenco Marques, 49 
Maputoland, 46 
Mossel Bay, 16 
Plettenberg Bay, 18 
Polana, 49 
Pondoland, 34 
Ponte Mahone, 48 
Ponte Maone, 48 
Port Alfred, 23 

Port Elizabeth, 20 
Port Nolloth, 2 

Port Shepstone, 35 
Port St. Johns, 33 
Richards Bay, 43 
St. Helena Bay, 4 
St. Lucia Bay, 44 
St. Sebastian Bay, 15 
Saldanha Bay, 5 
Simons Bay, 11 
Swakopmund, 1 
Transkei, 28 

Tugela River, 41 
Tsitsikama, 19 
Umgazi River, 32 
Umgeni River, 39 
Umhlanga, 40 
Umkomaas River, 36 
Umtata River, 31 
Walfish Bay, 1 

Xora River, 30. 
Zambezi River, 53 
Zululand, 42 


J. L. B. SMITH 


ICHTHYOLOGICAL PAPERS 
1931—1943 


J. LL.B. SMITH 
ICHTHYOLOGICAL PAPERS 


1931—1943 


Vol. 1 


Edited by 
Margaret M. Smith 


PUBLISHED BY THE J. L. B. SMITH INSTITUTE OF ICHTHYOLOGY 
RHODES UNIVERSITY . GRAHAMSTOWN 
1969 


Printed by Cape & Transvaal Printers Ltd., Cape Town. 


VOLUME 1 


Contents 
PAGE 


New and little known fish from the south and east coasts of Africa. Rec. Albany Mus. ] 
4 (1): 145-160 Pl. 16. January 1931 


An interesting new myctophid fish from South Africa. Trans. R. Soc. S. Afr. 21 (2): 17 
125-127 Pl. 9. April 1933 


The South African species of the genus Hemirhamphus Cuv. Trans. R. Soc. S. Afr. 21 
21 (2): 129-150 Pls. 10-12. April 1933 


The growth changes of Pteroplatea natalensis, G and T. Trans. R. Soc. S. Afr. 22 43 
(1): 83-87 Pl. 4. August 1934 


Marine fishes of seven genera new to South Africa. Trans. R. Soc. S. Afr. 22 (1): 49 
89-100 Pls. 5-6. August 1934 


The Triglidae of South Africa. Trans. R. Soc. S. Afr. 22 (4): 321-336 Pls. 16-23. 61 
December 1934 ' 


The fishes of the family Mugilidae in South Africa. Ann. S. Afr. Mus. 30 (5): 587-644 77 
Pls. 15-22. February 1935 


The “‘galjoen”’ fishes of South Africa. Trans. R. Soc. S. Afr. 23 (3): 265-276 Pls. 135 
13-17. September 1935 


New and little known fishes from South Africa. Rec. Albany Mus. 4 (2): 169-235 147 
Pls. 18-23. Mav 1935 


Fresh-water fishes of eastern Cape Province in A guide to the vertebrate fauna of the 215 
eastern Cape Province South Africa. Part II Reptiles, Amphibians, and Freshwater 
Fishes: 119-141 Pls. 29-34. 1937. Albany Museum, Grahamstown 


The South African species of the family Aluteridae. Rec. Albany Mus. 4 (2): 358-364 238 
Pls. 40-42. May 1935 


PAGE 


The genus Tripterodon Playfair. Trans. R. Soc. S. Afr. 23 (4): 303-310 Pls. 21-23. 245 
February 1936 


Two interesting new fishes from South Africa. Trans. R. Soc. S. Afr. 24 (1): 1-6 253 
Pls. 1-2. June 1936 


New gobioid and cyprinid fishes from South Africa. Trans. R. Soc. S. Afr. 24 (1): 259 
47-55 Pls. 3-5. June 1936 


New records of South African fishes. Ann. Natal Mus. 8 (2): 167-197 Pl. 11. May 269 
1937 


Acknowledgements 


I wish to express my appreciation: 


for the permission granted by the following to reproduce the papers in this volume: 


1. The Director, Albany Museum, Grahamstown; 

2. The Editor, Royal Society of South Africa, Cape Town; 
3. The Acting Director, Natal Museum, Pietermaritzburg; 
4. The Editor, Nature, London; 


to Miss M. Igglesden for many hours of work compiling the index; 
to the staff of the Department of Ichthyology for their encouragement and assistance. 


Margaret M. Smith, 
Grahamstown. 
June, 1968. 


JAMES LEONARD BRIERLEY SMITH was born in Graaff-Reinet, Cape Colony, on 
September 26, 1897, and died in Grahamstown on January 8, 1968. 

His love of fishes dated back to when, as a small boy at Knysna, he caught his first small 
silvery fish (Diplodus sargus). From then on angling and fishes became a passion that 
lasted until the end of his life. 

After an outstanding career in Chemistry, when by means of bursaries and scholarships, 
he put himself through Stellenbosch and Cambridge Universities, he returned in 1923 to 
lecture at Rhodes University College in Grahamstown. He was a brilliant, inspiring and 
much loved lecturer and teacher, and an indefatigable research worker. As such he came 
up against the “Old Guard” who had graduated to Chairs in the College from being 
school teachers, and looked askance at scientific research work. The turning point of his’ 
life probably came when he refused to spend his own precious time washing up students’ 
apparatus at the end of the year but offered to pay an unskilled person to do his share! 

When the Chair of Chemistry fell vacant, he was passed over. The only two members 
of the Senate who had doctorates, Professor S. Schonland, head of the Botany Depart- 
ment, and Professor J. E. Duerden, head of the Zoology Department, both fought bitterly 
to have Smith appointed to the Chair, but they were overruled, the excuse being that 
Organic Chemistry was a static subject of the past with no future! A Physical Chemist 
was appointed. Thus Chemistry’s loss proved ultimately to be Ichthyology’s gain. 

Time he would have spent on administration was given to research in Chemistry during 
official (and some unofficial) hours, and in Ichthyology during his spare time. 

He published his first short ichthyological paper in the Records of the Albany Museum 
in 1931, illustrating it himself with what he considered reasonable sketches. 

H. W. Parker, the herpetologist at the British Museum who was a student with him at 
Cambridge, wrote saying he was surprised to see a chemist publishing a paper on fishes. 
The text was quite good but the illustrations were terrible! For the next ichthyological 
paper he produced, he spent many arduous hours over the illustration which he redrew 
a number of times. The result (Myctophum (Nasolvchnus) florentii) was certainly worth all 
the trouble. He continued to illustrate his own papers until 1941. From 1943 with the 
Early Juvenile Stadia his wife, Margaret Mary, began illustrating his publications. Little 
did she realise where that was to lead her! | 

These two volumes represent the first phase of his ichthyological researches. During 
the war (1939-45) most of his time had of necessity to be given to teaching Chemistry, 
and at this period he published three Chemistry text-books. 

In September 1945 he was invited to produce a book on South African fishes. 

He had arrived at the parting of the ways. The newly formed South African Council 
for Scientific and Industrial Research made it possible for him to devote all his time to 
Ichthyology. He was given a grant of £800 p.a. to be renewed annually. It carried neither 
pension nor other benefits, and he lost his government pension due to him after 25 years 
of teaching because he refused to state that he was retiring from teaching on medical 
grounds. 

But to the end the Council for Scientific and Industrial Research which had taken over 
the old ‘‘Research Grant Board” that had assisted him with these early papers, supported 
his work in every way. 


In 1946 The Department of Ichthyology was founded and built up round his work. 
From 1946-1949 while busy writing The Sea Fishes of Southern Africa, he published 
seven more ichthyological papers in the Annals and Magazine of Natural History, 
London. These have not been included here as they are regarded as belonging to the 
second phase of his ichthyological work. Although A neutral solution of formaldehyde for 
biological purposes was published in 1947, it was read in 1944 and has been reprinted at 
the end. 

In two big monographs in volume 2, viz. The South African fishes of the families 
Sparidae and Denticidae and A living coelacanthid fish from South Africa, Smith’s personal 
copies have been used for reprinting and his own corrections have been reproduced. In 
the rest of the papers minor typographical errors have been corrected. 


1 


Records of the Albany Museum. Vol. IV. No. 1. pp. 145—160. 
Pl. XVI. January, 1931. 


New and little known fish from the south and east coasts 
of Africa. 


By J. L. B. SMITH, M.Sc., Ph.D. 
(With Plate XVI.) 


FAMILY GRAM MICOLEPIDAE. 
Prionolepis n.g. 


Body ovate, deep, very compressed. Anterior ridges on head 
and body strongly serrate. Scales small, embedded at right angles 
to the skin, with serrulate recurved upper edge. Whole of head 
and body scaly. Spinous and soft portions of dorsal and anal 
united; spines of dovsal, anal and ventrals stout, with serrated 
ridges. No spines on body at base of median fins, but alternate 
scales at base of dorsal and anal enlarged. Ventrals thoracic, 
reduced to a pair of spines and a few rudimentary rays. Mouth 
small, terminal horizontal. Gills 4, a slit behind the 4th. Gill 
membranes united to the isthmus. Gill rakers moderate, slender. 
Branchiostegals few, reduced. Pseudobranchiae absent. Lateral 
line obscure. Nostrils paired. 

This genus appears to be intermediate between Xenolepi- 
dichthys Gilchrist and Vesposus Jordan or somewhat closer to 
the latter, but is sharply distinguished from the other genera of 
the family by the character of the scales. This singular family 
now contains four monotypic genera, two of which are South 
African. 

Prionolepis hewitti n. sp. 

Body ovate, deep, very compressed. Dorsal profile elevated. 
Depth 13, length of head 3 in length of body. Maximum width 
of body at shoulder, 10 in length of body. Eye 2 in length of 
head, slightly greater than snout, twice interorbital width. Mouth 
small, terminal horizontal. Maxilla non-protractile, extends to 
almost below anterior nostril. A single series of small vertically 
implanted incisors in each jaw. Small teeth in bands on vomer. 
Dentigerous pharyngeals present. Nostrils paired, circular, close 


146 Smith—New South African Fish. 


together in a pentagonal depression before the orbit, edges of 
this depression serrate. Pre- and supra-orbital ridges serrate. 
2 serrated ridges from snout to base of spinous dorsal; similar 
ridges from the chin to the ventrals. 2 small ridges on snout, 
each with 3 very small recurved spines. Gills 4, a slit behind 
the 4th. Gill rakers moderate slender, about 13 on lower limb 
of anterior arch. Gill opening normal, membranes joined to the 
isthmus. Branchiostegals 3 much reduced. Pseudobranchiae 
absent. 

The scales are small, embedded at right angles to the skin, 
and have a denticulate recurved upper edge. The whole of the 
head and body is scaly. Lateral series 91., lateral tr. 14:18 
longitudinal series on cheek: lateral line tubules about 35. Lateral 
line gently curved, very obscure on posterior half of body. 
Alternate scales at bases of dorsal and anal much enlarged and 
sharply ridged. 

D V 29. Commences above hind margin of opercle. 1st spine 
longest slightly less than head, has 2 anterior and on each side 
two ridges, each with recurved spinelets. The remaining spines 
decrease in length to the 5th which is about half the length of 
the first. The 2nd—5th spines have on each side a ridge of 
recurved spinelets. The soft rays which are simple, uniform and 
somewhat shorter than the 5th dorsal spine, have on each side a 
row of recurved raylets. i 

A II 28. Commences below the 3rd dorsal spine. lst spine 
half the length of, and similar in structure to, the first dorsal 
spine. 2nd spine subequal and similar to the 5th dorsal spine. 
Rays slightly shorter than, but exactly similar to the soft dorsal 
rays. 

P 17, length of head, rays simple, base immediately below 
hind margin of opercle. 

V I (2 or 3), thoracic, rays rudimentary. Spines similar 
to and slightly shorter than the first anal spine. 

Caudal truncate, rays 26, similar in structure to dorsal and 
anal rays, each having a lateral row of recurved raylets on each 
side. Peduncle tapering posteriorly, subequal to eye. 


Prionolepis hewilti n. sp. X 2}. 


Below : scaling of right side, x about 7, 


148 Smith.—New South African Fish. 


Colour. Light yellowish; abdominal area and operculum dull 
silvery. A black blotch above the orbit, tapering to a point at 
the base of the spinous dorsal. A narrow black bar across the 
posterior part of the caudal peduncle. A black transverse band 
across the posterior third of the body, about half the depth at 
this point, and about one third wide as long. A narrow black 
irregular band below the posterior ? of soft dorsal and anal, 
following the dorsal and anal profiles. A few isolated angular 
dark spots on the body. Fins light yellow. 

Sex and condition cannot be determined without dissection, 
although the specimen is almost certainly juvenile. 

A single specimen, 40 mm. in length,* cast up during a storm 
at Great Fish Point. 


Type and only known specimen in the Albany Museum. 


*In all these descriptions, length of body excludes the caudal fin. 


FAMILY SOLEIDAE. 
Synaptura barnardi n.sp. (PI. XVI.) 


Dextral. Body lanceolate. Depth 23, length of head 4} 
in length of body. Eye 7 in head, 1? in snout, about twice inter- 
orbital width. Upper eye in advance of lower slightly more than 
half eye diameter. Lower and anterior border of lower eye with 
a filamentous fringe. Snout blunt, rounded, scarcely hooked. 
Cleft of mouth extends to below the centre of the lower eye: 
posterior margins of lips on coloured side fringed. Teeth very 
minute on jaws on blind side only. Anterior right nostril tubular: 
posterior with a fringed flap. Anterior left nostril surrounded 
by a fringed flap, developed behind covering a naked groove. 

Preopercle hidden below the skin. Opercular margin strongly 
fringed on blind side: lightly fringed on upper margin on 
coloured side. Opercular membrane joined to the upper margin 
of the base of the pectoral on the blind side. On the coloured 
side, the membrane is joined to the body well below the base of 
the pectoral and forms an extraneous fold which extends upwards 
obscuring more than half of the base of the pectoral, 


Smith—New South African Fish. 149 


D. 78. Commences on the snout in front of the lower margin 
of the upper eye: rays articulated, membrane slightly incised. 
Anterior rays short, graduated: middle anterior rays 33, posterior 
rays 4 in length of head. On the blind side, all the rays have a 
membranous fringe joined to the body at the base of the fin: 
fringe more pronounced anteriorly. Fin joined to the caudal. 

A. 57. Rays similar to the dorsal rays. 

Pectorals small but well developed: right pectoral 8 in length 
of head, left pectoral very slightly longer, about 7 in head. 

Scales small, strongly ctenoid on coloured side, a few on the 
head and along the lateral line bearing short filamentous process. 
11 about 110. Scales on the blind side rounded or very feebly 
ctenoid. Lateral line very gently curved, almost straight, extend- 
ing to above the preopercle on the right side. On the blind side, 
the lateral line is similar as far as the head, where it bifurcates, 
one branch turning down to near the corner of the mouth, the 
upper branch, at 180° to the lower, extending slightly obliquely 
back to near the base of the dorsal and then curving sharply 
forwards. 


Right side. Blind side. 
Scales of Synaptura barnardt n.sp. 


Caudal lanceolate, slightly rounded, 2 in length of head. 


Colour. On the right side the ground colour is light grey. 
Some very small isolated brownish spots on the head and body. 
Irregular very small white spots chiefly along the lateral line 
and on the anterior upper part of head and body. Median fins 
slightly darker at base, outer margins lighter: regular series of 


150 Smith—New South African Fish. 


light and of dark dots on the base of dorsal and anal. Caudal 
with close set dark spots, margin lighter, end brownish. Right 
pectoral light with reddish margin. Ventrals light. 


A single specimen 86 mm. in length from Great Fish Point, 
in very shallow water. 


Type in the Albany Museum. 


This species is very close to marginata Boulenger, but differs 
in the very small pectorals and in the presence of the curious fold 
in the opercular membrane on the right. side. 


It is noteworthy that many species of “Soles” are abundant 
in shallow water on the coast about Great Fish Point. The reefs 
of rock are very low at this point extending with very little fall 
some hundred yards below low water mark. Sand accumulates 
between the rocks, and fair numbers of “Soles,” up to a length 
of 500 mm., are obtained by wading out from knee to thigh deep 
in the water at low tide, and stabbing these patches of sand with 
a “spear,” resembling a pitchfork. The specimen described was 
secured in this manner. 


FAMILY SCORPIDIDAE. 
Neoscorpis n.g. 


Body oblong-ovate. Mouth small, terminal. The outer series 
of teeth in the jaws more or less enlarged and lanceolate. An 
upper and lower pair of dentigerous pharyngeals well developed. 
Minute teeth in bands on vomer and palatines, maxilla expanded 
posteriorly. Gill rakers moderate. Dorsal spines 6-8, in a deep 
groove, increasing in length posteriorly: soft portion longer than 
spinous, anterior rays subfalcate forming a prominent lobe. Anal 
spines graduated, soft rays similar to those of the dorsal. Anal 
dorsal and caudal scaly. Differs from Scorpis C. and V. in the 
absence of serrations on the preopercle, in the absence of lingual 
teeth, and in the smaller number of dorsal spines. 


A South African genus with one species, now recorded from 
Knysna to the Natal coast. 


Smith—New South African Fish. 151 


Neoscorpis lithophilus G. and T. 


Stone-fish (Natal), Butter-fish (Eastern Province). 
1908 Gilchrist and Thompson. Ann. 8. Afr. Mus. vol I, p. 162. 


1917 ibid. Ann. Durb. Mus. vol. I, pt. 4, 
p. 365. 
1925 Barnard. Ann. S. Afr. Mus. vol. XXI, p. 663, pl. 
XXVIII, fig. I. 


Depth 23 (Juv.) —2 (Ad.), length of head 34 (Juv.) —434 
(Ad.) in length of body. Eye 3—4 in length of head, in Juv. less 
than, in Ad. equal to snout; 1 (Juv.) —12 (Ad.) in interorbital 
width. Snout obtuse, interorbital prominent. Maxilla extends to 
below anterior border of eye. Teeth in bands in both jaws, outer 
row enlarged, lanceolate. Whole of head except interorbital space, 
- gnout and chin, scaly. Soft dorsal, soft anal and caudal scaly. 
Very fine naked groove at base of soft dorsal. Gill rakers 12-13 
on lower limb of anterior arch. Pyloric caeca very numerous and 
long. Sealy process in axil of ventrals ill defined in young exam- 
ples D VI—VIII, 20—22. Spines short and stout, increasing in 
size posteriorly. Anterior rays, almost length of head, subfalcate 
forming a prominent lobe. A III 23—26. Spines short and stout, 
graduated. 1st spine 3 and 8rd spine 1 in eye diameter. Anterior 
soft rays equal and similar to dorsal rays. Lateral line almost 
straight. Scales II 90—97, |. tr. we 

Colour. Silvery grey. In young often 7—8 faint dark cross 
bars equal to the interspaces. Interorbital darker with a broad 
light grey space on dorsal area. Ventrals light, remaining fins 
dark grey. Soft dorsal, soft anal and caudal with a black mar- 
ginal band about one-sixth of eye diameter. Membrane of dorsal 
spines black. Iris yellow. A semi-circular black mark on the 
hinder margin of the operculum. 

Locality. Knysna, Port Alfred, Great Fish Point, Isipingo, 
Natal coast. 

Distribution. South and south-east coasts of Africa. 

I have been fortunate in securing a well graduated series of 
specimens ranging in length from 44 to 468 mm. The number of 
dorsal spines appears to be commonly six, rarely seven or eight. 


152 Smith.—New South African Fish. 


This species is fairly common among rocks in shallow water; 
very young specimens appear to be most plentiful during the 
middle summer months. On the coast south of Natal, it is a 
somewhat rare capture on lines, but is frequently secured by net 
and spear at night with the aid of a light. It is esteemed as a 
food fish, the flesh being of fine texture and of excellent flavour. 
In April 1929, after a flood in the Great Fish River had caused 
considerable pollution of the sea for some miles west of the 
mouth, a number of specimens of this species was taken by hand 
in the shallow surf near the mouth of the river by a party of 
anglers. The specimens were alive but evidently suffering from 
partial suffocation induced by the amount of fine particles of 
clay in suspension in the sea water. At these times, the sea is 
coloured from red to deep brown, and relatively large numbers of 
immature fish of various species are cast up dead by the waves, 
while the larger specimens appear to desert temporarily the 
normal angling spots. 


FAMILY BLENNIIDAE. 
Blennius fascigula Barnard. (Pl. XVI.) 
‘Rocky’ (Eastern coasts). 
1925. Barnard. Ann. S.A. Museum, vol. XXI, p. 834. 


Depth 4—43, length of head 33—4 in length of body. Eye 
3$—4 in length of head, slightly greater than interorbital width, 
slightly less than snout. Profile of head very abruptly descending. 
Maxilla extends to below anterior third of the eye. Canines in 
both jaws, those in the lower jaw slightly larger: canines visible 
in even very small specimens. No occipital filaments. Nasal 
tentacles short, fimbriate. Supraorbital tentacles with a short 
flattened stalk, widening above and bearing 5—6 filaments, total 
length equal to the eye. Interorbital concave, groove extending 
to the snout.. A shallow transverse groove behind the orbits: 
another very shallow transverse groove before the base of the 
dorsal. Gill membranes united forming a fold across the throat. 
LL. anteriorly a double row of pores. 


Smith—New South African Fish. 153 


D. XII. 19—22, commences above the hinder margin of the 
preopercle, very slightly notched between the spinous and soft 
portions. Dorsal not joined to the caudal. 


A. II. 20-—-22. The two spines in the male enveloped in thick 
spongy skin. 1st spine obscure in females. Caudal rounded. 

Colour. Ground colour yellowish, 7 (Type 5) vertical cross 
bars about equal to the interspaces from the middle of the side 
extending on to the base of the dorsal. A variable longitudinal 
series of dark spots in pairs below the bars. Irregular dark spots 
on the lower part of the body. Small dark spots on the head, on 
the dorsal and on the base of the pectoral. Anal dark with 
projecting ends of the rays light. A dark tapering spot at the 
base of each alternate anal ray in some cases. Caudal light 
immaculate. A dark bar across the chin. Two dark bars across 
the throat; the hinder bar sometimes bifurcates on the side, and 
in some specimens a faint cross bar shows across the base of the 
ventrals. Usually a dark spot between the first and second dorsal 
spines. 

Length up to 85 mm. 

Locality. Cove Rock, East London. 

The species is founded upon a not too well preserved juvenile 
specimen from an unknown locality, and the original description 
is therefore necessarily somewhat inadequate. Neither the type 
nor my specimens show three clear cross bars across the throat, 
besides the one on the chin. A number of specimens from this 
locality which agree generally with this amended diagnosis of 
the species, show no sign of bifurcation of the hinder band on 
the throat, while in some cases the two bands have almost run 
into one large blotch. Pending the collection of further material 
these specimens are at present assigned to this species. The 
specimen figured is one in which the hinder band does not 
bifurcate. 

It may be remarked that specimens collected recently on 
the south-eastern coasts of Africa indicate that a revision of the 
African species of this genus is desirable, since certain diag- 
nostic features employed for the delimitation of species appear 
to be somewhat inconstant, and would, if maintained with 


154 Smith.—New South African Fish. 


material recently collected, involve an unnecessary multiplication 
of species. 

Further material is at present being collected with a view 
to such revision. 


FAMILY CLINIDAE. 
Clinus agilis n. sp. (Pl. XVI.) 
‘Klip-fish’ (South-west). ‘Rocky’ (Eastern coasts). 


Body moderately elongate, compressed. Dorsal profile gently 
curved from snout to base of anterior dorsal spines, with a slight 
hump above the hind margin of the preopercle. Depth 4—5 
length of head 34—4 in length of body. Eye 3—4 in head, 13 
times snout, almost twice interorbital width. 


Snout moderately sharp, sub-conical: very slight inter- 
orbital prominence. A slight groove from snout to interorbital. 
Two very shallow v-shaped grooves on occiput. Mucous pores on 
head prominent. Cleft of mouth oblique: jaws sub-equal: maxilla 
extends to below anterior third or centre of eye. A broad band 
of small teeth in both jaws, outer row enlarged: a curved band 
on vomer. A tentacle over the eye consisting of a flattish stalk 
with 2—38 cirri at end: length varies from 2—3 in eye. Minute 
nasal tentacles. 

D XXXIV—XXXV, 3—4. Commences above hind margin of 
preopercle. 1st, 3rd and 4th spines are shortest, subequal, 4—2 
of eye: 2nd spine 14—-14 times Ist. First three spines widely 
spaced, membrane deeply notched between 3rd and 4th spines. 
Remaining spines gradually increase in length, the last spine 
being 2—2i times the first. Membrane scarcely incised. An- 
terior ray slightly longer than last spine: remaining rays gradu- 
ated, last ray % of first in length, membrane joined to peduncle 
at base of caudal. 

A II, 21—24. 1st spine 3 of eye, 2nd spine 14 times as long 
as the first. Rays graduated, first ray 3, last ray 2 in length of 
head, ends project somewhat beyond membrane. 

P, 12, slightly less than head. 


V. I. 2-38, Inner (8rd) ray very small when present. 


11 


Smith—New South African Fish. 155 


Caudal rounded, of length of head. Peduncle slender, twice 
as long as deep. 

Colour (alive). Mottled grey-green. Seven irregular darker 
cross bands extending in some cases obliquely on to the base of 
the dorsal. Narrow dark band at base of caudal. Dark patch on 
nape. Head spotted. Dark spots on anal, pectoral and caudal, 
fewer on dorsal. Occasionally irregular reddish blotches on dorsal 
(males). Reddish tinge on posterior margin of opercle. Some- 
times a black spot on membrane between Ist and 2nd dorsal 
spines. 

Locality. Knysna estuary. The types are seven specimens in 
the Albany Museum collection ranging in length from 55 to 65 
mm. 

At Knysna this active and shy little fish lives in the sea- 
grass on the mud-banks of the river and is captured with 
difficulty. 

The species is very close to acuminatus C. and V., but differs 
chiefly in the dorsal fin formula and in the characters of the 
anterior portion of the spinous dorsal. 

Since our diagnosis of the species acuminatus is based 
largely on specimens collected from the Cape Peninsula, described 
by Gilchrist and Thompson (Ann. S.A. Mus. 1908, vol. VI, pt. 2, 
p. 124), it is possible that with collection over wider limits, the 
present species agilis may be shown to be a variety or sub-species 
of acuminatus. 

It may be noted here that I have found many species of 
Clinus hitherto recorded only from about the Cape Peninsula, 
occurring regularly along the coast as far east as East London: 
of these, cottoides C. and V., dorsalis Blkr, and superciliosus 
Linn. appear to be the most plentiful, while brachycephalus C. 
and V., and capensis C. and V. are fairly common, being 
especially plentiful in the rock-pools at and near East London: 
anguillaris C. and V., striatus G. and T., mus G. and T., pavo 
G. and T., brevicristatus G. and T., fucoruwm G. and T. and 
ornatus G. and T. have also been found, 


12 


156 Smith—New South African Fish. 


FAMILY SCORPAENIDAE. 
Amblyapistus marleyi Regan. 


Regan, Ann. Durban Mus., 1919, vol. II, pt. 4, p. 202. Text 
fig. 5. 

Barnard, Ann. S.A. Mus., 1925, vol. XXI, p. 917. 

Depth 22, length of head 4 in length of body. Eye 3 in head, 
twice interorbital and almost twice snout. Anterior profile of 
head almost vertical, concave. Maxilla extends to below the 
anterior third of the eye. Small villiform teeth on jaws, and in a 
crescentic band on vomer. Palate edentulous. Nostrils tubular, 
situated one above the other immediately before the lower margin 
of the orbit; anterior covered by a plain flap. Gills 4, no slit 
behind the fourth. Membranes free from the isthmus. Gill- 
rakers reduced to mere knobs, 7 or 8 on the lower part of the 
anterior arch. 

Preorbital produced backwards into a strong spine, reaching 
to below the posterior margin of the orbit; a smaller spine below 
at the base of the larger. Four preopercular spines, the lower 
three small and blunt, the upper larger, slightly less than the 
preorbital spine, reaching to the opercular margin. Two spines 
on the upper margin of the operculum, completely covered by 
skin. 

Lateral line almost straight, tubules 23. Rudimentary cycloid 
scales, not imbricated, below the skin on the anterior parts of 
the body, visible externally on the belly and peduncle. Skin 
smooth and soft. 

D XV, 8. lst spine short, subequal to eye, in advance of 
anterior margin of orbit. 2nd and 3rd spines longest, subequal, 
1} times length of head: bases of 2nd and 8rd spines apart by 
an eye diameter, 14 times as far apart as bases of 3rd and 4th 
spines. Membranes between 2nd and 8rd spines trapeziform, 
widening upwards. 6th spine shortest, less than half length of 
head, last spine 2 of 2nd spine. Membrane of soft dorsal joined 
to the peduncle at the base of the caudal. 5th ray longest. 

A III, 6. Commences below the 14th dorsal spine. Spines 
graduated, fairly stout. lst spine shorter than, 3rd spine almost 
twice, an eye diameter. Membrane of soft anal joined to the 


13 


Smith.—New South African Fish. 157 


peduncle an eye diameter away from the base of the caudal. 
3rd ray longest. P. 12. Rounded, reaches to above the 2nd anal 
spine. V. I. 5, reaches to the hinder margin of the vent. Caudal 
rounded, equal to head, rays 12. 

Colour. Brown, marbled and spotted with darker. Several 
dark blotches on the median fins. A light spot below the 9th 
dorsal spine, just above the lateral line. 

A single specimen, 147 mm. in length, from Mr, H. W. Bell 
Marley, Durban. 

Regan’s figure (loc. cit) shows neither the more or less 
trapeziform membrane between the 2nd and 8rd dorsal spines, 
nor the variation in the basal spacing of the anterior dorsal 
spines. Regan distinguishes this species from taentonotus, C. 
and V., by the greater length of the preorbital and upper pre- 
opercular spines. Barnard (loc. cit) quotes the latter species as 
having the 2nd and 3rd dorsal spines of the length of the 
head, and uses. this in his key to the South African species of 
the genus. Day (Fish of India, p. 157, pl. XX XVIII, fig. 5) 
quotes the 2nd and 3rd dorsal spines of taenionotus as being 
equal to the depth of the body, which his dimensions make to 
correspond with about 1+ times the length of the head. His 
figure (loc. cit) on the other hand shows these spines to be 
2 of the length of the head. In view of this discrepancy, a re- 
examination of the type of taenionotus appears desirable. 


FAMILY SYNANCIIDAE. 
Caracanthus zeylonicus Day. 

1869. Day. Proc. Zool. Soc, p. 515 (Amphiprionichthys 

zeylonicus ). | 

1878. Day. Fish of India, p. 158. Pl. XXXVIII, fig. 6. 

(Micropus zeylonicus) . 

Body deep compressed. Dorsal profile elevated, snout steep, 
blunt. Depth 2, length of head 21in length of body. Eye 4 in 
length of head, 14 times interorbital width, slightly less than 
snout. Interorbital space slightly convex, preorbital deep. 

Mouth moderate, terminal, somewhat oblique. Small villi- 
form teeth in bands on jaws only. Tongue small, adnate to floor 
of mouth, Maxilla extends to below anterior third of eye. Nostrils 


14 


158 Smith—New South African Fish. 


paired, tubular. Gills 4, no slit behind the fourth. Gill rakers 
moderate, slender, 13 on lower limb of anterior arch. Gill mem- 
branes united to the isthmus. 

Pseudobranchiae present. Branchiostegals 5. 

Preorbital produced into a laterally projecting spine, which 
is directed downwards and backwards. Preopercle with five 
spines, the upper pair small and blunt, the lower pair at the angle 
enlarged and recurved. A similar stout spine on the interopercle; 
subopercle with a blunt spine. One moderate and two very small 
blunt spines on the opercle. A finely serrated ridge runs from 
above the orbit along the nape. | 

Arborescent muciferous canals on the head above the pre- 
opercle. 

D VII, 18. Deeply notched between the spinous and soft 
portions. 1st spine minute, remaining spines increase in length 
to the 4th, which is 13 times the diameter of the eye, then 
decrease rapidly to the 7th, which is scarcely one third the 
diameter of the eye. Rays articulated, branched. The first two 
rays are subequal to the eye, the 3rd to the 9th almost twice as 
long, while the remainder are shorter than the first ray. 

A II, 11. Commences below the middle of the soft dorsal. 
Spines minute, not externally visible. Rays articulated, branched, 
subequal to eye. P. 13, rounded, rays articulated. Ventrals thor- 
acic, reduced to a pair of spines scarcely visible. 

Caudal rounded, rays articulated. Peduncle stout, deeper 
than long. Lateral line a continuous tube with 13 single tubules 
opening dorsally. The L. 1 follows the dorsal profile, and is 
obsolescent on the anterior part of the caudal peduncle. 

Scales absent, papillae on head and body, skin soft and 
velvety. 

Colour. Uniform dark brown, becoming lighter ventrally. 

A single specimen, 37 mm. in length, from Great Fish Point. 

This appears to be the first record from South Africa. 

The species is recorded by Day (loc. cit.) as occurring at 
Malabar and Ceylon, and its presence so far south is very inter- 
esting. Further search will doubtless reveal its presence on the 
eastern coasts of Africa. 


15 


Smith—New South African Fish. 159 


FAMILY TETRODONTIDAE, 
Tropidichthys oxylophius n. sp. (Pl. XVI.) 


Body deep, moderately compressed. Dorsal profile elevated. 
Snout concave. Profile strongly concave below chin. Back sharply 
ridged with apex slightly nearer tip of snout than base of caudal. 
Nostrils imperforate immediately before the orbit. Interorbital 
space strongly concave. Moderately prominent ant-orbital ridge. 
Depth 2, length of head 2 in length of body. Eye 34 in head, 
equal to the interorbital width, slightly less than snout. Eye 
midway between tip of snout and apex of ridge on back. Mouth 
small: upper teeth projecting beyond lower. Teeth subequal. 
Small 2-rooted spines which lie flat on the skin on most of head 
and body. Head spinulose except post-orbital and chin. From 
above and below centre of orbit, the anterior spines are directed 
forwards and upwards: the posterior spines are directed mostly 
backwards. Immediately before the orbit is a larger spine, 3 
times as long as the others, directed upwards. 

Most of the body is spinulose: no spines on the posterior 
third of the body, on the peduncle, on a small area behind the 
pectoral or on the top of the dorsal ridge from the apex to the 
base of the dorsal fin. On the lower part of the body, the roots 
of the spines are produced below the skin, giving this area a 
somewhat reticulate appearance. Skin rugose. 

D. 9, rays simple, well back, on posterior third of body. 

A. 9, rays simple, commences below posterior margin of base 
of dorsal. D. and A. subequal in length to diameter of eye. 

P. 18, rays simple, equal to snout. Caudal rays 9, branched. 
14 in length of head: subtruncate. : 

Colour. Uniform dark brown above, uniform light orange 
below. Slightly lighter ovate mark with light margin and with 
a short white oblique line below base of dorsal. Extremely fine 
dark lines running obliquely from eye to snout. Fins light yellow. 

A single specimen 32 mm. in length from Kareiga river 
mouth, near Port Alfred. 

Type in the Albany Museum. 

The occurrence of a species of this genus as far west as 
Port Alfred is rather interesting, no other species having been 


16 


160 Smith—New South African Fish. 


recorded south-west of Natal. It may be here noted that I have 
found many species hitherto regarded as confined to the eastern 
coasts of Africa extending regularly along the southern coast at 
least as far west as Knysna. 


The author desires to acknowledge gratefully financial 
assistance received from the Carnegie Grant in aid of Research, 
through the Research Grant Board of South Africa, which 
defrayed part of the expenses incurred in the collection of 
specimens. 


Rec.: Albany Mus. Vol. IV | Plate XVI 


Tropidichthys oxylophius n. sp. x 14. 
Blewnius fascigula Barnard x 75. 
Clinus agilis n. sp. x 17. 

Synaptura barnardi n. sp. X yo. 


17. 


Transactions of the Royal Society of South Africa. Vol. XXt. 
Part Il. pp. 125-127. Pl. IX. April, 1933. 


AN INTERESTING NEW MYCTOPHID FISH FROM SOUTH 
AFRICA. 


By J. L. B. Smira. 


(With Plate IX.) 


Family MycTopPHIDAE. 
Genus Myctophum Raf. 


The subdivision of the numerous polymorphous species of this genus has 
proved attractive, if somewhat troublesome. Goode and Bean (Ocean. 
Ich., 1895, p. 71) proposed an elaborate scheme, which has little practical 
or taxonomic value, since it is based mainly upon features which can have 
but little natural significance in these fishes. A simple and more natural 
scheme, based upon the nature and arrangement of the luminous organs, 
in the form in which it has been suggested by Brauer (Zool. Anz., 1904, 
vol. xxvili; Wiss. Erg. Deutsch. Tief-See Exp. Valdivia, 1906, vol. xv, 
pt. 1) has been accepted by most systematists. Brauer considers that 
there is but a single genus, Myctophum Raf., but recognises the four sub- 
genera Myctophum Raf. s.s., Lampanyctus Bon., Diaphus Hig & Hig, and 
Lampadena G. & B. - 

Since all species hitherto discovered have fallen more or less satisfactorily 
into one or other of the subgenera, these have acquired a possibly fictitious 
taxonomic significance, and Parr (Bull. Bing. Oc. Coll., 1928, vol. in, 
Art. 3, p. 49) has proposed the raising of the subgenera to full generic 
rank. This may prove to be of doubtful value, since, not only is our 
knowledge of these fishes far from complete, but renewed subdivision of 
Lampanyctus and Diaphus is almost certain to be proposed. 

The remarkable specimen described below does not fall within the 
limits of any one of these proposed four genera, but would appear to call 
in question the full generic distinction of Lampanyctus from Diraphus, 
since it to some extent bridges the gap between them, but also possesses 
certain features characteristic of neither. This would appear to involve 
either renewed subdivision of Diaphus, or the institution of yet another 
subgenus of the single wide genus Myctophum. The latter alternative 
appears preferable, and is here adopted. 


18 


126 Transactions of the Royal Society of South Africa. 


Nasolychnus, new subgenus. 


No luminous plates on peduncle. Luminous scales present or absent. 
Antorbital organs enlarged, probably always continuous above and below 
nostrils. Photophores circular, without dividing septum; five, or more 
or less than five, precaudal. The two infrapectoral organs probably never 
in line with the first pectoral organ. Spongy, probably luminous, tissue 
between the ventral bases, not corresponding with the scale pockets. 
Maxilla extending well behind eye, posteriorly dilated. 

This subgenus is closely related to Diaphus, from which, however, it 
differs in several features. 


Myctophum (Nasolychnus) florenti n. sp. 


Body elongate, moderately compressed, width 1-7 in depth. Dorsal 
profile even, snout blunt, almost vertical before eye. Depth 5-1, length 
of head 3-9 in length of body. Eye 4 in head, 1-8 times snout, 1-2 in 
interorbital width, and 2-6 in the postorbital part of the head. Strong 
median ridge on snout. Preopercle margin very oblique. Mouth large, 
maxilla extends 1-7 times eye diameter behind eye, posteriorly dilated. 
Minute teeth in both jaws and on palatines, pterygoids, and tongue: a few 
rudimentary teeth on sides of vomer. Gill rakers 25, slender, 2 in eye. 

D 19, commences above the ventrals, 5th ray longest, 1-7 in head, last 
ray 4 in head. Edge of fin straight. Base of dorsal as long as head. 
Adipose dorsal just behind end of anal base. 

A 19, commences below the 14th dorsal ray, 5th ray longest, 2:1 in 
head, last ray 6 in head, edge of fin gently concave. Base of anal 1-25 in 
length of head. 

P 12, inserted in lower third of side, 3-3 in head, does not reach to ventral 
base. V 9, inserted twice as far from base of caudal as from centre of eye, 
1-7 in head, reaches to vent. Caudal deeply forked, densely scaled. 

Photophores circular, without dividing septum. (Brauer’s terminology 
is hereemployed.) PLO much nearer lateral line than pectoral base. Upper 
PVO above pectoral base, lower behind upper. The PLO and the two 
PVO form an almost straight series with the 2nd (? 3rd, vide infra) PO. 
5 PO, first twice as far from second as remainder from one another (there 
may have been 6 PO, since on each side the scale behind the anterior PO 
is missing, see figure). VLO twice as far from ventral base as from lateral 
line. 5 VO in a gently curved series, 3rd highest. 4 SAO, the anterior 
three forming a gently curved oblique series with the posterior VO. Upper 
SAO in lateral line, forming a straight oblique series with posterior (4th) 
SAO and first antero-AO. 7+9 AO. 2 Pol, upper almost in lateral line, 
in a straight oblique series with last antero-AO. On one side 4+1 Pre, 


19 


An Interesting New Myctophid Fish from South Africa. 127 


separate from AO, last separate, highest, just below end of lateral line: 
on the other side there are 1 +2+1 Pre, the posterior as before, the remainder 
being arranged 1+2, with two rudiments between. Branchiostegal organs 
3, obscure. On one side 3 opercular organs, upper small ; lower two large, 
contiguous or partly fused, expanded ventrally. Upper organ absent on 
other side. Antorbital organs conspicuously enlarged, continuous above 
and below nostrils. A cuneiform patch of white, probably luminous tissue 
between ventral bases, about half eye diameter in length, not covered by 
scales and not corresponding with the scaling. No luminous plates. No 
true luminous scales, but most scales on the body and the top of the head 
with a central irregularly shaped light patch, probably luminous. Dorsal 
sheath scales mostly with central luminous patch. Edges of fontanel 
white, probably luminous. 

Scales cycloid, 1.1. tubes 46, 1.1. scales not enlarged, |.tr. 4 (from origin 
of dorsal), about 15 predorsal, 4-5 scales on cheek. 

Colour.—Uniform brown, antorbitals bright silvery. Photophores 
opalescent with black tissue before and behind. Proximal third of ventrals 
hight. 

Length.—136 mm. 

Type.—A ripe female, in the Albany Museum. 

Locality —Bushman’s River, near Port Alfred, found on the shore by 
Miss Florence Hewitt. 

The photophores on the cheek, and the Prec, as indicated above, are not 
the same on both sides. The 4+1 Pre is evidently normal, while the 
1+2+1 on the right side is most probably not, since there are two rudi- 
mentary or defective organs between the first and the second. 

It is an open question whether the photophores here described as the 
posterior (4th) SAO, and as the lower Pol, might not be regarded as the 
first and the ninth antero-AO respectively, elevated above the rest. Since 
the remaining AO are in two regular horizontal series, it appears preferable 
to regard these others as belonging to the SAO and to the Pol respectively. 

Considering that this remarkably large specimen was cast up by waves, 
although some of the scales are missing, it is in a remarkable state of 
preservation, so as to lead one to conclude that even a somewhat fragile 
dividing septum, such as is present in Diaphus, would still be intact in 
some of the photophores, had it originally been present. 

The eggs are about 0-6 mm. in diameter, and are certainly exceedingly 
numerous. 


ALBANY MUSEUM, 
GRAHAMSTOWN. 


‘ds -u s2zuasoyf (snuyohjosp ay) wnydopoh W 


Plate IX. 
Neill & Co., Ltd. 


tee 


Trans. Roy. Soc. 8. Afr., Vol. X XI. 


J. L. B. Smith. 


21. 


Transactions of the Royal Society of South Africa. Vol. XXI. 
Part Il. pp. 129-150. Pls. X—XII. April, 1933. 


THE SOUTH AFRICAN SPECIES OF THE GENUS 
HEMIRHAMPHUS CUV. 


By J. L. B. Smira. 
Genus Hemirhamphus Cuv. 
(With Plates X-XII and one Text-figure.) 


Cuvier, Reg. Anim., 1817, 11, p. 186. 

Body elongate, subcylindrical or compressed. Lower jaw produced 
into a long beak, with teeth only in the part opposite the edges of the 
premaxillae. A cutaneous fringe from corner of mouth on each side to 
apex of beak: a median bilobed sac expanded at origin below the chin, 
tapering forward to below tip of beak. Premaxillae, fused with maxillae, 
produced, forming a flattened triangular upper jaw, naked or scaly. Teeth 
small, conical, bi- or tricuspid, in bands in both jaws. Ventrals abdominal, » 
nearer to base of caudal than tip of upper jaw. Pectorals inserted above 
middle of side, not longer than head. Dorsal originates far back, much 
nearer base of caudal than base of pectoral. Caudal forked or emarginate, 
the lower lobe longer. Scales cycloid, moderate or large, deciduous. 
Lateral line ventro-lateral in position, commences on chest, ends before 
base of caudal: a superior branch to pectoral base. Lateral line tubes 
with few or many inferior tubules. Gull-openings wide, gill-rakers numerous, 
well developed. Third upper pharyngeals fused. Air-bladder simple or 
cellular. | 

Subgenera of Hemirhamphus Cuv. have several times been proposed. 
Gill (Proc. Ac. Nat. Sci. Phil., 1859, p. 131) instituted the genus Hyporham- 
phus for species with ventrals inserted nearer head than caudal base, and 
air-bladder simple. Fowler (‘‘ Fishes Oceania,’ Mem. B. P. Bishop Mus., 
1928, x, p. 75) has proposed the subgenus Rhynchorhamphus for species 
with triangular part of upper jaw longer than wide, beak very long, and 
ventrals nearer caudal base than head. To Hemirhamphus (s.s.) are then 
assigned species with ventrals nearer caudal base than head, upper jaw 
wider than long, and air-bladder cellular. | 

The position of the ventrals appears to be a most unsatisfactory feature 
‘on which to found an allied genus, or even a subgenus of Hemirhamphus, 


22 


130 Transactions of the Royal Society of South Africa. 


for in the species we find regular variation of this position from midway 
between caudal base and tip of pectoral to midway between caudal base 
and eye, several species having the ventrals inserted exactly midway 
between caudal base and head. The shape of the upper jaw and the 
length of the beak have little to recommend them as features for a similar 
purpose, since it is shown in the present work that these vary considerably 
with the growth of the fish. 

The South African species alone show the doubtful utility of the sub- 
genera as defined above: delagoae Brurd. has the simple air-bladder of 
Hyporhamphus, but the ventrals inserted as in Hemirhamphus ; knysnaensis 
n. sp. has the ventrals inserted midway between caudal base and head, 
thus falling midway between Hypohamphus and Hemirhamphus, the 
triangular part of the upper jaw in adults longer than wide, as in Rhyncho- 
rhamphus, but the beak is not very long, while the juvenile stages have the 
upper jaw wider than long, or as wide as long, according to the stage of 
growth. 

I cannot therefore accept the present forms in which the above sub- 
genera appear. 

Generic subdivision may more sharply and with greater justification 
be based upon other features, such as the presence or absence of scales on 
the triangular part of the upper Jaw, which is constant at all stages, or upon 
the nature of the growth changes observed during the early and mid- 
Juvenile stages. As will be seen below, these stadia of far Forsk. show 
primary increase in the relative length of the ventrals, followed by recession, 
together with very considerable changes in the shape of the dorsal fin. In 
knysnaensis n. sp., on the other hand, no such changes with growth are 
observed ; further, the triangular part of the upper jaw of knysnaensis is 
scaly, but naked in far. These two species may be regarded as sub- 
generically distinct, but until complete information on these points is 
available it would appear inadvisable to subdivide Hemirhamphus Cuv., 
and I do not here propose to attempt such a step, since it would serve no 
useful purpose as far as the South African species are concerned. 

The South African species have received somewhat scant attention, 
and stood in need of critical revision. -Two new species, knysnaensis and 
wmprovisus, are here described, and balinensis Blkr., which has not previously 
been recorded from our region, is now included. | 

Two species, marginatus Forsk. and schlegeli Blkr., which have been 
recorded from the neighbourhood of, but not strictly within the limits 
of, our region (after Barnard, Ann. S.A. Mus., 1925, xxi, p. 4, south of 
Mossamedes, west coast, and of Mozambique, east coast), are included, 
since with more intensive collection they are sure to be discovered here. 

In South Africa the species of Hemirhamphus are generally known as 


23 


The South African Species of the Genus Hemirhamphus Cuv. 131 


‘“ Naald-vis ” or “ Needle-fish,” also as ‘‘ Half-beaks.’’? They are of little 
or no economic significance, most of them being rather small and seldom 
taken in number. far is occasionally taken in moderate numbers by 
dragnets in the summer months, but is not generally in demand as a food- 


fish. 


Terms employed in the Descriptions. 


Since the beak and the caudal fin of preserved specimens are frequently 
damaged, the dimensional relationships are not here stated in total length 
only. The following limits are employed :— 

Length of Body.—Measured from caudal base (v.i.) to tip of triangular 
part of upper jaw. | 

Length of Head.—Measured from hind margin of operculum to tip of 
upper jaw. 

Length of Beak.—Measured from below tip of upper jaw to apex of 
beak. 

Base of Caudal.—This is taken as the base of the mid-caudal rays, 
generally obscured by scales. 

Lateral Rows of Scales.—Counted from above the hind margin of the 
operculum to caudal base, not including scales extending on to the caudal 
rays. 

Lateral Line.—F¥igures indicate the number of scales bearing tubules, 


and the count is taken as follows, e.g. 5007098’ which indicates 4 from 


pectoral base down to junction, 2 from chest to junction, 27 from junction 
to ventral base, and 28 posteriorly from ventral base. 


Key to the South African Species. 


A. Upper jaw naked. Ventrals nearer base of caudal than base of pectoral. 
1. Black blotches on side. Body subquadrangular in cross-section, not very 
compressed . . . . . . . . . . . far. 
2. No blotches on side. Body fairly compressed. . . . marginatus. 
B. Upper jaw scaly. Ventrals inserted midway between base of caudal and base of 
pectoral, or nearer the latter. 
1. Mid-caudal rays not longer than eye. Caudal deeply forked. Dorsal scaly. 
a. 50-52 lateral rows of scales. 
Pectoral and ventral not scaly . . . . . . dussumiert. 
b. 58-61 lateral rows of scales. 
Pectoral and ventral scaly. 
i. Beak 6:3 in total length. Preorbital 1-3 times longer than deep, not 
as long as eye . . . delagoae. 
ii. Beak 4-5 in total length. “Preorbital 1. 8-1: 9 times longer than deep, 
as long as eye . . . . . . .  balinensis. 


24 


132 Transactions of the Royal Society of South Africa. 


2. Mid-caudal rays longer than eye. Caudal emarginate. Dorsal not scaly. 
a. Ventrals nearer hind margin of preopercle than base of caudal. Upper jaw 


wider than long, with 3-4 transverse series of scales, middle series over 


median ridge improvisus. 


6. Ventrals nearer base of caudal than hind margin of preopercle. Upper 
jaw not wider than long, with 2 transverse series of scales, median 


ridge not covered by scales. 
i. Upper jaw longer than wide. 55-58 lateral rows of scales. South 


and east coasts . knysnaensis. 


ii. Upper jaw: as wide as long. 50-53 ‘lateral rows of scales. West 


coast schlegelt. 


Hemirhamphus far Forsk. 
(Plate XII, fig. 1.) 


1853. Bleeker, Nat. T¥dschr. Ned. Ind., v, p. 89 (fasciatus). 

1866. Bleeker, Atlas Ich., vi, p. 54, pl. vi, fig. 3. 

1878. Day, Fishes of India, p. 516, pl. cxx, fig. 3. 

1908. Gilchrist, Ann. S.A. Mus., vi, p. 266 (commerson1). 

1913. Weber, Siboga-Exp. Fische, p. 131 (fasczatus). 

1922. Weber and de Beaufort, Fishes Indo-Aus. Arch., iv, p. 157, fig. 55 
and p. 158 (marginatus, juvenile = fasciatus Blkr.). 

1925. Barnard, Ann. 8.A. Mus., xxi, p. 262. 


Adults. 


Body elongate, robust, width 1-3-1-5 in depth. Sides scarcely convex, 
form, with almost flat ventral surface, a right-angled ventro-lateral edge. 
Depth 6-8, length of head 4-3-5-3, in length of body. Lower jaw 
3-3°5 in body, 4-4-5 in total length. Eye 4-4-4 in head, 1-2-1-5 in 
snout, 1-6—1-8 in postorbital part of head, and 1-1-1-5 in interorbital width. 
Width of snout before eyes § of its length. Head slightly depressed, 
interorbital gently convex, with 2 shallow longitudinal grooves. Top of 
head scaly to above nostrils. A slight medio-longitudinal ridge on the 
triangular part of the upper jaw, which latter is thick and fleshy, scaleless, 
and 1-3-1-9 times as wide as long. Width of lower jaw below tip of snout 
almost an eye diameter. Fairly stout tricuspid teeth in 3-4 rows in lower 
and 2-3 rows in upper jaw. Beak stout, cutaneous fringes about half the 
width of the beak: subgular sac prominent. Preorbital 1-3 times deeper 
than long, depth about half-eye diameter; covered by a single scale. 
Opercle with one large and two smaller scales. A single series of scales 
on the cheek, extending forward to below the corner of the mouth. 

Gill-rakers 20-23, fairly stout, 3-5 in gill-filaments which are 1-5 in eye. 
Branchiostegals 12-13, membrane and mentum scaleless. 


25 


The South African Species of the Genus Hemirhamphus Cuv. 133 


D 13-14, arises twice as far from base of caudal as from base of ventral. 
Anterior rays elevated, 3rd longest 1-7-1-9 in head, remainder decrease 
rapidly to the penultimate, which is about 9 in head, the fin having a 
deeply concave edge. The last ray is elongate, about 5 in head, with hinder 
branch free from membrane. Anterior. fourth of fin densely scaled. 

A 10-13, commences below the 6th or 7th dorsal ray. Anterior rays 
elevated, 3rd longest 2-5 in head, remainder decrease to the penultimate, 
which is about 9 in head, the fin having a gently concave edge. The 
hinder branch of the last ray is slightly longer than the preceding rays. 
Anterior third of fin densely scaled. Base of anal 1-6-2 in length of base 
of dorsal (rarely more than 1:8). 

P 12, 1-3-1-7 in head, scaleless. 

V 6, scaleless, 2-2—2-8 in head, inserted midway between base of caudal 
and almost the tip to distal fourth of pectoral. 1st and last rays longest, 
subequal, inner four rays graduated, shorter than Ist and last; last ray 
about 1-3 times the length of the penultimate. 

Caudal deeply forked, upper lobe shorter, 1-3-1-4 in length of lower ; 
mid rays shorter than or equal to eye. 

Peduncle fairly robust. 

Air-bladder cellular. 


Scales.—Lateral rows 55-58, 1.1. , predorsal 37-39, 


3 +26 — 27 +21 — 22 
6-7 between dorsal and lateral line. Lateral line tubes highly arborescent 
ventrally (Plate X, A), more elaborated in anterior scales. 

Colour.—Blue-green above, silvery below, dorsal scales light, with dark 
margin; opercle bronzy. Beak and membranes black, tip of beak red. 
A silvery-plumbeous stripe with blue upper margin from base of caudal 
along side curving gently over pectoral base and ending on shoulder: 
widest midway. 4-10 more or less distinct dusky blotches on side chiefly 
above lateral stripe. Dorsal, most of caudal, anterior part of anal, and 
ventrals, except last ray, dusky. Pectorals light. Ripe fishes usually have 
reddish blotches on the abdomen. 

Length.—200-533 mm. 

Locality. Eastwards from False Bay, entering tidal rivers from Breede 
River, Port Beaufort to the Zambesi River, Rhodesia, occurring even in the 
Mazoe River, a tributary of the Zambesi, in fresh water miles from the coast. 

Distribution.—Almost circumtropical. 


Juveniles. 


. Very young specimens show considerable divergence from the adult 
form chiefly in the size and shape of the fins and of the jaws. The following 


26 


134 Transactions of the Royal Society of South Africa. 
table indicates certain of the features of change with the growth of 
the fish :— 
Ee nnn Tn inteal SanEEEEE SERaEISDT SEnSnissAttesn nnn ea a SS iE: as 
Total length (mm.) . 12| 18 | 24 | 33 | 53 | 60 | 72 | 115) 135] 195) 295 | 533 
Depth in body . .|>10|>10/ 10 |10}9 |9 {9 {84)8 18 | 7 | 65 
Head in body . . lea. 4 loa. 4 | 4:5 | 4:5) 4:5] 4:5 | 4:8 | 4-8] 4-8) 4:3 | 4-5 | 4-9 
Beak in total length . |None |ca. 20] 9 6 |5 |45] 48144) 44] 4:3] 461 5-0 
Triangular upper jaw, 

wide as long . . | .. lea. 7 jea. 7 Ica. Tica. Sica. 5) 4 | 2-5 | 2-5] 2-0] 1-4} 1-4 
Eye in head . . lea. 2 |ca.2-5} 3 13 | 3-2} 3:3/3-3)3-5/3-7)4 | 4 | 43 
Eye in snout. . .. |cea.0-4/ca.0°5| 0-6 | 0-7 | 0-8 | O-B | 1-1] 1-1] Ld | 14] 15 
Eye in postorbital part 

head. . . ca.1-3) 1-5 | 1-6} 1-5 | 1-7) 1-7] 1-7) 1-7) 1-7] 1-7) 1:8 
Pectoral in head . jea.l-5jca.1-5} 15 | 1:5) 1-4) 1-4) 1-4] 1-4] 1-4] 1-6) 1-4) 1-7 
Ventral in head . . joa. 4 joa. 3 | 2:0 | 1-8) 1-6] 1-6] 1-7] 1-5] 1-6] 2:0) 2-4] 2:8 
Upper caudal lobe in 

lower. . . jea.l-O} 1-1 | 1-2 | 1:6] 1-8] 1-6] 1-6) 1-7) 1:5] 1-4) 1-4) 1:3 
Shortest caudal rays 

ineye . . . |ca.0-5| 0-7 | 0-7 | 0-8] 0-8 | 0-8 | 0-9 | 0-9} 0-9] 1:0) 1-1) 11 


Features which do not change appreciably with growth are the shape 
of the anal fin and the relative size and shape of the pectorals. 


The Mouth and Jaws. 


There is no prolongation of the lower jaw in fishes not exceeding 15 mm. 
in length: the adult proportions are attained when the total length is 
50-60 mm. Fishes 100-200 mm. in length have relatively longer beaks 
than larger examples. (See above table.) The mouth is at first almost 
vertical and changes to nearly horizontal with the growth of the upper jaw, 
which commences only when the lower jaw is fully prolonged. Very young 
fishes have a single row of teeth, many of which are conical, though tricuspid 
teeth are observed at all stages: the number of rows increases with age. 


Ventral Fins. 


In very small fishes the ventrals are inserted midway between the base 
of the caudal and the middle to distal third of the pectoral. The point 
of insertion moves somewhat towards the caudal with growth of the fish. 
The ventrals are fairly small in the youngest stadia, increase in relative 
length until the fish is about 100 mm. in length, and thereafter decrease. 
In specimens 50-120 mm. in length the ventrals reach the base of the anal 
or beyond. The shape of the fin also undergoes modification: in fishes 
not longer than 60 mm. they resemble in shape that of the ventrals in an 
adult of the Exocoetid genus Cypsilurus Swnsn.: the 3rd ray longest, 
Ist shortest, 6th intermediate between lst and 3rd, remainder graduated : 
the distal third of the fin being subacutely triangular in shape. With 


27 


The South African Species of the Genus Hemirhamphus Cuv. 135 


growth of the fish, the Ist, 5th, and 6th rays increase in relative length, 
while the 3rd becomes shorter, so that when the fish is about 200 mm. in 
length, the 5th ray is longest, the 4th and 6th slightly shorter, subequal, 
Ist ray shortest. Further growth results in the change to the adult form | 
in which the Ist and 6th rays are longest, subequal, the 2nd ray shortest, 
edge of fin concave. 

Dorsal Fin. 


When the length of the fish does not exceed 50 mm. the posterior 
dorsal rays are about twice as long as the anterior, and when laid back 
reach on to the caudal: the edge of the fin is straight. The relative length 
of the rays changes with growth: at about 100 mm. total length the 
anterior and posterior rays are subequal, the mid rays shorter, the anterior 
margin of the fin being concave: at about 150 mm. the anterior rays are 
1-5 times the posterior, and the anterior part of the fin has the subfalcate 
shape of that of the adult. The posterior rays laid back no longer reach 
to the caudal. When the fish is about 200 mm. in length the shape of the 
fin approximates to that of the adult. 


Coloration. 


Very small fishes (35 mm.) are dusky green, with dendritic specks 
uniformly distributed over the body, approximating anteriorly to the 
scaling. Beyond this stage faint cross-bars appear, and with growth 
increase in density. In fishes exceeding 50 mm. in length there are 10 
cross-bars on the body, one across the base of the caudal rays and a wide 
blotch over the distal portion of the lower caudal lobe: the caudal bars 
and the ventral portion of the bars on the side are rust-red in colour. 
The cross-bars on the body are prominent in fishes of 60-200 mm. in length, 
the ventral portion, and the caudal bars fading with growth. 

In all but the smallest specimens the ventrals are black, the pigmentation 
retreating somewhat distally with growth, while the first ray becomes light. 
The dorsal and anal are at first almost wholly dark, the pigmented area 
retreating distally with growth, until in larger specimens the anterior 
part of each fin and a broad marginal band only are dark. The pectorals 
are uniformly unpigmented. The lateral stripe is not visible in specimens 
of less than 30 mm. total length. Fishes 50-100 mm. in length frequently 
have dark blue spots on the head. Beak and membranes black: tip of 
beak red in advance of mid-juvenile stage. 

H. fasciatus Blkr. is undoubtedly the juvenile form of far. Weber and 
de Beaufort’s diagnosis (loc. cit.) of fasciatus as juvenile margenatus Forsk. 
was evidently based upon the relative lengths of the bases of the anal 
and dorsal fins, which these authors state to be in far base anal 2 in base 

VOL. XXI, PART II. 10 


28 


136 Transactions of the Royal Society of South Africa. 


dorsal, and 1-5-1-7 in marginatus. In our specimens of far, the base 
of the anal is usually 1-6-1-8 in the length of the base of the dorsal, 
occasional specimens showing 1-8-2. Our specimens show also further 
divergence from the usual diagnosis of far: the last ray of the ventrals 
and of the dorsal are elongate, longer than the penultimate ; the number 
of scales in lateral series and the number of predorsal scales are greater : 
the anterior portion of the dorsal is more subfalcate in shape, while the 
body is not as deep nor the head as long as in specimens from other areas. 
It may be noted that the figures of Day (loc. cit.) and of Weber and de 
Beaufort (loc. cit.) are not strictly in accordance with the respective 
descriptions, while none of our specimens have the markedly fusiform 
body indicated in these figures, but are deepest across the middle of the 
pectorals, even fully ripe fishes. 

I have examined several specimens from the Indo-Pacific and find 
that they have the base of the anal 1-8-2 in length of dorsal base and 
51-55 lateral rows of scales. From the nature of the other features men- 
tioned above, it would appear that most descriptions of far have been 
based upon old, preserved, and possibly damaged specimens. 


Breeding Habits, Etc. 


In the Knysna and neighbouring rivers specimens of far are first 
observed early in October, all ripe adults. The numbers increase rapidly, 
attaining a maximum in December and January. Ripe and spent adults 
are found until March, after which only a very occasional specimen is 
encountered, and none are present after the end of March. The young 
appear in November, become exceedingly abundant by January, and 
disappear entirely by March. 

The eggs of far are demersal, 2-95-3-1 mm. in diameter, translucent, 
and provided with numerous glutinous filaments not longer than the 
diameter of the egg. The larger females produce about 12,000 eggs per 
season at Knysna. The breeding fishes, which appear to travel even 
beyond the tidal area in the rivers, congregate in shallow water over 
grassy banks, where two or more may frequently be seen slowly circling 
over a restricted area. In one case, two specimens so engaged were 
captured, and on examination, one, the larger, proved to be a ripe female, 
and the other a ripe male. | 

H. far may also be taken by allowing a fine line, baited with a small 
“shrimp ” or “ mud-prawn,” to drift in the surface of the water: these 
larger specimens are exceedingly lively and provide most excellent sport 
on light tackle. The flesh is of fine texture and flavour, but is not gener- 
ally esteemed on account of the presence of numerous fine bones. 


29 


The South African Species of the Genus Hemirhamphus Cuv. 137 


Hemirhamphus marginatus Forsk. 


1775. Forskal, Descr. Anim., p. 67. 

1922. Weber and de Beaufort, Fishes Indo-Aus. Archip., iv, p. 157. 

1928. Fowler, “ Fishes Oceania,’ Mem. B. P. Bishop Mus,, x, p. 78 
(record only). 

Body fairly compressed, width 2-3 in depth. Depth 5-3-6-5, length 
of head 3-7-4-2, in length of body. Lower jaw 3-3-3-5 in body, 4:5-5-1 in 
total length. Eye 3-8 in head, 1-3 in snout, 1-4-1-7 in postorbital part of 
head, and equal to or slightly greater than interorbital width. Head 
slightly depressed, interorbital gently convex, with 2 shallow longitudinal 

grooves. A slight medio-longitudinal ridge on the triangular part of the 

upper jaw, which latter is more convex above than in other species, and 
1-3-1-5 times as wide as long, scaleless. Width of lower jaw below tip 
of snout slightly more than half an eye diameter. Tricuspid teeth in 
2 rows in upper and 3-4 rows in lower jaw. Beak slender. Preorbital 
1-2 times longer than deep, # eye diameter in length. Single series of 
scales on cheek. 

Gill-rakers 23, 2-3 in gill-filaments, which are 1-6 in eye. 

D 13-14, arises 1-5-2 times as far from base of caudal as from base of 
ventral. Anterior rays elevated, about 1-9 in head, edge of fin deeply 
concave. Densely scaled anteriorly. | 

A 10-13, commences below the 6th dorsal ray. Anterior rays elevated, 
about 3 in head, edge of fin concave, last ray elongate. Scaly anteriorly. 
Base of anal 1-5-1-7 in length of base of dorsal. 

P 11, 1-2—1-3 in head, scaleless. 

V 6, scaleless, 2-6 in head, inserted midway between base of caudal 
and distal third of pectoral, first and last rays longest, subequal. 

Caudal deeply forked, upper lobe shorter, 1-3 in length of lower; mid 
rays shorter than eye. Air-bladder at least partly cellular. 

3-4 

Scales.—Lateral rows 53-56, 1.1. 5349799. LL? predorsal 34-38, 
6 between dorsal and lateral line. Lateral line tubes arborescent ven- 
trally. Arborescent muciferous canals on scales of head in some specimens. 

Colour.—Blue-green above, silvery below. Plumbeous lateral stripe 
curving up over pectoral base. Beak and membranes black. Dorsal 
dusky, dark anteriorly. Caudal dusky, darker marginally. Pectorals 
and ventrals light. 

Length.—Up to 350 mm. 

Locality.—Gold Coast. 

Distribution.—Almost circumtropical. 

This species has not yet been recorded from within our area, but is 


30 


138 Transactions of the Royal Society of South Africa. 


sure to be discovered on our eastern coast. I have examined specimens 
from the Indo-Pacific, and one specimen, loaned by the British Museum, 
from Accra. It is very closely related to far, from which it differs in the 
absence of the lateral blotches as well as in the highly compressed body, 
in the narrower interorbital, and in the fairly convex upper surface of the 
upper Jaw. 

Hemirhamphus dussumeri C. and V. 


1846. Cuvier and Valenciennes, Hist. Nat. Poiss., xix, p. 33. 
1925. Fowler, Proc. Ac. Nat. Sci. Phil., lxxvu, p. 203. 
1925. Barnard, Ann. 8.A. Mus., xxi, p. 263. 


Body elongate, subcylindrical. Depth 9, head 4:5, in length of body. 
(Lower jaw 4-5-5 in body, 5-6-3 in total length.) Eye 4 in head, 1-5 in 
snout, 1-5 in postorbital part of the head, slightly less than interorbital 
width (slightly less than or equal to interorbital). Width of snout before 
eye # of its length. Head moderately depressed, interorbital flat, scaled, 
three transverse series. A medio-longitudinal ridge on triangular part of 
upper jaw, which is 1-3 times as wide as long, scaly, 5 longitudinal and 3 
transverse series of scales. Width of lower jaw below tip of snout 3 of 
eye diameter. Small tricuspid teeth in fairly wide bands in both jaws. 
Beak stout, cutaneous fringes narrow, subgular sac small. Preorbital 
1-5 times longer than deep, covered by a single scale. Two series of scales 
on cheek, scaling extends forward to below tip of snout. A single line of 
pores on chin. Opercle naked, scales evidently fallen off. 

Gill-rakers 28, 3 in gill-filaments, which are 1-5 in eye. 

D 15 (14-16), commences slightly nearer base of ventral than base of 
caudal. Anterior rays elevated, almost 3 in head; remainder decrease, 
edge of fin concave. Fin scaly. 

A 15 (13-15), commences below the third dorsal ray. Similar in shape 
to dorsal. Fin scaly. Base of anal slightly less than base of dorsal. 

P 11, 2 in head, not scaly. 

V 6, not scaly, 3-3 in head, inserted midway between base of caudal 
and centre of opercle. 

Caudal deeply forked, lower lobe longer. Mid rays 3 of eye diameter. 

Air-bladder simple. 


Scales.—Lateral rows 52 (50-52), lateral line , predorsal 33 


3 
3 +22 +427 
(32-34), 6 scales between dorsal and lateral line. 

Colour (preserved).—Light brown, evidently bright green-blue above, 
silvery below. Fins light. Numerous close-set blue spots on head and 
snout and in a mid dorsal line to the origin of the dorsal. 

Length.—Up to 300 mm. 


31 


The South African Species of the Genus Hemirhamphus Cuv. 189 


Distribution.—East coast of Africa, Indo-Pacific. 

I have seen only one specimen of this species, from Aden, lent by the 
British Museum. 

It would appear that a revision of this species is highly necessary, for 
not only do descriptions vary rather widely, but authors differ in the 
nominal species admitted as synonyms. 

Day (Fishes of India, 1878, p. 515) includes dussumiert, described by 
Bleeker (Atl. Ich., 1866, vi, p. 56) in the synonymy of reynaldi C. & V. 
This is accepted by Weber and de Beaufort (Fishes Indo-Aus. Arch., 
1922, iv, p. 155), who consider both synonyms of dussumieri C. & V., and 
who somewhat arbitrarily admit erythrorinchus le Sueur to the synonymy of 
dussumiert C. & V., but do not give expression to the admitted priority of 
the former name. Fowler (“ Fishes Oceania,” suppl., 1931, p. 319) considers 
dussumert C. & V. a synonym of erythrorinchus le 8. 

Klunzinger (“ Fische Rothe Meer,’ Verh. Zoo. Bot. Ges. Wien, 1871, 
p. 584) gives a description of dussumiert C. & V. which does not agree with 
many other descriptions of this species, as indicated by Day (loc. cit.), 
whereas Weber and de Beaufort accept it as dussumiert C. & V. 

In my opinion, many specimens of erythrorinchus le 8. have been 
described as dussumiert C. & V., but it would appear that this latter species 
may well be maintained as distinct from the former, provisionally differ- 
entiated by the presence of scales on the dorsal and anal fins, postorbital 
part of head more than 1-4 times eye, mid-caudal rays less than eye, and 
ventrals nearer head than caudal base. 

Fowler’s Delagoa Bay specimens (loc. cit.) are most probably dussumiert 
C. & V. (as here defined), although this author states the relation of neither 
the postorbital part of the head, nor of the mid-caudal rays, to the eye, 
while his specimens have D 13, and only 18 gill-rakers on the lower part 
of the anterior arch. 

In so far as I am able to determine, erythrorinchus le 8. does not fall 
within or near our area. 


Hemirhamphus delagoae Brnrd. 
(Plate XII, fig. 2.) 


1925. Barnard, Ann. S.A. Mus., xxi, pl. x, fig. 6. 

Body elongate, slightly compressed, sides scarcely convex. Width 
1-25 in depth. 

Depth 10, length of head 4-8, in length of body. Lower jaw 4-9 in 
body, 6-3 in total length. © 

Eye 3-8 in head, 1-4 in snout and in postorbital part of the head, slightly 
greater than the interorbital width. Width of snout before the eyes 


32 


140 Transactions of the Royal Society of South Africa. 


about % of its length. Head depressed, indications of scales on occiput. 
Slight medio-longitudinal ridge on triangular part of upper jaw, which 
latter is twice as wide as long, scaly, 4-5 longitudinal and 3-4 transverse 
series of scales. Width of lower jaw below tip of snout about § of eye 
diameter. | 

Uni-, bi-, and tricuspid teeth present in anteriorly tapering bands, 
in 3-5 rows in each jaw, inner teeth mostly tricuspid, outer mostly uni- 
cuspid. Beak stout, cutaneous fringes 4-} as wide as beak; subgular 
sac moderate. Preorbital 1-3 times longer than deep, length 1-6 in eye, 
covered by a single scale with 2 central-branched pores. Three large 
scales with branched pores on opercle. 3-4 series of scales on cheek, 
scaling extends forwards to below the tip of the snout. 

Gill-rakers 24, moderately slender, 1:3 in gill-filaments which are 
3 in eye. Branchiostegals 12, membrane scaly. Mentum broad and 
densely scaled to symphysis, about 20 longitudinal and 4 transverse series 
of scales. 

D 15, arises considerably in advance of midway between bases of caudal 
and ventral. Anterior rays elevated, 2-2 in head; remainder decrease, 
edge of fin concave. Densely scaled, heaviest anteriorly. 

A 16, commences below the 2nd dorsal ray. Shape similar to dorsal, 
anterior rays about 2-5 in head. Densely scaled. Base of anal 1-25 in 
base of dorsal. 

P 12, 1-8 in head. Base densely scaled, scaling extends over half the 
length of the fin. 

V 6, scaled, densely on basal portion, length 3in head. Inserted midway 
between base of caudal and base of pectoral. Anterior rays longest, last 
slightly longer than penultimate. Edge of fin concave. Caudal deeply 
forked, upper lobe 1-3 in lower. Mid rays slightly less than eye. Peduncle 
fairly stout. 

Air-bladder simple. 


4 
Scales.—Lateral rows 61, 1.1. 2597 098° predorsal 40, 6 rows between 


dorsal and lateral line. Lateral line tubes with numerous inferior tubules 
(Plate X, C). 

Colour.—Greenish-blue above, silvery below. Nape, occiput, snout, 
and opercle dusky. Lower jaw and membranes black, tip of beak red. A 
narrow plumbeous lateral stripe from the base of the caudal to axil of 
pectoral, widest below origin of dorsal. Anterior dorsal rays dusky, remain- 
ing fins light. 

Length.—295 mm. 

Locality —Delagoa Bay. 

Lype.—In the South African Museum (No. 12303). 


33 


The South African Species of the Genus Hemirhamphus Cuv. 141 


The original description is inaccurate in many particulars, nor does the 
figure agree with the description. 

This species is very closely related to balinensis Blkr., but differs in the 
shorter beak, in the size and shape of the preorbital, and in the much wider 
upper jaw. The eye is relatively larger in delagoae than in balinensis. It 
ig an open question whether these differences alone justify the maintenance 
of delagoae as distinct from balinensis, the shorter beak of the former species 
being the only difference of note. 

Fowler (Ann. Nat. Mus., vi, pt. 2, p. 250) considers that delagoae is 
identical with unifasciatus Ranz. The errors in the original description 
have evidently misled Fowler, for delagoae is well differentiated from 
unifasciatus by the position of the ventrals alone. 


Hemirhamphus balinensis Blkr. 


1858. Bleeker, Nat. Tydschr. Ned. Ind., xvu, p. 170. 
1922. Weber and de Beaufort, Fishes Indo-Aus. Arch., iv, p. 152. 


Body moderately compressed, width about 2 in depth. Depth 9-10, 
length of head 4:5, in length of body. Lower jaw 3-4 in body, 4-8-4-9 in 
total length. Eye 4:5 in head, 1-5-1-6 in snout, 1-6-1-9 in postorbital 
part of head, equal to the interorbital width. Width of snout before the 
eyes % of its length. Head depressed, occiput scaly. Slight medio- 
longitudinal ridge on triangular part of upper jaw, which latter is 1-1-1-:3 
times as broad as long, scaly, 4-5 longitudinal and 3-4 transverse series 
of scales. Bi- and tricuspid teeth present in anteriorly tapering bands, 
in 3-4 rows in each jaw. Beak slender, cutaneous fringes narrow, subgular 
sac moderate. Preorbital 1-8-1-9 times as long as deep, as long as eye, 
covered by a single scale with 2 central pores. 2-3 series of scales on 
cheek, scaling extends forwards along sides of beak almost half-length of 
beak. Gill-rakers 22-26, slender, 2 in gill-filaments, which are about 2 in 
eye. Branchiostegals 12, membrane scaly. Mentum moderate, densely 
scaled to symphysis, about 20 longitudinal and 4 transverse series of scales. 

D 15-16, arises considerably in advance of midway between bases of 
caudal and ventral. Anterior rays elevated, about 3 in head, remainder 
decrease to the penultimate, last ray elongate. Edge of fin concave. 
‘Densely scaled, anteriorly. 

A 16-18, commences below the 2nd dorsal ray. Shape similar to— 
dorsal. Densely scaled. Base of anal as long as base of dorsal. 

P 12, base scaly. V 6, 3-5 in head, inserted midway between caudal 
base and hind margin to middle of opercle. Edge of fin concave. Base 
scaly. Caudal deeply forked, mid rays as long as eye, lower lobe longer. 

Air-bladder simple. | 


34 


142 Transactions of the Royal Society of South Africa. 


3 | 
— _ |, _—__———_,, predorsal 39-41, 
Scales.—Lateral rows 61-64, | 3098-99 428-99 p 


6 between dorsal and 11. Lateral line tubes with numerous inferior 
tubules, resemble closely those of delagoae (see Plate X, C). 

Colour (preserved).—Brown, darker above. Narrow dark stripe from 
base of caudal to axil of pectoral. Fins light. 

Length.—Up to 220 mm. 

Locality.—Mombasa. 

Distribution.—Indo-Pacific. 

The inclusion of this species in our fauna-list is based upon a specimen, 
lent by the British Museum, labelled dussumiert C. & V. from Mombasa. 
This specimen agrees generally with the diagnosis of balinensis Blkr., but 
differs in the rather light scaling on the vertical fins. Circumstances 
prevented a comparison of this specimen with two specimens of balinensis 
Blkr. subsequently kindly presented by Dr. de Beaufort, but there is little 
doubt that the Mombasa specimen is conspecific. The above description 
is based chiefly upon the two Indo-Pacific specimens. 


Hemirhamphus improvisus n. sp. 
(Plate XI, fig. 1.) 


Body elongate, robust, slightly compressed, sides slightly convex, 
width 1-3in depth. Depth 8:5, length of head 4-5, in length of body. Lower 
jaw 3-4 in body, 5 in total length. Eye 3-9 in head, 1-6 in snout, and in 
postorbital part of the head, 1-2 in interorbital width. Width of snout 
before the eyes 1:5 in length of snout. Head depressed, interorbital flat, 
with 2 shallow longitudinal grooves. No scales on occiput or interorbital, 
but these may have fallen off. Scarcely perceptible medio-longitudinal 
ridge on triangular part of upper jaw, which latter is slightly broader than 
long, scaly (fig. 1, B), 6-7 longitudinal and 3 or 4 transverse series : scaling 
extends over medio-longitudinal ridge. Width of lower jaw below tip of 
snout § diameter of eye. 

Tricuspid teeth in 5-6 rows in lower, and in 6-7 rows in upper jaw. 
Beak fairly stout, cutaneous fringes half as wide as beak; subgular sac 
moderate. Preorbital deeper than long, depth 1-8 in eye, a central pore. 
indicating a scale, evidently fallen off. 2 large and 1 small scales on 
opercle. 2-3 series of scales on cheek, scaling extending forward on man- 
dible to below middle of upper jaw. Gill-rakers 20, slender, 1-5 in gill- 
filaments, which are about 2 in eye. Branchiostegals 10: membrane 
naked ; mentum narrow, naked. 

D 14, arises very slightly in advance of midway between bases of caudal 


35 


Lhe South African Species of the Genus Hemirhamphus Cuv. 143 


and ventral. Anterior rays elevated, 4th longest about 2 in head; re- 
mainder decrease, edge of fin concave. Scaleless. 

A 15, arises below 2nd dorsal ray. Similar in shape to dorsal, anterior 
rays slightly longer than longest dorsal rays. Fin scaly anteriorly and 
along base. Base of anal 1-2 in base of dorsal. 

P 18, not scaled, 1-5 in head. 

V 6, not scaled, length 2-6 in head, 2nd ray longest, Ist and 3rd equal, 
remainder decrease. Inserted midway between base of caudal and middle 


Fia. 1.—Diagram to show type of scaling on upper jaws of A. H emirhamphus knysnaensis 
n. sp. and B. Hemirhamphus improvisus n. sp. 


of preopercle. Caudal slightly emarginate, mid rays 2-5 times eye. Lower 
lobe slightly longer than upper. Peduncle moderately slender. 

Air-bladder simple. 

Scales.—Lateral rows 53, Ll. _ 4 predorsal 38, 6 rows between 

2 +22 + 25 

dorsal and lateral line. Lateral line tubes with 2-4 inferior tubules. 

Colour.—Greenish above, silvery below. Nape, occiput snout, and top 
of opercle dark. Lower jaw and membranes black, tip of beak red. 
Median fins with dusky margins. A faint blotch on upper middle caudal 
rays. Remaining fins light. A silvery lateral stripe with dark upper 
border from base of caudal to axil of pectoral. 

Length.—218 mm. 

Locality.—Delagoa Bay. 

Type in the South African Museum (No. 16368). 

H. improvisus is closely related to schlegeli Blkr. and to knysnaensis 
n. sp., but is sharply distinguished from these by the character of the 
scaling on the triangular part of the upper jaw. In improvisus the middle 
of the 3 longitudinal series of scales is disposed centrally over the median 
ridge on the upper jaw, while the other two species have only 2 longitudinal 
series of scales which do not cover the median ridge (fig. 1, A and B). 


36 


144 Transactions of the Royal Society of South Africa. 


H. improvisus is further distinguished from schlegeli and knysnaensvs by 
the more anterior position of the ventrals and by the more obtuse upper 
jaw. The interorbital is markedly wider and the dorsal fin originates 
further back in improvisus than in schlegeli or knysnaensis. 


Hemirhamphus knysnaensts n. sp. 
(Plate XI, fig. 2.) 


1916. Thompson, Mar. Bio. Rep., ili, p. 266 (calabaricus). 
1925. Barnard, Ann. §.A. Mus., xxi, p. 262 (calabaricus part). 
1929. Fowler, Ann. Nat. Mus., vi, pt. 2, p. 250 (unifasciatus). 


Adults. 


Body elongate, slender, slightly compressed, about 14 times as deep as 
wide. Depth 9-3-10-9, length of head 4-3-5, in total length. Eye 
4-4-3 in head, 1-5-1-6 in snout, 1-5-1-7 in postorbital part of the head, 
and subequal to the interorbital width. Width of snout before eyes about 
% of its length. Head depressed, with 2 shallow longitudinal grooves. 
Top of head scaly. Medio-longitudinal ridge on triangular part of upper 
jaw, which latter is slender, 1-1-1:3 times longer than wide, scaly, 4-5 
longitudinal and 2 transverse series of scales: scales not overlapping 
medio-longitudinal ridge (fig. 1, A). Width of lower jaw below tip of 
snout about half-eye diameter. Minute curved conical teeth (unicuspid) 
in 2-3 rows in lower and 3-4 rows in upper jaw; a few tricuspid inner 
teeth in larger specimens. Beak slender, cutaneous fringes as wide as 
beak; subgular sac small. Preorbital slightly longer than deep, less 
than eye, covered by a single scale with central pore. Opercle with one 
small and two large scales. 2 series of scales on cheek, scaling extends 
forward along side of mandible about an eye diameter beyond below the 
tip of the snout. 

Gill-rakers 25-28, slender, 2 in gill-filaments, which are slightly longer 
than 2in eye. Branchiostegals 10, membrane and mentum scaleless. 

D 16-17, arises considerably in advance of midway between bases of 
caudal and ventral. Anterior rays slightly elevated, 3rd longest, about 
3 in head; remainder decrease to the last, which is about 5 in head. 

Kdge of fin slightly concave. Scaleless. 

A 16-18, commences below the 2nd or 3rd dorsal ray. Shape similar 
to that of dorsal. Length of base slightly less than that of dorsal. Scale- 
less, or few basal scales in large examples. 

P 11, conspicuously short, 1-8-2 in head. Scaleless. 

V 6, scaleless, about half the length of the pectoral, inserted midway 


37 


Lhe South African Species of the Genus Hemirhamphus Cuv. 145 


between base of caudal and base of pectoral to hind margin of operculum ; 
usually nearer base of caudal than hind margin of head. Anterior rays 
longer. 

Caudal moderately forked, upper lobe shorter, 1-2 in lower: mid rays 
1-6-2 times eye. Peduncle slender. 

Air-bladder simple. 


3 
Scales.—Lateral rows 55-58, 1.1, —__—_______, predorsal 34-35, 
3+ 26 —28 +22 —24 


5-6 rows between dorsal and lateral line. Lateral line tubes with one to 
four inferior tubules (Plate X, B). 

Colour.—Blue-green above, silvery below. Nape, occiput, snout, and 
opercle dusky. Beak and membranes black, tip of beak red. A narrow, 
plumbeous lateral stripe from upper edge of pectoral base to base of caudal, 
broadest (2 in eye) below origin of dorsal. In ripe males this stripe extends 
on to and over the lower caudal lobe. Anterior portion of dorsal dark, 
lighter posteriorly. A dark spot at the base of each dorsal and anal ray. 
Caudal with marginal area dusky. Remaining fins light. 

Length.—95-220 mm. 

Distribution.—South and east coasts of Africa. From False Bay to 
Natal, entering rivers, extending into fresh water. 

Types from Knysna in the Albany Museum. 


Juveniles. 


The young of this species are very slender and show main divergence 
from the adult form in the extent of the prolongation of the lower jaw 
and to a lesser extent in the shape of the caudal. The dorsal, anal, and 
pectoral fins approximate throughout in size and shape to those of the 
adult. The ventral fins are searcely visible in very small specimens 
(< 12 mm.), but thereafter rapidly attain and remain at the adult 
proportions. 

When 7 mm. in length the young are transparent, with a few dusky 
specks on the dorsal surface. The mouth is very oblique, the lower jaw 
not prolonged, and the triangular part of the upper jaw many times wider 
than long. The caudal is symmetrical, almost truncate, and there ig no 
trace of the lateral stripe. The lower jaw is first observed to project 
when the fishes exceed 10-12 mm. in length, and thereafter increases 
rapidly in relative length, the proportions of the adult in this feature 
being attained when the total length of the fish exceeds 30 mm. 

The relative increase in the length of this feature is shown on p. 146. 

At 10-12 mm. length the specks on the body are darker, and the 
first signs of the lateral stripe are seen as a series of short, dark longitudinal 


38 


146 Transactions of the Royal Society of South Africa. 
Total length of Lower jaw in 
specimen (mm.). total length. 
16 16 
20 10 
23 6:6 
27 55 
31 5-0 
33 4-6 


lines. The lower caudal lobe is now the longer. The lateral stripe is 
visible as such in specimens exceeding 20 mm. in length, and the caudal 
is markedly asymmetric. The mouth becomes less oblique, and the pro- 
longation of the upper jaw commences when the length exceeds 25 mm. 
At about 30 mm. total length the triangular upper jaw is about 4 times 
wider than long; at 70 mm. it is almost as long as wide; at 80-100 mm. as 
long as wide; thereafter gradually more acute, until in specimens over 
200 mm. in length it is as much as 1-3 times longer than wide. At 
all stages the beak is black, the red tip appearing in the later juvenile 
stages. 

This species is closely related to schlegeli Blkr. and improvisus n. sp., 
from which it is distinguished by the more acutely triangular part of the 
upper jaw as well as by the greater number of scales in a lateral series, 
and of dorsal and anal rays. 

It is also related to georgi C. & V., from which it differs in the shorter 
beak, as well as in the more anterior position of the ventrals and in the 
greater number of dorsal and anal rays. 

The two specimens from False Bay, very inadequately described by 
Fowler (loc. cit.) as unifasciatus Ranz, are most probably knysnaensis. 


Breeding Habits, Etc. 


Hl. knysnaensis appears to be present in the tidal rivers of South Africa 
throughout the year, extending into fresh water. Ripe fishes are first en- 
countered in October and are most numerous.in November and December. 
The young are exceedingly abundant in these months, and are present until 
February, after which they disappear. It is curious that the ripe fishes of 
this species have not been observed to congregate in any special areas, but 
are encountered in the surface of deep as well as shallow water, from two 
to four swimming one behind the other. Females predominate markedly 
in number. 

The eggs are demersal, 1-6 mm. in diameter, almost transparent, and 
densely covered with glutinous hair-like filaments, considerably longer 


39 


The South African Species of the Genus Hemirhamphus Cuv. 147 


than the diameter of the egg. The larger females produce about 10,000 
eggs per season. 


Hemirhamphus schlegeli Blkr. 


1862. Bleeker, Mem. Soc. Holl. Haarlem, p. 120, fig. 1. 

1866. Gunther, Cat. Fish. Brit. Mus., vi, p. 266 (calabaricus). 

1866. Gunther, Ann. Mag. Nat. Hist. (3), xvili, p. 427 (calabaricus- 
schlegelt Blkr.). 

1925. Barnard, Ann. S.A. Mus., xxi, p. 262 (calabaricus part). 


Body elongate, subcylindrical, about 14 times as deep as wide. 

Depth 9-9-4, length of head 4:4, in length of body. Lower jaw 3-3-3-6 
in body, 5-5-2 in total length. Eye 3-8-4 in head, 1-2-1:3 in snout, 1-5 
in postorbital part of the head and subequal to the interorbital width. 
Width of snout before eye 3 of its length. Head depressed, interorbital 
flat, with 2 shallow longitudinal grooves, scaly, 3 transverse series. Medio- 
longitudinal ridge on triangular part of upper jaw, which latter is slender, 
as long as wide, scaly, 3-4 longitudinal and 2 transverse series of scales, not 
overlapping median ridge. Width of lower jaw below tip of snout about 
half eye diameter. Small conical, bi- and tricuspid teeth in 3 rows in each 
jaw. Beak slender, cutaneous fringes almost as wide as beak; subgular 
sac small. Preorbital almost as deep as long, covered by a single scale with 
central pore. Three series of scales on cheek, scaling extends forward to 
below corner of mouth; a single line of peres on the side of the chin. 

Gill-rakers 22-26, slender, 2-3 in gill-filaments, which are about 1-5 in 
eye. Branchiostegals 10, membrane scaleless. 

D 14, arises considerably in advance of midway between bases of caudal 
and ventral. Anterior rays slightly elevated, 3rd longest, 2-5 in head, 
remainder decrease, last 5 in head. - Edge of fin gently concave. Scaleless. 

A 15-16, commences below 2nd dorsal ray. Shape similar to dorsal ; 
slightly higher than dorsal. Length of base slightly less than that of 
dorsal. Scaleless. 

P 11, 1-7-1-8 in head, scaleless. 

V 6, scaleless, half the length of the pectoral, inserted midway between 
base of caudal and hind margin to middle of opercle. Anterior rays longer. 

Caudal emarginate, upper lobe shorter, 1-3 in lower. Mid rays 1-5-1-7 
times eye. Peduncle slender. 


Scales.—Lateral rows 50, 1.1. predorsal 35, 6 rows between 


4 
2424 +22’ 
dorsal and lateral line. Lateral line tubes with 2 or 3 inferior tubules. 
Air-bladder simple. | 

Colour (preserved).—Brownish yellow, darker above. A dark lateral 
stripe from base of caudal to axil of pectoral, widest below origin of dorsal. 


40 


148 Transactions of the Royal Society of South Africa. 


Anterior dorsal rays and margin of caudal dusky. Remaining fins light. 
Beak black. 

Length.—140-215 mm. 

Locality.— Angola, Cameroons. 

Distribution.— West coast of Africa. 

This species does not strictly belong to the South African fauna, but is 
included here since it is almost certain to be discovered within our regian. 

Fowler (Ann. Nat. Mus., 1929, vi, pt. 2, p. 250) considers schlegeli 
identical with unifasciatus Ranz. This is scarcely possible, since this 
latter species has a very short beak, ventrals inserted midway between 
caudal base and eye, and dense scaling on the dorsal and anal fins. 


Notes on the H emirhamphidae. 


These fishes appear to live almost exclusively in the surface of the 
water, and show high specialisation of habit with adaptive modifications 
in external form. 

The leaping powers of the species of Hemirhamphus are markedly 
characteristic. These fishes emerge from the water with the body rigidly 
straight, and the leap is terminated either by falling flat upon the surface, 
or, in the case of the smaller specimens, with the caudal first entering the 
water. I have not yet seen any of these fishes ending a leap by diving head 
first into the water. The ventrals are fully extended laterally during the 
leap. As many as six or seven immediately consecutive leaps have been 
observed, the pause between each for renewal of the impulse being extremely 
brief. Adult specimens of H. far cover 10-15 feet in each leap, and are 
able to progress up to 100 feet along the surface in this manner, at an 
average speed of at least 25 feet per second. Very young specimens of 
this species are apparently unable to emerge from the water and are very 
easily captured, even by day. H. knysnaensis, a smaller species, progresses 
in a similar fashion along the surface, but the body is held at an angle of 
about 45° to the surface, the leaps are very short, 1-2 feet, and are 
seldom repeated more than 4 or 5 times. Half-grown specimens of this 
species show an interesting stage in the development of this power. When 
startled, these fishes shoot up until the body as far as the ventrals is out 
of the water at an angle of about 45° to the surface: with ventrals 
widely extended, the body is held in this position, and urged forward along 
the surface for a few yards at a surprising rate, the caudal vibrating with 
extreme rapidity. As far as has been observed, none of these fishes, small 
or large, attempt to escape attack, even from above, by diving, which 
would appear to indicate that they do not readily sink from the surface for 
any purpose. 


41 


Lhe South African Species of the Genus Hemirhamphus Cuv. 149 


The leaping powers of these fishes have led Schlesinger (Phy. and Eth. d. 
Scombr., Verh. Zoo. Bot. Ges. Wien, 1909, lix, pp. 302-339) to conclude 
that the closely related Exocoetidae have developed from the Hemirham- 
phidae. It would appear more reasonable to suppose that both forms have 
developed simultaneously from a common surface-dwelling ancestral type, 
the enlargement of the “ beak” of the Hemirhamphidae and the pectorals 
of the Hzocoetidae representing later specialisation. The very young 
stages of species of Hemirhamphus have no “ beak,” while the pectorals of 
the very young of certain species of the Exocoetidae are relatively consider- 
ably less elongate than those of the adult fishes. 

The adoption of a purely surface life in the case of the Hemirhamphidae 
(and of the Hzocoetidae) is further indicated by the ventro-lateral position 
of the lateral line in these fishes. 


Functions of the “ Beak.” 


Schlesinger (loc. cit.) has concluded that the prolonged lower jaw of the 
Hemirhamphidae is employed to plough up the bottom mud in search of 
food, and he discounts the statements of other investigators that the 
stomach contents of these fishes consist solely of surface plants and 
organisms. | 

It has been previously indicated that there are reasons for believing 
that the Hemirhamphidae are exclusively surface-dwellers. I have ex- 
amined microscopically the stomach contents of many specimens of various 
species of Hemirhamphus, and it is significant that in no case was any trace 
of sand-particles observed. The aliments generally appear to consist of 
surface plants, larval crustacea and fishes, and pelagic eggs. The Mugilidae 
are also surface-dwellers, but feed not only upon surface organisms, but 
also upon material obtained by skimming or stirring up the bottom mud and 
eel-grass (Zostera) in shallow water ; sand particles are invariably present 
in the aliments of the Mugilidae, frequently in large proportions. It 
appears inconceivable that the Hemirhamphidae could eliminate all trace 
of sand particles from food obtained by stirring up mud with the beak. 

Further, the tip of the beak of a live Hemirhamphus is soft, and the 
distal portions of the lateral cutaneous fringes extremely delicate and fairly 
vascular. These fishes rapidly succumb when the beak is damaged, and it is 
perhaps significant that among many hundreds of specimens I have taken 
alive none have been found with a broken beak. It would therefore 
appear extremely unlikely that the ‘“‘ beak ’’ is employed in the manner 
suggested by Schlesinger. 

From sustained observations, chiefly by night with the aid of a powérful 
light, it appears not unlikely that the prolonged lower Jaw, with its cutaneous 


42 


150 Transactions of the Royal Society of South Africa. 


fringes, is employed by these fishes as an organ for the detection of the 
minute surface organisms which constitute their food. They lie in the 
surface of the water with the cutaneous fringes of the beak fully extended 
laterally. At frequent intervals the head may be seen to be jerked slightly 
to one side, the body urged forward, and motions concomitant with the 
assimilation of some particle ensure. Specimens will lie undisturbed in the 
light of a lamp, if cautiously approached ; the body may be touched with 
a very fine wire without occasioning undue alarm, but if the beak is only 
lightly touched, the fish withdraws violently. 

The function of the subgular sac remains obscure. The contents of the 
distended sac (rarely so observed after capture) appear to be liquid, and 
not gaseous, but sufficient could not be obtained for investigation. 


I wish to express my gratitude to Mr. J. Hewitt, Director of the Albany 
Museum, and to Dr. K. H. Barnard, Assistant Director of the South African 
Museum, for advice and assistance, and for the loan of material and 
literature ; to Mr. J. R. Norman of the British Museum for the loan of, 
and to Professor de Beaufort of Amsterdam, for the donation of, valuable 
specimens. Also for financial assistance from the Carnegie fund, through 
the Research Grant Board of South Africa, which defrayed part of the costs 
of the investigation. 


ALBANY Museum, 
GRAHAMSTOWN. 


Trans. Roy. Soc. §. Afr., Vol. XXT. Plate X 


LATERAL LINE SCALES, SEMI-DIAGRAMMATIC. 


A, Hemirhamphus far Forsk. 
B, Hemirhamphus knysnaensis n. sp. 
C, Hemirhamphus delagoae Brnrd. 


J. L. B. Smith. Neill & Co., Ltd. 


Plate XI. 


Trans. Roy. Soc. 8. Afr., Vol. X XI. 


‘(ayeur) ‘ds ‘u siswanushuy snydunysmazy ‘2 
‘ds ‘u snswmosdut snydumnysuazy *y 


zy) \ 


Py 


Neill & Co., Ltd. 


J. L. B. Smith, 


Plate XIT. 


Trans. Roy. Soc. 8. Afr., Vol. X XT. 


‘pruig avobojap snydunysvwmayy °% 
‘ysiog wof snydupy.uwey *T 


o 


xh 
RIN 


(io 
wo 
i 


Neill & Co., Ltd. 


J. L. B. Smtth. 


43. 


Transactions of the Royal Society of South Africa. Vol. XXII. 
Part |. pp. 83-87. PI. IV. August, 1934. 


THE GROWTH CHANGES OF PTEROPLATEA 
NATALENSIS, G. anv T. 


By J. L. B. Smira, 


(With Plate IV and one Text-figure.) 


Famity DASYBATIDAE. 
Genus PreropLaTEA, Miller and Henle. 


1913. Garman, Mem. Mus. Comp. Zool. Harv., p. 412. 
1925. Barnard, Ann. 8.A. Museum, vol. xxi, p. 80. 


To the diagnosis of this genus add the following: “An ovoid flap of 
the iris extending over the pupil from the upper margin, increasing with 
age. A wide but low crenulate cutaneous flap on floor and on roof of mouth, 
behind the dental laminae, may be present. A small tentacle, on the inner 
posterior margin of the spiracle, sometimes present, decreasing with age, 
possibly becoming obsolete in large sexually mature individuals.’ 


Pteroplatea natalensis, G. and T. 


1911. Gilchrist and Thompson, Ann. S.A. Mus., vol. xi, pt. 2, p. 56. 

1925. Barnard (loc. cit.), p. 81. 

The holotype of this species is a juvenile male from Natal, 280 mm. 
across the disc. It possesses a small pointed tentacle on the inner posterior 
margin of the spiracle, while the disc is markedly triangular in shape, with 
undulate anterior margin. The caudal spine has been stated not to be 
serrate, but this is an error. The serrations are present, but are hidden 
under the skin or by a coating of mucus. 

The only other species recorded from South Africa is mecrura, Bl. Schn. 
The identification is based upon two large specimens, one stuffed, one cast 
(from the Agulhas Bank), inthe S.A. Museum. This species has no tentacle, 
and the disc is subrhomboidal in shape. 

Very little appears to be known about the habits and development of 
species of this genus, nor, in so far as I have been able to ascertain, have 
the growth changes of any species been previously described. 

The present work is based upon the examination of a graduated series 


44 


84 Transactions of the Royal Society of South Africa. 


of specimens, ranging from 104-1830 mm. across the disc, secured in the 
course of experimental netting in the Knysna estuary. 

These fishes appear to enter the river chiefly during the late summer 
and early winter months, but have been observed throughout the year. 
Of the specimens mentioned above, only two were sexually mature. These 
were gravid females, 1790 and 1830 mm. across the disc respectively, and 
each proved to hold eight embryos. The smaller of the two was taken 
near the mouth of the river, while the larger was captured about seven 
miles higher, where the salinity of the water is very considerably lower 
than that of the sea. The netters state that very large as well as very 
small specimens are by no means infrequently taken in the autumn. It 
is possible that gravid females may seek out the more protected waters of 
the river to give birth to the young. Since, however, sexually immature 
individuals are as frequently present, it is possible that the entering of the 
river may be purely fortuitous, for the mouth of the Knysna River is both 
wide and deep. 

In so far as may be judged from the stomach contents, the food of these 
fishes appears to consist of small fishes (Sparus sarba Forsk. and Mugil sps.) 
and crustacea, chiefly crabs. 

The larger of the two gravid females to which reference is made above 
might easily have been mistaken for a specimen of micrura, since there was 
no trace of a spiracular tentacle, and the disc was much less triangular in 
shape than that of the type of natalensis. The embryos it contained 
(5 2, 3 3) are, however, quite apparently conspecific with natalensis, since 
these all possess the spiracular tentacle, and are even more markedly 
triangular in shape than the type. These embryos (104-117 mm. across 
the disc) are far from mature: the appendicular yolk-sac is large, the caudal 
spine is unossified and adnate, while chromatophores are not visible, and 
filamentous external accessory gills are present. In the smaller of the two 
females (1790 mm.) the spiracular tentacle has been reduced to a mere knob, 
while the disc resembles that of the larger female. The embryos (6 2, 2 3, 
265-280 mm. across the disc) are in this case mature. The yolk has been 
fully absorbed, external gills are absent, the coloration is fully developed, 
and the caudal spine is ossified, serrate and free. All possess the spiracular 
tentacle. One of these is shown in Pl. IV, fig. 2. 


Embryonic Forms. 


The width of the disc in the embryos is 1-6-1-7 times the length (tip of 
snout to level of hind margin of disc). The mid-point of the line joining 
the pectoral apices is 3-5-5-0 times as far from the tip of the snout as from 
the hind margin of the disc (in the type specimen 4:3). There does not 


45 


Lhe Growth Changes of Pteroplatea natalensis, G. and T. 85 


appear to be any appreciable change in the shape of the disc from the early 
to the later stages of embryonic development. The interorbital width is 
0-9-1-3 in the snout (for convenience measured from midway between the 
centres of the eyes to the anterior point of the disc). In the very small 
specimens the eyes are pedunculated, and the developing iridal flap shows 
as a small but sharp convexity at the upper margin of the pupil. The 
caudal is 1-3-1-8 in the length of the disc, with a moderate cutaneous fold 
above and below. The ossification of the caudal spine is evidently one of 
the later processes of embryonic development, possibly taking place while 
the whole spine is adnate to the body of the caudal. The accessory external 
filamentous gills of the very small specimens are considerably longer than 
the disc. The coloration of the mature embryos is very similar to that of 
half-grown individuals, except that numerous round spots are more obvious. 


Post-natal and Mature Stadia. 


The disc changes from subtriangular to subrhomboidal with growth, 
becoming also relatively broader (fig. 1, A-F). The posterior margin of 


Fig. 1.—Diagram to illustrate the change in shape of the disc of Pteroplatea natalensis, 
G. and T., with growth. Width of disc in mm.: A, 105; B, 280; C, 435; D, 610; 
E, 1080; F, 1830. 


the disc is gently convex, sometimes finely scalloped. The anterior mar- 
gin of the pectoral is gently undulate, with concavity about midway 
between pectoral apex and snout. The pectoral apices become less 
obtuse with growth. The interorbital width is equal to or slightly 
greater than the snout (see above). The eyes are prominent, and the 
iridal flap in the larger specimens covers most of the pupil. Length of 
spiracle about 2-5 in interorbital width, which is 10-12 in disc width. The 


46 


86 Transactions of the Royal Society of South Africa. 


spiracular tentacle in young and half-grown individuals is slender and 
pointed, 8-10 in spiracle length; apparently diminishes with age. Caudal 
small, diminishing in relative length with growth (the accidental shorten- 
ing of this organ by injury is frequently observed in Dasybatids). Cutaneous 
fold, above and below, moderate. One or, more usually, two serrated 
spines near caudal base, short and stout. No dorsal fin or, as observed 
in one case, a mere rudiment, reduced to a small fold of thickened skin 
immediately anterior to the caudal spines. Skin smooth. Teeth tesselate, 
each rhomboidal base with a flattened triangular retrorse point. The 
number of the teeth in each row and the number of rows increases from 
40 rows of 8-10 each in the smaller to 110 rows of 15-20 each in the largest 
specimens. Anterior and posterior margins of each lamina undulate: the 
upper lamina with anterior median convexity, the lower with median anterior 
concavity (Pl. IV, A and B). (Barnard’s comment, loc. cit., p. 81, on the 
paragraph about the dental laminae in the original description is rather 
severe. The meaning of this rather obscurely worded paragraph becomes 
clear when the laminae are examined.) Colour uniform brown above, or 
with darker or lighter vermiculations and blotches, light below. Tail with 
alternate dark and light annulations, becoming obscure in adults. 

The following table indicates the chief changes which take place with 
growth :— 


Sex. Ft | M.2 |] M8 F. F. F.4 F, F. 
Width disc in mm. . | 105 | 265 | 280 | 435 | 610 | 1080/1790} 1830 
Length disc in width 1-6 | 1-7 | 1-75] 1-85] 1-9 | 2-1 | 2-1 | 2-0 
Caudal in disc length . | 1:5 | 1:8 | 2:0 | 2-5 | 2-0 | 2-1 | 2:5 | 3-3 
Pectoral line ® . | 50 | 4:2 | 4:3 | 3-2 | 2-6 | 1-95] 1-6 | 1-6 
In fig. 1 ; ; .| A - B C D 1D) .. F 


1 Most triangular immature embryo. 

2 Plate IV, fig. 2. 8 Type specimen. 4 Plate IV, fig. 1. 

* The number of times that the mid-point of the line joining the pectoral apices is 
farther from the tip of the snout than from the posterior margin of the disc. 


This species is very closely related to altavela Linn., from the tropical 
Atlantic, from which it apparently differs in the markedly triangular disc, 
in the more obtuse pectoral apices and in the more more undulation 
of the anterior margins of the pectorals. These differences are slight, 
especially since altavela appears to be of somewhat variable form. Garman 
(loc. cit., p. 415) considers canariensis Val., valencienni Dum., and vaillanti 
Rochebr. conspecific with altavela. It is possible that the earlier figures 
of species are not always to be relied upon, but I have seen those of these 


47 


The Growth Changes of Pteroplatea natalensis, G. and T. 87 


synonyms, and they differ very considerably one from the other. It is not 
unlikely also that very different stadia have been figured. 

I may here call attention to the fact that descriptions of the species of 
this genus rarely include the precise size of the specimens described, which, 
in view of the evidence here adduced, considerably diminishes their value. 
It would not indeed be surprising if natalensis were to be found conspecific 
with altavela. Barnard, loc. cit., probably had some such idea, since he 
remarks that the adult of natalensis, when found, would probably be found 
referable to a previously known species. I do not feel justified in pro- 
nouncing natalensis a synonym of altavela, since I have seen no authentically 
named specimens of the latter. 

It may be remarked that the growth changes of natalensis indicate 
that a revision of this genus, based upon adequate material, would prob- 
ably result in a somewhat drastic limitation of the number of species. It 
is evident that the shape of the disc, the length of the caudal, and even 
the presence or absence of the spiracular tentacle are, unless many equl- 
valent stadia are compared, not as reliable a guide to the differentiation of 
species as may hitherto have been supposed. 

At Knysna, natalensis is named “Backwater” by the netters: when 
caught, the fish lies flat against the net and so renders the retraction 
extremely arduous. The species is for this reason an extremely unwelcome 
capture, and also, in addition, since: its presence in the net appears to 
terrify the other fishes, great numbers of whom escape by jumping over the 
net. At Knysna this species is seldom taken on lines, but farther east 
it is not an infrequent capture. A large specimen of presumably this 
species was recently taken on a line in the Keiskama River, and in Natal 
waters it is no infrequent capture, large specimens proving formidable 
antagonists to the angler. 


I wish to express my gratitude to Dr. Barnard, Assistant Director of 
the S.A. Museum, for his kind assistance in regard to material and literature 
in the S.A. Museum. Also to the Carnegie Research Fund (through the 
Research Grant Board of South Africa) for financial assistance. 


ALBANY MUSEUM, 
GRAHAMSTOWN, 
August 1933. 


Trans. Roy. Soc. 8. Afr., Vol. XXII. 


Plate IV, 


pase 


} 


1. Half-grown female, x 


2. Mature male embryo. 


PTEROPLATEA NATALENSIS, G. and T. 
1 


10° 


A. Upper dental lamina of a specimen 1830 mm. disc width. 
B. Lower dental lamina of same. 


Neill & Co., Ltd. 


49. 


Transactions of the Royal Society of South Africa. Vol. XXII. 
Part |. pp. 89-100. Pls. V and VI. August, 1934. 


MARINE FISHES OF SEVEN GENERA NEW TO 
SOUTH AFRICA. 


By J. L. B. Smrra. 


(With Plates V and VI, and one Text-figure.) 


The following new genera are described below:— 


Family StromatErpaE, Papyrichthys. 
Family BarracnoipipaL, Batrichthys. 


Family SYNGNATHIDAE. 


M icrophis brachyurus Blkr. 


1888. Day, Fishes of India, p. 680, pl. clxxiv, fig. 3 (Doryichthys bleekeri 
Day). 

1922. Weber and de Beaufort, Indo-Aus. Fishes, iv, p. 44, fig. 21. 

Body moderately slender, depth 4-5 in head, almost as wide as deep. 
Head 4-8 in length to caudal base. Tail (without fin) slightly shorter 
than half distance from snout tip to vent. Eye 10, snout 1-6, and post- 
orbital 3-4 in length of head. 

Operculum with a slight medio-longitudinal ridge for the entire length; 
on one side 3, on the other 5, similar shorter, curved, radiating ridges 
below, all feebly serrate. Snout very compressed, with supero-median. 
ridge. Supra-orbital ridge continuous from temporal region on to snout. 
All ridges on head and snout feebly serrate. . 

Rings 22+23. Each body plate has 14-18 smooth parallel vertical 
ridges, which sometimes become furcated and reticulate below. Across 


_ Nee. 


COINS EEE EEE PELE 
Basse ecee a eS cee 


—lan__. 
Fia. 1.—Microphis brachyurus Blkr. 


the middle of each plate are 3 longitudinal curved ridges. The edge of 
each plate is serrulate, the serrae corresponding in number with the vertical 
ridges; last spinelet in each plate slightly enlarged. | 
The supero-lateral body ridge is discontinuous with the supero-lateral 
caudal ridge, and extends to the 7th caudal ring. The medio-lateral body 


50 


90 Transactions of the Royal Society of South Africa. 


ridge curves down below the origin of the dorsal and unites with the 
infero-lateral caudal ridge on the 2nd caudal ring. The infero-lateral body 
ridge is not continuous with the infero-lateral of the caudal, and curves 
on to the ventral surface of the caudal, ending in the 2nd caudal ring. 
The supero-lateral caudal ridge curves down (forwards) below the hind end 
of the dorsal to medio-lateral, and extends as far as the hind margin of 
the last body ring. 

D 40, base not raised, as long as snout, originates on penultimate 
body ring, on 9 rings (2+7). 

A 5, very small, inserted below anterior dorsal rays. 

P 19, slightly longer than eye, emarginate. Caudal small, about 3 in 
snout, rounded. No brood pouch visible, presumably a female specimen. 

Colour.—Light brown, margins of plates lighter. Fins light; caudal 
dark. 

Length.—160 mm. 

A single specimen, from Durban, presented by HE. C. Chubb, Esq., 
Curator of the Durban Museum. 

M. brachyurus has been recorded from East Africa to the Indo-Pacific. 
The only other species of this genus hitherto recorded from Africa is smathu 
Dum., from the tropical west coast, which is stated to be fluviatile. 

M. brachyurus is closely related to this species, differing only in certain 
minor features, such as the longer snout and the more posterior insertion 
of the shorter dorsal. it is possible that wider collection may show that 
the two forms are identical. It appears preferable to maintain the Indian 
Ocean form as a separate species until such time as the recorded areas may 
be linked up. 


Family PLECTORHYNCHIDAE. 
Scolopsis vosmert (Bloch.). 
(Plate V, A.) 


1878. Day, Fishes of India, p. 87, pl. xxiii, fig. 1. 


Dorsal profile even, body compressed, ovate. Depth 2-1, length of 
head 2-9 in length of body: eye 3-1 in head, slightly greater than snout 
and than interorbital width, 1-1 in postorbital part of head. Preopercular 
margin concave, strongly serrulate, spinelets directed obliquely outwards. 
Preorbital produced backwards into a prominent flat spine, extending 
almost to below hind margin of pupil: 5 graduated smaller spines below. 
Suborbital feebly serrated. Muciferous pores on anterior part of head. 
Mouth small, terminal, maxilla extends to below anterior margin of eye, 
almost entirely concealed beneath preorbital. Villiform teeth in bands 
in both jaws, tapering towards the sides, which have only a single row of 


51 


Marine Fishes of Seven Genera new to South Africa. 91 


teeth. Palate and tongue edentate. Gill-rakers 5, very short and stout; 
3 in gill-fringes, which are 2-5 in eye. 

D X 9, commences above middle of operculum, not notched. 1st spine 
shortest, 4-4in head; 2nd 3-1; 3rd 2:6; 4th and 5th longest, 2-4; last 2-8 in 
head. Anterior rays equal to last spine. 

A III 7, commences below the 2nd dorsal ray. 2nd spine very stout, 
longest slightly longer than 4th dorsal spine. Vertical fins not scaly. 

P 18; 1:3 in head, rounded. 

V 15, 1-25 in head, inserted below 6th dorsal spine; 1st ray filamentous, 
reaches to base of 2nd anal spine. 

Caudal forked, upper lobe longer. 

Scales ctenoid, 1.1. 41, ltr. 34/14; 1.1. tubes bifurcate on anterior scales. 
Some posterior 1.1. scales have 2-3 superior branches each ending in a 
pore. 1.1. scales smaller than body scales, 4-5 series on cheek. 

Colour.—Yellow-brown, with a light band over the nape and a light 
stripe along the side. Dorsal dark anteriorly, other fins light; axil of 
pectoral and upper hind margin of operculum black. 

Length.—107 mm. 

Locality.—Presumably Port Alfred, having been found together with 
other unlabelled fishes among the old collection of the Albany Museum. 

Distribution.—Indo- Pacific. 

This is the first authentic record of a species of this genus from South 
Africa. As mentioned by Barnard (Ann. S.A. Mus., 1925, xxi, p. 669), 
Playfair and Gunther record Scolopsis monogramma K. and v. H. from 
Mozambique, but without citing any authority. 

I have seen no recent work dealing with the species of this genus, but 
the older differentiation, based apparently largely upon coloration, appears 
none too sound, and is probably in need of revision. 

The presence of this species on our southern coast is noteworthy. The 
stretch of coast from Algoa Bay to East London has not before received 
much attention from ichthyologists, and has proved a rich collecting 
ground. Numerous species formerly believed to be confined to the western 
and south-western coasts (among others Congiopodus torvus (Gron.) and 
many species of Clinus Cuv.) have been found commonly occurring there, 
while many species normally accounted as tropical (e.g. among others 
Petroscirtes tapeinosoma (Blkr.), Monoceros Unicornis (Forsk.), and Albula 
vulpes (Linn.)) have also been found. 

The occurrence of the present species is therefore not in any way 


exceptional. 


52 


92 Transactions of the Royal Society of South Africa. 


Family BROTULIDAE. 


Grammonus opisthodon n. sp. 
(Plate VI, B.) 


Body compressed, moderately elongate, tapering uniformly from nape. 
Depth 4, length of head 3-2 in length of body. Eye 7, snout 4, interorbital 
3-2, and postorbital 1-7 in length of head. Nostrils widely separated, 
anterior on snout. No barbels or spines on head, except for a strong 
spine, and a second, weaker, diverging on opercle. Preopercle margin 
below skin, with an obtuse ridge at the angle. Muciferous pores on head, 
very marked on chin. Snout obtuse, somewhat swollen. Mouth terminal, 
large, slightly oblique. Maxilla extends well behind eye, its length 1-7 in 
length of head, posteriorly dilated to a width equal to eye, with a retrorse 
spine on the hinder lower edge. A single series of fairly large curved 
conical teeth in each jaw: no villiform teeth. Two separate obtuse 
vomerine knobs, each with two large conical curved teeth. Palatines, 
ather bones, and tongue edentate. Gull-membranes separate, free from 
isthmus. Gills 4, a slit behind the 4th. Giull-rakers 3, long and spinose, 
plus 5 spinous flattened ridges anteriorly. Pseudobranchiae absent. 
Branchiostegals 6. Vent postmedian. 

D 65, originates 1-6 times as far from caudal base as from tip of snout, 
over middle of pectoral. 

A 35, originates 1-5 times as far from tip of snout as from caudal base. 
Median fins confluent with caudal. 

P 23, 1-5 in head, rounded, base enlarged. 

Ventrals each reduced to a single filament, about 4 in head, inserted 
below the opercle. 

No sign of any scales, or of lateral line. 

Caudal small, pointed. 

Length.—50 mm. 

Locality.—Port Alfred. 

Colour.—Brownish, with numerous small black spots. 

This interesting small specimen is unfortunately not in very good 
condition, and the absence of scales and lateral line cannot therefore be 
regarded as established. Since it does not appear to fall into any genus of 
this family of which I have a description, I forwarded the specimen to 
Mr. Norman of the British Museum, who, at my request, compared it 
with the Brotulids in their collection. It is apparently identical with 
none of these. 

I have provisionally assigned it to Grammonus Gill, with the diagnosis 
of which it agrees broadly. It differs in many features, however, the 


53 


Marine Fishes of Seven Genera new to South Africa. . 93 


most marked being the presence of the maxillary spine, and in the absence 
of villiform teeth in the jaws and on the vomer, as well as in the absence 
of scales and of lateral line. 

Generic distinction in the Brotulidae appears to be rather fine, and the 
present species may well serve as the type of a new genus. This is, how- 
ever, largely dependent upon the discovery of specimens in good condition, 
in which the presumed absence of scales and lateral line may be confirmed. 
This is the first record of a species of this genus from South Africa. Two 
other species are known, both from the Mediterranean: ater Risso is 
probably bathybial and is said to come inshore to spawn; the other is 
armatus Doederlein, about which little is known. 


Family STROMATEIDAE. 
Papyrichthys n.g. 


Body ovate, very highly compressed, especially at the bases of the 
dorsal and anal fins, these regions being distinct from the body proper. 
Mouth large, slightly oblique. Maxilla excluded from margin of upper- 
jaw, almost entirely concealed beneath preorbital. A single row of 
moderate fine-pointed teeth in each jaw. Palatal teeth present or absent. 
Branchiostegals 5. Giull-opening wide, membranes free, rakers well 
developed. Pseudobranchiae present. 

Dorsal and anal very long, rays free from enveloping skin. Pectorals 
inserted low down. Ventrals of one short spine and five rays, longer than 
pectorals, inner rays longest, extending well behind origin of anal. Body 
and part of head covered with very small cycloid scales, probably always 
more than 100 in lateral series. Pores on head and body few or absent. 
Lateral line high, more or less following the dorsal profile. Air-bladder 
present. Pyloric caeca moderately numerous, branched. Vertebrae 40. 

Genotype pellucidus Liitken. | 

This genus is related to Psenes C. and V., but it is well differentiated — 
by the very minute scales, the size and shape of the ventrals, the very 
highly compressed body, and the greater number of vertebrae. I have 
examined a specimen of Psenes indicus Day, which apparently does not 
possess the many-branched pyloric caeca such as are found in Papyrichthys. 
This may prove a characteristic feature of this latter genus, although 
Regan (Ann. Mag. Nat. Hist., 1902 (7), X, p. 118) has given reasons why 
the nature of the caeca is not very reliable in defining genera in this family. 
Psenes is stated by Regan (loc. cit.) to have 25 vertebrae, but while he 
placed pellucidus in this genus, he had not seen a specimen, and merely 
followed Liitken’s diagnosis. | 

The very highly compressed plastic body, and the dentition, would 


54 


94 Transactions of the Royal Society of South Africa. 


indicate that Papyrichthys is related also to Schedophilus Cocco, but it is 
distinguished from this genus by the absence of pores on the body and 
of the thickened porous skin on the occiput, as well as by the course of the 


lateral line. 
Papyrichthys pellucidus Liitken. 


(Plate VI, A.) 


1895. Goode and Bean, Ocean. Ichth., p. 221, pl. lx, fig. 228 (Psenes 
pellucidus Liitken). 
1902. Regan, Ann. Mag. Nat. Hist. (7), X, p. 125 (Psenes p.). 


Body soft and flabby, very compressed, maximum width at shoulder 
about 10 in length. Dorsal profile gently sloping from nape. Regions 
containing the dermal rays at bases of dorsal and anal fins very highly 
compressed, almost translucent, sharply defined. Snout very blunt, 
almost vertical, with slight mesethmoidal ridge. 

Depth 2:8, length of head 3-0 in length of body. Eye 4-0, interorbital 
(convex) 5, snout 3-4, and postorbital length 2-1 in head. Interopercle 
with traces of marginal spinules. Other bones of head entire. Preorbital 
with anterior trifid ridge. Bones of head soft and cavernous. A few 
pores on chin and head. None on body. Nostrils close together, much 
nearer snout tip than eye. Mouth large, terminal, slightly oblique, jaws 
equal. Maxilla extends to below the anterior third of the eye; end of 
maxilla slightly expanded. A single comb-like row of moderate, fine, 
pointed teeth, slightly recurved, in each jaw; those in lower jaw slightly 
larger. Palatal bones edentate, but signs of small teeth on vomer. Tongue 
free, edentate. Gill-rakers 13, longest slightly shorter than gill-filaments, 
3 in eye, not very close set, spinulate below. Pseudobranchiae well 
developed. Branchiostegals 5. Pyloric caeca 10, each lobate, many- 
branched. 

D XII 34 (spines all broken, very weak), originates slightly behind 
preopercular margin. Anterior rays slightly shorter than mid-posterior, 
which are 1-7 in head. Hindmost rays slightly shorter. Bases of spines 
enveloped in skin; rays quite free, joined by membrane. 

A 35 (or I 34), originates below the 4th soft dorsal ray. Shape similar 
to dorsal; hinder rays of same length as corresponding dorsal rays. Rays 
free from skin. 

P 19, inserted low down, base slightly oblique. Fin rounded, 1-4 in 
head. 

Ventrals I 5, inserted below the pectoral base, 1-3 in head, inner rays 
longest, reach to the base of the 5th anal ray. 

Caudal damaged (forked according to G. and B., loc. cit.), 10+16 +10. 
Peduncle longer than deep, expanded posteriorly. 


55 


Marine Fishes of Seven Genera new to South Africa. 95 


Scales cycloid, very small, over whole body. Head apparently naked, 
but damaged. (On cheek, according to G. and B.’s fig., loc. cit.) Lateral 
line ascends to run. parallel with dorsal profile, almost an eye-diameter 
below. 1.1. about 125 (very approximately; G. and B.’s fig. shows 150). 

Colour (preserved).—Uniform light brown. Fins light; dorsal, anal, 
and ventrals with darkish margin. 

Length.—120 mm. 

A single specimen from Durban, presented by E. C. Chubb, Esq., 
Curator of the Durban Museum. 

There appears to be very little doubt that the present specimen should 
be assigned to pellucidus. 

The specimen is not in very good condition, but most of the features 
may be discerned. 

G. and B.’s figure shows a deeper body, more scales, a shorter snout, 
and the outer ventral rays longer than the inner. These differences are 
not significant, and may be due to the difficulty of observing accurately 
in so small a specimen as the type (about 60 mm. total length). 

This species has been recorded from the Atlantic (32° N. x 76° W.) 
from a depth of 528 fathoms, and its presence in Natal waters is of interest. 


Schedophilus medusophagus Cocco. 
(Plate V, C.) 


1876. Gunther, Challenger Rep., xxu, p. 46. 
1895. Goode and Bean, Ocean Ichth., p. 214, fig. 223. 
1902. Regan, Ann. Mag. Nat. Hist. (7), X, p. 196 (Lirus m.). 


Body very compressed, ovate, extremely soft and pliable. Regions 
at bases of dorsal and anal, containing dermal rays, distinct and differen- 
tiated. Dorsal profile evenly convex. Snout slightly swollen, somewhat 
blunt. Interorbital strongly convex. 

Depth 2-6, length of head 3-9 in length of body. Eye equal to snout 
and to interorbital width, 4:6, postorbital length 1-8, in length of head. 
Preorbital depth 1-7 in eye. Preopercle with flat spinules round angle. 
Traces of spines on interopercle. Remaining bones entire. Occiput and 
snout with thickened porous integument. Almost whole of head except 
operculum densely pitted with pores: radiating series round orbit. Nostrils 
close together, midway between snout tip and anterior margin of eye: 

Gill-openings wide, membranes free. Rakers fairly long, 15 on lower 
part of anterior arch, longest slightly shorter than gill-filaments, 2 in eye. 
Rakers spinulose on basal half. Oesophagus with 14 papillate sacs. 
Vertebrae 42 (16 + 26). 


56 


96 Transactions of the Royal Society of South Africa. 


Mouth large, slightly oblique; jaws equal. Maxilla extends to below 
behind centre of eye: upper margin slips, for the whole length, beneath 
the preorbital. Supplemental bone not obvious, maxilla not, or scarcely, 
expanded posteriorly. Teeth in a single comb-like series in each Jaw, 
larger and smaller teeth alternating, larger teeth subspatulate. Palate 
and tongue edentate. 

D 52, commences above hind margin of head. First few rays appear 
spinate, but all but tip is hidden in scaly skinny sheath. Rays increase 
gradually in length to the middle of the fin; there about 3 in head, remain 
subequal to the posterior, which are slightly shorter. The whole of the 
fin with a scaly, skinny basal investment; longest rays emerge for less 
than half of their length. Fin low, rays inclined obliquely backward, 
apparently not fully erectile. 

A 35, commences below the 20th dorsal ray; shape of fin, length of rays, 
and skinny basal investment as for dorsal. 

P 20, inserted very low, 2-2 in head, base not oblique. A scaly, skinny 
investment almost half of length of fin. Upper rays longer. 

V 15, inserted very slightly in advance of the pectorals, 3in head. Last 
ray joined to belly bya membrane. Base of fin in scaly, skinny investment. 

Caudal gently rounded, 1-7 in head. Peduncle as deep as long. Scales 
cycloid, fairly small. 1.1. 135, very slightly curved anteriorly, becomes 
straight below the 12th dorsal ray, running obliquely down to peduncle. 
Whole of body and head, except interorbital and chin, scaly. A pore below 
almost every scale on body. 

Colour.—Uniform brown (preserved). 

Length.—230 mm. 

A single specimen, found among fishes collected by S.S. “ Pickle,” 
labelled “Natal Fishes,” unclassified, presented by the Director of the 
Government Fisheries Survey to the Albany Museum. 

Regan (loc. cit.) regards Schedophilus Cocco as a synonym of Lirus 
Lowe, and states that Lirus has 25 vertebrae and a small supramaxilla. 
There is no doubt that the two genera are closely related, but the greater 
number of vertebrae of Schedophilus, as well as other minor features, would 
appear to Justify the maintenance of this genus as distinct from Lirus. 
Meristic variation is alone of doubtful validity in defining genera, but 
since the other species of Lirus all presumably have 25 vertebrae, it would 
appear reasonable in this case. Further, I can find no signs of any supra- 
maxilla in my specimen, nor is any mentioned in any description, or shown 
in any figure of medusophagus to which I have access. 

It is with considerable doubt that the present specimen is assigned to 
medusophagus, since it differs in many significant features. The heavy 
dermal investment of the greater part of the dorsal and anal, as well as the 


57 


Marine Fishes of Seven Genera new to South Africa. 97 


similar partial investment of the pectorals and ventrals, is not mentioned 
in descriptions of that species. Giinther (fide G. and B., loc. cit.) states 
that the hinder dorsal and anal rays are not “erectile into the vertical 
position.” This is true of my specimen, but the reason for this is the 
skinny sheath. Further, both dorsal and anal are longer, and there are more 
rays in the present specimen, but medusophagus is stated to vary widely in 
these counts, so that the present differences may not be of specific significance. 

medusophagus has been recorded from the Atlantic and from Samoa. 
Its presence in Natal waters is of interest. 


Family ScoRPAENIDAE, 
Setarches gtinthert Johnson. 
(Plate VI, C.) 


1876. Giinther, Challenger Rep., xxii, p. 19, pl. i, C (fidjiensis); p. 19 
(parmatus). | 
1895. Goode and Bean, Ocean. Ichth., p. 263; p. 264, fig. 249 (parmatus). 


Body moderately compressed, greatest width at nape. Head fairly 
depressed, interorbital almost plane. Depth 3-1, length of head 2-3 in 
length of body. Eye 6, snout 2-4, interorbital 4-5, and postorbital 2 in 
length of head. Head ridged and spinose. Preopercular stay prominent. 
Two long diverging spines on opercle. Suprascapula and coracoid exposed, 
each terminating in a flat spine. Three spines at angle of preopercle, 
slightly less than eye, the upper two on one side almost parallel (Pl. VI, A), 
the lower diverging: on the other side the three are equally divergent. 
Two others on lower margin, anterior smaller. Three preorbital spines 
projecting downwards over the maxilla. The anterior small and antrorse; 
the other two retrorse, graduated, the posterior the longer. Two low, 
flattened occipital spines, one similar postorbital spine, on each side, 
One antorbital spine projecting back over the eye, and one small retrorse 
nasal spine. Mouth large, somewhat oblique, maxilla highly expanded 
posteriorly, extends to below hind margin of orbit. Jaws subequal, no 
marked bony tubercle at symphysis of lower jaw. Upper jaw notched 
opposite symphysis. A narrow band of villiform teeth in each jaw. A 
chevron-shaped band, discontinuous at apex, of similar teeth on vomer. 
Similar teeth on palatines. The premaxilla has a supero-median expansion 
which slides under the maxilla. Premaxillary process short and curved. 
Nostrils tubular, close together, nearer anterior margin of eye than snout 
tip. Gill-openings wide, membranes free. Branchiostegals seven. Pseudo- 
branchiae present. Gill-rakers long, slender, 10 on lower part of anterior arch, 
anterior longest, almost equal to eye: gill-filaments short, about 3 in eye. 

VOL, XXII, PART I. 7 


58 


98 Transactions of the Royal Society of South Africa. 


D XII 10, originates slightly in advance of hind margin of head. Ist 
spine equal to eye, then increase to the 4th, which is longest, 2-6 in head: 
thereafter decrease to the 11th, which is 6 in head. 12th spine more than 
twice llth. Soft dorsal rounded, midrays longest, almost 2 in head. 

A III 5, originates below the 3rd dorsal ray. 

P 22, 1-3 in head, base fairly broad, extends to below the 4th dorsal 
ray. The upper ray and on one side 3, and on the other the 4 lower rays 
simple, the remainder branched. Ventrals 1-9 in head, do not reach vent; 
Ist ray longest. | 

Caudal almost truncate, 1-1 in head. Peduncle compressed, as deep 
as long, equal to eye. 

Scales minute, cycloid. About 95 in lateral series. Lateral line raised, 
26-27 perforations. Opercle and preopercle (above and below suborbital 
ridge) scaly. Rest of head naked. 

Colour.—Light red-brown. Fins light. 

Length.—190 mm. 

Locality Presumably Natal, being found among fishes labelled “ Natal 
Fishes,” unclassified. Collected by S.S. “ Pickle’’; presented to the Albany 
Museum by the Director of the Government Fisheries Survey. 

The three nominal species of this genus differ from one another only 
in unimportant details, and the types vary considerably in size. The 
type of guenthert, from Madeira, is 9 inches long; of fidjzensis, from the 
Fijis (315 fms.), 3 inches; while the type of parmatus, from New England 
(180 fms.), is only 2 inches in length, besides being damaged. 

The spination of the cephalic bones is evidently variable. In my 
specimen the arrangement of the preopercular spines is on one side inter- 
mediate between the arrangement on the other and the arrangement in 
gtinthert, in which they are stated to be parallel. The anterior (3rd) pre- 
orbital spine is not much exposed, and may not be obvious in the other 
specimens, especially in the smaller. 

The differences in depth of body and in the relative sizes of the fins 
of the various species described can be explained by the difference in size 
between the specimens, and cannot be regarded as of specific significance. 

Bathybial species are usually cosmopolitan, and the presence of this 
species in Natal waters, while of interest, is not in any way exceptional. 


Family BaTRacHOIDIDAE. 


Batrichthys n.g. 


Three lateral lines, each pore with minute cutaneous appendage above 
and below; middle line obscure. Jaws with tapering bands of small 
conical teeth. Larger teeth on vomer and palatines. Opercle with 2 


59 


Marine Fishes of Seven Genera new to South Africa. 99 


spines. Subopercle with 2 spines; lower shorter, parallel or slightly 
divergent. Gill-opening fairly wide, extending from above the axil to 
below pectoral base. Gill-rakers few, short, obtuse or tubercular. Pharyn- 
geal teeth equal or graduated in size. Three dorsal spines. Caudal free. 
No axillary foramen. Interorbital almost plane. Frontal ridges feeble. 
Vertebrae 29. 

Genotype albofasciatus n. sp. 


This genus is closely related to Coryzichthys Ogilby (from the Indo- 
Australian area), which is stated to have very restricted gill-openings, 
embracing only the upper half of the pectoral base, and the pharyngeal 
teeth are unequal in size. Ogilby (Ann. Queens. Mus., 1908, No. 9) * 
places gangene Ham-Buch. in Coryzichthys, despite the fact that Day’s 
figure (Fishes of India, p. 269, pl. lx, fig. 1) shows this species to have gill- 
opening at least as large as the pectoral base. Day’s figure (which is 
named grunniens, probably in error) shows 5 ventral rays, which must 
be erroneous, and he neither described nor figured any lateral line system. 

There appears to be little doubt that gangene and the species described 
below are congeneric. | 

This genus differs considerably from the two genera, Batrachoides Lac. 
and Marcgravia Jord., the only two hitherto found in South Africa. 


Batrichthys albofasciatus n. sp. 
(Plate V, B.) 


Body robust, compressed behind. Head depressed, interorbital flat, 
skin slightly rugose. Depth 5-1, length of head 3 in length of body. Eye 
5-5, snout 4, interorbital width 3-7, length of postorbital 1-6 in length of 
head. Six small barbels, about } eye, on each side of chin, and one at 
hind end of maxilla, and one at each side of upper jaw opposite symphysis. 
Conspicuous mucus pores on head, each with a minute cutaneous appendage 
above. Two short fimbriate nasal tentacles. No supra-orbital tentacles. 
Mouth large, maxilla extends to below posterior margin of orbit. Small 
curved conical teeth in posteriorly tapering bands in both jaws. Six 
larger conical teeth across vomer, with 4 widely-spaced smaller teeth 
anteriorly. A single series of similar teeth on palatines. Pharyngeal 
teeth curved, conical, of uniform size on upper; posterior teeth on lower 
pharyngeals slightly larger than anterior, no marked inequality in size. 
~ Gill-membranes fused with isthmus, membrane expanded, fused with chest 
to below behind pectoral base. Gill-opening wider than the whole width 
of the pectoral base. Branchiostegals 5. Gill-rakers short, tubercular, 


* T have not seen this paper. 


60 


100 Transactions of the Royal Society of South Africa. 


apically dilated or bifurcated, 8 on lower part of anterior arch. Two 
acute opercular spines, upper larger, divergent. Two similar on sub- 
opercle, upper larger, almost parallel. Vertebrae 29 (12 +17). 

D III+19, A 14. Soft dorsal and anal rays increase in length pos- 
teriorly, hind rays 2-5 in head. 

P 18, 1:5 in head, base heavy, sublobate. No foramen. 

Ventral 1-5 in head; spine in thickened skin. 

Caudal rounded, 1:4 in head. | 

No scales. Three lateral lines. Upper and lower distinct, middle 
obscure, consisting each of a row of pores, each with a minute cutaneous 
appendage above and below. Upper runs from opercular margin up to 
the soft dorsal and along the base of the latter. The middle line curves 
down from the same origin as upper and runs along the middle of the side, 
becomes obsolete before anal origin. The lower runs from the lower 
margin of the pectoral base down to the origin of the anal, and then along 
the base of the latter. | 

Colour.—Brown, irregularly mottled with darker, and with dark spots 
on the body, head, and all fins. Five slightly sinuous white cross-bars, 
hinder 4 extending on to soft dorsal. 

Length.—120 mm. 

Locahty.—Great Fish Point. 

Type.—In the Albany Museum. 

As indicated above, it is open to doubt whether this species should not 
be assigned to Coryzichthys Ogilby. The specimen described above is 
evidently bathybial, since it was thrown up after a storm at Great Fish 
Point, and the distended air-bladder was found to have filled almost the 
whole mouth. 

While bathybial forms are generally more or less cosmopolitan, it 
appears preferable to retain this new genus until such time as the oppor- 
tunity occurs to compare the genotype with Indian or Australian species. 

albofasciatus is very closely related to gangene, differing in fin formulae, 
in certain dimensional relationships, and in the absence of supra-orbital 
tentacles. This latter difference may prove to be sexual, and it is not 
unlikely that the two may prove to be conspecific. 


I wish to express my gratitude to Dr. Barnard, Assistant Director of 
the South African Museum, for his kind assistance and for the loan of 
literature. Also to the Carnegie Research Fund (through the Research 
Grant Board of South Africa) for financial assistance. 


ALBANY Museum, 
GRAHAMSTOWN, 
October 1933. 


Trans. Roy. Soc. 8. Afr., Vol. XXIT. Plate ¥. 


A. Scolopsis vosmeri (Bloch.). 
B. Batrichthys albofasciatus n.g. et sp. 
C, Schedophilus medusophagus Cocco. 


Neill & Co., Lid. 


Trans. Roy. Soc. 8. Afr., Vol. XXII. Plate VI. 


A. Papyrichthys pellucidus Liitken n.g. 
B. Grammonus opisthodon n. sp. 
C, Setarches giinthert Johnson. 


a, Preopercular spines on left side. 


Neill & Co., Ltd. 


61. 


Transactions of the Royal Society of South Africa. Vol. XXII. 
Part IV. pp. 321-336. Pls. XVI—XXII1. December, 1934. 


THE TRIGLIDAE OF SOUTH AFRICA. 
By J. L. B. Smira. 


(With Plates XVI-XXIII and one Text-figure.) 


(Read May 16, 1934. Revised MS. received June 18, 1934.) 


Famity TRIGLIDAE. 


Four genera of this family are admitted by Barnard (Ann. S.A. Mus., 
1925, vol. xxi, p. 938) to the South African fauna list, viz. Trigla Art., 
Peristedion Lac., Lepidotrigla Gnthr., and Chelidonichthys Kaup. 

An examination of the material in the collections of the South African 
Museum and of the Albany Museum indicated that a revision of the South 
African species of this family was necessary. 

A new sub-genus of Trigla, Trigloporus, and three new species, are 
described below. 


Key to the South African Genera. 


I. Lateral line without spines. 
A. Scales small to moderate, less than 70 in lateral line . Lepidotrigla. 
B. Scales very small, more than 70 in lateral line. . Chelidonichthys. 


II. Lateral line spinose. 
A. Body covered with bony plates, each bearing a spine. 
Preorbitals produced. . . . . . Peristedion. 
B. Body scaly. Preorbitals not produced. Whole body 
covered by a complex muciferous tube system. 
Lateral line scales with several spines. . . Trigla (Trigloporus). 


Genus LEPIDOTRIGLA Gnthr. 


1925. Barnard, Ann. S.A. Mus., vol. xxi, p. 938. 


Body covered with scales of moderate size, ctenoid, or ctenoid above 
lower third of side, cycloid ventrally. Breast naked, or partly scaly. 
Lateral line bifurcates at caudal base, the branches extending on to the 
caudal lobes. Lateral line tubes with one long oblique dorsal and a similar 
ventral branch, and one to six intermediate smaller branches, each ending 


in a pore. | 
No spines on lateral line scales, which are somewhat larger than the 


62 


322 Transactions of the Royal Society of South Africa. 


body scales, being vertically elongate. Teeth in jaws and usually on 
vomer; palatines edentate. Three free pectoral rays. A row of spinose 
plates along each side of the base of the dorsal fin. 

This diagnosis, where amended, is based upon South African material 
only. 

The species of this genus appear to be fairly numerous, but they have 
received relatively little attention from systematists, and the relationships 

o not appear to be well established. 

Of the features upon which differentiation may be based, the number 
and nature of the preorbital spines appear to be of special significance, 
but these have apparently received little attention, and they have not 
been described as minutely as they merit. The shape of the spinous 
dorsal is also of assistance, but is evidently somewhat variable even in one 
species, e.g. in natalensis G. and T. the 3rd spine varies from 1-0-1-2 times 
the length of the 2nd. The number of serrae on the Ist dorsal spine may 
also prove of assistance in differentiation. 

These fishes appear to be found below the 20-fathom line, and being 
of small size and relatively few in number, have at present no economic 
significance, in South Africa at least. All of our specimens have been 
obtained in Natal waters, and may ultimately be found in other parts of 
the Indo-Pacific area. 

The South African species first obtained and described by Gilchrist 
and Thompson (Ann. 8.A. Mus., 1914, vol. xiii, pp. 75, 76) were faurei 
and natalensis. Barnard (Ann. 8.A. Mus., 1925, vol. xxi, p. 938) later 
united these species under faurei, which has page-preference over 
natalensis. 

A careful study of the material believed to have been used by the 
original authors reveals that natalensis should be revived, but chiefly upon 
features not noticed by Gilchrist and Thompson. As will be shown below, 
it is not indeed certain that the specimen now described as faurei is one 
of those described by the original authors. 

A new species, multispinosus, well differentiated from our and from 
other Indo-Pacific species, has been added. 


Key to the South African Species. 


I. No keel on preopercle. Six or fewer preorbital spines. Pectorals 
longer than head. 
A. A patch of scales on breast. Two outer preorbital spines 


abruptly longer than the subequal inner spines . . faurei. 
B. Breast naked. Two outer preorbital spines not abruptly 
longer than the graduated inner spines. . . . natalensis. 


Ii. A strong keel on lower margin of preopercle. Ten or more spines 
on each preorbital. Pectorals shorter than head . - multispinosus. 


63 


The Triglidae of South Africa. 323 


In the following descriptions “length of head” is the distance from 
the hind margin of the upper portion of the opercular flap (i.e. excluding 
the projecting portion of the opercular spine) to the tip of the longest 
preorbital spine. This is in most cases practically the same as the distance 
from the hind margin of the opercular flap to the mid-point of the snout 
between the preorbital extensions, measured obliquely. 

Total length is measured from the tip of the longest preorbital spine 
to the apex of the mid-caudal rays. 


Lepidotrigla faurer G. and T. 
(Plate XVI, A; Plate XVIII, A; Plate XIX, A, C.) 


1914. Gilchrist and Thompson, Ann. §.A. Mus., vol. xiii, p. 75. 
1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 938. 


Depth 4-1, length of head 3-1 in length of body. Eye 3-9, interorbital 
4-9, snout (to tip of preorbital spine) 2-6, distance between apices of longest 
preorbital spines 3-5 in head. Dorsal profile of snout steep, almost straight. 
Snout between the projecting preorbitals scarcely emarginate. Pre- 
orbitals each with 6 spines, the outer 2 large, subequal in length, the 
inner 4 subequal, very abruptly smaller than the outer (Pl. XVIII, A). 
No preopercular spine or keel. Two inconspicuous spines on supero- 
anterior margin of orbit. A short transverse groove at the supero-posterior 
margin of each orbit. Nuchal spines feeble, reach beyond the base of the 
3rd dorsal spine. Opercular and humeral spines moderate. Jaws with 
bands of small conical teeth. Vomerine teeth doubtful. Palatines edentate. 
Pyloric caeca cannot be determined as the intestines have decomposed. 
Seven gill-rakers plus several rudiments. 

D VIII+16, inserted above the hinder third of the opercular flap. 
Ist, 2nd, and 3rd spines serrate, 38 serrae on Ist spine, those on 3rd spine 
very small. Ist spine 2-0, 2nd 1-7, 3rd 1-6 in length of head. 4th-6th 
dorsal rays longest; 5th ray 1-9 in head, 5-7 in length of body; 14th ray 
2-3 in head, 7-0 in length of body. 14th ray laid back just reaches upper 
margin of caudal base. 

A 16, inserted below the soft dorsal origin. Lower than soft dorsal. 

P 11+3, 1-13 times head, 2°75 in length of body ; tip reaches to below 
the base of the 7th dorsal ray, and to below the 29th lateral line scale. 
First upper and lower three of connected rays simple. Longest free ray 
1:3 in head, tip does not reach ventral tip. 

Ventrals 1-2 in head, reach to base of 3rd anal ray. 

Caudal slightly emarginate, peduncle 1-4 times eye. 

Scales moderate, ctenoid. dorsally to level of humeral spine, graduated 
into cycloid on ventral area (Pl. XIX, A and C). Lateral line scales 60; 


64 


324 Transactions of the Royal Society of South Africa. 


tubes with 3-8 branches, longest above and below. No pores on body. 
22 spines along dorsal base. Breast with a median patch of moderate 
cycloid scales. | 

Colour.—Preserved, uniform red-brown. Inner surface of pectorals 
dark, with light margin above and below. 

Length.—143 mm. (S.A. Mus., No. 11803). 

Locality.—Natal, off Tugela mouth, in 20-63 fathoms. 

In the South African Museum are eight adult specimens of Lepidotrigla, 
these having been registered as the material, collected by the s.s. “‘ Pieter 
Faure,” upon which Gilchrist and Thompson based their diagnoses of faurea 
and of natalensis. These eight specimens have the following numbers, 


lengths, etc.:— 


S.A.M. No. . . | 11797 | 11797 | 11797 | 11803 | 11812 | 11812 | 11812 | 11812 
Total length . | 178 | 174 | 160 | 143 | 167 | 162 | 155 | 135 
Body length . | 154 | 149 | 134 | 120 | 144 | 136 | 1380 | 120 
Dorsal spines . 9 9 8 8 9 8 8 8 


According to the South African Museum register, two of the three of 
No. 11797, plus No. 11803, are Gilchrist and Thompson’s orthotypes 
of faurei, and one of those of No. 11812 is the holotype of natalensis. 

According to Gilchrist and Thompson (loc. cit.), the three types of 
faurei were of length 120 mm., 120 mm., and 146 mm. None of the above 
are as small as 120 mm., unless these authors recorded body and not total 
length. Otherwise, some of their measurements were erroneous, or there 
has been a loss of specimens. 

In the case of natalensis the holotype was stated to be 120 mm. in 
length, and it is possibly the smallest of those of No. 11812. In any case, 
there has obviously been some lack of care in the designation and pre- 
servation of Gilchrist and Thompson’s orthotypes, which is particularly 
regrettable in the present case, since a careful study of the eight specimens 
reveals that No. 11803 is well differentiated from the remaining seven, 
which are undoubtedly conspecific. 

The original descriptions of fauret and of natalensis reveal nothing 
of significance whereby these two species may be distinguished from each 
other. L. faurei was stated to have “two strong spines” on each pre- 
orbital, and ctenoid scales, and natalensis to have “a row of strong spines” 
on each preorbital, and cycloid scales. 

None of the eight South African Museum specimens have only two 
preorbital spines, but in No. 11803 the two outer are very abruptly longer 
than the subequal inner spines (Pl. XVIII, A). In none of the remaining 


65 


The Trighdae of South Africa. | 325 


seven specimens are the two outer spines so abruptly differentiated from 
the inner (Pl. XVIII, Band C). Further, in all of the specimens, the scales 
above the lateral line are ctenoid, but cycloid on the ventral surface. 

It is therefore impossible to select with certainty the orthotypes of 
Gilchrist and Thompson, but there is no alternative except to revive 
natalensis. In view of the nature of the preorbital spines of No. 11803, 
this (a ripe female) is here designated the holotype of faurei, and the 
remaining seven specimens (male and female) listed above may be designated 
the lectotypes of natalensis, or, alternatively, No. 11812 of length 135 mm. 
might be selected as the orthotype. 

L. fauret is well differentiated from natalensis not only by the nature 
of the preorbital spines, but also by the presence of scales on the breast, 
and by the longer pectorals and soft dorsal rays. The scales of fawres are 
more strongly denticulate than equivalent scales of natalensis. 

It is, however, remarkable that among over 50 juvenile specimens, 
taken with these eight adults, none show any signs of a scaly breast, and 
should no further specimens of fawrei (as here defined) be discovered, it 
may be suspected that it is merely an abnormal specimen of the form now 
recognised as natalensis. 


Lepidotrigla natalensis G. and T. 
(Plate XVI, B; Plate XVIII, B, C; Plate XIX, B, D, H, F.) 


1914. Gilchrist and Thompson, Ann. 8.A. Mus., vol. xii, p. 76. 

1925. Barnard, Ann. §.A. Mus., vol. xxi, p. 938 ( faures part ?). 

Depth 4-0-5-0, length of head 3-0—-3-2 in length of body. Eye 3-4—4-0, 
interorbital 4-0-5-0, snout (to tip of preorbital spine) 2-2-2-5, distance 
between apices of longest preorbital spines 3-2-3-7 in head. Dorsal profile 
of snout steep, almost straight. Snout between the projecting preorbitals 
almost straight to emarginate. Preorbitals each with 4-6 spines, the outer 
the largest, graduated to the inner, no abrupt change in size (Pl. XVIII, 
B and C). The form and arrangement of these spines are somewhat 
variable, and some specimens have 1-2 outer smaller spines at the base 
of the largest. In very young specimens there are rarely more than 2-3 
spines visible, and these are highly incurved, as if the spinose preorbital 
extension develops from the inner side with growth. No marked pre- 
opercular spine in adults; a small pungent spine in very young specimens, 
which becomes more or less obsolete with growth. No preopercular keel. 
Two inconspicuous spines on supero-anterior margin of orbit; there are 
2-3 moderately prominent spines in juveniles which diminish with growth. 
A short transverse groove at the supero-posterior margin of each orbit, 
obscured on one or both sides in some specimens by an outgrowth of the 

VOL. XXII, PART IV. 23 


66 


326 Transactions of the Royal Society of South Africa. 


bony integument. Nuchal spines fairly strong, reach from almost below 
to beyond the base of the 8rd dorsal spine. Opercular and humeral 
spines moderate. 

Jaws and vomer with bands of small conical teeth. Palatines edentate. 
Gill-rakers 7-9 plus several rudiments. Pyloric caeca 6-8. 

D VIII-IX +16, inserted above the hinder third of the operculum. 
Ist, 2nd, and 3rd spines serrate, 42-46 serrae on first spine, those on 3rd 
very small. 1st spine 1-7-2:1, 2nd 1-6-1-8, 3rd 1-5-1-7 in length of head. 
3rd spine 1:03-1-2 times second. 4th-6th dorsal rays longest; 5th ray 
2-1-2:5 in head, 6-2-7-0 in length of body. 14th ray 2-6-2-7 in head, 
8-0-8-6 in length of body. 14th ray when laid back does not reach the 
upper margin of the caudal base. 

A 16, inserted below origin of soft dorsal, shape similar to that of soft 
dorsal, but rays shorter. 

P 11+3, 1-0-1-08 times head, 3-0-3-2 in length of body, tip reaches to 
below the base of the 3rd—5th dorsal ray, and to below the 24th-25th 
lateral line scale. First upper and lower three of connected rays simple. 
Longest free ray 1-4 in head, does not reach ventral tip. 

Ventrals 1-1—-1-2 in head, reach origin of anal or just beyond. 

Caudal slightly emarginate, peduncle 1-2—1-4 times eye. 

Scales moderate, weakly ctenoid dorsally to level of humeral spine, 
graduated into cycloid on ventral area (Pl. XIX, B, D, and E). Lateral 
line scales 58-61, tubes with 3-6 branches, with longest above and below 
(Pl. XIX, F). Sometimes a few pores on body, most marked ventrally, 
occasional scales perforated. 23-25 spines along dorsal base. Breast 
naked. 

Colour.—Preserved, uniform red-brown. Inner portion of pectorals 
dark with light margin above and below. 

Length.—35-178 mm. 

Localhity.— Natal, off Tugela mouth, up to 63 fathoms. 


Lepidotrigla multispinosus n. sp. 
(Plates XVII and XX.) 


Dorsal profile of snout undulate, abruptly descending, concave before 
eyes, convex on anterior part, with abrupt descent at tip. Depth 4, 
length of head 3-1 in length of body. Eye 4-0, interorbital 4-8, snout 
2-0, postorbital part of head 3-1 in length of head. 

Rostrals, with slight, widely diverging, medio-longitudinal ridge, 
length 11 in head (measured between parallels of mid-point of snout and 
tip of process). At the apex are small spines, mostly invested; on the 
left process there are 5 small outer spines and 12 inner, the first (outer) 


67 


The Triglidae of South Africa. 327 


of the 12 by far the largest; on the right are 10 spines, with the outer 
longest (Pl. XVII, C). Distance between the apices of the longest rostral 
spines 3 in head. Lateral keel of preorbital very faint, ending below and 
not continuous with the preopercular ridge. A distinct ridge on lower 
margin of preopercle, 1:3 times eye, slightly oblique downwards to the 
angle, at which there is no spine. Two weak spines above the anterior 
part of the orbit; one larger above the hind margin. On each side one 
weak occipital, and two nuchal spines, the hinder the larger; apex of this 
just reaches the base of the 3rd dorsal spine. Interorbital deeply concave. 
The granules on the cheek radiate fan-wise backwards from below the front 
margin of the eye; similarly on the*opercle from the base of the spine. 
On the preorbital the arrangement is irregular. A short deep groove at 
the supero-posterior margin of the orbit continuous, angularly bent back- 
wards across the occiput. Humeral spine 3, free distal portion 6 in head. 

Mouth sub-inferior, lower jaw included, maxilla 2-6 in head. Villiform 
teeth in bands in both jaws, wider in upper jaw. Vomerine teeth doubtful. 
None on other bones. Tongue adnate. Large pores on lower surface 
of mandibles. No barbels. Branchiostegals 7. Membranes free from 
isthmus. Gill-rakers 5, longest 5 in eye, plus 4 anterior rudiments. 
Pyloric caeca 6, moderate. 

D VIII+15, inserted above the hind margin of the operculum. Ist 
spine 2-1, 2nd and 3rd subequal, 1-8 in head. 1st spine with 24 anterior 
spinules, 2nd and 3rd weakly spinate. Base of Ist dorsal 1-8 in head. 
Longest soft ray (5th-8th) 3-5 in head. Base of soft dorsal 1-1 times head. 
On one side 23, on the other 25 spines below the dorsal. 

A 16, originates below the origin of the soft dorsal. 

P 11+38, 1:2 in head, reaches to below the base of the 7th dorsal ray. 
Longest (upper) free ray, reaches almost to ventral tip, 1-4 in head. 

V I, 5, reach to base of 3rd anal ray. 

Caudal almost truncate, peduncle twice eye. 

Scales (mostly shed) relatively small, ctenoid to below the pectoral 
base, ventral scales with a single large denticle (Pl. XX, A, B, and C). 1.1.59. 
Each tube with an upper and a lower oblique branch, and from one to four 
smaller intermediate branches, each ending in a pore (Pl. XX, D). Breast 
naked. 

Colour.—Preserved, uniform red-brown. Pectorals dark. 

Length.—160 mm. 

Type in the Albany Museum. 

This specimen was among certain fishes presented to the Albany Museum 
by the Director of the Government Fisheries Survey. It was without 
serial number or locality, although probably from Natal. 

It is very clearly distinct from our other species, the preopercular keel, 


68 


328 Transactions of the Royal Society of South Africa. 


the number of the preorbital spines, the long peduncle, the smaller scales, 
and the relatively short 3rd dorsal spine being distinctive features. This 
species appears to be well differentiated from all others by several features, 
notably the deep excavation between the long multidentate preorbital 
extensions and by the preopercular keel. 


Genus CHELIDONICHTHYS Kaup. 


1925. Barnard, loc. cit., p. 939. 


Scales very small. Three free pectoral rays. Palatines edentate. A 
row of spinose plates along the base of the dorsal fin. Lateral line without 
spines, bifurcating at the caudal base. Preorbitals produced into spines. 

Kurope, Africa, Japan, and Australia. 

This genus appears to be quite valid, but is apparently not recognised 
by some systematists. 

As far as the South African species are concerned, Barnard’s treatment 
requires no revision, the descriptions being in the main accurate and 
detailed, and the differentiations clearly defined. 


Key to the South. African Species. 


I. Breast naked. Preorbitals ending in several (usually incurved) spines. 
A. Eye not greater than interorbital width . . . . - capensis. 
B. Eye distinctly greater than interorbital width . . . . kumu. 
Il. Breast pitted. Preorbital ending in one large straight outer spine, with 
inner concealed fused spines. . . . . . . -  quekettr. 


capensis C. and V. appears to be well differentiated from any European 


species, as indicated by Barnard (loc. cit.). This species is of moderate 
economic significance as a food-fish in South Africa. 


Fic. 1.—Preorbital spines of Chelidonichthys queketti Rgn. 
X =Inner fused spines, concealed by membrane. 


Barnard (loc. cit.) has compared our specimens of kumu L. and G. with 
Australian specimens, and his opinion that they are conspecific is here 
accepted. 

quekettt Rgn. is apparently an endemic species. Previous workers do 
not appear to have noticed that, besides the large outer preorbital spine, 


69 


The Triglidae of South Africa. 329 


there is on each side, concealed within the basal skinny membrane, an 
inner process, which has obviously been produced by the coalescence of 
from four to eight smaller spines (fig. 1). 


Genus PERISTEDION Lac. 


1895. Goode and Bean, Ocean Ichth., p. 470. 
1913. Weber, Siboga Exp. Monogr. 57, p. 511. 
1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 944. 


Body covered with bony plates, each bearing a spine. Mouth inferior, 
lower jaw included, with barbels on the lower margin, the outer the longest. 
Kach preorbital greatly enlarged and produced forwards as a spatuliform 
process, with lateral keel continuous along lower margin of preopercle, 
which ends in a spine. No teeth. Dorsal single or divided. Two free 
pectoral rays. 

Two species have been found in our area, adent Lloyd, and what has 
hitherto (though with doubt by Barnard, loc. cit., p. 946) been accepted 
as gracile G. and B. This latter species has now been found not to be 
conspecific with gracile, and is here described as webert n. sp. 

Many other species will doubtless be found in our area with more 
intensive collecting, and it is not unlikely that at least some of the relatively 
numerous Indo-Pacific forms will be among them. 


Key to the South African Species. 


I. No spines on dorsal surface of snout. 1.1. 34 . . . .  webert. 
II. One median and 4 smaller spines on dorsal surface of snout. 1.1. 30 . adeni. 


Peristedion weberi Nn. Sp. 
(Plate X XI.) 


1924. Gilchrist and von Bonde, Fish. Mar. Surv. spec. Rep. ui, p. 22 
(gracile). 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 945 (gracile ?). 

Dorsal profile abruptly descending before upper half of eye, thereafter 
slopes more gently. Depth 5-4 (5-0), length of head including rostral 
process 2:4 (2-1), rostral process 8-5 (7:5) in length of body. (The figure 
in brackets gives length in distance from base of caudal to tip of maxilla.) 
Eye 5:5 (4-2), snout 2-9 (2-2), snout plus rostral process 1-6 (1-2), rostral 
process 3-5 (2-7), interorbital width 6-5 (5:0) in length of head (including 
rostral process; the figure in brackets is length in length of head without 


rostral process). 
Interorbital deeply concave, supraorbital ridge moderate, terminating © 


—70 


330 Transactions of the Royal Society of South Africa. 


in a flat spine, followed by two retrorse spines. A small serrated suborbital 
keel terminating in a feeble blunt spinose projection. The widely diverging 
rostral processes are continued backwards as serrated ridges, each ter- 
minating at the angle of the preopercle, hind margin scarcely acute (PI. XXI, 
Band C). No nasal or frontal spines. No spine at base of rostral process. 
Numerous pores on head and along lower edge of mandible. 8-9 groups 
of bi- or trifid barbels on each side of the lower jaw, longest fimbriate, equal 
to orbit. (Pl. XXI, A, pencilled in on photograph.) 

Maxilla 1-5 times eye. No teeth in jaws or on palate. Tongue adnate. 
Gill-rakers 20, longest 2:5 in eye. 30 spinose plates on dorsal, 28 on 
ventral surface. The anterior ventral plates are 2-3 times as long as 
broad, 7.e. the width of the two plates is 1:15 in the length of one plate 
(Pl. XXI, D). A rectangular portion of the second plates, five times wider 
than long, fits into a recess in the anterior plates. 

D VIII, 20, originates an eye-diameter behind eye, above the hind 
margin of the operculum. 4th spine longest. Ist spine 3-5, 2nd and 
3rd subequal 3-2, 4th 2-9 in length of head (without rostral process). Ist 
ray 0:4, 2nd 4-3, 3rd 3-6, 4th-8th longest 3-2 in length of head (without 
rostral process); thereafter the rays decrease. 

A 20 (or I, 19), inserted below the base of the 4th dorsal ray. Shape 
and length of raysasin dorsal. Base of anal 5-6 times longitudinal diameter 
of eye. 

P 11+2, 2:2, longer detached ray 1-6, lower 1:8 in head (without 
rostral process). Upper detached ray reaches well beyond ventral tip. 
Ventrals 1-9 in head (without rostral process) reach to vent. Membrane 
for-half-length of last ray joined to body. 

Caudal emarginate, mid-rays 1-1 times eye. 

The lateral line descends abruptly from the shoulder to the middle of the 
side. 34 spinose plates in the lateral line, 4 of these on the downward 
portion. Lateral line tubes bifurcate, ending in a pore above and a pore 
below the spine (Pl. XXI, E). The spines on the anterior 23 plates are 
simple, those on the 11 posterior plates have an anterior antrorse and a 
posterior retrorse point (Pl. XXI, E). 

Colour.—Preserved, uniform light brown. Traces of dark mottling on 
distal half of pectoral. 

Length.—185 mm. 

Locahity.— Off Delagoa Bay, in 260 fathoms. 

Type in the Albany Museum. 

A single specimen, labelled P. gracile G. and B., presented to the Albany 
Museum by the Director of Fisheries Survey. 

According to the Fisheries Survey records (Mar. Bio. Rep., 1921. (2), 
p. 10), 8 specimens, subsequently identified as Peristedion gracile G. and B. 


71 


The Triglidae of South Africa. 331 


(Gilchrist and von Bonde, Mar. Bio. Rep., 1924, 11, p. 22), were obtained 
off Delagoa Bay, in 260 fathoms; the specimen described above was one 
of these. : 

| Barnard (loc. cit., p. 947) was very doubtful of the validity of this 
identification, chiefly because gracile is'a Mexican species. This has led 
me to make a critical examination of the specimen in the Albany Museum. 

The original description of gracile (Goode and Bean, Ocean. Ichth., 
1896, p. 473, fig. 387) is based upon a single type specimen about 125 mm. 
in length. (There are several discrepancies in this description.) 

Although reasonably close to gracile, the specimen described above is 
evidently not conspecific. The shape of the spinous dorsal, the longer 
anal base, the bispinose posterior lateral line plates, and the number of 
plates, as well as the ridge below the eye, all distinguish the present specimen 
from gracile. There are also fewer gill-rakers, but Goode and Bean may 
possibly have given the number (26) on the whole arch. The original 
description of gracile does not say whether the rostral processes are divergent, 
but the two species have much in common. This is another case of remark- 
ably close relationship between Indo-Pacific and Atlantic species. 

weber is also related to rivers-andersoni Alcock, and to nierstraszt Weber, 
from the Indo-Pacific. I have seen only the original description and 
figures of the former species (Alcock, 1894, J. Asiat. Soc. Bengal, xii, 
pt. 2, p. 12, pl. vi, figs. 2, 2a, 2b). Dr. de Beaufort of Amsterdam has 
kindly lent me one of the co-types of the latter species. An examination 
of this indicates that nierstraszi is of doubtful validity, since the specimen 
agrees in most particulars with the description of rwers-andersonz, especially 
if the difference in size between the types of the two species be taken into 
account, e.g. the posterior lateral line plates of this specimen of nierstraszt 
are bispinose, and there is a small but distinct spine on the dorsal surface 
of the base of each rostral process. 

Since I have not seen a specimen of rivers-andersons, [| cannot venture 
to state that nierstraszi is conspecific, although it appears quite possible. 

In any case weberi is quite definitely distinct from either, in the nature 
of the more slender and diverging rostral processes alone. 

In regard to gracile, it would not indeed be surprising if a re-examination 
of that species showed that weberi is even more closely related than appears 
here. It is highly probable that gracile actually has bispinose posterior 
lateral line plates. 


Peristedion adeni Lloyd. 
(Plate XXII.) 


1907. Lloyd, Rec. Ind. Mus., vol. i, p. 8. 
1908. Alcock, Illustr. Zool. Invest. Fishes, pl. xl, figs. 1, la. 


72 


332 Transactions of the Royal Soctety of South Africa. 


1922. Gilchrist, Fish. Mar. Surv. Spec. Rep. iu, p. 78. 
1925. Barnard, loc. cit., p. 945. 
21925. Fowler, Proc. Ac. Nat. Sci. Phil., vol. Ixxvii, p. 256. 


Dorsal profile of snout gently sloping before eye, scarcely concave. 
Depth 5-1, length of head (without rostral process) 2-6, (with rostral process) 
2-2 in length of body (without rostral process). Eye 5-5, interorbital width 
4-7, snout (with rostral process) 1-4, (without rostral process) 1-7, rostral 
process (broken) about 6 in length of head (excluding rostral process). 

Interorbital moderately concave, supraorbital ridges moderate, with 
a moderate spine on the posterior margin. On each side of the nape a 
large spine, and a smaller one below. One median frontal spine. On 
each side one small antorbital and one small nasal spine. The preorbital 
process is large, flattened, with inner edges subparallel, outer margins 
converging; the outer margins produced meet at an angle of 30°. Width 
of process at base 1-2 in eye. The lateral edge of the preorbital process 
is continued backwards as an undulate keel, terminating in the large pre- 
opercular spine (Pl. XXII, B). There isa short and narrow suborbital keel. 

Mouth large, maxilla 2:2 times eye. No teeth, tongue adnate. On 
each side of the lower jaw there are 7-8 single small barbels, and one large 
outer, fimbriate, 2-8 times eye. Gill-rakers 17, longest 3 in eye, anterior 
rakers short but graduated. 

21 spinose plates on ventral, 24 on dorsal surface. The anterior 
ventral plates are together 1-1 times wider than the length, i.e. each 
plate is 1-8 times longer than wide (Pl. XXII, C). 2nd plates each slightly 
wider than long. A quadrangular portion of the 2nd plates, five times 
wider than long, fits into a recess in the anterior plates. 

D VII+15, inserted above hind margin of opercle, 3rd spine longest, 
3-1 in head (without rostral process). 18 spinose plates along base. 

A 15, inserted below the base of the 2nd dorsal ray, base five times 
longitudinal diameter of eye. 

P11+2, 1-9, upper detached ray 2-6 in head (without rostral process). 
Tip of pectorals reaches to below the 11th lateral line plate. 

Ventrals 2-0 in head (without rostral process), just reach vent. Mem- 
brane joined to body. 

Caudal damaged. 

The lateral line descends abruptly from the shoulder to the middle of 
the side. 30 lateral line plates, 5 of these on the downward portion. 
The anterior 20 unispinose, the next 7 bispinose, and the last 3 unispinose 
(Pl. XXII, D). The lateral line pores do not show externally. 

Colour.—Preserved, uniform light brown, pectorals darker. 

Length.—About 225 mm. 


73 


The Trighdae of South Africa. 333 


Locality.— Natal. 

A single specimen (S.A.M., No. 16223, from Mr. Bell Marley). 

The specimen described above was identified as adent Lloyd by Dr. 
Barnard (loc. cit.), with which identification I am provisionally in agreement. 

There are, however, certain variations from the original diagnosis, the 
most marked of which is the presence of the mid-posterior bispinose lateral 
line plates. These are not as obvious as in the species (webert) previously 
described, and may have escaped notice. 

The original description makes no mention of the antorbital and nasal 
spines. Barnard (loc. cit.) mentions these, but does not draw attention 
to the fact that they were not present in the type diagnosis. 

In this specimen the outer margins of the preorbital processes would 
meet at a more acute angle than those of the type, 1.e. they are less markedly 
convergent, but this feature is not of importance, and might vary con- 
siderably. Further, there is no mention of the suborbital keel which is 
present in this specimen. 

These combined differences might, if established, justify the separation 
of the present specimen from adeni, but as I have not access to any type 
of that species, it would be unwise to venture this step. 


Genus Tricia Art. 
1925. Barnard, loc. cit., p. 943. 


Body covered with very small scales, ctenoid above, cycloid on ventral 
surface. Three free pectoral rays. Lateral line scales spinose. A row 
of spinose plates along the base of the dorsal fin. No palatine teeth. 
Lateral line bifurcates at the caudal base. 


Trigloporus new sub-genus. 


Distinguished from Trigla (sensu stricto) by the whole body (excepting 
the chest) being covered with a complex system of reticulate tubes (each 
ending in a pore), between rectilinear transverse multiporose tubes, as 
well as by the multispinate nature of the lateral line scales. The pectorals 
are also probably longer than in Trigla (Trigla). 

Kurope and Africa. 

Genotype, africana n. sp. 

This sub-genus is proposed with considerable doubt as to its validity, 
since I have seen, besides lineata Gmel., only one other Kuropean species 
of Trigla (sensu lato), viz. gurnardus Linn. The extraordinarily highly 
developed mucus canal system of lineata and of africana (which are clearly 
congeneric in all respects), as well as the spination of the lateral line scales, 
would appear to justify full generic distinction of the latter two species 


74 


334 Transactions of the Royal Society of South Africa. 


from gurnardus. Against this, however, is an account of the species of 
Trigla by Smitt (Skand. Fish., 1892, pt. 1, p. 194), of which I have seen 
a brief abstract, in which he states that the species of this genus show 
practically all stages of development of the mucus canal system intermediate 
between that of gurnardus and that of lineata. Despite this, the difference 
between these species is very striking, but since material to establish or 
refute the absolute accuracy of Smitt’s statement is not available, as a 
compromise, T'rigloporus is proposed as a sub-genus of Trigla. 


Trigla (Trigloporus) africana n. sp. 
(Plate XXIII.) 
1925. Barnard, loc. cit., p. 943 (lineata Gmel.). 


Dorsal profile of snout fairly steep, gently concave. Depth 4-5-5, 
length of head 3-5-3-7 in length of body. Eye 4-3-4-5, interorbital 4-0-4:5, 
snout 2-2, postorbital part of head 2-5, preorbital depth 2-4—2-5 in length 
of head (measured from the hind edge of the operculum at the spine to 
the snout tip). Front of snout rounded, or with slight median concavity. 
The preorbitals are not produced, and there are no obvious spinules. 
Interorbital deeply concave. Two to four backwardly radiating antero- 
supraorbital ridges, each ending above the orbital margin in a small spine. 
One to three minute supero-postorbital spinate ridges, without any cross 
groove behind. One to three nuchal spines, the infero-posterior the 
longest, just reaches below the base of the Ist dorsal spine. Humeral 
spine very short, apex reaches beyond the hind margin of the operculum, 
a distance equal to 5-5-6-2 in head. No keel on preorbital, the lower edge 
of which is scarcely serrate. One set of backwardly radiating granules 
from anterior margin of preorbitals; three radiating sets on cheek, lower 
above end of maxilla, median largest in centre of cheek, and a smaller 
group below hind margin of eye. Preopercular and opercular spines very 
feeble. | 

Maxilla extends to below the hind margin of the eye. No vomerine 
teeth. The pores on the lower surface of the rami are arranged in 5-6 
more or less regular longitudinal series, close set, but not aggregated into 
groups. 

D X+15-16, inserted above slightly in advance of hind margin of 
operculum. Ist spine 1-4-1-5, 2nd, longest, 1-3-1-4, 3rd 1-4-1-5 in head. 
Ist spine with a few rudimentary spinules or granules on the lower 
anterior edge. Soft rays 2-5 in head. 

A 15-16, inserted below the soft dorsal, rays similar to dorsal. 

Pectorals 1-25-1-3 times head; tip reaches to below the 4th—5th dorsal 


75 


The Trighidae of South Africa. 335 


ray. Longest free ray 1-1-1-2 in head, tip reaches to the posterior third 
of the ventral. | 

Ventrals 1-0-1:2 times head, reach to the base of the 2nd—4th anal 
ray. | 

Scales very small, strongly ctenoid above the lateral line (Pl. XXIII, C), 
grade into cycloid on ventral surface. The whole of the body is covered 
with the reticulate system of mucus canals described earlier, the transverse 
rectilinear tubes resembling lateral folds of the skin between the scale 
rows (Pl. XXIII, B). 67-70 scales in the lateral line, each scale bearing 2-5 
spines, each much smaller than the single spine on the lateral line scales 
of the species gurnardus Linn. Breast partly or wholly scaly. 

Gill-rakers 7-8. Pyloric caeca 7-9. 

Colour.—Red-brown, mottled and spotted, with 4-7 faint cross-bars. 
Spinous and soft dorsal and caudal with irregular blotches. Ventrals 
reddish. Pectorals bluish, with irregular cross-mottlings; free rays with 
several brownish blotches. 

Length.—Up to 230 mm. 

Localities.—Cape St. Blaize, 26-33 fathoms; Algoa Bay, 40 fathoms; 
Port Alfred, cast up on the shore. 

Holotype, from Port Alfred, in the Albany Museum. 

The South African specimens have hitherto been regarded as identical 
with the European species lineata Gmelin. 

Since I had access to no recent detailed descriptions of that species, 
I forwarded a specimen of our species to Mr. Norman of the British Museum, 
requesting him to compare it with the European form, and, if there appeared 
any doubt of conspecificity, to send me material for comparison. He has 
kindly sent this, and there appears to be little doubt that the South African 
species is well differentiated. 

The following table indicates some of the numerous features by which 
the two species are differentiated :— | 


lineata africana. 
Pectoral times head _. ; ; 1-6 1-2-1:3 
Humeral spine in head . ; ; 3-4 5-5-6-2 
Eye in head ; ; ; ; 3°4 4-3-4:5 
Pores on rami ; ; , . | In groups of 5-6 | In uniform rows 
Breast ; ; . Naked Scaly 
Lower margin of preorbitals ; ; Serrate Smooth 
Dorsal spines ; 12 10 
Anterior margin of first dorsal spine | Serrate whole Feebly serrate 


length basally 


76 


336 Transactions of the Royal Society of South Africa. 


Besides these, the dorsal profile of the snout of lineata is much steeper, 
and the angle subtended by the corners of the mouth at the symphysis is 
much less than in our species, while the spination and granulations on 
the head of lineata are much more marked than in africana, and the spinous 
dorsal is markedly larger. 


I wish to express my gratitude to Dr. Barnard, Assistant Director of the 
South African Museum, for the very considerable assistance he continually 
affords me in providing extracts from literature not available here, as 
well as for freely lending material from the South African Museum collection. 
Also to the Carnegie Research Fund, through the Research Grant Board 
of South Africa, for generous financial assistance. 

ALBANY MUSEvuM, 


GRAHAMSTOWN, 
June 1934. 


Plate XVI. 


pue 


“WH 
K) sisuappyou njbr.uopideT 


L squosoider oinsy yore MOTEq SUI] OI, 
ze "L pue 


0 rainof op br.yoprdaT 


V 


Neill & Co., Lid, 


J. L. B. Smith. 


Trans. Roy. Soc. S. 


J. iL. B. Smith, 


Afr., Vol. XXII. 


| . 
HE 


Plate XVII. 


Neill & Co., Ltd. 


C. Preorbital spines. 


B. Dorsal view of head. 


The line below each figure represents 1 cm. 


(Type.) 


Ispinasus N. sp. 


A. Lepidotrigla mult 


rans. Roy. Soc. 8. Afr., Vol. XXTI. Plate XVIII. 


Preorbital spines, x 5, of Lepidotrigla species: 


. fauret G. and T. 
B and C. natalensis G. and T., showing limits of variation. 


J. iL. B. Smith. Veill d Co., Ltd 


Trans. Roy. Soc. 8. Afr., Vol. XXII. 


‘i 


- 


: 


‘ ih \ 
' i “ ‘ 


iN 
4 \ 
a 


Plate XIX. 


‘ 


Scales of Lepidotrigla species: 


A. Above lateral line, below spinous dorsal, 
of faure: G. and T. 

B. Equivalent scale of natalensis G. and T. 

C. Ventral, behind breast, of faurer G. and T. 


The line below each scale 


J. L. B. Smith. 


D. Equivalent scale of natalensis G. and T. 


E. From behind pectoral base of natalensis 
G. and T. 
F. 31st lateral line scale of same. 


represents 1 mm. 


Neill & Co., Lid. 


Trans. Roy. Soc. 8S. Afr., Vol. XXII. Plate XX 


Scales of Lepidotrigla multispinosus n. sp. 


A. Above lateral line, below spinous dorsal. _ C. Ventral, behind breast. 
B. Behind pectoral base. D. 30th lateral line scale. 


The line below each scale represents 1 mm. 


J. L. B. Smith. Neill & Co., Ltd. 


Plate X XI. 


ni 
Ha 


Reet 
wana 


A. Peristedion weberi n. sp. D. Anterior ventral plates. 


B. Ventral surface. KE. 21st (right)-27th lateral line spines from above. 
C. Dorsal view of head. The line with each figure represents 1 cm. 


J, LL. B. Smith. Veill & Co., Ltd. 


Trans. Roy. Soc. 8. Afr., Vol. XXII. Plate XXII 


A. Peristedion adeni Lloyd. C. Anterior ventral plates. 
B. Dorsal view of head. D. 20th (right)—24th lateral line spines from above. 


The line below each figure represents | cm. 


J. LL. B. Smith. Neill & Co., Lid. 


Trans 


Roy. Soc. 8. Afr., Vol. XXII. 


J. L. B. Smith, 


Plate XXIII, 


Neill & Co., Lid, 


B. A portion of the skin below the lateral line, showing a few lateral line scales. 
The line below B and below C represents 1 mm. 


i (Type.) 


° S1IZe e 


Nat 


1cana Nn. sp. 


Trigla (Trigloporus) afr 
C. Scale from below the spinous dorsal, above the lateral line. 


A. 


77. 


Annals of the South African Museum. Vol. XXX. Part 5. 
pp. 587-644. Pls. XV—XXII. February, 1935. 


19. The Fishes of the Family Mugilidae in South Africa.—By J. L. B. 
SMITH, M.Se., Ph.D., Hon. Curator of Fishes, Albany Museum, 
Grahamstown. 


(With Plates XV-XXII and 17 Text-figures.) 


[Members of the family Mugilidae—known in South Africa as ‘‘ Harders’’—are 
an important edible commodity. Pappe refers to their value, and an enthusiastic 
praise of them occurs in the memoirs of Lady Anne Barnard. From the culinary 
point of view one species of Harder appears to be as good as another, but for 
scientific purposes it is important to have the species occurring in our waters 
defined as accurately as possible. In this paper Dr. J. L. B. Smith—a worthy 
successor of Dr. Andrew Smith in the early part of last century—has tackled a 
very difficult problem in systematics, which previous writers have left severely 
alone.—Ep.] 


Family MUGILIDAE. 


The South African species all fall within the single wide genus 
Mugil Linn. | 
Genus Mueix Linn. 


1861. Giinther, Cat. Fish. B.M., vol. i, p. 409, and p. 466 (Myzus). 

1884. Jordan and Swain, Proc. U.S. Nat. Mus., vol. vii, p. 261 
(Liza). 

1916. Boulenger, F.W.F. Africa, vol. iv, p. 78. 

1920. Athanassoupoulos, Ann. Mus. Civ. Genoa (3), vol. viii, 
p. 254 ff. 

1922. Weber and de Beaufort, Fish. Indo-Aust. Archip., vol. iv, 
p. 229, and p. 264 (Myzxus). 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 302, and p. 311 
(Myzxus), and p. 1023. 

Body elongate, sub-cylindrical, more or less compressed posteriorly. 
Head usually somewhat depressed, more or less cuneiform in trans- 
verse section, with rounded apex below, generally completely scaly. 
Scales usually ctenoid, sometimes cycloid, ventral scales more 
markedly denticulate. Sometimes a secondary cycloid squamation 
in the investing integument of the scales. No lateral line, but most 
scales with one or more pits or canals, sometimes more numerous 
on dorsal scales. Mouth small, terminal or sub-inferior, protractile. 
Maxilla almost or entirely hidden beneath preorbital, excluded from 
margin of upper jaw. Upper lip narrow or fleshy, with or without 

VOL. XXX, PART 5. 40 


78 


588 Annals of the South African Museum. 


papillae. Minute recurved teeth, generally compressed and apically 
dilated, unicuspid, tricuspid, or spatulate, in one or more rows 1n 
upper jaw, present or absent. Very minute slender pointed teeth 
in lower jaw rarely present. Villiform teeth present or absent on 
vomer, palatines, pterygoids, and tongue. The buccal membrane 
sometimes with apically dilated cilia. A transverse concavity before 
the vomer, obscured in some species. Tongue adnate to floor of 
mouth. 

Eyes fairly large, usually with adipose eyelids, rudimentary or 
highly developed, better developed in adults. 

Two dorsal fins; the first, of 4 pungent spines, with a pointed 
basal scaly process inserted near middle of body, membrane from 
last ray joined to body. Second dorsal of 1 weak spine and 7-10 
rays, inserted above last fourth of body. Anal of 3 weak spines and 
8-11 rays, inserted below second dorsal. Pectorals inserted slightly 
or considerably above middle of side, more or less falcate, with or 
without an elongate scaly axillary process. Ventrals abdominal, 
joined by a membrane to the body and to one another. An elongate 
cuneiform interventral scaly process. Caudal feebly emarginate to 
forked. Fins, except first dorsal, more or less scaly. 

Third and fourth upper pharyngeals fused, enlarged, with con- 
voluted adipose base. Giull-openings wide, membranes separate, free 
from isthmus. Gill-rakers very numerous, long and close-set; lower 
pharyngeal rakers functional. Stomach tough and muscular, gizzard- 
like; pyloric caeca few. 

Mugil sensu lato is one of the earliest described genera of fishes, 
but the natural relationships of the species appear to be but poorly 
understood, and more or less unsatisfactory division of the wide 
genus has several times been proposed. 

Myzxus Guthr. (loc. cit.) was proposed for species with teeth in the 
jaws, and (sometimes?) on the palate. The validity of this differ- 
entiation has been accepted or rejected by systematists without 
precise definition of the criteria upon which they base their opinions. 
It is even not unusual to find an author stating that specimens of 
Mugil sensu stricto (i.e. accepting Myrus) have teeth in the jaws. 
Further, systematists frequently imply that teeth are absent from 
the palatal bones of certain species, whereas even a casual examination 
of their specimens would reveal that teeth are present. 

The inaccuracy of many statements about the dentition is probably 
in part due to the relatively small mouth of Mugil species, which 
renders the examination of the palatal bones, especially in small 


719 


Lhe Fishes of the Family Mugilidae in South Africa. 589 


and preserved specimens, a troublesome matter. In illustration may 
be quoted the fact that I have found no mention of lingual teeth, 
which are by no means infrequently present. 

In the South African species there appear to be all degrees between 
the entire absence of teeth, and the state where most of the normally 
dentigerous bones are fully dentate. Many of our species have teeth in 
the jaws, and on the palatal bones, at least as well developed as those 
present in, e.g., Elops Linn. In this latter genus the mouth is large. 

In so far as our species are concerned, division of Mugil sensu lato 
on the nature of the dentition alone would be not only of question- 
able value, but also exceedingly difficult to define and justify. It is 
not unlikely that this may well apply to all cases where Myzxus has 
been recognised. 

Jordan and Swain (loc. cit.) have proposed the genus Liza for 
species which do not possess adipose eyelids (genotype: capito Cuv.), 
while to Mugil sensu stricto are assigned those with eyelids. It has 
already been pointed out by Jacot (Sci. Rep. Tohok. Imp. Univ., 
1930, vol. v, No. 4, p. 827) that division on this feature is of doubtful 
value, since there is to be found almost every degree between obsolete 
and fully developed eyelids. Further, it may be indicated that 
adipose eyelids are, sometimes at least, better developed in the adult 
than in the juvenile stadia, and, also that as far as the South African 
species are concerned, the adults of all possess definite, if not always 
highly developed, eyelids. | 

It may be noted that species (such as cephalus Linn.) which possess 
well developed eyelids probably always have an edentate vomer, 
fairly concave anteriorly, while the transverse concavity anterior to 
the vomer is not usually obscured (all of the American species of 
Mugil appear to be of this type). These differences are nowhere 
constant and sharp, nor can they be considered as a basis upon which 
the genus may be divided. 

As far as the South African species are concerned, it would be 
exceedingly difficult to justify the separation of the three with well 
developed eyelids from the remainder, and I do not propose to attempt 
it. (But see note on scaling of cephalus, p. 19.) 

No representatives of the closely related genera Agonostomus Benn. 
and Cestraeus C. and V., appear to have been found in ourarea. The 
latter at least appears to be well differentiated from Mugil. 

The majority of museum collections of Mugil species appear to be 
in a somewhat chaotic state, and few systematists care to undertake 
positive identifications upon which reliance may be placed. 


80 


590 Annals of the South African Museum. 


The early descriptions are hardly ever of real diagnostic value, 
and the diagnoses of many species cannot be anything but provisional. 
Added to this, differentiation between the species of this genus has 
proved to be one of the most difficult problems which face the system- 
atist; many of the features which in other groups provide a basis 
for differentiation here appear to vary little between the different 
species. The habits and environment of the different species show 
little variation. It is found that such cases of obscured differentia- 
tion sometimes occur in genera in which there is this uniformity of 
habit and of habitat (e.g. Hpinephelus Blch.). 

In recent literature there is a general tendency to reduce the 
number of nominal species of Mugil. At the same time, it is remark- 
able that the differentiation of those closely related is still frequently 
based chiefly upon features, which can, by the examination of large 
numbers of specimens, be shown to vary widely within any one, and 
to overlap between related, species. The present confusion may 
be attributed partly to the fact that where ordinary features have 
apparently failed to show differentiation between what have quite 
obviously been different species, systematists have assigned a purely 
fictitious value to characters which are inconstant and unreliable. 
The depth of the body, and of the caudal peduncle, the position of 
insertion of the pectorals, and too restricted and undefined limits for 
scale-counts, are, among others, comparatively useless features in this 
genus. On the other hand, it is singular to find how many most 
significant characters have been entirely overlooked. The nature of 
the scales and of the teeth are of importance, but have not received the 
attention they merit. The changes which take place with increase 
in size have not been properly considered, and many nominal 
species may prove to be merely different stadia of others. It is 
remarkable that the highly characteristic form of the ventral 
fins has not, in so far as I can determine, been regarded as worthy 
of special description. This is a feature more of generic than specific 
significance. 

Where significant features have been overlooked, it is not unusual 
to find that two closely related species may be confused. This is 
probably a frequent occurrence with Mugil species, and is in part due 
to the fact that the features of doubtful validity upon which differ- 
entiation is based often vary widely, and the limits can actually 
never be clearly defined. 

A case in point is that of auratus Risso, which has been included 
in our fauna-list upon. the authority of Boulenger (loc. cit., p. 86), 


81 


The Fishes of the Family Mugilidae in South Africa. 591 


who identified * as this species a specimen in the South African 
Museum from East London. In this he was followed by Barnard 
(loc. cit., p. 308). Specimens which agree more or less with the usual 
diagnosis of auratus have been found to be the most abundant form 
from Knysna eastwards. When numbers of these were examined, 
it was obvious that two well-differentiated species were present, 
neither of which is auratus, nor can either be identified with any 
existing species. The one species (canaliculatus n. sp.) is easily 
recognised by the multicanaliculate dorsal scales, the other (éri- 
cuspidens n. sp.) by the relatively large tricuspid teeth, besides 
many other features in each case. 

The South African species have stood in need of critical revision. 
Of the fourteen (incl. Myzus) hitherto admitted to our fauna-list, 
four, viz. auratus Risso, speiglert Blkr., saliens Risso, and cunnes ius 
C. and V., have now been found not valid, while strongylocephalus Rich., 
oligolepis Blkr., tricuspidens n. sp., and canaliculatus n. sp. have been 
added, and euronotus A. Smith, up to now regarded as a synonym of 
saliens, has been revived. M. diadema G. and T. has been found to 
be identical with compressus Gnthr., and ceylonensis Gnthr. with 
buchanant Blkr. Myzxus barnardi G. and T. has been found to be a 
synonym of Mugil cephalus Linn. 

It may be as well to state that of definite purpose there have been 
omitted from most of the descriptions those details which, by the 
examination of a number of specimens, have been found to vary 
widely, and which can have but little significance. Among these 
may be mentioned the nature of the dorsal and ventral profiles, of 
the interorbital, and the degree of compression of the body. Varia- 
tion in these may result from different methods of preservation, as 
well as in part from varying degrees of sexual maturity. 

It should be emphasised that a critical revision of Mugil. species, 
based upon an adequate world collection, is long overdue. Such a 
revision would be of the greatest value to systematists, if only because 
Mugil is almost cosmopolitan in distribution, and local specimens are 
contained in probably every Museum collection throughout the world. 

It is frankly admitted that the identification of many of our species 
with those from other ‘parts is provisional only. Not only are the 
majority of descriptions in the somewhat scanty literature available 
to me of little diagnostic value, but it has not been possible to secure 


* It may be remarked that there is no record of this identification in the 8.A. 
Museum. Boulenger identified other species from specimens in this Museum, of 
which records have been kept. 


82 


592 Annals of the South African Museum. 


the desired number of identified specimens from other parts for 
examination and comparison. The final pronouncement of the 
identity of our species must be left to some worker who has at his 
disposal adequate world material. 

The chief aim of the present work has been to establish clearly the 
differentiation of the South African species. 

It will be noticed that there are very few positive statements about 
synonymy, where such would be based upon descriptions only. I 
have nevertheless made an exhaustive study of descriptions of our, 
and of related, species, but in most cases the lack of significant detail 
would render opinions based upon these of little value. | 

The localities given in the present work are those from which the 
specimens were actually obtained. The sizes given are those of the 
specimens examined. 

Certain characteristic features of Mugil species which have received 
little or no attention from systematists, but which appear to be of 
considerable taxonomic significance, are described below. 


GROWTH CHANGES. 


There is the usual variation with age in the relative size of the eye, 
and of the dimensional relationships of the parts of the head. The 
shape of the dorsal and of the anal fin undergoes considerable modi- 
fication with growth in some species, e.g. buchanani (q.v.). The 
anterior dorsal and anal rays, the caudal lobes, the pectorals, the 
pectoral axillary scale, and the adipose eyelids, all appear to increase 
somewhat in relative size with growth. The exposed surface on the 
chin increases in size in some species, e.g. buchanan, and the scaling 
on the vertical fins appears to become denser. The origin of the 
first dorsal frequently appears to move towards the snout, as if there 
is a somewhat greater increase of the posterior than of the anterior 
half of the body with growth. 


SCALES. 


The scales may be cycloid or ctenoid, and in some cases (e.g. 
euronotus) both are present. The denticulations are generally larger, 
and greater in extent, on scales from the ventral area. The ventral 
scales are always more elongate than those from the dorsal area. 

The form of equivalent scales has been found to be sometimes 
highly characteristic, and in several cases immediately diagnostic. 
The multicanaliculate scales of canaliculatus enable this species to 


83 


Lhe Fishes of the Family Mugilidae in South Africa. 593 


be distinguished at a glance from all others from South Africa. The 
dorsal scales of capito are, from the early mid-juvenile stadia, denticu- 
late, whereas those of the closely related species euronotus are cycloid. 
Further, the mucus canals of the former species are long and narrow, 
whereas those of the latter are short and wide. 

The young of those species I have examined have cycloid scales, 
the denticulations appearing as a small mid-posterior patch, which 
rapidly extends over the whole area (Pl. XIX, A-E, for capite). 

Two species, cephalus Linn. and strongylocephalus Rich., are remark- 
able in possessing two distinct squamations. The main scales are 
large. In the investing mesodermal integument is found a secondary 
squamation of minute cycloid scales, which are visible upon the sur- 
face of the primary scales (Pl. XV, A and B). This is especially well 
developed over the occipital area, while, in cephalus, the largest of 
these scales of secondary origin are to be found in the thickened dermal 
investment of the axillary scales of the pectoral and ventral fins. 

Enlarged photographs of the scales of most of the South African 
species are reproduced in Plates XVII-XX, XXII. 


ScaLy Basat Process oF First DORSAL. 


One feature of significance is the relative length of the pointed 
scaly process at the base of the first dorsal fin. Not only does the 
relative length of this appear to be remarkably constant at all stages 
in any one species, but it differs between the species, so as to afford 
in some cases a reliable guide to differentiation. Its use was ap- 
parently first proposed by Ninni * (Considerazioni sul genere Mugil, 
Venezia, 1909), but I have not seen this paper, and do not know how 
he proposed to use it. Later authors have apparently not considered 


this feature of any value. 


DENTITION. 


The teeth are always very small, and are in some species very 
- minute, so as to resemble partly or wholly ossified dermal cilia, but 
they are always, in the upper jaw at least, definitely sub-labial, with 
the bases adnate to the premaxilla. These premaxillary teeth vary 
both in size and shape, and have been found to be frequently char- 
acteristic, and in some species immediately diagnostic: the tricuspid 
teeth of tricuspidens are larger than those of most others, and enable 


this species to be easily distinguished. 


* Vide Athanassoupoulos, Ann. Mus. Civ. Gen., 1920, xlviii, p. 255. 


84 


594 Annals of the South African Museum. 


The teeth appear to be equally developed at all stages. In older 
individuals they may be partly hidden by the development of infra- 
labial spongy tissue, which has probably given rise to the idea that 
teeth are sometimes better developed in the young, since they are 
then frequently exsert and more easily visible. 

In order to determine the nature of the premaxillary teeth, when 
they are partly or completely hidden, the following procedure has 
been found satisfactory. With a sharp pair of fine scissors a thin 
strip of the upper jaw is snipped off, and soaked for some minutes in 
rectified spirit. The strip is then placed on a slide and allowed to 
dry thoroughly. This causes retraction of the enveloping tissue, 
and leaves the teeth clearly visible. 

The teeth possess elongated, dilated, bases. 

The teeth of certain species have brown apices. This applies 
especially to those such as euronotus and tricuspidens, in which they 
are apically dilated, like those of the fresh-water Cichlidae. It is 
interesting to note that these species are almost wholly fluviatile. 

Enlarged photographs of characteristic forms of premaxillary 
teeth are reproduced in Plate XVII, C-G. 

Palatal and lingual teeth are usually villiform or obtusely conical. 
In old specimens they may be obscured by a layer of mucus, but may 
be detected by means of a dissecting needle. Those on the tongue 
are usually present in adjoining patches round the anterior margin, 
and occasionally also over the slight median ridge of this organ 
(fig. 8, A). 


VENTRAL FINS. 


As has been noted previously, the form of the ventrals is highly 
characteristic, and may prove of use as a generic feature. 

The last ray of each fin is connected for a part of its length to the 
body by a membrane, which is also joined to that from the opposing 
fin (fig. 1, A and B). This forms a hollow pouch, sub-triangular in 
cross-section. Over this pouch, and of the same length, or shorter, 
projects a cuneiform inter-ventral scaly process, consisting usually 
of five or six series of scales, the apical scale being elongate and 
pointed (fig. 1, sp.). 

The precise function of this peculiar structure is not clear. It is 
possible that the increased rigidity imparted to the distended ventrals 
may play some important part in the leaping powers of these fishes. 
When the. fish moves rapidly, the pressure of the water upon 
the inclined plane of the obliquely extended ventrals would tend 


85 


Lhe Fishes of the Family Mugilidae in South Africa. 595 


to divert the anterior part of the body upwards. The larger the 


ventrals, and the more anterior their insertion, the greater will be 
this effect. 


Fie. 1.—Diagram to show the structure of the ventral fins of Mugil species— 
A, lateral view; B, ventral view. mv, membrane connecting ventrals *; mb, 
membrane joined to body; sp, interventral scaly process. 


GILLS AND GILL-RAKERS. 


The gill-rakers to the three inner arches are set at right angles to 
the vertical plane of the gill-arches. The rakers themselves are 
extremely close-set and form a plane surface, the rigidity of each 
plane being assisted by the enmeshing of setiform processes which 
are present on the adjacent basal portions of each raker. The rakers 
do not interdigitate with those from the adjacent arch, but those of 
each side of each arch form a gently curved edge, which coincides 
exactly with that from the adjacent arch (Pl. XV, C). 

The lower pharyngeal area is divided by a raised medio-longitudinal 
ridge, of which a longitudinally grooved anterior dilation is immedi- 
ately posterior to the basibranchial cartilage. Hach half of the 
lower pharyngeal area is concave, the enlarged upper pharyngeals 
fitting exactly into the two concavities. The lower pharyngeal 
bones are themselves very thin, long, curved, and fairly narrow: 
along the middle of the upper surface of each is a cartilaginous ridge, 
which bears on each side of the apex, as a continuous curved plane, 
lamellae exactly similar in structure to, but longer than, the rakers 
on the functional gill-arches. The edge of the exterior series meets 
that of the inner series of rakers from the inner functional gill-arch. 
The inner edge of the inner series of these lamellae is adnate to the 
cutaneous margin of the medio-longitudinal pharyngeal ridge, while 


* Slightly exaggerated. 


86 


596 Annals of the South African Museum. 


the lower margins of both series of lamellae are adnate to the upper 
margin of the pharyngeals, the outer series projecting some distance 
into the branchial cavity (Pl. XV, C). 

It may be presumed that these pharyngeal lamellae have been 
developed from the true rakers originally present on the arch now 
modified to form the pharyngeals. If so, it is an interesting example 
of a surviving integral portion of a highly modified structure, having 
retained the original form and function despite the profound struc- 
tural and functional modification of the main structure. 


DIMENSIONAL RELATIONSHIPS, ETC. 


In order that dimensional relationships and scale-counts shall 
have their full value, it is essential that the precise limits should be 
defined. In the present paper the following have been employed :— 

Length of Head.—This is measured with dividers in a straight line 
from the tip of the retracted snout to the hindmost point of the 
opercular margin on the level of the upper margin of the pectoral 
base, 2.e. 1t is measured diagonally, and not in profile. 

Head without snout is measured from the hind margin of the head 
to the anterior margin of the orbit. 

Length of Pectoral—This is employed as an important diagnostic 
character, and is measured from the body to the tip of the pectoral, 
when the latter is held at right angles to the side. 

Caudal Base.—This is taken as the base of the mid-caudal rays, 
which is obscured by the scaling. The body scales diminish very 
little in size up to this point, the scales on the basal portion of the rays 
being generally abruptly smaller. In fresh specimens the bases of 
the rays are easily visible if the caudal be distended and viewed 
against a light. In preserved specimens this point is less easily 
ascertainable, but may generally be determined by similar means. 
The diagnostic scale-counts employed in the present work may quite 
easily be made with sufficient accuracy. 

Scales : Lateral Series.—-This is taken as the number of scales in 
the first continuous series above the axil of the pectoral, from directly 
above the hind margin of the head (see Length of Head, above) to the 
caudal base (q.v.). The first series above the axil is not usually 
continuous, being interrupted at the 3rd or 4th scale, whereas 
the next series above is usually regular, often starting from above 
the hind margin of the head in a gentle downward curve. 

Scales : Transverse Series.—This count can have but little signifi- 


87 


Lhe Fishes of the Family Mugilidae in South Africa. 597 


cance, since the number of transverse series varies very little between 
the species. The counts given in this work are taken from before 
the origin of the first dorsal to the mid-line of the belly. 

Scales: Predorsal.—All previous counts of the predorsal scales 
have been taken from the origin of the first dorsal to the snout. 
Since the dorsal cephalic squamation is rarely ever regular, counts 
between these limits are of little value in the absence of precise 
definition of the method of counting. 

The number of predorsal scales does not appear to be of any special 
significance, as apart from the number in lateral series, but the 
number of series between the 
origin of the first dorsal and 
the point above the hind margin 
of the head has in each case 
been recorded. 

Angle of Lower Jaw.—This 
does not alone appear to be of 
any special significance, but 
as most workers make some 
statement about the nature 
of. this feature, it appears 
advisable to follow suit. The 
majority of statements appear Fic. 2.—Diagram of chin of a specimen of 
to have been based upon casual Mugil. To show how the angle of the 

mouth is taken. 

estimations, which can . have 

but little value, and actual diagrams in illustration of the angle of 
the mouth do not infrequently fail to agree with corresponding 
statements. I have therefore judged it wise to record the actual 
angle subtended by the corners of the mouth at the symphysis, 
which is not affected by the nature of the outline of the jaw, 
this being sometimes rounded, and sometimes angular. This angle 
has been measured by means of a simple goniometer devised and 
constructed for this purpose. 

Origin of first Dorsal (1) to Snout, (2) to Caudal Base, (3) te Hind 
Margin of Caudal Rays.—These are measured with dividers, one point 
at the anterior point of the base of the first dorsal spine, and the 
other (1) at the tip of the snout, (2) at the mid-point (lateral) of the 
caudal base, and (3) at the actual hind margin of the mid-caudal 
rays, respectively. 

Length of Pointed Basal Scaly Process of Furst Dorsal.—This is 
measured from the anterior point of the base of the first dorsal spine _ 


88 


598 Annals of the South African Museum. 


to the hindmost point (apex) of the scaly process. When the two 
processes are unequal in length, the longer is measured. 

Length of snout is measured obliquely from the tip of the snout to 
the anterior margin of the orbit. 

Total length is measured from the tip of the snout to the hind 
margin of the mid-caudal rays. 

Other measurements are taken in the usual manner. 

The interventral scaly process varies very little in relative length 
between the species, averaging three in head. 


Key to the South African species. 


I. Adipose eyelids well developed, the posterior covering more 
than half of the posterior width of the iris, in some cases 
reaching to hind margin of pupil. 

A. Scales 37-42. Pectorals not longer than head without 


snout.* 
1, Anal rays 7-8. 2nd dorsal not scaly. Anterior 
eyelid well developed . cephalus. 
2. Anal rays 9. 2nd dorsal scaly. Anterior ‘eyelid 
feeble. . robustus. 
B. Scales 33-35. Pectorals longer than head without 
snout * . . . strongylocephalus. 


IJ. Adipose eyelids narrow or rudimentary, better visible in 
adults, round the outer margin of the eye, covering not 
more than half of the posterior portion of the iris. 
A. Prominent papillae in several series on lower margin of 
upper lip, which is very deep at snout tip. 
(Scales 37-40) . . . . . . crentlabis. 
B. No papillae on upper lip, which is not, or “scarcely, 
more than 4 of eye deep at snout tip. 
1. Scales 41-49. 
a. Pectorals not longer than head without 
snout.* Palatine teeth present. 
x. Scale at base of first dorsal 6-5-8 in 
distance from origin of first dorsal 
to snout tip. Soft dorsal com- 
pletely scaly . euronotus. 
y. Scale at base of first dorsal 4-5 in 
distance from origin of first dorsal 
to snout tip. Soft dorsal not 
scaly posteriorly . . capito. 
b. Pectorals longer (in adults much longer) than 
head without snout.* No palatine teeth. 
x. Teeth comparatively large, tricuspid. 
Pectorals 1-2-1-3in head. Maxilla 
wellexposed . . . . tricuspidens., 


89 


The Fishes of the Family Mugilidae in South Africa. 599 


y. Teeth not tricuspid, small. Pectorals 
1-0-1:1 in head. Maxilla com- 
pletely concealed . . . seheli. 

2. Scales 29-40. 

a. End of maxilla concealed. <A scaly process 
in axil of pectoral, longer than } length 
of fin. 

x. Pectorals longer (in adults much longer) 
than head without snout.* (1-0— 
1-1 in head.) 

a. Seales 38-40 (42). Longest dor- 


sal ray shorter than ventrals seheli. 
B. Scales 33-35. Longest dorsal 
ray longer than ventrals buchanani. 


y. Pectorals not longer than head without 
snout.* (1-4-1-5 in head, scales 
37-39) . . . . . robustus, 
b. End of maxilla exposed. No process, or a 
very short and blunt one, in axil of 
pectoral. 
x. Ventrals longer than head without 
snout.* Scale at base of first 
dorsal shorter than % of postorbital — 
part of head. (Scales 29-32) compressus. 
y. Ventrals shorter than head without 
snout.* Scale at base of first 
dorsal longer than % of postorbital 
part of head. 
a. Scales 33-35. Pectorals very 
little longer than head with- 
out snout.* Predorsal scales 
not multicanaliculate . . macrolepis. 
B. Scales 36-39. Pectorals longer 
than head without snout,* 
usually as long as entire head. 
Predorsal scales multicanali- 
culate . . . . canaliculatus, 
3. Scales 26-28. 
a. Analrays 8. Caudal almost truncate. Pec- 


torals partly or wholly black . waigiensis. 
6b. Anal rays 9. Caudal emarginate. Pec- 
torals light . . . . . oligolepis. 


Note on Key.—It will be observed that two species (seheli and robustus) each 
occur twice in the Key. It appears to be difficult to avoid the use of scale-count 
as a primary diagnostic character, and seheli forms an unfortunate bridge between 


* i.e. distance from hind margin of head to anterior margin of orbit. The 
hind margin of the head is always taken on the level of the upper margin of the 
base of the pectoral. 


90 


600 Annals of the South African Museum. 


the two main groups. The method here employed obviates this difficulty. In 
the case of robustus the eyelids shrink on preservation, and might not be accounted 
large enough for the species to fall in Group I (see note under robustus). 

Another Key, embracing five Indo-Pacific species likely to be found here, will 


be found at the end of this paper. 


Mugil cephalus Linn. 
(Plate XV.) 


1861. Giinther, Cat. Fish. B.M., vol. ii, p. 418 (constantiae), and 
p. 419 (cephalotus). 

1888. Day, Fish. India, p. 353, pl. xxv, fig. 3 (oeur). 

1916. Boulenger, F.W.F. Africa, vol. iv, p. 82, fig. 48 (oeur). 

1918. Athanassoupoulos, Ann. Mus. Civ. Gen. (3), vol. vin, p. 264. 

1922. Weber and de Beaufort, Fish. Indo-Aust. Archip., vol. iv, 


p. 253. 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 302, and p. 311 (Myzus 
barnardt). | 

1930. Jacot, Sci. Rep. Imp. Univ. Tohok. (4), vol. iv, No. 4, 
p. 825 fi. | 


Snout very broad, bluntly rounded or obtusely angular. Depth 
3-5-4, length of head 3-3-4 in length of body. Eye 3-8 (Juv.)-5, snout 
3°3-4:8, interorbital width 2-2-4, postorbital length 1-8-2 in length 
of head. Adipose eyelids well developed, completely encircling pupil, 
exposed surface of iris small or none (adult), aperture in membrane 
round or vertically elliptical. Nostrils 2-5 in eye diameter apart, 
posterior as far from front margin of eye as anterior from profile of 
snout tip. Lower margin of preorbital not bent or notched, obliquely 
truncated, maxilla almost or quite concealed, exposed portion in- 
creases with age. Angle of lower jaw 65-88°, outline of lower jaw 
sub-angular, rounded, or undulate. Symphysial knob double. 
Upper lip thin, width at apex of snout 6 in eye. Curved compressed 
teeth in one, few, or many series in each jaw. Villiform teeth on 
pterygoids. Palatines, vomer, and tongue edentate. Prevomerine 
groove distinct. Exposed area on chin long and wide. 

D IV+I, (6-)8. First dorsal inserted 1:0-1-1 times further from 
caudal base than from tip of snout, 1:3-1-4 times as far from the 
hind margin of the mid-caudal rays as from the snout tip. First 
- spine 1-6-1-8, base of first dorsal 2-0-2-5 in head. Distance from 
origin of first to origin of second dorsal 1-0-1-3 (J.) in head. First 
dorsal inserted above the 13th—-14th, second above the 24th-26th 
lateral scale. Pointed sheath scale extends behind origin of first 


9] 


The Fishes of the Family Mugilidae in South Africa. 601 


dorsal 2-8-3-8 in head, 2-6-3-0 in distance from origin of first to 
origin of second dorsal, 5-3-6-4 in distance from origin of first dorsal 
to snout tip. Longest soft ray 1-6-1-8, base of second dorsal 2-0-2-6 
in head. Last ray much longer than penultimate, edge of fin concave. 
Second dorsal scaly basally only. 


Fie. 3.—Mugil cephalus Linn. 


Note.—In this and other text-figures of species, the lateral row of dots indicates 
the number and disposition of the lateral rows of scales. 

The dimensional relationships involving the length of the head may appear to differ 
between the text and the figures, but the head in the former is not measured in profile 


(see p. 596). 

A III, 8. Inserted in advance of second dorsal, below the 22nd- 
25th lateral scale. Longest ray 1-5-1-8 in head, shape of fin similar 
to that of soft dorsal. Scaly for anterior 2. 

P 17, 1-3-1-5 (J.) in head, tip reaches to the 10th—12th lateral scale. 
Inserted 2-3-3-5 times as far from the ventral as from the dorsal 
profile. Axillary scale long and pointed, 3:3-4-8 (J.) in head. 

Ventrals 1-5-1-6 in head, inserted below in advance of, behind, or — 
at, midway between origin of first dorsal and hind margin of head. 
Edge of fin truncate. Axillary scale 3-2-4 in head. 

Caudal deeply forked, upper lobe longer, increasing with age, 
mid-rays 1-7—1-9 in head. 

Scales ctenoid, predorsal scales 1-:0-1:2 times wider than long 
(Pl. XV, A). Mucus canal long and narrow, often oblique. Most 
scales, especially those on occipital region, and axillary scales, with 
a superimposed secondary squamation of minute cycloid scales, 
developed in the integument, more noticeable in large specimens * 

* I have examined specimens from Japan and America, and this secondary 
scaling is well developed, indicating that it is characteristic of the species and 


not confined to South African specimens. 


92 


602 Annals of the South African Museum. 


(Pl. XV, Aand B). Lat. ser. 39-42, l.tr. 14-15, 2 cheek scales, 13-14 
predorsal to above hind margin of head. 

Colour.—Silvery, darker above. Sometimes longitudinal stripes. 
Fishes from brackish water generally darker in colour. 

Localities—Lakeside (Cape Peninsula), Knysna River, Kabeljaauws 
River, Port Elizabeth (Zwartkops River), Kowie River, Great Fish 
River, Buffalo River, Mazeppa Bay, Durban, Sinkwazi, Kosi Bay. 
Also Japan, Peru, Chesapeake Bay (N. America). 

Length.—Up to 630 mm. 

Thirty-four specimens, from 55 mm. up, examined. 

The synonymy of this species is rather complex. oeur Forsk. is 
regarded by most authors as conspecific with cephalus, but Boulenger 
(loc. cit.) regards the former as distinct in that the angle of the lower 
jaw is acute, whereas that of cephalus is stated to be obtuse. I have 
seen a juvenile specimen in the S.A. Museum which Boulenger 
identified as cephalus. This specimen has the mouth open, and it 
appears as if the mandibles are set at an obtuse angle, whereas when 
the mouth is closed the angle is 84°. I have seen no specimens in 
which the angle of the mouth is obtuse. It is curious that Giinther 
(loc. cit.) gave no diagram of the mouth of cephalus, or of related 
species stated to have a mouth of obtuse angle, whereas he gave 
numerous diagrams of mouths of acute angle. 

I have examined several specimens of cephalus from America 
(kindly donated by Dr. George 8. Myers of the U.S. National Museum), 
and these are in all respects identical with ours. The angle of the 
mouth falls within the limits of variation in our specimens. 

M. cephalus (as here defined) is quite obviously a somewhat poly- 
morphous species. In so far as | have observed, it is almost always 
fluviatile, and does not commonly occur in the sea. It is possible 
that purely local forms may show minor variations from the general. 
The teeth in the jaws, the extent of the exposure of the maxilla, and 
the shape of the mouth are all extremely variable. In some specimens 
_ there is a single series of teeth, while others from the same locality 
have several or many rows in both jaws. The angle of the mouth 
varies from 65-88°, while the lower jaw may be sub-angular, rounded, 
or anteriorly undulate. Further, the angle of the mouth does not, 
in my specimens, become more obtuse with age. Some of the largest 
have the angle 65-70°. 

I have endeavoured to find some constant basis among these 
variations for the establishment of sub-species, but there appears to 
be no combination of these, or of these with other characters, which 


93 


Lhe Fishes of the Family Mugilidae in South Africa. 603 


would justify this step. With more intensive study it may be possible 
to discover combinations of features which will establish definite sub- 
species. It is not unlikely that a very detailed study of specimens 
from all parts of the world would probably yield interesting results. 

The extraordinary secondary cycloid squamation briefly described 
above alone merits special attention, and may ultimately prove of 
importance in the division of the genus, especially as it is present in 
strongylocephalus, which also possesses well-developed eyelids. These 
smaller scales are quite obviously ossified, and mesodermal in origin. 

Myzxus barnards G. and T. is undoubtedly merely a juvenile cephalus. 
Barnard (loc. cit.) suspected this, but, possibly because the specimen 
is damaged, missed the scaly process which is present in the axil of 
the one undamaged pectoral. This process is always much smaller 
in the very young stadia. 

M. cephalus attains a large size. In brackish vleis, and in the quiet 
upper reaches of lagoons and estuaries, specimens up to ten pounds 
in weight are not infrequently encountered. This species possesses — 
very considerable leaping powers, which are not, however, as great 
as those of tricuspidens, while the type of leap is also different: 
cephalus leaps with the head well up, and the body curved, whereas 
the former species jumps much further, with the body more or less 
straight and parallel with the surface of the water. 


Mugil robustus Guthr. 
(Plates XXI, A, and XXII, A, B.) 


1861. Giinther, Cat. Fish. B.M., vol. iti, p. 482. 

1916. Boulenger, F.W.F. Africa, vol. iv, p. 92, fig. 54. 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 305. 

Body markedly robust anteriorly, tapering posteriorly. Snout 
fairly broad, bluntly rounded, upper lip forms oblique truncated 
margin. Depth 4-0, length of head 3-9 in length of body. Eye 
4-1-4:3, snout 3-7, interorbital width 2-5, postorbital part of head 
1-8 in length of head. Adipose eyelids very fragile, but well developed, 
especially the posterior, which extends almost to the hind margin of 
the pupil; the anterior covers about half of the width of the iris. 
(With preservation the eyelids appear to shrink considerably.) 
Nostrils 2-8 in eye diameter apart, the anterior as far from snout tip 
profile as the posterior from the front margin of the eye. Lower 
margin of preorbital not notched, sharply bent downwards over the 
angle of the mouth, end scarcely serrate, slightly convex. Mazxilla 

VOL. XXX, PART 5D. | 4] 


94 


604 Annals of the South African Museum. 


completely concealed. Angle of lower jaw 110-112°; outline of jaw 
angular. Symphysial knob single. Upper lip fairly thin, width at 
apex of snout4ineye. Noteethin any part of mouth. Prevomerine 
groove well marked. Exposed area on chin long and very narrow. 
D IV+I, 8. First dorsal inserted 1-05-1-08 times further from 
caudal base than snout tip, 1:4 times as far from the hind margin of 
the mid-caudal rays as from snout tip. First spine 1-7, base of first 
dorsal 1-9-2 in head. Distance from origin of first to origin of second 
dorsal 1-:0-1-1 in head. First dorsal inserted above the 12th, second 
above the 23rd—24th lateral scale. Pointed sheath scale extends 
behind origin of first dorsal 2-3-2-4 in head, 2-2 in distance from 
origin of first to origin of second dorsal, 1-3 in postorbital part of 


Fic. 4.—Mugil robustus Gnthr. (see note, fig. 3). 


head, and 4-3-4-4 in distance between origin of first dorsal and snout 
‘tip. Longest soft ray 2-0, base of second dorsal 2-3-2-5 in head. 
Last ray slightly longer than penultimate. Fin scarcely falcate, 
edge concave; completely scaly. 

A III, 9. Inserted in advance of second dorsal, below the 22nd—23rd 
lateral scale. Longest ray 2-1 in head; shape of fin similar to that of 
soft dorsal; completely scaly. 

P 15, 1-4—1-5 in head, shorter than head without snout, tip reaches 
to the 10th lateral scale. Inserted 2:4 times as far from the ventral 
as from the dorsal profile. Axillary scale large and pointed, 3-5 in 
head, 2-5 in pectoral. 

Ventrals 1:6 in head, inserted below midway between origin of 
first dorsal and hind margin of head. Edge of fin subtruncate. 
Axillary scale 3-0 in head. 

Caudal moderately forked, upper lobe longer, mid-rays 1-8 in head. 

Scales cycloid or very feebly denticulate, or with scalloped edge, 
predorsal scales slightly longer than wide (Pl. XXI, A and B). No 


95 


Lhe Fishes of the Family Mugilidae in South Africa. 605 


secondary squamation. Lat. ser. 37-39, l.tr. 12-13, three cheek scales, 
12 predorsal to above hind margin of head. 

Colour.—Bright silvery, slightly dusky above. Sometimes a golden 
opercular spot. Traces of faint longitudinal stripes. Axillary spot 
very distinct. | 

Localitves.—Isipingo lagoon, Durban Bay, Kosi Bay. 

Length.—Up to 230 mm. | 

Four specimens, from 190 mm. up, examined. 

This appears to be a well-defined, but comparatively scarce and 
localised species. 

Previous descriptions do not agree very well. Giinther (loc. cit.) 
specifically mentions the well-developed adipose eyelids, whereas 
Boulenger (loc. cit.) neither describes them nor shows them in his 
figure. Barnard (loc. cit.) possibly never saw a specimen, and may 
have been misled by Boulenger’s work. | 

As has been indicated above, the adipose eyelids of robustus are 
abnormally thin, and tend to shrink with preservation, especially 
if the specimen is permitted to become even superficially dry. Even 
so, the eyelids are then so well marked as to merit special mention. 

It would be strange to find any species as strictly localised as this 
would appear from its recorded area. Day (Fish. India, 1888, 
p. 356) described as caeruleomaculatus Lacep., a species, which, except 
for the absence of adipose eyelids, agrees exactly with the diagnosis 
of robustus. I have not seen the original description of caeruleo- 
maculatus, but Giinther’s diagnosis (loc. cit., p. 445) of that species, 
while rather brief, agrees with that of Weber and de Beaufort (Fish. 
Indo-Aust. Archip., 1922, vol. iv, p. 250), and fits quite well a specimen 
of this species (from India) which I have examined and which is 
quite distinct from robustus. It is not unlikely that Day may have 
examined preserved specimens with eyelids so shrunken as to have 
misled him. It is extremely likely that Day’s specimens were 
actually conspecific with robustus, in which case this species extends 
from Africa through Mauritius to the Indo-Malayan area, which 
appears reasonable. 

According to the Indian netters on the Natal coast, robustus is never 
very plentiful, but relatively large numbers appear on the coasts in May. 
Unfortunately, little reliance can be placed upon their identifications. 

M. robustus is closely related to cephalus as well as to seheli. From 
the former it is distinguished by the extra anal ray, by the nature of 
the eyelids, by the shape of the mouth, and by the scaly median fins. 
From the latter by the much shorter pectorals, and by the presence 


96 


606 Annals of the South African Museum. 


of eyelids, as well as in dimensional relationships. It could scarcely 
be confused with any other South African species. 


Mugil strongylocephalus Rich. 


(Plates XVI, A, and XVIII, A, B.) 


1861. Giinther, Cat. Fish. B.M., vol. iii, p. 425, and p. 428 
(longimanus). 

1888. Day, Fish. India, p. 349, pl. lxxiv, fig. 3 (cunnesius). 

1922. Weber and- de Beaufort, Fish. Indo-Aust. Archip., vol. iv, 
p. 239 (longumanus). 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 302 (cunnestus). 

1925. Fowler, Proc. Ac. Nat. Sci. Phil., vol. Ixxvii, p. 208 
(longimanus). 

Depth 3-8, length of head 3-6 in length of body. Kye 3-7, snout 
3:6, interorbital width 2-3-2-5, postorbital 1-8-2-0 in length of head. 
Adipose eyelids well developed, almost encircling pupil, posterior 
more prominent, almost reaching pupil, aperture in membrane 
elliptical. Nostrils + of eye diameter apart, anterior midway between 
profile of snout tip and anterior margin of eye. Lower margin of 
preorbital slightly bent, scarcely notched. End of preorbital narrow, 
serrated, maxilla almost concealed. Angle of lower jaw 92-96°, 
outline of jaw angular, or very slightly rounded (adults). Symphysial 
knob double. Upper lip thin, width at apex of snout } of eye. 
Very minute pointed teeth in a single row in upper jaw. Villiform 
teeth on pterygoids and possibly also on tongue. Traces of minute 
teeth on vomer. Palatines and lower jaw edentate; _ exposed area 
between rami of mandibles short and narrow. | 

D IV+I, 8. First dorsal inserted 1-0-1-06 times as far from the 
tip of the snout as from caudal base, 1-2-1-3 times as far from the 
tip of the mid-caudal rays as from tip of snout. First spine 1-8-2-1, 
base of first dorsal 1-9-2-2 in head. Distance from origin of first to 
origin of second dorsal 1-15-1-25 in head. First dorsal inserted above 
the 10th—-12th, second above the 19th-20th lateral scale. Pointed 
sheath scale extends behind first dorsal 2-2-2-5 in head, 1-8-2-2 in 
distance from origin of first to origin of second dorsal, 4-0-4-6 in 
distance from origin of first dorsal to tip of snout. Second soft Tay 
1-6-1-8, base of second dorsal 2:7-2-9 in head. Last ray longer than 
penultimate, fin not much elevated anteriorly, edge concave. Soft 
dorsal at least partly scaly. 


A III, 9, inserted in advance of second dorsal, below the 17th-18th 


97 


The Fishes of the Family Mugilidae in South Africa. 607 


lateral scale. Second ray 1-6-1-8 in head; shape of fin similar to that 
of dorsal; scaly. - 

P 15-16, 1-06—-1-25 in head, tip reaches to 11th—12th lateral scale, 
inserted 2-6-3 times as far from the ventral as from the dorsal profile. 
Axillary scale, bluntly rounded or pointed, 3-3-5 in head. 

Ventrals 1-6-1-7 in head, inserted below 1-1-1-2 times as far from 
the origin of the first dorsal as from the hind margin of the head. 
Kdge of fin gently rounded. Axillary scale 2-7-3-0in head. Ventrals 
and pectorals scaly on basal half. 

Caudal moderately forked, mid-rays 1-8 in head; scaly. 

Scales cycloid; mucus canal long and narrow. Predorsal scales 


las 


a 

ony 

’ 

.¢ , 
YQ I) S/S, / 
rd 


——.. 
——] 


tem 


————— 


Fic. 5.—Mugil strongylocephalus Rich. (see note, fig. 3). 


1-1 times as wide as long (Pl. XVIII, A and B). Lat. ser. 33-35, Ltr. 
11-12, 3-4 cheek scales, 11-12 predorsal to above hind margin of head. 
A few small secondary elongated cycloid scales upon the scales of 
the nuchal region. (Also found in a specimen from India.) 

Colour.—Silvery, slightly darker above. Caudal with dark margin. 
Pectoral axil black. 

Localities.—Isipingo lagoon, Durban, Beira, Bay of Bengal. 

Length.—Up to 195 mm. 

Seven specimens, from 140 mm. in length up, examined. 

I have not seen the original description, and the diagnosis of these 
specimens as strongylocephalus, being based on Giinther’s description 
(loc. cit.) of the type (and of others?), is provisional only. 

There appears to be a somewhat hopeless confusion in regard to 
specimens described as engeli Blkr., kelaartii Guthr., longimanus Guthr., 
and strongylocephalus, the types of all of which come from the Indo- 
Pacific. I have seen only Giinther’s description of the latter species. 
The majority of authors agree in placing kelaartw in the synonymy 


98 


608 Annals of the South African Museum. 


of engeli, but this appears to be doubtful. The former species was 
described from very young specimens, in which the pectorals are 
generally shorter than in the adult, and yet the pectorals of these 
juvenile types are stated to be actually longer than those of adult 
engelt. A critical revision of these two species would almost certainly 
establish that kelaartii is distinct from engel. 

M. kelaartii is held to be distinct from longimanus ( fide Giinther) 
mainly because the pectorals of the former are somewhat shorter. 
Here again the difference in size between the type-specimens would 
easily account for the slightly shorter pectorals of the former species, 
and the two are most probably conspecific. Further, the upper 
lip of longimanus is stated by Giinther (loc. cit.) and by Weber and 
de Beaufort (loc. cit.) to be rather thick, whereas Day’s figure (loc. cit.) 
shows a thin lip. The specimens described above have a thin lip, 
and, with the exception of the point of insertion of the first dorsal, — 
agree exactly with Day’s figure. 

The. head and chin of strongylocephalus as figured by Giinther agree 
exactly with those of my specimens, and I can find nothing of import- 
ance in which they differ from Giinther’s description of that species. 
I have examined a specimen from the Bay of Bengal, kindly lent by 
the Indian Museum, Calcutta, labelled cunnesius C. and V., which 
agrees in all particulars with my specimens, and with Giinther’s 
description and figures of the head of strongylocephalus. 

M. kelaarti and longimanus are held to differ from strongylocephalus 
in that the maxilla of the former two is entirely concealed, whereas 
the tip of that of the latter remains visible. Giinther does not state 
the size of the type of the latter species, but it 1s, from what he says, 
presumably an adult. I have found that the extremity of the 
maxilla is generally more exposed in large specimens. In my speci- 
mens, especially in the smallest, when the mouth is pressed shut, 
it appears as if the maxilla is entirely hidden, but a careful examina- 
tion reveals that the extremity always remains visible. This slight 
difference can alone scarcely justify the maintenance of kelaartit and 
longimanus as distinct from strongylocephalus, and in my opinion they 
are most likely conspecific. Fowler’s Delagoa Bay specimen appears 
to be unquestionably conspecific. 

This species, which appears to be widely distributed in the Indo- 
Pacific, will probably be found to be fairly common in Natal waters 
with more intensive collection. 

It is easily distinguished by the well-developed eyelids, the long 
pectorals, and the scale-counts from all other South African species. 


99 


The Fishes of the Family Mugilidae in South Africa. 609 


Mugil crenilabis Forsk. 


¢ 1861. Giinther, Cat. Fish. B.M., vol. iii, p. 458. 

1888. Day, Fish. India, p. 355. 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 307. 

Depth 3-8-4-3, length of head 3-2-3-4 in length of body. Eye 
3-8-4, snout 3:5, interorbital width 2-3-2-5, and postorbital length 
1-9 in length of head. Adipose eyelids rudimentary. Nostrils 1 eye 
diameter apart, anterior midway between front margin of eye and 
profile of snout tip. Lower margin of preorbital bent, deeply emar- 
ginate, end dilated, serrae large. Maxilla completely concealed. 


Fic. 6.—Mugil crenilabis Forsk. (see note, fig. 3). 


Angle of mouth 94—96°, outline of jaw sub-angular. Upper lip thick, 
half, or slightly less than half, eye diameter at snout tip; lower margin 
with 5-6 series of fleshy tubercles, the lower with apical branches. 
Lower lip with expanded rugose plicate fringe. Exposed area on 
chin small and narrow. No teeth visible in jaws, or on palatal 
bones. Minute teeth on tongue. 

D IV+I, 8. First dorsal inserted 1-0-1-06 times as far from snout 
tip as from caudal base, 1-3 times as far from the hind margin of the 
mid-caudal rays as from snout tip. First spine 2:1-2-3, base of first 
dorsal 4 in head. Distance from origin of first to origin of second 
dorsal 1-3-1-4 in head. First dorsal inserted above the 12th, second 
above the 24th lateral scale. Pointed sheath scale very short, 
extends behind origin of first dorsal 4-6 in head, 3-3 in distance from 
origin of first to origin of second dorsal, 8 in distance from origin 
of first dorsal to snout tip. Longest soft ray 1-8, base of second dorsal 
3-5 in head. Second dorsal sub-faleate, edge concave; scaly. 


100 


610. Annals of the South African Museum. 


A III, 9, inserted in advance of second dorsal, below the 23rd 
lateral scale. Longest ray 1-8 in head. In shape similar to dorsal. 

P 16-17, 1-3-1-4 in head, tip reaches 12th lateral scale, inserted 
9-5-3 times as far from the ventral as from the dorsal profile. No 
axillary scale, or a very indistinct short process. 

Ventrals 1-7 in head, inserted below midway between hind margin 
of head and origin of first dorsal. Edge of fin truncate. Axillary 
scale 3-5-4 in head. 

Caudal emarginate, lower lobe slightly longer, mid-rays 1-7 in head. 

Scales cycloid, but with rudimentary scalloping on posterior 
margin, indicating that adults will probably have ctenoid scales. 
Predorsal scales about as wide as long. Lat. ser. 37-40, l.tr. 13-14, 
3-4 cheek scales, 12-13 predorsally to above hind margin of head. 

Colour.—(Preserved.) Uniform light brown. 

Locality.—Durban. 

Length.—Up to 56 mm. 

Two specimens, 54 and 56 mm. in length, examined. 

This is the only species from South Africa with tubercular lip, and 
it is easily distinguished from our others by this feature alone. 

It is evidently fairly rare: I have seen none but the two 8.A. 
Museum specimens described above. It has been stated to attain 
a length of over 200 mm., and to be fairly widely distributed in 
the Indo-Malayan area. I am not quite certain that our specimens 
are actually crenilabis, but they are very small and not too well 
preserved, so that I am unable to venture any definite opinion on the 
matter. The above description is taken as a composite from both 
specimens. _ 

Probably ruppellia Gnthr. (loc. cit., p. 458) is not different. 


Mugil euronotus Smith. 


(Plates XVI, E; XVII, E; XIX, G, H.) 
1849. Smith, Illvs. 8.A. Pisces, pl. xxix, fig. 2. 
1861. Giinther, Cat. Fish. B.M., vol. iii, p. 443 (saliens part). 
1861. Boulenger, F.W.F. Africa, vol. iv, p. 85 (saliens part). 
1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 307 (saliens part). 
Depth 3-9-4-6, length of head 3-2 (J.)-4-5 (Ad.) in length of body. 
Kye 4:0 (J.)-4:8 (Ad.), snout 3-2 (Ad.)-3-7 (J.), interorbital width 
2-2-2-6, and postorbital length 1-8-2-0 in length of head. Adipose . 
eyelids rudimentary, better visible in adults. Nostrils 1 of eye 


ro) 


. . 6 
diameter apart, anterior nearer profile of snout tip than anterior 


101 


Lhe Fishes of the Family Mugilidae in South Africa. 611 


margin of eye. Lower margin of preorbital gently curved down- 
wards, not notched, serrate. Maxilla not, or only extreme tip, 
exposed. Preorbital scaly. Angle of lower jaw 95-98°. Outline 
of jaw rounded or angular. Symphysial knob single. Upper lip 
fairly thick, width at apex of snout 3-3-5 in eye. Relatively large, 
close-set, flattened, recurved, dilated spatulate teeth with a notch at 
each side of apex (strangulated) (Pl. XVI, E) in a single series in upper 
Jaw, similar in all stadia. Small cilia sometimes in lower jaw. Villi- 
form teeth on vomer, palatines, pterygoids, and tongue. Exposed 
surface on chin long and narrow in juveniles, long and wide in adults. 

D IV +I, 8. First dorsal inserted nearer caudal base than tip of 
snout, 1:04-1:07 times as far from the latter as from the former, 


Fic. 7.—Mugil euronotus A. Smith (see note, fig. 3). 


1:18-1:22 times as far from the hind margin of the mid-caudal rays 
as from the tip of the snout. First spine 1-7-1-9, base of first dorsal 
2-4-2-7 in length of head. Distance from origin of first to origin of 
second dorsal 1-1-1-3 (Ad.) in head. First dorsal inserted above the 
15th-17th, second above the 28th-30th lateral scale. Pointed 
sheath scale extends behind origin of first dorsal 3-5-4-0 in head, 
65-8 in distance from tip of snout to origin of first dorsal, 2-9-3-5 
in distance from origin of first to origin of second dorsal. Second 
soft ray 1-9-2-5, base of second dorsal 2-0-2:2 in head. Last ray 
slightly longer than penultimate, fin little elevated anteriorly, edge 
gently concave. Second dorsal completely scaly. 

A III, 9, inserted in advance of second dorsal, below the 26th-27th 
lateral scale. Second ray 1-8-2-1 in head, last ray longer than pen- 


ultimate, edge of fin slightly concave; scaly. 
P 17, 1-4-1-5 in head, tip reaches to the 10th—12th lateral scale, 


102 


612 Annals of the South African Museum. 


inserted 1-9-2-6 times as far from the ventral as from the dorsal 
profile. No axillary scale. 

Ventrals 1-5-2-2 (J.) in head, inserted below midway between base 
of pectoral and origin of first dorsal, or nearer the latter. Edge of 
fin almost truncate. Axillary scale 4-3-5 in head. 

Caudal moderately forked, upper lobe longer in adults; mid-rays 
1-8-2-0 in head. _ | 

Scales predorsally cycloid, becoming ctenoid on sides and belly. 
Predorsal scales slightly wider than long (Pl. XTX, G and H), |r. 43-45, 
ltr. 14-15. Four cheek scales, 16-17 predorsal to above hind margin 
of head. 

Colour variable. In sea, light dusky above, silvery below. In 
fresh or brackish water, almost black above, shading through dusky 
to light below. 

Locality—Knysna River, freshwaters of the Hastern Province, 
Port Alfred (river), Fish River, Buffalo River, Durban (harbour ?), 
Sinkwazi River. 

Length.—Up to 300 mm. 

Seventeen specimens, from 55 mm. up, examined. 

Plesiotypes, from Knysna, in the Albany Museum. 

This has proved a very troublesome species. I was at first inclined 
to consider our specimens conspecific with saliens Risso, which is 
apparently so closely related to capito Cuv. that the majority of 
workers have found the greatest difficulty in differentiating at all 
clearly between them (see p. 616). Yet a careful examination of my 
specimens revealed so many striking differences from the latter species 
that I felt 1t was impossible for previous workers to have missed them. 

I have sent a specimen to Mr. Norman of the British Museum, who 
has kindly compared it with their specimens of saliens. He has 
stated that theirs are rather small, but that the specimen I sent him 
is unquestionably different. He has also compared this with the 
various badly stuffed types of A. Smith, but is unwilling, in view of 
the condition of the latter, to give any opinion. I have examined 
specimens from Italy, among which were reputed saliens, but my 
specimens are unquestionably not conspecific. 

In so far as | am able to judge from Smith’s figure (loc. cit.) ewronotus 
was probably identical with the present species. The first dorsal 
(in the figure) is inserted slightly behind midway between the base 
of the caudal and the tip of the snout, and the scale at the base of the 
first dorsal is shown to be about 8 in the distance from the origin 
of the first dorsal to the tip of the snout. Further, the maxilla is 


103 


Lhe Fishes of the Family Mugilidae in South Africa. 613 


drawn as if the tip would be hidden, or only just exposed, when the 
mouth is closed. The typical scaly second dorsal and the relatively 
large eye are not shown. Smith also mentions the presence of a row of 
“small criniform teeth in the upper jaw.” Besides this, Smith’s name 
possibly has reference to the markedly broad nuchal region,* which 
is characteristic. I am therefore provisionally reviving euronotus. 
This is preferable to instituting a new species, since ewronotus is so 
numerous and widely distributed in our fresh and brackish waters 
that it is more than likely that A. Smith actually secured a specimen. 

M. euronotus is very easily distinguished from all other South > 
African species by many features, chief of which are the very short 
scaly process at the base of the first dorsal fin and the characteristic 
premaxillary teeth. These, with the scaly nature of the soft dorsal 
and the markedly larger eye, serve to distinguish this species immedi- 
ately from capito in all stadia. 

On our southern coasts, in my experience, euronotus rarely occurs 
in the sea. I have caught and identified well over a thousand 
specimens of Mugil, from the sea, from estuaries, and from fresh 
water. Only two specimens of this species have been found in the 
sea; in each case near the mouth of a tidal river, and in one case 
after a flood. In tidal estuaries ewronotus is not usually found near 
the sea but high up the river, where the salinity of the water is low. 
Curiously enough, in Natal waters the species appears to be as 
commonly found in the sea itself. 

In the Eastern Province euronotus occurs in most of the fresh 
waters, in most cases in isolated pools which have no connection 
with the sea. The species appears to thrive in dams, into which it 
has been introduced. It appears to breed freely in such waters, and 
many farmers ensure a regular supply of fresh fish by stocking dams 
or pools on their farms. | 


Mugil capito Cuv. 


(Plates XVII, C; XIX, A-F.) 


1861. Giinther, Cat. Fish. B.M., vol. m1, p. 439. 

1916. Boulenger, F.W.F. Africa, vol. iv, p. 83, fig. 49. 

1918. Athanassoupoulos, Ann. Mus. Civ. Gen., vol. xlvi, p. 26. 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 304. 

* euronotus (=S. x S.E. wind) may, however, refer indirectly to the part of 
our area, 1.e. the south and south-eastern coastal regions, in which this species 
occurs, although Smith states that it inhabits the seas of the eastern and western 
coasts. Smith may have meant eurynotus. 


104 


614 Annals of the South African Museum. 


Depth 3-8-4-4, length of head 3-6-4-2 (J.) in length of body. Kye 
4-7 (J.)-6:3 (Ad.), snout 3-2-3-6, interorbital width 2:2-2-7, and 
postorbital length 1-8-1-9 in length of head. Adipose eyelids rudh- 
mentary, scarcely visible in juveniles, better developed in adults, 
but never extending further than the outer rim of the iris. Nostrils 
4 of eye diameter apart, anterior midway between anterior border 
of eye and profile of snout tip. Lower margin of preorbital scarcely 
bent, sometimes with a very small notch, lower and hinder margins 
serrate, scaly. End of maxilla well exposed. Angle of lower jaw 
93-103°, outline of jaw subangular in juveniles, more rounded in 


Fie. 8.—Diagram to show the dentition of Mugil capito. A, tongue; B, upper 
jaw and palate. M, maxilla; Pl, palatines; Pm, premaxilla; Pt, pterygoids; 
Vm, vomer. Dentate areas dotted. 


adults. Symphysial knob double. Upper lip fairly thin, width at 
snout apex 3-Dineye. Slightly flattened, recurved, subspatulate teeth 
in a single fairly widely spaced row in the upper jaw (Pl. XVII, C). 
Lower jaw edentate. Villiform teeth on vomer, palatines, pterygoids, 
and tongue (fig. 8). 

D IV +I, 8-9. First dorsal inserted nearer tip of snout than base 
of caudal, 1-06—-1-1 times as far from the latter as from the former, 
1:30-1:38 times as far from the hind margin of the mid-caudal rays 
as from the tip of the snout. First spine 1-7-2-0, base of first dorsal 
2-1-2-5 in head. Distance from origin of first to origin of second 
dorsal 0-95-1-2 (J.) in head. First dorsal inserted above the 14th— 
16th, second above the 28th—30th lateral scale. Pointed sheath scale 
extends behind origin of first dorsal 2-3-2-6 in head, 2:0-2-6 in distance 
from origin of first to origin of second dorsal, 4-5 in distance from 


105 


Lhe Fishes of the Family Mugilidae in South Africa. 615 


origin of first dorsal to tip of snout. Longest soft ray 1-9-2-3, base 
of second dorsal 2-9-3-5 in head. Last ray slightly longer than 
penultimate, fin slightly elevated anteriorly, edge gently concave. 
Second dorsal scaly basally and anteriorly only. 

A III, 9. Inserted slightly in advance of second dorsal, below the 
26th-29th lateral scale. Longest ray 1-8-2-3 in head. In shape 
fin resembles dorsal. Scaly anteriorly and basally. 

P 16-18, 1-45-1-75 in head (usually 1-6—1-7), tip reaches to the 
10th-12th lateral scale. Fin scarcely ever as long as postorbital 
plus eye, inserted 1-5-2-3 times as far from the ventral as from the 


N- 


Fic. 9.—Mugil capito Cuv. (see note, fig. 3). 


dorsal profile. Axillary scale 2-7 (Ad.)-4-5 in length of pectoral, 
obscure in young specimens. 

Ventrals 1-6 (J.)-1-9 in head, inserted below midway between origin 
of first dorsal and hind margin of head, or slightly behind or before. 
Axillary scale 3-3-3-6 in head. Edge of fin truncate. 

Caudal moderately forked, upper lobe slightly longer, mid-rays 
2-1-2-5 in head. 

Scales ctenoid (Pl. XIX, E and F), predorsal scales 1-0-1-2 (Ad.) 
times as long as wide. Mucus canal long and narrow. Very young 
fishes have cycloid dorsal scales, the denticulations develop with 
growth (Pl. XIX, A-E), lr. 44-48, ltr. 15-16; 4-5 cheek scales, 15-17 
predorsal to above the hind margin of the head. 

Colour.—Greenish to dull brown above, silvery below. Sometimes 
indistinct longitudinal streaks. Opercles usually with golden blotch. 

Localities.—Walfisch Bay, Lambert’s Bay, Table Bay, False Bay, 
Cape Agulhas, Port Beaufort, Knysna, Plettenberg Bay, Port Eliza- 
beth, Port Alfred, Great Fish Point, East London, Mazeppa Bay, 
Durban, Sinkwazi. Also in tidal rivers. 


106 


616 Annals of the South African Museum. 


Length.—Up to 405 mm. 

Fifty-five specimens, from 50 mm. up, examined. 

It appears to be reasonably certain that our specimens should 
be assigned to capito Cuv. 

I have received a number of specimens from the Zoological Station, 
Naples, among which is one very likely conspecific with the species 
I have here designated capito. Specimens of capito, auratus, and 
saliens were included, but the preservative employed in the package 
had unfortunately destroyed all the labels, so that I am unable to 
say which specimen was actually identified in Naples as capito. I 
have also examined a specimen of reputed capito from Holland, but 
this is not conspecific; if it 1s correctly named, our species is not 
capito. This specimen has pectoral 1:35 in head and has no axillary 
pectoral process, and in general outlines resembles auratus Risso 
rather than capito or saliens. 

Since, as is indicated below, I find it impossible from the literature 
to find any certain basis for the differentiation of saliens from 
capito, the most that can be said is that our specimens are probably 
identical with the latter species. 

Boulenger identified a specimen (No. 12048) in the 8.A. Museum 
as capito Cuv. He also identified as saliens Risso another specimen 
(No. 10157, from Table Bay), which I cannot by any means whatso- 
ever differentiate from the former; the latter even possesses a well- 
developed scaly process in the pectoral axil, the absence of which in 
saliens Boulenger (loc. cit.) makes his Key characteristic for differentia- 
tion from capito. | 

Giinther (loc. cit.) appears to have been satisfied that specimens 
from the Cape were identical with the European capito, but he re- 
marked that Smith’s specimens were badly stuffed, and of little use 
as types or for comparison—a fact which has recently been confirmed 
by a private communication from Mr. Norman of the British Museum. 

If one may Judge from the literature, a certain amount of mystery 
surrounds the identity of saliens. Giinther (loc. cit., p. 443) did not 
appear to be very certain of this species, and the features upon which 
he based his differentiation of saliens from capito (and from auratus) 
are inconstant and unreliable. Boulenger (F.W.F. Africa, loc. cit.) 
was obviously uncertain of saliens, and it may be remarked that this 
is the only African Mugil species of which he gives no figure. Barnard 
has evidently merely followed Boulenger in regard to the differentia- 
tion of saliens from capito. 

Athanassoupoulos (loc. cit.) has endeavoured to elucidate this, 


107 


Lhe Foshes of the Family Mugilidae in South Africa. 617 


but his conclusions have shed little light upon the problem, for he 
relies in his Key chiefly upon the supposed fact that the mouth of 
sahens is more convex than that of capito, which is at best of little 
practical value, and unlikely to be constant even were more precise 
details provided. This author has also proposed to use as diagnostic 
features certain dimensional relationships which would have to be 
tested over a wider range of stadia before their value can be accepted. 

Differentiation between capito and saliens, based solely upon the 
presence or absence of the axillary process of the pectoral, has been 
accepted by many systematists, but does not appear to be absolute. 
Most authors state that this axillary process is present in capito 
but absent in saliens, whereas Athanassoupoulos says that the latter 
species actually has a short process in the axil. As far as the South 
African specimens are concerned, a large process is apparent only 
in adults of capito. Juveniles have a very short process in the axil, 
and in a long and regular series of all stadia it may be seen that the 
size and length of this process increases regularly with age. In pre- 
served juvenile and half-grown specimens it is often exceedingly 
difficult to be certain whether the process is present or not, and so 
inconstant and unreliable is this feature that I should not venture 
to use it as a sole basis for differentiation in the present case. 

Not only does it appear certain that salens, as distinct from 
capito, does not occur in our area, but I have come, from the literature 
at my disposal, to doubt the validity of that species. At all events, 
it would appear that those who have specimens of capito, and of 
reputed saliens, must present stronger evidence for the maintenance 
of the latter species than has hitherto appeared. Athanassoupoulos 
states that the basal scale of the first dorsal of salens is shorter than 
the base of this fin, whereas in capito it is slightly shorter to slightly 
longer than the base, but he has given no quantitative data. This 
may eventually prove to be the key feature of any established 
differentiation. Among the specimens from Naples are two which 
agree in some respects with the general diagnosis of saliens. There 
is no process, or a very small one, in the axil, and the preorbital is 
deeply notched; the mouth is more obtuse, and the scale at the base 
of the first dorsal is relatively longer than in capito, while the pectorals 
are 1-3-1-35 in head. I cannot venture to make any statement about 
the identity of these specimens, but they are certainly different from 
any species from South Africa which I have examined. 

M. capito appears to be found throughout the greater part of our 
area, being most abundant on the West coast and round the Cape as far 


108 


618 Annals of the South African Museum. 


as Port Beaufort, in which parts it is the most important Mugil 
species. 

It is distinguished from other South African species by the scale- 
count, the very short pectorals, the scarcely obtuse mouth, and the 
well-exposed maxillary. | 


Mugil tricuspidens n. sp. 


(Plates XVII, A, F, G; XVIII, G, H.) 


1849. Smith, Illus. §8.A. Pisces, pl. xxx, fig. 1 (capensis C. & V.). 

1853. Pappe, Edible Fish. C.G.H., p. 27 (multilineatus). 

1861. Giinther, Cat. Fish. B.M., vol. i, p. 443 (salvens part). 

1925. Barnard, Ann. §.A. Mus., vol. xxi, p. 308 (auratus part). 

Depth 4-0-4:5, length of head 4-0 (J.)-4-4 (Ad.) in length of body. 
Eye 4:6 (J.)-5-4 (Ad.), snout 3-4-3-7, interorbital width 2-0-2-5, post- 
orbital length 1-8-2-0 in length of head. Adipose eyelids rudi- 
mentary, scarcely visible in juveniles, clearly visible in adults, 
posterior better developed. Nostrils 4 of eye diameter apart, 
anterior midway between front margin of eye and tip of snout 
profile. Lower margin of preorbital undulate, lower and hinder edge 
serrate. End of maxilla clearly visible. Angle of lower jaw 103-108°, 
outline of jaw rounded. Symphysial knob indistinctly double. 
Upper lip thin, width at apex of snout 3-4ineye. Flattened, apically 
dilated, recurved, tricuspid teeth (Pl. XVII, G) in a single series in 
upper jaw: in juveniles the teeth are more dilated, and the central 
cusp is spatulate (Pl. XVII, F). When viewed in fresh specimens, 
usually the central cusp only shows. The lip must be pushed back 
before the basal cusps are to be seen. The relatively large size of 
the teeth, and the wide spacing of the central cusps are distinctive 
characters. Lower jaw edentate. Villiform teeth on vomer, ptery- 
goids, and tongue. Palatines edentate. In adults the membrane of 
the roof of the mouth, and of the tongue, have close-set, apically 
dilated, tricuspid cilia. 

D IV +I, 8. First dorsal inserted 1-0 (J.)-1-07 times as far from 
base of caudal as from tip of snout, 1-25-1-35 times as far from hind 
margin of the mid-caudal rays as from tip of snout. First spine 
1-9-2-1, base of first dorsal 2-1-2-5 in head. Distance from origin 
of first to origin of second dorsal 1-0-1-1 times head. First dorsal 
inserted above the 15th or 16th, second above the 28th or 29th 
lateral scale. Pointed sheath scale extends behind origin of first 
dorsal, 2-5-2-8 in head, 5-4—5-8 in distance from snout tip to origin 


109 


The Fishes of the Family Mugilidae in South Africa. 619 


of first dorsal, 2-4-2-8 in distance from origin of first to origin of 
second dorsal. Second soft ray 1-5-1-8, base of second dorsal 2-6-2:8 
in head. Last ray longer than penultimate, fin anteriorly elevated, 
sub-falcate, edge concave. Second dorsal scaly only anteriorly and 
basally. 

A III, 9, inserted slightly in advance of second dorsal, below the 
27th—28th lateral scale. Second ray 1-6-1-8 in head, last ray longer 
than penultimate; edge of fin concave; scaly. 

P 18, 1-2-1-3 in head, tip reaches to the 11th or 12th lateral scale, 
inserted 1-6-2-0 times as far from the ventral as from the dorsal 


\ 


-" 
-- 
- 
oo 
~~ 
~~ 
~ 


Fic. 10.—Mugil tricuspidens n. sp. (see note, fig. 3). 


profile. No marked axillary scale in juveniles; a short, blunt, curved 
scale in adults, movable only in fresh specimens. _ 

Ventrals 1-7-1-8 in head, inserted below midway between base of 
pectoral and origin of first dorsal, edge of fin gently rounded 
Axillary scale 3-1-3-5 in head. | 

Caudal forked, upper lobe longer in adults, mid-rays 1:7-1-9 in head. 

Scales ctenoid: mucus canal short, oblique. Predorsal scales 
nearly as wide as long (Pl. XVIII, G and H), |r. 43-48, Ltr. 14-15. 
Four cheek scales, 16-17-predorsal to above hind margin of head. 

Colour.—Greenish above, silvery on sides and below. 7-8 very 
distinct longitudinal dusky streaks corresponding with the scale rows, 
visible in all but the very youngest stadia. Opercles golden or bronzy. 

Localities.—Mossel Bay, Knysna River, Zwartkops River, Buffalo 
River, Mazeppa Bay, Durban. | 

Length.—Up to 550 mm. 

Sixteen specimens, from 60 mm. in length up, examined. 

Types, from Knysna, in the Albany Museum. 

There is very little doubt that the specimens described by Smith 

VOL. XXX, PART 9. 42 


110 


620 Annals of the South African Museum. 


(loc. cit.), as capensis C. and V., belong to this species. The sole 
diagnostic feature in the description is the scale-count, which is 
valid also for capito or saliens. The figure, however, leaves no doubt 
about the identity with tricuspidens. 

It may be noted that in A. Smith’s pl. xxx, fig. 1 is below and fig. 2 
above. Pappe (loc. cit.) had evidently not noticed this, for he has 
obviously confused multilineatus with capensis. 

Boulenger (loc. cit.), presumably having seen both the type of 
capensis C. and V. and Smith’s specimen, stated that this latter is 
not capensis C.and V. Giinther (loc. cit.) accepted Smith’s diagnosis, 
but regarded the latter species as identical with euronotus Smith, 
and stated that both are identical with saliens Risso. 

The original description of capensis (C. and V., Hist. Nat. Poiss., 
vol. x1, p. 108) is so vague and brief that it is quite impossible even 
to guess what species was actually described. 

I am therefore provisionally naming this tricuspidens n. sp., and 
must leave the final pronouncement of the validity of this step to 
some worker who may be able to examine adequate material, in- 
cluding the type of capensis C. and V. 

As this species is normally estuarine, it is possible that it may 
prove to be endemic. 

It is well differentiated from our other species. by numerous features, 
chiefly by the relatively large tricuspid teeth, while the characteristic 
longitudinal stripes show up well even in preserved specimens. 

It may be noted that net fishermen at Knysna constantly dis- 
tinguish this species as the “Streepharder,” naming large specimens 
(unfortunately in common with large specimens of all species) 
“Springer.” 

M. tricuspidens does not appear to be anywhere very numerous nor 
specially gregarious, and, so far as I am aware, occurs only in tidal 
estuaries. Juvenile specimens are seldom encountered, and since 
the species is characterised by most extraordinary leaping powers, 
large numbers are rarely taken by the nets. I have at night in a 
boat frequently pursued adults of this species, which are exceedingly 
difficult to capture. When startled, large adults will leap anything 
up to 40 feet, rising 7 to 8 feet in the air, and the leap may be repeated 
six or seven times. The species may be clearly distinguished at 
night, when in the air, by means of a powerful light, the longitudinal 
stripes showing up clearly against the light silvery body. 

Specimens occasionally jump into a boat which carries a light ; 
large adults weighing 53 lb. have been taken in this manner, and I 


111 


Lhe Fishes of the Family Mugilidae in South Africa. 621 


have known a man to be knocked from his seat by the impact of 
one of these fishes on his chest. 

At Knysna ripe females are encountered in the late autumn and 
early spring. Specimens are usually encountered at night in shallow 
water on mud-banks, and are exceedingly shy. I have occasionally 
been able to approach specimens which have continued to circle over 
the mud, clearly visible in the light of the lamp, but the least move- 
ment of the light, or any noise in the boat, results in the characteristic 
leap. On one occasion a dozen or more large specimens broke water 
round the boat, and for some seconds the air appeared to be full .of 
silvery bodies, and the plunging leaps produced a considerable volume 
of sound. 


Mugil sehelt Forsk. 


(Plates XVI, C, and XVIII, C, D.) 


1888. Day, Fish. India, p. 355. 

1916. Boulenger, F.W.F. Africa, vol. iv, p. 91, fig. 53. 

1922. Weber and de Beaufort, Fish. Indo-Aust. Archip., vol. iv, 
p. 252. 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 306. 

Depth 3-8, length of head 3-8 in length of body. Eye 4-5, snout 
4-2, interorbital width 2-4, and postorbital length 1-8 in length of 
head. Adipose eyelids rudimentary. Nostrils 4 of eye diameter 
apart, anterior nearer the profile of the tip of the snout than the 
anterior margin of the eye. Lower margin of preorbital bent, 
notched, and serrate; scaly. Maxilla completely concealed. Angle 
of lower jaw 103°, outline of jaw angular. Upper lip thin, width at 
apex of snout 4 in eye. Symphysial knob double. No teeth in 
jaws visible. Vomer with traces of fine teeth, tongue with patches 
of villiform teeth. Palatines edentate. Exposed area on chin very 
short and narrow, would probably increase with age. 

D IV+I, 8. First dorsal inserted 1-1 times as far from caudal base 
as from the tip of the snout, 1-4 times as far from the tip of the 
mid-caudal rays as from the tip of the snout. First spine 2-1, base 
of first dorsal 2-9 in head. Distance from origin of first to origin of 
second dorsal 1-05 in head. First dorsal inserted above the 13th, 
second above the 26th lateral scale. Pointed sheath scale extends 
behind origin of first dorsal 2-4 in head, 2-1 in distance from origin 
of first to origin of second dorsal, and 4-2-5 in distance from origin 
of first dorsal to tip of snout. Longest soft ray shorter than the 
ventral fin and than the distance from hind margin of head to the 


112 


622 Annals of the South African Museum. 


centre of the eye, 1:8; base of second dorsal 2-7 in head. Last ray 
longer than penultimate, fin scarcely falcate anteriorly, edge moder- 
ately concave. Soft dorsal scaly. 

A III, 9. Inserted slightly in advance of second dorsal, below the 
25th lateral scale. Longest ray 1-7 in head, shape of fin similar to 
that of second dorsal; scaly. 

P 18, 1-1 in head, tip reaches to the 13th lateral scale, fin inserted 
3:3 times as far from the ventral as from the dorsal profile. Axillary 
scale long and pointed, 3-2 in length of head. Most of fin scaled. 

Ventrals 1-6 in head, longer than longest dorsal ray, inserted below 


Fie. 11.—Mugil seheli Forsk. (see note, fig. 3). 


midway between hind margin of head and origin of first dorsal or 
slightly nearer the former. Edge of fin gently rounded. Axillary 
scale 3in head. Fin almost completely scaly. 

Caudal deeply forked, upper lobe longer, mid-rays 1-7 in head; scaly. 

Scales more or less cycloid, traces of denticulations on exposed 
area. Mucus canal long and narrow (Pl. XVIII, Cand D). Predorsal 
scales as wide as long. Lat. ser. 39-41, ltr. 14, 3 cheek scales, 
14 predorsally to above the hind margin of the head. 

Colour.—Silvery, darker above. Axil of pectoral black. 

Localhity.—Durban, Chilka Lake, Bay of Bengal.* 

Length.—Up to 170 mm. 

Three specimens, from 167 mm. in length up, examined. 

Judging from the literature many authors are uncertain of the 
diagnosis of seheli. The majority agree in a scale-count of 38-42 
and in stating that the maxillary is hidden. Fowler (Proc. Ac. Nat. 
Sei. Phil., 1925, vol. lxxvi, p. 209) describes as seheli two specimens 


* A specimen kindly lent by the Director of the Indian Museum, Calcutta. 


113 


The Fishes of the Family Mugilidae in South Africa. 623 


from Delagoa Bay, and states that the scales are 33-35, but omits 
to mention whether the maxillary is hidden or exposed. These 
specimens can hardly be seheli; the description agrees closely with 
that of strongylocephalus Guthr., and it is possible that Fowler may 
have overlooked the adipose eyelids, although in the same paper he 
describes a specimen of this latter species (as longimanus Gnthr.) 
from the same locality. (But see note under canaliculatus.) 

M. caeruleomaculatus Lac. is by many authors held to be a synonym 
of seheli. I have, however, examined a specimen of the former species 
from India,* and it is quite clearly distinct. 

M. sehelt is not very abundant in our area, nor does it appear to 
extend south of Natal. It is apparently widely distributed in the 
Indo-Pacific. 


Mugil buchanani Blkr. 


(Plates XVI, D, and XX, C, D.) 


1861. Ginther, Cat. Fish. B.M., vol. iii, p. 446 (ceylonensis). 

1888. Day, Fish. India, p. 358. 

1916. Boulenger, F.W.F. Africa, p. 93, fig. 55 (ceylonensis). 

1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 305 (ceylonensis). 

1928. Fowler, Fish. Oceania, p. 123. 

Snout very broad and short, bluntly rounded anteriorly. Depth 
3°3-3'7, length of head 3-3 (J.)-4-0 (Ad.) in length of body. Hye 3-2 
(J.)-5 (Ad.), snout 3-5 (J.)—4-0 (Ad.), interorbital width 2-0-2-3 (J.), 
and postorbital length 1-7-1-9 in length of head. Adipose eyelids 
rudimentary, better visible in adults. Nostrils 4 of eye diameter 
apart, anterior as far from profile of snout tip as posterior from 
_ anterior margin of orbit. Lower margin of preorbital slightly bent, 
not, or slightly, notched; end truncated, lower and hinder edge 
serrated, scaly. End of maxilla completely concealed. Angle of 
lower jaw 110-122°, outline of jaw almost angular. Symphysial 
knob double. Upper lip thin, width at apex of snout 5ineye. Very 
minute ciliiform teeth in a single series in each Jaw in very young 
specimens, none visible in half-grown or adults. Vomer and palatines 
edentate. Pre-vomerine groove distinct. Villiform teeth on ptery- 
goids and round the anterior margin of the tongue. Space between 
rami of mandibles on chin almost absent in juveniles, gradually 
enlarges with age; long and wide in large adults. 

D IV+I, 8. First dorsal inserted 0-90 (J.)-1:15 (Ad.) times as 


* Kindly lent by the Director of the Indian Museum, Calcutta. 


114 


624 Annals of the South African Museum. 


far from the caudal base as from tip of snout, 1-23 (J.)-1-43 times as 
far from the hind margin of the mid-caudal rays as from the tip of 
the snout. First spine 1-7-1-9, base of first dorsal 2-0-2-2 in head. 
Distance from origin of first to origin of second dorsal 1:0—1-2 (J.) in 
head. First dorsal inserted above the 9th-12th, second above the 
20th-23rd lateral scale. Pointed sheath scale extends behind the 
origin of the first dorsal 2-6—2-9 (J.) in head, 2-4-2-6 in distance from 
origin of first to origin of second dorsal and 5-0-6-6 in distance from 
origin of first dorsal to tip of snout. Second soft ray longer than 
ventrals, and than distance from hind margin of head to centre of 
eye, 1:25-1-4 (J.); base of second dorsal 2-6-3 (J.)in head. Last ray 


fem. — 


Fig. 12.—Mugil buchanani Blkr. (see note, fig. 3). 


longer than penultimate; in adults the fin is anteriorly elevated, 
falcate, edge deeply concave. Second dorsal densely scaled. 

A III, 9. Inserted opposite origin of second dorsal. Second ray 
1-2-1-4 (J.) in head, last ray longer than penultimate; in adults the 
fin is anteriorly elevated, falcate, edge deeply concave, densely scaled. 

There is considerable alteration in the shape of the soft dorsal 
and anal fins with growth. In very young specimens (< 100 mm.) 
the anterior rays are fairly long, but the fin is not markedly falcate, 
since the middle rays are relatively longer than in the adult, and the 
edges of the fins are feebly concave. As the size of the fish increases, 
the anterior rays become relatively longer and the middle rays 
shorter, the fin assuming the anteriorly falcate shape when the 
length of the fish is more than +120mm. The scaling of the fins also 
increases from the very young to this size. In the former the basal 
scaling only is plain, there being apparently a mere sprinkling of 
light scales over the distal portions of the fins. 

P 17-18, 1-0-1-15 (J.) in head, tip reaches 11th—12th lateral] scale, 


115 


Lhe Fishes of the Family Mugilidae in South Africa. 625 


inserted 2-8-3-3 times as far from ventral as from dorsal profile. <A 
large scaly axillary process, 2-8-5-2 (J.) in length of head. 

Ventrals 1-5-1-7 in head, inserted below midway between hind 
margin of head and origin of first dorsal, or nearer the latter; edge 
of fin truncate. Axillary scale 2-8-3-4 in head. 

Caudal deeply forked, upper lobe longer, mid-rays 1-7 (J.)—2-0 in 
head, densely scaled. 

Ventrals and pectorals scaly basally. 

Scales cycloid, ventral scales very finely denticulate. Longitudinal 
length of predorsal scales 1-1 times width (Pl. XX, C and D), Lr. 33-36, 
l.tr. 13; 9-11 predorsal to above hind margin of head, 3-4 cheek scales. 

Colour.—Bright silvery, darker above. Indistinct longitudinal 
_ stripes. Axil of pectoral black, except in very small specimens. 

Localities.— Knysna, Durban, Chinde, Celebes.* 

Length.—Up to 385 mm. 

Kleven specimens, from 68 mm. up, examined. 

It appears to be fairly certain that ceylonensis Gnthr. is a synonym 
of buchanant Blkr. This was Day’s opinion (loc. cit.) after examining 
the types of both species, with which Fowler (loc. cit.) 1s apparently 
in agreement. 

This species has, beyond the early juvenile stages, a characteristic 
shape, which distinguishes it at a glance from all other South African 
species; buchanan and compressus are the only two species with 
markedly falcate dorsal and anal fins. M. buchananz is easily distin- 
guished from the latter by the very blunt rounded snout, by the shape 
and length of the ventrals, and by the concealed maxilla. 

It is probably widely distributed in the Indo-Pacific area. From 
the outlines this is a swift pelagic species. 

Probably many specimens now assigned to caeruleomaculatus Lac. 
will be found to be conspecific with buchanani (see notes under 


robustus). 
Mugil compressus Guthr. 


(Plates XVII, B, and XX, KE, F.) 
1861. Giinther, Cat. Fish. B.M.,. vol. ii, p. 451. 
1911. Gilchrist and Thompson, Ann. S.A. Mus., vol. xi, p. 42 


(diadema). 
1916. Boulenger, F.W.F. Africa, vol. iv, p. 94 (macrolepis part). 


1925. Barnard, Ann. §.A. Mus., vol. xxi, p. 309 (diadema). 


* A specimen, 107 mm. in length, kindly lent by Dr. de Beaufort, Curator of 
the Zool. Museum, Amsterdam. 


116 


626 Annals of the South African Museum. 


Depth 4-0-4-3, length of head 4-0-4-4 in length of body. Hye 
5-1-6, snout 3-3-4, interorbital width 2-0-2-2, length of postorbital 
part of head 1-7-1-8 in length of head. Adipose eyelids rudimentary, 
visible in adults. Nostrils + of eye diameter apart, anterior slightly 
behind midway between anterior margin of orbit and profile of snout 
tip. Lower margin of preorbital curved, not, or very slightly, notched ; 
serrate, scaly. End of preorbital obliquely truncated, edge slightly 
convex. End of maxilla clearly visible. Angle of lower jaw 105°, 
outline of jaw angular. Symphysial knob double. Upper lip thin, 
width at apex of snout 3 in eye. Very small recurved slightly com- 
pressed teeth in two series in upper jaw, the posterior series well 
back. Lower jaw edentate. Villiform teeth on pterygoids, and in 
patches round the anterior margin of the tongue. Vomer and 
palatines edentate. Exposed space between the rami of the mandibles 
long and narrow. 

D IV +I, 8. First dorsal inserted 0:95-1:05 times as far from tip 
of snout as from caudal base, 1-25 times as far from the hind margin 
of the mid-caudal rays as from the tip of snout. First spine 1-6-1-7, 
base of first dorsal 2-1-2-2 in head. Distance from origin of first to 
origin of second dorsal 0-9-1-:2 times head. First dorsal inserted 
above the 10th-llth, second above the 20th-22nd lateral scale. 
Pointed sheath scale extends behind origin of first dorsal 2-8-3:3 in 
head, 6-6-8 in distance from origin of first dorsal to snout tip, 3-2-3-5 
in distance from origin of first to origin of second dorsal. Highest 
soft ray longer than distance from hind margin of head to centre of 
eye, 1:2-1:3; base of second dorsal 2-6-2:9 in head. Last ray longer 
than penultimate, fin anteriorly elevated, falcate, edge deeply con- 
cave. Second dorsal densely scaled. First ray much longer than 
distance from hind margin of head to centre of eye. 

A Ill, 9. Inserted slightly in advance of second dorsal, below the 
19th-21st lateral scale. Second ray 1-1 in head, last ray longer 
than penultimate, fin anteriorly elevated, falcate, deeply concave. 
Densely scaled. 

P 16, 1-2 in head, tip reaches 8th—-9th lateral scale, inserted 2-6-3 
times as far from the ventral as from the dorsal profile. No axillary 
scale. | 

Ventrals 1-25 in head, longer than head without snout, inserted 
below 1:3-1-4 times as far from first dorsal origin as from hind margin 
of head. First and second rays elongate, fin sub-falcate. Axillary 
scale 3-8 in head. 


Caudal deeply forked, upper lobe longer, mid-rays 1:6-1-7 in head 


117 


Lhe Fishes of the Family Mugilidae in South Africa. 627 


Scales large, predorsal very finely denticulate, ventral scales more 
distinctly so. Predorsal scales as wide as long (Pl. XX, E and F), 
lr. 29-32 (Giinther 28), l.tr. 11, 10-11 predorsal to above hind margin 
of head. Five cheek scales. | 

Colour (Preserved).—Uniform light brown; silvery in life. Hind 
edge of scales dark. 

Localities.—Port Elizabeth, Durban, St. Lucia Bay, Kosi Bay. 

Length.—Up to 600 mm. 

Five specimens, all adults (one stuffed, Port Elizabeth Museum), 
examined. 

It has earlier been indicated that the majority of Giinther’s descrip- 
tions of Mugil species (loc. cit., pp. 417-460) are scarcely full enough 


i) — 
OF a ;; 


\ 


Fic. 13.—Mugil compressus Gnthr. (see note, fig. 3). 


to be of much diagnostic value. But that of the Australian species, 
compressus Gnthr., is an exception. I have very little hesitation in 
pronouncing diadema G. and T. synonymous. Nevertheless, as I have 
seen no Australian specimens, this diagnosis is provisional only. 

It must be confessed that I was led to search for some Indo-Pacific 
form, with which diadema might prove identical, by the outlines of 
this species, which indicate a swift, pelagic fish, possessing great leap- 
ing powers; likely to be widely distributed, but difficult to capture. 

In certain minor details only does the diagnosis of compressus 
differ from that of deadema. Giinther states that the former has 
98 series of scales: diadema has 29-32; this is well within normal 
limits of variation. The exposed surface on the chin of compressus 
is stated to be very short and narrow, in diadema it is long and 
narrow. I have found that the extent of this exposed area varies 
in one species, and increases with age. Long preservation in spirits 


118 


628 Annals of the South African Museum. 


might account for the highly compressed body of Giinther’s speci- 
men, a feature he regards as significant, but which I have observed 
in old spirit-preserved specimens of all species. In no significant 
feature does compressus differ from diadema. 

The outstanding characteristics which, with the small number 
and large size of the scales, immediately distinguish compressus from 
all other species are the very elongate ventrals, longer than the 
head without the snout, inserted much nearer to the hind margin of 
the head than to the origin of the first dorsal (see note under ventral 
fins, p. 595). 

Fowler (Fishes Oceania, Mem. B.P. Bishop Mus., 1928, vol. x, 
p. 125) considered compressus identical with macrolepis. Boulenger 
(loc. cat.) originally considered diadema synonymous with macrolepis, 
but later recognised (fide Barnard, loc. cit.) the former as distinct. 
Fowler has evidently missed the significant paragraph about the 
ventrals in the original description of compressus. In 1926 Fowler 
(Proc. Ac. Nat. Sci. Phil., vol. Ixxvii, p. 210) suggested that diadema 
is a synonym of oligolepis Blkr., which is not likely. It is scarcely 
possible that Fowler’s specimen is a juvenile compressus, since he 
stated that the pectoral was 1-6 and the ventral 1-4 in head, neither 
of which agrees with this species. 

It may be remarked that the elongate anterior dorsal, anal, and 
ventral rays are probably marked only in advance of the early 
juvenile stages. I have seen no young specimens, but these will 
probably prove difficult to distinguish from similar stadia of macro- 
lepis. The relative length of the scale at the base of the first dorsal 
will probably be of use in distinguishing juveniles. 


Mugil macrolepis Smith. 


(Plate XX, A, B.) 


1849. Smith, Ilus. 8.4. Pisces, pl. xxviii, fig. 2. 

1861. Giinther, Cat. Fish. B.M., vol. iii, p. 447 (smithit). 

1916. Boulenger, F.W.F. Africa, vol. iv, p. 94, fig. 56. 

1925. Barnard, Ann. S.A. Mus., vol. xxi, p. 309 (pl. xu, fig. 2, 
non macrolepis). 

Body of characteristic shape, usually with a false appearance of 
extra width between the anal and the soft dorsal fins. 

Depth 3-6-3-8, length of head 3-8 in length of body. Eye 4-1-4:3, 
snout 3-5-4, interorbital 2-2, postorbital length 1-9-2 in length 
of head. Adipose eyelids rudimentary but clearly visible in adults. 


119 


The Fishes of the Family Mugilidae in South Africa. 629 


Nostrils % of eye diameter apart, anterior midway between anterior 
margin of eye and profile of snout tip. Lower margin of preorbital 
bent downwards; not, or scarcely, notched; lower and hinder edge 
serrate. Hnd of maxilla exposed. Angle of lower jaw 105-108°, 
outline of jaw gently rounded, symphysial knob single. Upper lip 
thin, width at snout apex 5 in eye. Very small, slightly spatulate 
teeth, with a notch at each side of apex, in two rows in upper jaw; 
teeth fairly close-set in each row, the hinder row well behind the 
anterior. Villiform teeth on vomer, pterygoids, and anterior margin 
of tongue. Lower jaw and palatines edentate. Groove before vomer 
distinct. Exposed area on chin fairly short and narrow. 

D IV +I, 8. First dorsal inserted 1-04-1-08 times as far from the 


Fig. 14.—Mugil macrolepis A. Smith (see note, fig. 3). 


tip of the snout as from caudal base, 1-15—1-25 times as far from the 
hind margin of the mid-caudal rays as from tip of snout. First spine 
1-7, base of first dorsal 1-9-2-1 in head. Distance from origin of first 
to origin of second dorsal 1-0-1-:1 in head. First dorsal inserted 
above 13th—-14th, second above the 24th—25th lateral scale. Pointed 
sheath scale extends behind origin of first dorsal 2-5 in head, 5-5-8 
in distance from snout tip to origin of first dorsal, 2-3-2-5 in distance 
between origin of first and origin of second dorsal. Second soft ray 
shorter than hind margin of head to centre of eye, 1-7, base of second 
dorsal 3:1 in head. Last ray longer than penultimate, fin little 
elevated anteriorly, edge gently concave. Second dorsal completely 
scaly in adults. 

A III, 9, inserted slightly in advance of second dorsal, below the 
93rd lateral scale. Second ray 1:7 in head, last ray longer than 
penultimate, in shape resembles second dorsal. Completely scaled. 


120 


630 Annals of the South African Museum. 


P 16, 1-25-1-35 in head, tip reaches to the 10th—11th lateral scale, 
inserted twice as far from the ventral as from the dorsal profile. No 
axillary scale, or a very short blunt one, in adults. 

Ventrals 1-5 in head, shorter than head without snout, inserted 
below 1-0-1-2 times nearer hind margin of head than origin of first 
dorsal. First ray very slightly longer than remainder, edge of fin 
almost straight. Axillary scale 3-3-5 in head. 

Caudal slightly forked, mid-rays 1-8-1-9 in head. 

Scales finely ctenoid; predorsal scales about 1-2 times as long as 
wide. Mucus canal long and narrow (Pl. XX, A and B), Lr. 33-35, 
ltr. 12; 3 cheek scales, 12 predorsal to above hind margin of head. 

Colour.—Bright silvery, slightly darker above. 

Locality.—Mazeppa Bay, Durban, Isipingo River, Sinkwazi Lagoon, 
Kosi Bay. | 

Length.—Up to 305 mm. 

Thirty-four specimens, 66 mm. in length up, examined. 

There seems to be little doubt about the identity of the specimens 
described above. 

This 1s a very characteristic species. It is easily distinguished 
from all others from South Africa by the number of seales, by the 
absence of the long scaly process from the axil of the short pectoral, 
and by the exposed maxillary. 

Boulenger (loc. cit.) considered troscheli Blkr. a synonym of macro- 
lepis, while Fowler (Fishes Oceania, Mem. B.P. Bishop Mus., 1928, 
vol. x, p. 124) placed both troscheli and borneensis in the synonymy 
of macrolepis. On the other hand, Weber and de Beaufort (Fish. 
Indo-Aust. Archip., 1922, vol. iv, pp. 248, 249) considered troscheli 
and borneensis distinct from one another, and (evidently, since they 
made no mention of it) also from macrolepis. A careful analysis of the 
various descriptions appears to support Fowler’s conclusion. If this 
is correct, then macrolepis is widely distributed in the Indo-Pacific. 
Probably olwvaceus Day (Fish. India, p. 357) is not different. 

M. macrolepis appears to be fairly abundant on the Natal coast. 


Mugil canaliculatus n.sp. 


(Plates XVI, B; XVII, D; XVIII, E, F.) 
1925. Barnard, Ann. 8.A. Mus., vol. xxi, p. 303 (speiglert); p. 308 
(auratus, part). 7 
Depth 3-9-4-3, length of head 4-0 (J.)-4-5 in length of body. Eye 
4-0 (J.)-4-6, snout 3-3-9, interorbital width 2-3-2°5, and postorbital 


121 


The Fishes of the Family Mugilidae in South Africa. 631 


length 1-5-2-0 in length of head. Adipose eyelids visible even in 
juveniles, clearly visible in adults, posterior better developed than 
anterior, covering almost half of iris posteriorly. Nostrils } of eye 
diameter apart, anterior nearer tip of snout profile than front margin 
of eye. Preorbital deeply notched and bent downwards, end dilated 
and rounded, lower and hinder edges serrate. End of maxilla well 
exposed. Angle of lower jaw 108-112°, outline of jaw angular. 
Symphysial knob double. Upper lip thin, width at apex of snout 
3-4 ineye. Very fine recurved, compressed, apically truncated teeth 
(Pl. XVII, D) in a single series in upper jaw; juvenile and adult 
teeth identical. Lower jaw edentate. Villiform teeth on vomer,. 
palatines, pterygoids, and tongue. Exposed area on chin fairly 
long. 

DIV+I, 8. First dorsal inserted 1-0—1-05 times as far from caudal 
base as from tip of snout, 1-25-1-3 times as far from the hind margin 
of the mid-caudal rays as from the tip of the snout. First spine 
1:8-2:1, base of first dorsal 2-4-2-8 in head. Distance from origin 
of first to origin of second dorsal 0-95-1:1 times head. First dorsal 
inserted above the 13th—-14th, second above the 24th—-25th lateral 
scale. Pointed sheath scale extends behind the origin of the first 
dorsal 2-0-2:4 in head, 2:0-2-4 in distance from origin of first to 
origin of second dorsal, 4-3-4-9 in distance from origin of first dorsal 
to tip of snout. Second soft ray 1-6-1-7, base of second dorsal 
2-2-2°6 in head. Last ray longer than penultimate, fin anteriorly 
very slightly elevated, edge gently concave. Second dorsal scaly 
only anteriorly and basally. 

A III, 9. Inserted only slightly in advance of second dorsal, 
below the 23rd—25th lateral scale. Longest ray 1-7-1-8 in head, fin 
not much elevated anteriorly; scaly. 

P 16, 0-95-1-:15 (J.) in head, tip reaches to the 10th lateral scale, 
inserted 1-9-2-7 times as far from the ventral as from the dorsal 
profile. No axillary scale in juveniles, a very small obscure curved 
scale in adults. 

Ventrals 1-5-1:6 in head, inserted below 1-1-1:3 times further 
from the hind margin of the head than from the origin of the first 
dorsal. Edge of fin almost truncate. Axillary scale 2-8-3-0 in head. 

Caudal moderately forked, upper lobe longer in adults, mid-rays 
1-7-1-9 in head. 

Scales, dorsal weakly, ventral strongly, ctenoid. Predorsal seales 
slightly longer than wide, multicanaliculate to about the third row 
down (Pl. XVIII, E and F). Canalisation appears to increase with 


122 


632 Annals of the South African Museum. 


age, young fishes having 2-3, large adults up to 14 wavy canals on one 
scale. Lat. rows 36-39, l.tr. 13-14, 3-4 cheek scales, 13-14 pre- 
dorsal to above the hind margin of the head. 

Colour.—Dusky above, silvery below. Opercles dull. 

Localities —Knysna, Plettenberg Bay, Port Alfred, Great Fish 
Point, East London, Mazeppa Bay, Durban, Delagoa Bay. Also in 
tidal rivers. 

Length.—Up to 285 mm. 

Forty-three specimens, from 90 mm. up, examined. 

Types, from Knysna, in the Albany Museum. 

It is probable that this species must previously have been described, 
but I cannot yet with certainty assign it to any known species. 


Fic. 15.—Mugil canaliculatus n. sp. (see note, fig. 3). 


canaliculatus is very close to, if not actually identical with, hoefleri 
Stndnr., from Senegambia. Beyond the absence of the adipose 
eyelids in this latter species, there appears to be little difference 
between them. I have unfortunately been unable to obtain one of 
Steindachner’s types for comparison, but Dr. Pietschmann of Vienna 
has kindly sent me an accurate drawing of one of the few remaining 
predorsal scales on the only scaled type of hoefleri, and this scale 
resembles those of canaliculatus in being multicanaliculate. Never- 
theless, as I have not seen any of these West African types, and in 
view of the widely separated recorded areas of these species, it would 
appear better to maintain both for the present. 

It is most likely that canaliculatus occurs in the Indo-Pacific, but 
I have not been able to recognise it from the descriptions of any 
species from this area. Barnard (loc. cit.) had identified one of the 
S.A. Museum specimens as speigleri Blkr., and others as auratus 
Risso. I have examined a specimen of speigleri from India, and 
canaliculatus is quite definitely distinct. I have also examined 


123 


Lhe Fishes of the Family Mugilidae in South Africa. 633 


specimens from Italy, among which were reputed auratus, and 
canahculatus, while related, is certainly different. The latter has 
fewer scales, longer pectorals, and better developed adipose eyelids 
than the northern species. 

It is probable that it is canaliculatus which Boulenger (F.W.F. 
Africa, p. 88) identified as auratus (from East London). Boulenger’s 
figure of auratus (loc. cit., fig. 50) might well pass for the former 
_ species. If the marked adipose eyelids and the scale-counts are 
overlooked, it would evidently be easy to confuse these two species, 
although I have forwarded a specimen to Mr. Norman of the British 
Museum, and he states that it is quite distinct from their specimens 
of auratus. 

Fowler (Proc. Ac. Nat. Sci. Phil., 1925, p. 209) has described two 
specimens from Delagoa Bay as seheli Forsk., which cannot be that 
species, and are possibly canaliculatus. 

M. canaliculatus is exceedingly abundant on the South and East 
coasts, at least as far as Delagoa Bay. It enters tidal rivers, but 
does not appear to ascend very far. At Knysna it may be seen that 
canaliculatus abounds up to about five miles from the mouth of the 
river; in the higher part of this area tricuspidens is also found, 
together with cephalus. Both of these latter species extend several 
miles farther up the river to the point where the water is only 
slightly saline. Beyond this stage, cephalus and euronotus are found. 

M. canaliculatus does not appear to attain a large size. It is appar- 
ently only those species which are largely fluviatile, such as cephalus 
and tricuspidens, which grow very large. 

At Knysna and Great Fish Point, ripe females of canaliculatus are 
observed during August and September. 

M. canaliculatus does not apparently possess any marked leaping 
powers. This may have some connection with the markedly | 
posterior insertion of the ventrals. 

The canalisation of the scales, and the long pectorals, immediately 
distinguish this from all other South African species. 


Mugil waigiensis, Q. and G. 
(Plate XX, G, H.) 
1861. Giinther, Cat. Fish. B.M., vol. ii, p. 435. 
1888. Day, Fish. India, p. 359, pl. Ixxiu, fig. 4. 
1916. Boulenger, F.W.F. Africa, vol. iv, p. 97, fig. 59. 
1922. Weber and de Beaufort, Fish. Indo-Aust. Archip., vol. iv, 
p. 244. 


124 


634 Annals of the South African Museum. 


1925. Barnard, Ann. §.A. Mus., vol. xxi, p. 310. 
2 1928. Fowler, Fishes Oceania, p. 124, fig. 27. 

Dorsal profile flat, interorbital flat, head very depressed at occiput. 
Depth 4, length of head 3-1-3-3 in length of body. Kye 3-8-4-1, 
snout 3-0-3-6, interorbital width 2-1-2-2, and length of postorbital 
1-9-2-1 in length of head. Adipose eyelids rudimentary. Nostrils 
3-5 in eye diameter apart, anterior midway between profile of snout 
tip and anterior margin of eye. Lower margin of preorbital bent, 
not, or slightly, notched, serrated. End of maxilla exposed. Angle 
of mouth 95-97°, outline of lower jaw angular. Symphysial knob 


Fic. 16.—Mugil waigiensis Q. and G. (see note, fig. 3). 


single. Upper lip thin, width at apex of snout } of eye. No teeth 
in jaws or on palate. Exposed area on chin short and narrow. 

D IV+I, 8, first dorsal inserted 1-1-1-18 times as far from tip 
of snout as from caudal base, 1-3 times as far from hind margin of 
mid-caudal rays as from tip of snout. First spine 1-9-2, base of 
first dorsal 3-3-5 in head. Distance from origin of first to origin 
of second dorsal 1-25 in head. First dorsal inserted above the 8th, 
second above the 16th-18th, lateral scale. Pointed sheath scale 
extends behind origin of first dorsal 3-3-4 in head, 6-8 in distance 
from origin of first dorsal to snout tip, 2-6-3-5 in distance from 
origin of first to origin of second dorsal. Second soft ray 1-6-1:8 
base of second dorsal 3-6 in head. Edge of fin scarcely concave. | 

A III, 8, inserted in advance of second dorsal, below the 15th—-16th 
lateral scale. Second ray 1-6 in head. Shape similar to dorsal. 

P 16, 1-3 in head, tip reaches 7th—8th lateral scale, inserted 1-7— 
2-2 times as far from ventral as from dorsal profile. No axillary scale. 

Ventrals 1-5-1-6 in head, inserted below 1-3 times as far from the 


125 


The Fishes of the Family Mugilidae in South Africa. 635 


origin of the first dorsal as from hind margin of head. Edge of fin 
gently rounded. Axillary scale 3-5-4 in head. 

Caudal almost truncate, lobes equal, mid-rays 1-3-1-4 in head. 

Scales ctenoid, predorsal scales as long as wide, mucus canals short, 
lanceolate (Pl. XX, G and H); lr. 26-28, 1.tr. 9-10, 8 predorsal to 
above hind margin of head, 3 cheek scales. 

Colour (Preserved).—Light brown, probably silvery in life. Pec- 
torals partly or wholly dark. Vertical fins with dark margins. 
Dark longitudinal streaks. 

Localities. —Chinde, Delagoa Bay. 

Length.—Up to 100 mm. 

Three specimens, from 44 mm. up, examined. 

Kasily distinguished from all other South African species by the 
small number of scales, together with the feebly emarginate caudal 
and the markings. Fowler’s specimen from Delagoa Bay (Proc. Ac. 
Nat. Sci. Phil., 1925, vol. Ixxvu, p. 209), described as oligolepis Blkr., 
does not appear to be very different from waigrensis (see notes under 
oligolepis). This author’s figure of wazgiensis (loc. cit.) differs in many 
respects from my specimens, and from most descriptions. The 
pectorals are shown to be about 1-8 in head, the first dorsal is inserted 
nearer the snout tip than the caudal base; the distance from the origin 
of the first to the origin of the second dorsal is about equal to the head, 
and there are 30 rows of scales: Day’s figure (loc. cit.) is also rather 
singular in many respects. Barnard (loc. cit.) states that the maxilla 
is concealed. This is an error, if based on the 8.A. Museum specimens. 

The synonymy of this species appears to be somewhat extensive. 
A revision of material from all parts of its recorded area might show 
that several related species have been confused. 

M. waigiensis is apparently widely distributed, and fairly common, 
throughout the whole of the Indo-Pacific region. It is not very 
common on our coasts. 

From the outlines, this is probably a somewhat sluggish species. 


Mugil olygolepis Blkr. 
(Plates X XI, B, and XXII, C, D.) 


1861. Giinther, Cat. Fish. B.M., vol. i, p. 452 (melinopterus, 
C.and V.?2). 
1888. Day, Fish. India, p. 358, pl. lxxvi, fig. 2. 
1992. Weber and de Beaufort, Fish. Indo-Aust. Archip., vol. iv, 
p. 245, and p. 246 (melinopterus, C. and V.). 
VOL. XXX, PART 5. 43 


126 


636 Annals of the South African Museum. 


1925. Fowler, Proc. Ac. Nat. Sci. Phil., vol. Ixxvu, p. 209. 

Body moderately robust, well compressed posteriorly. Head broad 
and depressed, snout slightly rounded. Front profile of snout fairly 
blunt, formed by upper lip. Depth 3-5, length of head 4-0 in length 
of body. Eye 4-1, snout 3-5, interorbital 2-2, postorbital length 2-0 
in length of head. Adipose eyelids moderate, anterior weak, posterior 
better developed, covering about 4 of the iris. Nostrils ~ eye 
diameter apart, anterior slightly behind midway between anterior 
border of eye and snout tip profile. Lower margin of preorbital 
bent slightly downwards, not notched, edge strongly serrate. End of 
maxilla well exposed. Angle of lower jaw 106°, outline of jaw 
angular, margins slightly rounded; symphysial knob single. Upper 
lip fairly thin, width at snout apex 4 in eye. Minute curved teeth 
in a single close-set series in upper jaw; lower jaw, vomer, and pala- 
tines edentate. A few small patches of minute teeth on the outer 
margin of the tongue. The medio-longitudinal ridge on the tongue 
higher than in other species. Prevomerine groove not very convex 
posteriorly. Exposed area on chin lanceolate, long and narrow, 
with anterior constriction. 

D IV+I, 8. First dorsal inserted exactly midway between snout 
tip and caudal base, 1-4 times farther from the tip of the mid-caudal 
rays than from the tip of the snout. First spine 1-6, base of first 
dorsal 2:0 in head. The spines are very much stronger than those of 
any other South African species. They are also more close-set, the 
4th not remote from the others, being apically almost adnate to the 
3rd (this may possibly be a deformity). When the spinous dorsal 
is folded down, the exposed parts of the spines, excepting the anterior 
margin of the first, are scaly. Distance from origin of first to origin 
of second dorsal equal to head. First dorsal inserted above the 9th, 
second above the 19th lateral scale. Pointed sheath scale extends 
behind origin of first dorsal 2-1 in head, 4:4 in distance from snout 
tip to origin of first dorsal, and 1-1 in the postorbital part of the head. 
Longest soft ray 1-6, base of second dorsal 2:7 in head. Last ray 
very little longer than penultimate, fin not falcate, edge gently con- 
cave. Second dorsal completely scaly, with heavy basal scaly sheath. 

A III, 9, anterior half of base in advance of second dorsal, inserted | 
below the 16th lateral scale. Longest ray 1-5 in head, last ray scarcely 
longer than penultimate, edge of fin gently concave. Densely scaly, 
especially basally. 

P 15, 1-25 in head, tip reaches to the 8th lateral scale, inserted twice 
as far from the ventral as from the dorsal profile. No axillary scale. 


127 


The Fishes of the Family Mugilidae in South Africa. 637 


Behind and below the upper part of the base of the fin is a small 
scaled cutaneous projection. Fin scaly on basal half. 

Ventrals 1-4 in head, inserted below 1-1 times nearer hind margin 
of head than origin of first dorsal. Edge of fin almost truncate, very 
slightly emarginate. Axillary scale 3-5 in head. Interventral scaly 
process rather wide and heavy. 

Caudal forked, mid-rays 1-7 in head. 

Scales very-finely ctenoid; predorsal scales slightly longer than wide, 
mucus canal rather narrow (Pl. XXII, C and D). On the lateral 
scales the mucus canal posteriorly communicates with a system of rudi- 
mentary canals or grooves, more or less arborescent. Lat. ser. 27, 


Fic. 17.—Mugil oligolepis Blkr. (see note, fig. 3). 


ltr. 10. Three cheek scales, 8-9 predorsal to above the hind margin 
of the head. 

Colour.—Olive grey above, lighter below. Tips of dorsals darkish. 
Tip of upper lobe and hind margin of caudal dusky. Weak axillary 
spot. Faint streaks along the scale rows. 

Locality—Isipingo Lagoon, near the sea. 

Length.—206 mm. 

A: single specimen examined. 

A most rare and elusive species, sought for almost three years 
without success until recently. Not known as a separate species to 
the Indian netters in the neighbourhood of Durban. 

This specimen is very probably conspecific with that described 
by Fowler (loc. cit.) from Delagoa Bay. Fowler’s specimen had a 
narrower interorbital, while the markedly robust dorsal spines, 
obvious in my specimen, were not mentioned by him. Further, 
Fowler stated that the ventrals were 1:4 and the pectorals of his 


128 


638 Annals of the South African Museum. 


specimen were 1-7 in head, but this may be an error. I have not 
seen any species of Mugil which has the pectorals so markedly shorter 
than the ventrals. Further, variation in length of the pectoral 
from 1-7 to 1:25 (my specimen) in head is far too wide for any one 
species. 

Day’s description and figure (loc. cit.) do not agree, and the latter, 
though most likely drawn from a juvenile, does not agree very well 
with my specimen. Nevertheless it is very likely that they are 
conspecific. 

No descriptions of oligolepis, to which I have access, mention the 
posterior eyelid, which is very clear in my specimen. 

It would not indeed be surprising to find that the synonymy of this 
species is somewhat extensive. Giinther’s account of melinopterus 
C. and V. (loc. cit.) fits my specimen almost exactly, whereas his 
account of oligolepis (loc. cit., p. 449) does not. 

M. nepalensis Guthr. (loc. cit., p. 424), of which I have seen no figure, 
and none but the original description, appears to be very closely 
related to, if not identical with, olgolepis. 

There appears to be little of significance in Weber and de Beau- 
fort’s descriptions (loc. cit.) of oligolepis and of melinopterus to warrant 
their maintaining the two as distinct. In the length of the pectoral 
my specimen agrees with their account of olkigolepis, whereas in the 
presence of the adipose eyelid it agrees with their melinopterus. It 
would appear that these two species are synonymous, or, at any rate, 
the specimens described by these authors are all of one species. 

It may be noted that all descriptions of oligolepis that I have seen 
have been based on apparently juvenile specimens. A careful study 
of adequate material will probably show that oligolepis is merely the 
juvenile form of melinopterus. 


SPECIES LIKELY TO BE DISCOVERED IN SoutH AFRICA. 


There are five species, widely distributed in the Indo-Pacific, 
which with more intensive collection will probably be found in our 
area. It seems desirable to indicate these, and to give a brief account 
of them and of their synonymy. 

Those which occur in the Red Sea, or nearer our area, have been 
selected. Of these I have examined specimens of caeruleomaculatus 
Lacep. and of spezglert Blkr. only. 

An abbreviated composite Key, to enable these species to be recog- 
nised, is appended. 


129 


The Fishes of the Family Mugilidae in South Africa. 639 


Mugil tade Forsk. 


1861. Giinther, Cat. Fish. B.M., vol. iii, p. 426 (parsia H-B), 
and p. 427 (belanak Blkr.), and p. 428 (planiceps C. and V.). 

1922. Weber and de Beaufort, Fish. Indo-Aust. Archip., vol. iv, 
p- 236. 

1928. Fowler, Fishes Oceania, p. 122. 

Adipose eyelids present. Maxilla exposed. Pectorals shorter 
than head without snout, with short axillary scale. Origin of first 
dorsal nearer snout tip than caudal base. Caudal feebly emarginate. 
DIV +I, 8-9, A III, 9, Lr. 33-35. 

Distribution.—Indo-Pacific (Red Sea). 


Mugil sperglert Blkr. 

1861. Ginther, loc. cit., p. 435. 

1888. Day, loc. cit., p. 348. 

1922. Weber and de Beaufort, loc. cat., p. 241. 

1928. Fowler, loc. cit., p. 123. 

Adipose eyelids present. Maxilla visible. Pectorals 1-1-:15 in 
head, with long axillary scale. Scale at base of first dorsal as long as 
postorbital part of head. Origin of first dorsal nearer to snout tip 
than caudal base. Soft dorsal and anal densely scaled. Caudal 
forked. DIV+I, 8; ATIITI, 9; Lx. 40-42. 

Distribution.—Indo-Malayan area (Red Sea). 


Mugil cunnesius C. and V. 


1861. Giinther, loc. cit., p. 434. 

1922. Weber and de Beaufort, loc. cit., p. 242. 

1928. Fowler, loc. cit., p. 123. 

Adipose eyelids present. Maxilla exposed. Pectorals shorter 
than head without snout, with long axillary scale. Origin of first 
dorsal nearer snout tip than caudal base. Soft dorsal and anal 
scaleless. DIV+I, 8; A III, 9; lr. 42-43. 

Distribution.—Indo-Malayan area (Red Sea). 


Mugil labiosus C. and V. 


1861. Giinther, loc. cit., p. 454. 

1922. Weber and de Beaufort, loc. cit., p. 259, fig. 67. 
1928. Fowler, loc. cit., p. 126. 

Upper lip very thick, with a single series of papillae. 


130 


640 Annals of the South African Museum. 


No adipose eyelids. Maxilla exposed (but stated to become hidden 
in large specimens?). Pectorals as long as head, with short axillary 
scale. First dorsal about midway between caudal base and snout 
tip. Caudal emarginate. DIV+I, 7-8; A III, 9-10; lr. 34-36. 

Distribution.—Indo-Malayan area (Red Sea). 


Mugil caeruleomaculatus Lac. 


1922. Weber and de Beaufort, loc. cit., p. 250. 

No adipose eyelids. Maxilla concealed. Pectorals 1-1 in head, 
with long axillary scale. First dorsal nearer snout tip than caudal 
base, or midway. Scald at base of first dorsal long, 1-8 in head, 
about as long as postorbital part of head. Caudal forked. No 
exposed area on chin. DIV+I,8; ATIII, 9; Lr. 36-38. 

Distribution.—Indo-Pacific (Zanzibar ?). 


ABBREVIATED CoMPOSITE Key. 


I. Adipose eyelids well developed, covering most of'the iris 


posteriorly. 
A. Scales 33-36. 
1. Pectorals longer than head without snout . strongylocephalus. 
2. Pectorals not longer than head without snout . tade. 
B. Scales 38-43. 
1. Anal with 7-8 soft rays . . . . . cephalus. 
2. Anal with 9 soft rays. 
a. Maxilla concealed . . . . . robustus. 
b. Maxilla exposed. 
i. Soft dorsal completely scaly . . spergleri. 
ii. Soft dorsal not scaly . . . CUNNESLUS. 


II. Adipose eyelids small or rudimentary, covering not more 
than half of the posterior portion of the iris. 
A. Upper lip very thick, almost half eye diameter at 
| snout tip, with papillae on lower margin. 
1. Papillae in 5 or 6 series. Pectorals 1-3—1-4 in 


head. Scales 37-40 . . . crenilabis. 
2. Papillae in one series. Pectorals 1-1-1 in head. 
Scales 34-36 . . . labiosus. 


B. Upper lip not more than + eye deep at snout tip, 
without papillae. 
1. Dorsal scales multicanaliculate. (Pectorals 1- 
1-1 in head) . . . . . . canaliculatus. 
2. Dorsal scales not multicanaliculate. 
a. Scales 41-49 (caudal forked). 
i. Pectorals not longer than head with- 
out snout. 


131 


The Fishes of the Family Mugilidae in South Africa. 641 


x. Soft dorsal completely scaly . euronotus. 
y. Soft dorsal not completely scaly . capito. 
ii. Pectorals longer than head without | 
snout. 
x. Teeth tricuspid. Maxilla  ex- 
posed . . . . tricuspidens. 
y. Teeth not tricuspid. Maxilla 
concealed . . . . sehelt. 


b. Scales 29-40 (caudal forked). 
i. End of maxilla concealed. 
x. Pectorals longer than head with- 
out snout. 
a. Scale at base of first dorsal 
shorter than ? of post- 
orbital part of head. 

* Scales 38-42 . . sehelt. 
** Scales 33-36 . . buchanan. 

B. Scale at base of first dorsal 

about as long as post- 


orbital part of head . caeruleomaculatus. 
y. Pectorals not longer than head 
without snout . . . robustus. 


ii. End of maxilla exposed. 
x. Ventrals longer than head with- 


out snout . . . COMPressUs. 
y. Ventrals shorter than head with- 
out snout . . . . macrolepis. 


c. Scales 26-28. 
i. Caudal almost truncate. Pectorals 
black . . . ; ; waigiensis. 
ii. Caudal emarginate. Pectorals light .  olagolepis. 


HaBIts, BREEDING HABITS, ETC. 


The general habits of Mullets are fairly well known, since these 
fishes live mainly inshore, on the surface and in shallow water. They 
are fairly easily captured and appear to thrive in aquaria. 

In the latter may be seen how they project the mouth as a scoop 
and suck in sand or mud, triturate this for a time, and finally reject 
what has proved inedible. 

Mullets appear to be largely herbivorous, but will eagerly feed upon 
soft flesh, such as the liver or intestines of fishes, even of their own 
kind. They will also take insects which have fallen into the water. 
At Knysna I have observed a shoal feeding upon fallen flying ants 
which were over the water. Despite this, the species do not appear 
to take an artificial fly. 


132 


642 Annals of the South African Museum. 


The intestinal contents nearly always consist very largely of sand. 
The intestine is very long, and the unabsorbed residue from the 
alimentary tract is little else but sand and shell particles. In the 
stomach itself, besides sand and mud, green algae, and fragments 
of marine plants, may occasionally also be found eggs, larval crustacea 
and fishes. 

In tidal estuaries these fishes appear to congregate, especially 
at night, upon the sand- or mud-banks, where the mud and the eel- 
grass (Zostera) both teem with many lowly forms of life. From a 
boat I have watched shoals of small half-grown Mugil feeding. 
They swim in a compact body, facing the current, moving at the rate 
of a few inches a minute. They constantly suck up the mud, or the 
slime on the grass, retain it for perhaps 20 or 30 seconds, and eject 
the hard portion. The movements of such a shoal are extremely 
erratic. In the van are often to be seen a number of individuals who 
shoot forward some inches and then drop back into the main body. 
The shoal will veer as a whole to one side, or will move rapidly forward 
for a few feet and then resume the slow advance. Occasionally a 
shoal will be seen to break up, dart some distance back, and cover 
the same area at a slow pace. 

These small shoals nearly always consist of individuals of more 
or less constant size. On one occasion only did I see a large specimen 
feeding with a shoal of others of very much inferior size. A lucky 
cast with a throw-net secured this specimen, which proved to be 
cephalus, while the others were canaliculatus. 

I have never been fortunate enough to see a shoal of large specimens 
feeding in this manner. At night, with a powerful light, I have 
frequently been among large numbers of adults, but only occasional 
specimens came into the light; these were merely swimming idly 
and appeared uneasy, some sheering off wildly for no apparent reason, 
while others would swim until almost against the side of the boat 
before taking fright. 

Mullets seem to be timid, but exceedingly curious. From a high 
rock, overlooking moderately deep water, I have dropped stones into 
the middle of a shoal. The fishes scatter widely, but almost immedi- 
ately turn and circle in a dense cloud in the disturbed area. If a 
handful of crushed liver be thrown amongst them, the same perform- 
ance results, and a fierce mélée ensues until all has been consumed. 
Any sudden movement of an exposed part of the observer results in 
the rapid departure of the shoal, if the fishes be of any size. 

“Harders” are generally captured by means of nets, but specialised 


133 


The Fishes of the Family Mugilidae in South Africa. 643 


methods of angling are also employed with success. In tidal rivers 
very small hooks mounted on fine gut, buoyed with small corks and 
baited with dough, or with various fancy concoctions, are employed. 
In the sea at various places, liver or fish bait on tiny hooks is generally 
used. A large “Harder” or “Mullet,” especially tricuspidens, pro- 
vides magnificent sport on light tackle. Successful angling depends 
upon a close study of the habits of these fishes, and requires consider- 
able skill and patience; this sport has not yet found favour with the 
majority of anglers in South Africa. 

In the fresh waters of the Eastern Province “Springer” (ewronotus) 
fishing is largely indulged in. To the line near the hook are fastened 
a number of small corks in a series, so as to keep 4 or 5 feet of the 
line on the surface. At the end, on some inches of fine gut, is mounted 
a small hook. The favourite lure is a ‘“‘flying-ant.”” The fishes 
usually bite well towards evening, and large catches are frequently 
made. 

In the Eastern Province the majority of the species appear to 
breed in September and October. The shoals come into shallow water 
all along the coast, and at night lie in the shallows, where the eggs 
are presumably shed and fertilised. At this time the fishes appear 
to be much less timid, and, by the aid of a light, may at night be 
scooped up in numbers with a landing net. I have in this fashion 
secured numbers in pools at the edge of the surf, both males and 
females, fully ripe. The females appear to outnumber the males 
by as much as ten to one. 

It has frequently been stated that ripe fishes seek out tidal estuaries 
for the purpose of spawning. The evidence I have been able to obtain 
does not substantiate this belief, which is probably only partly correct. 
The fishes are certainly more plentiful in such waters at these times, 
but I have observed that they are also more numerous in the shallow 
water of the sea itself. 

In estuaries, when every fish taken was ripe, I have at night, from 
a boat, with the aid of a light observed from two to four fishes slowly 
circling over shallow water on mud-banks. On some occasions, in 
still water, I have been able to keep them in the illuminated area for 
as long as ten minutes. On occasions there has appeared to be an 
ejection of faintly opalescent matter which instantly disappeared, 
but this has not been observed to be followed by the ejection of milt. 
I have several times captured, with a throw-net, two or three speci- 
mens so engaged, and on each occasion one of the fishes proved to 
be a ripe male, the remainder females, so it may be presumed that 


134 


644 Annals of the South African Museum. 


they were engaged in spawning. It is, nevertheless, singular that 
tow-netting over these presumed spawning areas has in no case 
resulted in a catch containing eggs which could be shown to be 
those of ripe Mugils. 

The somewhat oily flesh of the “Harder” is of fine texture and 
delicate flavour, and probably of relatively high calorific value. 

These fishes are, especially in the Western Province of South 
Africa, highly esteemed, and a large “‘Harder,’’ baked whole, is 
undoubtedly a culinary delicacy. 

Vast numbers are caught annually, chiefly by drag-nets, whole 
shoals being encircled in the surf. On the west coast fair numbers 
are taken by floating gill-nets, anchored near the shore. 

Large numbers are salted and dried, and these form an important 
part of the diet of the poorer section of the coastal population. 

The flesh of those fishes taken far up in estuaries is generally 
slightly less palatable, while those taken from the inland waters of 
the Hastern Province have a distinctly unpleasant “muddy” flavour. 


I wish to express my gratitude to the Director of the South African 
Museum for his kindness in assisting with the loan of the whole of 
the 8.A. Museum collection of Mugil species, and of literature. To 
the Research Grant Board of South Africa (Carnegie Fund) for 
generous financial assistance, which has defrayed the greater part 
of the expenses incurred in the investigation. Also to Messrs. H. J. 
Koch and B. Hindson for valuable collections from Natal. 

I must also acknowledge my indebtedness to Dr. C. von Bonde, 
Director of the Government Fisheries Survey, and to Mr. Bell-Marley, 
Principal Fisheries Officer of Natal, for permission to net in preserved 
waters. 

ALBANY MUSEUM, 


GRAHAMSTOWN, 
July 1934. 


Ann. S. Afr. Mus., Vol. XXX. 
Plate XV, 


A, Latero-occipital scale of Mugil cephalus Linn., from specimen 405 mm. in length, 
to show secondary scaling. The lines radiating from S indicate rows of 
minute superimposed scales. 

B, A portion of the integument, from a medio-lateral scale of the same species 
(length 275 mm.), showing the small cycloid scales embedded therein. 

C, View of the branchial and lower pharyngeal regions of the same species (length 
405 mm.). G, Rakers of the four branchial arches; P, Pharyngeal rakers. 


J. L. B. Smith, — Neill & Co., Ltd. 


Ann. S. Afr. Mus., Vol. XXX. Plate XVI 


SAR 
ARE 


Mugil species. A, strongylocephalus Rich.; B, canaliculatus n. sp.; C, seheli 
Forsk.; D, buchanani Blkr.; E, euronotus A. Smith. “The line below each 
figure represents 1 cm. | 


J. L. B. Smith. ; Neill & Co., Ltd. 


Ann. 8. Afr. Mus., Vol. XXX. Plate XVII 


eo 


A, Mugil tricuspidens, n. sp.; B,* Mugil compressus, Gnthr. . 

C-G, premaxillary teeth of Mugil species. The total length of the specimens from 
which the teeth were taken is given in brackets. The line below each figure 
represents a tenth of a millimetre. 

C, capito Cuv. (280 mm.); D, canaliculatus n. sp. (210 mm.); E, euwronotus Smith 
(230 mm.); F, tricuspidens n. sp. (60 mm.); G, tricuspidens n. sp. (405 mm.). 


* Copied by permission, from a photograph taken by the Director of the Natal 
Museum. 


J. LI. B. Smith. Neill & Co., Ltd. 


Plate XVIII, 


Ann. 8. Afr. Mus., Vol. XXX. 


*( 


"SJOYORIG UL UDAIS ST UdY¥R} SI oTVOS OY} YOIYM UWWOIZ UTUTOOdS 944 Jo YASUI] [e404 OUT, ‘WW YT 
oul] oyy, ‘aired yoo jo 4siy 


WwW FEZ) ‘ds ‘u snzojnoWouvd 


‘A pue a *( 


WUE LOT) “YsIog yoyas “q pue C$ 


ey} ale sefvos [esiopaiq ‘soioeds pln jo 


( 


(‘urut 99z) ‘ds -u suapidsnois3 ‘FY pue 5 


‘uu C6| 


) 


yoy snpoydasojhbuo.s 


6 


q pue Vv 


syuosoidel oTeos YORe MOTOq 


sovos j[erzuoaysod-ptu pue yesiopoid yp 


Neill & Co., Ltd. 


JL. B. Smith, 


Plate XIX. 


Ann. 8S. Afr. Mus., Vol. XXX. 


‘oules Jo ayvos [erquoAqsod-prul ‘Fy 

‘(*WUW CZZ) YYWIG snjgouoima jo sjeos jestopeid yyy “H f(wWU OEZ) “AND opidvd Jo oTeoS yeryuUoAysod-pru “_ { ("UIT 
O€@) H ‘(wu OFT) Gd ‘Cm OZ) O (Cw OTT) gq “(CUM C6) Y +'AnD opdno Jo sopeos jessoperd yyy ‘oatsnpoutr q-V 

"WU [ SyueseIder o[eOs YORE MOTOq OUT] OYJ, ‘SJoYOVIG UI USATS st Udyey 


SI 9[BoS oY} YOIM worl, uswtosds 944 Jo YYSue, [e{04 oY, ‘soloods pbnyy Jo sapeos yeayuoaysod-pru pue yesropeid yyF 


Neill & Co., Lid. 


J.L. B. Smith. 


Ann. 8. Afr. Mus., Vol. XXX. 


‘(WU OOT) “D PUL "OH sisuarbimm “Fy pue H 
(‘WU OEE) “1yQUD snssaudwoo ‘q pue q f("wUE 0OZ) “TAI wwounyong ‘qT pue O ‘(WUE LTT) YPWg sidajospm ‘gq pue V 
"SJOYORIG Ul USATS SI UdYV4 SI BTVOS OY} YOIYM WOI, UsUMTOods oY} Jo ygsugy [e}0} 94, “WU T syuesoidel oyvos Yove MOTOG 
oul] 84, ‘aeos [esiopord oy} st ared yoo jo 4siy oy, ‘soroods piubnypy jo soyeos [eryueaysod-prur pue jesiopeid yyL 


Neill & Co., Ltd. 


J. L. B. Smith. 


Plate X XI. 


Ann. 8. Afr. Mus., Vol. XXX. 


‘td | Squsssider oinsy 


wg, sadajobyo pbngy “a + 


e 


YoRed MOTI IT] OU, 


yyy snysnqgos qonqy SV 


, Ltd. 


ill & Co., Lt 


Ne 


J. iL. B. Smith. 


Ann. S. Afr. Mus., Vol. XXX. Plate XXII. 


mal , cco 
Scales of AMugil species. 


A, 7th predorsal of robustus Gnthr. (200); B, mid-postventral of same; C, 7th 
predorsal of oligolepis Blkr. (206); D, mid-postventral of same. 


The line below each figure represents 1 mm. The length of the specimen, in 
millimetres, from which the scales were taken, is given in brackets. 


J. L. B. Smith. Neill & Co., Ltd. 


135. 


Transactions of the Royal Society of South Africa. Vol. XXIII. 
Part Ill. pp. 265—276. Pls. XIlI—XVII. September, 1935. 


THE “GALJOEN” FISHES OF SOUTH AFRICA. 


By J. L. B. Sirs. 


(With Plates XIII-XVII.) 
(Read May 15, 1935.) 


Famity DICHISTIIDAE nov. 


Body compressed, ovate or elevated. Mouth terminal, slightly protractile. Maxilla 
expanded distally, with large supero-median expansion (Pl. XVII, F); extremity not 
entirely covered by preorbital. No supramaxilla. Premaxillary pedicels moderate 
(Pl. XVII, G), not reaching frontals. A small pointed supero-median expansion on 
each premaxillary ramus, sliding behind maxilla. A single series of elongated com- 
pressed incisiform teeth in each jaw, with similar, but very much smaller teeth behind 
(PI. XVII, Cand G). Palate and tongue edentate. 

Gill-membranes united forming a moderate fold across the throat, narrowly fused 
below with isthmus. 

Two nostrils, close together. 

Air bladder not constricted, nor posteriorly bifurcated. 

Parietal crests very low: no transverse ridge at posterior margin of frontals. Supra- 
occipital fairly elevated, anterior edge thickened. A strong subocular shelf. 

Vertebrae 25(10 +15): precaudals with parapophyses from the fourth. Ribs, inserted 
behind and above level of parapophyses, sessile except last three. 

Spinous dorsal of stout spines, shorter than soft fin. Anterior to the spinous dorsal 
several overlapping antrorse recumbent spines below the skin, the anterior (inferior) 
resting on the apex of the supraoccipital, the posterior ankylosed with the basipterygials 
of the first two exposed spines. Soft fins densely scaly, scaling not to margin. 

Seales ctenoid, strongly adherent, deeply embedded. 


The various taxonomic positions hitherto assigned to the genus Dichistius 
Gill would appear to have been selected in more or less arbitrary fashion. 
Regan (Ann. Mag. Nat. Hist., 1913 (8), vol. xii, p. 127) stated that it 
“probably pertained” to the Girellidae, in which family it was later placed 
by Barnard (Ann. 8.A. Mus., 1927, vol. xxi, p. 635). Fowler (Proc. Ac. 
Nat. Sci. Phil., 1925, vol. Ixxvu, p. 233) at first considered Dichistius to fall 
in the Sparidae; later (Bull. U.S. Nat. Mus., 1933, vol. xu, p. 216) in the 
Kyphosidae, while more recently (Proc. Ac. Nat. Sci. Phil, 1934, 
vol. Ixxxvi, p. 476) for one of the species he has proposed a new genus 


placed in the Seorpididae. 
A critical examination of the genus Dichistius has shown that there is 


136 


266 — Transactions of the Royal Society of South Africa. 


little to choose between these various diagnoses, since this genus appears 
to differ in about equal degree from the general diagnosis of each of those 
four families, showing perhaps least affinity with the Girellidae. It would, 
however, appear to differ from any one of the families Kyphosidae, Sparidae 
and Seorpididae by characters at least as significant as those which are 
generally accepted as justifying the separate existence of those three 
families. 

The dentition, the exposed maxilla, and the nature of the vertebral 
column and appendiculars, among other features, would appear to rule out 
its inclusion in the Girellidae. The strong subocular shelf, and the nature 
of the vertebral column are not Kyphosid, while these features, together 
with the united gill-membranes, would scarcely allow of its inclusion in the 
Scorpididae. It is also well differentiated from the Sparidae by the rela- 
tions of the maxulary bones alone. 

Dichistius would actually appear to have closest affinity with the 
Scorpididae. In some respects, however, it would appear to be a connecting 
link between that family and the Kyphosidae, e.g. according to Regan 
(loc. cit., p. 126) the Seorpididae all have 25(10+15) vertebrae, with para- 
pophyses from the 3rd or 5th precaudals: the Kyphosidae (Regan, loc. cit.) 
have 24(10+14), with parapophyses from the 4th. Duchistius has 25 
vertebrae, with parapophyses from the 4th. The dentition also is inter- 
mediate between that of the Kyphosidae and that of the Scorpididae, 
approximating, however, more closely to that of the former family. 

From the skeletal and structural evidence I have obtained, it would 
appear as reasonable to unite the families Scorpididae and Kyphosidae as 
to place Dichistvus in either. It is now proposed to regard Dichistius as 
the type of a separate family. It is frankly admitted that this diagnosis 
can at best be provisional only, since I have at my disposal only one genus, 
Kyphosus Lac. of the KypHosipaz, and only Neoscorpis J. L. B. Smith of 
the Seorpididae. For further information in regard to skeletal features 
in those families I have relied upon Regan (loc. cit.). It may here be in- 
dicated that Neoscorpis differs from Regan’s diagnosis (loc. cit., p. 126) of 
the Scorpididae in that there is no subocular shelf. (I have examined a 
head of Parascorpis Blkr., and a strong subocular shelf is present.) Further, 
the hydrostatic organ is posteriorly bifurcated into two caudal extensions 
after the fashion of that of Kyphosus (that of Parascorpis does not, fide 
Dr. Barnard, in litt.). Nevertheless Neoscorpis agrees in the main with 
Regan’s diagnosis of the Seorpididae, in which it undoubtedly falls. 

It may be indicated that the nature of the obsolescent portion of the 
dorsal fin in Perciform fishes is a feature which has apparently not been 
assigned its true taxonomic significance, and which may assist in elucidating 
hitherto obscure relationships. 


137 


The “Galjoen” Fishes of South Africa. 267 


In Kyphosus there are 11 exposed dorsal spines. Anterior to and quite 
separate from these, below the skin, is the rudiment of a single-spined, 
probably separate, dorsal, the spine being so far obsolete that it cannot 
be determined whether it was originally antrorse or retrorse. In Neoscorpis 
there is evidence of the existence at some period of two separate spinous 
dorsals, of which the first was composed of three spines, the anterior of 
which was antrorse and procumbent, the middle spine either antrorse or 
retrorse, while the posterior spine was likely normal and retrorsely depres- 
sible. The posterior spinous dorsal of Neoscorpis is also evidently under- 
going reduction. That fin would appear at some stage to have consisted 
of 10 exposed spines, proportioned much as the 10-spined fin of Dichistius. 
From 2 to 4 of the anterior spines have already become obsolete, and reduced 
to subcutaneous dilations of the basipterygial apices. In some specimens 
the anterior exposed spine is so minute that its presence may be ascertained 
only upon dissection. 

In Dichistvus, immediately anterior and adjacent to the exposed 10- 
spined dorsal, is a subcutaneous structure of four overlapping antrorse 
procumbent spines, of which the posterior (and superior) is fused with the 
united basipterygials of the first two external spines (PI. XVII, B), while 
the anterior rests upon and over the supraoccipital apex. It may be noted 
that an exactly similar structure is present in Plataz Cuv., Drepane Cuv., 
and Tripterodon Plyfr. Dichistvus is related to this group of allied genera 
in other ways, and in some respects falls intermediate between these genera 
and the natural group of the Scorpididae, Kyphosidae and Sparidae. 

The species of the single genus of the Dichistiidae are confined to the 
southern African region, and are found only in shallow water. 


Genus DicHistius Guill. 


1888. Gill, Proc. U.S. Nat. Mus., vol. ii, p. 68. (Type capensis Cuv. Dichistius pro- 
posed to replace Dipterodon Cuv., preoce.) 

1927. Barnard, loc. cit., p. 635. (Dipterodon Cuv.) 

1933. Fowler, Bull. U.S.N. Mus., vol. xii, p. 215. (Coraconus Gron.) 

1934. Fowler, Proc. Ac. Nat. Sci. Phil., 1934, vol. Ixxxvi, p. 476. (Drepanoscorpis.) 


Body compressed, elevated or ovate. Mouth terminal, slightly oblique, protractile. 
Maxilla not completely covered by preorbital. Posterior teeth in each jaw incisiform, 
very small, in two rows, inserted on roof and floor of mouth well behind the exterior 
incisors. Pharyngeal teeth enlarged, obtuse, molariform (Pl. XVII, D and E). 

Gill-rakers moderate in size and number. Branchiostegals 7, basal membrane scaly. 
Gill-membranes scaly, united, forming a fold across the throat, narrowly united with 
isthmus. — 

Two nostrils, close together, the anterior much the larger, each with a plain flap. 
Preorbital of moderate and uniform depth, lower margin entire, posterior portion scaly. 


Preopercle margin serrate (Pl. XVII, A). . . 
Dorsal with deep scaly sheath, of 10 stout spines, the middle the highest, deeply 


138 


268 Transactions of the Royal Socrety of South Africa. 


notched between spinous and soft portions. Soft dorsal longer than spinous, anterior 
rays forming a prominent lobe, sometimes sub-falcate. Soft fin densely scaly for at 
least the basal half. Anal of three spines, first short and stout, remaining two subequal 
in length. Soft anal of 13 to 14 rays, the anterior elevated forming a lobe. Scaly as for 
dorsal. Ventrals inserted well behind pectorals, of moderate size, slightly shorter than 
pectorals. Caudal slightly forked. ll fins scaly. 

Scales firmly adherent, small, ctenoid, lateral line tubular scales much smaller than 
adjacent scales. 


The original generic name was Coracinus Gron. [cauda lunata Gron., 
1763]. 

According to Mr. Norman of the British Museum, by Opinion 89 of the 
International Commission, Gronovius is ruled out, unless his name can be 
shown to have been adopted by some later accepted binomial author. 
There appears to be no record of any such usage between 1763 and 1888, 
when Gill (Joc. cit.) proposed Dichistvus. In view of the generally accepted 
Rules and Opinions of the International Commission, therefore, it appears 
that Dichistius Gill must replace Coracinus and Dipterodon. 

Fowler’s Drepanoscorpis (loc. cit.) is somewhat difficult to understand. 
His description of the species Drepanoscorpis gilchristi agrees in all parti- 
culars with his own earlier descriptions of Dipterodon capensis C. and V., 
and of Coracinus capensis C. and V. It is possible that Fowler recognised 
gilchristi as being identical with those specimens earlier described as capensis, 
but he makes no mention of this. There is no room for doubt that gilchristi 
Fowler is identical with the species multijasciatus Pell. (see below). Since 
that species and capensis are emphatically congeneric, Fowler’s genus falls 
into the synonymy of Dichistius. 

The original species of this genus must have been one of the earliest 
typical Cape fishes to be recognised by the early settlers, as it is fairly 
abundant in shallow water. Only one species was recognised until Pelle- 
grin (Bull. Soc. Zool. Fran., 1914, vol. xxxix, p. 231) described multifasciatus 
(type locality Madagascar) and distinguished it from the common Cape 
capensis C. and V. This diagnosis has not been accepted by subsequent 
workers, partly because Pellegrin laid stress upon certain features which 
can hardly ever be of much taxonomic significance. Nevertheless as is 
shown below, multefasciatus is a perfectly valid species, being clearly differ- 
entiated from capensis by numerous features. 

Examination of material from all parts of the South African coast has 
revealed that a third species, falcatus n. sp., is also present. This may 
immediately be distinguished from the other two species by its general 
outline, but it has proved exceedingly difficult to reduce the differentiation 
to an immediately appreciable quantitative form. This species is exceed- 
ingly close to capensis, and may even be regarded as merely a sub-species. 


139 


The “Galjoen” Fishes of South Africa. 269 


Since trinomialism does not yet appear to be generally accepted ichthyo- 
logical practice, falcatus is here described as a distinct species. 

The following table indicates some of the features in which differentia- 
tion between the three species is found. 


capensis. falcatus. mulirifasciatus. 
Depth peduncle in body depth . 3-1-3-2 3-0-3-2 3:9-4-1 
Pectorals in body length . 4-0-4:1 3-6-3-8 3°5-3°8 
Ventrals in body length . . 4-5-5-0 4-0-4:5 3:7-4:0 
Preorbital depth in head . 6-5-7:2 6-0-6-6 5°3-5°6 
Base of soft dorsal with length 0 of head Less than. Less than. Greater than. 
Lateral rows of scales . . 75-78 73-78 83-86 
Dorsal profile curve . . . | Broken at eye. Smooth. Smooth. 
Anterior dorsal rays . . . Not falcate. Falcate. Not falcate. 


Besides these features, there are numerous others, which are indicated 
below in the descriptions. The distribution of the three species is of 
interest: capensis appears to occur from the Cape to about Kast London, 
falcatus n. sp. from about Knysna to Natal, and multifasciatus from about 
Algoa Bay to Madagascar. 


Key to the Species. 


I. Least depth of peduncle 3-0~3-2 in body depth. 
Body more than twice as long as deep. Cross bars, if present, of 
uniform width. 
A. Pectorals 4:0-4:1 in length of body. Anterior dorsal rays not 


falcate . . capensis. 
B. Pectorals 3-6-3: 8 in length of body. "Anterior dorsal rays 
falcate . . . . falcatus. 


II. Least depth of peduncle 3° 9-4. ‘lin depth of body. 
Body less than twice as long as deep. Cross bars alternately wide 
and narrow . . . . . . . . . multifasciatus. 


The length of the pectoral is measured from the body to the tip of the 
fin when the latter is held at right angles from the side. 


Dichistius capensis (C. and V.). 
(Plate XIIT.) 


1831. Cuvier and Valenciennes, Hist. Nat. Poiss., vol. vil, p. 276, Pl. 188. 

1908. Gilchrist and Thompson, Ann. S.A. Mus., vol. vi, p. 165. (Dipterodon capensis 
C. and V.) 

1914. Gilchrist, Mar. Bio. Rep., vol. ii, p. 90. (Dipterodon capensis.) 

1927. Barnard, Ann. S.A. Mus., vol. xxi, p. 635. (Dipterodon capensis, part), Pl. 25, 
fig. 2. 


VOL. XXIII, PART III. 19 


140 


270 Transactions of the Royal Society of South Africa. 


Body ovate, compressed but robust, maximum width 1-6-1-7 in head. Dorsal profile 
gently sloping to interorbital, steeper to snout tip. Snout distinctly blunt, almost 
vertical from prominent interorbital in some specimens. : 

Depth 2-2, length of head 3-2-3-3 in length of body. Eye 3-8 (Juv.)—4:5, snout 3-0-3-1, 
interorbital 2-5-2-6, postorbital part of head 1-7-1-8 in length of head. Preorbital 
uniformly deep, depth above end of maxilla 6-4—7-2 in head. The distance from the 
upper margin of the dorsal sheath at base of 4th dorsal spine to the nearest point of 
the lateral line distinctly less than post-orbital part of head. 

Mouth moderate, terminal, slightly oblique, maxilla extends to below or almost 
‘below anterior border of orbit: in small specimens below anterior part of eye. Distance 
from hind margin of maxilla to upper margin of upper lip at snout tip 2-1 in postorbital 
part of head. Upper lip deep and fleshy, depth at snout apex 2ineye. 20-26 compressed 
elongate incisiform teeth in upper jaw, 20-24 similar in lower jaw. Two rows of about 
8 each of similar smaller teeth, concealed in fleshy pads, behind the outer series in each 
jaw. Palate and tongue edentate. 

Gill-membranes united, forming a smooth fold across the throat, fused anteriorly 
below with isthmus. Gill-rakers moderate, 5-6 +13-15, 1-8-2 in gill-filaments, which 
are 1-0-1-3 in eye. 

D X, 18-19, inserted above midway between pectoral and ventral origins or slightly 
nearer the pectoral, deeply notched between spinous and soft portions. Spines fairly 
stout. First spine 10-14, 2nd 4-6, 3rd 2-5-3-3, 4th and 5th 2-2 in length of head, thereafter 
graduated shorter. Soft rays anteriorly elevated forming a blunted or pointed lobe, not 
falcate. Lobe extends beyond tip of tenth dorsal spine 1-7—2-1 in head, equal to or less 
than postorbital part of head. Base of soft dorsal shorter than head. 

A III, 13-14, inserted below the origin of the soft dorsal. Spines stout, first 5-9-6-6, 
second 4-4-5, third 45-48 in length of head. Soft rays anteriorly elevated forming a 
broad lobe, which extends 1-9-2-2 in head beyond the apex of the third anal spine. 

Pectorals 4-0-4:1 in length of body, tip reaches below more than an eye diameter 
short of the origin of the soft dorsal. 

Ventrals 4-5-5-0 in length of body, tip does not reach origin of anal; in large specimens 
the tip does not reach beyond the vent. 

Caudal moderately forked, peduncle 1-1-1-2 times as long as deep. Least depth of 
peduncle 3:1-3-2 in depth, 6-7-7-0 in length of body. 

Scales ctenoid, larger on posterior part of body, all deeply embedded. Lateral line 
gently curved, tubular scales smaller than adjacent scales. Lateral rows 73-78, tubular 
scales 60-64; 28-30 above lateral line obliquely back from base of first dorsal spine. 
Whole body and head, except muzzle, scaly. Median fins scaly; on soft dorsal and anal 
scaling extends about half length of rays and is not very dense beyond the basal third. 
Bases of alternate doisal spines scaly. Pectorals and ventrals scaly about three-fourths. 

Colour.—Rather variable when taken from the water; either bright silvery, or with 
light or dark irregular blotches, or wholly dark. Uniform or with 7-9 uniformly wide 
cross bars, wider than the interspaces. Dorsal spines, dorsal and anal lobes, and ventral 
rays black or dusky. After death the colour is usually uniform dusky or almost black, 
with or without cross bars. 

Locality.—Walfish Bay to Table Bay (Gilchrist); Cape Peninsula to Port Alfred, very 


occasionally entering tidal rivers, then rarely beyond the reach of the waves. 
Length.—Up to 600 mm. 


Hight specimens from 140 mm. up examined. 
D. capensis is the well-known “Galjoen” (Ang. Galleon) of the Cape. 


141 


The “Galjoen” Fishes of South Africa. 271 


It is a somewhat polymorphous species, and sub-species will most likely 
eventually be proposed. It may immediately be distinguished from 
multifascratus by the shallower body, by the heavier peduncle, and by the 
body markings, besides other features: from falcatus by the markedly 
shorter fins, as well as by the characteristic snub nose. 

The colouration of the live fish is extremely variable, even in specimens 
from one restricted area. The live fish may be uniform, or blotchy, or 
mottled in various shades of silver to black. Cross bars are rarely observed 
on the live fish in the summer months, generally appearing, if at all, when 
the fish is moribund. The banded form is commoner eastwards from 
Knysna, and is more frequently observed during the winter months. 
When the fish is moribund the melanophores evidently expand fully, for 
whatever the colour of the live fish, soon after death it is almost black. 
It is possible that there is a nocturnal barred colour phase, such as I have 
observed in the species at present recognised as Sparus sarba Forsk, but 
as the “Galjoen” does not appear to feed during the night, information 
on this point is not available. Gilchrist (loc. cit.), who studied the colour 
phases of this species in an aquarium, is rather vague about the banded 
phase, and does not appear to have examined the specimens at night. 

Gilchrist had in the aquarium noticed the great ease with which this 
species could turn and reverse its progress. These fishes must be extremely 
powerful, for they evidently swim freely even in turbulent surf. The 
general habits and environment of this species have frequently been out- 
lined, and detailed repetition is unnecessary, but one or two characteristics 
call for special mention. From a rocky ledge overlooking moderately 
deep water at the edge of a reef, I have been able to attract several of this 
species by dropping into the water pieces of “Red-bait” (an Ascidian: 
Pyura stolonifera). These the fishes would seize right in the surface of 
the water. After some time, the bait was thrown on to the reef itself, and 
after several times seizing bait which had been swept back by the receding 
wave, on several occasions one of these fishes actually came over the reef 
in the surge of the wave in pursuit of a piece of bait, seized this, and returned 
to deeper water before the wave had subsided. 

That this species will perform a feat of this nature is evidently well 
known. One angler has reported having seen one of these fishes actually 
left stranded on an isolated rock by one wave, and returning to the water 
by the next. Another who had noticed these fishes passing over a rock 
in the surges of the waves states that he actually hooked a fish by casting 
his bait on to the rock itself. In the Eastern Province of South Africa this 
species is frequently taken by anglers in the surf at high tide, being often 
actually hooked on a spot which at low tide would be above the level of 
the water. From this characteristic has resulted the Hastern Province 


142 


272 Transactions of the Royal Society of South Africa. 


name of “Highwater.” It has been observed that when this species 1s 
hooked in relatively shallow water over broken rocky ground, it will 
occasionally jump clear of the water, more often than not thereby freeing 
itself from the hook. 

This species appears to feed very largely upon the common mussel 
(Mytilus edulis), the intestinal residues generally consisting chiefly of 
crushed mussel shells. The stout anterior teeth of the “Galjoen” are 
admirably adapted for wrenching or twisting shellfish from the rocks, 
and these mollusca are doubtless crushed between the pharyngeals, which 
bear heavy crushing molariform teeth (Pl. XVII, D and E). Seaweeds also 
at times form a large part of the diet of this species. 

Ordinary ‘‘Red-bait”’ is the common bait employed on lines, but large 
specimens have also been taken on fish bait, as well as on “‘Sea-cat”’ 
(Polypus sp.). 

It may be noted that in areas much frequented by anglers this species 
has diminished most markedly in numbers within the last thirty or forty 
years, so that where large catches were at one time of common occurrence, 
the capture of even a single specimen may now be an event. On the other 
hand, in places difficult of access and so rarely fished, it is still possible to 
catch large numbers of these fishes at any normal time. There is every 
indication that this species moves within a comparatively restricted area, 
so that intensive fishing may well produce a local scarcity. 

The spawning habits of this species are unknown. It is curious that 
the specimens examined have all been immature males. It might be 
suggested that anglers in the Western Province should examine fishes 
caught in the winter months and forward specimens of ripe females to the 
South African Museum. 


Dichistius falcatus n. sp. 
(Plates XIV and XVII.) 


Body ovate, fairly compressed, maximum width 1-6 in head. Dorsal profile gently 
curved, interorbital prominence very slight, snout little steeper below. 

Depth 2-0—2-1, length of head 3-0-3:3 in length of body. Eye 4-3-5-0, snout 3-0-3:3, 
interorbital 2-5, postorbital part of head 1-7-1-8 in length of head. Preorbital uniformly 
deep, depth above end of maxilla 6-0-6-6 in head. The distance from the upper margin 
of the sheath at the base of the fourth dorsal spine to the nearest point of the lateral 
line is equal to or less than the postorbital part of the head. 

Mouth terminal or even occasionally sub-inferior, slightly oblique, maxilla extends to 
below anterior border to first third of eye, extremity not covered. Distance from hind 
margin of maxilla to upper margin of upper lip at snout tip 1-7—-1-9 in postorbital part of 
head. Upper lip deep and fleshy, depth at snout apex 1-8-2 in eye. 20-24 compressed 
rather elongate incisiform teeth in a single row in each jaw, anterior subtruncate, lateral 


143 


The “Galjoen” Fishes of South Africa. — 273 


teeth more acute. Two series of about eight each of similar but much smaller teeth 
behind the outer series in each jaw, concealed in fleshy pads. Palate and tongue edentate. 

Gill-membranes united forming a fold across the throat, narrowly united below with 
isthmus. Gill-rakers 5-6 + 13-14, 2-2-5 in gill-filaments, which are 1-1-1 in eye. 

D X, 19-20, inserted above slightly nearer pectoral than ventral base, deeply notched. 
Spines fairly stout, heteracanth, fold into deep scaly sheath. First spine 10-12, 2nd 
4:3-4:5, 3rd 2-5-2-6, 4th 2-1-2-2, 5th 2-0-2-1 in head, thereafter shorter to the 9th. Soft 
rays anteriorly elevated, falcate or subfalcate, edge of fin deeply concave. The tip of the 
lobe when depressed usually reaches above beyond the apex of the depressed last anal 
ray, and extends beyond the apex of the 10th dorsal spine 1-5-1-7 in head, usually longer 
than postorbital part of head. Base of soft dorsal shorter than head. 

A III, 13-14, inserted below slightly before origin of soft dorsal, spines moderately 
stout, first 5-5-6, 2nd 3-3-3-5, 3rd 4-0-4:3 in head. Soft rays anteriorly elevated forming 
a subfalcate lobe, which extends 1:6—1-:7 in head beyond the apex of the 3rd anal spine, 
edge of fin deeply concave. Base of soft anal 3 in depth of body. 

Pectorals 3-6-3-8 in length of body, tip reaches almost below origin of soft dorsal. 

Ventrals 4-0-4-5 in length of body, apex reaches, or almost reaches, base of first anal 
spine. 

Caudal moderately forked, peduncle 1-1-1-2 times as long as deep. Least depth of 
peduncle 3-0-3-2 in depth of body, 6-6 in length of body. 

Scales ctenoid, somewhat larger on posterior part of body, generally larger below 
lateral line than above. Lateral line gently curved, lateral rows 73-78, tubules about 
65, 28-30 above lateral line obliquely back from base of first dorsal spine. About 50-60 
predorsal, and about 25-30 on cheek to preopercle edge. Whole body and head, except 
muzzle, scaly. Median fins densely scaly, on soft dorsal and anal posteriorly scaling 

‘scarcely extends more than half length of membrane. Scaly sheath of spinous dorsal 
continued along basal portion of alternate spines. Pectorals scaly three-fourths. 
Ventrals with outer rays only scaly. 

Colour.—Dark brown to black with or without. 8-10 darker cross bars about equal 
to eye, wider than interspace, first over nape, last on peduncle. Fins black, caudal 
dusky. Iris bronzy. 

Length.—Up to 400 mm. 

Locality.—Knysna, Port Alfred, Great Fish Point, Cove Rock (East London), Transkei 


Coast. 


/ 


Type from Great Fish Point (Pl. XIV) in the Albany Museum. Seven 
specimens from 200 mm. up examined. 

This species is admittedly very close to capensis and was separated only 
after long and detailed study. It is possible that specimens may ultimately 
be obtained which will show complete transition between the two forms 
here recognised. 

Nevertheless the form of falcatus is characteristic of the coast of the 
south-eastern Cape, and after brief study there is no difficulty in distinguish- 
ing it from capensis. All the fins are markedly longer, even the spinous 
dorsal, the mouth is slightly larger, and always extends to the eye, while 
the shape of the snout is characteristic, being more rounded. Further, 
this species is more often taken with cross bars than capensis, and the silvery 
colour phase occasionally found in that species is apparently very un- 


144 


274 Transactions of the Royal Society of South Africa. 


common, or unknown, in falcatus, most specimens being very dark in 
colour. 

It appeared at one time as if sexual dimorphism might account for the 
difference between capensis and falcatus, since all my specimens of the 
former appeared to be males, and most of the latter, with the exception of 
one or two too immature or decomposed for positive determination, 
females. Nevertheless, this hypothesis is scarcely tenable, since to my 
knowledge the form of falcatus does not occur near the Cape Peninsula, 
while capensis grows to a larger size than that species, which would be 
remarkable if capensis proved to be the typical form of the male. 

The habits and environment of falcatus are very similar to those of 
capensis, except that the former appears to prefer less turbulent water, 
being generally taken from deeper water in gullies or on reefs. The chief 
food appears to be the common mussel, together with occasional small 
crustacea and seaweeds. 

In some respects falcatus appears to fall midway between capensis and 
multifasciatus, and has probably contributed to the non-recognition of 
the latter. It is well differentiated from multifasciatus as very brief study 
shows. | 

Ripe females of falcatus are taken on the coast of the Eastern Province 
chiefly during August. The eggs are pelagic, very numerous, and less 
than 1 mm. in diameter. 


Dichistius multifasciatus (Pell.). 


(Plates XV and XVI.) 


“Kolayeleetye”; ‘“‘Ntombiyeleetye.”’ (Native names.) 

1914. Pellegrin, Bull. Soc. Zool. Fr., vol. xxxix, p. 231. 

1925. Fowler, Proc. Ac. Nat. Sci. Phil., vol. xxvii, p. 233. (Dipterodon capensis C. 
and V.) 

1927. Barnard, loc. cit., p. 635. (Dipterodon capensis C. and V., part.) 

1933. Fowler, Bull. U.S. Nat. Mus., vol. xii, p. 216. (Coracinus capensis C. and V.) 

1934. Fowler, Proc. Ac. Nat. Sci. Phil., vol. lxxxvi, p. 476, fig. xlii. - (Drepanoscorpis 
gilchristt.) 


Body deep, margins subangular, very compressed, maximum width 1-8-2-0 in head, 
3:0—3-3 in depth of body. Dorsal profile of snout fairly steep, in a more or less even curve 
from first dorsal, with very slight interorbital prominence. 

Depth 1-8-1-9, length of head 3-0-3-2 in length of body. Eye 4:0-4:3, snout 2-8-3-0, 
interorbital 3-1, postorbital part of head 1-8-1-9 in length of head. Preorbital uniformly 
deep, depth above end of maxilla 5-3-5-6 in head. The distance from the upper margin 
of the sheath at the base of the fourth dorsal spine to the nearest point of the lateral . 
line is 1-15—1-2 times the postorbital part of the head. | | 

Mouth moderate, terminal, slightly oblique, maxilla extends to below nostrils or to 
below anterior part of orbit, extremity not covered. Upper lip rather deep and fleshy, 


145 


Lhe “Galjoen” Fishes of South Africa. 275 


depth at snout apex 2:3-2:5 in eye. 30-34 elongate teeth in single outer series in upper 
jaw, anterior more or less truncated, posterior more slender and acute. Two series of 
similar but much smaller teeth behind the outer, concealed in fleshy pads, and very fragile. 
Similar teeth in lower jaw. Palate and tongue edentate. 

Gill-membranes united forming a moderate fold across the throat, narrowly united 
with isthmus below. Gill-rakers moderate, 5-6 + 14-15, 1-8-2-2 in gill-fllaments, which 
are ]-2—1:3 in eye. 

D X, 21-23, inserted above nearer base of pectoral than base of ventral, deeply notched 
between spinous and soft portions. Spines stout, heteracanth, fold into deep scaly 
sheath. First spine 10-15, 2nd 4-5-6, 3rd 2-6—-3-2, 4th 2-2-2-5, 5th 2-2-2-4 in head, 
thereafter graduated shorter. Soft rays anteriorly elevated forming a short blunted or 
pointed lobe not falcate, which extends beyond apex of 10th spine much less than length 
of postorbital part of head: Base of soft dorsal longer than (1-1-1-2 times) head. The 
tip of the depressed anterior rays scarcely reaches beyond the base of the last anal ray. 

A III, 13-14, inserted below the origin of the soft dorsal. Spines stout, first 5-6—6-0, 
2nd 3-3-3-5, 3rd 4:0 in head. Soft rays anteriorly elevated, slightly longer than dorsal 
rays. Lobe extends 1-9-2-1 in head beyond apex of last anal spine. Edge of fin deeply 
concave. Base of anal 2-7-3 in depth of body, greater than postorbital. 

Pectoral 3-5-3:8 in length of body. 

Ventrals 3-7—4-0 in length of body, apex usually reaches well beyond the base of the 
first anal spine. 

Caudal moderately forked, peduncle 1-2—1-3 times as long as deep. Least depth of 
peduncle 3-9-4-1 in depth of body, 7-4—7-6 in length of body. 

Scales ctenoid (Pl. IV, A) somewhat larger on posterior part of body. Lateral line 
scales largely concealed, smaller than adjoining scales, tubes simple (Pl. IV, B). Lateral 
line gently curved. Lateral rows 83-86; tubules about 70, but almost impossible to 
count with accuracy; 28-31 above lateral line, counted backwards from base of first 
dorsal spine. About 50-55 predorsal, and about 20 on cheek to edge of preopercle. 
Whole body and head scaly, except snout in advance of nostrils, anterior part of pre- 
orbitals, and chin. Median fins densely scaled. On soft dorsal and anal scaling extends 
at least 3 of length of membrane, even of posterior rays, and is very dense. Pectorals 
scaly for proximal three-fourths. Ventrals similarly between outer rays. Bases of 
alternate dorsal spines scaly. | | 

Colour.—Light silvery brown, with seven darker cross bands about equal to light 
interspace, equal to or wider than eye; between each pair a much narrower stripe. First 
bar over nape, last over peduncle. Head dark. Membrane of spinous dorsal black, 
hinder margin lighter. Hind margin of caudal black. Anal spines light, soft fin dark. 
Ventral spine light, distal $ of fin dark. Pectoral dark distally. Iris bronzy. 

Length.—Up to 285 mm. 

Locality Knysna (Juvenile 55 mm.); Great Fish Point; East London to Durban; 


Madagascar (Type locality). 


The body shape of this species is characteristic (Pl. XV), and enables 
it to be identified at a glance even in preserved or dried specimens from 
which all trace of marking has faded. The heavy scaly investment of the 
median fins prevents much movement of the rays, and in those specimens 
I have examined the dorsal and anal rays have remained fully erect. 

This is a much more uniform species than capensis, and in so far as I 
am aware, unbanded specimens are never encountered. The alternate 


146 


276 Transactions of the Royal Society of South Africa. 


wide and narrow bands are very characteristic, and very vivid in the live 
fish, or in fresh specimens. | 

There can be no question of the validity of multifasciatus, and it is 
remarkable that subsequent workers should have rejected Pellegrin’s 
perfectly plain diagnosis of the differentiation from capensis. The very 
shallow peduncle and the long base of the soft dorsal are immediately 
diagnostic, even in juveniles, as are in effect also the body shape and 
markings. 

Fowler (loc. cit.) stated that Drepanoscorpis has only a single series of 
teeth. The two inner series of teeth are actually very easy to overlook, 
being concealed within fleshy pads, and the teeth are small and fragile. 
They are incisiform, and not conical in shape, as usually stated. There are 
several inconsistencies between the description and figure of galchriste (loc.cit.). 

D. multifasciatus is rarely if ever encountered west of Algoa Bay. A 
single small specimen was taken in a rock pool at Knysna. 

This species does not appear to grow as large as the other two, 300 mm. 
apparently being an outside length. 

In so far as can be ascertained, it is unusual for multifasciatus to be 
captured on lines at the same time as the other species or under equivalent 
conditions. The latter may be taken in broken water, relatively shallow, 
on sandy as well as rocky bottoms, whereas multifasciatus appears to 
prefer exclusively rocky haunts and clearer water, being generally taken 
in deeper water, often in gullies at low tide. It has been noticed that when 
this species is taken, other fishes, especially larger fishes, are generally absent. 

On the eastern coast of South Africa, where all three species may be 
found, the flesh of multifasciatus is generally regarded as superior to that 
of the others, being lighter in colour and of less pungent flavour. The 
natives of this area constantly distinguish multifasciatus from falcatus 
(known as ““Damba”’) by names indicating its fondness for rocky haunts. 
The two native names mean “Pick up a stone’’ and ‘‘ Stone-maiden.”’ 


I wish to express my gratitude to Dr. Barnard, of the South African 
Museum, for the loan of material. To numerous friends, especially J. Rex 
Meterlerkamp, Hsq., of Knysna, for valuable assistance rendered in securing 
material. Also to the South African Research Grant Board (Carnegie 
Fund) for generous financial assistance. 


The Council desires to acknowledge the receipt of a grant from The 
Research Grant Board towards the cost of publishing this paper. 


ALBANY MuSEvUM, 
GRAHAMSTOWN, 
April 1935. 


rans 


Roy. Soc. 8. Afr., Vol. XXIII. Plate XIII. 


is (C. and V.) (Juvenile). 


istius capens 


Dich 


J. L. By. Smith. Neill & Co., Lid. 


Trans. Roy. Soc. 8. Afr., Vol. XXIII. Plate XIV. 


estrus falcatus n. sp. (Type). 


chi 


2 


J. L. B. Smith. Neill & Co., Lid, 


Plate XV. 


Trans. Roy. Soc. 8. Afr., Vol. XXIII. 


Dichistius multifasciatus 


Pell 


° 


Neill & Co., Ltd. 


J. L. B. Smith. 


Trans. Roy. Soc. 8. Afr., 


J. iL. B. Smith, 


ol. XXIII. 


Plate XVI. 


Neill & Co., Ltd. 


ltifasciatus (Pell.). 


. 


wtius MU 


Scales of Dichi 


t to lateral 1 


B. 20th lateral line tubular scale. 


The line below each scale represents 1 mm. 


ine. 


jacen 


4th lateral scale adj 


9 


A. 


30 mm. total length. 


imen 2 


From spec 


Trans. Roy. Soc. 8. Afr., Vol. XXIII. Plate XVII. 


CM. 


Dichistius faleatus n. sp. 


A. Preopercle. 3B. Anterior dorsal spines. C. Mandible from behind: the line 


indicates posterior teeth. D. Middle upper pharyngeals. 
EK. Lower pharyngeals. F. Maxilla. G. Premaxilla. From specimen 285 mm. 


length. 


_ Neill d& Co., Ltd 


J. L. B. Smith, 


147. 


Records of the Albany Museum. Vol. IV. Part II. pp. 169-235. 
Pls. XVIII—XXII1. May, 1935. 


New and Little Known Fishes from South Africa. 
[With Plates XVIII—XXIII, and 5 text figures.] 
By J. L. B. SMITH. 


Family MYLIOBATIDAE. 
Myliobatis cervus n. sp. [Text fig. 1.] 
“PYL-STERT”’ (Knysna.) 

Disc about 1.75 times as wide as long. Pectoral tips moder- 
ately pointed, sub-falcate. Snout not very blunt, rounded, with 
apical point. Flanges on side of head, connecting rostrals with 
pectorals, very narrow. A circular flap of the iris projecting 
over most of the pupil from above. Males with a small conical 
horn above the orbit. Dorsal small, projects beyond hind margin 
of base, originates 3-34 lengths of base behind posterior margin 
of ventral base, 1-14 lengths of base behind end of ventrals. 
Males with two serrated caudal spines, posterior longer, females 
with one or two spines. Caudal 14-2 times as long as disc. Central 
series of teeth 4-5 times as wide as long. Skin smooth, no 
tubercles. Colour uniform brown. 

Size, up to 4 ft. across the disc (females), males usually 
much smaller. 

Localities: Cape Agulhas, Knysna estuary, Bushmans River, 
Port Alfred, Great Fish Point. 

Types (one male, one female) in the Albany Museum, from 
Knysna. 

I have examined a large number of specimens of both sexes, 
and find that the females would all be referable to aqwila Linn, 
except for the shape of the snout, which is somewhat more 
pointed in cervus than in aquila. This is apparently the only 
difference between the females. The orbital horn of the males of 
cervus is very slight. Full specific distinction of cervus from 
aquila is of perhaps doubtful validity, since the females of these 
species cannot easily be distinguished one from another. 

Cervus is distinguished from the related species tobyei BI., 
by the wider disc and by more posterior insertion of the dorsal. 
According to Garman (Mem. Mus. Comp. Zoo. vol XXXVI, p, 
431), cornuta Gnthr. is a synonym of tobyez. 


148 


170 South African Fish.—Smith. 


Aquila is more common at Knysna than cervus: both are 
present in fair numbers during the summer months. The females 
appear to seek out the shallow banks in the river, where the 
young are born. One large female after capture gave birth to 
seven young, which were 200-230 mm. across the disc. These 
“rays” are evidently preyed upon by various species of “sharks.” 
On one occasion a large specimen of aquila was observed to take 
refuge in a few inches of water on the edge of a bank in the 
river, violently beating the water with the pectorals in an effort 
to scare off a large “shark” which lay close in, with part of the 
body out of the water. The specimen of aguila was secured and 
proved to be an adult female, 3 young being born after capture. 
On another occasion a large female of aquila was taken on a 
line, and proving too large to lift into the boat, was tied to the 
bows. Several young were born and drifted down with the tide. 
Shortly afterwards a large “shark,” some 9ft. in length, appeared, 
and seized one of the newly-born young. 


a 


10cm. 


Text fig. 1, Myliobatis cervus n, sp. Male. 


149 


South African Fish.—Smith. 171 


Family SYNGNATHIDAE. 


Hippocampus capensis Bler. 


1900. Boulenger, Mar. Invest. $.A. vol. i, p. II, Pl. III, fig. 2. 
1916. Thompson, Mar. Bio. Rep. vol. III, p. 90. 
1925. Barnard. Ann. S.A. Mus. vol. XXI, p. 293. 


Head (snout to gill opening) 1.5-1.9 in trunk (gill—opening 
to last body ring). Trunk 1.9-2.1 in tail. Snout, short and 
straight, 2.6-8, eye 4.5-5.5, postorbital part of head 2.1-2.4 in 
length of head. Coronet reduced, almost obsolete, a very small — 
tubercle at each end of median occipital keel. Faint radiating 
ridges on opercle. 

Rings 11+83-86. Tubercles on head and body small and 
obtuse but distinct, the supero-lateral on the first, fourth, seventh, 
tenth and last body rings enlarged. Tubercles on fourth, sixth 
and eighth tail rings usually enlarged and less obtuse than the 
remainder. Plates smooth, ridges distinct. No filaments. 

D 16-18, sub-dorsal rings 2-3+1. Ridge below dorsal about 
half eye diameter in height posteriorly. 3 enlarged sub-dorsal 
tubercles, one on the 10th, and two on.the last body ring. 

A 4; P 15-17. 

Colour: Light to dark brown, uniform, or with dark spots. 
Dorsal fin with black sub-marginal band. 

Length: Up to 110 mm. 

Locality: Knysna. 

Plesiotypes from Knysna in the Albany Museum. 

As indicated by Barnard (loc. cit.), Boulenger (loc. cit.) 
states that there are 10 body rings, whereas his figure shows 11. 
A large number of specimens, ranging from 33-110 mm. in 
length, has been examined, and the number of body-rings is 
constantly 11. The original description is also inaccurate in 
other particulars, notably in stating that tubercles are not present 
on the head and body, whereas the figure shows them to be 
present. 

This species is rather scarce even in the Knysna river, which 
is the only locality from which it has been recorded, Specimens 


150 


172 South African Fish.—Smith. 


have several times been found in the stomach of the “Dasje.” 
(Diplodus capensis A. Smith.) 

Capensis is closely related to kuda Blkr, but the rudimentary 
coronet and the very short snout distinguish it sharply from 
the latter species. 


Family CLUPEIDAE. 
Sardinella melanura Cuv. 


1888. Day, Fishes of India, p. 636, Pl. CLXIV, fig. 5 
(atricauda) . 
1925. Fowler, Proc. Ac. Nat. Sci. Phil., vol. LXXVII, p. 194. 


Body compressed, belly cultrate, with scutes. Depth 3.7-4, 
length of head 4.2-4.3 in length of body. Eye 3.5, interorbital 
4.0, snout 3.8, and postorbital 2.4 in length of head. Interorbital 
concave, grading forward into an obtuse mesethmoidal ridge. 
Posteriorly converging parietal ridges. Arborescent muciferous 
canal system on cheek. Adipose eyelids well developed, both 
extending to pupil. Mouth very oblique, lower jaw projects, 
maxilla extends to below anterior margin of pupil. 

No teeth visible in jaws, or on vomer. Traces of fine teeth 
on palatines, and clearly visible on tongue. Gill rakers 37-40, 2 
in gill filaments, which are slightly less than eye. Branchios- 
tegals 6. 

D 17-18, inserted 1.4 times further from caudal base than 
snout tip, midway between snout tip and middle of anal. Longest 
ray, 3rd, 1.6 in head, decreasing posteriorly, last ray longer than 
penultimate. Base of dorsal 1.7 in head. 

A 18-21, inserted 3.3 times further from snout tip than 
caudal base, longest, anterior, rays 5 in head. Last ray much 
longer and thicker than penultimate. Base of anal 1.5 in head. 
P. 15, 1.4 in head, inserted low down, below the hinder third of 
the sub-opercle. V. 8, 2.1 in head, inserted below the 3rd dorsal 
ray. Caudal deeply forked. 

Scales cycloid, not deciduous, 1]. 44-48, Itr. 10-11; 16-17 
preventral, 14 post-ventral scutes. Head naked. Dorsal and ana! 
with heavy scaly basal sheath. 


151 


South African Fish.—Smith. 173 


| Colour: Dark green-black above, silvery below. Interorbital, 

snout and tip of lower jaw dark. Dorsal and caudal dusky. Re- 
maining fins light. 

Length: Up to 160 mm. 

Locality: Durban, taken in the surf. 

Distribution: East coast of Africa to the Indo-Pacific. 

These specimens described above appear to be conspecific 
with melanura, except that the dorsal appears to be rather far 
forward. | 

This is the commonest Clupéid at Durban, at any rate in the 
winter months. 


Family OPHICHTHYIDAE. 
Ophichthys marginatus Peters. 


1866. Gunther. Fish. Zanz. p. 128 (Ophiurus m.). 
1870. Gunther. Cat. Fish. Brit. Mus. viii, p. 64 
(Ophichthys m.). 


Body cylindrical. Head small, 6.3 in body from gill-opening 
to vent, 18 in total length. Cleft of mouth extends behind eye, 
3 in length of head. Eye small, 2.5 in snout. Tail 1.7 in total 
length, 1.7 times body from gill-opening to vent. Anterior nasal 
tube longer than eye (not very short). Snout blunt and moder- 
ately swollen at extremity, lower jaw much shorter than upper. 
Teeth moderate, conical, pointed, biserial in both jaws. 

Pectorals short, 3.1 in head. The dorsal commences above 
the middle of the pectoral and is very low, being received into a 
deep groove. Anal similar to dorsal. 

Skin on head deeply corrugated longitudinally. 

Colour: Uniform light brown, head darker. 

Length: 430 mm. 

Locality: Knysna. 

Distribution: East coast of Africa. 

This specimen differs from the descriptions of marginatus, 
notably in the greater length of the tail, and in the not markedly 
short anterior nasal tube. It is at present provisionally assigned 
to marginatus, with the diagnosis of which it agrees closely. 


152 


174 South African Fish.—Smith. 


It is singular that this species has not previously been dis- 
covered in the waters of Natal. Its presence as far south as 
Knysna is interesting and confirms the view expressed elsewhere 
that the Indo-pacific species extend more generally westwards 
along our coast than has hitherto been suspected. 

Family MYCTOPHIDAE. 
Myctophum (Diaphus) elucens Brauer. [Text fig. 2.] 

1906, Brauer, Tiefsee Exp. vol. XV, p. 219, fig. 140. 

1913, Gilbert, Mem. Carneg. Mus. VI, p. 98, Pl. XII, fig. 2. 

(gigas). 

1928, Parr. Bull. Bing. Ocean. Coll. III, p. 121 and p. 188 ff. 

Depth 4.3-5.3, length of head 3.4-3.8 in length of body. Eye 
4, snout, 6.4-7, interorbital width 3.3, and postorbital 1.6 in length 
of head. 

Mouth large, extends well behind eye, cleft of mouth 1.4 in 
head. Maxilla scarcely dilated posteriorly. Gill rakers 6+12-18, 
longest 2 in eye. 

D 14-15, originates 1.38 times further from caudal base than 
from tip of snout. 

A 15, originates below end of dorsal. 

P 12, 2.38 in head, does not reach ventral base. 

V 10, inserted below dorsal origin, 1.8 in head, inner rays 
_ longer, scarcely reaches vent. 

Antorbital organs very large. Lower antorbital continuous 
above and below nostrils, occupying most of snout, narrowly 
separated in front along the mesethmoidal ridge. Upper antor- 
bital smaller, rounded, situated immediately above the orbit, 
separated from the lower organs by a moderately wide black 
septum. 

Photophores with a curved black septum. PLO twice as 
far from lateral line as from pectoral base, with a luminous 
scale below. PVO 2, both below pectoral base, form an oblique 
straight series with first PO. PO 5, 4th elevated above the rest 
to level of pectoral base. VLO midway between lateral line and 
ventral base, slightly behind ventral origin. VO 5, 2nd and 3rd 
elevated above the rest, 3rd highest. SAO 8, form an oblique, 


153 


South African Fish.—Smith. 175 


equally spaced series with last VO, highest three times as far 
from ventral profile as from lateral line. Pol. 2.3 times as far 
from anal base as from lateral line. AO 6+5; antero-AO form 
a gently curved series, convex below, first and last at same level: 
postero AO straight. Pre 4, widely separated from AO, form a 
curved series, last high up, twice as far from ventral profile as 
from lateral line. | 

Scales: 1.1. 36, no luminous scales except a single scale below 
PLO. Scales mostly shed. 

Colour (preserved) uniform brown. Antorbitals bright 
silvery. 

Two specimens, 145 and 165 mm. in length, precise locality 
unknown, but probably Natal, having been found among unclassi- 
fied fishes from Natal presented to the Albany Museum by the 
Director of the Fisheries Survey. | 


ee oe we ww ee 
o « « . 

co 6 wee, ar . 

° ee & © © © © ee wo es a 

-2 @ e 


Text fig. 2. Myctophum elucens Brauer. 


These specimens are in fairly good condition except for the 
absence of most of the scales. 

Parr (loc. cit) appears to be fully justified in uniting gigas 
with elucens. , 

The specimens described above are clearly conspecific, differ- 
ing in minor features, such as the distinctly smaller eye. 

This species has been recorded from just south of the 
equator on the East Coast of Africa, so that its presence in the 
waters of Natal is to be expected. With further collecting it may 
be discovered on our South Coast. 


154 


176 South African Fish.—Smith. 


Family ALEPISAURIDAE. 
Alepisaurus ferox Lowe. 


1895. Goode and Bean, Ocean. Ichth. p. 117, fig. 142. 
1925. Barnard, Ann. S.A. Mus. vol. XXI, p. 250, Pl. X, fig 2. 


The only specimen hitherto recorded from South Africa was 
from East London, described by Barnard (loc. cit.) 

Another specimen, 900 mm. in length, has since been cast up 
at the Mbotyi river mouth (Transkei) and has been forwarded 
(dried and shrunken) to the Albany Museum by Mr. W. W. 
Roberts. : 

This specimen is undoubtedly conspecific with ferox, as it 
agrees with the general diagnosis in all particulars which can 
be determined with certainty; D 40, A 17, P 14, V 9. 

Eye 23 in snout: head 6 in body. Gill rakers about 20 (prob- 
ably a few more) fine, bifid. 

Barnard’s opinion (loc. cit.) that there is only one species 
in this genus is probably correct. It may be noted that the finder 
reports that this specimen did not have the characteristic en- 
larged upper caudal lobe (both now broken). This is not 
important, as the feature may prove to be of secondary sexual 
significance only. 


Family CORYPHAENOIDIDAE. 
Lionurus nasutus n. sp. 


Body compressed. Snout pointed, flattened above and below, 
projects almost an eye diameter beyond vertically above the tip 
of the upper jaw. : 

Depth 7.4, length of head 4.9 in total length. Eye 3.3, 
longitudinal length of orbit 3, of snout and interorbital 3.2, and 
postorbital 2.5 in length of head. Upper jaw length 3.1 in head. 
Barbel 1.4 in eye. Mouth sub-lateral, maxilla extends to below 
posterior margin of pupil. Minute pluriserial conical teeth in 
upper jaw. Similar teeth in 2-3 series at symphysis, in a single 
row on side of mandibles. Vomer, palatines and tongue edentate. 

Gill-openings wide, rakers 138, tubercular. First gill-slit 
reduced by a membranous fold. Branchiostegals 7. 


155 


South African Fish.—Smuith. 177 


Vent in scaleless fossa slightly behind insertion of ventrals. 

D 2, 9+ many, originates above the hind margin of the oper- 
culum, Ist spine minute, 2nd 1.5 in head, serrated along entire 
front edge. First ray almost as long as 2nd spine. 

A ca. 140, originates below the distal part of the pectoral, 
anterior rays slightly longer than posterior, graduated. P. 17, 
1.7 in head, V, 8, 2.5 in head. 

Body scales with 6-7 rows of spinules, posterior spine longest. 
(Fig. 3.) Cycloid scales on shoulder girdle. Scales on head with 
spiny tubercles arranged in quincunx. Six series of scales between 
dorsal origin and lateral line. 

Colour: (preserved) Uniform light brown. 

Length: 243 mm. 

Type, a single specimen in the Albany Museum. 

Locality: Precise locality unknown, having been found among 
some unlabelled specimens, collected by S. S. Pickle, presented 
to the Albany Museum by the Director of the Fisheries Survey. 

This species differs from the other South African species of 
Lionurus: the snout is much more elongated, and the gill-rakers 
are greater in number. 

I am unable to identify this specimen with any species 
described in the literature at my disposal, but as there are 
numerous nominal species, nasutus may eventually be found to 
be identical with some known form. | 


Text fig. 3. To show the arrangement of the spinules on 
a body scale of Lionurus nasutus n. sp. x 15. 


156 


178 South African Fish.—Smith. 


Family ATHERINIDAE. 
Iso natalensis Rgn. [Pl]. XIX, fig. C.] 


1919, Regan, Ann. Durb. Mus. vol. II, p. 200, fig. 3. 
1925, Barnard, Ann. S.A. Museum, vol. XXI, p. 300. 


Body very compressed, maximum width at shoulder 2.9 in 
depth, tapers ventrally. Ventral edge cultrate. Breast sharply 
keeled: a median cultrate thickened fold of skin along belly. 
Dorsal profile even, gently convex from snout to caudal, with 
slight concavity above eye. Snout bluntly rounded. Body tapers 
rapidly behind ventrals. Depth 3.4, length of head 5 in length of 
body. Eye 3.1 in head, equal to snout, 1.2 in interorbital width 
and in the postorbital part of the head. 

Mouth very oblique, almost vertical, lower jaw shorter, 
included in upper, edge of upper jaw curved. Maxilla, excluded 
from margin of upper jaw, extends to below posterior third of 
snout. Minute curved conical teeth in a single row in each 
jaw. Similar teeth on vomer and palatines. A large cutaneous 
process in roof and in floor of mouth across symphysis. Lips 
moderate, finely papillose. Opercular bones entire. Gill-openings 
wide, membranes free. Gill-rakers 11, slender, almost as long as 
eye, 1.5 in gill-membranes. No slit behind last gill. Branchioste- 
gals 5. Pseudo-branchiae present. 

D VI+1, 15. First dorsal originates slightly in advance of 
midway between tip of snout and base of caudal, well behind mid- 
way between ventral and anal origins. Spines very slender, first 
longest, 3.5 in head, remainder decrease, 6th spine very short, 
membrane scarcely incised. Second dorsal originates 1.7 times 
as far from tip of snout as from base of caudal, above the 7th 
anal ray; spine subequal to eye. Anterior rays slightly elevated, 
2.2 in head, edge of fin gently concave, last and penultimate rays 
slightly longer than preceding rays. Base of first dorsal 2.6, of 
second dorsal 1.1 in head. 

A 1, 26. originates midway between caudal base and hind 
margin of eye. Spine short. Anterior rays slightly elevated, first 
longest, 1.9 in head. Edge of fin gently concave. Base of anal 1.3 
times head. 


157 


South African Fish.—Smith. 179 


P 18, 1.4 in head, upper margin of base close to dorsal profile. 

V.1, 5, 1.9 in head, inserted slightly more than twice as far 
from caudal base as from tip of snout. 

Caudal 8+18+8, forked, peduncle twice as long as deep. 

Scales cycloid, small, no basal striae. Scaling extends to base 
of anal and obliquely up as far as midway between tip of 
pectoral and origin of first dorsal. Lat. ser. 74, 6 from origin 
of first dorsal to lateral stripe, and 6 series across lateral stripe. 
Hight series across peduncle. Head, belly and nape scaleless. 

Colour. (Alive). Light blue-green above, silvery white be- 
low. A bright silvery lateral stripe with dark upper margin, 1.3 
times as wide as eye, from pectoral base to peduncle, tapers 
posteriorly from second dorsal. A silvery blotch on caudal base. 
Opercular bones dull silvery. Nape dusky. Fins light. 

Length: 73 mm. 

Locality: Knysna. 

Distribution: South and East coasts of Africa. 

Plesiotype in the Albany Museum. 

A single specimen, taken in the surf at Knysna Heads. 

In the summer months at Knysna, a cold current, of tempera- 
ture 50°F., frequently sets inshore, numbing innumerable fishes. 
This small specimen did not appear to be inconvenienced by this 
cold water, since it was able to swim at a remarkably rapid rate, 
and was captured with great difficulty. A larger specimen, at 
least 100 mm. in length, was also observed, but it evaded all 
attempts at capture. 

The above diagnosis differs in several particulars from that 
of the holotype, chiefly in the presence of 6 spines in the first 
dorsal and in the greater number of anal rays, as well as in the 
shape and size of the mouth. Since the original description is 
somewhat meagre I submitted the Knysna specimen to Mr. 
Norman, of the British Museum, who compared it with the type 
of natalensis. He is of opinion that, making allowance for the 
difference in size, the two are conspecific. The type is some- 
what damaged, and Regan’s figure (loc. cit.) is not quite accurate. 


158 


180 South African Fish.—Smith. 


Family GADIDAE. 
Merluccius capensis Cast. 


A juvenile specimen of this species has been received from 
Mr. A. M. Vardy, which was taken some half mile up the Gamtoos 
river. Also Mr. Vardy states that a large adult specimen was 
taken in the same river a few years ago. In view of the normally 
bathybial habit of this species, the occurrence in a relatively 
shallow tidal river is remarkable, all the more so since the waters 
of this river normally carry a considerable amount of suspended 
earthy matter. 


Family CETOMIMIDAE. 
Cetomimus picklet Gilch. 


1922, Gilchrist, Fish. Mar. Surv. Spec. Rep. III, p. 56, Pl. TX, 
fig. 1 (Pelecinomimus picklei). 


Body moderately compressed. Depth, behind head 7, at 
origin of dorsal 7, midway along body 9, in length of body. Head 
3.3 in body. Snout, about equal to interorbital width, 2.4 in head. 
Eye rudimentary, just above maxilla. Mouth very large, cleft 
1.2 in head. Maxilla, excluded from margin of upper jaw, dilated 
posteriorly. Minute compressed triangular teeth in 3-7 series in 
lower and 3-5 series in upper jaw, bands tapering posteriorly, 
symphysis edentate in both jaws. Similar teeth on vomer, 
palatines, pterygoids, and on all gill-arches. Tongue rudiment- 
ary. Gill-opening wide, gill-filaments few, about 7 in head. No 
gill-rakers. Branchiostegals 8. Caudal damaged, peduncle fairly 
compressed. Lateral line a single wide tube with about 10 large 
pores between opercle and caudal base. Large pores on head. 
Suborbital produced into a bifid spine which projects laterally 
above the hind end of the maxilla. Preopercle with a flat spine at 
angle just behind angle of mouth. 

The skin on the body is very loose and flabby, scaleless, nor 
are any traces of scale pockets to be observed. 

D 16, commences 3.3 times as far from snout as from base 


of caudal, posterior rays longest, 2.4 in head. Length of base of 
dorsal 1.3: in base of anal. 


159 


South African Fish.—Smith. 181 


A 16, commences slightly in advance of the dorsal, shape 
similar to dorsal. 

P 18, inserted low down on side, 2.3 in head. 

Colour: Uniform black. | 

Length: 66 mm. 

The above description is based upon a re-examination of the 
type (and only known) specimen of mcklei. The original descrip- 
tion is very brief and misleading. Gilchrist (loc. cit.) states that 
scales are present, ‘‘very long scales,” and mentions this feature 
as a characteristic of the genus Pelecinomimus Glch. As indicated 
above, there is absolutely no trace of any scales, and it is difficult 
to account for Gilchrist’s postive statement to the contrary. Both 
the skin and the lateral line are similar to those described for 
other members of this family. Further, the body does not taper 
uniformly from the head, but is narrowest midway, after the © 
manner of that of regani (Zugm.), but the median construction 
is not so pronounced in this case. The size and position of the 
eye are, in these of all fishes, scarcely features of generic signi- 
ficance, and there appears at present to be no valid reason for the 
maintenance of Pelecinomimus as distinct from the type genus. 

It may be indicated that Barnard (Ann. S.A. Mus. 19285, vol. 
XXI, p. 252) who did not see the specimen, was doubtful of the 
accuracy of the statement about the presence of scales. 


Family EXOCOETIDAE. 
Cypselurus hewitti n. sp. 


Body moderately compressed, sub-quadrangular in section, 
1.6 times deeper than wide. Depth 5.6, length of head 4.1 in 
length of body. Eye 4 in head, 1.4 in interorbital, slightly 
greater than snout and 1.8 in the postorbital part of the head. 
Head moderately depressed, interorbital gently concave with two 
slight longitudinal grooves. Top of head scaly. Minute conical 
teeth in a single series in each jaw. Vomer and _ palatines 
edentate. Maxilla extends to below nostrils. A pair of flat fringed 
tapering barbels at symphysis of lower jaw, 2 in head, width at 
base 8-9 in head. Gill-rakers 21, 2 in gill-filaments which are 1.6 
in eye, Branchiostegals 10. 


160 


182 South African Fish.—Smith. 


D 18, arising 2.5 times as far from base of caudal as from 
base of ventral. Edge of fin highly convex. 1st ray 1.4, 2nd 1.7, 
3rd longest, 1.1, last ray 4 in head; fin scaleless. 

A. 10, commences below 6th dorsal ray. Edge of fin convex. 
Ist ray 3.5, 6th longest, 2.2, last ray 4 in head. Length of base of 
anal 1.5 in base of dorsal. Fin scaleless. 

P 14, reaches almost to base of caudal, first ray simple, 3rd 
ray longest. Base of fin scaly. 

V 6, 1.3 times head, inserted midway between base of caudal 
and hind margin of preopercle, 3rd ray longest, Ist shortest: a 
few basal scales. Caudal deeply forked, upper lobe shorter, 1.5 in 
lower. Mid-rays 1.4 in eye. 

Air-bladder simple. 

Scales: lateral rows 56. L.l. 56, 28 from chest to base of 
ventral. 32 pre-dorsal scales, 8 above lateral line. 1.1. tubes with 
5-6 postero-inferior tubules. Scales with 5 radiating basal striae. 

Colour: Brownish above, lighter below. Pectorals dark with 
upper margin light. Distal ? of ventrals dark. Dorsal and anal 
dark with light anterior margin. Caudal light. Barbels dark. 

Length: 215 mm. 

Locality: Port Alfred. 

Type in the Albany Museum. 

The present species resembles bahiensis Ranz, but differs in 
the presence of the mental barbels, in the absence of palatine 
teeth and in the larger number of scales. It may be noted that 
Weber and de Beaufort (Fishes Indo-Aus. Archip. 192, IV, p. 
190) state that bahiensis has 48 lateral rows of scales, while 
Fowler (Fishes Oceania, Mem. B. P. Bish. Mus. 1928, X, p. 84) 
gives 48-57. 


Family TRACHICHTHYIDAE. 
Hoplostethus gilchristi n. sp. [Pl. XXII, fig. E.] 


Body ovate, deep, moderately compressed. Dorsal profile 
even, snout fairly blunt. Depth 2.2, length of head 2.6 in length 
of body. Eye 4.2 in head, slightly less than snout, 1.4 in inter- 
orbital width. | : 


161 


South African Fish.—Smith. 183 


Head bony, ridged, ridges covered by thin skin, with mucifer- 
ous pores and canals. Ridges mostly minutely serrate. A strong 
spine at the angle of the preopercle. Indications of a suboper- 
cular spine. No spine on opercle. A large suprascapular spine, 
connected to the side of the nape by a large scale, which has a 
denticulate hinder margin. Nostrils circular, close together, 
posterior very large. 

Mouth large, very oblique, moderately protractile. Maxilla 
extends to well behind eye, to about below the upper angle of 
the preopercular limb. Maxilla excluded from the margin of the 
upper jaw. Supramaxilla present. Lower jaw almost included in 
upper. Villiform teeth in both jaws in bands, wider anteriorly. 
An edentate symphysial notch in upper jaw, into which fits a 
corresponding dentate knob on the upper surface of the lower 
jaw. A median tubercle on chin. Vomer edentulous. A single 
series of villiform teeth on the palatirnes, triserial in a small 
anterior patch. | 7 

D VI, 17 (or V, 18), commences over the 4th lateral line 
scale, behind the base of the pectoral. 1st spine very short, 
remainder increase to the 6th, which is about 4 in head. Soft 
rays higher than the spines: mid-posterior longest, almost 2.5 
in head. Posterior margin of fin rounded. Anterior rays in- 
distinctly articulated. 

A III, 11, commences below the base of the 10th dorsal ray. 
First two spines very short, 8rd abruptly longer, about 5 in head. 
Soft rays longer than the last spine: mid-posterior longest, about 
2.7 in length of head. Posterior margin of fin rounded. 

P 20, 1.5 in length of head, extends to above the vent. Base 
scaly, oblique. 

V 1, 6, inserted below the base of the pectoral, 2 in head, 
not reaching to vent. 

Caudal deeply forked, lobes lanceolate, VII+21+VII. Peduncle 
longer than deep. 

Seales strongly ctenoid, those of the lateral line vertically 
elongate, twice as high as wide. 33 scales in lateral line. Large 
perforation in each |.l. scale covered by a denticulate flap hinged 
anteriorly. Body scales are one third linear of 1.1. scales. About 


162 


184 South African Fish.—Smith. 


11 scales above and 23 below 1.1. Abdominal scutes 18, very 
feeble. Mentum and lower portion of preopercular limb scaly, 
rest of head naked. Gill-openings very wide, membranes free. 
Gills 4, no slit behind the 4th. Gill-rakers 14, long and fairly 
stout, 2.5 times the gill-filaments and 4.5 in head. Pseudo- 
branchiae present. Branchiostegals 8. : 

Colour: (preserved) uniform red-brown. Inside of mouth 
and gill cavity black. 

Type in the Albany Museum. 

A single specimen, 238 mm. in length, taken in 300 fms. 
off Durban by S. S. “Pickle,” was presented to the Albany 
Museum by the Director of the Govt. Fisheries Survey. 

It appears not unlikely that the late Dr. Gilchrist had recog- 
nised this specimen as differing from the commoner mediter- 
raneus (C & V), since it had evidently been separated from 
specimens of this latter species. 

Gichristi is intermediate between mediterraneus and at- 
lanticus (Coll.). It differs from the latter in the larger body 
scales as well as in other characters. It differs markedly from 
mediterraneus in the much shorter paired fins, as well as in the 
numerically greater but much weaker abdominal scutes. 

In the specimens of mediterraneus I have examined, a long, 
slender opercular spine is present, which resembles closely that 
found in Gephyroberyx darwini (Lowe). The present species has 
no opercular spine. 


Family GRAM MICOLEPIDAE. 
Xenolepidichthys dalgleishi Glch. [Pl. XVIII, fig. A.] 


1922, Gilchrist, Fish. Mar. Surv. Spec. Rep. vol. III, p. 73, 
pl. xii, fig. I. 
1925, Barnard, Ann. S.A. Mus. vol. XXI, p. 370. 


Body very compressed, deep, subrhombic, dorsal profile steep, 
almost straight from snout to dorsal origin. Depth 1.2 times 
length of body: head 2.7 in length of body. Eye 2.5 in head, 
slightly greater than snout and 1.6 times interorbital width. 


163 


South African Fish.—Smith. 185 


Supraorbital ridge serrate, terminating in a tri-dentate 
laterally projecting ridge, one third of eye diameter, above the 
hind margin of the eye. Mouth very small, terminal, scarcely 
protractile. Maxilla extends to almost below anterior margin 
of orbit. Gill-membranes united, scarcely united with isthmus. 
Gill-rakers short, 15 on lower part of anterior arch. 

D. V. 27, originates above hind margin of head. First spine 
minute, second with a double series of barbs behind, filamentous, 
_ 2.4 times length of body; 3rd spine as long as body, 4th half as 
long as the second, 5th shorter. Soft rays about 2.2 in head, 
slightly longer posteriorly. A series of short spines, being 
enlarged scales, along the base of the fin. 

A II, 27, originates below the 10th dorsal ray. First spine, 
serrulate behind and in front, very elongate, filamentous, twice 
the length of the body: second spine broken at base, probably 
also elongate. Soft rays similar to dorsal rays. Short spines 
along base of fin. 

P. 14, 1.8 in head. 

V 1.6, 2 in head. 

Caudal emarginate, upper lobe longer. 

Seales vertically elongate, 1.1. 88, strongly arched anteriorly. 

Colour: Bright silvery-yellow, with round black spots ar- 
ranged more or less in three longitudinal rows, the upper follow- 
ing the dorsal, the lower the ventral profile. Two dark bars across 
the hinder part of the peduncle. Hinder margin of caudal dusky. 
Base of anterior dorsal spines dark. 

Length: 43 mm. 

Locality: Great Fish Point. 

This specimen was cast up by the waves after a storm, and 
is in an excellent state of preservation. Specimens of this species 
are sometimes picked up on the shore in the neighbourhood of 
Port Alfred, but this is apparently the smallest specimen hitherto 
secured. The extremely elongate 2nd dorsal and Ist anal spines 
are possibly a juvenile characteristic, but may also persist in 
the adult stage, since the shorter 1st anal spine of larger fishes 
resembles the broken proximal pertion of a longer growth, 


164 


186 South African Fish.—Smith. 


Family OPISTHOGNATHIDAE. 
Opisthognathus marrostomus n. Sp. [Pl. XX,fig. B.] 


Body compressed, moderately elongate. Head large, rounded, 
interorbital concave. Snout very blunt, abruptly descending be- 
fore eyes. 

Depth 4.3, length of head 2.9 in length of body. Eye 4 in 
head, twice snout and interorbital width, 2.8 in postorbital part 
of head. Preopercle without free hind margin, hidden below skin. 
Mouth very large, oblique, cleft extends to well behind the 
posterior margin of the orbit. Maxilla enlarged, with supra- 
maxilla, extends to almost below the hinder margin of the oper- 
culum. Very slight membranous connection with cheek. Each 
jaw with a single series of fine conical teeth along the sides: 4 
series in upper and three series in lower in a narrow symphysial 
patch. Vomer and palatines edentulous. A single pair of dentiger- 
ous upper pharyngeals, teeth blunted, conical. Lower pharyngeals 
apparently fused, dentigerous. Tongue rounded, free. Gill-open- 
ing wide, membranes free from isthmus. Gills 4, a slit behind 
the fourth. Gill-rakers 25, slender, 1.4 times gill-filaments and 2 
in eye. Pseudobranchiae present. Branchiostegals 6. 

D XI, 18, commences in advance of the hind margin of 
operculum. Spines feeble, Ist 3.5 in head, remainder increase in 
length to the 7th and 8th which are 2 in head, remainder de- 
crease to the last which is 2.4 in head. Anterior rays very slight- 
ly shorter than last spine, remainder increase to the 19th which 
is 2.2 in head, thereafter decrease to the last. Posterior 8 rays 
branched. Whole fin enveloped in thin skin. 


A 16 (or II, 14), commences below the base of the last 
dorsal spine. First two rays spiniform. 1st ray less than an eye- 
diameter, remainder increase to the 14th which is 1.8 times an 
eye-diameter. Posterior 7 rays branched. Whole fin enveloped 
in thin skin. 

P 20, rounded, 2.7 in length of head. 

V I, 5, half the length of the head, inserted in advance of the 
base of the pectoral. Bases close together; fin enveloped in skin 
spine feeble and indistinct. 


165 


South African Fish.—Smith. 187 


Caudal 15, rounded, 2.0 in length of head. Peduncle deeper 
than long, very compressed. | 

Scales cycloid, very small. Lateral line commences on head 
above hind margin of preopercle, and has transverse branches 
ending in pores. These branches are arborescent on the anterior 
portion of the lateral line, becoming simple posteriorly. Lateral 
line extends below ? of the dorsal, ending below the 8th dorsal 
ray. 

Head naked, skin densely pitted, with many arborescent 
muciferous tubules. 

Colour: Brown, mottled with darker. Lower margin of, and 
a patch on hind portion of, maxilla black. Ventrals, hinder 
margin of caudal and portions of dorsal, dusky. Remaining fins 
light. A dark blotch on the dorsal between the 5th-9th spines. 

Type: A single specimen 160 mm. in length, from Natal, 
precise locality unknown, found among certain fishes collected by 
S.S. “Pickle” in Natal waters, presented by the Director of the 
Govt. Fisheries Survey to the Albany Museum. 

This species is close to nigromarginatus Ruppel. It differs 
from this species in several features: in the number of dorsal 
spines, in the wider interorbital, in the extent of the lateral line, 
and in the slightly smaller expansion of the maxilla, together | 
with the absence of a membranous connection between the cheek 
and the hinder extended portion of the maxilla. 


Family CALLIONYMIDAE. 
Synchiropus monacanthus n. sp. 


Dorsal profile flat, with an abrupt interorbital prominence, 
sloping before eyes. Body moderately compressed, sub-cylindrical. 
Head (tip of snout to gill-opening) 1.2 times as long as broad, 
fairly depressed. | 

Depth 6.7, length of head (to gill-opening) 3.6 in length of 
body. Eye 3 in head, slightly greater than snout and 6 times the 
interorbital width. Preopercular spine 5 in head, no antrorse 
spine at base, a single denticle, 1/3 length of spine, on distal third, 
directed inwards and forwards. Eyes partly lateral. 


166 


188 South African Fish.—Smith. 


Mouth moderate, fairly protrusible obliquely downwards. 
Maxilla not concealed beneath the preorbital, extends to barely 
below the anterior border of the eye. Slender pointed teeth in 
bands in both jaws, larger anteriorly, those in the upper Jaw 
directed horizontally backwards. Other bones edentulous. Tongue 
absent. 

Gill-membranes united with isthmus. Gill-opening very 
small, reduced to an ovoid foramen, almost lateral. Gill-rakers 
11, small, pointed. Anterior third of first gill-arch adnate to inner 
surface of operculum. Pseudobranchiae present. Branchiostegals 
q. 

D IV 8, commences above the gill-opening. Spines filament- 
ous, lst longest, equal to head, remainder decrease to the 4th 
which is 2 in head. Second dorsal separated from the first by a 
space equal to the length of the base of the first dorsal. Rays 
bifid, filamentous, Ist ray 3 in head, 2nd-4th slightly shorter, 
remainder increase to the last, which is twice the first. 

A 7, commences below the base of the 3rd dorsal ray. Rays 
bifid, first about 3 in head, remainder increase to the last which 
is twice the first. 

P 20, lanceolate, almost as long as head. 

V I, 5, inserted below the first dorsal, as long as head. Bases 
moderately separated by a space less than half of the length of 
the fin. Last ray joined by a membrane to the pectoral base. 

Caudal lanceolate, 1.4 times head. Peduncle 1/3 as deep as 
long. The lateral line is continuous on to the head and across 
the nape, with branches on the head. The lateral line, from above © 
the gill-opening, first curves sharply down towards the pectoral 
base, then in a gentle curve up towards the anterior dorsal rays, 
then down to the middle of the body, thereafter straight to the 
base of the caudal: there are above and below, small oblique 
branches opening each by a pore. 

Colour: (Preserved) light brown, marbled and spotted with 
darker. Top of head dark. Upper ventral rays, tip of caudal and 
margins of dorsal and anal dusky. 

Type in the Albany Museum. 

A single male specimen, 148 mm. in length from Port Alfred. 


167 


South African Fish.—Smith. 189 


This species is related to lineolatus (C. & V.), from the Indo- 
Pacific, but differs in having only one denticle on the preoper- 
cular spine. 


Family SERRANIDAE. 
Serranus knysnaensis Gilch. 


1904, Gilchrist, Mar. Invest. S.A., vol. iii, p. 2, pl. XIX. 
1925, Barnard, Ann. S.A. Mus. vol. xxi, p. 461. 


Body moderately compressed. Dorsal profile flat, with slight 
prominence above eyes. Depth 3.3, length of head 2.8 in length 
of body. Eye 3.5 in head, 1.5 times interorbital width, slightly 
greater than snout, and 1.8 in the postorbital part of the head. 
Mouth large, lower jaw prominent, maxilla extends to almost 
below centre of eye. Small curved conical teeth in narrow bands 
in both jaws, on vomer and on palatines. Canines small, fewer 
and larger in lower jaw. Tongue edentate. 

Preopercle serrate, serrae not enlarged at angle. Middle 
opercular spine largest. Preorbital shallow, 3 in eye. Gill-rakers 
11, very slender, far apart, 1.5 in gill-filaments, which are 3 in 
eye. 

D X 14, commences above hind margin of operculum: Ist: 
spine shortest 6.3, 2nd 4.2, 3rd-5th subequal longest 2.4 in head. 
Remainder decrease to the last. No notch. Anterior rays almost 
1.5 times last spine. Fin scaly. 

A III 7, commences below base of last dorsal spine; 2nd spine 
as long as, but stouter than the third. 

Pectoral rounded, 1.4 in head. Ventrals 1.5 in head, inserted 
behind pectorals, reach vent. | 

Caudal emarginate, peduncle longer than deep. 

Scales stenoid. 1.1. 71, ltr. 7/19, 1.1. tubes anteriorly with 
ascending branch. L.|. follows dorsal profile. 10 scales on cheek. 
Scaling on head extends to above eye. 

Colour: (alive) Red-brown above, lighter below. Two dark 
brown longitudinal stripes, 2/3 width of eye, upper from eye just 
below lateral line along side to peduncle, thence over extending 
on to the base of the upper caudal lobe: lower from pectoral base 
to the base of lower caudal lobe. Two rust-red diagonal streaks 


168 


190 South African Fish.—Smith. 


from preorbital down across cheek. A red blotch on the lower 
opercular margin. Red bars along lower jaw. Hind margin of 
caudal red, fin spotted with red. Pectoral reddish. Dorsal and 
anal suffused with red, with longitudinal series of darker red 
spots. 

Length: 65 mm. 

A single specimen taken at Knysna. | 

It is singular that no specimens of this species have been 
reported since its discovery in 1904 to the present, and the fish 
is not known to the oldest angling residents at Knysna. The 
absence of markings on the type is not remarkable, since the 
brilliant colouration of the live fish has practically disappeared 
with preservation, though the longitudinal stripes are still clearly 
visible. 


Epinephelus flavocaeruleus Lac. [Pl. XXII, fig. A.] 


1878, Day Fishes of India, p. 15, pl. iii, fig. I. 
1925, Barnard, Ann. S.A. Museum, vol. XXI, p. 475. 


Body compressed, dorsal profile, from snout to base of dorsal, 
undulate, convex above eyes. Depth 2.4, length of head 2.6 in 
length of body. Eye 4.3 in head, equal to snout and interorbital 
width. Preopercle margin serrate, serrae enlarged at angle, one 
large spine with 2 smaller below. Lower opercular spine further 
back than upper. Mouth oblique lower jaw projects, maxilla 
extends to below hind margin of pupil. Villiform teeth in 
posteriorly tapering bands in each jaw, canines small. A single 
series of teeth on side of mandible. Gill-rakers 15, longest as 
long as gill-filaments, which are 2.2 in eye. 

D XI, 17, arises over hind margin of operculum, not notched. 
1st spine, shortest, 5, 2nd 3.2, 3rd 3.0, 4th longest, 2.8 in length 
of head: remainder decrease slightly. Anterior rays longer than 
last spine, 3rd ray longest, 2.6 in head. Fin scaly. 

A III 8, arises below base of 5th dorsal ray. 2nd and 38rd 
spines sub-equal, 2nd stoutest. 

Pectorals rounded, 1.6 in head. Ventrals 1.8 in head, reach to 
vent. 

Caudal truncate. 


a7 -I4 


169 


South African Fish.—Smith. 191 


Scales on body ctenoid, cycloid on head and nape. 1.). 73 
lr. 187, 25 rows above ll. | 

Colour: Body dull-blue from snout posteriorly to above end 
of base of anal, behind this bright yellow. Anterior portion of 
dorsal base dark blue, remainder of dorsal, whole of anal, of 
pectorals and all of ventrals, excepting dark distal third of 1st 
ray, bright yellow. 

Length: 116 mm. 

Locality: Port Alfred 

Distribution: Indo-Pacific. 

A juvenile specimen of this species, a rare migrant to the 
more southerly portions of the South African coast. The colour- 
pattern is most striking and vivid, with the bright canary-yellow 
peduncle and caudal. 


Family HOPLEGNATHIDAE. 
Hoplegnathus robinsoni Ren. [Pl]. XVIII, fig. C.] 


1916, Regan, Ann. Durb. Mus. vol. i, pt. 8, p. 168. 
1925, Barnard, Ann S.A. Mus. vol. XXI, p. 506. 


Body ovate, very compressed, width 3.5-4 in depth. Depth 1.8-2, 
length of head 3 in length of body. Eye 3-4.2 in head, 1.1-1.3 in 
interorbital width. | 

Mouth terminal, moderate, lower jaw projects slightly. In 
the young the teeth are incisiform, closely set in a single row 
in each jaw. With growth the teeth become fused, with develop- 
ment of a flattened inner flange on the roof and floor of the mouth, 
bearing marginal tubercles. No other teeth. Maxilla extends to 
below the anterior border of the eye. Anterior nostril circular 
with a plain flap. Concealed spine on posterior margin of opercle, 
hind margin serrate in young. Preopercle m-yeune-serrate, serrae 
enlarged at angle. Gill-membranes more or less united, free from 
isthmus. Gill-rakers 16-1% slender, 2 in eye. | 

D XI, 21-22, originates above pectoral base, first spine 
shortest 8-10 in head, remainder increase rapidly to the 4th, 
and thereafter very gradually, last spine longest. The anterior 
soft rays are elevated, becoming sub-falcate in the adult, decrease 


170 


192 South African Fish.—Smith. 


rapidly after the 7th, which in the adult is about as long as head. 
The base of the soft dorsal is slightly longer than that of the 
spinous portion of the fin. _ 

A III, 14-16, third spine strongest. Anterior soft rays 
longest. Shape of fin resembles that of the dorsal. 

P. 15-16, 1.8-2 in head, gently rounded. 

V I, 5, without axillary process, 1.2-1.5 in head, reaches 
beyond origin of anal. 

Caudal emarginate, peduncle as deep as long, about 3 in head. 
Seales strongly ctenoid, l.r. about 110, 30 above 1.1. 1.1. tubules 
82-90. Interorbital naked except for a few scales along anterior 
border of eye. | 

Colour: (alive) Yellow, with 5 anteriorly convex black cross- 
bars, Ist through eye, 2nd from anterior dorsal spines through 
opercle and base of pectoral extending over the ventral; 3rd 
broadest, on middle of body extending on to dorsal and anal, 4th 
on hinder part of body, 5th narrow, on peduncle at base of caudal. 
Pectorals light. Caudal with dark margin, very marked in the 
adult. 

Length: 32-185 mm. 

Locality: Algoa Bay, Knysna, Natal. 

This species has been known only from the type. Regan (loc. 
cit.) states that the interorbital of the type is scaly and that the 
middle dorsal spines are the longest. In my specimens the 
interorbital is scaleless, except for a few scales along the upper 
margin of the orbit, and the posterior dorsal spines are longest. 

The genus Scarostoma (Kner) differs from the type genus 
in having the interorbital scaly and the spinous dorsal not more 
than 1.4 times as long as the soft portion of the fin. Robinson 
thus bridges the gap between these genera, and it would appear 
scarcely necessary to maintain Scarostoma as distinct. 

Specimens of this species are rarely encountered and, as I 
have observed, anglers who secure any would be inclined to regard 
them as slightly abnormal specimens of Diplodus cervinus Lowe, 
which the present species resembles closely in shape and coloura- 
tion. | 


171 


South African Fish.—Smith. 193 


Family CEPOLIDAE. 
Acanthocepola cuneatus n. sp. [Pl. XXI, fig. A.] 


Body very compressed, elongate, band-like, tapering poster- 
iorly. Greatest depth across pectoral base. Profile gently sloping 
from nape, snout somewhat blunt and swollen. Muciferous pores 
and canals on head. Depth 6, length of head 6.7 in length of body. 
Eye 3.4 in head, 1.4 times snout and 1.1 times the interorbital 
width. 

Six preopercular spines: largest spine at angle, 4 smaller 
on lower limb, anterior three below skin, one small spine above 
the angle below skin. Mouth moderate, very oblique, lower jaw 
projects. Maxilla excluded from upper jaw, extends to below the 
posterior third of the eye. Small conical teeth in a single row in 
the upper jaw. A single series of similar teeth on the sides of 
the lower jaw, with 2 short series of smaller teeth on a pair 
of symphysial knobs, which latter fit into a corresponding recess 
in the upper jaw. Three pairs of dentigerous upper pharyngeals. 
Tongue rounded, free. 

Nostrils double. Vent far forward, immediately before the 
base of the anal, below the base of the pectoral. 

Gill-opening wide, gill-membranes free from isthmus. Gills 4, 
a slit behind the 4th. Gill-rakers 30, long and slender, slightly 
longer than gill-filaments, about 5 in head. Branchiostegals 
6; pseuodbranchiae present. 

D 67, commences above the middle of the opercle. Ist ray 
shortest, about 5 in head, remainder increase to the middle, 
thereafter decrease somewhat in length. Middle rays about 1.6 
in head. Rays indistinctly articulated. 

A 74, commences below the base of the 6th dorsal ray. 
Anterior rays increase to the 12th which is about 2.3 in head, 
thereafter sub-equal. Articulations indistinct. Dorsal and anal 
confluent with caudal. 

P 20, 1.7 in head, rays branched. A thickened fold of skin 
from the axil of the pectoral to the upper angle of the operculum. 

V I 5, inserted slightly in advance of below the hinder 
margin of the preopercle, 1.4 in head. 1st ray filamentous, 


172 


194 South African Fish.—Smith. 


reaches to the base of the 6th anal ray. Ventral bases united, 
last ray joined to the body by a membrane. 

Scales cycloid, minute. Free margin of scales forming a 
right angled point. Scales on hinder portion of body slightly 
larger than those on anterior part. Lateral series about 285. 
Lateral line commences at the shoulder and runs obliquely up 
to the base of the 4th dorsal ray, thereafter along the base of the 
dorsal, finally curving abruptly down on to the base of the caudal. 
The lateral line scales are of curious shape, anteriorly ridged. 
(Pl. XXI, fig. A). A large scale with arborescent muciferous 
tubules above the upper angle of the operculum. Head naked, 
except cheeks, interopercle, and a small patch on upper anterior 
margin of opercle. Limb of preopercle naked. Mentum naked, 
transversely rugose. Five median scales before the base of the 
dorsal. 

Colour: Red-brown. Fins light. Dark oval blotch between 
the 8th-14th dorsal rays. 

Type: A single specimen, 262 mm. in length, from Natal, 
precise locality unknown, the specimen being among certain 
fishes obtained by the 8.8. “Pickle,” labelled “Natal fishes, un- 
classified,’ donated by the Government Fisheries Survey to the 
Albany Museum, 1931. 

Barnard (Ann. S.A. Mus.., 1925, vol. XXI, p. 503) included 
Acanthocepola limbata (C. & V) in the South African fauna list 
on the examination of a dried specimen from Delagoa Bay. It 
is not unlikely that this (fide Barnard, since lost) was identical 
with that here described, since both possess the blotch on the 
anterior dorsal spines. It must, however, be noted that Barnard 
(loc. cit.) admits indica Day to the synonymy of limbata, but 
states that the latter has no lateral line, whereas Day (Fishes of 
India, 1878, p. 796) states that indica has a lateral line on the 
anterior third of the body. The lateral line of the present species 
is certainly obscure, but nevertheless quite definite. More careful 
observation may show that the other species of this genus also 
possess complete lateral lines. 

Cuneatus is very much deeper anteriorly, and the body tapers 
more rapidly, than is the case with the other species of the 


173 


South African Fish.—Smith. 195 


family. This feature is so marked, that should the other species 
of Acanthocepola Blkr. be found to have incomplete lateral lines, 
generic distinction for cuneatus would appear justifiable. 


Family LUTIANIDAE. 
Caesio algoae n. sp. 


Body compressed. Dorsal profile even, sharply concave over 
eye. Interorbital gently convex. Depth 2.7, length of head 3.5 in 
length of body. Eye 3.2 in head, 1.3 times snout and equal to the 
interorbital width. Posterior nostril slit-like. Preorbital depth 3.3 
in eye. Maxilla narrow, almost concealed by preorbital, does not 
extend below anterior border of orbit. Small curved conical 
teeth in a single series in each jaw, none on vomer, palatines or 
tongue. A lanceolate space between the rami of the lower jaw, 
densely scaled to apex at symphysis. Head, excepting interorbital, 
snout and sides of lower jaw, densely scaled. Naked interorbital 
and snout densely pitted with minute pores. None of opercular 
bones serrate. 

Branchiostegals 6, 4 scales on membrane between the 2nd 
and 8rd (anterior) rays, remainder of membrane naked. Gill- 
rakers 19, slender, 1.5 in gill-filaments which are 2 in eye. 
Pseudobranchiae present. Pyloric caeca 4. 

D XI, 18, commences over the ventral base. Ist spine about 
5 in head, 4th and 5th longest 2.5 in head, remainder decrease. 
Anterior rays slightly longer than last spine, decrease pos- 
teriorly. Membrane of soft dorsal basally scaly. 

A III, 12, commences below the base of the 2nd dorsal ray, 
Ist spine short, 2nd stout, sub-equal in length to the 3rd which is 
2.5 in head. Anterior rays slightly longer than 8rd spine, there- 
after decrease slightly. Membrane of soft rays basally scaly. 

P 17, falciform, as long as head. Base scaly. Ventrals in- 
serted behind pectoral base, do not reach to vent. Caudal deeply 
emarginate, base densely scaled. 

Seales strongly ctenoid, 1.1. 72, l.tr. 9/16. 11 series on cheek. 

Colour: Rosy above, silvery below. Occiput dusky. Black 
spot in axil of pectoral. Margin of dorsal dusky, remaining fins 
light. 


174 


196 South African Fish—Smith. 


A single mature male specimen, 200 mm. in length, from off 
Algoa Bay in 60. fms. 

Type in the Albany Museum. 

This specimen is very close to axillaris Blgr., from which 
it differs in having an extra ray in the dorsal and anal fins, in 
the greater number of lateral line scales and in the much nar- 
rower preorbital. 

A revision of this genus will probably show that nominal 
species exist; wider collection may show the present species to 
fall within the limits of variation of axillaris. 


Family CARANGIDAE. 
Decapterus lajang Blkr. [Pl. XXI, fig. D.] 


Body elongate, moderately compressed, width 1.3 in depth. 
Depth 6.8, length of head 3.8 in length of body. Eye 3.6 in head, 
1.1 in snout, 1.5 in postorbital part of head and slightly greater 
than interorbital width. Snout sub-conical. Dorsal profile even. 
Interorbital gently convex, scaleless, with a medio-longitudinal 
ridge parallel with which on each side are several fine longitudinal 
grooves. Preorbital, depth almost two thirds of an eye-diameter, 
with radiating muciferous tubules, branching distally. Preopercle 
margin even, with radiating grooves and tubules. Opercular 
membrane not serrate. Mouth oblique, lower jaw projects. 
Maxilla expanded posteriorly, hind margin concave, extends to 
below the anterior margin of the orbit. Fine teeth in a single 
series in the lower jaw, in a longitudinal bi- or tri-serial band on 
tongue, and in a patch on each side of vomer. Palatines with a 
few minute rudimentary teeth in a single widely spaced series. 
No teeth in upper jaw. Adipose eyelids well developed, anterior 
and posterior united above and below, leaving an oval median 
opening almost as wide as pupil. Gill-rakers 36, slender, 1.2 in 
gill filaments, which are 1.5 in eye. 

D I+ VITII+1, 37+1, originates twice as far from base of caudal 
as from tip of snout. Ist spine minute, procumbent, directed 
forwards. 3rd spine of second dorsal longest, 2.1 in head, last 
spine very short. Anterior portion of soft dorsal elevated, falcate, 
2nd ray longest, 2.5 in head: remainder decrease to the 8th, which 


175 


South African Fish.—Smith. 197 


is 5.5 in head, thereafter sub-equal in length. The detached finlet 
is 4.6 in head, reaches beyond caudal base. Fin ‘scaleless, with 
low scaly sheath, flanking fairly deep groove. 

A II +1, 28+1, originates below the 6th dorsal ray. First two | 
spines separate, sub-equal, about 9 in head. Second anal similar 
in shape to soft dorsal, 2nd ray longest, 3 in head. Finlet as 
long as dorsal finlet, opposite which it is inserted. 

P 22, falcate, 1.5 in head, shorter than head without snout. 

‘V.1.5, 2.3 in head, bases narrowly separated, inserted below 
hind margin of pectoral base, Ist ray longest. Caudal forked, 
peduncle moderately depressed. 

Scales very small, lateral line scales enlarged. Lateral line 
gently undulate, curves down from shoulder, up to below origin 
of soft dorsal, down to below 15th dorsal ray, thereafter straight. 
94 scales from shoulder to below the 24th dorsal ray, then follow 
29 spinose scutes which increase in size to the 14th—20th largest, 
height 2.5 in eye, thereafter decrease. 12 scales with tubules on 
caudal base. The tubes on the anterior (non-spinose) scales have 
a recurved branch above and below (Pl. 1Va). 9 series of scales 
above lateral line below origin of dorsal. 2 series on cheek. 
Opercle naked. 

Colour: Blue-green above, silvery below. A dusky spot on 
hind margin of operculum. Axil of pectoral black. Fins light. 

Length: Up to 170 mm. 

Locality: Knysna, (Durban). 

Distribution: Indo-pacific. 

This specimen is provisionally assigned to lajang Blkr., from 
most descriptions of which, however, it differs in certain parti- 
culars, notably in the more elongate body, in the relatively larger 
eye and in the shorter pectoral. The very feeble dentition of the 
fishes of this genus appears to be the feature upon which 
differentiation is mainly based. The palatine teeth in this 
specimen are so exceedingly feeble that their presence may be 
ascertained only with the greatest difficulty. It is not improbable 
that this specimen may prove to be the juvenile of macrosoma 
Blkr., to which lajang is very closely related. 


176 


198 South African Fish.—Smith. 


The presence of any species of this genus as far South as 
Knysna is interesting. None of the netters to whom I showed 
the specimen could recollect having seen any other. 

I have since found juveniles of this species very plentiful 
at Durban. 

Family CHAETODONTIDAE. 
Chaetodon marleyi Regn. [Pl. XXII, fig. C.] 


1921, Regan, Ann. Durb. Mus. III, pt. 1, p. 1. 
1925, Fowler, Proc. Ac. Nat. Sci. Phil. LX XVII, p. 251. 
1925, Barnard, Ann. S.A. Mus. XXI, p. 618. 


Body ovate, very compressed. Dorsal profile steep from nape, 
slightly convex at nape, concave from occiput to snout. Depth 
1.4-1.6, length of head 2.4-2.6 in length of body. Eye equal to 
snout, to postorbital part of head and to interorbital width, 3 in 
length of head. 

Snout pointed, sub-conical. Mouth very small, maxilla ex- 
tends to below anterior nostril. 

D XI, 23-25, Ist spine almost antrorse, slightly shorter 
than eye. Remainder increase to the 4th and remain sub-equal. 
Soft rays anteriorly equal to last spines; gradually decrease. 

A III 17-19, inserted below the base of the last dorsal spine. 

Pectorals equal to ventrals 1.3-1.4 in head. Ventrals extend 
to or beyond vent. 


Seales strongly ctenoid, 1.1. 41-48, I.tr. The  scale- 


7 
rows above the lateral line run back and up. “On the first 
third below the lateral line the rows run back and up. Those 
on the mid-posterior part of the body only are horizontal, the 
rows on the lower part of the body run back and slightly down. 
Five cheek scales. Whole of head and body scaly. Dorsal and anal 
densely scaled. 

Colour: (Alive) Orange yellow, with scale centres slightly 
darker accentuating the scale rows. A dark median bar from nape 
to snout through interorbital. A dark curved band from nape 
through eye to isthmus, less than eye. A broader band from the 
2nd-4th dorsal spines through pectoral base extending oyer 
ventrals and a similar band from the origin of the soft dorsal 


177 


South African Fish.—Smith. — 199 


to the anterior part of the soft anal. A dark bar across peduncle, 
and a lighter bar across the base of the caudal rays. A white- 
edged black ocellus on the 3rd-10th dorsal rays, about the size of 
the eye. Pectorals and unbarred part of the dorsal, anal and 
caudal, yellow. Ventrals with outer part black, inner yellow. 

Length: 26-60 mm. 

Locality: Knysna, Algoa Bay, Kariega and Kowie Rivers, 
Port St. Johns, Durban. | 

I have examined a number of specimens, but in none are 
any but a few mid-posterior portions of the scale-rows below the 
lateral-line horizontal. Nor have I seen any in which the ocellus 
covers the last dorsal spine, as stated by Barnard (loc. cit.) 

This is the only South African species with 11 dorsal spines, 
a number which appears remarkably constant. This count, to- 
gether with the brilliant and characteristic markings, enable this 
species to be distinguished at a glance. 

It is singular to find a species of this genus commonly 
occurring so far south. It has been stated previously that at 
Knysna, in summer, it is not uncommon for very cold water to 
appear suddenly. The presence of marleyi in the river is an 
infallible sign that such an onset is imminent, and when the cold 
water actually enters the river, dead specimens of this species 
are usually thrown ashore. 


Family RHACHICENTRIDAE. 


Rhachycentrum canadus Linn. 


“Sergeant fish’ (Port Elizabeth). ‘Prodigal Son” (Natal). 
1920, Robinson, Rep. Nat. Fish. 1919, p. 48 (Hlacate nigra). 
1925, Barnard, Ann. 8.A. Mus. vol. XXI, p. 511. 


Body elongate, very slightly compressed or sub-cylindrical. 
Dorsal profile even, almost horizontal. Head depressed, snout 
sharp. | 

Depth 6.5-7, length of head 4.7-5 in length of body. Eye 
about 7 in head, 2.5-3 in snout and 8-3.5 in interorbita] width. 


178 


200 South African Fish.—Smith. 


Interorbital slightly convex, skin rugose. Mouth large, 
terminal, slightly oblique, apparently fairly protractile. Villiform 
teeth in broad bands on jaws, vomer and palatines. Maxilla 
extends to below the anterior border of the eye. Lower jaw 
projects somewhat. Nostrils paired, close together, slightly nearer 
anterior margin of orbit than tip of snout. 

D VIII+1, 31-32. Spines short and stout, sub-equal, 10 in 
length of head, widely spaced, connected to the body by a mem- 
brane, and depressible each in a separate groove. Anterior rays 
elevated, sub-falcate; 4th-6th rays longest, % of the length of 
the head. Remainder decrease rapidly to the 9th and thereafter 
gradually to the last, which is about 1/3 of the length of the 4th. 

A II, 24-28. Spines small and fairly stout, enveloped in skin. 
Anterior soft rays elevated, sub-falcate, 2nd-4th longest, about 
half the length of the head. Shape of soft fin similar to that of 
the soft dorsal. | 

P 20-21, sub-falcate, slightly shorter than head. 

V I, 5, less than half the length of the pectoral; inserted 
below the middle of the base of the pectoral. 

Caudal deeply emarginate, almost lunate. Peduncle moder- 
ately slender, 14 times as long as deep. 

Scales: Very small. Lateral line single, almost straight. 
Cheek scaly. 

Colour: Dark above, lighter below. Fins dark. 

Length: Up to 1115 mm. 

Locality: Algoa Bay. 

The above description is based upon three stuffed specimens 
in the Port Elizabeth Museum. 

Barnard (loc. cit.) queries the validity of Robinson’s record 
(loc. cit.) of the above species in Natal waters, which query he 
bases upon the possibility of confusion of the generic names 
Elacate and Elagatis. Further, Elagatis bipinnulatus (Q & G) 
is stated by Barnard to have the colloquial name “Prodigal Son.” 

Apart from this, it does not appear likely that the present 
species could be confused with E. bipinnulatus (Q & G), since 
the two differ markedly in shape and in external features. 

This species is an occasional capture in Algoa Bay. 


179 


South African Fish.—Smith. | 201 


Family PLECTORHYNCHIDAE. 
Pomadasys operculare (Plyfr.) 


“Knoorhaan” (Port Beaufort to Keurbooms River), “Tiger” 
(Eastern Province), “Spotted grunter’ (East London 
to Natal). 

1866, Playfair & Gunther, Fish. Zanz. p. 24, Pl. 24, fig. I. 

1888, Day, Fish India, p. 76. 

1925, Fowler, Proc. Acad. Nat. Sci. Philad. vol. LXXVII, 
p. 231. | 

1925, Barnard, Ann. 8. African Mus. vol. XXI, p. 675. 


Body moderately compressed. Dorsal profile even, concave 
above eyes, with a nuchal prominence, increasing with age. 

Depth 3.3-3.6, length of head 3-3.1 in length of body. Eye 
6.9-7.2 in length of head, 2.8-3 in snout and 1.8-2 in interorbital 
width. Postorbital part of the head 2.2 in length of head. Snout 
sub-conical, slightly depressed. 

Mouth small, terminal, slightly oblique. Maxilla extends to 
below midway between the tip of the snout and the anterior 
border of the eye. Maxilla almost concealed beneath preorbital. 
Villiform teeth in bands in both jaws; other bones edentulous. 

Rudimentary, but distinct, adipose eyelids, more evident in 
larger examples. Preorbital naked, almost twice as long as deep. 
Nostrils close together just before the orbit, anterior oval, 
posterior circular or ovoid. 

Whole of head, except preorbital, chin, and snout in advance 
of anterior nostril, scaly. 

Preopercle denticulate, with two large flat blunted projections 
at the angle: hinder margin of limb concave. No obvious opercular 
spines. An enlarged scale above the opercle on the shoulder. 
Gill-rakers 15-16, moderate, 3 in length of the gill-filaments, 
which are sub-equal to eye. Branchiostegals, 7. 

D. X-XI, 13-16. (commonest count XI, 14) deeply notched, 
commences vertically above or slightly behind the base of the 
pectoral. Spines depressible in a deep groove. First spine short, 
about half an eye diameter. Second twice the first, third and 
fourth sub-equal about 3 in head. Remaining spines decrease to 


180 


202 South African Fish.—Smith. 


the last. The anterior soft rays are somewhat elevated, the first 
and second longest, sub-equal in length to the sixth dorsal spine. 
Remainder decrease in length. 

A III, 8-10. (commonest count III, 9.) commences below 
the base of the third dorsal ray. Fin depressible in a deep groove. 
First spine very short and stout; second longest and very stout, 
third spine slightly shorter and much less stout than second. 
First soft ray longest, fifth shortest, the edge of the fin being 
slightly concave. 

P 18-19, falcate, about equal to the head, reaches to above 
the base of the anal. Basal membrane slightly scaly. Large 
rounded scaly process in axil. V I, 5, inserted below the dorsal 
origin, 1.9 in length of head, membrane scaly. 

Caudal deeply emarginate, upper lobe usually slightly the 
larger, densely scaled; peduncle longer than deep. 

Scales: ctenoid on body, cycloid on head. 1.]. 51-61, gently 
curved, extends on to the caudal. l.-tr. a About ten series 
on the cheek, and seven series on the operculum. Two series of 
scales on sheath below dorsal spines, four below soft dorsal and 
anal. 

Colour: (Alive) Bright silvery, slightly bronzy above. About 
190 dark spots on side above the level of the base of the pectoral, 
those below the lateral line occasionally in undulating rows. A 
black axillary spot. Dark blotch on hinder margin of operculum. 
A black triangular mark before the base of each dorsal ray: a 
similar mark before the apex of each dorsal spine. Membrane 
of ventral and anterior membrane of anal dusky. Caudal light, 
with a narrow dark posterior margin. Remaining fins light. 

Length: Up to 770 mm. 

Locality: All tidal rivers from Breede River, Port Beaufort 
to Natal. 

The above description is based upon the examination of 
over a hundred fresh adult specimens, taken in various rivers, 
and ranging from 400-770 mm. in length. Of these, one only had 
D X, 16, and was evidently abnormal. 

Previous descriptions appear to have been based upon im- 
mature specimens, and to have been none too accurate. 


181 


South African Fish.—Smith. 203 


Barnard (loc. cit.) does not accept records of this species 
from anywhere east of East London, and suggests confusion 
with bennetti (Lows). Operculare is abundant in all tidal rivers 
as far west as the Knysna and the Brak Rivers, and is well known 
in the tidal portion of the Breede River, Port Beaufort. 

Operculare Plyfr. is closely related to P. swillum (C & V), 
from the tropical west coast of Africa. The two species have 
hitherto been maintained as distinct on the difference between 
the dorsal and anal fin formulae, and on the scale counts. The 
former species is now shown to fall within the limits for the 
variation in the fin formulae of the latter. In regard to scale 
counts, unless the precise limits are stated, it is possible, in the 
case of operculare at least, to record wide variation. The lateral 
line extends on to the caudal, as many as seven scales bearing 
tubules being found beyond the base of the fin. Further, in the 
transverse series, three different counts are possible, and may 
be taken: 1. series from in advance of the dorsal fin; 2. series 
from below the spinous dorsal including the sheath scales; 3. the 
same count not including the sheath scales, which are very small. 

For suillum (C & V), the usual count as given is I.tr. 
iT while Pellegrin (Poiss, eux d ouc. ]’Afr. oce. p. 257) 


gives ai This latter count was most likely taken after the 


manner of 3. above. 

I have not been able to secure a specimen of swillum, which is 
not represented in collections in South Africa, nor in that of the 
British Museum. Nevertheless, the evidence here adduced renders 
it most unlikely that operculare Plyfr. can be maintained as 
distinct from suillum (C & V), as previously suspected by Day 
(loc. cit.). 

P. operculare ranks high among the estuarine game fishes of 
South Africa. It is extremely vigorous, and being rather shy 
and wily, is a much-prized capture. Specimens weighing 16 lbs. 
have been recorded, though any above 10 lbs. are rare. 

It appears to feed chiefly on crustacea, as well as upon the 
small creatures which inhabit the mud and sand banks of es- 
tuaries. These latter the “Tiger” is believed to secure by forcing 


182 


204 South African Fish.—Smith. 


its pig-like snout into the mud, and blowing them from their 
holes. The tails of these fishes may frequently be seen in estuar- 
ies waving above the surface of the water on shallow banks. 
This habit is shared by Pagellus lithognathus (C & V), the 
“White steenbras.” 

The name “Grunter,” applied to most of the species of this 
genus, arises from the sounds which most of them emit after 
capture. It is generally stated that these sounds are produced 
by contractions of the air-bladder. In operculare and bennett 
Lowe at least this is not the case. The sounds are produced by 
the rasping together of the upper and lower dentigerous 
pharyngeals. 

Day (loc. cit.) states that the flesh of these fishes is not 
highly esteemed in India. This is remarkable, since in South 
Africa operculare is rightly considered one of the most delicate 
food-fishes. 

Family SPARIDAE. 
Pagrus nasutus Cast. {Pl. XXIII, fig. B.] 

Body compressed, width 3 in depth. Dorsal profile from 
nape slightly concave above eye, with a moderate preorbital 
prominence. 

Depth 2.2, length of head 2.7 in length of body. Eye 3.6 in 
head, 1.3 in snout, 1.6 in postorbital part of head, equal to 
interorbital width and 1.4 times depth of preorbital. Preorbital 
longer than deep. Posterior nostril oval, almost as large as 
anterior. Mouth fairly large, moderately oblique, maxilla extends 
to below the anterior margin of the eye. 

Four canines in upper, 6 in lower jaw, the smaller teeth in 
the sides of the jaws variable in size. Gill-rakers 11, short and 
stout. 

D XII, 10, notched between spinous and soft portions. First 
spine 6, 4th longest, 2.5, last spine 3.3 in head. Soft rays higher 
than spines, edge of soft dorsal rounded. The whole fin may be 
contained in a deep groove with a scaly sheath. 

A III, 8, inserted below the first dorsal ray. Second spine 
longest and stoutest. Soft rays higher than spines, edge of fin 
rounded. Fin with deep scaly sheath. 


183 


South African Fish.—Smith. 205 


Pectoral almost as long as head, base scaly. 

Ventrals 1.6 in head, reach to vent: 

Caudal forked, base densely scaled. 

Scales ctenoid; those above the lateral-line much smaller 
than those below. 1.1. 61, ltr. 10/20. Limb of preopercle scaly. 
14 series on cheek. 10 series on sheath below soft dorsal. Scaling 
on head extends to above the posterior nostril. 

Colour: Red-brown,. with irregular lighter nebulous patches, 
largest over nape. Ventrals, axil of pectoral, and anterior soft | 
rays of dorsal and anal black. Snout and nape dusky. Iris bronzy. 

Length: 140 mm. 

Locality: Port Alfred. 

It is with difficulty that this species may be recognised in so 
small a specimen, since there are numerous differences from the 
familiar form of the adult. The dentition alone is misleading, 
for the lateral molars are acute, and do not differ appreciably 
in form from the lateral teeth of some species of Dentex Cuv. 
In fact, for some time I considered this specimen to fall in that 
genus, but somewhat larger examples recently secured leave no 
doubt of its identity. | 

It may be noted that Fowler (Bull. U.S. Nat. Mus., 1983, 
vol. 12, p. 145 ff.) has united under Sparus Linn. the species of 
Sparus and those placed by Barnard (loc. cit. p. 692-704) in 
Pagrus Cuv. There are numerous objections to Fowler’s diag- 
nosis, and in my opinion these two groups are not congeneric. 


Family AMPHIPRIONIDAE. 
Dascyllus axillaris n. sp. [Pl. XXII, fig. D.] 


Dorsal profile even, with a moderately prominent inter- 
orbital bulge. Body ovate deep, compressed. 

Depth 1.6, length of head 3.4 in length of body. Eye 2.3 in 
head, 1.2 times interorbital width, and 1.8 times snout. Pre- 
orbital 3 in eye diameter. | 

Maxilla extends to below the anterior third of the eye, and is 
almost concealed beneath the preorbital. Mouth small, oblique, 
moderately protractile. Villiform teeth in bands in both jaws, 
the outer row enlarged, conical; other bones edentulous. Opercle 


184 


206 South African Fish.—Smith. 


not serrate: two flat spines on upper hinder marign, the lower 
more acute. Preopercle, sub- and pre-orbitals serrate. Suborbital 
with free lower margin. Preorbital slightly longer than deep. 

Gill-rakers 13 on lower limb of the anterior arch, slender, 
2 in length of the gill-filaments, which are 4.5 in the length of 
the head. Branchiostegals 5, concealed. 

D XII, 16, commences above the hind margin of the oper- 
culum. 1st spine short, 2.3 in head. 2nd and 3rd spines longest, 
sub-equal, about 1.2 in head. Remainder decrease to the last, 
which is 1.5 in head. The first soft ray is slightly longer than the 
last dorsal spine. The rays increase in length to the fifth and 
sixth, and thereafter decrease rapidly to the last, which is 
shorter than the first dorsal spine. The fin is scarcely notched, 
and the soft rays have a highly convex margin. Both spinous 
and soft dorsal scaly. 

A II, 14, commences below the base of the last dorsal spine. 
First spine short, sub-equal to but stouter than the first dorsal 
spine. Second spine three times the length of the first. Soft 
rays longer than the spines, the shape of the fin resembling 
that of the soft dorsal. Anal rays densely scaled. 

P 20, sub-equal to head, basal portion lightly scaled. 

V I, 5, inserted below the anterior margin of the base of the 
pectoral, reach to base of the Ist anal spine. The ventrals have a 
pointed axillary scale and an enlarged pointed scale between the 
bases. Caudal sub-truncate or very slightly emarginate, basal 
portion scaly. 

Seales: ctenoid, large. Lat. series 28,  l.tr. ts. Two 
lateral lines, tubules 12, The upper lateral line follows the 
dorsal profile and ceases below the base of the 6th dorsal ray. 
The lower lateral line, which runs along the middle of the 
peduncle, consists of 5 tubules and several rudimentary canals 
on bordering scales. The whole of the head is scaly. The scales 
on the snout, interorbital and nape are small. 3 or 4 vertical 
series of scales on cheek, 5 horizontal series on opercle. 

Colour: Uniform red-brown. Ventrals and anal spines black. 
Margin of soft dorsal and anal, apices of dorsal interspinous 


185 


South African Fish.—Smith. 207 


membrane and the whole of the caudal, dusky. Pectorals light. 
A black spot in the axil of the pectoral. 

A single specimen 87 mm. in length, from Great Fish Point. 

Type in the Albany Museum. 

The South African species of this genus are not well known, 
and the three hitherto admitted are distinguished chiefly by the 
colouration. 

The present species differs from these in this character, 
chiefly in the presence of the marked axillary spot, as well as in 
the scale counts, fin formulae and shape of the anterior portion 
of the dorsal fin. The scaly membrane of the spinous dorsal is 
most likely a generic feature, but does not appear to have been 
mentioned in descriptions of species of this genus. 

Since specimens are but rarely seen, and most of the descrip- 
tions based on but a single specimen, it is not unlikely that a 
revision of adequate material would result in a very severe 
limitation of the number of species in this genus. 


Family SCOMBRIDAE. 
Neothunnus itosibi J. & E. [Text fig. 4.] 


1926, Jordan & Evermann, Occ. Papers. Cal. Ac. Sci. XII, 
p. 22, pl. 6. 


Text fig. 4. Neothunnus itosibi J. & E. 


Body fusiform, scarcely compressed, as wide as deep. Depth 
5.0, length of head 3.9 in length of body. Eye 12 in head, 4.6 in 
snout and in interorbital width, 7 in the postorbital part of the 
head. Maxilla extends to below the anterior margin of the eye. 


186 


208 South African Fish.—Smith. 


D XV, 1449. originates behind the base of the pectoral. 
Anterior spines highest, 1.8 in head, decrease rapidly to the last 
which is about 1/6 of the length of the first. Anterior soft rays 
very elongate, filamentous, 2.4 in length of body, reach to caudal! 
base. 9 detached finlets small, sub-equal. Interspinous membrane 
scarcely incised. 

A II, 1249, originates below the 14th dorsal ray, shape 
similar to that of dorsal, filamentous rays 2.2 in length of body, 
reach beyond caudal base. Finlets 9. 

P 382, sub-falcate, reaches almost to base of soft dorsal 
(probably longer, but tips broken in specimen). 

Ventrals inserted below dorsal origin. 

Entire body scaly, scaling forming a corselet anteriorly. 
Lateral line gently curved anteriorly, following dorsal profile. 

Caudal lunate, peduncle very slender, with a single large 
lateral keel. 

“Colour: Silvery brown below, darker above. Head and spin- 
ous dorsal dark. 

Lengh: 1650 mm. 

Locality: Algoa Bay. 

Distribution: Pacific (Japan). 

This remarkable specimen was taken on rod and line from 
the Port Elizabeth breakwater in June, 1932. A school of these 
fishes suddenly appeared, but only the specimen captured would 
take any of the baits offered, the effective lure being a whole 
“Elft” (Pomatomus saltator Linn). 

A very large fish was taken in a net in the Knysna lagoon 
shortly after the capture of the Algoa Bay specimen. It had 
unfortunately been cut up and eaten before I heard of it, but 
from descriptions was undoubtedly conspecific with the Algoa 
Bay specimen, and may have been a straggler from the main 
school. Those who were concerned in the capture of the Knysna 
specimen stated that it appeared to be affected in some way 
though no external injury was apparent. Another specimen was 
encountered at sea by a fishing-boat from Knysna, 

The critical internal diagnostic features of the stuffed Algoa 
Bay specimen had unfortunately not been preserved, and the 


187 


South African Fish.—Smith. 209 


aksolute identity of the specimen must to some extent remain 
uncertain, but I have provisionally assigned it to itosibi, with 
the somewhat meagre description of which the external features 
agree fairly closely, except that the latter species has 8 anal 
finlets, whereas this specimen has 9. 

Its presence in our waters is very remarkable, but is in line 
with the roving habits of the Scombridae. 

Fowler (Proc. Ac. Nat. Sci. Phil. 1933, vol. LXXXV, p. 163, 
Pl. 12) has described a new genus, Semathunnus, differentiated 
from Neothunnus by the greater length of the dorsal and anal 
lobes. This appears a rather slender basis for generic distinc- 
tion and will scarcely prove acceptable. It is not unlikely that the 
various species of Neothunnus (and of Semathunnus?) will 
ultimately all be found to be identical: one single circumtropical 
form. 

Family TEUTHIDIDAE. 
Monoceros unicornis Forsk. 


1931, Smith. Rec. Alb. Mus. IV, p. 145, fig. (Prionolepis 
hewitt.) 

The small type of P. hewitti was placed in the Grammicolepi- 
dae largely on account of the grammicolepid type of scaling. A 
slightly larger specimen, obviously conspecific, has since been 
found, in which developmental changes make it obvious that the 
above was a mal-identification, and that the specimens are most 
likely the juvenile forms of wnicornis Forsk. 

These very interesting juveniles are at present being investi- 
gated, chiefly in regard to the squamation, and full details of this 
and of the specimens will be published later. 


Family PLATACIDAE, 
Tripterodon orbis Plyfr. 
1866, Playfair and Gunther, Fish Zanz. p. 42, fig. 
1909, Gilchrist and Thompson, Ann. §.A. Mus. VI. p. 228 
(Chaetodipterus faber non Brouss.) 
1917, Gilchrist and Thompson, Ann. Durb. Mus. I, p. 367 
(ibid). | 


188 


210 South African Fish.—Smith. 


1922, Norman, Ann. Mag. Nat. Hist. (9), IX, p. 321 
(Chaetodipterus orbis Bl.). 

1925, Fowler, Proc. Ac. Nat. Sci. Phil. LXXVII, p. 242. 

1925, Barnard, Ann. S.A. Mus. XXI, p. 603 (Chaetodipterus 
goreensis non C & V): and p. 604 (Chaetodipterus 
orbis non Bloch): and p. 725 (T. Orbis.) 


The above synonymy has resulted from the examination of 
the specimens in the South African Museum: all of these are 
clearly conspecific with Tripterodon orbis, Plyfr., and Chaetodip- 
terus Lac. is thus shown not to occur in South Africa. Dr. 
Barnard, of the S.A. Museum, is in agreement with this finding. 

Norman’s record (loc. cit.) of C. orbis Bloch from South 
Africa was due to a confusion of entries in the Register of the 
British Museum. Barnard had accepted Norman’s record. 

This little-known species is of great interest, since its 
systematic position is as yet undetermined. It is at present 
provisionally placed in the Platacidae. 

A special study of this species is now in progress and the 
full text will be published later. 


Family SCOMBRIDAE. 
Scomberomorus lineolatus C & V. 
1925, Barnard, Ann. 8.A. Museum, XXI, p. 803. 


Depth slightly more than length of head. Eye 6.5, snout 
3.2, interorbital 3.2, postorbital 1.9 in length of head. Body 
moderately compressed, snout fairly pointed. Maxilla extends 
to below the hind margin of the pupil. A single row of com- 
pressed acutely triangular teeth in a single row in each jaw. 
Very fine teeth in patches on vomer, palatines and pterygoids. 
Branchiostegals 6. 

D XVI, 17+?, Spines slender, separated. 

A II, 16+?, inserted below the middle of the soft dorsal. 
Soft dorsal and anal anteriorly elevated, sub-falcate. Pectoral 
1.9, ventrals 4 in head. 

The lateral line runs slightly downwards, somewhat un- 
dulate in parts, to the middle of the side below the anterior finlets. 


189 


South African Fish.—Smith. 911 


Colour: Deep blue above, silvery below. 3-4 series of longi- 
tudinally elongated dusky spots below the dark dorsal area. 
Pectoral, soft dorsal and tip of anal dusky. 

Length: About 400 mm. 

Locality: Port Alfred. 

A damaged specimen, with the body severed behind the 3rd 
dorsal finlet, thrown up on the shore near Port Alfred. The 
remainder leaves no doubt as to the identity with lineolatus 
C& V. 

This species is fairly common, at certain seasons, in Natal 
waters where it is known as “Snoek.” 

This is the first record from this area, though it may be 
noted that the related, but larger, species commersoni Lac 1s 
well known at Algoa Bay. | 


Family BLENNIIDAE. 
Blennius cornutus (Linn.). 


1908, Gilchrist and Thompson, Ann. S.A. Mus. vol. VI, pt. 2, 
p. 1038 (seullyz). 
1916, Gilchrist, Mar. Bio. Rep., vol. III, p. 11. 
1918, Thompson (ibid), vol. IV, p. 145 (cornutus), p. 145 
(scullyz). 
1925, Barnard, Ann. §8.A. Mus. vol. XXI, p. 883 (cornutus), 
p. 836 (scullyi). 
According to the descriptions of cornutus and of scully, 
the latter species differs from the former in the presence of a 
dark spot between the 1st and 2nd dorsal spines, and in the 
shorter supraorbital tentacles. An examination of the type and 
co-types of scullyi, and a comparison of these with material 
recently collected from the Cape to Natal, show that these 
supposedly specific characters are inconstant and unreliable, 
there being complete gradation from the one form to the other. 
The dorsal spot is present in all specimens (cornutus or scullyt) 
that I have examined, but is relatively obscure in large examples, 
and tends to disappear in preserved material. There is no sharp 
line of division in the length of the supraorbital tentacles in a 


190 


212 South African Fish.—Smith. 


series of either sex. In the Blenniidae, the variable markings 
are scarcely of specific value. 


Blennius bifilum Gnthr. [Pl. XX, fig. C.] 


1861, Gunther, Cat. Fish. Brit. Mus. vol. III, pp. 225, 261. 

1917, Gilchrist and Thompson, Ann. Durb. Mus. vol. i, pt. 4, 
p. 414. 

1923, von Bonde, Fish Mar. Surv. Spec. Rep. i, p. 34, pl. IT, 
fig. I (Salarias sexfasciatus). 

1925, Barnard, Ann. S.A. Mus. vol. XXI, p. 833 (bifilum) ; 
and p. 845 (Salarias sexfasciatus). 


Body moderately compressed. Profile of head abruptly 
descending, snout very blunt. 

Depth 5, length of head 4-44 in length of body. Eye 3.2-4 in 
head, 1-14 times snout and 1.6-2.2 times interorbital width. 
A simple tapering nasal tentacle about 9 in length of head. A 
pair of slender simple tapering tentacles on the nape, about 3 
in head. No tentacles above eye. Mouth moderate, maxilla ex- 
tends to below the anterior third of eye. A single row of incisors 
in each jaw, no canines visible. Gill-membranes united forming 
a fold across the throat. 

D XI, 17-18, commences above midway between the hinder 
margins of the opercle and the preopercle, moderately to deeply 
notched (juv.). Anterior spines about 3 in head. Middle rays 
sub-equal to anterior spines. Last ray joined to peduncle at base 
of caudal. 

A II, 18, commences below the base of the last dorsal spine. 
Spines small and feeble, soft rays longer. P. 15, rounded, equal 
to head. 

V I, 2, inserted below the hinder margin of the preopercle, 
slightly less than head. Caudal lanceolate, slightly longer than 
head. 

Lateral line obscure, consisting of a single row of tubules 
on the anterior third of the body. 


Colour: Brown, with six dark wide cross-bars extending on 
to the dorsal. 


191 


South African Fish.—Smith. 213 


A dark bar across the pectoral base and a dark spot on the 
base of the middle pectoral rays. Two dark bars across the 
throat, of which the hinder bifurcates, with posterior limbs 
extending on to the ventral bases. Two dark marks on opercle 
and a black spot below and behind eye. Caudal spotted, other fins 
light. 

Three specimens, 30 mm. in length up, from St. Lucia Lake, 
presented by W. Bell-Marley, Esq. 

Distribution: East Coast of Africa. 

The markings on the smallest specimen (figured) are very 
vivid, the gular bars being especially marked. These cross- 
bars and the blotches on the pectoral are much fainter in the 
larger specimens, and have disappeared on preservation in alcohol. 

I have examined specimens of bifilum in the S.A. Museum 
and find distinct traces of these markings, which do not appear 
to have been noticed previously. 

I have not been able to examine the type of S. sexfasciatus 
vB., but the somewhat meagre description and the figure (loc. cit.) 
are in close agreement with the diagnosis of bifilum, and with the 
specimens of this latter species that I have examined. The pearly 
spots on the body of sexfasciatus are also present in some speci- 
mens of bifilum, and there is one in the S.A. Museum with the 
spots corresponding with those shown in the figure of 
sexfasciatus. 

It may be indicated that the young of various species of 
Blennius (Linn.) have somewhat movable teeth, especially those 
in the lower jaw, but never to the same extent as the teeth of 
species of Salarias. Barnard (loc. cit., p. 845) was evidently 
doubtful of the validity of sexfasciatus. 


Petroscirtes tapeinosoma Blkr. [Pl]. XXI, fig. B.] 


1877, Gunther, Mus. Godeffroy, Journ. vol. 6, pt. II, p. 195, 
pl. 115, fig. D. 

1906, Seale, B.P. Bishop Mus. Occ. Papers vol. 4, no I, p. 89. 

1928, Fowler, Fishes Oceania, Mem B.P. Bish. Mus. vol X, 


p. 430. 


192 


214 South African Fish.—Smith. 


Body elongate, compressed. Depth 8.5, length of head 5.6 
in length of body. Eye 3.1 in head, slightly greater than snout 
and than interorbital, 1.6 in postorbital part of the head. Snout 
pointed, dorsal profile gently sloping, dorsally depressed, projects 
half eye diameter beyond anterior margin of upper jaw. Mouth 
inferior, cleft extends to almost below centre of eye; teeth in a 
single immovable series in each jaw, a very large curved 
canine on each side in lower jaw. No tentacles. Gill-openings 
reduced, about half eye diameter. 

D 46, commences slightly in advance of gill-opening, anterior 
rays shortest, increase posteriorly last rays shorter, fin ends 
near caudal base. Longest rays 1.5 times eye. 

A 31, commences below the 17th dorsal ray, ends near caudal 
base. Longest rays 1.8 times eye. 

P. 12, 1.8 in head. 

V 3, thoracic, reduced, middle ray longest, 3 in head. 

Caudal forked. 

Colour: (Alive) Indigo-blue above, light blue below. A fine 
light blue longitudinal stripe from hind margin of eye, following 
dorsal profile. A wide dark blue lateral stripe, from snout and 
head, divided by 16 narrow light blue cross-bars tapering on 
peduncle and continued on the middle of the caudal. Head to lower 
edge of pupil indigo-blue; a narrow light blue stripe from lower 
margin of snout to pectoral base, bordered below by a fine darker 
line. Mouth, lower part of snout, throat, ventrals and caudal 
bright orange. Anal blue-black with light margin. Dorsal] with 
broad blue-black marginal band. Pectoral yellow. 

A single specimen, 80 mm. in length, from Great Fish Point, 
thrown up by the waves, together with innumerable live speci- 
mens of Champsodon capensis Rgn., during a storm. 

Distribution: Red Sea, Indo-Pacific. 


The colour pattern of the live fish is exceedingly beautiful 
and vivid. 

This is the first record of this species from South Africa. 
It is remarkable that it has not previously been discovered in 
the tropical portion of our East Coast region, 


193 


South African Fish.—Smith. 215 


Family GoBIIDAE. 


Psammogobius n.g. 


Differs from Gobius Art. only in the very wide and unre- 
stricted gill-opening, the membranes being completely free from 
the isthmus. The ventrals are elongate, reaching to the base of 
the anal or beyond. Genotype knysnaensis n. sp. 

In view of the prolific suggested sub-genera of Gobius, the 
introduction of a new genus closely allied to Gobius may not 
appear very desirable, but the character of the gill-opening in 
Psammogobius merits the institution of the genus. 


Psammogobius knysnaensis n. sp. 


Body cylindrical, scarcely compressed posteriorly. Head not 
very broad, scarcely depressed. Depth 5-5.5, length of head 3.3- 
3.5 in length of body. Eye 8.3-8.6 in head, and sub-equal to snout. 
Interorbital very narrow, eyes almost contiguous, dorso-lateral. 

Mouth moderate, oblique, lower jaw projects strongly. 
Maxilla extends to below the anterior margin of the eye. Teeth 
fine, pointed, conical, recurved and depressible, in 3-4 series in 
each jaw, inner series largest. No large canines. Tongue bilobed, 
free. 

Muciferous pores on head; 2-3 horizontal series of papillae 
on cheek. Anterior nostril tubular. Gill-opening wide, mem- 
branes completely free from isthmus. Gill-rakers short and stout, 
anterior club-shaped, 7-8 on the lower part of the anterior arch. 
Branchiostegals 4. Pseudobranchiae reduced, not obvious. No 
shoulder flaps. 

D VI+I, 9, commences vertically above the middle of the 
pectoral, the two fins clearly separated. The spinous dorsal is 
not elevated, the anterior spines are about 3/4 of the depth of 
the body. Spines scarcely project beyond the membrane. The 
soft rays are slightly higher than the anterior spines. 

A J, 10. Rays about half of the depth of the body. 

P 17-18, rounded, no free rays, almost 3/4 of the length of 


the head. 


194 


216 South African Fish.—Smith. 


Ventrals not adnate to belly, very long, reaching to or beyond 
the origin of the anal. 


Caudal rounded, about 2/3 of the length of the head. 


Seales ctenoid. Lat. series 29-31. trans. series 9-10. Scaling 
on head extends to vertically above midway between the hinder 
margins of the opercle and preopercle. 10-12 median series of 
scales before the base of the dorsal. Head otherwise naked. 
A small patch of cycloid scales on the throat immediately before 
the base of the ventrals, one large median scale, the others 
smaller, scarcely visible, obscured by a thick coat of mucus. 


Colour: Grey-brown, mottled and streaked. 5 irregular cross- 
bars. 6-7 vertical fine white cross-bars on the lower part of the 
side (rarely absent). Dorsal and anal with rows of spots. In 
some males a bright green ocellus between the 4th-6th dorsal 
spines. 

Types from Knysna in the Albany Museum. 

Length: Up to 61 mm. 


The above description is based upon a number of specimens 
ranging from 25-61 mm. in length, collected in the tidal portions 
of the Breede, Knysna, Keurbooms, Bushmans and Kowie Rivers. 


This species lives on the sand banks in these rivers. The 
colouration harmonises well with the surroundings, and the 
species is able to bury itself in the sand with extraordinary 
rapidity. On the sand banks, these small fishes advance with the 
tide, resting in the fringes of the tiny ripples; when startled, 
they dart away a short distance, and easily sink below the surface 
of the sand, when they may be secured by scooping up the sand. 
Specimens are frequently left stranded by the tide, but appear to 
suffer no inconvenience, remaining buried in the moist sand until 
the water returns. 


The white transverse streaks found on the majority of 
specimens are very characteristic, and apart from the obvious 
character of the unrestricted gill-opening, serve to distinguish 
at a glance this species from the numerous species of Gobius Art. 


195 


South African Fish.—Smith. 217 


Family CLINIDAE. 
Clinus heterodon C. & V. [Pl. XXII, fig. F.] 
1908, Gilchrist and Thompson, Ann S.A. Mus. vol. VI, p. 186 
(graminis). : 

1925, Barnard, Ann. S.A. Mus. vol. XXI, p. 863. 

Body fairly compressed, deepest at vent. Depth 3.4-3.6, 
length of head 4 in length of body. Eye equal to snout 5.6-6, 
interorbital 7-7.5, and postorbital 1.7 in length of head. Snout 
moderately pointed, sub-conical. Maxilla extends to below the 
anterior margin of the orbit or slightly further. A single row 
of fairly stout blunt conical teeth in the outer margin of each 
jaw. A double patch of smaller similar teeth behind the outer 
teeth at the front of each jaw. A single chevron-shaped row 
across vomer. Gill-membranes united, forming a pronounced fold 
across the throat. No tentacle over the eye. 

D XXXIII, 5 (or III+XXX, 5) inserted above the hind 
margin of the preopercle. First 3 spines elevated forming a low 
crest, 2nd spine longest 2.3 in head. 4th spine shortest equal to 
eye. Remainder increase gradually to the ante-penultimate. 
Penultimate and last spine somewhat abruptly graduated longer, 
last spine equal to second spine of crest. All spines with a single 
cirrus at apex, largest on crest. Soft rays abruptly longer than 
spines, membrane joined to peduncle. 

A II 22-23, inserted below the 18-15th dorsal spine. P 12, 
rounded, ventrals I, 3, last ray small but distinct. Caudal sub- 
truncate. | 

Scales cycloid, small, but larger than in most other species 
of this genus, clearly imbricated. Lateral line distinct to peduncle. 

Colour: Brilliant and variable. Yellowish to almost black, 
with vermiculations and mottlings in bronzy and red tints. A 
darkish wavy band on body below dorsal, traces of cross blotches 
in some specimens. 

Length: Up to 186 mm. 

Localities: Bushman’s River mouth; Great Fish Point; St. 
Lucia Bay, Zululand. 


196 


218 South African Fish—Smith. 


A somewhat rare species, only recently discovered in the 
Eastern Province. Its occurrence as far north as Zululand is 
especially remarkable. 

Venustris G & T and taurus G & T have also been obtained 
recently at Knysna, and robustus G & T, at East London. 


Clinus laurentii. G & T. 


1908, Gilchrist & Thompson, Ann. S.A. Mus., vol. vi, pt. 2, 
p. 120. 
1917, zbid Ann. Durban Mus. vol. i, pt. 4, p. 414. 
1925, Barnard, Ann. S.A. Mus., vol. xxi, p. 866 (Petraites l.). 


Body moderately compressed. Dorsal profile even, gently 
sloping to snout, slightly convex before eyes. Depth 4.2, length 
of head 4 in length of body. Eye 5 in head, 1.4 in snout and twice 
interorbital width. Snout pointed, moderately depressed. Mouth 
moderate, slightly oblique, lower jaw projects, lips thick. Maxilla 
extends to below the anterior third of eye. Villiform teeth in 
bands in both jaws, outer series enlarged. A single band of 
curved teeth on vomer. A simple flap over the anterior nostril. 
No supraorbital tentacles. Gill-membranes united forming a fold 
across the throat. Gill-rakers reduced, 5 on lower limb of the 
anterior arch. 

D III, XX VII, 5, commences slightly in advance of the hinder 
margin of the preopercle. The first three spines form a crest, 
separated by a deep notch from the rest of the fin; 4th spine 
almost free from the membrane from the 3rd spine. 1st and 2nd 
spines sub-equal, almost half of the length of the head, 3rd spine 
3.0 tn head. 4th spine shortest, 6 in head. Remaining spines 
increase to the last, which is twice as long as the 4th. Anterior 
soft rays elevated, 1st ray sub-equal in length to the 1st dorsal 
spine, remainder decrease slightly. Last ray joined by a mem- 
brane to the peduncle near the base of the caudal. 

A II, 21, commences below the base of the 15th dorsal spine. 
Spines stout, second longer. Anterior soft rays longer than the 
spines, 2.6 in head; remainder increase to the 19th, last ray 
short, not joined to the peduncle, 


197 


South African Fish.—Smith. 219 


P 12, rounded, 1.3 in length of head. V I, 3, half length of 
head inner ray small but stout. 

Caudal slightly rounded, half of the length of the head. 

Scales very small. Lateral line descends abruptly over the 
end of the pectoral, complete to peduncle, tubules simple. 

Colour: (preserved) Light brown-yellow with traces of 
darker mottling. 

A single specimen 150 mm. in length, from Port St. Johns. 

Barnard (loc. cit.) has separated laurentii G & T, together 
with argentatus Risso and mentalis G & T from the other species 
of Clinus Cuv., and has placed them in the genus Petraites Ogilby. 
This genus resembles Clinus in all respects save that the anterior 
crest of the dorsal is separate from the rest of the fin. 

In the Clinidae, the character of the highly developed dorsal 
fin has served as a basis for generic sub-division, but the partial 
application accepted by Barnard is scarcely justifiable. It may 
be pointed out that the South African species of Clinws appear 
to fall into certain well defined groups; for example, anguillaris 
C & V and striatus G & T are closely related, and differ so mark- 
edly in general features from other groups of closely allied 
species, such as supercilioysus Linn., ornatus G.& T and laurentii 
G & T, as to render generic distinction from these at least as 
sound as that of laurentii from supercilioysus. In regard to the 
latter species, I find that the further East specimens are obtained, 
the higher is the dorsal crest, and the lower, on the average, is 
the point of attachment of the membrane from the third to the 
fourth dorsal spine. 

Certain specimens of supercilioksus in the Albany Museum, 
collected from Great Fish Point to the Transkei coast, have the 
dorsal crest as high as the length of the head, while the membrane 
from the 3rd dorsal spine barely reaches the base of the 4th. 
So marked is the division in the dorsal fin, that, barring the 
presence or absence of orbital tentacles, there can be no doubt 
that supercilioksus and laurentii should be placed together in the 
same genus. 

The generic sub-division of Clinus, based largely upon the 
character of the dorsal fin, is widely accepted, and the South 


198 


220 South African Fish—Smith. 


African species of this genus may possibly with advantage so 
be divided. 


Text fig. 5. Clinus laurentii G & T. 


Family BROTULIDAE. 
Brotula palmietensis n. sp. [Pl. XXI, fig. C.] 


Body elongate, compressed, tapering posteriorly. Snout 
rounded, blunt, profile gently sloping from nape. Depth 6.4, 
length of head 5 in length of body. Eye 3.6 in head, 1.7 times 
snout and 2.5 times interorbital width. Preopercle without free 
margin, covered by skin. A sharp spine on the opercle. Nostrils 
double. Mouth moderate, lower jaw shorter, included in upper. 
Maxilla extends to below the hind margin of the eye. A single 
row of fine conical teeth in both jaws, those in the lower slightly 
larger: other bones edentulous. Three pairs of mandibulary 
barbels, sub-equal, about 10 in head. Two pairs maxillary and 
one pair nasal barbels; outer maxillary barbels longest, about 7 
in head. Gill-membranes free from isthmus. Gills 4, a slit behind 
the 4th. Gill-rakers reduced to mere knobs, 6-7 on the lower 
part of the anterior arch. Pseudobranchiae absent. 

D ca 100, commences over the middle of the pectoral, rays 
sub-equal. 

A ca 80, commences below the middle of the body. Dorsal 
and anal confluent with caudal, which is pointed. 

P 24, slightly less than head; rays simple. 

V 2, reduced, inserted slightly in advance of below the hind 
margin of the preopercle, slightly more than half of the length 
of the head. Inner ray short, adnate to outer filamentous ray. 


199 


South African Fish.—Smith. 221 


Scales cycloid, very small. 1.1. ca. 80, curves up over the 
pectoral, then slightly down and thereafter runs obliquely towards 
the upper margin of the base of the caudal. 8 scales above and 
16 below lateral line. Head naked except for a wedge-shaped area 
on occiput with apex over the middle of eye, merging behind 
into the general scaling. 

Colour: Silvery-brown, with numerous fine spots. Abdominal 
area and preopercle densely spotted. Snout and nape dusky. 
Fins light. Dorsal and anal with a small spot at the base of each 
ray, and with a dusky spotted border. 

Type in the Albany Museum. 

A single specimen, 50 mm. in length, almost certainly juven- 
ile, from Great Fish Point, taken in a rock pool. This is the first 
record of a species of this genus from South Africa. 


Family STROMATEIDAE. 
Psenes whiteleggii Waite. [Pl. XIX. fig. A.] 


1902, Regan. Ann. Mag. Nat. Hist. (7), vol. X, p. 126. 


Body ovate, very compressed. Profile of head deeply concave 
at interorbital, abruptly descending before eyes. Depth 2.4, 
length of head 2.4 in length of body. Eye 2.7 in head, 1.6 times 
snout and 1.3 times interorbital width. Bones of head not serrate. 
Two flat ill-defined spines at upper hinder margin of operculum. 
Lateral line on head freely branched. 

Mouth small, terminal, oblique. Maxilla extends to below 
the anterior third of the eye, almost entirely concealed beneath 
the preorbital. A single row of minute curved conical teeth in 
each jaw, other bones and tongue edentulous. Nostrils ovoid, 
large. Gill-membranes free from isthmus. Gill-rakers 16, moder- 
ate, slightly shorter than the gill-filaments, 10 in length of head. 
Pseudobranchiae present. 

D XI-I, 20, commences above the middle of the opercle, 
deeply notched, 1st XI spines form an almost separate fin. 1st 
spine short, 3rd-5th longest, almost 2 in head, remainder decrease 
to the 10th and 11th, which are small, 12th spine abruptly longer. 


200 


222 South African Fish.—Smith. 


Soft rays higher than the last spine, slightly shorter than the 
4th spine, decrease slightly posteriorly. 

A III, 19, commences below the base of the last dorsal spine. 
1st spine very small, remainder increase. Soft rays sub-equal, 
slightly longer than 3 in head. 

P 20, rounded, 1.6 in length of head. 

V I, 5, inserted below the middle of the base of the pectoral, 
1.6 in length of head, reach to the base of the 2nd anal spine. 
Last ray joined to body by membrane. 

Caudal deeply forked. 

Scales thin, cycloid, moderate, rather deciduous. Lateral 
line 60-63, gently curved following dorsal profile, slight downward 
bend on peduncle: tubules simple, 4-5 series of scales above 
lateral line, those below mostly shed. 

Colour: Light silvery-brown, darker above, nape dusky. 
Dendritic spots over most of the body and fins. Ventrals, caudal 
lobes, first dorsal, margins of pectorals, 2nd dorsal and anal dusky. 
Pectorals immaculate. The spots are denser in 3 areas on the 
body giving the impression of three faint cross bands, the first 
very wide, from head almost to vent, the second below the middle 
of the soft dorsal and the last on the peduncle. 

Distribution: South-east coast of Africa: Australia. 

A single specimen 30 mm. in length from Great Fish Point. 
A shoal of juvenile specimens was found in a rock-pool after a 
violent storm at sea, but owing to a mishap, all except this one 
specimen were lost. 

In the members of this family growth changes occur in 
the body and fins. This little specimen at first sight resembles 
Cubiceps natalensis (G & vB), but the absence of lingual and 
vomerine teeth, the highly compressed body and the position of 
the rather elongate ventrals leave no doubt that it falls in the 
genus Psenes (C & V). Minor differences from whiteleggii exist, 
but they are scarcely of specific significance. | 

The discovery of this Australian species on our coasts is in- 
teresting. 

A specimen 42 mm. in length has since been found at the 
Bushman’s River (near Port Alfred). 


201 


South African Fish.—Smith. 223 


Family SCORPAENIDAE, 
Coccotropus jubatus n. sp. [Pl]. XVIII, fig. B.] 


Body compressed. Dorsal profile even to eye, sharply concave 
before eye. Snout very blunt, abruptly descending. Depth 2.7, 
length of head 2.8 in length of body. Eye 5.2 in head, 1.4 in snout, 
and slightly greater than interorbital width. Head armoured. 
2nd suborbital produced across cheek. Preorbital produced into 
2 recurved spines, the larger above. 4 flat spines on the pre- 
opercle, the lower 3 each with a fleshy papilla above. 3 opercular 
spines, slender, below the skin. Mucous pores and granulations 
on head. » | 

Lower nostril tubular; upper circular, small, scarcely larger 
than a mucous pore. Mouth small, terminal, slightly oblique. 
Lower jaw projects slightly. Maxilla extends to below anterior 
quarter of eye. Supra-maxilla present. Villiform teeth in bands 
in both jaws and on vomer. Other bones edentulous. Gills 34. 
Gill-rakers 7, reduced to mere knobs. Pseudobrancheae absent. 
Gill-membranes free from isthmus. Gill-opening wide. Branch- 
iostegals 3. 

D XVI, 8, commences over anterior border of the eye. 
Anterior spines elevated forming a crest. 1st spine shortest 
about 5, 2nd 1.6, 8rd 1.4, in length of head. 4th spine sub-equal 
to 2nd, the remainder decrease to the 7th, remain equal to the 
9th, while 10th-16th are sub-equal, and slightly longer than the 
9th. The soft rays are simple, and increase in length posteriorly: 
the last ray is abruptly shortest and is joined to the peduncle 
at the base of the caudal. 

A II, 7. Commences below the base of the 12th dorsal spine. 
Spines short, curved, moderately stout. Ist anal spine 6, 2nd, 4 
in length of head. Rays longer than the spines, the last not 
joined to the base of the caudal. 

P 9, rays simple, almost as long as head, reaches to vent. 

V I, 2, very oblique, bases moderately separated. Inserted 
slightly in advance of pectorals. Spine almost half length of 
head, rays simple, inner ray half length of spine, joined for 2/3 
of its length by a membrane to the belly. | 


202 


224 South African Fish.—Smith. 


Caudal 13. Pointed, as long as head. Peduncle as deep a8 
long. 

Scales absent. Head rugose. 1.]. tubules 18, extend to peduncle. 

Colour: Brown with traces of darker marbling. Streaks on 
head, one from interorbital through eye to margin of preopercle, 
Fins light. 

Type: A single specimen 57 mm. in length from Natal, 
precise locality unknown, the specimen being among certain fishes 
obtained by the S.S. “Pickle,” labelled “Natal fishes, unclassified,” 
donated by the Director of the Government Fisheries Survey to 
the Albany Museum. 

The systematic position of this specimen is somewhat doubt- 
ful. In some respects it would appear closely related to the 
Aploactidae, or at any rate midway between this family and the 
Scorpaenidae. Since I had not sufficient material or literature 
to enable me to decide this point, I sent the specimen to Mr. 
Norman of the British Museum, who considers that it is a 
Scorpaenid, probably falling in Coccotropus Kaup. I am pro- 
visionally accepting this diagnosis, although the elevated anterior 
portion of the dorsal does not appear to be characteristic of this 
genus. 


Family SCORPAENIDAE. 


Scorpaena kowiensis. n. sp. 


Body very slightly compressed. Snout moderately pointed, 
slightly depressed. Depth 2.7, length of head 2.5 in length of 
body. Eye 3 in head, equal to snout and twice interorbital width. 
Head ridged and spinose. 2 flat spines on operculum; a large 
spine at angle of preopercle, 4 small spines below. 3 preorbital 
spines, one pointing forward, the other two downward and 
backward over the maxilla. 3 small postorbital, 3 temporal, 3 
-supraorbital, one nasal and two suprascapular spines. Spinate 
suborbital ridge. A flat spine on shoulder girdle above axil of 
pectoral. Interorbital concave with a slight medio-longitudinal 
ridge. No supraorbital tentacles. No skinny flaps on head or body. 


203 


South African Fish.— Smith. 225 


Mouth moderate. Maxilla extends to below the anterior third of 
the eye. Villiform teeth in both jaws in bands, which are wider 
anteriorly. Villiform teeth on vomer and on anterior portion of 
palatines. Gill-openings wide, membranes free from isthmus. 
Gill-rakers reduced, 5, short and stout, plus four rudiments, on 
lower limb of anterior arch. Pseudobranchiae present. Branch- 
iostegals 7. 

D XII, 9, deeply notched, commences over the base of the 
pectoral, in advance of the hind margin of the opercle. 1st spine 
short, remainder increase to the 4th, which is 3 in head: there- 
after decrease to the 11th, which is shorter than the lst. The 
last spine is abruptly longer than the 11th. Soft rays higher than 
the spines. 

A ITI, 5, commences below the base of the 2nd dorsal ray. 
lst spine short, 2nd longest, 2 in head. Soft rays longer than 
the spines. 

P 15, rounded, 1.5 in head. 10 lower rays simple, upper 5 
branched. V I, 5, bases close together. Inner ray short, connected 
for half its length by a membrane to the belly. 

Caudal 22, rounded. 

Scales ctenoid, moderate. Lateral series of scales 39, tubules 
24, ltr. 4. Head naked, indications of rudimentary scales 
below skin. 

Colour: Red-brown, head and nape darker, mottled and 
spotted. Pectorals and dorsal brownish, spotted. Dark bar across 
the base of the caudal. Other fins light. 

Type in the Albany Museum. 

A single specimen, 59 mm. in length, from Port Alfred. 

This species is closely allied to haplodactylus (Blkr.) which 
has been recorded from Natal (Gilchrist and Thompson, Ann. 
Durb. Mus. 1917, vol. I, p. 408), but specimens from South 
African waters have apparently not been described. 

Kowiensis differs from haplodactylus chiefly in having only 
10 simple pectoral rays, and in the absence of supraorbital 
tentacles and of skinny flaps on the body. 


204 


226 South African Fish—Smith. 


Family PLATYCEPHALIDAE. 


Platycephalus indicus (Linn.). [Pl]. XX, fig. D.] 
“Vlei stok-vis” (Knysna). 


1878, Day, Fishes of India, p. 276 (P. insidator, Forsk.) 
1925, Fowler, Proc. Ac. Nat. Sci. Phil. vol. LXXVII, p. 255. 
1925, Barnard, Ann. S.A. Mus. vol. XXI, p. 932. 


Body elongate, very depressed. Depth 11, length of head 3.7 
in length of body. Eye 7.4 in head, 2.2 in snout, and 1.2 in 
interorbital width. Width of head at level of base of preorpercle 
spines 1.3 in length. Head very depressed, ridged, ridges mostly 
smooth. Interorbital concave. A small blunted spine before the 
inner margin of the eye. Postorbital longitudinal ridge with two 
small spines, anterior an eye diameter behind eye, posterior mid- 
way between hind margins of eye and operculum. Two preoper- 
cular spines, upper gently curved, 2 in eye, lower slightly longer, 
straight. One opercular spine, blunt and feeble. Two supra- 
scapular spines, lower strong. A naked skinny flap on opercular 
membrane below preopercular spines. Nostrils circular. No 
tentacles. A sub-triangular flap of the iris extends over half the 
pupil from inner margin. Maxilla extends to below middle of eye, 
lower jaw projects strongly. A band of villiform teeth, widening 
abruptly anteriorly in upper jaw, 2-3 anterior teeth on each side 
much enlarged. A single series of small conical teeth in lower 
jaw: a curved row on vomer single anteriorly, multiple posterior- 
ly with enlarged inner teeth. A single series of curved conical 
teeth on palatines. Tongue edentate free. 

Gill membranes free, rakers 6 plus 4 or 5 rudiments. Pyloric 
caeca 9. 

D I+VII+1+12-13, originates 2.3 times as far from base of 
caudal as from tip of snout. 1st spine short, 15 in head. Second 
dorsal elevated, 1st and 2nd spines 2.0, 3rd 2.5, last 13 in head, 
membrane scarcely incised. 

Behind the second dorsal is a small recumbent spine. Soft 
rays anteriorly elevated, lst and 2nd longest, 2.4 in head, re- 
mainder graduated, 12th 4.0, last 7.5 in head. Membrane deeply 
incised between rays. 


205 


South African Fish.—Smith. 227 


A 12-13, inserted below origin of soft dorsal. 1st and last 
rays sub-equal, 5 in head; rays increase to the 5th and remain 
sub-equal to the penultimate. Membrane deeply incised between 
rays. 

P 18, more or less rounded, 6th ray longest, 2 in head. Tips 
of lower five rays free from membrane. 

V I, 5, inserted below 1st dorsal spine, last (inner) ray 
longest 1.5 in head. 

Caudal sub-truncate, peduncle moderately depressed. 

Scales strongly ctenoid, 8 radiating striae; lateral line 
smooth. 1.1. 128, 12 rows above. 

Colour: Yellow-brown above and on sides, with very faint 
darker cross-bars, white below. Dark spots on lower side of head 
and on opercular membrane. Dorsal spines and rays spotted or 
annulate. Membrane light. Pectorals and ventrals brown, marbled 
and spotted with darker. Caudal dark below, middie and upper 
parts yellow, with large black spots above. 

Length: Up to 670 mm. 

Locality: Knysna, Port Elizabeth, Port Alfred. 

Distribution: South and East Coast of Africa to Indo- 
pacific. 

The numerous synonyms of indicus would appear to indicate 
that it is an exceedingly variable species. 

Barnard (loc. cit.) states that the ridges on the head are 
quite smooth. Day (loc. cit.) mentions the small spine above the 
eye. The two specimens here described have two small but distinct 
recumbent spines on each postorbital ridge, besides the supra- 
orbital spine. Fowler (loc. cit.) states that his specimens have 
ll. 76 or 77; such wide variation is scarcely possible, and the 
statement is possibly erroneous. 

The iridal flap is conspicuously triangular in shape. Day 
(loc. cit. p. 274) states that there are two, a superior and an 
inferior flap, both semi-circular. There is no inf?rior flap in my 
specimens. | 

It is an open question whether our specimens merit specific 
distinction from those of the Indo-pacific, but a comparison of 


206 


228 South African Fish.—Smith. 


specimens from both localities would be necessary to justify such 
a step. 

Indicus is not an infrequent capture at Knysna, and a fish 
taken in the Breede River, Port Beaufort, described to me by 
a netter, was probably this species. At Knysna it is not in- 
frequently taken also on lines, but it is regarded locally as 
inedible. The name “Vlei-stokvis” is derived from the shape of 
the posterior portion of the body and the tail, which resembles 
somewhat those of the “Stok-vis,” (Merluccius capensis Cast.) 


Family MONACANTHIDAE. 
Monacanthus setifer Benn. [Pl. XIX, fig. B.] 


1925, Barnard, Ann. S.A. Mus., vol. XXI, p. 958. 


Body very compressed, back concave between the dorsals. 
Depth (back to lower margin of ventral flap) 1.5-1.8, or (origin 
of dorsal to origin of anal) 1.6-1.8, length of head 2.8-3.0 in body. 
Eye 3.4-8.7 in head, 2.4-2.8 in snout and 1.0-1.2 in the inter- 
orbital width. Gill-opening oblique, slightly less than an eye 
diameter, posterior margin slightly behind eye, in a straight 
line with hind margin of eye and origin of dorsal spine. Profile 
of snout steep, gently concave. Scales papilliform, with stellate 
apical dilation, causing skin to feel rather rough. 

D 1+ 31-32. Spine stout, 1.6-1.7 in head, rough in front, with 
2 rows of barbs behind, originates above hind margin of pupil. 
Soft dorsal low, inserted at highest point of dorsal profile, rays 
Increase to the 4th, which is 2.7 in head, remain sub-equal to 
the 16th-18th and thereafter decrease somewhat. 

A 33-35, originates below the 5th dorsal ray, slightly lower 
than dorsal but of similar shape. 

P 14-15, 2.6 in head. Ventra] spine movable. 

Caudal 12, rounded, peduncle slightly longer than eye. 

Colour: (Alive) Greenish brown, uniform, or with obscure 
longitudinal series of dark blotches. Dorsal, anal and pectoral 
yellow-brown. Caudal uniform brown-green, sometimes dusky 
posteriorly. 

Length: 94-195 mm. 

Locality: Knysna, Port Alfred, Great Fish Point. 


207 


South African Fish.—Smith. 229 


These specimens are provisionally assigned to setifer Benn., 
from which however they may prove to be distinct. The species 
in this family are known to exhibit sexual dimorphism (Ebina, 
J. Imp. Fish. Inst. Tokyo, 1932, XXVII, p. 15) and are in general 
somewhat variable. 

The present specimens would appear to differ from the 
typical form of setifer in the greater number of anal rays, in 
the anteriorly low soft dorsal, as well as in this latter fin being 
inserted at distinctly the highest point of the dorsal profile. Day 
(Fishes of India, 1888, p. 692) states that setifer has the dorsal 
spine inserted above the posterior part of the eye, which is the 
highest point of the dorsal profile; and that there are 12 rays 
in the pectoral and 9 in the caudal. There is a small specimen 
(65 mm.) in the 8.A. Museum, in which the spinous and soft 
dorsal are on the same level, but all other South African speci- 
mens I have seen have the soft dorsal on a higher level than the 
spine, P 14-15 and C 12. Further, our specimens have not the 
two dark cross-bars on the caudal, and the body is deeper than is 
usually the case in setifer. . 

None of our specimens have the filamentous dorsal ray (lst. 
Barnard; 2nd Day) occasionally observed in setifer. 

Individually, these differences would scarcely justify the 
separation of our specimens from setifer, but collectively they 
would indicate that it might be justifiable. 

Setifer is only occasionally found at Knysna, but is fairly 
common eastwards from Port Elizabeth. 


Family TETRODONTIDAE. 
Tetrodon lagocephalus Linn. [Pl. XX, fig. A.] 
1925, Barnard, Ann. 8.A. Mus. vol. XXI, p. 966. 


Body elongate, fairly compressed. Depth (uninflated) 3.8, 
length of head 38.8 in length of body. Eye 5.8 in head, 2.5 in 
snout, 2.3 in interorbital and in post-orbital part of head. 

Belly from below middle of snout to vent with fixed 4-rooted 
spines: skin otherwise naked and smooth. Head less than dis- 
tance from origin of dorsal, Lateral cutaneous fold distinet, 


208 


230 South African Fish.—Smith. 


more obvious posteriorly. A small longitudinal fold of skin on 
upper part of peduncle. Nostrils paired, flush with surface. 
Mucus canals on head distinct. Whole of interorbital bony, 
expanded laterally over eye. 

D 13, originates about midway between base of caudal and 
tip of pectoral, faleate, 2 in head. Base raised, lobate. Last ray 
joined to body by membrane. 

A 18, arises slightly behind dorsal, same size and shape as 
dorsal. 

P 15, scarcely faleate, 1.7 in head, lower four rays abruptly 
shorter. Caudal emarginate, posterior margin undulate, lobes 
pointed, lower 1.2 times upper. 

Colour: Indigo-blue above, light blue below. Spinose belly 
almost white. A series of small black spots along side. Dorsal 
and caudal dark blue-black. Anal light, posterior margin dusky. 
Pectoral blue-black above, lower four rays light with dusky 
margin. A dusky blotch behind vent, behind anal and on lower 
part of peduncle. 

Length: 550 mm. 

Locality: Knysna River. 

Distribution: Atlantic, Mediterranean, South African seas 
to Natal. 


Barnard (loc. cit.) considers that the specimen somewhat 
scantily described by Fowler (Proc. Ac. Nat. Sci. Phil. 1925, vol. 
LXXVII, p. 267) as stellatus Schn. may possibly be lagocephalus. 
Fowler states definitely that almost the whole body is minutely 
spinulose, and that the caudal is convex behind. Stellatus is 
generally accepted (also by Barnard) as a synonym of aerostaticus 
Jenyns, and Fowler’s specimen is very likely this latter species, 
although this author does not mention the presence of the 
characteristic black ring round the vent. 

Lagocephalus is evidently a very rare visitor at Knysna, for 
the specimen described was shown to many local netters and 
line fishermen, but none could recollect having previously seen 
another, 


209 


South African Fish.—Smith. 231 


Family OSTRACIONTIDAE. 
Lactophrys quadricornis (Will.). (Pl. XXII, fig. B.] 
1865, Bleeker, Atl. Ich. V, p. 32. 


Depth 3, head 6.5 in total length. Preorbital depth equal 
to head, 3 times eye. Carapace three-angled, dorsal ridge acute, 
without spine. A stout spine, directed backwards, on each ventral 
ridge, below the dorsal fin. A conical spine, as long as eye, before 
each eye. The carapace forms a continuous bridge across the 
caudal peduncle behind the dorsal, terminating in an oblique 
median spine above, pointed below, laterally deeply concave. 
Interorbital deeply concave. Anterior opening of carapace slight- 
ly less than eye. 

D 10, rays branched. A 10, inserted behind dorsal, rays 
branched. P 11, base almost horizontal, rays branched. Caudal 
4 in total length, rays much branched, peduncle free, 2/3 length 
of head. Whole body granular, with fused margins of hexagonal 
scutes naked. 

Colour: Light brown, rows of dark spots on upper part of 
side. Four longitudinal bars on snout and cheek. Fins light, 
spotted. 

Length: 290 mm. 

Locality: Algoa Bay, (Cape of Good Hope, Bleeker). 

Distributon: Tropical Atlantic, Indian Ocean. 

A single specimen only, in the Port Elizabeth Museum. 

This species must be a very rare migrant, for it has not 
been recorded from South Africa since 1865 (Bleeker loc. cit.) 


Lactophrys gibbosus Linn. 


1888, Day, Fishes of India, p. 695, Pl. CLXXXI, fig. 4, 
(Ostracion turritus) . 

Body 3-ridged. Dorsal ridge sharp, elevated into a strong, 
slightly recurved, compressed spine, much higher than dorsal fin. 
At the anterior end of each supraorbital ridge is a fairly strong 
recurved compressed spine, projecting slightly outwards. On 
each ventro-lateral ridge are 4 compressed recurved spines, the 


210 


232 South African Fish.—Smith. 


anterior just behind the pectoral, the posterior below the middle 
of the dorsal. 

Height of body 2, length of head 3.5 in length to caudal base. 

Eye 2.6, preorbital depth 1.7 in length of head. Median cross- 
section of body, exclusive of ridges, an equilateral triangle. In- 
terorbital deeply concave. 

Carapace forms a continuous bridge above and below 
peduncle, pointed (60°) above, rounded below. 

D.9, A.9, former inserted well in advance of latter. P 11, 
almost vertical. Caudal rounded, 1.4 in head. 

Colour: Brownish, with numerous round blue spots, mostly 
one in each scute. 

Length: 195 mm. 

Locality: Durban. 

Distribution: East coast of Africa to the Indo-Malayan area. 
New York (fide Day loc. cit)? 

This is the first record of the above species from South 
Africa. 

It is a little difficult to understand why Lactophrys Swnsn. 
has been accepted by many systematists as different from 
Ostracion, whereas Lactoria has not. The species of Lactophrys 
that I have examined all show distinct traces of the two dorso- 
lateral ridges clearly present in Ostracion (s.s). 

Whichever of the 3 forms (i.e. carapace 3-, 4- or 5-angled) 
be selected as typical, the remaining two are clearly minor varia- 
tions only, and seem scarcely to merit full generic distinction. 


Family BALISTIDAE. 
Balistes conspicillum Blch. 


1888, Day, Fishes of India, p. 689, 


19038, Jordan & Fowler, Proc. U.S. Nat. Mus., vol. 25, p. 256 
(Pachynathus c.) 


Depth about 2, length of head 3 in length of body. Eye 4 in 
snout, 5.5 in head. 

A marked groove before eye: enlarged scutes behind gill- 
opening. Mouth set in naked ring round end of snout and chin. 


211 


South African Fish.—Smith. 2338 


D.11I+26, inserted behind eye, above gill-opening. First 
Spine very stout, 1.4 in head, uniformly granulate anteriorly. A 
23, inserted well behind dorsal. 

42 series of scales. Lateral line curves up from the shoulder 
then down below the soft dorsal origin. About 33 transverse 
series of scales between the soft dorsal and the vent. Cheek 
scales smaller than body scales. Each scale on mid-hinder part of 
body and tail with a retrorse spine: anteriorly there are 8-¥ 
rows with small spines, graduated posteriorly to 2-3 rows with 
much larger spines on the peduncle. 

Colour: Brownish, a light band across the snout before the 
eyes, originating each side below the anterior margin of the eye. 
A more or less triangular light patch below the cheek. 18-20 
light round-oval patches along the lower half of the side. A 
narrow light ring round the mouth. 

Length: 190 mm. 

Locality: Presumably Port Alfred, having been found among 
other more common species from this locality, in the old collection 
of the Albany Museum. 

Distribution: Indo-pacific. 

I have since been informed by Mr. Chubb, Director of the 
Durban Museum, that he possesses a specimen of this species 
from Durban Bay. 


Family ONCHOCEPHALIDAE. 
Halieutea fitzsimonsi G. & T. [Pl. XXIII, fig. B.] 


1916, Gilchrist & Thompson, Mar. Bio. Rep., p. 58, fig. 
(Halieutichthys f.) 

1921, Regan, Ann. Mag. Nat. Hist. (9) VII, p. 419 
(liogaster). | 

1922, Gilchrist, Fish. Mar. Survey. Spec. Rep. III, p. 79 
(liogaster). 

1923, von Bonde ibid, Spec. Rep. 1, p. 36 (uogaster). 

1925, Barnard, Ann. §.A. Mus. XXI, p. 1009 (Halieutichthys 
f.) and p. 1010 (liogaster). 

The type of this species is in the Port Elizabeth Museum. 

The original description gives the length as 158 mm. The actual 


212 


234 South African Fish.—Smith. 


specimen I examined is 188 mm. in length, but the difference is 
probably due to typographical error, not uncommon in the earlier 
descriptions of South African fishes. There are also only four 
anal rays and not nine, as given by Gilchrist and Thompson, 
which is probably a similar error. 

The markings described by the original authors, while some- 
what faded, may clearly be discerned in the type. Two other 
specimens from Algoa Bay, and another from East London, are 
clearly conspecific with the type of fitzsimonsi, the two pairs 
of rings, one dark and one light, being very obvious in fresh 
specimens. 

It is however singular that none of the four specimens 
examined have any trace of vomerine or of palatine teeth, while 
gill-rakers are present in each case, being clearly visible on 
distension of the oral cavity. There are, however, two pairs of 
dentigerous upper pharyngeals, which are placed rather far 
forward, owing to the distortion of the usual positions of the 
bones of the palate by longitudinal compression, such as is found 
in fishes of this type. It is possible that Gilchrist and Thompson 
mistook these pharyngeal for palatal teeth, although their 
orientation in regard to a similar pair of lower dentigerous 
pharyngeals immediately reveals their true nature. 

The gill-rakers are short and stout, apically dilated, 6 on 
the lower limb of the anterior arch. 

This species thus falls in the genus Halieutea C & V, whieca 
is distinguished from the related genus Halieutichthys Poey by 
the absence of palatal teeth, and by the presence of gill-rakers. 
It may be remarked that Barnard (loc. cit.) was quite evidently 
doubtful of the validity of G & T’s diagnosis, for the latter genus 
is found in the West Indies. 


A. comparison of these specimens with liogaster Rgn shows 
that it is not necessary to maintain the latter ag distinct, since 
the two dark post-orbital spots of fitzsimonsi represent the only 
difference. One specimen of liogaster in the S.A. Museum has a 
dark post-orbital spot on one side only. The spots tend to dis- 
appear in preserved specimens, 


213 


South African Fish.—Smith. 235 


One of the larger specimens has two tentacles on the anterior 
margin of the depression containing the gill-opening. The outer 
is trifid, the inner simple. The lower surface of the live fish is 
coral-red, but the colour fades rapidly on preservation. 


I wish to express my indebtedness to Dr. Barnard, Assistant 
Director of the S.A. Museum for assistance, and for the loan of 
material and literature: to Mr. J. Hewitt, Director of the Albany 
Museum, whose wide and unfailing knowledge is ever available | 
and invaluable: and to the Carnegie Fund (through the Research 
Grant Board of South Africa) for generous financial assistance. 


Albany Museum, 
Grahamstown, 
March, 1934. 


Plate XVIII. 


Rec. Alb. Mus., Vol, IV. 


As 
Say 
a 

: _ : 
SSR 
Saas 
SRG 


i Y 
a 


a 
a : 7 


: 


ii 


ce 


a 
sa 


ubatus sp. nov. 


° 


sj 


Coccotropu 


B. 
gn. 


ilehr. 


Xenolepidichthys dalgleishi 


A. 


insont 


gnathus rob 


’ 


Hople 


c, 


FR 
| 
ee 


Rec. Alb. Mus., Vol. IV. 


- _ 


. 
__ 


ae 


ne fe . ¥ 


A. Psenes whitelegii Waite. 
C. Iso natalensis Rgn. 


a 


a . : 


l 


* 
i 


Monacanthus setifer Benn. 
Median body scale x 16. 


y 


_ 


oy 


XIX 


Rec. Alb. Mus., Vol. IV. 
Plate XX. 


A. Tetrodan lagocephalus Linn. B. Opisthognathus macrostomus sp. nov. 
C. Blennius bifilum Gunth. D. Platycephalus indicus Linn. 


ate XX] 
ec. Alb. Mus., Vol. IV. 


Oro\eMtemormenet 


ages 


Pt 
ar — i 
Maienieertn tances aie 


A. Acanthocepola cuneatus sp. nov. B. Petroscirtes tapeinosoma Blkr. 
a. Lateral line scale. I). .Decapterus lajang Blkr. 
: C. Brotula palmietensis sp. nov. a. Anterior lateral line scale x 8. 


ec. Alb. Mus., Vol. IV. 
Rec. A us., Vol. I\ Plate XXII. 


A. KEpinephelus favocaeruleus Lac. B. Lactophrys quadricornis Linn. 
C. Chaetodon murleyi Ren. | D. Dascyllus axillaris sp. nov. 
EK. Hoplostethus gilchristi sp. nov. . F. Clinus heterodon C. & V. 


(The line below each figure represents one centimetre.) 


Rec. Alb. Mus., Vol. IV. 


Plate XXII] 


# 


a 


sab A gen nent % 


A. Pagrus nasutus 


B. Halieutea fitzsimonsi G. & T. 


215. 


From A Guide to the Vertebrate Fauna of the Eastern Cape 
Province of South Africa. Part ll. Reptiles, Amphibians and 
Fresh—water Fishes. pp. 119-141. Pls. XXIX—XXXIV. 1937. 
Albany Museum, Grahamstown. 


FRESH-WATER FISHES OF EASTERN CAPE 
PROVINCE. By Dr. J. L. B. Smrru. 


Class PISCES. 


(Fishes, breathing by gills.) 


Sub-class TELEOSTOMI. 
(Fishes other than ‘“ Sharks ”’ or “‘ Rays.’’) 


Order TELEOSTEI. 
(True Bony Fishes.) 


About 40 genera and over 200 species of fishes are known 
to occur in the fresh waters of South Africa. By far the 
greatest number of those are found in Natal, Transvaal, and 
Rhodesia, relatively few in the Cape Province, while only 28 
species are known from the waters of the South-eastern Cape. 

Our knowledge of even the scanty ichthyfauna of the 
latter region is far from complete, for no systematic investiga- 
tion has apparently ever been undertaken there. Additional 
records, and probably even new species, await discovery, 
although the conditions in the South-eastern Cape preclude 
the possibility of the existence of varied and abundant fish 
life. There are no natural extensive permanent bodies of 
water. The existing waters are generally subject to rapid 
and extreme variation between drought and inundation, 
both of which take heavy toll of fish life. Inevitably also, 
the devices of man still further reduce the already depleted 
numbers. Drought-restricted pools are swept by nets, while 
unscrupulous persons not infrequently destroy all life by 
the use of explosives. 

Our rivers are normally merely a series of disconnected 
pools, which during a drought may largely disappear, or 
become so foul or saline that only a few species may survive. 
During floods a large proportion of those fishes which have 
survived are swept out to sea and destroyed. 

It is not at all unlikely that in our natural waters the 
numbers of individuals, and even of species, are becoming 
steadily less. This is regrettable, but as our inland fishes 
are not ever likely to be of economic significance there is 

119 


216 


little possibility that any serious effort will be made to 
readjust matters. 

Besides indigenous species, there are present also in our 
area certain others, which have been introduced from the 
Northern Hemisphere, chiefly as angling fishes. The Carp, 
the Trout, and most recently the Black Bass, are the most 
common. The former two have long been established, and 
in certain parts of South Africa have been flourished and 
multiplied ; in the case of the Carp often at the expense 
of the natural fauna. 

At least one of our fishes, the ‘‘ Geel-vis”’ (Barbus 
holubi) attains a fair size, and ranks high as a game-fish. 
- Large sums have been expended upon those imported species, 
and it is to be regretted that a part of that money has not 
been diverted to the propagation of species such as the 
‘“‘ Geel-vis.’”’ In our area, where the Trout is only moderately 
successful, the “ Geel-vis,’ with little attention, would 
probably flourish and provide excellent angling. 


Systematic Account. 


Below will be found a brief account of the main external 
features and dimensional relationships which are chiefly 
employed in diagnosis and identification of fishes. 

Fins are of three kinds: they may be composed of bony 
spines, or of soft articulated rays, or they may be rayless 
or fleshy, known as adipose. Fin formulae are written in 
abbreviated form: D=dorsal fin; A=anal fin. Spines are 
indicated by Roman, rays by Arabic numerals. D. XII. 10, 
indicates 12 spines and 10 soft rays in a single undivided 
dorsal fin. D.VI.+1, 8. indicates two separate dorsal fins, 
the first of 6 spines, the second of 1 spine and 8 rays. 

Scales are broadly of two kinds: cycloid scales have a 
smooth external margin ; ctenoid scales have the hind margin 
coarsely or finely toothed or serrate. Various scale-counts 
are employed—l.].=lateral line; 1.tr. =lateral-transverse 
series; e.g. 1]. 37 indicates that there are 37 scales in the 
lateral line from its origin at the shoulder to the base of the 
caudal fin. When there is no lateral line the number of 
lateral series of scales between the same limits is taken. 
When there are two lateral lines the count is written, e.g. 


b] 


17 which indicates 15 scales in the anterior (or upper) 


217 


lateral line, and 17 in the posterior (or lower) lateral line. 
4-5 |, 
Ltr. 79 indicates 4-5 transverse series of scales between 


the top of the body and the lateral line, and 7-9 series 
between the lateral line and the belly. 

Adipose eyelids are translucent skinny eyelids which 
generally cover only part of the eye. They are usually not 
easily seen in live or fresh specimens. | | 

Short barbels are easy to overlook; gentle probing with 
a needle will disclose whether they are present or not. 


TExtT-FIG. 1.—Diagram of bony fish, showing chief dimensions. 


B. Length of body. D. Depth of body. H. Length of head. 
1. Spinous dorsal fin. 2. Soft dorsal fin. 3. Adipose dorsal 
fin. 4. Caudal fin (rounded). 5. Anal fin. 6. Ventral (pelvic) 
fin. 7. Pectoral fin. 8. Lateral line. 9. Preopercle. 10. Opercle. 
11. Adipose eyelids. 12. Snout. 13. Nasal barbel. 14. Maxil- 
lary barbel. 15. Mandibulary barbel. 


Gills and gill-rakers.—Gill-rakers are finger-like pro- 
cesses on the upper surface of the cartilaginous gill-arches. 
On the lower surface of these are the vascular (usually red) 
gill-filaments. The number of gill-rakers on the lower 
(longer) limb of the outer gill-arch is counted. This count 
may generally be made merely by lifting the operculum (or 
gill-cover), when the outer gill-arch generally is fully exposed. 


Dimensional Relationships. 


Length of body: measured from the tip of the snout to 
the base of the caudal fin [Text-fig. 1, B]. 

Depth of body: the greatest depth of the body, wherever 
it may occur [Text-fig. 1, D]. 

Length of head: measured from the tip of the snout to 
the hind margin of the operculum, or gill-cover [Text-fig. I, H]. 

Dimensional relationships are remarkably constant in 


218 


any one species, and are important in diagnosis and identi- 
fication. 

The relationships are expressed in an abbreviated form, 
e.g. “depth 3, head 3-4 in body length” indicates (a) that 
the maximum depth of the body is contained 3 times in the 
length of the body, i.e. the depth is 4 of the length ; obviously 
the more slender the body the greater will be the depth 
cipher; (b) that the length of the head varies from 3-4 
of the body length. Also ‘‘ eye 5-5 in head ” indicates that 
the head is 5-5 times the length of the diameter of the eye. 


Key to the Families of Fresh-water Fishes of the 
Southern-eastern Cape. 


I, Caudal fin united with dorsal and anal fins. 


A. Barbels present. Spine in pectoral fin. Plotosidae (p. 130). 

B. Barbels and fin-spines absent. Anguillidae (p. 131). 
II. Caudal fin free, separate from dorsal and anal. 

A. A spine in the pectoral fin. Bagridae (p. 129). 


B. No pectoral spine. 
X. A single dorsal fin (spine +rays, or rays 


only). 
a. No lateral line. 
x. Caudal deeply forked. Clupeidae (p. 139). 
y. Caudal truncate or rounded. Galaxiidae (p. 140). 
6. A single lateral line. 
x. Barbels present. Cyprinidae (p. 122). 


y. Barbels absent. 
a. Soft dorsal and anal elevated Monodactylidae 


anteriorly. (p. 139). 
B. Soft dorsal and anal not 
elevated anteriorly. Centrarchidae (p. 141). 
c. Two lateral lines. 
x. 3-4 anal spines. Cichlidae (p. 132). 
y. 5 or more anal spines. Anabatidae (p. 134). 
Y. Two separate dorsal fins. 
a. Posterior dorsal adipose. Salmonidae (p. 140). 
b. Posterior dorsal normal. 
x. 4 spines in first dorsal. Mugilidae (p. 135). 
y. 6 spines (weak) in first dorsal. Gobiidae (p. 137). 


Family CYPRINIDAE. Members of this family are found 

throughout Africa; most are small, but some species attain 

a fair size. asily recognised by the scaly body, and by the 

presence of one or two barbels. Three genera occur in the 
South-eastern Cape. 


I. Mouth more or less normal, lips do not form a sucking disc. 
A. Dorsal fin short, base shorter than head. Barbus. 
B. Dorsal fin longer, base longer than head. Cyprinus. 
II. Mouth inferior, lips form a sucking disc. Labeo. 


219 


Genus BARBUS Cuvier. 


Smallish mouth, sometimes overhung by snout. Usually 
one or two barbels on each side. Dorsal and anal fins short, 
the former sometimes with the last spine enlarged, and 
serrate behind. Scales fairly large; lateral line present, 
sometimes incomplete. In some species breeding males 
develop spiny tubercles on the head. 

The species are numerous and usually abundant in the 
fresh waters of Africa and Asia. - Most are small, but some, 
e.g. the “‘ Mahaseer”’ of India and our own “ Geel-vis,’’ 
attain a large size, and are angling fishes of the first rank. 

Some of the smaller species (e.g. paludinosus) occur in 
vast numbers, and being carnivorous probably play a con- 
siderable part in combating the mosquito menace in tropical 
and sub-tropical areas. Those species multiply even in 
restricted waters, so that they might perhaps with advantage 
be used to stock waters in mosquito-infested areas in South 
Africa. 


Key to the Eastern Cape Species. 


I, A single barbel on each side. 
A. Last dorsal spine serrated behind. trevelyani. 
B. Last dorsal spine not serrated. 
x. Barbel less than to equal to eye diameter. anoplus. 
y. Barbel 14 times eye diameter. senticeps. 
II. Two barbels on each side. 
A. Last dorsal spine serrated behind. 
x. Ventral base entirely in advance of dorsal. paludinosus. 
y. Ventral base partly in advance of dorsal. brookingi. 
z. Ventral base below middle of dorsal. capensis. 
B. Last dorsal spine not serrated. 
x. Striae on exposed surface of scales more or 
less parallel, longitudinal. 


a. Lips produced into lobes. gilchristi. 
B. Lips normal. holubi. 
y. Striae on exposed surface of scales radiating 
fan-wise. 
a. Lateral line complete. 
i. Base of ventral in advance of dorsal. __ vulneratus. 
ii. Base of ventral below dorsal. burchelli. 
8. Lateral line obsolete posteriorly. hemipleurogramma. 


Barbus trevelyani Gunther. [Pl. 29, Fig. 1.]—Depth 
3-4-4, head 3-7-4 in body length. A single barbel on each 
side, slightly less than eye. D.III.7, third spine strong, 
serrated behind on upper part only. A.III.5. Pectoral 
1-2 in head, not reaching ventral. Scales radiately striate 


220 


33-36, l.tr. 5-6 34 between lateral line and ventral base. 


Brown-yellow, darker above. A dark lateral stripe above 
lateral line, ending near caudal base in a round dark spot. 
L.R.S.t 4 inches. 

Recorded from the Buffalo River near Kingwilliamstown. 

Barbus anoplus Weber. ‘ Rooi-vlerk”’ or “ Red-fin.” 
[Pl. 29, Fig. 2.*]}—Depth 3-5-4-5, head 3-4 in body length. 
Eye 3-5 in head. A single barbel on each side, slightly less 
than or equal to eye. D.III.7, third spine feeble, not 


TExtT-Fic. 2.—Barbus senticeps (J. L. B. Smith). 


serrate. <A. III. 5, pectoral 1-2 in head, not reaching ventral. 
5-7 

Scales radiately striate, 32-38 l.tr. — 4-5 between lateral 
—d 


line and ventral base. Lateral line occasionally not complete 
posteriorly. Brown-yellow, darker above, some specimens 
almost golden. A narrow lateral streak above lateral line, 
sometimes ending in a spot. Usually a bright red spot at 
bases of some or all fins except caudal. L.R.S. 5 inches. 

Common throughout the fresh waters of the Cape, often 
found in dams and reservoirs; found also in the Free State, 
Natal, and Transvaal. 

This little fish thrives in even small garden ponds, and 
is able to survive fairly long journeys in vessels containing 
water. Breeding males often have spiny tubercles on the head. 

Barbus senticeps J. L. B. Smith. [Text-fig. 2.}—Depth 
equal to head, 3-4 in body length. Eye 5 in head. A single 


+ L.R.S. = Largest recorded size. 
* Figures so marked after Boulenger, ‘‘ Freshwater Fishes of Africa.” 


221 


barbel on each side, 1-5 times eye. Large spiny tubercles 
on the snout and chin. D.III. 7, third spine feeble, not 
serrate, A. III. 5. Pectorals 1-3 in head, reach ventral base. 
Scales radiately striate, 30, ltr. 8, 3 between lateral line 
and ventral base. Dark olive-brown above, lighter below. 
_ An obscure dark lateral stripe. Tubercles reddish. L.R.S. 
34 inches. . | 

Known only from the type, from Kromme River, Assegai 
Bosch. | 

Barbus paludinosus Peters. ‘‘ Gilliminkie.” [Pl. 30, 
Fig. 4.*]—Depth 3-5-4-5, head 3-4 in body length. Eye 
3-5-4-5 in head. Two barbels on each side, anterior short, 
posterior less than or equal to eye. D. III. 7, last spine 
very strong, strongly serrated behind. A.III.5. Pectoral 
1-1-1-5 in head, reaches ventral base. Ventral entirely in 
advance of dorsal. Scales radiately striate. 1.1. 33-36, Ltr. 
6-7 
5-6" 
silvery, darker above. Sometimes a dusky lateral streak, 
L.R.S. 34 inches. | 

Common in reservoirs and dams round Grahamstown 
and throughout the Eastern Cape. Extends to Transvaal 
and Rhodesia. Occurs in vast numbers in even small dams 
and thrives in garden ponds. Consumes mosquito larvae 
with avidity. | 

Barbus brookingi Gilchrist. [Pl. 30, Fig. 3.*]—Depth 
equal to head, 4 in body length. Eye 4 in head, 2 barbels 
on each side, posterior longer, 1:3 in eye. D.III. 7, last 
spine enlarged, somewhat feebly serrated behind. A. III. 5. 
Pectoral 1:5 in head, not reaching ventral. Ventral base in 


3-4 between lateral line and ventral base. Yellow or 


6 : 
advance of dorsal. Scales radiately striate, 33, l.tr. 5 3—4 


between lateral line and ventral base. Brown-yellow, a 
faint dark lateral streak ending in a dark spot. L.R.S. 
4 inches. 7 
~ Known only from the type from East London. A doubt- 
ful species, probably identical with the preceding. 
Barbus capensis A. Smith. ‘‘ Witte-vis ” or “ Silver-fish.”’ 
[Pl. 30, Fig. 1.*]—Depth 3-7-4-4, head 3-5 in body length. 
Eye 5-7 in head. Two barbels on each side, about equal, 
1-2-1-7 times eye diameter, D. III. 8, last spine strong, 


222 


serrated behind. A. III. 6. Pectoral 1-3-1:7 in head, not 
reaching ventral. Ventral below middle of dorsal. Scales 


6-7 . 
radiately striate, 39-40, l.tr. a 4 between lateral line and 


ventral base. Dirty white or silvery, darker above. L.R.S. 
14 inches. 

Common in the rivers of the Cape Province, extends 
to Natal. 

This is one of the species of Barbus which is of some 
significance as a food and angling fish. It takes baits such 
as worm, crickets, or grasshoppers freely at times, will even 
rise to artificial fly, and on light tackle provides good sport. 
Fairly large numbers are taken by drag-net in the river 
pools of the Eastern Province. When taken from clear 
water the flesh is very tasty. The numerous fine bony 
ribs are troublesome, but in parts where fresh marine fish 
is not easily obtained this species is highly esteemed. 

Barbus gilchristi Blgr. [Pl. 29, Fig. 3.*]—Depth 4:5, 
head 3:8 in body length. Eye 5 in head. Two barbels on 
each side, posterior longer, slightly longer than eye. Lips 
very thick, both produced into median lobes. D. III. 8, 
third spine feeble, not serrate. A.II.5. Pectoral 1-7 in 
head. Ventral below dorsal. Scales longitudinally striate, 


6 
42, ltr. 7 4 between lateral line and ventral base. Brown 


above, lighter below, dusky spots above lateral line. L.R.S. 
7 inches. 

Known only from Barkly East. 

Barbus holubi Stndnr. ‘“‘ Geel-vis,’ ‘‘ Yellow-fish.”’ 
[Pl. 31, Fig. 1.*]—Depth equal to head, 3-4 in body length. 
Kye 3-6 in head. ‘Two barbels on each side, posterior usually 
longer, 1-1-5 times eye. D.IV. 8, fourth spine strong but 
not serrate. A.III.5. Pectoral almost as long as head. 
Ventral below front margin of dorsal. Scales longitudinally 


a 6-7 
striate, 34—43, Ltr. ay? 3-4 between lateral line and ventral 


base. Golden yellow, darker above. L.R.S. no precise data, 
probably over 30 inches. 

Occurs in rivers throughout the Union of South Africa, 
also in South-West Africa. 

A game fish, providing excellent sport. Rises to arti- 


223 


ficial fly. An excellent bait is a small cricket or grasshopper 
suspended a few inches below the surface of the water by 
a small float. This species is: fairly voracious, and takes 
most baits when on the feed. 

The habits and breeding habits of this species might 
repay study. Many waters which do not suit imported game 
fishes might perhaps with advantage be stocked with this 
species, which is our best indigenous fresh-water game fish. 

Barbus vulneratus Cast. [Pl. 31, Fig. 2.*]—Depth equal 
to head, 4 in body length. Eye 4-5 in head. Two barbels 
on each side, posterior longer, equal to eye. D. III. 7, third 
spine feeble, not serrate. A. III.5. Pectoral 1-2 in head. 
Ventral in advance of dorsal. Scales radiately striate, 32-34, 


D 
l.tr. Part 3-4 between lateral line and ventral base. Brown 


above, lighter below. A longitudinal series of spots, increas- 
ing in size posteriorly. L.R.S. 34 inches. 

Found in the Cape Province. Recorded also from the 
Transvaal. 

Barbus burchelli A. Smith. [Pl. 30, Fig. 2.*]—Depth 
equal to head, 3-4 in body length. Eye 4-6 in head. Two 
barbels on each side, posterior longer, up to 1-4 times eye. 
D. III. 7, last spine feeble, not serrate. <A. III. 5. Pectoral 
1:3 in head. Base of ventral below dorsal. Scales radiately 


striate, 33-38, I.tr. =, 4 between lateral line and ventral 


base. Brown, darker above, lighter below. L.R.S. 5 inches. 

Common in the South-west Cape. Now recorded also 
from the Buffalo River, East London. Will probably be 
found throughout the Eastern Province. 

Barbus hemipleurogramma Blgr. [PI. 29, Fig. 4.*]— 
Depth equal to head, 3-4 in body length. Eye 3-4 in head. 
Two barbels on each side, posterior longer, up to 1-5 times 
eye. D.III. 7-8, last spine feeble, not serrate. A. II. 5. 
Pectoral 1-3in head. Ventral below dorsal. Scales radiately 


4 ; 
striate, 27—29, l.tr. 15’ 2-3 between lateral line and ventral. 


Lateral line incomplete, tubules on 8-15.scales. Coppery 
red, darker above, light below. Six or more dark spots 
along the side. L.R.S. 3 inches. | 

Baakens River, Port Elizabeth. Also from the Transvaal. 


224 


Genus CYPRINUS Linn. 


The common Carp of Europe, Cyprinus carpio Linn. has 
long been established in South Africa, where it not only 
thrives but often does so at the expense of our indigenous 
fishes. It is well known that it is unprofitable to liberate 
the young of species such as the Trout in waters containing 
Carp. 

Several curious varieties of the Carp exist, notably the 
“Mirror Carp,’ which has only a few very large isolated 
scales. This form is found in the Great Fish River. 

Carp are reputed to live to a great age, and to attain 
a large size; specimens weighing over 40 lb. are not un- 
common. The Carp does not rank high as an angling fish, 
being rather sluggish. When taken from clear water the 
flesh is quite palatable. Many farmers stock their dams 
with Carp, which generally multiply rapidly, even in com- 
paratively muddy or foul water. 


Genus LABEO Cuv. 


Body slightly compressed, with moderate to small 
scales. Mouth fairly large, overhung by the snout, the lips 
forming a sort of sucking disc. No true teeth, but there 
is a sharp horny cutting edge in each jaw. Barbels present 
or absent. Generally silvery or yellowish. 

Fishes of moderate size, usually found in turbid water. 
The flesh is soft and tasteless, while fine bones are very 
numerous. - 

Only two species have been found in the South-eastern 
Cape. 

Key to the Species. 


Two barbels on each side. 
I. 44-50 series of scales. capensis. 
II. 56-60 series of scales. umbratus. 


Labeo capensis A. Smith. [Pl. 31, Fig. 3.*]—Depth 
3-4, head 4-5 in body length. Eye 6-8 in head. Two 
barbels on each side, longest equal to eye. D. IIJ-IV. 10-11. 
A. I{I.5. Pectoral nearly as long as head. Ventral below 
8-9 
11-12 
white or pinkish below. L.R.S. 17 inches. 


dorsal base. 44-50 scales, Ltr. Olive-brown above, 


225 


Rivers and reservoirs, extending to Natal and Transvaal. 
When the Great Fish River comes down in flood many 
thousands of this species are thrown up dead on the shore 
near the mouth of the river. | 

Labeo umbratus A. Smith. [Pl. 31, Fig. 4.*]—Depth 
3°5-4:5, head 4-4-5 in body length. Eye 5-6 in head. Two 
barbels on each side, neither as long as eye. D.III. 9. 
A. III. 5. Pectoral nearly as long as head. Ventral below 
dorsal. 56-60 scales lL.tr. et 
14-16 
below. Small dark spots on upper part of side. L.R.S. 
10 inches. 

Rivers of the Cape Province, also in Natal and Transvaal. 


Brown above, yellowish 


Family BAGRIDAE. “ Cat-fishes,”’ ‘Barbels.’ A large 
family of numerous genera and species, found in the tropical 
and temperate waters of the world, generally of sluggish 
habit. They are easily recognised by the slimy scaleless body, 
together with the well-developed barbels, sometimes branched, 
and by the presence of a serrated spine in the pectoral fin. 
These fishes appear to be able to survive even prolonged 
droughts, and, it has been stated, may remain dormant in 
the dried mud from a pool, to revive when the water returns. 
Many of the species live for a considerable time out of water. 
Species of Clarias, if wrapped in moist cloth or straw, will 
survive long journeys. 


Only two genera have been recorded from the Eastern 
Province. 


Key to the Genera. 


I. A single dorsal fin, spineless. Clarias. 
II. Two dorsal fins, the first with a spine, the second adipose. Gephyroglanis. 


Genus CLARIAS Gron. 


Body elongate, with a single long spineless dorsal. Head 
depressed, bony. KEyesmall. Four pairs.of barbels. Minute 


teeth in jaws. | | 
Five species in South Africa, only one recorded from 


the Eastern Province. 

Clarias gariepinus Burch. “ Barbel” or “ Barber.” 
—Depth 5-7, head 3-3-8 in body length. Eye very small. 
Head granular above, with two median pits. One nasal one 


226 


maxillary, and two mandibular barbels, the latter three 
nearly as long as head. D. 65-80, originates above tip of 
pectoral. A. 50-60. Both fins reach nearly to caudal, which 
is free and rounded. Pectoral with serrated spine. Reddish 
or brown or slaty above, sometimes mottled lighter below. 
Attains a weight of over 100 lb. 

Extends from the Cape as far as Mozambique. 

This is our largest fresh-water fish, but is so sluggish 
as to be of little interest to anglers.. Usually taken on set 
lines, using worm or meat as bait. Very plentiful in the 
Orange River and tributaries. When those rivers are low 
natives secure numbers of fishes by spearing in the muddy 
bottoms of pools. Individuals over 100 lb. in weight are 
not uncommon. The flesh is palatable, but is something of 
an acquired taste, being reddish and rather tough. These 
fishes are very tenacious of life; after being exposed to the 
air during a whole day even in summer they will revive 
within a few moments when placed in water. After removal 
from the fish the heart will continue to beat for some time, 
especially if placed in water. 


Genus GEPHYROGLANIS Bler. 


Body moderately compressed. Two dersal fins, both 
short, the first normal, with a spine, the posterior adipose. 
Four pairs of barbels. Minute teeth in jaws. 

Only one species in South Africa. 

Gephyroglanis sclateri Blgr. [Pl. 33, Fig. 1.*]—Depth 
4-5-6, head 4 in body length. Eye 6-7 in head. Nasal 
barbel minute, maxillary and 2 mandibular barbels longer, 
but longest only half head length. First dorsal of one smooth 
spine and 7 rays. Adipose dorsal long and low. A. 16-17. 
Pectoral with serrated spine. Caudal free, forked. Dark 
brown above, lighter below. L.R.S. 15 inches. 

Eastern Cape to Natal and Transvaal. 


Family PLOTOSIDAE. A family of curious fishes, mostly 
marine and estuarine, of the Indo-Pacific region. Some 
species live in fresh water. 


Only one genus and species in South Africa. 

Plotosus anguillaris Bloch. “ Barbel eel.” [Pl. 34, 
Fig. 2.*]—Body more or less compressed, tapering posteriorly, 
scaleless. Depth 6-8, head 4-5-5 in body length. Eye 


227 


6-7 in head. Mouth large, with conical teeth in upper jaw. 
On each side one nasal, one maxillary, and two mandibulary 
barbels, moderately long. Two dorsal fins, the first short, 
with a serrated spine ; the second long, confluent round caudal 
with anal. Pectoral with a serrated spine. Brownish to 
darker above, slightly lighter below. Sometimes 2 or 3 
bluish or lighter longitudinal streaks. L.R.S. 16 inches. 

Occurs in the sea and also high up in tidal rivers, even 
to fresh water, from the Cape to Natal. 

These fishes should be handled with care, for the serrated 
spines inflict most painful wounds which may even prove 
dangerous. There are no actual poison-glands, but the 
mucus enveloping the spines appears to be highly toxic. 


Family ANGUILLIDAE. ‘Eels.’”’ Eels are well known and 

easily recognised. It has been shown by the late Dr. J. 

Schmidt of Copenhagen that only one species of eel, 

Anguilla mossambica Peters, occurs in the fresh waters of 
South Africa. 


Anguilla mossambica Peters.—Body very elongate, with 
minute scales embedded in the skin. Pectorals present. 
Dorsal and anal confluent round tail. Mouth large, snout 
pointed, teeth small. Brown or olive above, lighter below. 
L.R.S. probably 60 inches. 

Almost all fresh waters of the Cape. Extends right 
through to the Pacific region. 

It has been found that the adult eels of the Northern 
Hemisphere leave the rivers and seek out the depths of the 
sea near the West Indies, where the eggs are laid. The 
young hatch out at a depth of 100-150 fathoms, where they 
remain until they are about 1 inch long. At this stage 
the young eel, formerly known as Leptocephalus, is unlike 
the adult. The body is deeper, very highly compressed, 
and quite transparent, being not much thicker than paper. 
This stage last for about three years, during which time 
vast hordes of the young of the European species travel 
across the Atlantic, a large proportion naturally falling a 
prey to fishes and other creatures. The body now changes 
to the typical cylindrical form of the adult, and the “ elvers,”’ 
or young eels, ascend the rivers. It is believed that the 
spent adult eels die after spawning. 

The life-history of the African eel has not yet been worked 


228 


out. It has been suggested that it probably has a cycle 
similar to that of the European eel, and the deeps near 
Madagascar have been suggested as a possible spawning 
ground. On the other hand, evidence is being accumulated 
which indicates that our eels may have a totally different 
life-history, and it is not even improbable that they do not 
return to the sea at all. 

Leptocephalus stadia of many marine eels are fairly 
common in the sea, but no marine Leptocephalus referable to 
mossambica have yet been found; actually no larval stage 
of that species is known. The Madagascan-deep theory 
would mean that the Leptocephalus would be able to survive 
the sudden change from the warm waters of the east to the 
cold waters of the Atlantic, since it has been stated that 
mossambica has been found in the Orange River (though 
not yet in other rivers of the west coast). 

In the quiet coastal fresh waters of the south-eastern 
coasts of South Africa mossambica attains a large size, 
specimens weighing more than 50 lb. being not uncommon, 
which prove most formidable antagonists to the angler. 

This species has been observed to travel overland from 
one pool to another. 


Family CICHLIDAE. A family of numerous genera and very 
numerous species, especially abundant in tropical Africa. 
Most are moderate or small, but some attain a fair size. In 
the region of the Great Lakes of Central Africa certain species 
are of economic importance as food-fishes. 

The spawning habits of some species (e.g. genus Tilapia) 
are of interest. The eggs are laid in a rough nest, usually 
a hollow in mud or gravel, and after fertilisation are carried 
in the mouth of the female. Even after they have hatched 
the young are cared for by the mother, in whose mouth they 
seek refuge on the approach of danger. This is unusual, for 
in most other cases where parental care of the young is shown 

by fishes it is the male who undertakes this duty. 


Genus TILAPIA A. Smith. 


Body compressed, with fairly large scales. Mouth fairly 
large, with rows of bi- or tricuspid teeth, apically brown. 
Two lateral lines. 


229 


Of the 30 species present in South Africa only two have 
been recorded from the Eastern Province. 
In the Transvaal Tilapia species are known as ‘‘ Kurper.”’ 


Key to the Species. 


I. 16-20 gill-rakers: D.XV-—XVII. 10-12. natalensis. 
II. 9-12 gill-rakers: D.XIII-XV. 9-11. - sparrmani., 


Tilapia natalensis Weber. ‘‘ Mud-fish.”’ [Pl. 32, Fig. 1.*] 
—Depth 2-2-2-8, head 2-7-3 in body length. Eye 3-5 in 
head. Mouth extends almost to below eye. Teeth with a 
basal cusp, 3-6 rows in upper jaw. 16-20 gill-rakers. 
D. XV-XVII. 10-12, posterior rays forming a pointed lobe. 
A. III. 9-10, soft fin similar to dorsal. Caudal truncate. 

3—4 WW 17-21 
13-16" © 11-17 
above, lighter to dusky below. Dark blotch on opercle. 
Dark spots in two series along sides. Young with dusky 
cross-bars. Soft dorsal and anal with rows of dusky spots. 
L.R.S. 15 inches. 

Cape Province through Natal and Transvaal, as far as 
Rhodesia. 

This species is abundant in the brackish and fresh waters 
of the coastal region of the Eastern Province. The young 
may often be seen in shallow water among reeds or rushes. 
The flesh is palatable. 

Tilapia sparrmani A. Smith. [Pl. 32, Fig. 2.*]—Depth 
2-2:5, head 3 in body length. Eye 3-5 in head. Mouth 
large, extending almost below eye. Teeth with single basal 
cusp, 3-6 rows in upper jaw. 9-12 gill-rakers. D. XIII-XV., 
posterior rays forming a pointed lobe. A. III. 8-10, posterior 
rays similar to dorsal. Caudal subtruncate. Scales cycloid, 

2-3 |, 16-19 
27-29 series, |.tr. Qc1’ LI: 9219" 
with lighter margins, dusky below. Young with 7-9 dark 
cross-bars, which fade with growth. A blotch on the opercle. 
Soft dorsal, anal, and caudal with rows of dark spots. L.RS. 
6 inches. 

Comparatively scarce in the Eastern Province, but one 
of the most widely distributed species, being found from 
the Eastern Cape in most waters up to Rhodesia. 


Brownish 


Scales cycloid, 27-32 series, |.tr. 


Olive-brown above, scales 


230 


Genus HAPLOCHROMIS Pfeff. 


A genus of doubtful validity, very close to Tilapia. 
Some of the species have conical teeth, but the dentition 
varies with age. Only one species in our area. 

Haplochromis moffati Castlenau. [Pl. 32, Fig. 3.*]— 
Depth equal to head length, 2-5-2-7 in body length. Eye 
3-5-4-5in head. Teeth in 3-4 series, outer larger and conical, 
inner bi- or tricuspid. 7-10 gill-rakers. D. XITI-XYV. 9-11. 


3 
A. III. 8-10. Scales ctenoid, 26-31 series, l.tr. TET 
10-12 a . , 
Ll. Fc3° Olive or brown, with or without lighter dots, or 


with faint cross-bars; sometimes a lateral stripe. Soft 
dorsal, anal and caudal with series of light and dark spots. 
L.R.S. 5 inches. 

Extends from the Eastern Cape to the Congo. 


Family ANABATIDAE. Small fishes, found in Africa and 
Asia, characterised by the possession of an accessory breathing 
organ, a rudimentary “‘lung’’ in a cavity above the gills, 
which enables them to live for a considerable time out of water. 


Genus ANABAS Cuvier. 


Body scaly. A single long dorsal. Two lateral lines. 
One species Anabas testudineus (usually scandens), from 
India, is commonly known as the “‘ Climbing Perch,”’ from 
its supposed habit of climbing trees in search of food. 

These fishes do not attain any size, but they are car- 
nivorous and extremely voracious. They thrive even in small 
permanent pools. 

Only two species are recorded from the Eastern Province. 


Key to the Species. 


I. D. XII-XIV. 7-10; scales 26-29, 2 capensis 


9-10 
Il. D. XV-XVIT. 10; scales 31-35, ae bainsii. 


Anabas capensis C. & V. “ Kurper,” “‘ Rocky.” [Pl. 33, 
Fig. 3.*]—Depth 3, head 2-5-3 in body length. Mouth 
large, with small conical teeth. D. XII-XIV. 7-10. A. VI- 


231 


VIII. 8-10. Head scaly except snout. Scales ctenoid, 26-29 
34 il 15-18 
9-10° " 9-14" 
above, yellowish on sides and below. Radiating streaks from 
eye. Dark spot on opercle. L.R.S. 10 inches. 

A species vicinus Blgr. has been recorded from the 
Baakens River, Algoa Bay, but it is identical with capensis. 
Capensis extends from the Cape Peninsula throughout the 
Cape midlands; it occurs in waters round Grahamstown, 
but does not appear to extend east to the Great Fish River. 

A most pugnacious and voracious small fish ; a specimen 
was observed attempting to swallow a Barbus anoplus quite 
half its size. This species doubtless consumes a large pro- 
portion of the liberated fry of the Trout. 

Large numbers may be caught on almost any bait, and 
provide good sport for the juvenile angler. The flesh of the 
larger specimens is quite palatable. 

Anabas bainsii Cast. [Pl]. 33, Fig. 2.*]—Depth equal to 
head, 2-8 in body length. Mouth large, with small conical 
teeth. D. XV-XVII. 10. A. VI-VIII. 9-10. Caudal rounded. 
Head except snout scaly. Scales ctenoid, 31-35 series, |.tr. 

6—7 
13-15 
or more radiating streaks from eye. Dark spot on opercle. 
L.R.S. 6 inches. a 

Kingwilliamstown and East London area. Apparently 
rather a localised species, will probably be found elsewhere. 


series, l.tr. Caudal rounded. Olive-brown 


15 
Ll. 15 Dark brown above, dusky yellow below. One 


Family MUGILIDAE. “ Mullets,” ‘“Harders,” “ Springers.” 
Body compressed, scaly. Two dorsal fins. Shoal fishes, 
mostly of small to moderate size, though some species attain 
a length of over 36 inches. Some species are cosmopolitan, 
others circumtropical, and a few localised. Many are of 
considerable economic significance since they occur in vast 
numbers. They are largely herbivorous, some omnivorous. 
Most species are marine or estuarine, but a few live exclusively 
or partly in brackish or fresh water. — 


Only one genus in South Africa. 


Genus MUGIL Linn. 


The first dorsal of four spines. Some species with well- 
developed adipose eyelids. Scales moderate or large. Mouth 


232 


very small, teeth minute or absent. No lateral line, but 
most scales with pits or canals. Fourteen species are found 
in South Africa, but only two extend regularly into fresh 
water. 


Key to the Species. 


I. Large adipose eyelids, cover most of iris. cephalus. 
II. Adipose eyelids rudimentary. euronotus. 


TEXtT-FIie, 3.—A. Mugil cephalus, Linn. 3B. Mugil euronotus, A. Smith. 
The dots indicate lateral rows of scales. 


Mugil cephalus Linn. ‘ Harder,” ‘“ Mullet,” “ Springer.”’ 
[Text-fig. 3, A.]|—Depth 4-5, head 4in body length. Nostrils 
wide apart. Snout flat and wide. Adipose eyelids well 
developed, encircle pupil D.IV.+1, 7. A. III. 8. Pectorals 
short, with long pointed scale in axil. 39-42 series of scales, 
ltr. 14-16. Olive or brown above, silvery on sides and 
below. Sometimes streaks along sides. Axil of pectoral 
dark. L:R.S. 30 inches. 


Circumtropical. Extends well up tidal rivers and often 


233 


found in landlocked vleis and lagoons. In quiet waters 
specimens over 10 lb. in weight are not uncommon. This 
species does not usually appear to congregate in shoals. 

Mugil euronotus A. Smith. “ Springer.’ [Text-fig. 3, B.] 
—Depth 4, head 3-4 in body length. Nostrils close together. — 
Adipose eyelids rudimentary. D. IV.+1, 8. <A. III. 9. 
Pectorals short, no scaly process in axil. 43-45 series of 
scales, ltr. 14-15. Dark above, shading through dusky to 
silvery below. L.R.S. 12 inches. 

Occurs in most of the fresh waters of the coastal region 
of the Eastern Province, plentiful not far from Grahamstown. 
Angling for these fishes is a popular sport, special tackle 
being employed; the favourite lure is the “‘ flying-ant,’’ or 
the larger termite. The flesh of fishes from muddy water 
has a rank flavour. 


Family GOBIIDIAE. ‘Gobies.’’ Small carnivorous fishes, 
usually found in temperate and tropical seas, some entering 
fresh water. 


A single genus Gobius is found in our fresh waters. 


Genus GOBIUS Art. 


Ventrals (pelvics) united, forming a feeble sucking disc 
(by which the fish may adhere to stones or any smooth 
surface). Scales ctenoid, often obseured by a thick coating 
of mucus. Mouth large, teeth small, conical, sometimes 
posterior enlarged canines. Gill-openings restricted. Two 
dorsal fins, the first of six spines. Three species in the fresh 
waters of the Eastern Province. | 


Key to the Species. 


I. Lower jaw projecting beyond upper. 


A. Caudal pointed, longer than head. : callidus. 
B. Caudal rounded, not longer than head. giuris. 
Il. Upper jaw projecting over lower. aeneofuscus. 


Gobius callidus J. L. B. Smith. [Text-fig. 4.]—Depth 5, 
head 3 in body length. Eye 5 in head. Papillae in un- 
dulating rows on cheeks. Lower jaw projects slightly. 
D. VI.+1, 8, last soft rays as long as head. A.I1. 7, soft 
rays resemble soft dorsal. Pectoral as long as head. Caudal 
1-2 times head. Twenty-eight series of scales, l.tr. 8. Head 


234 


naked. Dark olive above, dusky below. A blotch on 
spinous dorsal, and spots on soft dorsal and caudal. L.R.S. 
4inches. Known only from the type, taken in the Bushman’s 
River, Alicedale. 


¢2 >. NO . 
ANY SNS 
eh SS : 
Se? > 


Trext-Fic. 4.—Gobius callidus J. L. B. Smith. 


Gobius giuris Ham. Buch.—Depth 4-5, head 3 in body 
length. Eye 4-7 in head. Papillae in rows on cheeks. 
Lower jaw projects strongly. Mouth extends below eye. 
D. VI.+1, 9, soft rays in second dorsal not nearly as long 
as head. A.I.8, soft rays similar to dorsal. Pectoral 
shorter than head. Caudal shorter than or equal to head. 
30-36 series of scales, ltr. 9-11. Scales on top of head. 
Brown-olive, with or without blotches on sides. Pectoral, 
dorsal, and caudal spotted or barred. L.R.S. 15 inches. 

Equally at home in the sea or in fresh water. One of the 
most widely distributed species. Occurs from about Knysna 
eastwards along the coast to Madagascar, India, and China. 
In all estuaries and lagoons and in many inland waters in 
South Africa. This is the largest of the Gobies. It breeds 
freely in captivity, and in India is kept in tanks and ponds, 
and is eaten with relish by the natives. | 

Gobius aeneofuscus Peters.—Depth 5-6, head 3-5 in body 
length. Eye 4-7 in head. Mouth extends to below eye, 
upper jaw extends beyond lower. D. VI.+1, 10; A.I. 10. 
Caudal shorter than head. 58-64 series of scales; L.tr. 
16-18. Head naked. Olive-brown above, bronzy on sides, 
light below. Obscure mottlings on body. Two oblique lines 
from eye to mouth. Soft dorsal and caudal spotted or 
barred. L.R.S. 10 inches. 

Fresh waters of the Eastern Cape, extending through 
Natal, Transvaal to Rhodesia. Also in fresh waters of 
Madagascar. 


235 


Family CLUPEIDAE. Herrings and relatives. 


Genus GILCHRISTELLA Fowler. 


Compressed, almost transparent body. Scales soft, 
easily shed. Pectorals low down. Ventrals before dorsal. 
A single short dorsal, anal short or long. Teeth minute or 
absent. Adipose eyelids absent. A silvery lateral stripe. 

Gilchristella aestuarius Glch.—Depth equal to head, 4 in 
body length. Eye 3 in head. No adipose eyelids. Lower 
jaw projects, mouth extends to eye. D. 14-16, inserted 
nearer caudal base than snout tip. A. 18-21, originates 
below end of dorsal. 40 series of scales; Ltr. 10. Caudal 
forked. Silvery, back darker, brilliant silvery lateral stripe. 
L.R.S. 3 inches. 

This small Clupeid occurs in the sea, but is found mostly 
in estuaries. It occurs also in landlocked brackish and fresh- 
water lagoons. These fishes swim in large shoals, and are 
preyed upon by numerous aquatic birds. The fishes in the 
shoals have a curious habit of swimming with the head up 
and the snout just out of the water. On the approach of 
danger there is a silvery flash and the whole shoal dives a 
few feet, to reappear a few moments later at the surface. 
Sometimes known as “ Whitebait.”’ 


Family MONODACTYLIDAE. Small, deep-bodied, compressed 

fishes of tropical and temperate waters, some of which are 

equally at home in the sea and in fresh water. A single dorsal. 
Bright silvery, with dark cross-lines on young. 


Genus MONODACTYLUS. 


Body deep, scaly. Mouth small, protractile. Colour 
silvery. One species extends from the Cape eastwards, 
found in the sea and in fresh waters of the Kastern Cape 
and beyond. 

Monodactylus falciformis Lacep. ‘‘ Moony,” ‘“ Cape 
Lady.” [Pl. 34, Fig. 1.*]}—Body ovate, deeper in the adult, 
depth 1-5-2, head 3 in body length. Eye 2in head. D. VIII. 
27-30. A. III. 27-30. Front rays of soft dorsal and anal 
elevated. Ventrals very small. 1.1. scales 50-60. Bright 
silvery. Lobes of soft dorsal and anal dusky. Young with 
numerous narrow dark cross-bars. L.R.S. 10 inches. 


236 


Fresh waters of the Eastern Cape, even near Grahams- 
town. Often caught by “Springer” anglers. The flesh is 
rather insipid and deteriorates rapidly. 


Family GALAXIIDAE. Small fishes of the Southern 
Hemisphere, fairly closely related to the Salmonidae (of the 
Northern Hemisphere). 

Some species live and reproduce in fresh water, others 
are catadromous, 7.e. return to the sea to spawn, the reverse 
of some of the Salmonidae, which ascend rivers to spawn 

(anadromous). 
Genus GALAXIAS Cuvier. 


Body elongate. A single short dorsal fin. Mouth 
moderate, with small teeth. 

Only one species from the South-eastern Cape. 

Galaxias zebratus Cast. [Pl]. 34, Fig. 3.*]|—Depth 5-6-5, 
head 3-5-4 in body length. Eye 4 in head. Mouth extends 
to below eye. Teeth of more or less even length. Lower 
jaw projects slightly. D. III-IV. 7-8, base shorter than head 
length. A. III-IV. 7-8. Spines in both fins feeble. Caudal 
slightly rounded. No lateral line. Grey-brown, mottled to 
form irregular cross-bars. L.R.S. 24 inches. 

Inhabitants of clear fresh waters, plentiful in the Cape 
midlands, scarce farther east. 


Family SALMONIDAE. ‘“ Salmon,’ ‘‘ Trout.’’ One species 
of these carnivorous fishes from the Northern Hemisphere 
has long been established in certain parts of South Africa. 
The fishes thrive in the clear waters of the Western Cape, and 
in the mountain streams of the Eastern Cape and Natal. 

The somewhat impermanent turbid waters of the South- 
eastern Cape do not suit them as well, and even in large 
permanent reservoirs they have not been as successful as was 
anticipated. 

T'wo sub-species Salmo fario fario Linn, the Brown Trout, 
and Salmo fario irideus Gibbons, the Rainbow Trout, have 

been firmly established. 


Genus SALMO Linn. 


Body elongate. Two dorsal fins, the posterior adipose. 
Mouth large, teeth conical. Body with a heavy coating of 
mucus. | 


237 


Two sub-species in South Africa. 


A. Caudal truncated. Dark spots on body. fario fario. 
B. Caudal slightly emarginate. A longitudinal purplish stripe. fario irideus. 


Family CENTRARCHIDAE. The large-mouth Black Bass, 

Micropterus salmoides Lacep., is a recent importation. It has 

been introduced into waters in numerous parts of South 

Africa and, where records are available, appears everywhere 
to flourish. 


Micropterus salmoides Linn. ‘‘ Black Bass.’? Depth 
about 3 in body length. Mouth large, extending below eye. 
Spinous dorsal short, almost separate from soft fin, 3rd and 
4th spines longest, hinder spines shorter. Soft dorsal and 
anal rounded, not elevated anteriorly, of 11-12 rays. Caudal 
rounded truncate. Superficially resembles the ‘“‘ Kurper,”’ 
but has a deeper body, only one lateral line, and a shorter 
spinous dorsal. | 

Reported to attain a weight of 20 lb. (America) ; average 
weight 3-4 lb. Already firmly established in several parts 
in the South-eastern Cape. This species spawns fairly 
young, probably from the age of 2 or 3 years, and the eggs 
and fry are guarded by the male parent. 

The Black Bass is an active predaceous fish, which 
appears to take almost any bait or artificial lure. It has 
been found that while the Trout may almost be prevented 
from multiplying by the voracious Kurper (Anabas capensis), 
that latter species may actually be exterminated by the more 
voracious Black Bass. 


PLATE XXIX. 


a 
Sau 
_ 


(From ** Catalogue Fresh-Water Fishes of Africa,’ British Museum.) 


(1) Barbus trevelyani Gnthr. (p. 123); (2) Barbus anoplus Weber 
(p. 124) “ Red-fin” ; (3) Barbus gilchristi Blgr. (p. 126); (4) Barbus 
hemipleurogramma Blegr. (p. 127). 


PLATE XXX. 


From “‘ Catalogue Fresh-Water Fishes of -\frica,” British Museum.) 


Barbus capensis Smith (p. 125) “ Silver-fish ”’ ; Barbus 
burchelli Smith (p. 127); Barbus brookingi Glch. (p. 125); 
Barbus paludinosus Peters (p. 125) ‘‘ Gilliminkie.” 


PuateE XXNI, 


(From ‘ Catalogue Fresh-W ater Fishes of Africa,” British Museum.) 
(1) Barbus holubi Stndnr. (p. 126) ‘‘ Yellow-fish”’; (2) Barbus 


vulneratus Cast. (p. 127); (3) Labeo capensis Smith (p. 128); 
(4) Labeo umbratus Smith (p. 129). 


SS Sa ahah f Sane 
Nun ania 3 ‘ : Reon 


sia 
at : 
Haat 
ae 


ee 
une ana 


: 


(From “ Catalogue Fresh-Water Fishe 


1) Tilapia natalensis Weber (p. 133) ‘‘ Mud-fish”’; (2) Tilapia 
sparrmani Smith (p. 133); (3) Haplochromis moffati Cast. (p. 134). 


of Africa,’ British Musenm.) 


Pirate XXXII. 


ee 


[CS om 
Paes oo : i a 
|. _. : 


. 
: 


a 
A 


, 
: _ 


ee 
Ra e 
an 

| 


i a i 
me 


(From ‘ Catalogue Fresh-Water Fishes of Africa,” British Museum.) 


— (1) Gephyroglanis sclateri Blgr. (p. 130) ‘‘ Barbel”’; (2) Anabas 
bainsii Cast. (p. 135); (3) Anabas capensis C. & V. (p. 134) ‘‘ Kurper.” 


Prats XXXIV. 


(From ‘Catalogue Fresh-Water Fishes of Africa,’ British Museum.) 


(1) Monodactylus falciformis Lacep. (p. 139) “ Moony”; (2) 
Plotosus anguillaris Blch. (p. 130) ‘‘ Barbel-eel”’; (3) Galaxias 
zebratus Cast. (p. 140). 


238. 


Records of the Albany Museum. Vol. IV. Part II. pp. 358-364. 
Pls. XL—XLII. May, 1935. 


The South African Species of the Family Aluteridae. 


By J. L. B. SMITH. 
[With Plates XL—XLII.] 


The highly specialised Division Sclerodermi has been 
critically examined by comparatively few workers. The most 
comprehensive investigation appears to be that of Regan (Proc. 
Zool. Soc. Lond. 1902, vol. II, p. 284 ff.). Other works, to which 
I have access, deal only with regional fauna. The relationships 


of the various groups do not appear to be well understood, and 
the Division seems badly in need of revision. 

Regan’s classification of the groups was based chiefly upon 
internal features, but evidently he did not examine sufficient 
material to ensure that his generalisations were applicable to 
all the genera. At the same time, he stated that any attempt 
at further subdivision of the family Balistidae into new families 
would probably necessitate the raising of practically every genus 
then recognised to family rank. 

In recent work on these fishes, in addition to the Balistidae, 
the family Monacanthidae has been generally recognised. 

A graduated series of stadia of Aluteres monoceros Osb. 
has recently been obtained from South African waters, and work 
on this has led to a critical examination of the genus Aluteres 
Cuv. This has hitherto been accepted as falling in the Mona- 
canthidae (or by many, Monacanthinae). While Aluteres has 
most probably originated from the Monacanthidae, the juveniles 
showing many typically Monacanthid features, there are never- 
theless sufficient grounds to justify the separation of Aluteres 
from the Monacanthidae, with at least sub-family rank, at least 
as far as South African material is concerned. 


Mr. Fraser-Brunner, at the British Museum, who has under- 
taken the much-needed revision of the whole order Plectognathi, 
has informed me that my conclusions are in line with those to 
which he had independently come, and has urged the publication 
of the results embodied in the present work. It is now suggested 
that Aluteres be raised to full family rank, the Aluteridae. 


239 


South African Fish.—Smith. 359 


FAMILY ALUTERIDAE. 


Supraclavicle vertical. Precaudal vertebrae with moderate 
parapophyses, without epipleurals. No epipleurals to caudal 
vertebrae. Prezygapophyses of the much-reduced first vertebra 
ankylosed with the occipital condyles. Dentary fused with 
articular. Premaxillae not protractile, united with maxillae. 
Palatines free from ectopterygoid. Preorbital ossified. Ethmoid 
region very elongate, without nasal cavities. Spinous dorsal of 
two spines, probably of secondary origin, the first weak, with 
posterior basal dilation: the second rudimentary, with anterior 
basal dilation, which can lock the first in the erect position. Soft 
dorsal and anal long and low. A large oblique ossified sub- 
cutaneous process in advance of the basipterygials of the soft 
dorsal. (Pl. XLI, b.) Pelvis represented by a single long bone, 
attached to the pectoral arch, more or less movable. Ventrals 
rudimentary, immovable, probably dermal in origin, present in 
juveniles as a minute stellate spine, inserted below about the 
middle of the pelvis: become reduced or obsolete in large adults. 
Compressed incisor-like teeth in both jaws. The air-bladder has 
on each side a sub-conical posterior process, which penetrates 
some distance into the caudal region. 

Vertebrae 22 (7 +15): the haemal process of the first caudal 
vertebra much elongated, reaching to near the ventral margin, 
apically dilated. 

The entire absence of any trace of epipleurals, the presence 
of the ventral spine, and the nature of the air-bladder, do not 
appear to have been previously noticed, while it has hitherto been 
accepted that the number of vertebrae is 21. 

The large ossified process immediately anterior to the 
basipterygials of the soft dorsal, which is present also in those 
species of the Balistidae and of the Monacanthidae which I have 
seen, is possibly the coalesced basipterygials and fin spines of an 
obsolete spinous dorsal. | 

It is noteworthy that the bases of the fin-rays of the soft 
dorsal and anal fins do not articulate with the apical dilations 
of the basipterygials, but are widely separated from these, by a 


240 


360 South African Fish.—Smith. 


cartilaginous rod, for the entire length of each fin (Pl. 2, a). 
This rod is elliptical in cross-section, and has on each side, 
opposite each fin-ray, a groove, in which lies the muscle con- 
necting the rays with the basipterygials. Part of the anterior 
end of the dorsal rod fits into a supero-posterior excavation in 
the ossified predorsal process described above. 

The entire absence at all stadia of epipleurals,* the obsoles- 
cent ventral spine, and the peculiar structure of the hydrostatic 
organ, none of which is characteristic of the Balistidae or of 
the Monacanthidae, would appear to justify the separation of 
Aluteres from these families. 

It is possible that some of the relatéd genera, such as 
Pseudaluteres, may also be found to fall in the Aluteridae, but I 
have no material to determine this point. 

It is noteworthy that juveniles of this family show Several 
typical Monacanthid external features, which are partly or wholly 
lost or changed with growth. 


Genus Aluteres Cuv. 


1902. Regan, loc. cit. p. 290. 
1908. Jordan and Fowler, Proc. U.S. Nat. Mus.,.vol. 25, 
p. 274 and p. 275 (Osbeckia). 

1925. Barnard, Ann. S.A. Mus., voi. X XI, p. 960. 

With the characters of the family. 

Gill-rakers normal, small, lanceolate, fairly widely spaced. 
The skin is covered with minute close-set cilia (Pl. 3, E), so as 
to be soft and velvety to the touch. 

Caudal rounded or truncate (adults). 

The only two South African species, monoceros Osbeck and 
scriptus Osbeck, appear to be quite clearly congeneric. Jordan 
and Evermann (Check-List Fishes, 1896, p. 424, fide Barnard 
loc. cit.) proposed the genus Osbeckia for species with concave 
snout profile and caudal longer than head. These slender dis- 
tinctions, especially in view of the evidence here adduced, are 
inadmissible as of generic significance. 


*It may be remarked that in one case a single presumed epipleural 
to the 5th precaudal vertebra was found on one side only. 


241 


South African Fish.—Smith. 361 


In so far as can be determined, both species are valid. They 
are so closely related, especially in the juvenile. stadia, that the 
possibility of sexual dimorphism might be suspected. I have no 
sexually mature adults of either species, but in so far as I have 
been able to determine, among the adults of monoceros which I 
have examined are males and females, so that both species may 
be accepted. 

Numerous nominal species of Aluteres have been described, 
but many of these are obviously merely different growth stadia. 

In the case of monoceros at least, there is very considerable 
change in general features with growth, as will be described 
below. In the case of scriptus, these cannot be described, since I 
have seen only one specimen, but there is reason for presuming 
that at least similar growth changes occur in this species (vide 
infra). 

KEY TO THE SOUTH AFRICAN SPECIES. 
I. Caudal not longer than head; rounded or 
truncate (adult) = .. . ; ..  monoceros. 
II. Caudal longer than head, rounded . .. seriptus. 
Aluteres monoceros Osb. (Pls. XL—XLIL) 
? *Cantor, Cat. Mal. Fish, p. 353 (obliteratus). 


1888. Day, Fishes of India, p. 698. Pl. CLXXIX, fig. 2 
(synonymy). | 

1903. Jordan and Fowler, loc. cit., p. 274. 

1925. Barnard, loc. cit., p. 960. 

1928. Fowler, Fishes Oceania, p. 461. 

This species has so often been described that repetition of 
the obvious features is unnecessary. The more unusual features 
and the growth changes are described below. | 

The following table indicates the change with growth of 
certain of the dimensional relationships. 


Length, mm. a .. 68. 95. 123. 135. 300. 595. 
Depth in length of body 2.4 24 24 26 29 3. 
Eye in head.. a .. 40 5.0 5.0 5.0 6.0 6.5. 
Caudalinhead .. .. — 18 14 17 2.0 2.6 


* I have received a copy of this description from Mr. Norman, of the 
British Museum, 


242 


362 _ South African Fish.—Smith. 


The chief growth changes are the following :— 

Shape of body: In the very young the body is fairly deep, 
with the greatest depth across the shoulder. The breast is 
strongly convex. The body tapers more or less uniformly from 
the head to the caudal. With growth, the convexity of the breast 
becomes less, till in old adults there is a marked angle below the 
chin at the breast. The cuneate shape is modified in that the 
body becomes deeper posteriorly, until it is about of even depth 
from the first to the second dorsal. (Pl. XL, A-D.) 

Dorsal spine: In the young the first spine is relatively more 
elongate, and has four series of fairly large barbs (Pl. XLII, C). 
With growth the spine becomes relatively shorter, more slender, 
and the barbs degenerate into granulations. This spine is 
frequently deformed. 

Caudal fin: In the young the fin is relatively more elongate 
and strongly rounded. With growth, the fin becomes shorter, and 
the hind margin less convex, until finally the truncate fin of the 
adult results. 

Ventral spine: In the very young the ventral spine is fairly 
prominent, stellate (Pl. XLII, D). With growth, the apical pro- 
jections diminish, until in adults the whole is reduced to a 
minute more or less translucent knob, which can nearly always 
be located by loosening the skin of the hinder third of the pelvis 
and viewing against a light. In the very smallest specimens there 
appear to be signs of some connection between the base of the 
spine and the pelvis, but in all the other stadia, it is unquestion- 
ably dermal. The significance of the dermal origin of the spine 
was pointed out to me by Mr. Fraser-Brunner, who is dealing 
fully with the matter. 

Profile of snout: In the young the dorsal profile of the snout 
is fairly concave. With growth, the concavity lessens, until in 
large adults the profile is convex. 

Gill-rakers 22-25 on the lower part of the anterior arch, 
about 6 in gill-filaments, which are rather longer than the eye. 

The dorsal originates opposite or behind the anal. 

This species occurs on our coasts from the Cape eastwards. 
I have seen specimens from the Cape; Mossel Bay; Knysna: 


243 


South African Fish.—Smith. 363 


Port Alfred: Great Fish Point; East London; Durban, and 
Delagoa Bay. 


Aluteres scriptus Osb. (Pl. XLII, A.) 


1888. Day, loc. cit., p. 694. Pl. CLXXX, fig. 3. 

1908. Jordan and Fowler, loc. cit., p. 276 (Osbeckia 
scriptus) . 

1925. Barnard, loc. cit., p. 961. 

1928. Fowler, loc. cit., p. 461. Pl. XLVII, B. 


I have seen only one specimen of this species, 113 mm. in 
length, from Durban. From this, and from figures, it would 
appear that growth changes occur also in this species. 

In the juveniles, the maximum depth is through the breast, 
which is strongly convex. This convexity apparently diminishes 
with growth, although the body does not appear to lose the 
uniform tapering from the first dorsal to the caudal. 

The caudal apparently always remains rounded and longer 
than the head (1.2 times head), and the dorsal profile of the 
snout is always concave. 

At equivalent stadia, the body is always less deep than that 
of monoceros. | | 

The ventral spine resembles that of monoceros. The dorsal 
spine in my specimen (Pl. XLII, B) shows minute barbs set 
among granulations. In my specimen, the dorsal originates in 
advance of the anal. 

Gill-rakers 28 on the lower part of the anterior arch, about 
4 in gill-filaments, which are equal to the eye. 

This species is very seldom encountered on our shores. 


I wish to express my gratitude to the Carnegie Research 
Fund (through the Research Grant Board of South Africa) for 
financial assistance, which has defrayed a great part of the costs 
incurred in the investigation. Also to Dr. J. A. Wain of Cathcart, 
for numerous radiographs. | 

Plate XL. 

Four stadia of Aluteres monoceros Osb. The small arrow 

indicates the position of the ventral spine, 


244 


364 South African Fish.—Smith. 


Plate XLI. 
Aluteres monoceros Osb. Radiograph. 
a. Cartilaginous rod between fin-rays and basipterygials. 
b. Predorsal process. 
Plate XLII. 
A. Aluteres scriptus Osb. The arrow indicates the position of 
the ventral spine. 
B. First dorsal of same; enlarged. 
b. second spine. 
C. First dorsal of a juvenile Aluteres monoceros Osb. 
a. second spine. 
D. Ventral spine of same specimen (135 mm. in length). Pos- 
terior margin to the right. 
E. Portion of skin of the same specimen, showing cilia, 


Rec. Alb. Mus., Vol. IV. Plate XL 


Aluteres monoceros Osb. 


(The small arrow indicates the ventral spine.) 


Rec. Alb. Mus., Vol. IV. | Plate XLI. 


lem, 
teas | 


Aluteres monoceros Osb. 


a. Cartilaginous rod below dorsal and anal. 
b. Ossified predorsal process. 


Rec. Alb. Mus., Vol. IV. Plate XLII. 


D  ,__ 


A. Aluteres scriptus Osb. B.. Dorsal spine of same. 
C. Aluteres monoceros Osb. Dorsal spine. D. Ventral spine. EK. Portion of skin. 


245. 


Transactions of the Royal Society of South Africa. Vol. XXIII. 
Part IV. pp. 303—310. Pls. XXI—XXII|. February, 1936. 


THE GENUS TRIPTERODON PLAYFAIR. 


By J. L. B. Smira. 
(With Plates XXI-X XIII and one Text-figure.) 


(Read June 19, 1935.) 


The genus Tripterodon Plyfr. with the single African species orbis Plyfr. 
has hitherto not received more than brief attention, and the features 
which would indicate its taxonomic relationships do not appear to have 
been investigated in any detail. 

A careful examination of the skeletal and structural features of this 
genus has led to a comparison with the same features in the obviously 
related genera Platax Cuv. and Drepane Cuv.; with these is allied 
Evhippus Cuv. It has not been possible to obtain a specimen of the latter 
for dissection, but it is obvious from accounts dealing with the internal 
structure, as well as from the external features in a specimen examined, 
that this genus and the three others mentioned form a natural group. 

Although Platax, Ephippus, and Drepane are among the earliest de- 
scribed genera of fishes, there is some diversity of opinion among sys- 
tematists as to their taxonomic relationships. Regan (Ann. Mag. Nat. 
Hist., 1913 (8), vol. xii, p. 127), who based his opinions chiefly upon skeletal 
features, placed Platax with Ephippus in the Ephippidae, and separated 
Drepane as the type of the family Drepanidae; in this latter case he has 
not given any detailed reasons for his conclusion beyond that Drepane 
has no subocular shelf, while stating that in all other skeletal features the 
two families are identical. Plataz, with a feeble subocular shelf, in this 
respect falls midway between Drepane and Ephippus, the latter having 


a strong shelf. 
Barnard (Ann. S.A. Mus., 1927, vol. xxi, pp. 600 and 605) followed 


Regan’s classification. 

Fowler (Proc. Ac. Nat. Sci. Phil., 1925, vol. Ixxvu, p. 251) at first 
accepted Regan’s classification for Drepane, but has more recently (Bull. 
U.S. Nat. Mus., 1929, vol. viii, p. 25; and Proc. Ac. Nat. Sci. Phil., 1934, 
vol. lxxxvi, p. 479) accorded Drepane only sub-family rank in the Ephippidae, 
while separating Plataz as a monotypic family. In the absence of more 
precise reasons, based on skeletal data, for the recognition of the Drepanidae 


246 


304 Transactions of the Royal Society of South Africa. 


by Regan, Fowler’s later classification, although obviously based on external 
features only, might appear almost justifiable. 

The genus Tripterodon Plyfr. has been somewhat arbitrarily placed in 
various families. Fowler (Proc. Ac. Nat. Sci. Phil., 1925, vol. lxxvu, 
p. 242; and Bull. U.S. Nat. Mus., 1933, vol. xii, p. 202) placed Tripterodon 
in the Girellidae, whereas more recently (Proc. Ac. Nat. Sci. Phil., 1934, 
vol. Ixxxvi, p. 478) he has described the single species of that genus as a 
Chaetodipterus Lac., placed with Drepane in the Ephippidae, and has thus 
indirectly recognised that Tripterodon is closely related to Ephippus and 
to Drepane. Barnard (loc. cit., p. 725) placed Tripterodon in the Sparidae, 
and also confused Tripterodon with Chaetodipterus (loc. cut., pp. 603, 604) 
by accepting what have proved to be somewhat excusable malidentifica- 
tions by earlier workers on South African fishes. 

Examination has shown that the inclusion of T7'r:pterodon in the Girellidae 
has little to recommend it. Fowler has probably regarded the dentition 
as a feature of paramount significance, but there is little else in common 
and so much divergence in important features that this diagnosis cannot 
be accepted. 

In the majority of skeletal features which systematists have hitherto 
regarded as of chief significance, the four genera Plataw, Ephippus, Triptero- 
don, and Drepane are so closely allied that they might easily be regarded 
as all falling within one family. Nevertheless Drepane is distinguished 
by several features, notably the structure of the upper jaw, which, although 
apparently not regarded as important by earlier workers, would appear 
to justify separation from the remaining three genera by full family rank. 
According to the degree of taxonomic significance assigned to these special 
characteristics of Drepane it is not even unlikely that the Drepanidae 
may ultimately be raised to the rank of a division, with probable inclusion 
of the Chaetodontidae. 

With regard to the classification of the remaining three genera there 
are several alternatives, but the selection of any one of these will very 
largely depend upon the personal interpretation of the taxonomic signi- 
ficance of the relatively few features in which these genera differ one from 
the other. They may be regarded as each forming the type of a separate 
family, although the separation of Platax from the remaining two, as 
proposed by Fowler (loc. cit.), merely upon the somewhat abnormal external 
features of that genus, does not appear to be justified, and is not accepted 
here. Should, however, this distinction for Platax otherwise meet with 
general approval, T'ripterodon would then appear to have at least equal 
claims to recognition as the type of a separate family. It may here be in- 
dicated that Fowler’s key to the genera of the Ephippidae (Bull. U.S. Nat. 
Mus., 1929, vol. vill, p. 24) does not agree with his diagnosis of the family. 


247 


The Genus Tripterodon Playfair. 305 


The strong subocular shelf and the setiform dentition are distinctive 
characters which might reasonably be regarded as justifying full family 
rank for E'phippus as opposed to the closely related Platax and Tripterodon. 
Nevertheless, when all the available evidence is carefully weighed, it 
appears to be more in keeping with generally accepted limits to adopt 
the classification here proposed, i.e. in view of the general agreement in 
the majority of important features, the genera Plataz, Ephippus, and 
Tripterodon be placed in a single family, the Platacidae.* The special 
features of Ephippus are then assigned only sub-family rank, while Platax 
and Tripterodon are placed together in the sub-family Platacinae, as 
arranged below in the key to the genera of the family Platacidase. 

The following summary indicates the main features in which the 
Drepanidae may be regarded as well distinguished from the Plataeidae. 


I. Anterior maxillary processes produced, meeting beneath the anterior 

margin of the mesethmoid, over the premaxillary pedicels. Pre- 

maxillaries slender, highly expanded distally above and below. 

Premaxillary pedicels very long, oblique, entering the skull 

beneath the olfactory sacs, sliding in a deep excavation extending 

between the lower orbital margins. Posterior ribs extensively 

articulated with the parapophyses. Gill-membranes narrowly 

fused with isthmus. Pectorals much longer than head. Scales 

cycloid. (Caudal more or less rounded) . . . Drepanidae. 
II. Maxillaries normal, not meeting anteriorly over the premaxillary 

pedicels. Premaxillaries not or scarcely expanded distally: Pre- 

maxillary pedicels not penetrating the skull, entirely anterior to 

the orbital margins, sub-vertical. Posterior ribs not or scarcely 

articulated with the parapophyses. Gill-membranes broadly 

fused with isthmus. Pectorals shorter than head. Scales ctenoid. 

(Caudal truncate or nearly so) . . . . . . . Platacidae. 


In most other respects the families agree. 

The curious shape and disposition of the premaxillaries of the 
Drepanidae reveal the connection between that family and the Chaeto- 
dontidae. In the latter, although the premaxillary pedicels are fairly 
short, they are very oblique, and penetrate below the olfactory sacs. fn 
Chaetodon setifer Blch. there does not appear to be the extensive articula- 
tion of the ribs with the parapophyses indicated by Regan (loc. cit.) as 
characteristic of the family. 

It has earlier been indicated (Smith, Proc. Roy. Soc. 8.A., 1935, 
vol. xxiii, p. 267) that the genus Dichistius Gill. reveals affinities with the 
Platacidae, and connects that family with the natural group of the Scor- 
pididae and the Kyphosidae. 


* Platax (Cuv. Reg. Anim., 1817, vol. ii, p. 334) has page preference over Ephippus 
(ibid., p. 335), and should therefore determine the name of the family. 


248 


306 Transactions of the Royal Society of South Africa. 


Famity PLATACIDAE. 


Body compressed, elevated or ovate. Front profile steep. Mouth terminal, small, 
moderately or scarcely protractile, more or less horizontal. Maxillary moderate, without 
anterior forward expansion. Premaxillary pedicels fairly long. Jaws with bands of 
movable teeth, slender and setiform, or compressed and notched. Palate edentate, 
vomer sometimes feebly dentate. Preopercle not, finely, or coarsely, serrate (spinate in 
young?). Gill-membranes united, broadly fused with isthmus, openings lateral. Gill- 
rakers short, few. Branchiostegals 6 (or 7?). 

Dorsal long, of a few spines, soft dorsal longer than spinous. Soft fin scaly. Below 
the skin a stout antrorse procumbent spine, fused with the basipterygial of the first 
two exposed spines; anterior to this, three obsolescent antrorse spines, the anterior 
(inferior) overlapping the supraoccipital apex (Pl. XXII). 

Anal of three spines, with small antrorse recumbent spine on anterior basipterygial; 
soft fin scaly. Pectorals short. Ventrals fairly or very long, with axillary process. 
Caudal of 17 principal rays, of more or less truncated form. Scales ctenoid, small or 
moderate. Lateral line gently curved. 

Vertebrae 24 (10+ 14), with parapophyses from the 4th. Anterior ribs expanded 
proximally. An elevated occipital crest, anterior margin expanded. No parietal crests. 
Subocular shelf feeble or strong. Air-bladder posteriorly bifurcated into two caudal 
prolongations (Pl. XXIII). 

The three genera here mentioned occur in the Indo-Pacific. 


Key to the Genera. 


I. Ephippinae.—Subocular _ shelf Strong. Teeth  setiform. (Scales 
fairly large) . ; ; ; . . Ephippus. 

II. Platacinae. —Subocular shelf very feeble. Teeth compressed, tri- 

cuspid. (Scales moderate or small.) 

A. Dorsal spines graduated, none longer than longest soft rays. 

More than 50 series of scales. Preorbital not very deep. 
Premaxillary pedicels shorter than rami ; . Plataz. 

B. 3rd-5th dorsal spines elongate, longest longer than posterior 

spines and than longest soft rays. Less than 50 series of 

scales. Preorbital very deep. Premaxillary pedicels as long 
as rami . . . . . . . . . .  Lripterodon. 


Genus Epruiprus Cuv. 


I have seen only a single specimen of the species orbis Bloch, from India. 
This has not yet been found in the South African region, but will probably 
be discovered there with more intensive collecting. It occurs in the Indian 
region, but appears to be nowhere very plentiful. 

In external form, chiefly in the shape of the fins, this species resembles 
Tripterodon orbis Plytr. closely, but has a shallower preorbital and the 
body is more uniformly orbicular, while the setiform teeth are immediately 
diagnostic. As is shown below, the species goreensis C. and V. from the 
West Coast of Africa shows even more remarkable resemblance to TJ. orbis 
in external form. 


249 


Lhe Genus Tripterodon Playfair. 307 


Genus Puatax Cuv. 


Premaxillary pedicels short, almost vertical. Supraoccipital and anterior dorsal 
basipterygial becoming greatly enlarged with age. Preopercle margin entire in adult. 


Sometimes a few fine teeth on vomer. Teeth rather slender, but apically somewhat 
dilated and sharply tricuspid. 


On casual examination the teeth appear setiform. The dentition in 
this genus is intermediate in form between that of Ephippus and that of 
Tripterodon. 

Fowler (Bull. U.S. Nat. Mus., 1929, vol. viii, p. 17) has stated that 
the air-bladder is simple. In those specimens I have examined the air- 
bladder has a median longitudinal septum, and is bifurcated into two 
caudal extensions, which are more pronounced in large specimens. 

There appear to be fairly extensive growth changes and many nominal 
species have been proposed, but the majority of recent workers appear 
to accept only two, pinnatus Linn. (tera Cuv.=) and orbicularis Bloch 
(vespertilio Cuv.=). Diagnoses of these species do not always agree, and 
there appears to be some confusion with regard to the features upon which 
differentiation is based. 

As far as South African material is concerned, positive identification 
with either species, 2.¢e. if both are recognised, is difficult. 

It would appear not unlikely that there is but a single rather poly- 
morphous species, though I have not sufficient material to enable me to 
decide this. 


Genus TRIPTERODON Plyfr. 


1866. Playfair, Fishes of Zanzibar, p. 42. 
1927. Barnard, loc. cit., p. 725. 
1933. Fowler, Bull. U.S. Nat. Mus., vol. xii, p. 202. 


Premaxillary pedicels as long as rami, but not reaching frontals, almost vertical; 
superior lateral processes very small. Maxillaries normal, short and deep, moderately 
expanded distally. Subocular shelf extremely feeble, merely a slight pointed downward 
expansion of the second suborbital. | 

Vertebrae 24 (10 +14), first very much reduced, not ankylosed with the skull. Para- 
pophyses from the fourth precaudal. Ribs all sessile. Epipleurals extremely fine, hair- 
like. Haemal spines of anterior caudal vertebrae with median longitudinal expansion 
below the extensions of the air-bladder (Pls. XXII and XXIII). Post-temporal forked, 
not ankylosed. Post-clavicle long and slender. Pectoral radials 5, upper three bobbin- 
shaped, only the lowest (5th) abutting on the coracoid (fig. 1). 

Branchiostegals 6. Pseudobranchiae absent (concealed). Teeth very compressed, 
slightly recurved, tricuspid. Pyloric caeca few (3-4). 


Other characters as outlined for the family. 
At present only a single species, orbis Plyfr., is known from the east 
coast of Africa. 


250 


308 Transactions of the Royal Society of South Africa. 


Tripterodon orbis Plyfr. 
(Pls. XXI-XXIIT) 
Tripterodon orbis Plyfr., Playfair, loc. cit., p. 42, pl. vii, fig. 1. Fowler, Proc. Ac. 
Nat. Sci. Phil., 1925, vol. xxvii, p. 242; and 1933, loc. cit., p. 202. Smith, Rec. Albany 


Mus., 1935, vol. iv, p. 209.* . 
Chaetodipterus orbis (non Bloch), Norman, Ann. Mag. Nat. Hist., 1922 (9), vol. ix, p. 321 


(record only, error in Brit. Mus. Register).t 


TExtT-ria. 1.—Left pectoral girdle of T'ripterodon orbis Plyfr. showing orientation of 
ventral basipterygium. 


Cl., cleithrum; Co., coracoid; Pc., post-clavicle; Pf., pectoral fin; Pt., post-temporal; 
h., radial; Sc., scapula; Scl., supracleithrum; Vb., ventral basipterygium. 


Chaetodtpterue goreensis (non C. and V.), Barnard, loc. cit., p. 603. Fowler, Proc. Ac. 
Nat. Sci. Phil., 1934, vol. lxxxvi, p. 478, fig. 43. 

Body more or less orbicular, more angular in adults, fairly compressed. Dorsal 
profile very elevated, slightly undulate, with moderate interorbital prominence, which 
increases with age. 

Depth 1-2-1-4, length of head 3-0-3-2 in length of body. Eye 2-7-3-7 (Ad.), snout 1-6 
(Ad)-1-8, interorbital 2-7-3-2 and postorbital 3-5-4-0 in length of head. Least depth of 
preorbital 1-0 (Juv.)—1-4 times eye. 

Nostrils wide apart, anterior rounded, posterior long and oval. Mouth horizontal, 
small, terminal, only slightly protractile. Maxilla extends to below anterior nostril. 


* In this paper Chaetodipterus orbis Bloch, recorded by Barnard, loc. cit., p. 604, 
has erroneously been included in the synonymy of 7’. orbis; actually this author had merely 
accepted the British Museum record. 

{ Private communication from Mr. J. R. Norman. 


251 


The Genus Tripterodon Playfair. 309 


Lips thick, especially upper, which is } eye deep at snout tip. All the teeth movable 
compressed tricuspid incisiform, in about 4 rows, the outer of 18-20 teeth in each jaw: 
the inner teeth embedded in fleshy pads. 

Palate and tongue edentate. Gill-membranes broadly fused with isthmus. Gill- 
rakers 9-10, short, about 4-5 in gill-filaments, which are 14-14 in eye. Preopercle serra- 
tions more marked in juveniles. 

D IX, 19-21, originates above hind margin of opercle. First spine very short ; second 
twice the first; third 1-5-2-2 (Juv.); fourth 1-3-2-0 (Juv.); fifth 1-0-1-2 times head ; 
remainder short, about equal to eye. Soft rays abruptly longer than last spine, 6th—8th 
longest, 1:1-1-4 in head. Scaling on soft fin very dense at base. Base of spinous dorsal 
about as long as head, of soft dorsal 1-4—1-5 times head. 

A III, 16-17, spines fairly short and stout, second longest. Anterior rays longest, 
1-2-1-4 in head, edge of fin almost straight. 

Pectorals rounded, 1-2~-1-6 (Ad.) in head. 

Ventrals 1-0-1-6 (Juv.) times head, first ray filamentous. 

Caudal with hind margin gently undulate, with median convexity. Least depth of 
peduncle about 2-6 in head. . 

Scales ctenoid (Pl. XXIII), lateral line scales smaller than adjoining scales but not 


12-13 
hidden (Pl. XXIII), 1.1. 43-45, Ltr. 
23-24 


cheek, not much smaller than opercular scales. 

Colour: Silvery brown, with dark cross-bars of varying width, first from nape through 
eye to chest, second from before dorsal origin, thereafter 3 pairs alternately narrow and 
wide, and ninth bar indistinct on peduncle. Bars generally narrower than interspace. 
Snout dark. Spinous dorsal, ventral, and anterior portion of soft dorsal and anal, black. 
Pectorals light, dark base. Soft dorsaland anal and caudal with obscure series of dark spots. 


(origin of first dorsal spine back), 7 series on 


Length.—Up to 300 mm. (Playfair). 

Locality.—Durban, Delagoa Bay, Zanzibar. 

Seven specimens, from 97-215 mm. in length, examined. 

A very characteristic species, easily recognised by the shape of the body 
and fins, by the restricted gill-openings, and by the apically brown, strongly 
tricuspid teeth. 

It is not remarkable that this species has been confused with Chaeto- 
dipterus goreensis C. and V. (Cuvier and Valenciennes, Hist. Nat. Poiss., 
1831, vol. vii, p. 125, pl. 178), for in general external features the two 
species resemble one another so closely that only detailed comparison could 
show the difference between them. 

In so far as may be judged from the description of the type of goreensis 
(an adult) the sole external difference of any significance between that 
species and Tripterodon orbis is in the dentition, the former having typical 
brush-like setiform teeth. Besides this the preorbital in goreensis does not 
appear to be as deep as in orbis, while C. and V.’s figure shows the cheek 
scales to be very much smaller than those on the opercle, which is not the 
case in orbis, the 3rd and 5th dorsal spines not filamentous, and the upper 


lip not as deep as it is in orbis. 
VOL. XXIII, PART IV. 22 


252 


310 Transactions of the Royal Society of South Africa. 


A later description of goreensis by Pellegrin (Ann. Inst. Ocean., 1914, vol. 
vi, p. 55, figs. 7 and 8) does not mention the dentition, shows the preorbital 
to be deeper, and the 3rd—5th dorsal spines to be short in the juvenile and 
elongate and filamentous in the adult stage, which latter growth change 
appears rather exceptional. 

So similar are goreensis and orbis that, had Cuvier not made the positive 
statement about the setiform dentition of the former, I should have had 
little hesitation in regarding the two as conspecific, for the remaining 
differences are unimportant, and the conflicting statements about the lengths 
of the 3rd-5th dorsal spines may probably be attributed to accidental 
shortening of the fragile filamentous extensions of these spines. 

In so far as my specimens are concerned, the chief growth changes 
observable are that the preorbital becomes markedly deeper, the inter- 
orbital convexity greater, the 3rd—5th dorsal spines shorter, and the 
ventrals shorter with increase in size of the fish. In juveniles the ventrals 
reach well along the base of the soft anal, in adults barely beyond the origin 
of the spinous fin. I have not seen any very young specimens. 

A re-examination of the type of goreensis might be of interest. 

Fowler’s specimen described as goreensis (loc. cit.) might be either 
that species or orbis, since he does not mention the nature of the dentition. 
There is, however, little doubt that the specimen is conspecific with orbis 
as here accepted. In the description Fowler stated that the 4th dorsal 
spine varies from 1-6-2-3 in length of body, whereas in the figure it is 
shown to be much shorter, about 3. Also the shape of the anal fin in the 
figure agrees neither with that shown for goreensis (C. and V., loc. cit., and 
Pellegrin, loc. cit.) nor with that of my specimens of orbis. 

T. orbis is only an occasional capture in the nets at Durban, and is 
apparently hardly ever encountered during the winter months. There 
appears to be no record of its occurrence any distance south of Durban. 
I have recently received a large collection of representative fishes from 
the coast of Tanganyika Territory (collected during the summer months), 
and since orbis was not among them it may be concluded that it is not 
abundant even in the neighbourhood of the type locality. 

Nothing is known of the habits of this species, but they probably 
resemble those of Plataxz and Drepane, and it probably lives chiefly among 
reefs. None of the specimens I have examined have been sexually mature. 


I wish to express my gratitude to Dr. Barnard, Assistant Director of 
the South African Museum, for the loan of material and literature. Also 
to the Carnegie Research Fund (through the Research Grant Board of 
South Africa) for generous financial assistance. 


ALBANY MuSsEvUM, 
GRAHAMSTOWN, 
May 1935. 


Plate X XI. 


Trans. Roy. Soc. 8. Afr., Vol. XXIII. 


Smith. | Neill & Co., Ltd. 


Trans. Roy. Soc. S. Afr., Vol. XXIII. Plate XXII. 


Tripterodon orbis Plyfr. x 1. 


Stained and cleared specimen showing skeleton, with upper half of the paired structures removed. 


Smith. Neill & Co., Lid. 


Trans. Roy. Soc. 8. Afr., Vol. XXIII. | Plate XXIII. 


Tripterodon orbis Plytr. 


A. Showing structure of air-bladder. B. 7th lateral line scale. C. Scale above lateral 
line below spinous dorsal. Scales from specimen 185 mm. in length. The lines below 


B and C represent 1 mm. 
Smith. . Neill & Co., Ltd. 


253. 


Transactions of the Royal Society of South Africa. Vol. XXIV. 
Part |. pp. 1—6. Pls. | and Il. June, 1936. 


TWO INTERESTING NEW FISHES FROM SOUTH AFRICA.* 


By J. L. B. Smrra. 
(With Plates I and II, and two Text-figures.) 


(Read September 18, 1935.) 


Family SCYMNORHINIDAE. 


Regan (Ann. Mag. Nat. Hist., 1908 (8), vol. 11, p. 40) defined the family 
SQUALIDAE so widely as to include even genera such as Pristiophorus M & H, 
which are to-day generally accorded full family rank. Scymnorhinus 
Bonap. and Echinorhinus Blnv., and their relatives, without fin-spines, 
appear to merit family distinction from the SquaLipAE. Actually certain 
authorities favour even recognition of the families ECHINORHINIDAE and 
SCYMNORHINIDAE as distinct one from the other. Although this appears to 
set rather narrow limits for family distinction, it is accepted here. 

Scymnorhinus Bonap. replaces Scymnus Cuv., preoccupied (Kugelman 
1814, a genus of Beetles). These two genera are sometimes accepted as 
synonyms of Dalatias Raf. (1810), but according to Jordan and Evermann 
(The Genera of Fishes, 1917, p. 97) this is not established, and is not accepted 
here. 

Scymnorhinus has not previously been recorded from South Africa. A 
new species of that genus is described below. 


Scymnorhinus brevipinnis n. sp. 


No anal fin, no fin-spines, no nictitating membrane. 

Greatest depth of body 9-10 in total length. First gill-slit about mid- 
way between tip of snout and origin of first dorsal. Longitudinal diameter 
of orbit 7-5-8-5, vertical diameter 19-22, prespiracular distance 4:5, preoral 
length 6-:2-6:4, width of mouth 6-0-6-3 in distance from tip of snout to 
origin of first dorsal. Inner internasal distance 1-8 in preoral length. 
Interspiracular distance 1-5 in prespiracular length. 

Mouth transverse, lower lip thick, no labial folds. A straight longitu- 
dinal groove from each corner of mouth. 19 compressed triangular teeth 
in lower jaw. The central (anterior) tooth erect, the remainder oblique, 

* The Council desires to acknowledge the receipt of a grant from the Carnegie 


Corporation through the Research Grant Board towards the cost of printing this paper. 
VOL. XXIV, PART I. 1 


254 


2 Transactions of the Royal Society of South Africa. 


increasingly so posteriorly. Edges serrate, 19-26 fine serrae on each edge 
(fig. 2). Behind the single row of functioning teeth are four overlapping 
rows of inwardly depressed teeth in succession. 19 series of teeth in upper 
jaw, narrow and pointed: central tooth almost vertical, remainder oblique, 


edges entire. The second row close behind the first, and behind these three 
more rows in succession. | 

First dorsal inserted midway between origin of second dorsal and hind 
margin of eye, 2-2—2-3 times nearer origin of pectorals than origin of second 
dorsal, 1-7 times than origin of ventrals, and twice as far from hind margin 
of caudal as from snout tip. Second dorsal 1-2 times further from hind 


es eer -—— oo 


“Ne 
. 
ve 


’ 
H ‘ 
' 
‘ 
i : 
. ' 
‘ 
+ [om 
ns a . 
eA 
. vl » 
. ; . 4 
‘ 
. ’ 
coy 
er 
‘ ! 
. . . 4 
ae | Set 
“ i 
s 
1 
A 


_ Text-rica. 2.—Dentition of Scymnorhinus brevipinnis n. sp. 


A. Anterior teeth of lower jaw. 3B. Posterior tooth of same. 
C. Anterior teeth of upper jaw; those in second row with interrupted outline. 
: 7 (All x 2:3.) 


margin of caudal than from origin of first dorsal. Three-quarters of ventral 
base in advance of second dorsal. First dorsal about as long as second, 
slightly less than prespiracular length. Base of first dorsal 1:5, base of 
second dorsal 1-2 in length of first dorsal. Length of ventrals 1-3 in, base 
of ventrals 1-3 times, length of first dorsal. Pectorals 1-3-1-4 times length 
of first dorsal. 

Depth of gill-openings about equal to longitudinal diameter of eye. 
Nostrils more or less rounded, separated, no cirri or oronasal grooves. A 
slight depression in lower surface of snout between nostrils. Spiracles 
large. 


255 


Lwo Interesting New Fishes from South Africa 3 


Caudal lobes small: greatest transaxial width of caudal fin scarcely 
greater than prespiracular length. 

The whole body and head covered with small scales, each more or less 
quadrate, having a dorsal ridge terminating in a stout spine, usually also 
one or two minute basal spines. 

Colour.—Uniform dark brown, slightly lighter below. Tips of fins 
slightly lighter. 

Length.—T70-1100 mm. 

Holotype in the Albany Museum. 

Three specimens examined. 

These interesting fishes were taken by the trawler ‘‘ Algoa Bay” at 
a depth of 110-150 fathoms, some 35 miles south of Cape Recife, about 
34° 32’ §., 25° 42’ E., and were presented by Mr. H. D. Jackson, manager 
of Messrs. Irvin & Johnson at Port Elizabeth, who has also previously 
donated valuable specimens. 

S. brevipinnis is very closely related to the only other species of the 
genus that has yet been described, licha Bonn., which has been recorded 
from the western North Atlantic and the Mediterranean, from fairly deep 
water. Descriptions of that species available to me are exceedingly meagre, 
the only figure seen being that of Goode and Bean (Ocean. Ichthy., 1895, 
p. 7, fig. 4,* Scymnorhinus lichia). S. brevipinnis appears to differ from — 
licha in numerous dimensional relationships, notably in the smaller fins, 
the caudal lobes being markedly lower. Also Giinther (Cat. Fish. B.M., 
1861, vol. viii, p. 426), and others, have stated that the teeth of the lower 
jaw in licha are oblique in juveniles but erect in adults. In my specimens 
of brevipinnis, of which the larger are presumably adults, all but the central 
symphysial erect tooth of the lower jaw are oblique, the posterior teeth 
very markedly so. 

In any case, since a number of South African species of fishes, until 
recently held to be identical with those from the northern hemisphere, have 
upon detailed comparison of suitable material proved to be distinct, it is 
felt expedient to maintain brevipinnis as distinct from licha until such 
time as equivalent material from all the recorded areas may be compared. 

Jordan and Fowler (Proc. U.S. Nat. Mus., 1903, vol. xxvi, p. 63, Dalatias 
licha) have given a very brief account of a stuffed specimen of Scymnorhinus 
from Japan, which they have diagnosed as identical with the Atlantic 
species. This has been accepted by Regan (loc. cit.). The few details 
given by Jordan and Fowler agree neither with my specimens nor with 
descriptions and the figure of licha. It was stated, for example, that in the 
Japanese specimen the anterior margin of the mouth is before the anterior 
border of the eye. 

* The text gives fig. 3. 


256 


4 Transactions of the Royal Society of South Africa. 


Although these fishes are bathybial, and so with little to hinder wide 
distribution, they are obviously very sluggish in habit, and it is not unlikely 
that a re-examination of the specimen from Japan will establish that it is 


a distinct species. 


Family TRICHONOTIDAE. 


No member of this family has previously been recorded from South 
Africa. Records have hitherto been from the Indo-Australian area. 

A new species of the genus Taeniolabrus Stndunr. from South Africa is 
now described. 


Taenolabrus marley n. sp. 


(Plates I and IT.) 


Body very elongate, sub-quadrangular in cross-section, somewhat deeper 
than wide. 7 

Depth 12, length of head (tip snout) 5, or (tip lower jaw) 4-7, in length 
of body. Eye 6, snout 3-2, postorbital part of head 1-7 in length of head 
(tip snout). 

Head little depressed, snout very sharp, dorsal profile even. Hyes 
prominent, sub-dorsal in position, projecting above the profile, but vision 
chiefly lateral. Over the upper margin of the pupil is a small extension 
from the iris which divides below into radiating stripes, all of a golden 
colour. Anterior adipose eyelid moderately developed. Interorbital very 
narrow, with a small longitudinal ridge. Nostrils paired, rounded, minute. 
Lower margin of preorbital slightly concave above maxilla. Several pores 
and canals on preorbital and snout. Preorbital depth somewhat less than 
longitudinal diameter of orbit. Subopercle and interopercle much enlarged, 
covering branchiostegals. | 

Gill-opening large, membranes free from isthmus, unite beneath the 
hind edge of maxilla. Branchiostegals 5. Pseudobranchiae present. 
Gill-rakers 3+21, very fine and slender, 1-3 in gill-filaments which are 2 
in eye. 

_ Vent normal, in front of anal origin. 

Mouth large, almost horizontal, slightly protractile. Maxilla partly 
concealed beneath preorbital, extends to below middle of orbit. The lower 
jaw projects strongly, forming a prominent mental lobe or chin, the upper 
jaw almost included. Lower lips finely papillose. Minute curved villiform 
teeth in a narrow band in each jaw. Upper jaw with a large dentate 
symphysial knob, teeth recurved, slightly enlarged. Minute teeth on 
vomer in an angular patch with anterior median projection. Similar small 
teeth in a narrow band on palatines. Tongue edentate, apically dilated, 
free. 


257 


Two Interesting New Fishes from South Africa. 5 


The intestinal tract is simple, being merely siphonal, the stomach 
elongate: no pyloric caeca. Liver bilobed, the lobes even, extending almost 
above vent. The presence of a gall-bladder cannot be established. No 
air-bladder. 

D 47 (or IIT 44) originates slightly behind the base of the pectoral. 
Anterior three rays filamentous, spiniform. First two rays detached from 
remainder, first 2-0, second 1-8 times head. Third ray 1-5 times head. 
The fourth ray is 2 in head and simple but articulated. The remainder are 
mostly branched, often very indistinctly so. From the fourth, the rays 
increase in length slightly to the 10th and are thereafter subequal. 

A 39 (or I 38, the fig., Pl. I, shows 37 in error) originates below the 
10th dorsal ray, less than a head length behind head, below the tip of 
the pectoral. Ist ray spiniform, ‘short: remainder bifurcated or branched. 
Rays gradually increase posteriorly to almost length of dorsal rays. 

Pectorals 1-4 in head, inserted fairly low, of 12 branched rays; tip 
reaches to 14th lateral scale, above origin of anal. 

Ventrals I, 5, inserted close together, in advance of pectorals. Spine 
weak: inner ray simple, remainder branched. Ist ray shortest, remainder 
increase to 4th, which is filamentous, the total length being 1-4 times the head, 
the filament is as long again as the fin, which reaches the origin of the anal. 

Caudal broadly hastate, of 13 branched rays, slightly longer than head. 

Scales cycloid, longer than wide, with angular hind margin, the edges 
meeting at an angle of about 80° (Pl. IJ). Lateral line scales with a deep 
notch in hind edge, and a short vertical groove above the tube (PI. II). 


41 
Lateral line straight, tubules simple. 1.1. 57, Ltr. a? 10 predorsal scales, 
1 | 


end above hind margin of preopercle. Head naked except for five scales 
below the orbit down hind margin of preorbital and 3-4 more behind angle 
of mouth along anterior margin of preopercle. 


Colour (Alive): ‘‘Above amber, studded with turquoise blue and red 
dots. The lower surface faint rosy. Anals and ventrals with numerous 
dark red dots. Eye with about twenty golden lines, umbrous red above.”’ 

(Preserved): Light yellow-brown above with 11 faint darker cross-bars, 
wider than eye, not reaching to ventral surface, the first over nape extending 
on to opercle, the last below the last dorsal rays. _ Series of dusky bordered 
ocelli along the body, 6 series anteriorly, becoming fewer posteriorly: one 
ocellus on each lateral line scale. Similar ocelli in about three series along 
side of head. <A few dark dots along lower surface of body. Dorsal faint 
dusky with 6-8 series of ocelli, anterior two spines with alternate light and 
dark annular rings. Irregular dark spots, larger posteriorly, over fin. 
Caudal similar. Anal similar, but lighter, and fewer ocelli. Pectorals 
light. Ventrals with signs of darker markings. A few dusky mottlings 
on nape and muzzle. Iris, iridal flap, and radiating lines golden. 


258 


6 Transactions of the Royal Society of South Africa. 


Length.—190 mm. 

Locality. Durban (6th June 1935). 

Type in the Albany Museum. (H. W. B. M. 969). 

This interesting specimen was presented by Mr. H. W. Bell Marley, 
Principal Fisheries Officer of Natal, after whom it is named. His colour 
notes of the live fish are given above. Mr. Bell Marley has also sent another 
specimen which, while almost certainly conspecific, nevertheless differs 
markedly in the nature and development of the fins. This smaller specimen 
(155 mm. total length) has D 45, A 37 and 1.1.55. The first three dorsal 
rays are only slightly longer than the remainder, the third being actually 
the longest, about 1-7 in head, while the remaining rays are relatively 
shorter than those of the type, and are almost all simple. Further, the 
dorsal originates about an eye diameter behind the pectoral base. The 
caudal is only 2 as long as the head, and broadly rounded. The anal is 
similar to that of the type, but the rays are shorter, and the origin of the 
fin is almost a head length behind the head, well behind the pectoral tip. 
The 4th ventral ray is only slightly extended, the total length being 1-4 
in head, and even the filamentous tip does not reach near the vent. There 
are 3+19 gill-rakers, and the vomerine angular patch has no anterior 
extension. It is difficult to decide whether these two specimens are con- 
specific or not. The nature of the dorsal and caudal fins alone might 
justify distinction, but in general features, shape, markings, etc. there is 
such marked similarity that specific distinction for the smaller specimen 
would appear venturesome. Neither specimen shows any trace of sexual 
development, so that immaturity may be the cause of the numerous 
variations. 

Very few species in this family have been described, and very little 
is known about them. It would certainly appear that a revision of the 
genera and species is necessary. 

TI. marley appears to be well differentiated from all others by having the 
anterior three dorsal rays and the 4th ventral ray elongated, as well as by 
the scales on the cheek, and by the number of scales in transverse series. 

The Ammodytid-like body and head would appear to indicate a mode of 
existence resembling that of those fishes. 

This is a noteworthy addition to the ichthy-fauna list of South Africa. 


i wish to express my gratitude to the Research Grant Board of South 
Africa (Carnegie Fund) for generous financial assistance. 


ALBANY Museum, 
GRAHAMSTOWN, 
August 1935. 


Trans. Roy. Soc. 8. Afr., Vol. XXIV. Plate I. 


(Type.) 


Taeniolabrus marleyi n. sp. Nat. size. 


:. SCOOT Si 
D5. wae Sieh ores 
De 


ae 


J. L. B. Smith, del. Neill & Co., Lid. 


Plate IT. 


Trans. Roy. Soc. 8. Afr., Vol. XXIV. 


ainqny oy} 04 qusoelpe 


‘od A4 JO opls YYSII WOIy YO 


JAOOIS (VSIOP SOJBOIPUL MOLIW *1Z x 
‘ds ‘u wapipu snaqnjowuany, yO 


"GG x “aUl] [e1o}ze] BAOQ® o[VOS YIFZ— Jy bry 


‘g[vos oul] [eloqel YIPZ—"Hfa'7 


S 


apeog 


Neill & Co., Ltd. 


J. LL. B. Smith. 


259. 


Transactions of the Royal Society of South Africa. Vol. XXIV. 
Part |. pp. 47—55. Pls. I1I-V. June, 1936. 


NEW GOBIOID AND CYPRINID FISHES FROM SOUTH AFRICA.* 


By J. L. B. Smira. 
(With Plates III-V and three Text-figures.) 


(Read October 16, 1935.) 


Family GoBIIDAE. 
Gobtus vonbondei n. sp. 
(Text-fig. 1 and Plates III and V, fig. A.) 


Body fairly compressed, markedly so posteriorly. Dorsal profile even, 
moderately sloping before eyes, snout blunt, rounded. 

Depth equal to length of head, 3-5 in length of body. Eye 4:3, inter- 
orbital 4-0, snout 3-3, and postorbital part of head 2-0 in length of head. 
Least depth of preorbital 1-8 in eye. | 

No barbels. Series of papillae on cheeks ; below the eye a row with 
short vertical transverse series. A series along the lower margin of the 
preorbital extending beyond angle of mouth on to preopercle. A single 
series in a shallow groove along the lower surface of each mandible. Nostrils 
small, posterior circular, anterior with plain flap. A large pore above the 
anterior nostril. 

Mouth moderate, slightly oblique, jaws equal, maxilla extends to below 
anterior margin of pupil. Teeth in 3-4 rows in each jaw, with an outer 
enlarged series of 14—-15, caniniform. An outwardly flaring curved canine 
tooth on each side of lower jaw. Palate and tongue edentate. Tongue 
anteriorly truncated, adnate. A papillose cutaneous process across roof 
of mouth. 

Gill-opening restricted, membranes broadly fused with isthmus from 
below posterior half of operculum. 5 gill-rakers on lower margin of anterior 
arch, of curious claw-like design. The upper limbs of the anterior two 
arches have a forward extension bearing three large claw-like processes, 
which have below numerous shorter papilliform processes. 

D VI+I, 14. First dorsal originates just behind pectoral base, twice as 
far from caudal base as snout tip. Spines elongate, filamentous, Ist 1-4, 


* The Council desires to acknowledge the receipt of a grant from the Carnegie 
Corporation through the Research Grant Board towards the cost of printing this paper. 


260 


48 Transactions of the Royal Society of South Africa. 


Qnd 1-2, 3rd 1-0, 4th 0-85, 5th 1:3, and 6th 2-1 in length of head. Base of 
first dorsal 1-4 in head. Membrane from 6th spine reaches origin of second 
dorsal. Second dorsal inserted slightly nearer caudal base than snout tip, 
above the 20th lateral scale, spine 2-2 in head, anterior rays 2:0 in head, 
increase to the 7th (longest) 1-5 in head, thereafter graduated shorter. 
Base of second dorsal 1-2 times head. Hind margin of fin does not reach 
the caudal. 

A I, 12. - Originates below the origin of the second dorsal. Spine 1:5 
times eye; first ray, shortest, 1-7 in head; penultimate, longest, 1-3 in head. 
When laid back the hind margin of the fin reaches on to the caudal fin. 
Base of anal equal to that of second dorsal. 

Pectoral rounded, of 17 rays, 1:1 in head, tip reaches below origin of 
second dorsal. 

Ventrals united, as long as pectorals. 

Caudal broadly rounded, of 15 principal rays. Peduncle about as long 
as deep. 

Scales moderate, mostly ctenoid (Plate III, A), but those on the nape, 
head, pectoral base, and chest, cycloid. Lat. ser. 52, l.tr. 18, predorsal 
23, which end above posterior third of eye. Scales in advance of pectoral 
base smaller than posterior scales. About ten series of scales across upper 
portion of opercle, rest of head (except dorsally to interorbital) naked. 

Colour.—Brownish, slightly lighter below. Five dark cross-bars, 
narrower than eye, much narrower than interspaces; first below middle of 
first dorsal, second below anterior dorsal rays, third below middle dorsal 
rays, fourth below posterior dorsal rays, last across peduncle. Between 
each pair a narrower faint stripe. Faint stripes below origin of first dorsal, 
over nape, and behind eye. Opercle dark below. A faint dark smudge 
below anterior margin of eye. A large black spot at upper margin of 
opercle above pectoral base, another similar on upper side of caudal base 
(ocelli in life 2); three small dark spots along base of dorsal; a similar spot 
on upper part of hind margin of caudal, and four spots along lower margin 
of caudal. Dorsal, anal, and ventrals dusky, other fins light. 

Length.—133 mm. 

Locality.—Natal. 

Type (g) in the Albany Museum. 

Named after Dr. C. von Bonde, Director of the South African Govern- 
ment Fisheries Survey, who has donated numerous valuable specimens to 
the Albany Museum. | 

The type is a ripe male. Conspecific is also a smaller ripe female, 
120 mm. total length. A certain amount of sexual dimorphism is exhibited 
by this species. The body of the female tapers to the peduncle shightly 
more than that of the male, while there are only the two large spots, one 


261 


New Gobioid and Cyprinid Fishes from South Africa. 49 


at the pectoral the other at the caudal base, the smaller spots on the dorsal 
and caudal being absent. Chiefly the sexes differ in the nature of the 
dorsal fins. The first dorsal is longer based, but not as high in the female, 
while in the second dorsal, the posterior rays of that of the female are the 
longest, the shape of the fin resembling that of the anal. Also the ventral 
is slightly shorter, scarcely reaching the vent. 

D VI+I, 14, AI, 13; lat. ser. 53, ltr. 19; predorsal 23. 6 gill-rakers. 


X 


—_ 


Fie. 1.—Gobius vonbondei n. sp. (Female). The line represents 1 cm. 


G. vonbonder is easily distinguished from all other South African species 
by the high dorsal fin, as well as by numerous other features. 


Gobius gulosus n. sp. 
(Text-fig. 2.) 


Body fairly slender, sub-cylindrical, compressed posteriorly. Dorsal 
profile even, with slight concavity before eyes. Snout moderately sharp. 

Depth 5-0, length of head 3-1 in length of body. Eye 5:4, snout 3-0, 
interorbital 6-0, and postorbital 2-1 in length of head. Least depth of 
preorbital equal to eye. 

No barbels. 6-7 undulating series of papillae on cheeks. Various 
other series; along preopercle margin, double, behind eye to occiput, and 
along head above upper margins of opercle and preopercle. A series along 
each lower margin of opercle, with a third forming a triangle. A deep 
groove containing papillae on each side of snout before eye. Several pores 
on snout. Chin and lower surface of rami of lower jaw with papillae. 
Anterior nostril tubular. 

Head very depressed, maximum width behind eyes 1-2 in length, con- 
siderably wider than deep. Mouth large, oblique, lower jaw projects 


262 


50 Transactions of the Royal Society of South Africa. 


slightly. Maxilla extends below anterior border of orbit. Slender conical 
teeth in 4-5 rows in each jaw, outer series but slightly enlarged. Tongue 
adnate, truncated. Cutaneous process in roof of mouth narrow. Gill- 
opening restricted, membranes fused with isthmus from below middle of 
opercle. 9 gill-rakers on lower limb of anterior arch, fairly short, moder- 
ately stout. 

D VI+I, 8. First dorsal originates 1-8 times further from caudal base 
than snout tip, above the 5th lateral scale. Spines slender, Ist 3-0, 2nd 2-2, 
3rd (longest) 1-9, 6th 4-0 in length of head. Base of first dorsal 1-9 in head. 
Two dorsals close together. Second dorsal originates slightly nearer caudal 
base than snout tip, above the 13th lateral scale. Ist ray 2-0 in head, 


Fig. 2.—Gobius gulosus nu. sp. The small arrow indicates anterior limit of scaling. 
The line represents 1 cm. 


increase to penultimate (longest) almost as long as head, when laid back 
reaches well on to the caudal fin. Base of 2nd dorsal 1-4 in length of head. 

AI,7. Inserted below the 2nd dorsal ray, below the 15th lateral scale. 
Spine short, first ray 2-2, remainder increase to penultimate, longest, 1-2 in 
head, depressed reaches to caudal base. Base of anal 1-8 in head. 

Pectoral more or less lanceolate, of 16 rays, as long as head, tip reaches 
beyond anal origin, to 16th lateral scale. No free rays. 

Ventrals 1-4 in head, reach to vent, basal membrane wide. 

Caudal lanceolate, 1:2 times head; peduncle slender, almost 2-5 times 
as long as deep. | 

Scales moderate, mostly ctenoid, but cycloid on the nape, on the chest, 
on ventral surface of body, on pectoral base, and on caudal base. Lat. 
ser. 31, ltr. (between dorsal and anal) 8, predorsal 8-9, very small. 
Posterior scales of body much larger than those anterior to dorsal origin. 
Head entirely naked except few predorsal scales, which extend almost 
above hind preopercle margin. Chest only lightly scaled. 

Colour.—Uniform dark olive, nape darker, slightly lighter on belly. 
First dorsal with large dusky blotch over last three spines. Several rows 


263 


New Gobioid and Cyprinid Fishes from South Africa. 51 


of dusky spots along base of second dorsal; also over base of caudal. 
Ventrals dusky. Pectorals light. 

Length.—91 mm. 

Localhity.—Bushman’s River, Alicedale (Cape Province, about 30 miles 
from nearest point of coast). 

Type in the Albany Museum. 

A single specimen, presented by A. Cruden, Esq. 

This species would appear to fall into Stenogobius Bleeker, of which 
somewhat doubtful sub-genus no representative has hitherto been recorded 
from South Africa. 

G. gulosus is well differentiated from all other species from South Africa 
by various features. It is closest to the widely distributed species giuris 
Ham. Buch., but is distinguished chiefly by the elongated fins, by the 
absence of scales from the occiput, by the arrangement of the buccal 
papillae, and by the adnate tongue, besides other features. It is not even 
unlikely that gulosus has originated, by isolation, from giuris, or that both 
species have developed from a common ancestor. 

Extensive collecting in the fresh waters of the Hastern Province of South 
Africa will almost certainly result in the discovery of numerous new species 
of fishes, some of which at least will not improbably show relationship to, 
or even traces of origin from, known widely distributed forms. In that 
area it would appear that geographical and climatic changes have consider- 
ably affected the character of the larger rivers. Even those which at one 
time were presumably of continuous flow, and with few barriers to hinder 
the inland penetration of marine species, now have the inland freshwater 
without direct or continuous communication with the sea. Only during 
floods is there continuous water, but at those times the current is too swift 
to allow of counter migration. Actually the periodic inundations which 
occur in that region take heavy toll of the fish life of the fresh waters, for 
after any heavy spate innumerable dead fresh water fishes are thrown up 
on the shore in the neighbourhood of the mouth of any large river in the 
Eastern Province area. 

The Bushman’s River is tidal for over 20 miles, but beyond that consists 
at normal times of merely a series of disconnected pools. The course of 
the river from Alicedale to the sea would be at least 50 miles. Any upward 
migration in that river would appear impossible, and gulosus is most 
probably a localised fresh water species. 


264 


52 Transactions of the Royal Society of South Africa.’ 


Family HLEOTRIDAE. 
Eleotris imosus n. sp. 
(Plates IV and V, fig. B.) 


Body robust, moderately compressed, maximum width at shoulder 
1:2 in depth. Dorsal profile flat, with deep concavity before eyes. 

Depth 4-1, length of head 3:3 in length of body, Head 1-2 times wider 
than deep. Eye 6-5, interorbital 2-4, snout (measured obliquely to tip) 3-4, 
and postorbital part of head 1-6 in length of head. Least depth of preorbital 
1-8 in eye. (Length of head measured obliquely to snout tip.) 

Anterior nostril tubular, apically dilated, immediately above maxilla; 
posterior before the orbit. A large pore close above the anterior nostril. 
Several papillose, muciferous canals on cheek, two radiating down from the 
eye; from the anterior nostril along the lower preorbital margin another, 
which extends beyond the angle of the mouth well on to the preopercle. A 
groove behind the eye extends to the upper margin of the gill-opening; at 
the origin next the eye is a large pore; also several others in and above 
the course of the groove. 

Anal papilla broad and flat, about as long as eye. 

Gill-opening fairly wide, gill-membranes not united, but fused narrowly 
below with isthmus. Gull-rakers 12, anterior moderate, stout, posterior 
mere knobs. Pseudobranchiae well developed. Space between rami of 
lower jaw wide, transversely rugose. 12-14 apically dilated papillae in a 
series in a groove along each side of the lower surface of the lower jaw, 
separated by a small space from a posterior groove containing two similar 
papillae and ending in a large pore. 

Mouth large, oblique, lower jaw projects strongly, maxilla extends to 
almost below centre of eye. Uniform minute conical villiform teeth in 
each jaw in a band, anteriorly half as wide as eye: no canines, no enlarged 
outer series. Palate and tongue edentate. Tongue free, wide, rounded. 

D VI+18. First dorsal inserted midway between the tip of the snout 
and the hind margin of the base of the second dorsal, above the 6th lateral 
scale. Base of first dorsal 2 in head. Ist and 6th dorsal spines shortest, 
3rd 2:3, 5th 2:4 in length of head, 1-8 in depth of body. Second dorsal very 
close to first, separated by a space about 3in eye. Second dorsal originates 
above the 17th lateral scale, slightly nearer caudal base than hind margin 
of eye. First ray (shortest) 2-0, last ray (longest) 1-15 in head, tip of de- 
pressed fin reaches well beyond caudal base. Base of second dorsal 1-7 
in head. 

AnalI, 7. Inserted below the 19th lateral scale, below the base of the 
second dorsal ray, midway between the caudal base and hind preopercle 


265 


New Gobioid and Cyprinid Fishes from South Africa. 53 


edge. Base of anal 2 in head. Ist ray (shortest) 2-5, last (longest) 1-5 
in length of head. The depressed last ray reaches the caudal base. 

P14. 1-1in head, rounded, with heavy base, tip reaches to below the 
base of the first dorsal ray, to the 19th lateral scale. 

Ventrals inserted below pectorals, 1-4 in head, tip reaches just beyond 
vent, apex of longest ray slightly filamentous. 

Caudal 15, large, rounded lanceolate, just longer than the head (base 
obscured by scales). Peduncle 1-5 times as long as deep. 

Scales large, mostly ctenoid (Plate III, B), but the predorsal scales and 
those on the head, on the pectoral base and on the chest, cycloid. Lateral 
rows 41 to caudal base, several smaller scales beyond: l.tr. 13 (dorsal to 
anal); 12 vertical series on preopercle, 5 on opercle, 18 predorsal. Whole 
body and head scaly, except muzzle, lower margin of lower jaw and front 
half of cheek, naked. Caudal scaly basally, other fins not. 

Colour.—Uniform dark olive, muzzle slightly darker. Fins dusky; 
first dorsal with a few basal dark spots, second dorsal and caudal with rows 
of numerous dark spots. Second dorsal, caudal, anal, and ventral with 
very narrow light margin (in life faint orange). 

Length.—270 mm. 

Locality.—Isipingo lagoon, near the sea. 

Type in the Albany Museum. 

E. limosus is closely related to ophiocephalus C. and V. and to mada- 
gascariensis C. and V., which are indeed only doubtfully distinct one from 
the other, agreeing closely in all features except that the latter species has 
a more elongate caudal peduncle and an extra ray in the second dorsal 
(fide Boulenger, F.W.F., Africa, 1916, vol. iv, pp. 15-16). &. lumosus is 
differentiated by several features, notably by the greater size of the fins, 
while the two dorsals are much closer than in ophiocephalus and in mada- 
gascariensis. Further, the teeth of lamosus are of uniform size, there being 
no even slightly enlarged outer series. Day (Fishes of India, 1888, p. 312, 
pl. Ixvii, fig. 2) described and figured ophiocephalus C. and V.,and Boulenger 
(loc. cit.) has accepted his diagnosis, whereas Day’s specimens cannot 
possibly be conspecific with those described by Boulenger. Day stated 
that his specimens had 31-34 scales, whereas Boulenger gave 35-42. It 
may be noted that Day’s figure agrees in very few significant features with 
his description, the discrepancies being so marked as almost to appear as 
if some confusion of specimens had occurred. At all events, either the 
specimens described. by Day were not ophiocephalus C. and V., or, if they 
were that species, then Boulenger’s were not. Actually Boulenger’s 
description (loc. cit., p. 15) agrees very well with my specimen, and 
it is not unlikely that his specimens will be found conspecific with 


lamosus. 


266 


54 Transactions of the Royal Society of South Africa. 


In view of the habits of the ELEoTRIDAE it would be surprising to find our 
southern African species identical with those from India. 


Family CyPRINIDAE. 
Barbus senticeps n. sp. 
| (Text-fig. 3.) 


Body moderately compressed. Depth equal to length of head, 3-4 in 
length of body. Eye 5-0, snout 3-7, interorbital 3-0, and postorbital 1-8 
in length of head. | 

Snout somewhat blunt and swollen, with sharp concavity, part of a 
transverse groove, before eyes. Upper surface of snout and head with 


Fia. 3.—Barbus senticens n. sp. (Type). The line represents 1 cm. 


numerous spiny tubercles, each with a rounded enlarged horny base; those 
on the snout large, behind the eyes much smaller. A single barbel on each 
side, 1-5 times eye. 

Mouth inferior, lower jaw included, maxilla extends to below nostrils. 
7 gill-rakers on anterior arch. 

D Ill, 7. Inserted slightly nearer snout tip than caudal base. 
Anterior rays 1-3, posterior 2-5 in head. Edge of fin straight. 

A III 5. Inserted twice as far from snout tip as caudal base. Anterior 
rays 1-3, posterior 2-5 in head. Edge of fin gently concave. Pectorals 
1-3 in head, reach to ventral base. Ventrals inserted very slightly in 
advance of dorsal, 1-4 in head, reach to anal origin. Caudal deeply forked, 
peduncle 1-8 times as long as deep. 

Scales large, with a moderate number of widely radiating striae. 1.1. 30, 
l.tr. 8 (dorsal origin back): 3 between ventrals and lateral line, 10 around 
peduncle, 14 predorsal. 


267 


New Gobioid and Cyprinid Fishes from South Africa. 55 


Colour.—Dark olive-brown above, lighter below. An obscure dark 
lateral stripe. Tubercles reddish. 

Length.—T7 mm. 

Locality —Kromme River, Assegai Bosch (near Humansdorp, C. P.). 

Type in the Albany Museum. 

This species falls into the small group which have the dorsal spine simple 
and flexible, and only one barbel. It is near to anoplus Weber and to 
afer Peters, but differs in the much smaller eye, in the much longer barbel, 
more anterior insertion of the dorsal, fewer scales transversely, and several 
other features. 

Numerous new species of this genus doubtless await discovery in the 
Eastern Province, since the fresh water fauna of that region has not been 
seriously investigated. 


I wish to express my gratitude to the Research Grant Board of South 
Africa (Carnegie Fund) for financial assistance. 


ALBANY MUSEUM, 
GRAHAMSTOWN, 
September 1935. 


Trans. Roy. Soc. 8. Afr., Vol. XXIV. Plate III. 


The line represents 1 ¢m. 


Gobius ronbondei n. sp. (Male) Type. 


J. L. B. Smith, del. Neill dt Co. Ltd 


Trans. Roy. Soc. 8. Afr., Vol. XXIV. Plate IV. 


The line represents 1 cm. 


Eleotris limosus n. sp. (Type). 


J. L. B. Smith, del. Neill & Co., Ltd. 


Plate V. 


XXIV 


Trans. Rov. Soc. 8S. Afr.. Vol. 


"U SnsoUult) 


Std 


400] J 


ayeos 


eT 


a 


rad Ay, 049 


Woody Y 


oR 


¥ 


& 


x 


* 


ds 


U 1apuUoquo 


et 


Ss 


n 


190) 


JO O[BOS [V194%] 


otpew YT “V 


Neill & Co., Lta. 


J. L. B. Smith. 


269 


Annals of the Natal Museum, Vol. VIII, part 2. 


NEW RECORDS OF SOUTH AFRICAN FISHES. 167 


New Records of South African Fishes. 


By 
J. L. B. Smith, M.Sc., Ph.D., F.R.S.(S.A.), 


Albany Museum, Grahamstown. 


With Plate XI and 6 Text-figures. 


CONTENTS. 

PAGE 
Acanthidium quadrispinosum McCulloch . . . . 168 
Haplochilus katange Boulenger . . . . . . 170 
Pisoodonophis boro Ham-Buch . . . . . . 7 
Mugil crenilabis Forsk . . . . . . . . 172 
Holocentrum sammara Forsk . . . . . . 173 
Parapercis pulchella Day . . ; . ; ; . 174 
Caranx gymnostethoides Bleeker ; . . ; . 177 
Chetodon trifasciatus Mungo Park . . . .  . 178 
Chetodon xanthocephalus Bennett ; . ; . . 180 
Chetodon kleinii Bloch . ; . . . . 182 
Pristipomoides argyrogrammicusC.¢& V. . . . . 183 
Pomadasys stridens Forsk . . . . ; » « 185 
Pteragogus opercularis Peters . . . . ~-  . 187 
Thyrsitoides marleyi Fowler . . . . . . 189 
Sarda chilensis C. & V. . . . . . . . I9l 
Blennius fascigula Barnard . . . . . . . 193 
Petroscirtes tapeinosoma Blkr. . . | . 194. 
Note on the Distribution of Species of Clinus Cur, in South Africa. 194 
Nemacoclinus navalis Barnard. . . 2 . 195 
Gobius callidus nom.nov.  . . . . . ; . 197 


Wiruin the last few years there have been very numerous 
additions to the ichthyfauna list of South Africa, both of new 
species, and of known species new to this region. In particular 
a large number of known Indian and Indo-Pacific forms have 


VOL. VIII, PART 2. 13 


270 


168. J. L. B. SMITH. 


been found to be regular inhabitants of the waters of our eastern 
coasts. Very many of these fishes have been discovered by 
Mr. H. W. Bell-Marley, Principal Fisheries Officer of Natal, to 
whose enthusiasm and zeal South African ichthyology is deeply 
indebted. 

Most of the species described below were obtained by Mr. 
Bell-Marley. 


Family SQUALID. 
Acanthidium quadrispinosum McCulloch. Text-fig. 1. 


Acanthidium quadrispinosum McCulloch, Endeav. Fish., vol. 1, 
pt. 2, p. 100, pl. xiv, fig. 5, 1915. 


Head very depressed, especially snout. Snout tip to pectoral 
origin 4:3 in total length. Body anteriorly little, progressively 
more compressed posteriorly. Greatest width of head 1-8 in 
length. Orbit 4-5, snout 2-35, preoral length 1-7, interorbital 
3-6, width of mouth 3-6 in length of head (to pectoral origin). 

Snout with sharp edges, pentagonal in shape, with angles 
rounded, tapers sharply before nostrils. Nostril width 2:3 in 
internarial distance. Orbit equal to distance from nostrils. 
Centre of eye very slightly nearer pectoral origin than snout tip. 
Orbit slightly less than distance to anterior gill-slit. Spiracles 
2 in orbit, 1:5 times distance behind orbit. Gill-openings 
graduated wider posteriorly, last more than half orbit deep. 

Mouth gently curved. Upper teeth with broad bases, the 
single cusp broad at base, more or less erect, 26 in anterior 
series. 30 anterior teeth in lower jaw, bases broad, cutting 
edges almost horizontal. Lips very thin, not continuous across 
mouth. Labial grooves very oblique, slightly longer anteriorly, 
total length almost width of mouth. 

First dorsal inserted slightly nearer snout tip than hind 
margin of second dorsal base. Spine 1-2 in orbit, distal third 
exposed. Greatest height of soft fin only slightly more than 
spine; base of first dorsal (with spine) 2-4 times orbit. Base 
of first dorsal (with spine) half the interdorsal space. Edge of 
fin gently convex, posterior rays elongated. Origin of second 


271 


NEW RECORDS OF SOUTH AFRICAN FISHES. 169 


dorsal behind hind margin of ventral base, slightly nearer hind 
margin of caudal than origin of first dorsal. Spine 1-3 times 
orbit. Soft fin anteriorly elevated; greatest height: 1-5 times 
orbit; last portion elongated, reaches beyond the origin of the 


TEXT-FIG. 1. 


_ 
( 
== 


Lower surface of head of Acanthidium quadrispinosum McCull. 
xX 


lower caudal lobe; edge of fin emarginate. Base of fin (with 
spine) slightly less than base of first dorsal. Both dorsal spmes 
with two grooves on each side. 

Pectorals with gently rounded outer margins; inner extremity 
does not reach below first dorsal origin. Ventrals mserted 
slightly more than twice as far from pectoral origin as from 
origin of lower caudal lobe. Apex reaches to below the anterior 
fourth of the second dorsal. Caudal moderately deep; greatest 
transaxial width through lower lobe slightly greater than height 
of second dorsal. A slight notch between extremities of upper 
and lower lobes. 

Whole body and fins covered with minute scales; on the 
anterior part of the snout they are granular; on most of the rest 


272 


170 J. L. B. SMITH. 


of the body they bear spines, pedunculated, mostly tridentate. 
On the upper caudal surface the spines are broadly hastate, 
occasionally with small basal cusps. 

Cotour.—Grey-black, slightly lighter below. 

LenetH.—700 mm. 

Locatiry.—Off Algoa Bay in 150 fathoms. 

The single specimen described was presented by J. H. Jackson, 
Esq., Manager of Messrs. Irvin & Johnson, Port Elizabeth. 

This specimen is unquestionably conspecific with the Australian 
species quadrispinosum McCull., which has not previously 
been recorded from South Africa. 

A. natalense Gich. is of doubtful validity. The original 
description (Gilchrist, “Fish. Mar. Surv. Spec. Rep. III’, p. 49, 
pl. vu, fig. 2) is most inadequate, and I have not been able to 
examine the type. A. natalense is supposed to differ from 
quadrispinosum in that the former has the eye slightly nearer 
pectoral base than snout tip. In McCulloch’s figure (loc. cit.) 
of the latter species the centre of the eye is actually in that 
position, and Gilchrist’s figure shows so little difference in that 
particular character that natalense is at best a doubtful 
species. 


Family CyPRINODONTID. 
Haplochilus katange Boulenger. 


Haplochilus katange, Blgr.; Boulenger, F. W. Fishes Africa, 
vol. iil, p. 67, fig. 54, 1915. 


Body fairly compressed, depth 3-9, length of head 3-4 in 
length of body. Hye 3-6, snout 3-3, interorbital 2-2, and post- 
orbital length 2-3 in length of head. Mouth terminal, sub- 
vertical, lower jaw projects. Head depressed, interorbital but 
faintly convex. Preorbital depth 4 in eye. 

D. 10 inserted above the 5th anal ray, 1-8 times further from 
snout tip than caudal base. Penultimate ray longest, about 
twice eye. 

A. 14, inserted 1-6 times further from snout tip than caudal 
base. 8th or 9th ray longest, about twice eye. 


273 


NEW RECORDS OF SOUTH AFRICAN FISHES. 171 


Pectorals 1:3, ventrals 1-8 in head. Ventrals inserted 1-4 
times further from caudal base than snout tip; do not reach 
anal origin. Caudal rounded, 1-3 in head, peduncle twice as 
long as deep. 

Scales cycloid, 26-27 lateral series, 1. tr. 6 (ventral base up). 
15 predorsal to hind margin of head. No lateral line pits. 

CoLouR.—Olive-brown above, lighter on belly. Each scale, 
light centre with darker edge. Dark band from caudal base 
slightly below middle of side to pectoral base, continued faintly 
over operculum and through eye to snout. All fins except 
ventrals dusky. Caudal and anal with faint rows of darker 
spots. 

LENGTH.—33 mm. 

LocaLity.—Kula River, Zululand. 

A single specimen. (H. W. Bell-Marley. 3.1x.1935.) Previously 
reported from the Belgian Congo. 

There does not appear to be very much difference between 
katange and myapose Bigr. 


Family OPHICHTHYIDA. 


Pisoodonophis boro Ham-Buch. 


Pisoodonophis boro (Ham-Buch), Weber and de Beaufort 
Fishes Indo-Austr. Archip., vol. iii, p. 297, figs. 138, 140, 141, 1916. 

Head 3-7 in body to vent. Tip of snout to vent 1-8 in caudal. 
Eye about 3 in snout, almost 2 in interorbital, slightly nearer 
hind margin of mouth than snout tip. Mouth cleft about 4 in 
head, extends an eye diameter behind eye. Lower jaw shorter. 
Teeth obtuse, sub-conical, in several series in jaws and on palate. 
Gill-openings small, embrace pectoral base. 

Dorsal originates about a pectoral length behind tip of pectoral. 

CoLour.—Light brown above, lighter below. 

LencTH.—470 mm. 

Locatity.—Zululand coast. 

A single specimen. (H. W. Bell-Marley, vii.1935.) 

An Indian species, only once (Fowler, 1929) previously recorded 
from South Africa. 


274 


172 J. L. B. SMITH. 


Family Mueiuip2. 
Mugil crenilabis Forsk. 


Smith, Ann. S. Afr. Mus., vol. xxx, pt. 5, p. 609, fig. 6, 1935 (Revision of 
South African Mugilide). 


This species is exceedingly uncommon in our area, and despite 
a search of several years, no adult specimen has been obtained 
until recently. The description of crenilabis in the above 
revision was based on juvenile specimens. The large specimen 
now to hand, obtained by Mr. H. J. Koch at Isipingo, requires 
that the diagnosis of that species should be emended. 

Depth 4, length of head 4:6 in body length. Eye 4-6. mter- 
orbital 1-9, snout 4-0, and postorbital length 1-6 in head. Adipose 
eyelids rudimentary. Lower margin of preorbital bent and 
sharply emarginate. Upper lip at snout tip deeper than half 
eye, lower margin with 5-6 transverse series of fleshy tubercles. 
Lower lip with expanded rugose fringe. Maxilla entirely 
concealed. | 

D. IV + 1, 8, first dorsal inserted midway between snout tip 
and caudal base. Ist spine 1-9 in head, spines rather feeble. 
Origin of first to origin of second dorsal 1-1 times head. First 
dorsal originates over the 12th, the second over the 24th scale. 
Pointed sheath scale extends behind origin of dorsal 2:6 in head, 
and 5-5-6 in distance from origin of first dorsal to snout tip. 
Second dorsal sub-falcate anteriorly, densely scaly. 

A. III, 9, mserted below the 23rd lateral scale. In shape 
similar to dorsal, densely scaly. 

Pectorals longer than 1:1 times head, with fairly strong 
axillary scale, 3-5 in length of fin. Fin densely scaly. Ventrals 
1-5 in head, margins truncate. Caudal emarginate. 

Scales all cycloid, dorsal scales with moderate scalloping along 
exposed edges, grooves long and narrow. Lat. ser. 41, 1. tr. 13, 
3 cheek scales, 13 predorsal to above hind margin of head. 

Cotour.—Silvery olive, axil of pectoral and margin of upper 
ray black, pectoral dusky behind. 

LEeNneTH.—250 mm. 


275 


NEW RECORDS OF SOUTH AFRICAN FISHES 173 


LocaLity.—Durban. 

The main growth changes in crenilabis appear to be an 
increase in length of the pectorals (from 1-3 in to 1-1 times head), 
of the pointed scale at the dorsal base, and of the ventrals, 
while the two dorsals become more widely separated. Also the 
interorbital is wider, the postorbital is shorter, and a scaly 
process develops in the pectoral axil. 


Family HoLocentripz. 
Holocentrum sammara Forsk. Pl. XI. 


Holocentrum sammara (Forsk), Barnard, Ann. S. Afr. Mus., vol. 
xxi, p. 365, 1927. 


Depth 3-2-3-4, length of head 2-8 in length of body. Eye 2-7, 
snout 4:2, interorbital 4-1, and postorbital 2-6 in length of head. 

Preorbital serrate below; anteriorly an enlarged curved 
denticle. Preopercular spine short, not extending beyond 
subopercular margin. Two short divergent opercular spies, 
upper longer, apex beyond opercular margin. All free margins 
of opercular bones serrate. Exposed surface of post-temporal 
striate. Interorbital flat, with a longitudinal furrow, porous. 
Occiput with radiating longitudinal ridges. Posterior nostril 
without denticles. Mandibulary rami canalized and porous 
below. Mouth fairly large, horizontal, lower jaw projects 
strongly. Villiform teeth in bands in jaws, and on vomer and 
palatine. Maxilla extends to almost below centre of eye. 
Premaxillary pedicels reach to or near frontals. 8-9 gill-rakers 
on lower limb of anterior arch, slender, 2 m gill-filaments. 

D. XI, 12, inserted behind opercular margin. Ist spine 2:6, 
Qnd 2:2, 3rd and 4th 2:0 in head, thereafter graduated shorter 
to penultimate, with last spine somewhat longer. Soft rays 
anteriorly elevated, 2-4 in head. Base of spinous dorsal 1-1 
times, of soft dorsal 3-0 in head. 

A. IV, 8, inserted below middle of soft dorsal, 3rd spme 1-4 in 
head, very strong. Pectorals 1-8, ventrals 1:6 in head. Caudal 
deeply forked, lobes rounded, peduncle slender. 


276 


174 | J. L. B. SMITH. 


Scales strongly denticulate, not, or scarcely striate. 1.1. 42, 

. tr. ° ; 
7-8 
cheek scales, head otherwise naked. 

CoLour.—(Preserved specimen): Silvery red-brown grading 
lighter below. A dusky subdorsal band from nape to end of 
dorsal; a moderate dusky band embracing the course of the 
lateral line. Faint narrow lines along the scale rows. Top of 
head dusky, silvery red below. Pectorals and ventrals light. 
Spinous dorsal faint dusky with a large dark blotch between Ist 
and 4th spines. First dorsal ray, 4th anal spine, Ist anal ray 
and outer margins of caudal lobes dusky, fins otherwise faimt 
rosy. 

LenetH.—45-107 mm. 

Locauiry.—Durban : Delagoa Bay. 

According to Barnard (loc. cit.) the inclusion of this species 
in our ichthyfaunal list is based on a specimen presented by Sir 
A. Smith to the British Museum. The re-discovery of this species 
in our area after so long an interval is of interest. 


] 7 predorsal, end above hind margin of eye. 4-0 


Family PINGUIPEDIDS. 
Parapercis pulchella Day. Text-fig. 2. 


Parapercis pulchella Day; Fishes of India, p. 262, pl. Ixviii, fig. 2, 
1878-88. 


Body robust, cylmdrical. Snout conical. Depth 5-0, length 
of head 4 in body-length. Eye equal to snout, 3-6, post-orbital 
part of head 2-1 in length of head. Interorbital very narrow, 
about 3°5, preorbital depth 2 in eye. | 

Mouth moderate, lips moderately thick, lower jaw projects 
slightly, maxilla extends below anterior third of eye. Villiform 
teeth in a band in each jaw, outer row enlarged, caniniform, 
mid outer tooth largest. Fine teeth on vomer, palatines eden- 
tate. Six short gill-rakers on lower limb of anterior arch. 
Upper opercular spine extends beyond membrane margin, 
lower a group of 3 smaller spines. Preopercle margin undulate, 
more or less spinate round angle. 


TEXT-FIG. 2. 


By aT) f77 
My} a 


Parapercis pulchella Day. 


278 


176 J. L. B. SMITH. 


D. V, 21, inserted behind posterior opercular margin. 1st 
spine 10, 2nd 6, 3rd and 4th subequal 4-2 in head, 5th abruptly 
shorter. Ist ray shortest 4:0, 2nd 3-4 in head, thereafter 
increase to mid-posterior, last rays slightly shorter. Base of 
spinous fin 3 in, of soft fin twice, head length. 

A. I, 16, inserted below the 5th dorsal ray, rays slightly 
shorter than dorsal rays. 

Pectorals rounded, 1-4 in head, reach to below the 5th dorsal 
‘ray. Ventrals as long as head; 4th ray (imner) longest, reaches 
to anal origin. Caudal sub-truncate. 

Scales ctenoid, those on middle of side near lateral line with 


10-12 radiating strie. 1.1. 60, 1. tr. 2. (below spinous dorsal). 


11 or 12 predorsal, 5-6 series on cheek. Lateral line undulate, 
curves up over pectoral. Scaling on nape ends above preopercle 
margin. Muzzle naked. Caudal and pectoral with scales over 
base, other fins naked. 

CoLour.—(Preserved): Grey tinged faint reddish. 8-9 faint 
dusky cross-bars, fade out over lateral-line, reappear below. 
Two dusky to red areas on caudal base, another on pectoral 
base continued ventrally. Soft dorsal with a row of small 
black spots near apices of rays, also a row of larger dark spots 
along middle of rays. Fins light-dusky. Lower margin of 
iris light. 

(Fresh): The body variously speckled and marked with red 
or red-brown. The cross-bars not prominent. 

LEeNcTH.—65 mm. 

Locatiry.—Durban, rock-pool. 

A single specimen (H. W. Bell-Marley, 980). 

This obviously juvenile specimen is almost certainly con- 
specific with the Indian pulchella. There are minor varia- 
tions; e.g. the dorsal is inserted further back than is shown 
in Day’s figure (loc. cit.), but his figures frequently suffer 
from what may be a form of artistic distortion. Further, the 
markings are not quite identical, there being a longitudinal 
stripe along the soft dorsal in Day’s figure, whereas in my 
specimen there is a row of large dots. Also my specimen has 


279 


NEW RECORDS OF SOUTH AFRICAN FISHES 177 


one fewer anal ray than the Indian form. These variations do 
not justify specific distinction for the South African specimen. 
P. pulchella has not previously been recorded from South 
Africa. 

P. robinsoni Fowler (‘ Ann. Natal Mus.’, vol. vi, pt. 2, 
p. 261, fig. 5, 1929) may prove to be the adult of pulchella, 
but Fowler stated 31 (9/22) transverse series of scales for 
robinsoni; if that is correct then robinsoni is probably 
distinct. 


Family CaARANGIDA. 


Caranx gymnostethoides Bleeker. 
Caranx gymnostethoides Bleeker; Day, Fishes of India, p. 217, 
pl. xlviii, fig. 6, 1878-88. 

Body robust, moderately compressed, elongate. Dorsal 
profile steep from above eye, slightly concave before eyes, with 
sharp convexity above snout end. Nape with various series of 
arborescent canals. 

Depth equal to length of head, 3-3 in length of body. Eye 
5-0, snout 2-5, interorbital 3-0, and postorbital 2-3 in length 
of head. | 

Maxilla extends below anterior margin of eye. Small villiform 
teeth in bands in each jaw, no outer enlarged series; also 
minute teeth on vomer and palatines, and in a longitudinal 
median patch on tongue. Nostrils close together, twice as 
far from snout tip as anterior eye margin. 18-19 gill-rakers 
on lower limb of anterior arch, slightly longer than gill-filaments, 
1-6 in eye. 

D. VIII + I, 29, inserted an eye diameter behind pectoral 
. base. longest spine 2-5 in head. Anterior rays falcate, 1-6 in 
head; last ray enlarged, twice as long as penultimate. 

A. IL+1I, 24, anterior rays falcate, 1-6 in head, last ray 
enlarged. Anal inserted at level of hind margin of base of 
elevated soft dorsal. 

Pectorals falcate, 1:35 times head; tip reaches well beyond 
elevated part of anal. Ventrals 2-1 in head, do not reach 
vent. Caudal damaged, peduncle depressed. 


280 


178 J. L. B. SMITH. 


Scales moderate, Breast, from behind ventral base obliquely 
up halfway to pectoral base, naked. Top of head, interorbital 
and muzzle naked. 19 spinous scutes, extend forward to below 
the ante-penultimate dorsal ray. Lateral line makes a long curve 
up over the pectoral, straightens out below middle of soft dorsal. 

CoLour.—Olive above, silvery below. An indistinct small 
curved dark area near upper opercular angle. Dorsal, anal, 
ventrals, and apex of pectoral, brownish. Fins otherwise light. 

LEenGTH.—390 mm. 

Locatity.—Algoa Bay. 

A single specimen presented by the Port Elizabeth Museum. 

No specimen of gymnostethoides from South Africa has 
hitherto been described. Fowler (‘ Proc. Ac. Nat. Sci. Phil.’, 
vol. Ixxxvii, p. 381, 1935) has recorded a juvenile of that species 
from Natal, without description, and with brief notes which 
do not clearly indicate whether the identification was positive or 
not. In genera such as Caranx Lacep., in which the differen- 
tiation of juveniles is not easy, a full description of any specimen 
upon which a new record is based is desirable. 

The present specimen is almost certainly conspecific with 
gymnostethoides Bleeker} Day’s figure (loc. cit.) shows 
the anal fin to originate considerably further forward than in 
my specimen, and the whole of the area below the pectoral base 
naked. Also the scutes are shown larger, and to extend further 
forward than mm my specimen, while the anterior dorsal and 
anal lobes are neither falcate, nor as elevated as in my specimen. 

The occurrence of gymnostethoides as far south as Algoa 
Bay is noteworthy. 


Family CHZTODONTID. 
Chetodon trifasciatus Mungo Park. 


Chetodon trifasciatus Mungo Park; Fowler, U.S. Nat. Mus., Bull. 100, 
vol. viii, p. 65, 1929 (references and synonymy). 


Body ovate, dorsal profile only moderately steep, slightly 
undulate before eye. Snout moderately blunt, lower jaw 
projects slightly. 


281 


NEW RECORDS OF SOUTH AFRICAN FISHES. 179 


Depth 1-6-1-7, length of head 3-3 in length of body. Eye 
34-35, snout 4-2, interorbital 3-5, and postorbital 2-6-2-8 in 
length of head. Maxilla extends to below anterior nostril. Gill- 
rakers small, alternately very short, 15 on lower limb of anterior 
arch. | 

D. XIII, 22-23, imserted behind pectoral base. Spines 
increase to the 5th, thereafter subequal. Base of spinous fin 
about twice that of soft. Soft fin rounded. 

A. TI, 19, soft fin rounded. Pectoral 1-1, ventral 1-2 in 
head ; latter fin reaches vent. Caudal sub-truncate. 

Scales of equal size, rows slightly obliquely up and back. 
1. 1. 31, ends below soft dorsal origin. Lateral series 38, 1. tr. = 
from dorsal origin. 

CoLour.—Light grey-yellow. 3 curved dark cross-bars on 
head, anterior over muzzle, next, widest, from nape through 
eye to isthmus joining fellow, less than eye ; hindmost, narrowest, 
from below dorsal origin over postocular down to interopercle. 
A narrow light-blue line along each scale row on side. Soft 
dorsal with a narrow dark sub-marginal line; another lower. 
From peduncle upwards to soft dorsal a lanceolate dark patch, 
bordered yellow. Soft anal with a narrow dark marginal line, 
fin deep orange to a sub-basal heavy dark band, bordered yellow. 
Peduncle orange, to a dark bar across middle of caudal rays. 
Pectorals and ventrals light yellow. 

LeneTH.—35-80 mm. 

Locatiry.—East London: Delagoa Bay. 

A characteristic species, with contrasted vivid colour pattern, 
now recorded for the first time from South Africa. 

In recent years the normal flow of the Mozambique current 
along the shore at East London is periodically interrupted and 
replaced by much colder water. When the warmer water is 
inshore, numerous tropical species are found in the rock-pools, 
but are not observed when the colder water remains. Usually 
these tropical species are most abundant inshore just before the 
appearance of the colder water. 


282 


180: J. L. B. SMITH. 


Chetodon xanthocephalus Bennett. Text-fig. 3. 


Fowler, U.S. Nat. Mus. Bull. 100, vol. viii, p. 82, 1929 (references and 
synonymy); and Chetodon nigripinnatus Desjardins, loc. cit. 
p. 83. 


Body elevated ovate. Dorsal profile steep, deeply concave 
above eye to snout. Snout sharp, sub-conical. 

Depth 1-6, length of head 3-0 in length of body. Hye 3:8, 
snout 3-0, interorbital 3-2, and postorbital 2-4 in length of head. 
Maxilla extends between tip of snout and anterior nostril. 
Gill-rakers small, 10-11 on lower limb of anterior arch. 
Preopercle margin feebly serrate. 

D. XIV, 25, inserted above hind edge of opercle. Spines 
increase to the 6th, and thereafter remain subequal. Base of 
spinous fin about 1-6 times that of soft. Edge of soft fin broadly 
rounded. 

A. ITI. 24, edge of soft fin broadly rounded. 

Pectoral 1-4, ventral 1:3 in length of head, latter fin reaches 
vent. Caudal truncate. 

Scales above lateral line graduated upwards smaller than those 
below. Rows below lateral line more or less horizantal, slightly 
inclined upwards. Rows above lateral line run with dorsal 
profile. 1.1. 36, 1. tr. 11/19, from dorsal origin. 45 predorsal, 
7-8 across cheek. Soft dorsal and anal scaly to submarginal 
band. 

CoLouR.—(Preserved): Yellow-brown, nape dusky. Very 
narrow faint ocular cross-bar, visible just above and below eye. 
5-6 faint lines across upper part of body. Soft dorsal and anal 
dark to near margins, with famt darker marginal line, outer 
margin of fin light or yellow. Membrane of spinous dorsal 
basally dark, with posterior 6-7 spines in dusky extension from 
soft fin. Ventrals, caudal and pectorals light. (Living): “ Six 
curved violet limes across body. Below eye, back of pectoral to 
vent yellow. Dorsal, and ventrals rich chrome yellow with 
paler borders. Spimous dorsal orange red above, also caudal 
base. Margms of caudal yellow, centre violet.” (H. W. 
Bell-Marley.) 


TEXT-FIG 3. 


S: 


$4) 


* 


Had ayal) 


‘ “rth FY 
Taras acaad at 


ANS 


Chetodon xanthocephalus Bennett. 


284 


182 J. L. B. SMITH. 


Lenetu.—140 mm. 

Locatity.—Durban (H. W. Bell-Marley, 966). 

A rare species which has not been taken for very many years 
in South African waters. 

The brilliant coloration of the living fish fades on preservation 
so that unless the changes are followed, it might easily be 
supposed that the fresh and the preserved forms represent 
distinct species. Probably on that account the species has 
several synonyms; nigripinnatus Desjardins is certamly 
synonymous. 


Chetodon kleinii Bloch. 


Chetodon kleinii Bloch; Barnard, Ann. 8S. Afr. Mus., vol. xxi, p. 614 
(Mozambique), 1927; Fowler, U.S. Nat. Mus., Bull. 100, vol. viii, 
p. 113, 1929; Fowler, Proc. Ac. Nat. Sci. Phil., vol. lxxxvi, p. 480, 
fig. 45 (cingulatus Fwir.), 1934; Fowler, sbid., vol. Ixxxvii, p. 
396 (cingulatus), 1935. 


This species is exceedingly uncommon in our area. Recently 
two specimens have been obtained from Durban, the larger of 
which was obviously identical with that described by Fowler as 
cingulatus. It proved also identical with the Mozambique 
specimen described by Barnard as kleinii: The latter speci- 
men was therefore sent to the British Museum, and Mr. Norman 
has confirmed (in litt.) its identity with kleinii. 

The sole difference between cingulatus and kleiniiis that 
the former has not two diffuse broad bands across the body from 
the dorsal. Later (1935) Fowler has identified as cingulatus 
a specimen from Durban which shows signs of these sub-dorsal 
markings ; thus indirectly the identity of cingulatus with 
kleinii has been recognized. 

The smaller specimen from Durban is just emerging from the 
Tholichthys stage. The ocular band is only faintly marked, 
and is somewhat broader than in the adult. The body is other- 
wise of uniform colour. 


285 


NEW RECORDS OF SOUTH AFRICAN FISHES. 183 


Family LuTianip. 
Pristipomoides argyrogrammicus C.¢& V. Text-fig. 4. 


Pristipomoides argyrogrammicus (C. & V.); Fowler, U.S. Nat. 
Mus., Bull. 100, vol. xi, p. 189, 1931. 

Body elongate-ovate, moderately compressed. Depth equal 
to head, 3-6 in length of body. Eye 4:3, snout 3-1, interorbital 
3-4, and postorbital part of head 2-2 in length of head. Preorbital 
depth 1-6 in eye. An obscure transverse depression on head 
behind eyes. Posterior nostril circular, small. Preopercle 
margin indistinctly serrate. Exposed blade of post-temporal 
striate. | 

Mouth moderate, maxilla exposed, extends below anterior 
fourth of eye. An outer enlarged somewhat irregular series of 
conical or caniniform teeth in each jaw ; behind these an inner 
band of villiform teeth. Fine teeth in bands on vomer and 
palatines, incurved on latter. 14 gill-rakers on lower limb of 
anterior arch. : 

D. X, 11. first spine very slender, less than eye, 4th and 5th 
subequal, longest 2:6 in head, remainder shorten to 10th, 3 in 
head. Anterior soft rays as high as last spine, shorten pos- 
teriorly, last ray much enlarged, 1-9 in head, reaches almost to 
caudal base. No notch in dorsal fin. 

A. III. 8, soft fin similar shape to dorsal, last (long) ray 2 in 
head. Dorsal and anal naked. 

Pectorals 1-1 in head. Ventrals 1-3 im head, do not reach 
vent. Caudal lunate, scaly to near margin. 


ig 6 cheek scales, 
16 predorsal, end behind eye. 8 across opercle. Preopercle 
flange naked. Interorbital, snout, muzzle (and preorbital) naked. 

CoLour.—(Preserved): Reddish. Operculum, snout and top> 
of head darker. Scales, especially dorsal, with dusky margins. 
(Fresh) : “ Rosy, shot pale violet. Gamboge edges to scales give 
appearance of lines. Pectoral yellow. Outer rays of ventrals 
pale blue. Anal rosy with red patch anteriorly. Caudal red, 
gamboge tips.” (H. W. Bell-Marley.) 

VOL. VII, PART 2. 14 


Scales feebly ctenoid. 1.1. 61, 1. tr. 


TEXT-FIG. 4. 


Mey 


o) 


des argyrogrammicus (C. & V.). 


ipomoi 


Prist 


287 


NEW RECORDS OF SOUTH AFRICAN FISHES. 185 


Lenetu.—415 mm. 

Locauiry.—East London. 

A single specimen (H. W. Bell-Marley, No. 946). 

I have seen the South African Museum specimen from Natal, 
described by Barnard (‘Ann. 8. Afr. Mus.’, vol. xxi, p. 648, 1927) 
as filamentosus C. & V. That author stated palatine teeth 
to be absezit, but they are present in the specimen, and beiig 
incurved are not easy to detect. The South African Museum 
specimen and mine are unquestionably conspecific. The gill- 
raker count given is that of the former, the gills having been 
removed from mine. | 

Fowler’s treatment of the Indo-Pacific species of Pristi- 
pomoldes Blkr. (loc. cit., pp. 186-193) is not very satisfactory ; 
e.g. he uses the number of gill-rakers as his primary key- 
character, and gives 12-13 for microdon Stewndachner, and 15- 
16 for argyrogrammicus. Then, in a footnote to the 
description of the former species he states that microdon 
from Okinwa has 15, from Hawaii 16-17 gill-rakers. 

The South African species agree best with the diagnosis of 
argyrogrammicus C. & V., which is the oldest species. In 
Fowler’s descriptions, so little difference of any significance 1s 
shown between the four nominal species that it would not be 
surprising to find the eventual recognition of only one somewhat 
polymorphous form. 


Family PLECTORHYNCHIDS. 


Pomadasys stridens (Forsk). 


Pomadasys stridens (Forsk) ; Fowler, U.S. Nat. Mus., Bull. 100, vol. 
xi, p. 319, 1931 (references and synonymy). 


Body robust, slightly compressed, not very deep. Head 
rather broad, mouth small. Profile of snout low. Depth 3-1, 
length of head 3-0 in length of body. Hye 3-7, snout 3-1, inter- 
orbital 3-3, postorbital 2-1 in head length. Preopercle with 
straight hind edge, finely serrate. Maxilla extends to below 
posterior nostril. 9-10 stout gill-rakers on lower limb of anterior 


28 8 


186 J. L. B. SMITH. 


arch, 3-5 in gill-filaments, which are 1-7 in eye. A defined pit 
and two anterior pores on chin. 

D. XII. 14, inserted above just behind opercular margin. 
Spines moderate, Ist 7:5, 2nd 3-8, 3rd 2°8, 4th and 5th longest 
2:2 in head ; remainder decrease to the penultimate, and 12th 
spine slightly longer. Anterior rays longest subequal 3 in head. 
Edge of soft fin gently convex. Base of spinous fin 1-1, of soft 
fin 1-6 in head. 

A. III 7, inserted below the 4th dorsal ray ; 2nd spine 2:5 in 
head, longer and stronger than 3rd. Soft rays 2-2 in head, edge 
of fin gently concave, heavily scaled for ? length. 

Pectoral 1-1 in head, reaches above anal origin. Ventrals 
do not reach vent, 1-6 in head. Caudal moderately forked. 

Scales strongly ctenoid. Lateral series about 65, 1. tr. 0 
(1. 1. to belly). Tubular lateral line scales 52. About 56 predorsal 
scales, extend to above nostrils. 10-12 cheek scales, and 5-6 
more on preopercle flange. Anterior lateral line tubes bifid, 
single on peduncle. 

CoLtour.—Dark brown, slightly lighter below. An indistinct 
fairly broad dusky stripe from opercle to caudal; traces of 2 
narrow stripes above. Opercular spot doubtful. Fins all dusky, 
caudal darkest. 

LrenetH.—Up to 177 mm. 

LocaLity.—Durban. 

Two specimens taken on lines near Durban (H. W. Bell-Marley, 
4/35). 

This is the first record from South Africa of this Indian 
species. 

Fowler (° Proc. Ac. Nat. Sci. Phil.’, vol. Ixxvii, p. 230, 1925) 
described as striatus G. & T. a specimen from Delagoa Bay. 
Actually that specimen was probably stridens (see Barnard, 
‘Ann. 8. Afr. Mus.’, vol. xxi, p. 680, 1927). Also Fowler (‘ Proc. 
Ac. Nat. Sci. Phil.’, vol. lxxxvi, p. 467, fig. 39, 1934) has described 
two specimens as striatus G. & 7. which cannot be that species, 
but are obviously conspecific with my present specimens. 
Actually my specimens and Fowler’s differ in certain particulars 


289 


NEW RECORDS OF SOUTH AFRICAN FISHES. 187 


from the usual diagnoses of stridens, and it is possible that 
they represent a new species. This cannot be determined with 
certamty without comparison with an Indian specimen of 
stridens. I have compared the specimens described above with 
one of striatus G. & T. of the same size ; the two are certainly 
not conspecific. P. stridens is of much darker colour, has a 
smaller mouth, and a more robust but shallower body than 
striatus. Also the longitudinal stripes are much clearer in all 
stadia of the latter species than in stridens. 


Family Laprip”. 
Pteragogus opercularis Peters. Text-fig. 5. 


Pteragogus opercularis Peters; Barnard, Ann. S. Afr. Mus., vol. 
xxi, p. 754, 1927. 

Body oblong, compressed. Dorsal profile concave before eye, 
nape prominent. Depth 2-3, length of head 2-8 in body length. 
Kye 6-0, snout 3-0, interorbital 5-2, and postorbital 2-0 in length 
of head. Preorbital depth slightly greater than eye. Maxilla 
completely concealed beneath preorbital, extends to below 
anterior margin of pupil. 4 anterior caniniform teeth in the 
front of each jaw, the median pair small and erect, the outer 
curved, flaring out and back. 2-3 small posterior canines at 
hind edge of upper jaw. A single series of stout conical teeth 
along the side of each jaw, close-set, forming a cutting edge. 
Outer gill-arch more or less reduced, rakers 5 or 6 on lower 
limb, rudimentary. Gill-membranes hidden below interopercles, 
united, form a moderate fold across the throat. Preopercle flange 
spinate, 24-26 spinules down to lower angle, lower margin entire. 

D. XI, 9, inserted above middle of opercle, 1st spine 1-2 times 
eye, increase to the last, about 2-5 times eye. Membrane 
between spines extended, lobed. Anterior soft ray equal to last 
spine, increase to the 6th, 3 times eye, edge of fin rounded. 
Depressed rays reach caudal base. | 

A. III, 9, inserted below the last dorsal spines. Soft fin in 
shape and size like soft dorsal, depressed rays reach beyond 


caudal base. 


TEXT-FIG. 5. 


+ RT 


Pteragogus opercularis Peters. 


291 


NEW RECORDS OF SOUTH AFRICAN FISHES. 189 


P. 13, 1-8 in head, rounded. Ventrals with first ray filamen- 
tous, 1:8 in body length; tip reaches almost to caudal base. 
Caudal broadly rounded. 

Scales large, upper dorsal and lower ventral scales produced, 
forming a sheath for the dorsal and anal fins. Lateral line 
continuous, abruptly bent before peduncle. 1. 1. 24, 1. tr. 2/6 
(including dorsal sheath). Cheek scales concealed beneath thick, 
spongy skm. Opercular bones scaly, scales partly concealed. 
Lateral line tubes of anterior scales with an arborescent superior 
branch. 5 predorsal scales. 

CoLour.—Olive-green above, lighter below. Numerous dark 
spots on forehead and above and behind eye. A large ocellus 
covering most of opercle. Dark edges to scales above the lateral 
line. Fainter dark spots in a series above and below the lateral 
line tubes. A dark spot between each pair of the 1st-4th dorsal 
spines. Mid-posterior dorsal rays dusky. Caudal reddish. outer 
rays alternately ight and dark. Iris golden. 

Lenetu.—140 mm. 

Locatity.—Durban (H. W. Bell-Marley, 4/35). 

A single specimen of this species. Previously recorded from 
Mozambique by Peters in 1855, but not hitherto found in Natal 
waters. 


Family GEMPYLID. 
Thyrsitoides marleyi Fowler. 


Thyrsitoides marleyi Fowler, Ann. Natal Mus., vol. vi, pt. 2, p. 255, 
fig. 2, 1929. 


Skin very thin and papery. Body elongate, very compressed, 
tapers to peduncle. Depth 8°9, length of head 3:6 in length of 
body. Eye 7:0, snout 2-2, interorbital 5-4, and postorbital 2-6 
in head length. 

Mouth very large ; lower jaw projects strongly; maxilla extends 
below anterior margin of eye. A single row of pomted trenchant 
teeth in each jaw. Anteriorly in the upper jaw several very 
large slender fixed pointed somewhat compressed teeth, one 


292 


190 J. L. B. SMITH. 


smaller anteriorly, depressible. Vomer edentate, fine teeth on 
palatines. Tongue covered with minute teeth. Gill-rakers 
rudimentary, 4 or 5 fine bi- or trifid processes near angle on lower 
margin of anterior arch; remaining upper surface of arch with 
fine villiform teeth. Guill-membranes not united, free from 
isthmus. Nostrils wide apart, anterior small, circular, posterior 
slit-like. Free margin of preopercle round angle undulate, 
probably spinate in young. Interorbital with a longitudinal 
wide furrow. | | 

D. XVIII, 12 + 5, originates above middle of opercle. 
Anterior spines longest, Ist 2-8, 2nd 2-3, 3rd and 4th broken, 
5th 3-8 in head, thereafter shorter. Anterior rays falcate, 
3-5 in head. 

A. 13 + 4, inserted below soft dorsal; anterior rays falcate, 
somewhat lower than soft dorsal. Pectorals subfalcate, 2:5 in 
head. Ventrals close together, first (outer) ray longest, 3-5 in 
head. Caudal deeply forked, peduncle cylindrical. 

Body largely covered with very fine cycloid scales. densest on 
nape, along back from peduncle. A patch behind eye, head 
otherwise naked. Two lateral lines, lower branching from upper. 
Upper runs from upper angle of operculum to the base-of the 
16th (left) or 17th (right) dorsal spme. On the right side this 
line has a break with a small inferior branch below the 13th 
dorsal ray. 190 scales in upper line (24 to where lower branch 
originates). The lower line originates just behind the base of 
the 4th dorsal ray, and curves down to run along the middle 
of the side to the caudal base. About 310 scales in the lower 
branch. 

CoLour.—Slaty-blue, shading to grey on belly. Membrane 
of spinous dorsal, and tips of caudal and pectoral black. 

LenetTH.—590 mm. | 

Locatiry.—East London. 

A single specimen, presented by the Curator of the East 
London Museum. This is an interesting rediscovery of a species 
known hitherto only from the type, from which the present 
specimen differs only in minutie, despite the disparity in 
size. 


293 


NEW RECORDS OF SOUTH AFRICAN FISHES. 191 


Family ScomBRIDA. 
Sarda chilensis C.& V. Text-fig. 6. 


Sarda chilensis C. & V.; Barnard, Ann. 8. Afr. Mus., vol. xxi, p. 800, 
1927. 

Body fusiform, moderately compressed. Snout conical, 
pointed. Depth 4-5, length of head 3-5 in body length. Eye 
6-0, snout 2-8, mterorbital 3-4 and postorbital 2:0 in length of. 
head. Preorbital depth half eye. 

Mouth large, maxilla distally expanded, spatulate; extremity 
extends below hind margin of eye. 17-18 moderate compressed 
inwardly depressible acute teeth in a single series in each side of 
upper jaw. In lower jaw 4 or 5 large sub-conical depressible 
teeth anteriorly ; along each side 5-6 large compressed inwardly 
depressible teeth, with a few smaller mtermediate. Vomer 
edentate, palatines with fine sharp teeth. Gull-membranes 
united, free from isthmus. 8 gill-rakers plus one anterior 
rudiment on lower limb of anterior arch, longest 1-6 in gill- 
filaments, which are equal to eye. Adipose eyelids moderate, 
not to pupil. Interorbital broadly convex. with a slight median 
longitudinal furrow. 

D. XIX, 14 + 7, inserted above hind margin of operculum. 
Ist spine 2-8, 2nd and 3rd subequal, longest 2-7 in head ; 
remainder graduated shorter, last minute. Soft rays anteriorly 
falcate, highest 3-2 in head. Base of spinous dorsal 1-2 times 
head. 

A. III, 12 + 6, insertéd below end of base of continuous soft 
dorsal. In shape and size much as soft dorsal. 

Pectorals 2:6 in head, ventrals 3-2 in head, inserted below 
pectorals. Caudal lunate. Peduncle depressed, with a single 
stout lateral keel, base of caudal with two auxiliary keels, one 
above and one below end of peduncular keel. 

Scales minute, over whole body. Enlarged scales form a 
corselet over pectoral region, from isthmus up, embracing 
pectoral, extending up and forwards over nape; a long narrow 
extension back along base of spinous dorsal. Lateral line curves 
up over pectoral, thence obliquely down to peduncle. 


TEXT-FIG. 6. 


= 
SEHR 


eee 


* 
a 
; 

t 


Sarda chilensis C.& V. 


295 


NEW RECORDS OF SOUTH AFRICAN FISHES. 193 


CoLour.—Steel-blue to black above, lighter below. Head 
blackish above, with curved dark post-orbital bar. Body with 
6-7 narrow longitudinal dark stripes, longest along middle of 
side. 12-13 dusky cross-bars, slightly greater than interspaces 
fading out on abdomen. Spinous dorsal black anteriorly and 
posteriorly, mid-spines lighter. Soft dorsal dusky. Caudal 
dark with median light patch in each lobe. Pectorals with 
distal central dark patch, margin light. Ventrals dusky 
proximally. Anal and all finlets light. 

LeneTH.—310 mm. 

Locatiry.—Durban. 

A single specimen (a ripe female) (H. W. Bell-Marley, 978). 
Seldom taken in our area. 

The Atlantic species sarda C. & V. is generally held to be 
distinct from chilensis C & V., chiefly in having 21 dorsal 
spines and the maxilla extending behind the hind margin of the 
eye. The two species are evidently very closely related, and 
comparison of equivalent stadia may eventually reveal that 
they are identical. 


Family BLENNIID. 
Blennius fascigula Barnard. 


Blennius fascigula Brnrd; Smith, Rec. Alb. Mus., p. 152, pl. xvi, 
1931 ; Fowler, Proc. Ac. Nat. Sci. Phil., vol. lIxxxvu, p. 404, fig. 36, 
1935 (trifascigula). | 


A specimen of this species, from Durban, has been received 
from Mr. Bell-Marley, who also sent to Fowler the specimens 
upon which was based the species trifascigula Fowler as 
distinct from fascigula Barnard. Fowler has evidently not 
seen the account of specimens of fascigula from East London 
(Smith, loc. cit.). I have compared the specimen from Durban 
with those from East London, and with the type of fascigula, 
and they are all without doubt conspecific. trifascigula 1s 
most certainly identical with both the Durban and the Kast 
London specimens, and cannot therefore be maintained as 
distinct. 


296 


194 J. L. B. SMITH. 


Petroscirtes tapeinosoma Blkr. 


Petroscirtes tapeinosoma Blkr.; Smith, Rec. Alb. Mus., vol. iv, p. 
213, pl. xxi, B, 1935. 


The first record from South Africa of this Indo-Pacific form 
was a specimen taken at Great Fish Point in 1934. Two 
specimens have recently been obtained in a rock-pool near Kast 
London, so that the species is evidently established in our area. 


Family CLINID. 


NOTE ON THE DISTRIBUTION OF SPECIES OF CLINUS Cuv. IN 
Souto AFRICA. 


Until fairly recently it has been accepted that species of 
Clinus Cuv. were largely confined to the colder waters charac- 
teristic of the coasts of the west and south-western Cape. This 
was chiefly due to the fact that the only part of South Africa in 
which intensive collecting of these fishes had been carried out 
was that particular area. 

Within the last few years investigation has shown that 
Clinus species extend very much further east into the warmer 
waters of the Indian Ocean than had previously been suspected. 
Quite a number have been found to occur regularly along the 
coast eastwards from False Bay, at least as far as East London. 
Quite recently I have found a number of Clinus species 
exceedingly abundant as far east as the mouth of the Great Kei 
River. It is most probable that continued investigation will 
show that these species extend even still further eastwards. 
One species, heterodon, C. & V., has actually been traced as 
far as North Zululand (Smith, 1934). 

The results of recent investigations merit a revised account of 
the distribution of Clinus species in South Africa, and this is 
summarized below. 


297 


NEW RECORDS OF SOUTH AFRICAN FISHES. 195 


Clinus species. Distribution. 
acuminatus C. & V. Walfisch Bay to False Bay. 
anguillaris C. & V. False Bay to Great Fish Point (S.): 
Durban (B.M.). 
Knysna estuary. 
False Bay to Keimouth (S.). 
False Bay ; Great Fish Point (S.). 


agilis J. L. B. Smith 
brachycephalus C. & V. 
brevicristatusG. dT. . 


capensis C.¢d V. . 


cottoidesC. d& V.. 


dorsalis Blkr. 
fucorum G. & T. 


heterodonC. & V. 
laurentiiG. @ T. . 
latipennis C. & V. 


mentalisG. & T. 
musG.¢T.. 
ornatusG. & T. 
pavoG.d T. 


robustusG. & 7. 


striatus G. & T. 


superciliosus Linn. 


taurus G.d& 7. 


venustrisG. ¢& T.. 


Lambert’s Bay (S.) to East London 
(S.): Durban (B.M.). 

Lambert’s Bay (S.) to Durban (B.M.). 

Walfisch Bay to Keimouth (S.). 

False Bay: Port Alfred (8.). 

False Bay to North Zululand (S.). 

Port Alfred (S.) to Zululand (B.M.). 

False Bay. 

East London: Durban (B.M.). 

False Bay to Keimouth (S.). 

False Bay : Knysna (S.). 

False Bay (Great Fish Point): 
Keimouth (S.). 

Table Bay to East London (S.): 
Durban (B.M.). 

False Bay ; Port Alfred (S.); Durban 
(B.M.). 

Walfisch Bay to Durban (B.M.). 

Cape Peninsula: Plettenberg Bay (S8.). 

False Bay : Knysna (S.). 


Continuity of observed distribution is indicated by “to” between 
localities ; any considerable gap between recorded areas is indicated by a 


colon. 


(B.M.) = extreme locality fide H. W. Bell-Marley. 
(S.) = extreme locality fide J. L. B. Smith. 
Other limits as given by Barnard (° Ann. 8. Afr. Mus.’, vol. xxi, pp. 852-— 


866, 1927). 


Nemacoclinus navalis Barnard. 


Nemacoclinus navalis Barnard, Ann. 8. Afr. Mus., vol. xxx, p. 646, 


fig. 1, 1935. 


Body compressed, elongate. Depth 5-8, head 4:5 in body 
length. Snout pointed, sub-conical. Hye 4-3, snout 4:4, and 
interorbital 7-5 in head length. 


298 


196 J. L. B. SMITH. 


Mouth moderate, maxilla extends to below middle of eye. 
Teeth in jaw subequal, minute teeth in a band on vomer. A 
short, apically dilated, fringed supraorbital tentacle. Mucous 
pores on head opening by a series of transversely opposed pores. 

D. XXXIV, 1 originates over hind preopercle margin. Anterior 
three spines widely spaced, first two remote from third. 1st and 
4th spines shortest, scarcely greater than eye, remainder sub- 
equal about 1:7 times eye. Soft ray shorter than last spine; 
membrane reaches caudal base. 

A. II, 23; V, I, 3, inner ray small, adnate. Caudal rounded. 

Lateral line complete, almost to caudal base; tubes with 
transverse tubes above and below, each opening by a pore, 
exactly opposed. 80 cross tubules. 

CoLour.—Olive-green, mottled red-brown, fins greenish. 

LeNeTH.—66 mm. 

Locatity.—Great Fish Point. 

A single specimen, taken in a rock pool. 

This species has been known only from the type, taken at 
Simonstown. Its rediscovery at a point so far distant is of 
interest. Since the very intensive collecting of Clinids about 
the Cape Peninsula by the late W. Thompson, no new species 
have been taken in that area until navalis was discovered in 
1934. Durmg the past six years I have collected Clinid fishes 
most intensively from a stretch of rocks at Great Fish Point, so 
that had navalis been a normal inhabitant of that area, it 
could scarcely have escaped detection. Actually my specimen 
was taken only a short time after the discovery of the orthotype 
at Simonstown. In the circumstances it might be supposed 
that navalis has but recently appeared on our coasts. 

Barnard has regarded navalis as generically distinct from 
Clinus Cuv. on the character of the lateral line, which in that 
latter genus is generally stated to have the lateral line tubes 
simple, opening by a single pore. Actually several species 
(e.g. capensis C. & V.) have at least the anterior portion of 
the lateral line of structure similar to that in Nemacoclinus. 

Our species of Clinus Cuwv. are in need of revision, since they 
fall into groups which appear to merit generic distinction, on 


299 


NEW RECORDS OF SOUTH AFRICAN FISHES. 197 


grounds at least as valid as those accepted by Barnard for genera 
such as Nemacoclinus Brnrd., Clinoporus Brnrd., and 
Petraites Ogilby. | 


Family GopiuD4a. 


Gobius callidus nom. nov. 


Gobius gulosus J. L. B. Smith, Trans. Roy. Soc. 8.A., vol. xxiv, pt. I, 
p. 49, fig. 2, 1936. 
It has been found that the name gulosus has previously been 
used for an American species of the genus, callidus is therefore 
proposed to replace gulosus for the South African species. 


I wish to express my gratitude to the Research Grant Board 
of South Africa (Carnegie Fund) for financial assistance in this 
work. 


Albany Museum, 


Grahamstown ; 
August, 1936. 


EXPLANATION OF PLATE XI, 


Illustrating Dr. J. L. B. Smith’s paper, ‘‘ New Records 
of South African Fishes ”’. 


Holocentrum sammara fForsk. xX 1:7. 


Plicklk 


Ann. Natal Mus., Vol. VIII. 


ae oar | 


HOLOCENTRUM SAMMARA Forsk 


Ltd. 


Adlard & Son, 


INDEX TO VOLUMES | and Il 


Volume | ends at page 300 


VALID SCIENTIFIC NAMES are in plain capitals. 
Page in heavy type shows a description on that page. 


Synonyms and Malidentifications are in italics. 


An asterisk shows that there is an illustration on that page in the text. 


Acanthidium natalense 272 
ACAN THIDIUM QUADRISPINOSUM 270, 


ACANTHOCEPOLA 173 
AGANTHOCEPOLA CUNEATUS 171-3 PI. 


ACANTHOCEPOLA LIMBATA 172 
ACANTHOPAGRUS 307, 311, 313, 314-5, 


ACANTHOPAGRUS' BERDA 311, 314, 
315—6*—7, 320 Pl. 18 

ACANTHOPAGRUS BIFASCIATUS 311, 
315, 317-8*—320 Pls. 18, 24 

ACTINISTIA 389, 391, 393, 412*, 415 

ACUMINATUS, CLINUS 11, 297 

ACUTIPENNIS, GOBIUS 386 

ADENI, PERISTEDION 69, 71-3 Pl. 22 

AENEOFUSCUS, GOBIUS 233, 234 

AEROSTATICUS, (TETRODON) 208 

AESTUARIUS, (ACANTHOPAGRUS) 317 

AESTUARIUS, GILCHRISTELLA 235 

AENEUM, PACHYMETOPON 311, 363, 364 
Pls. 22, 29; p. 526 

AFER, BARBUS 267 

AFRICANA, TRIGLA (TRIGLOPORUS) 
73, 74-6 Pl. 23 

AGILIS, CLINUS 10, 11 Pl. 16; p. 297 

AGONOSTOMUS 79 

ATBOFASCIATUS, BATRICHTHYS 59, 60 


ALBULA VULPES 51 

albus, Dentex 370, 373 

ALEPISAURIDAE 154 

ALEPISAURUS FEROX 154 

ALGOAE, CAESIO 173 

ALTAVELA, (PTEROPLATEA) 46, 47 

ALUTERES 238, 240-1 

ALUTERES MONOCEROS 238, 240, 241-4 
Pls. 40, 41, 43 

ALUTERES SCRIPTUS 240-1, 243-4 Pl. 42 

ALUTERIDAE 238, 239 

AMBLYAPISTUS BINOTATA 543 

AMBLYAPISTUS MARLEYI 12 

Amblyapistus marleyi 543 

AMIA 501 

Amphiprionichthys zeylonicus 13 

AMPHIPRIONIDAE 183 

ANABAS 230 

ANABAS BAINSII 230, 231 Pl. 33 

ANABAS CAPENSIS 230-1, 237 Pl. 33 

ANABAS SCANDENS 230 

ANABAS TESTUDINEUS 230 

Anabas vicinus 231 

ANABATIDAE 218, 230 


ANGLICUS, CHR YSOBLEPHUS 311, 340-1, 
347-8*, 380 Pls. 19, 24; p. 526 

ANGUILLA MOSSAMBICA 227 

ANGUILLARIS, CLINUS 11, 197, 297 

ANGUILLARIS, PLOTOSUS 226 Pl. 34 

ANGUILLIDAE 218, 227 

ANOPLUS, BARBUS 219, 220, 231 Pl. 29; p. 
267 

APHAREUS 535 

APLOACTIDAE 202 

AQUILA, (MYLIOBATIS) 147, 148 

ARGENTATUS, (CLINUS) 197 

ARGYROGRAMMICUS, PRISTIPO- 
MOIDES 28S, 286*-7 

ARGYROPS 305, 311, 313, 333, 370, 525 

ARGYROPS FILAMENTOSUS 311, 333- 
335, 525 

ARGYROPS SPINIFER 311, 333, 334-5 Pls. 
20, 25; p. 525 

ARGYROZONA 374, 378 

argyrozona, Dentex 378 

ARGYROZONA, POLYSTEGANUS 308%, 
368, 374, 378 Pls. 21, 27 

ARMATUS, (GRAMMONUS) 53 

ATER, (GRAMMONUS) 53 

ATHERINIDAE 156 

ATLANTICUS, (HOPLOSTETHUS) 162 

atricauda, (Sardinella) 150 

auratus, Mugil 80-1, 106, 108, 120, 122-3 

AURATA, (SPARUS) 314, 320 

AURIVENTRIS, AUSTROSPARUS 308%, 
311, 319-325*, 326*-7, 362 Pls. 18, 23; p. 512 

auriventris, Austrosparus 545-548 

auriventris, Diplodus 325, 547 

AUROLINEATUS, GNATHODENTEX 398, 
541 

AUSTRALIS, (ACANTHOPAGRUS) 315, 
320 . 

AUSTROSPARUS 307, 311, 313, 315, 319- 
321, 327, 330, 545, 546-7 

AUSTROSPARUS AURIVENTRIS 308%, 
311, 319-325*, 326*—7, 362 Pls. 18, 23; p. 512 

Austrosparus auriventris 545-548 

AUSTROSPARUS GLOBICEPS 311, 320, 
321-2*-3, 327-8 Pls. 18, 23; p. 545, 546-9 

AUSTROSPARUS SARBA 311, 319-323-4*, 
327-8 Pls. 18, 23; p. 545-6, 547-9 

Austrosparus sarba 325 

AUSTROSPARUS TRICUSPIDENS 546, 
548-9 

AXELIA 425, 447 

AXELIA ROBUSTA 501, 505 

AXILLARIS, (CAESIO) 174 


Volume | ends at page 300 


AXILLARIS, DASCYLLUS 183-5 PI. 22 
AXINECEPS 374, 379 


BAGRIDAE 218, 225 

BAHIENSIS, (CYPSELURUS) 160 

BAINSII, ANABAS 230, 231 PI. 33 

BALINENSIS, HEMIRHAMPHUS 22-3, 
33-4 

BALISTES CONSPICILLUM 210 

BALISTIDAE 210, 238-240 

BARBUS 218, 219 

BARBUS AFER 267 

BARBUS ANOPLUS 219, 220, 231 Pl. 29; 


BARBUS BROOKINGI 219, 221 Pl. 30 
BARBUS BURCHELLI 219, 223 Pl. 30 
BARBUS CAPENSIS 219, 221 Pl. 30 

BARBUS GILCHRISTI 219, 222 Pl. 29 


BARBUS HEMIPLEUROGRAMMA 219, 
223 Pi. 29 


BARBUS HOLUBI 216, 219, 222 Pl. 31 
BARBUS PALUDINOSUS 219, 221 Pl. 30 
BARBUS SENTICEPS 219, 220*, 266* 
BARBUS TREVELYANI 219 Pl. 29 
BARBUS VULNERATUS 219, 223 Pl. 31 
barnardi, Myxus 81, 90, 93 

BARNARDI, SYNAPTURA 4, 5* Pl. 16 
BATRACHOIDES 59 
BATRACHOIDIDAE 49, 58 
BATRICHTHYS 49, 58-9 
BATRICHTHYS ALBOFASCIATUS 59, 60 


belanak, Mugil 129 

BENNETTI, (POMADASYS) 181, 182 

BERDA, ACANTHOPAGRUS 311- 
315-6*—7, 320 Pl. 18 

BIDENICHTHYS 566 

BIDENICHTHYS CAPENSIS 566 

BIFASCIATUS, ACANTHOPAGRUS 31], 
315, 317-8*—320 Pls. 18, 24 

bifasciatus, (Austrosparus) 323, 547 

BIFILUM, BLENNIUS 190-1 PI. 20 

BINOTATA, AMBLYAPISTUS 543 

bipinnulatus, Elacate 178 

BITORQUATUS, GYMNOCRANIUS 536, 
537 

bleekeri, Doryichthys 49 

BLENNIIDAE 8, 189, 295 

BLENNIUS 191 

BLENNIUS BIFILUM 190-1 Pl. 20 

BLENNIUS CORNUTUS 189 

BLENNIUS FASCIGULA 8 Pl. 16; p. 295 

Blennius trifascigula 295 

BLOCHI], PACHYMETOPON 311, 363-4 
Pls. 22, 28, 29 

BOOPSOIDEA 304-5, 311, 350, 351 

BOOPSOIDEA INORNATA 311, 351-2 Pls. 
22, 28 

borneensis, Mugil 120 

BORO, PISOODONOPHIS 273 

BOX 304-5 

BRACHYCEPHALUS, CLINUS 11, 297 

BRACHYURUS, MICROPHIS 49*, 50 

BREVIORISTATUS, CLINUS 11, 297 

BREVIPINNIS, SCYMNORHINUS 
253-4*-5 

BROOKINGI, BARBUS 219, 221 Pl. 30 

BROTULA PALMIETENSIS 198 Pl. 21 

BROTULIDAE 52-3, 198, 306, 566 


BUCHANANI, MUGIL 81-2, 89, 113-4*-5, 
131 Pls. 16, 20 
BURCHELLI, BARBUS 219, 223 Pl. 30 


caeruleomaculatus, Mugil 95, 113, 115 

CAERULEOMACULATUS, MUGIL 128, 
130-1 

CAESIO ALGOAE 173 

calabaricus, (Hemirhamphus) 36, 39 

calabaricus schlegeli, Hemirhamphus 39 

CALLIDUS, GOBIUS 223-4*, 299 

CALLIONYMIDAE 165 

CANADUS, RHACHICENTRUM 177-8 

CANALICULATUS, MUGIL 81-2, 89, 113, 
120—2*-3, 130, 132 Pls. 16, 17, 18 

CANARIENSIS, (PTEROPLATEA) 46 

CANESCENS, PACHYMETOPON 311, 
363-4 

CANTHARUS 304 

CAPENSIS, ANABAS 230-1, 237 Pl. 33 

CAPENSIS, BARBUS 219, 221 Pl. 30 

CAPENSIS, BIDENICHTHYS 566 

capensis, Champsodon 192 

CAPENSIS, CHELIDONICHTHYS 68 

CAPENSIS, CLINUS 11, 297-8 

capensis, Coracinus 138, 144 

CAPENSIS, DICHISTIUS 137-139, 143-6 PI. 
13; p. 301-2 Pl. 10; p. 557-8*-9 

CAPENSIS, DIPLODUS 150, 331 

capensis, Dipterodon 138-9, 144 

CAPENSIS, HIPPOCAMPUS 149, 150 

CAPENSIS, LABEO 224 Pl. 31 

CAPENSIS, MERLUCCIUS 158, 206 

capensis, Mugil 108, 110 

CAPITO, MUGIL 79, 83, 88, 102, 
103—4*-5*-7, 110, 13] Pls. 17, 19 

CARACANTHUS ZEYLONICUS 13 

CARANGIDAE 174, 279 

CARANTHUS 304-5 

CARANX 280 

CARANX GYMNOSTETHOIDES 279, 280 

CARPIO, CYPRINUS 224 

cauda lunata, (Coracinus) 138 

CENTRARCHIDAE 218, 237 

cephalotus, (Mugil) 90 

CEPHALUS, MUGIL 79, 81, 83, 88, 90, 
91*-5, 123, 130, 132 Pl. 15; p. 232* 

CEPOLIDAE 171 

CERVINUS, DIPLODUS 170 

CERVUS, MYLIOBATIS 147, 148* 

CESTRAEUS 79 

CETOMIMIDAE 158 

CETOMIMUS PICKLEI 158-9 

ceylonensis, Mugil 81, 113, 115 

Chaetodipterus 188 

CHAETODIPTERUS 246 

Chaetodipterus faber 187-8 

Chaetodipterus goreensis 188, 248, 250-2 

Chaetodipterus orbis 188, 250 

Chaetodon cingulatus 284 

CHAETODON KLEINII 284 

CHAETODON MARLEYI 176-7 Pl. 22 

Chaetodon nigripinnatus 282, 284 

CHAETODON SETIFER 247 

CHAETODON TRIFASCIATUS 280-1 

CHAETODON XANTHOCEPHALUS 
282-3* 

CHAETODONTIDAE 176, 246-7, 280 


Volume | ends at page 300 


CHALUMNAE, LATIMERIA 387-8*-390, 
391-4 Pls. 3-7, 403-5*—7*-8-409-415 Pls. 
1-3; p. 425-426-522 19 figs. 44 pls. 

CHAMPSODON CAPENSIS 192 

CHARAX 304-5 

CHEIMERIUS 368, 370 

CHEIMERIUS NUFAR 368, 370-1*-3 Pls. 
21, 27 

CHELIDONICHTHYS 61, 68 

CHELIDONICHTHYS CAPENSIS 68 

CHELIDONICHTHYS KUMU 68 

CHELIDONICHTHYS QUEKETTI 68* 

CHILENSIS, SARDA 293, 294*-5 

CHORISOCHISMUS DENTEX 563 

CHRYSOBLEPHUS 311, 314, 320, 331, 335, 
338, 339-40, 551 

CHRYSOBLEPHUS ANGLICUS 311, 340-1, 
347-8*, 380 Pls. 19, 24; p. 526 

CHRYSOBLEPHUS CRISTICEPS 311, 340— 
342*-7 Pls. 20, 26; pp. 551, 552*-6 

Chrysoblephus cristiceps 343 

CHRYSOBLEPHUS GIBBICEPS 311, 337, 
340-1, 345-7 Pls. 19, 26 

CHRYSOBLEPHUS LATICEPS 311, 340, 
341 Pls. 19, 26 

CHRYSOBLEPHUS LOPHUS 311, 340-1, 
346-7, 551 

CHRYSOBLEPHUS PUNICEUS 311, 340-1, 
343-4*_5, 526, 551-2, 554-5*-6 

CHRYSOPHRYS 304 

Chrysophrys haffara 547 

Chrysophrys natalensis 323, 547 

Chrysophrys sarba 547 

CICHLIDAE 84, 218, 228 

CILIARIS, PAGRUS 334 

cingulatus, Chaetodon 284 

CLARIAS 225 

CLARIAS GARIEPINUS 225 

CLINIDAE 10, 195, 296 

CLINOPORUS 299 

CLINUS 51, 197, 296-8 

CLINUS ACUMINATUS 297 

CLINUS AGILIS 10-1 Pl. 16; p. 297 

CLINUS ANGUILLARIS 297 

CLINUS BRACHYCEPHALUS 11, 297 

CLINUS BREVICRISTATUS 11, 297 

CLINUS CAPENSIS 11, 297-8 

CLINUS COTTOIDES 297 

CLINUS DORSALIS 11, 297 

CLINUS FUCORUM 297 

CLINUS HETERODON 195 PI. 22; pp. 296-7 

CLINUS LATIPENNIS 297 . 

CLINUS LAURENTII 196—8*, 297 

CLINUS MENTALIS 197, 297 

CLINUS MUS 1], 297 

CLINUS ORNATUS 11, 197, 297 

CLINUS PAVO 11, 297 

CLINUS ROBUSTUS 196, 297 

CLINUS STRIATUS 11, 197, 297 

CLINUS SUPERCILIOSUS 11, 297 

CLINUS TAURUS 196, 297 

CLINUS VENUSTRIS 196, 297 

CLUPEIDAE 150, 218, 235 

COCCOTROPUS 202 

COCCOTROPUS JUBATUS 201 Pl. 18 

COELACANTH = see LATIMERIA 
CHALUMNAE 

COELACANTHIDAE 388*-9, 393-4, 
413-4*_5, 420-422 


COELACANTHUS GRANULATUS 437 

COERULEOPUNCTATUS, POLYSTEGA- 
NUS 368, 373, 374-6 Pls. 22, 28 

commersoni, (Hemirhamphus) 24 

COMMERSONI, (SCOMBEROMORUS) 189 

COMPRESSUS, MUGIL 81, 89, 115—7*-8, 
131 Pls. 17, 20 

CONGIOPODUS TORVUS 51 

CONSPICILLUM, BALISTES 210 

conspicillum, Pachynathus.210 

constantiae, Mugil 90 

Coracinus 137-8 

Coracinus capensis 138, 144 

CORNUTA, (MYLIOBATIS) 147 

CORNUTUS, BLENNIUS 189 

CORYPHAENOIDIDAE 154 

CORYZICHTHYS 59, 60 

COTTOIDES, CLINUS 11, 297 

CRENIDENS 304-5, 311, 357, 360 

CRENIDENS CRENIDENS 311, 360, 361* 
Pls. 22, 26 

CRENIDENS, CRENIDENS 311, 360, 361* 
Pls. 22, 26 

CRENILABIS, MUGIL 88, 99*, 100, 130, 
274-5 

CRISTICEPS, CHRYSOBLEPHUS 311, 
340-1, 342*-7 Pls. 20, 26; pp. 551, 552*-6 

cristiceps, Chrysoblephus 343 

CRISTICEPS, POLYAMBLYODON 
(LEPTOMETOPON) 527*, 528-9*, 530*-2 
Pl. 50 

CROSSOPTERYGII 411, 412*, 415 

CUBICEPS NATALENSIS 200 

CUNEATUS, ACANTHOCEPOLA | 171-3 
Pl. 21 

CUNNESIUS, MUGIL 129, 130 

cunnesius, Mugil 81, 96, 98 

CURVIDENS, GYMNOCROTAPHUS 311, 
358* 

CUVIERI, (ACANTHOPAGRUS) 315 

CYMATOCEPS 311, 314, 337-8 

CYMATOCEPS NASUTUS 311-2, 338-9*, 
382 Pls. 20, 25; p. 512 

CYPRINODONTIDAE 272 

CYPRINIDAE 218, 266 

CYPRINUS 218, 224 

CYPRINUS CARPIO 224 

CYPSELURUS HEWITTI 159 

CYPSILURUS 26 


DALATIAS 253 

DALATIAS LICHA 255 

DALGLEISHI, XENOLEPIDICHTHYS 
162-3 Pl. 18 

DARWINI, GEPHYROBERYX 162 

DASCYLLUS AXILLARIS 183-5 Pl. 22 

DASYBATIDAE 43 

DECAPTERUS LAJANG 174-5 PI. 21 

DELAGOAE, HEMIRHAMPHUS 22-3, 31-4 
Pls. 10, 12 

DELAGOAE, NEMIPTERUS 542 

DENTATA, PORCOSTOMA 311, 349, 350 
Pls. 20, 25 

DENTEX 183, 304-5, 367-8, 370, 373, 381, 
399, 533, 536 

Dentex albus 370, 373 

Dentex argyrozona 378 

DENTEX, CHORISOCHISMUS 563 


Volume | ends at page 300 


Dentex filamentosus 399 

Dentex filosus 370, 372, 376, 556 

Dentex lineopunctatus 374-6 

DENTEX MATSUBARAE 367 

Dentex miles 370, 373 

Dentex natalensis 374 

Dentex nufar 370, 371, 373 

DENTEX PERONII 367 

Dentex rivulatus 533, 537, 541 

Dentex robinsoni 533, 537 

Dentex rupestris 371 

Dentex variabilis 37\ 

DENTICIDAE 303, 308*-9, 337, 366-8, 381, 
532-534, 556 

DERMATOPSIS 566, 568 

DERMATOPSIS KASOUGAE 566—7*-8 

DERMATOPSIS MICRODON 568 

DERMATOPSIS MULTIRADIATUS 568 

DIABOLUS, MOBULA 561, 569 

diadema, Mugil 81, 115, 117-8 

DIAPHUS 17-19 

(DIAPHUS) ELUCENS, MYCTOPHUM 
152—3* 

DICHISTIIDAE 135, 137, 301, 557 

DICHISTIUS 135, 136, 137, 138, 302 

DICHISTIUS CAPENSIS 137-139, 140, 143-6 
Pl. 13; pp. 301-2 Pl. 10; pp. 557-8 *-9 

DICHISTIUS FALCATUS 138-9, 141, 142-4 
146 Pls. 14, 17 

DICHISTIUS MULTIFASCIATUS — 138-9, 
141, 144-6 Pls. 15, 16 

DIPLOCERCIDES 421 

DIPLOCERCIDES KAYSERI 477 

DIPLODUS 304-5, 307, 311, 313-5, 319, 321, 
327, 330-1, 546 

Diplodus auriventris 325, 547 

Diplodus capensis 150 

DIPLODUS CERVINUS 170 

DIPLODUS SARGUS 311, 320, 331-2* Pls. 
19, 24 

DIPLODUS TRIFASCIATUS 311, 33), 
332-3* Pls. 19, 24 

Dipterodon \37-8 

Dipterodon capensis 138-9, 144 

DORSALIS, CLINUS 11, 297 

Doryichthys bleekeri 49 

DREPANE 137, 245-6, 252 

DREPANIDAE 245-247 

Drepanoscorpis 137-8, 146 

Drepanoscorpis gilchristi 138, 144, 146 

Dulosparus 334 

DURBANENSIS, SPARODON 311, 321, 
328-9 *-330 Pls. 18, 23 

durbanensis, Sparus 328 

DUSSUMIERI, HEMIRHAMPHUS 23, 
30-1, 34 


ECHINORHINIDAE 253 

ECHINORHINUS 253 

ECKLONIACHTHYS 561 

ECKLONIAICHTHYS SCYLLIORHINI- 
CEPS 561-2*-3 

Elacate 178 

Elacate bipinnulatus 178 

Elacate nigra 177 

Elagatis 178 

ELEOTRIDAE 264 

ELEOTRIS LIMOSUS 264-5 Pls. 4, 5 


ELEOTRIS MADAGASCARIENSIS 265 

ELEOTRIS OPHIOCEPHALUS 265 

ELOPS 79 

ELUCENS, MYCTOPHUM (DIAPHUS) 
152-3* 

EMARGINATUM, SPONDYLIOSOMA 
311, 362 Pls. 22, 29 

engeli, Mugil 97-8 

EPHIPPIDAE 245-6 

EPHIPPINAE 248 

EPHIPPUS 245-7, 248*-9 


. EPINEPHELUS 80 


EPINEPHELUS FLAVOCAERULEUS 168 
Pl. 22 
ERYTHRORINCHUS, 
PHUS) 31 
EURONOTUS, MUGIL 81-4, 88, 100-1 *-3, 
110, 123, 131, 133 Pls. 16, 17, 19; p. 232*, 233 
EXOCOETIDAE 41, 159 


(HEMIRHAM- 


faber, Chaetodipterus 187-8 
FALCATUS, DICHISTIUS 138-9, 141, 142-4, 


146 Pls. 14, 16 

FALCIFORMIS, MONODACTYLUS 235 
Pl. 34 

FAR, HEMIRHAMPHUS 22-3, 24-8, 30, 40 
Pls. 10, 12 


FARIO FARIO, SALMO 236, 237 

FARIO IRIDEUS, SALMO 236, 237 

FARIO, SALMO FARIO 236, 237 

fasciatus, Hemirhamphus 24, 27 

FASCIGULA, BLENNIUS 8 PI. 16; p. 295 

FAUREI, LEPIDOTRIGLA 62, 63-5 Pls. 16, 
18, 19 

FEROX, ALEPISAURUS 154 

fidjiensis, (Setarches) 57, 58 

FILAMENTOSUS, ARGYROPS 311, 333- 
335, 525 

filamentosus, Dentex 399 

filamentosus, Pristipomoides 287 

filosus, Dentex 370, 372, 376, 556 

FITZSIMONSI, HALIEUTEA 211-2 PI. 23 

fitzsimonsi, Halieutichthys 211 

FLAVOCAERULEUS, EPINEPHELUS 168 
Pl. 22 

FLORENTII, MYCTOPHUM (NASO- 
LYCHNUS) 18 Pl. 9 

frenatus, Gymnocranius 536, 537 

FUCORUM, CLINUS 11, 297 


GADIDAE 158, 306 

GALAXIAS 236 

GALAXIAS ZEBRATUS 236 PI. 34 

GALAXIIDAE 218, 236 

GARIEPINUS, CLARIAS 225 

GEMPYLIDAE 291 

GEORGII, (HEMIRHAMPHUS) 38 

GEPHYROBERYX DARWINI 162 

GEPHYROGLANIS 225, 226 

GEPHYROGLANIS SCLATERI 226 PI. 33 

GERMANUM, POLYAMBLYODON 311, 
366 

GERMANUS, POLYAMBLYODON 
(POLYAMBLYODON) 527, 528 

GIBBICEPS, CHRYSOBLEPHUS 311, 337, 
340-1, 345-7 Pls. 19, 26 

GIBBOSUM, PACHYMETOPON 526-7 


Volume | ends at page 300 


GIBBOSUS, LACTOPHRYS 209 

GIBBOSUS, POLYAMBLYODON (LEPTO- 
METOPON) 528 531 

gigas, (Myctophum) 152-3 

GILCHRISTELLA 235 

GILCHRISTELLA AESTUARIUS 235 

GILCHRISTI, BARBUS 219, 222 Pl. 29 

gilchristi, Drepanoscorpis 138, 144, 146 

GILCHRISTI, HOPLOSTETHUS 160-2 PI. 
22 

GIRELLIDAE 135-6, 246, 305, 360 

GIURIS, GOBIUS 233, 234, 263 

GLAUCOSOMA PEAOLOPESI 396-7* 

GLAUCUM, PACHYMETOPON 311, 363, 
365—6 

GLOBICEPS, AUSTROSPARUS 311, 320, 
321-2*—3, 327-8 Pls. 18, 23; p. 545, 546-9 

GNATHODENTEX 534-6, 541 

GNATHODENTEX AUROLINEATUS 398 

GOBIESOCIDAE 561 

GOBIIDAE 193, 218, 233, 259, 299, 385, 564 

GOBIUS 194, 233 

GOBIUS ACUTIPENNIS 386 

GOBIUS AENEOFUSCUS 233, 234 

GOBIUS CALLIDUS 233, 234*, 299 

GOBIUS GIURIS 233, 234, 263 

GOBIUS GULOSUS 261-2*-3, 299 

GOBIUS KEIENSIS 385-6* 

GOBIUS VONBONDEI 259-261 * Pls. 3, 5 

goreensis, Chaetodipterus 188, 248, 250-2 

gracile, Peristedion 69-71 

GANGENE, (CORYZICHTHYS) 59, 60 

graminis, (Clinus) 195 

GRAMMICOLEPIDAE 1, 162, 187 

GRAMMONUS 52 

GRAMMONUS OPISTHODON 82 PI. 6 

GRANDE, PACHYMETOPON 311, 363, 
365-6, 526 

GRANDOCULIS, (MONOTAXIS) 542 

GRANUALTUS, COELACANTHUS 437 

GRISEUS, GYMNOCRANIUS 532-3, 536, 
537, 539*-541 Pl. 58 

GRUNNIENS, (BATRICHTHYS) 59 

GUENTHERI, (SETARCHES) 58 

GULOSUS, GOBIUS 261-2*-3, 299 

GUNTHERI, SETARCHES 57-8 Pl. 6 

GURNADUS, (TRIGLA) 73-5 

GYMNOCRANIUS 305, 307-369, 532-536-7, 
541-2 

GYMNOCRANIUS BITORQUATUS 536, 
537 

Gymnocranius frenatus 536-7 

GYMNOCRANIUS GRISEUS 532-3, 536, 
537, 539*-541 Pl. 58 

Gymnocranius microdon 536-1 

Gymnocranius rivulatus 533 

GYMNOCRANIUS ROBINSONI 368, 369 
Pls. 21, 27; p. 532 

Gymnocranius robinsoni 533, 536-7, 541 

GYMNOCRANIUS RUPPELLII 533, 537, 
541 

GYMNOCROTAPHUS 304-5, 311, 357, 358, 
529 

GYMNOCROTAPHUS CURVIDENS 311, 
358* 

GYMNOSTETHOIDES, CARANX 279, 280 


haffara, Chrysophrys 547 


HALIEUTEA 212 

HALIEUTEA FITZSIMONSI 211-2 PI. 23 

HALIEUTICHTHYS 212 

Halieutichthys fitzsimonsi 211 

HAPLOCHILUS KATANGAE 272 

HAPLOCHILUS MYAPOSAE 273 

HAPLOCHROMIS 230 

HAPLOCHROMIS MOFFATI 230 Pl. 32 

HAPLODACTYLUS, (SCORPAENA) 203 

HEMIPLEUROGRAMMA, BARBUS 219, 
223 Pl. 29 . 

HEMIRHAMPHIDAE 40-2 

HEMIRHAMPHUS 21-2, 40-1 

HEMIRHAMPHUS BALINENSIS 22-3, 
33-4 

Hemirhamphus calabaricus schlegeli 39 

HEMIRHAMPHUS DELAGOAE 22-3, 31-4 
Pls. 10, 12 

HEMIRHAMPHUS DUSSUMIERI 23, 30-1, 
34 | 

HEMIRHAMPHUS FAR 22-3, 24-8, 30, 40 
Pls. 10, 12 

Hemirhamphus fasciatus 24, 27 

HEMIRHAMPHUS IMPROVISUS 22, 24, 
34-5*-6, 38 Pl. 11 

HEMIRHAMPHUS KNYSNAENSIS 22, 24, 
35*, 36-8, 40 Pls. 10, 11 

HEMIRHAMPHUS MARGINATUS 22-3, 
27-29 

HEMIRHAMPHUS SCHLEGELI 22, 24, 
35-6, 38, 39, 40 

HETERODON, CLINUS 195 PI. 22; p. 296-7 

HEWITTI, CYPSELURUS 159 

HEWITTI, PRIONOLEPIS 1, 3* 

hewitti, Prionolepis 187 

HIPPOCAMPUS CAPENSIS 149, 150 

hoefleri, Mugil 122 

HOLOCENTRIDAE 275 

HOLOCENTRUM SAMMARA 275 PI. 11 

HOLUBI, BARBUS 216, 219, 222 PI. 31 

HOLUBI, SARGUS 327 

HOPLEGNATHIDAE 169 

HOPLEGNATHUS ROBINSONI 169-170 
Pl. 18 

HOPLOSTETHUS GILCHRISTI 160-2 Pi. 22 

HORRIDA, SYNANCEIA 543-4 

Hyporhamphus 2\-2 


IMPROVISUS, HEMIRHAMPHUS 22, 24, 
34-6, 38 Pl. 11 

INDICA, (ACANTHOCEPOLA) 172 

INDICUS, PLATYCEPHALUS 204-6 PI. 20 

INDICUS, PSENES 53 

INORNATA, BOOPSOIDEA 311, 351-2 Pls. 
22, 28 | 

insidator, Platycephalus 204 

IRIDEUS, SALMO FARIO 236, 237 

ISO NATALENSIS 156-7 

ITOSIBI, NEOTHUNNUS 185* 


JACKSONI, TAENIOIDES 564—5*—6 
JABATUS, COCCOTROPUS 201 PI. 18 


KASOUGAE, DERMATOPSIS 566—7*-8 
KATANGAE, HAPLOCHILUS 272 
KAYSERI, DIPLOCERCIDES 477 


Volume | ends at page 300 


KEIENSIS, GOBIUS 385-6* 

kelaartii, Mugil 97-8 

KLEINII, CHAETODON 284 

KNYSNAENSIS, HEMIRHAMPHUS 22, 24, 
35*, 36-8, 40 Pls. 10, 11 

KNYSNAENSIS, PSAMMOGOBIUS 1934 

KNYSNAENSIS, SERRANUS 167 

KOWIENSIS, SCORPAENA 202-3 

KUDA, (HIPPOCAMPUS) 150 

kuhli, Mobula 569 

KUMU, CHELIDONICHTHYS 68 

K YPHOSIDAE 135-7, 247 305 

KYPHOSUS 136-7 


LABEO 218, 224 

LABEO CAPENSIS 224 PI. 31 

LABEO UMBRATUS 224, 225 PI. 31 

LABIOSUS, MUGIL 129-30 

LABRIDAE 289 

LACTOPHRYS 210 

LACTOPHRYS GIBBOSUS 209 

LACTOPHRYS QUADRICORNIS 209 Pl. 22 

Lactoria 210 

LAGOCEPHALUS, TETRODON 207-8 PI. 
20 

LAJANG, DECAPTERUS 174-5 Pl. 21 

LAMPADENA 17 

LAMPANYCTUS 17 

LANIARIUS, PTEROGYMNUS 311, 335, 
336*—7 Pls. 19, 25; p. 526 

LATICEPS, CHRYSOBLEPHUS 311, 340, 
341 Pls. 19, 26 

LATIMERIA 422-6, 461, 471, 501 

LATIMERIA CHALUMNAE 387-8*-390, 
391-4 Pls. 3-7; pp. 403-S*—7*-8, 409-415 
Pls. 1-3; pp. 425, 426-522 19 figs, 44 pis. 

LATIPENNIS, CLINUS 297 

LATOVITTATUS, MALACANTHUS 395 

LATUS, (ACANTHOPAGRUS) 315 

LAURENTII, CLINUS 196—8*, 297 

laurentii, Petraites 196 

LEPIDOTRIGLA 61, 64 

LEPIDOTRIGLA FAUREI 62, 63-5 Pls. 16, 
18, 19 

LEPIDOTRIGLA MULTISPINOSUS 62, 
66—7 Pls. 17, 20 

LEPIDOTRIGLA NATALENSIS 62, 64, 65 
Pls. 16, 18, 19 

LEPTOCEPHALUS 227-8 

LEPTOMETOPON 528 

(LEPTOMETOPON) CRISTICEPS, POLY- 
AMBLYODON 527*, 528—9*, 530*-2 Pl. 50 

(LEPTOMETOPON) GIBBOSUS, POLY- 
AMBLYODON 828, 531 

LETHRINIDAE 533-535 

licha, Dalatias 255 

LICHA, SCYMNORHINUS 255 

lichia, Scymnorhinus 255 

LIMBATA, ACANTHOCEPOLA 172 

LIMOSUS, ELEOTRIS 264-5 Pls. 4, 5 

LINEATA, (TRIGLA) 73-6 

LINEOLATUS, SCOMBEROMORUS 188-9 

LINEOLATUS, (SYNCHIROPUS) 167 

lineopunctatus, Dentex 374-6 

liogaster, (Halieutea) 211-2 

LIONURUS NASUTUS 154-5* 

LIRUS 56 

LITHOGNATHUS 311, 351-2, 354 


LITHOGNATHUS LITHOGNATHUS 311, 
350, 354, 355 

LITHOGNATHUS, LITHOGNATHUS 311, 
350, 354, 355 

LITHOGNATHUS MORMYRUS 311, 355, 
356* Pls. 20, 29 

LITHOGNATHUS, PAGELLUS 182 

LITHOPHILUS, NEOSCORPIS 7 

Liza 79 

longimanus, Mugil 96-8, 113 

LOPHUS, CHRYSOBLEPHUS 311, 340-1, 
346-7, 551 

lophus, Sparus 340 

LOUTI, VARIOLA 542 

LUTIANIDAE 173, 285, 396, 532-535 


MACROLEPIS, MUGIL 89, 118-9*, 120, 131 
Pl. 20 

macrolepis, Mugil 115, 118 

MACROPOMA 389, 425 

MACROSOMA, (DECAPTERUS) 175 

MADAGASCARIENSIS, ELEOTRIS 265 

madagascariensis, Sparus 315 

MALACANTHIDAE 395 

MALACANTHUS LATOVITTATUS 395 

MARCGRAVIA 59 

MARCOSTOMUS, OPISTHOGNATHUS 
(= MACROSTOMUS, O) 164 PI. 20 

MARGINATA, (SYNAPTURA) 6 

MARGINATUS, HEMIRHAMPHUS 22-3, 
27-29 

marginatus, (Hemirhamphus) 24 

MARGINATUS, OPHICHTHYS 151 

marginatus, Ophiurus 151 

MARLEYI, AMBLYAPISTUS 12 

marleyi, Amblyapistus 543 

MARLEYI, CHAETODON 176-7 PI. 22 

MARLEYI, TAENIOLABRUS 256-8 Pls. 1, 2 

MARLEYI, THYRSITOIDES 291-2 

MATSUBARE, DENTEX 367 

MEDITERRANEUS, (HOPLOSTETHUS) 
162 

MEDUSOPHAGUS, SCHEDOPHILUS 55-7 
PI. 5 

MELANURA, SARDINELLA 150-1 

melinopterus, Mugil 125, 128 

MENTALIS, CLINUS 197, 297 

MERLUCCIUS CAPENSIS 158, 206 

MESOPRION, (NEMIPTERUS) 401-2 

MICRODON, DERMATOPSIS 568 

microdon, Gymnocranius 536-7 

MICRODON, PRISTIPOMOIDES 287 

microlepis, Spondyliosoma 362 

MICROPHIS BRACHYURUS 49*, 50 

MICROPTERUS SALMOIDES 237 

Micropus zeylonicus 13 

MICRURA, (PTEROPLATEA) 43-4 

miles, Dentex 370, 373 

MOBULA DIABOLUS 561, 569 

Mobula kuhli 569 

MOBULIDAE 568 

MOFFATI, HAPLOCHROMIS 230 Pl. 32 

MONACANTHIDAE 206, 238-240 

MONACANTHINAE 238 

MONACANTHUS SETIFER 206-7 Pl. 19 

MONACANTHUS, SYNCHIROPUS 165 

MONOCEROS, ALUTERES 238, 240, 241-4 
Pls. 40, 41, 42 


Volume | ends at page 300 


MONOCEROS UNICORNIS 51, 187 

MONODACTYLIDAE 218, 235 

MONODACTYLUS 235 

MONODACTYLUS FALCIFORMIS 235 
Pi. 34 

MONOGRAMMA, SCOLOPSIS 51 

MONOTAXIS 534-5, 542 

MORMYRUS, LITHOGNATHUS 311, 355, 
356* Pls. 20, 29 

MOSSAMBICA, ANGUILLA 227 

MUGIL 44, 77-83, 85*, 87*, (130), 231 


Mugil auratus 80-1, 106, 108, 120, 122-3 

Mugil belanak 129 

Mugil borneensis 120 

MUGIL BUCHANANI 81-2, 89, 113-4*-S, 
131 Pls. 16, 20 

MUGIL CAERULEOMACULATUS | 128, 
130-1 


Mugil caeruleomaculatus 95, 113, 115 
MUGIL CANALICULATUS 81-2, 89, 113, 
120-2*-3, 130, 132 Pls. 16, 17, 18 


Mugil capensis 108, 110 

MUGIL CAPITO 79, 83, 88, 102, 103—4*-5*-7, 
110, 131 Pls. 17, 19 

MUGIL CEPHALUS 79, 81, 83, 88, 90-1 *-S, 
123, 130, 132 Pl. 15; p. 232* 

Mugil ceylonensis 81, 113, 115 

MUGIL COMPRESSUS 81, 89, 115—7*-8, 131 
Pls. 17, 20 

Mugil constantiae 90 

MUGIL CRENILABIS 88, 99*, 100, 130, 
274-5 

MUGIL CUNNESIUS 129, 130 

Muzgil cunnesius 81, 96, 98 

Mugil diadema 81, 115, 117-8 

Mugil engeli97-8 | 

MUGIL EURONOTUS 81-4, 88, 100, 101 *-—3, 
110, 123, 131, 133 Pls. 16, 17, 19; p. 232*, 233 

Mugil hoefleri 122 

Mugil kelaartii 97-8 

MUGIL LABIOSUS 129, 130 

Mugil longimanus 96-8, 113 . 

MUGIL MACROLEPIS 89, 118-9*, 120, 131 
Pl. 20 

Mugil macrolepis 115, 118 

Mugil melinopterus 125, 128 

Mugil multilineatus 108, 110 

Mugil nepalensis 128 

Mugil oeur 90, 92 

MUGIL OLIGOLEPIS 81, 89, 118, 125—7*-8, 
131 Pls. 21, 22 

Mugil olivaceus 120 

Mugil parsia 129 

Mugil planiceps 129 

MUGIL ROBUSTUS 88-90, 93-4*-5, 115, 
130-1 Pls. 21, 22 

Mugil ruppellii 100 

Mugil saliens 81, 100, 102, 106-8, 110 

MUGIL SEHELI 89, 95, 111-2*-3, 123, 131 
Pls. 16, 18 

MUGIL SPEIGLERI 128-9-130 

Mugil speigleri 81, 120, 122 

MUGIL S TRONGYLOCEPHALUS 81, 83, 
88, 93, 96—7*-8, 113, 130 Pls. 16, 18 

MUGIL TADE 129, 130 

MUGIL TRICUSPIDENS 81, 83-4, 88, 93, 
108—9*—10, 123, 131, 133 Pls. 17, 18 

Mugil troscheli 120 


MUGIL WAIGIENSIS 89, 123-4*-S, 131 
Pi. 20 

MUGILIDAE 41, 77, 218, 231, 274 

MUGILIDAE, Abbreviated key to 130 

MULLOIDES, NEMIPTERUS 399, 400*-2 

mulloides, Nemipterus 542 

MULTIFASCIATUS, DICHISTIUS 138-9, 
141, 144-6 Pls. 15, 16 

multilineatus, Mugil 108, 110 

MULTIRADIATUS, DERMATOPSIS 568 

MULTISPINOSUS, LEPIDOTRIGLA 62, 
66—7 Pls. 17, 20 

MUS, CLINUS 11, 297 

MYAPOSAE, HAPLOCHILUS 273 

MYCTOPHIDAE 17, 152 

MYCTOPHUM 17 

MYCTOPHUM (DIAPHUS) ELUCENS 
152-3* 

MYCTOPHUM (NASOLYCHNUS) FLO- 
RENTII 18 Pl. 9 

MYLIOBATIDAE 147 

MYLIOBATIS CERVUS 147-8* 

Myxus 77-9, 81 

Myxus barnardi 81, 90, 93 


NASOLYCHNUS 18 

(NASOLYCHNUS) FLORENTI, MYCTO- 
PHUM 18 PI. 9 

NASUTUS, CYMATOCEPS 311-2, 338-9*, 
382 Pls. 20, 25; p. 512 

NASUTUS, LIONURUS 154-5* 

NASUTUS, PAGRUS 182-3 PI. 23 

natalense, Acanthidium 272 

natalensis, Chrysophrys 323, 547 

NATALENSIS, CUBICEPS 200 

natalensis, Dentex 374 

NATALENSIS ISO 156-7 

NATALENSIS, LEPIDOTRIGLA 62, 64, 65 
Pls. 16, 18, 19 

NATALENSIS, PAGELLUS 311, 352-3*4 
Pis. 20, 29 

natalensis, Polysteganus 376 

NATALENSIS, PTEROPLATEA 43-45*, 47 
Pl. 4 

NATALENSIS, TILAPIA 229 Pl. 32 

NAVALIS, NEMACOCLINUS 297-8 

NEMACOCLINUS 298-9 

NEMACOCLINUS NEVALIS 297-8 

NEMIPTERIDAE 399, 533-535, 542 

NEMIPTERUS 399 

NEMIPTERUS DELAGOAE 542 

NEMIPTERUS MULLOIDES 399, 400*-2, 
542 

NEOSCORPIS 6, 136, 137 

NEOSCORPIS LITHOPHILUS 7 

NEOTHUNNUS 187 

NEOTHUNNUS ITOSIBI 185* 

nepalensis, Mugil 128 

NESIDES 421 

NIERSTRASZI, (PERISTEDION) 71 

nigra, Elacate 177 

nigripinnatus, Chaetodon 282, 284 

NIGROMARGINATUS, (OPISTHOGNA- 
THUS) 165 

NOCT, (DIPLODUS) 331 

NUFAR, CHEIMERIUS 368, 370—1*-3 Pls. 
21, 27 

nufar, Dentex 370-1, 373 


Volume | ends at page 300 


nufar, Polysteganus 371 


obliteratus, (Aluteres) 241 

oeur, Mugil 90, 92 

OLIGOLEPIS, MUGIL 81, 89, 118, 125—7*-8, 
131 Pls. 21, 22 

olivaceus, Mugil 120 

ONCHOCEPHALIDAE 211 

OPERCULARE, POMADASYS 179-182 

OPERCULARIS, PTERAGOGUS 289, 290* 

OPHICHTHYIDAE 151, 273 

OPHICHTHYS MARGINATUS 151 

OPHIOCEPHALUS, ELEOTRIS 265 

Ophiurus marginatus 151 

OPISTHODON, GRAMMONUS 82 PI. 4 

OPISTHOGNATHIDAE 164 

OPISTHOGNATHUS MARCOSTOMUS 
(= MACROSTOMUS) 164 PI. 20 

ORBICULARIS, (PLATAX) 249 

orbis, Chaetodipterus 188, 250 

ORBIS, TRIPTERODON 187-8, 245, 248- 
250*—2 Pls. 21-23; p. 542-3 

ORNATUS, CLINUS 11, 197, 297 

Osbeckia 240 

Osbeckia scriptus 243 

OSTRACION 210 

OSTRACIONTIDAE 209 

Ostracion turritus 209 

OXYLOPHIUS, TROPIDICHTHYS 15 Pl. 16 


PACHYMETOPON 305, 311, 357, 362, 364, 
526 

PACHYMETOPON AENEUM 311, 363, 364 
Pls. 22, 29; p. 526 

PACHYMETOPON BLOCHII 311, 363, 364 
Pls. 22, 28, 29 

PACHYMETOPON CANESCENS 311, 363-4 

PACHYMETOPON GIBBOSUM 526-7 

PACHYMETOPON GLAUCUM 311, 363, 
365, 366 

PACHYMETOPON GRANDE 311, 363, 
365-6, 526 

Pachynathus conspicillum 210 

PAGELLINAE 310-11, 350 

PAGELLUS 304-5, 311, 351, 352, 354 

PAGELLUS LITHOGNATHUS 182 

PAGELLUS NATALENSIS 311, 352-3*-4 
Pls. 20, 29 

PAGRUS 183, 304—5, 317, 319, 349 

PAGRUS CILIARIS 334 

PAGRUS NASUTUS 182-3 Pl. 23 

PALMIETENSIS, BROTULA 198 PI. 21 

PALUDINOSUS, BARBUS 219, 221 Pl. 30 

PAPYRICHTHYS 49, 53-4 
PAPYRICHTHYS PELLUCIDUS 54-5 PI. 

6A 

PARAPERCIS PULCHELLA 276-7*-9 

Parapercis robinsoni 279 

PARASCORPIS 136 

parmatus, (Setarches) 57-8 

Parsia, Mugil 129 

PAVO, CLINUS 11, 297 

PEAOLOPESI, GLAUCOSOMA 396-7* 

Pelecinomimus 159 

Pelecinomimus picklei 158 

PELLUCIDUS, PAPYRICHTHYS 54-5 P}. 6 

pellucidus, Psenes 53-4 


PENTAPODIDAE 534, 535, 342 

PENTAPODUS 534-4, 541 

PERISTEDION 61, 69 

PERISTEDION ADENI 69, 71-3 Pl. 22 

Peristedion gracile 69-7) 

PERISTEDION WEBERI 69-71, 73 Pl. 21 

PERONII, DENTEX 367 

PETRAITES 299 

Petraites 197 

Petraites laurentii 196 

PETROSCIRTES TAPEINOSOMA 51, 191-2 
Pl. 21; p. 296 

PETRUS 366-8, 380-1 

PETRUS RUPESTRIS 368, 381, 382* Pls. 21, 
27 

PICKLEI, CETOMIMUS 158-9 

picklei, Pelecinomimus 158 

PINGUIPEDIDAE 276 

PINNATUS, (PLATAX) 249 

PISOODONOPHIS BORO 273 

planiceps, Mugil 129 

PLATACIDAE 187, 247, 248, 542 

PLATACINAE 247, 248 

PLATAX 137, 245-249, 252 

PLATYCEPHALIDAE 204 

PLATYCEPHALUS INDICUS 204-6 PI. 20 

Platycephalus insidator 204 

PLECTORHYNCHIDAE SO, 179, 287 

PLOTOSIDAE 218, 226 

PLOTOSUS ANGUILLARIS 226 PI. 34 

POLYAMBLYODON 305, 311, 357, 362, 366, 
526-8 

POLYAMBLYODON GERMANUM #311, 
366 

(POLYAMBLYODON) GERMANUS, 
POLYAMBLYODON 527, 528 

POLYAMBLYODON (LEPTOMETOPON) 
CRISTICEPS 527*, 528, 529*, 530*—2 Pl. 50 

POLYAMBLYODON (LEPTOMETOPON) 
GIBBOSUS 528, 531 

POLYAMBLYODON (POLYAMBLYO- 
DON) GERMANUS 527, 528 

POLYPTERUS 501 

POLYSTEGANUS 367-8, 371, 373-4, 378-380 

POLYSTEGANUS ARGYROZONA 308%, 
368, 374, 378 

POLYSTEGANUS COERULEOPUNCTA- 
TUS 368, 373, 374-6 Pls. 22, 28 

Polysteganus natalensis 376 

Polysteganus nufar 37} 

POLYSTEGANUS PRAEORBITALIS 368, 
374, 379, 380* Pls. 21, 28 

POLYSTEGANUS UNDULOSUS 368, 374, 
376-7* Pls. 21, 28; p. 556 

POMADASYS OPERCULARE 179-182 

Pomadasys striatus 288-9 

POMADASYS STRIDENS 287-9 

POMADASYS SUILLUM 181 

POMATOMUS SALTATOR 186 

PORCOSTOMA 311, 314, 348 

PORCOSTOMA DENTATA 311, 349, 350 
Pls. 20, 25 

PRAEORBITALIS, POLYSTEGANUS 368, 
374, 379, 380* Pls. 21, 28 

PRIONOLEPIS 1 

PRIONOLEPIS HEWITTI I, 3* 

Prionolepis hewitti 187 

PRIONOSPARUS 546, 548 

PRISTIOPHORUS 253 


Volume | ends at page 300 


PRISTIPOMOIDES 287 

PRISTIPOMOIDES 
MICUS 285-6*-7 

Pristipomoides filamentosus 287 

PRISTIPOMOIDES MICRODON 287 

PSAMMOGOBIUS 193 

PSAMMOGOBIUS KNYSNAENSIS 193-4 

PSENES 53, 200 

PSENES INDICUS 53 

Psenes pellucidus 53-4 

PSENES WHITELEGGII 199, 200 Pi. 19 

PSEUDALUTERES 240 

PTERAGOGUS OPERCULARIS 289, 290*-—1 

PTEROGYMNUS 310-11, 314, 335 

PTEROGYMNUS LANIARIUS 311, 335, 
336*-7 Pls. 19, 25; p. 526 

PTEROPLATEA 43 

PTEROPLATEA NATALENSIS 43-5*-7 PI. 
4 

PULCHELLA, PARAPERCIS 276—7*-9 

PUNICEUS, CHRYSOBLEPHUS 311, 340-1, 
343-4*-5, 526, 551-2, $54—-5*-6 

PUNTAZZO 305, 311, 313, 330 

PUNTAZZO PUNTAZZO 311, 330 

PUNTAZZO, PUNTAZZO 311, 330 


ARGYROGRAM- 


QUADRICORNIS, LACTOPHRYS 209 PI. 
22 | 

QUADRISPINOSUM, ACANTHIDIUM 
270-1 *-2 

QUEKETTI, CHELIDONICHTHYS 68* 


regani, (Cetomimus) 158 

REYNALDI, (HEMIRHAMPHUS) 31 

RHABDODERMA 425 

RHABDOSARGUS 321, 546 

RHACHICENTRIDAE 177 

RHACHICENTRUM CANADUS 177-8 

Rhynchorhamphus 21-2 

RIVERS-ANDERSONI, (PERISTEDION) 71 

rivulatus, Déntex 533, 537, 541 

rivulatus, Gymnocranius 533 

ROBINSONI, (ACANTHOPAGRUS) 317 

robinsoni, Dentex 533, 537 

ROBINSONI, GYMNOCRANIUS 368, 369 
Pls. 21, 27 

robinsoni, Gymnocranius 533, 536-7, 541 

ROBINSONI, HOPLEGNATHUS 169-70 PI. 
18 

robinsoni, Parapercis 279 

ROBUSTA, AXELIA 501, 505 

ROBUSTUS, CLINUS 196, 297 

ROBUSTUS, MUGIL 88-90, 93-4*-S, 115, 
130-1 Pls. 21, 22 

RUPPELLII, GYMNOCRANIUS 533, 537, 
541 

ruppellii, Mugil 100 

rupestris, Dentex 371 

RUPESTRIS, PETRUS 368, 381-2* Pls. 21, 27 


Salarias 191 

Salarias sexfasciatus 190-1 © 

saliens, Mugil 81, 100, 102, 106-8, 110 
SALMO 236 

SALMO FARIO FARIO 236, 237 
SALMO FARIO IRIDEUS 236, 237 


SALMOIDES, MICROPTERUS 237 

SALMONIDAE 218, 236 

SALPA, SARPA 311, 360 

SALTATOR, POMATOMUS 186 

SAMMARA, HOLOCENTRUM 275 PI. 11 

SARBA, AUSTROSPARUS 311, 319, 320- 
323—4*, 327-8 Pls. 18, 23; p. 545-547-9 

sarba, Austrosparus 325 

sarba, Chrysophrys 547 

SARBA, SPARUS 44, 141, 547 

SARDA CHILENSIS 293, 294*—5 

SARDA SARDA 295 

SARDA, SARDA 295 

SARDINELLA MELANURA 150-1 . 

SARGUS, DIPLODUS 311, 320, 331-2* Pls. 
19, 24 

SARGUS HOLUBI 327 

SARPA 305, 311, 357, 359 

SARPA SALPA 311, 360 

SASSENIA 425 

SCANDENS, ANABAS 230 

Scarostoma 170 

SCATHARINAE 310-1, 357 

SCHEDOPHILUS 54, 56 

SCHEDOPHILUS MEDUSOPHAGUS 55-7 
Pl. 5 

SCHLEGELI, HEMIRHAMPHUS 22, 24, 
35-39, 40 

schlegeli, Hemirhamphus calabaricus 39 ; 

SCLATERI, GEPHYROGLANIS 226 Pl. 33 

SCOLOPSIS MONOGRAMMA 51 

SCOLOPSIS VOSMERI 50 

SCOMBEROMORUS LINEOLATUS 188-9 

SCOMBRIDAE 185, 187-8, 293 

SCORPAENA KOWIENSIS 202-3 

SCORPAENIDAE 12, 57, 201-2, 543 

SCORPIDIDAE 6, 135-7, 247 

SCORPIS 6 

SCRIPTUS, ALUTERES 240-1, 243-4 Pl. 42 

scriptus, Osbeckia 243 

SCULLYI, (BLENNIUS) 189 

SCYLLIORHINICEPS, ECKLONIA- 
ICHTHYS 561-2*-3 

SCYMNORHINUS 253 

SCYMNORHINUS BREVIPINNIS 253—-4*-5 

SCYMNORHINUS LICHA 255 

Scymnorhinus lichia 255 

SCYMNORHINIDAE 253 

SCYMNUS 253 

SEHELI, MUGIL 89, 95, 111-2*-3, 123, 131 
Pls. 16, 18 

Semathunnus 187 

SENTICEPS, BARBUS 219, 220*, 266* 

SERRANIDAE 167, 542 ' 


SERRANUS KNYSNAENSIS 167 


SETARCHES GUNTHERI or GUENTHERI 
57-8 Pl. 6 

SETIFER, CHAETODON 247 

SETIFER, MONACANTHUS 206-7 Pl. 19 

sexfasciatus, Salarias 190-1 

SINUOSA, WIMANIA 500 

SMITHII, (MICROPHIS) 50 

smithii, (Mugil) 118 

SOLEIDAE 4 

SPARIDAE 135-7, 182, 303-308*-10, 337, 
525-6, 532-534, 536, 545 

SPARIFORMES 534 

SPARINAE 3410-312-3 

SPARODON 311, 313, 315, 327-8 


Volume | ends at page 300 


SPARODON DURBANENSIS 311, 321, 
328-9*, 330 Pls. 18, 23 

SPARRMANI, TILAPIA 229 PI. 32 

SPARUS 183, 304-5, 313-321, 328-331, 334-5, 
340, 349, 545 

Sparus durbanensis 328 

Sparus lophus 340 

Sparus madagascariensis 315 

SPARUS SARBA 44, 141 

Sparus sarba 547 

SPEIGLERI, MUGIL 128, 129-30 

speigleri, Mugil 81, 120, 122 

SPINIFER, ARGYROPS 311, 333, 334-5 Pls. 
20, 25; p. 525 

SPONDYLIOSOMA 305, 311, 357, 362-3 

SPONDYLIOSOMA EMARGINATUM 311, 
362 Pls. 22, 29 

Spondyliosoma microlepis 362 

SQUALIDAE 253, 270 

STELLATUS, (TETRODON) 208 

STENOGOBIUS 263 

STRIATUS, CLINUS 11, 197, 297 

striatus, Pomadasys 288-9 

STRIDENS, POMADASYS 287-9 

STROMATEIDAE 49, 53, 199 

STRONGYLOCEPHALUS, MUGIL 81, 83, 
88, 93, 96-7*-8, 113, 130 Pls. 16, 18 

SUILLUM, POMADASYS 181 

SUPERCILIOSUS, CLINUS 11, 197, 297 

SYNAGRIS 399 

SYNANCEIA 543 

SYNANCEIA HORRIDA 543-4 

SYNANCEIA VERRUCOSA 543-4 

SYNANCIIDAE 13, 543 

SYNAPTURA BARNARD] 4, 5* Pl. 16 

SYNCHIROPUS MONACANTHUS 165 

SYNGNATHIDAE 49, 149 


TADE, MUGIL 129-30 

TAENIOIDES 564 

TAENIOIDES JACKSONI 564—5*—6 

TAENIOLABRUS 256 

TAENIOLABRUS MARLEYI 256*-8 Pls. 
1,2 

TAENIONOTUS, (AMBLYAPISTUS) 13 

Taius 367, 373 

TAPEINOSOMA, PETROSCIRTES 51, 191-2 
Pl. 21; p. 296 

TAURUS, CLINUS 196, 297 

teira, (Platax) 249 

TELEOSTEI 215-217* 

TESTUDINEUS, ANABAS 230 

TETRODON LAGOCEPHALUS 207-8 Pl]. 22 

TETRODONTIDAE 15, 207 

TEUTHIDIDAE 187 

Tholichthys 284 

THYRSITOIDES MARLEYI 291-2 

TILAPIA 228-9, 230 

TILAPIA NATALENSIS 229 Pl. 32 

TILAPIA SPARRMANI 229 Pj. 32 

TOBYEI, (MYLIOBATIS) 147 

TORVUS, CONGIOPODUS 51 

TRACHELOCHISMUS 561 

TRACHICHTHYIDAE 160 

TREVELYANI, BARBUS 219 Pl. 29 

TRICHONOTIDAE 256 

TRICUSPIDENS, AUSTROSPARUS 546, 
548-9 


TRICUSPIDENS, MUGIL 81, 83-4, 88, 93, 
108-9*-10, 123, 131, 133 Pls. 17, 18 
TRIFASCIATUS, CHAETODON 280-1 


TRIFASCIATUS, DIPLODUS 311, 331, 
332-3* Pls. 19, 24 

trifascigula, Blennius 295 

TRIGLA 61, 73-4 

TRIGLA (TRIGLA) 73 

(TRIGLA), TRIGLA 73 

TRIGLA (TRIGLOPORUS) 61, 73-4 

TRIGLA (TRIGLOPORUS) AFRICANA 75, 
74-6 Pl. 23 

TRIGLIDAE 61 

TRIGLOPORUS 61, 734 

(TRIGLOPORUS) AFRICANA, TRIGLA 
73, 74-6 Pl. 23 

(TRIGLOPORUS), TRIGLA 61, 734 

TRIPTERODON 137, 245-249, 304 

TRIPTERODON ORBIS 187-8, 245, 248- 
250*-2 Pls. 21-23; pp. 542-3 

TROPIDICHTHYS OXYLOPHIUS 15 Pl. 16 

troscheli, Mugil 120 

turritus, Ostracion 209 


UMBRATUS, LABEO, 224, 225 Pl. 31 

UNDULOSUS, POLYSTEGANUS 368, 374, 
376—7* Pls. 21, 28; p. 556 

UNICORNIS, MONOCEROS 51, 187 

unifasciatus, (Hemirhamphus) 33, 36, 38, 40 


VAGUS, (ACANTHOPAGRUS) 314 
VAILLANTII, (PTEROPLATEA) 46 
VALENCIENNI, (PTEROPLATEA) 46 
variabilis, Dentex 37} 

VARIOLA LOUTI 542 

VENUSTRIS, CLINUS 196, 297 
VERRUCOSA, SYNANCEIA 543-4 
vespertilio, (Platax) 249 

VESPOSUS 1 

vicinus, Anabas 231 

VONBONDEI, GOBIUS 259-261 * Pls. 3, 5 
VOSMERI, SCOLOPSIS 50 PI. 5 
VULPES, ALBULA 51 

VULNERATUS, BARBUS 219, 223 Pl. 31 


WAIGIENSIS, MUGIL 89, 123-4*-5, 131 PI. 
20 

WEBERI, PERISTEDION 69-71, 73 Pl. 21 

WHITEIA 425 

WHITELEGGII, PSENES 199, 200 Pl. 19 

WIMANIA 425, 447 

WIMANIA SINUOSA 500 


XANTHOCEPHALUS, 
282-3 * 

XENOLEPIDICHTHYS 1 

XENOLEPIDICHTHYS DALGLEISHI 
162-3 Pl. 18 


CHAETODON 


ZEBRATUS, GALAXIAS 236 Pl. 34 
zeylonicus, Amphiprionichthys 13 
ZEYLONICUS, CARACANTHUS 13 
zeylonicus, Micropus 13 


Agulhas Cape, 13 
Algoa Bay, 20 
Amanzimtoti, 37 
Bashee, 29 
Bazaruto Island, 51 
Beira, 52 
Bredasdorp Coast, 14 
Buffalo River, 26 
Bushmans River, 22 
Camps Bay, 8 

The Cape, 9 

Cape Agulhas, 13 
Cape of Good Hope, 9 
Cape Padrone, 21 
Cape Point, 9 

Cape Town, 7 
Chalumna River, 25 
Dassen Island, 6 
Delagoa Bay, 49 
Durban, 39 

East London, 26 
False Bay, 12 

Fish Point, 24 
Great Fish Point, 24 
Great Kei River, 27 
Inhaca, 47 
Inhambane, 50 
Isipingo, 38 

Kalk Bay, 10 

Kei Mouth, 27 

Kei River, 27 
Knysna, 17 

Kowie River, 23 
Kosi Bay, 45 


— 


OONOahWHh 


Swakopmund 
Walfish Bay 
Port Nolloth 
Lamberts Bay 
St. Helena Bay 
Saldanha Bay 
Dassen Island 
Cape Town 
Camps Bay 

The Cape 

Cape of Good Hope 
Cape Point 

Kalk Bay 
Simons Bay 
False Bay 

Cape Agulhas 
Bredasdorp Coast 
St. Sebastian Bay 
Mossel Bay 
Knysna 
Plettenberg Bay 
Tsitsikama 
Algoa Bay 

Port Elizabeth 
Cape Padrone 
Bushmans River 
Kowie River 
Port Alfred 
Great Fish Point 
Chalumna River 
Buffalo River 
East London 


SOUTHERN AFRICAN 


Kei Mouth 
Great Kei River 
Transkei 
Bashee 

Xora River 
Umtata River 
Umgazi River 
Port St. Johns 
Pondoland 
Port Shepstone 
Umkomaas River 
Amanzimtoti 
Isipingo 
Durban 
Umgeni River 
Umhlanga 
Tugela River 
Zululand 
Richards Bay 
St. Lucia Bay 
Kosi Bay 


Maputoland 49) 
Inhaca cceencee 


L 
Ponte Mahone M 
Ponte Maone 
Delagoa Bay 
Lourenco Marques 
Polana 
Inhambane 
Bazaruto Island 
Beira 
Zambezi River 


ALOR I, ACO R yy, — iglEteaey.. OOO Gy, 


Lamberts Bay, 3 
Lourenco Marques, 49 
Maputoland, 46 
Mossel Bay, 16 
Plettenberg Bay, 18 
Polana, 49 
Pondoland, 34 
Ponte Mahone, 48 
Ponte Maone, 48 
Port Alfred, 23 

Port Elizabeth, 20 
Port Nolloth, 2 

Port Shepstone, 35 
Port St. Johns, 33 
Richards Bay, 43 
St. Helena Bay, 4 
St. Lucia Bay, 44 
St. Sebastian Bay, 15 
Saldanha Bay, 5 
Simons Bay, 11 
Swakopmund, 1 
Transkei, 28 

Tugela River, 41 
Tsitsikama, 19 
Umgazi River, 32 
Umgeni River, 39 
Umhlanga, 40 
Umkomaas River, 36 
Umtata River, 31 
Walfish Bay, 1 

Xora River, 30. 
Zambezi River, 53 
Zululand, 42