FISH OF THE NAGA HILLS, ASSAM.

By SUNDER LAL HORA, D.Se., F.R.S.E., F.A.S.B., and DEV DEV

MUKERJI, M.Sc., Zoological Survey of India, Oalcutta.

(Plate VII)

INTR-ODUCTION.

In 1921, one l of us published an account of the " Fish and Fisheries

of Manipur with some observatio!,s on those of Naga Hills" and recorded
17 species from streams with rocky beds in the northern watershed
(Brahmaputra drainage) and 18 species from similar streams in the
southern watershed (Chindwin drainage) of the Naga Hills. The lists
included the 3 species mentioned by Day in the Fauna of B'ritish India
from the Naga Hills and the two species described by Chaudhuri2 from
this area. Only 3 species, viz., Ba'rbus hexastichus, Ba1'ilius barila and
Danio aequipinnatus, were known to be common to both the water-
sheds; this reduces the total of the species known from the Naga Hills
to 323, It may be noted here that under the name Naga Hills is included
all the country inhabited by the Naga tribes and not merely the district
to which the name is officially applied.
The interest of the fish fauna of the Naga Hills lies in the fact that
it contains representatives of the Assamese and the Burmese elements
in almost equal propo,tions. To study this point in detail one of us
requested Dr. J. H. Hutton, Deputy Commissioner of the Naga Hills,
in 1927 to make a collection of fish during the course of one of his tours
to an unfrequented part of the hills. Dr. Hutton agreed to this sugges-
tion and procured a valuable collection for the Zoological Survey of
India. Early this year Dr. B. Prashad and Dr: B. Chopra visited the
Naga Hills to study certain zoogeographical problems and in the course
of their investigations made an extensive collection of fishes.
The present paper deals mainly with these two collections.
Dr. Hutton made collections at the following places :-
Zhokami
Tekhubami } Brahmaputra drainage (\)ayang VaJJey).
Mekruchu
Mohumi
Purobami
Sahunyu
Melori Chilldwin drainage (Tizu Valley).
Lepori
_Yisisu
Laruri

1 Hora, Rec. Ind. Mus. XXII, pp. 165-214, pIs. ix-xii (1921).
3 Chaudhuri, Rec. Ind. Mus. VII, p. 443, pI. xl, figs. 4, 4a, 4b and pI. xli, fig~. 1,
la., lb; p. 441, pI. xl, figs. I, la, lb (1912).
3 In 1921, Dr. l\'lurray Stuart of the Geological Survey of India made a t:lmaU collec-
tion of :fish, comprising II species (Hora, Ree. Ind. Jlus. XXII, p. 743, 1921) from the
North-eastern border of Burma and the Naga Hills. As the exact Ioca,J~t.ie8 of the specie8
are not known, we have not included these forms in our treatment of the fish fauna
of the Naga Hills. It may be remarked that 4 out of the II species, 'viz., Ga:rra rJotyla,
Barbus ehrysopterus, BariUus t'agra and A mbassis nama., M·P not roprf>sented in ot.her
colleotions from the N aga. Hills.
[ 381 ]
382 Records of the Indian Museum. [VOL. XXXVII,

Drs. Prashad and Chopra made collections at the following places :-.
Kohima
Mao
Chakabama Brahmaputra drainage.
Kekrima.
Emilomi
Khezhabama
Phekrokejima.
Chepoketami Chindwin drainage.
Sakhai
Sakhalu
There are altogether 23 species in the two collections and of these
11, i.e., Amblyeeps mangois, Exostoma labidtum, Barbus tietol, Orosso-
ckilus latius, Garra lissorhynchus, Garra kempi, Oreinus moleswortki,
Psilorkynchus homaloptera, sp. nov., N emaehilus rupecola, N. subfusea,
and Ophicephalus gaehua, are recorded from the Naga Hills for the first
time. Though only one new species has been discovered in the collec-
tions under report, it has been possible to elucidate the precise specific
limits of Barbus hexagonolepis, N ernachilus manipurensis and N. subfusca,
while the geographical distribution of a number of species has been
more precisely determined. For instance, Exostoma labiatum, Barbus
clavatus, B. hexagonolepis, Garra lissorhynchus, G. kempi, G. naganensis
and Oreinus molesworthi, which were hitherto known from the Brahma-
putra drainage system, are now recorded from the Chindwin drainage
system also. Species like Brachydanio aeu~icephala and N emachilus
manipurensis, which were known only from the Chindwin drainage
system, have now been discovered in the Brahmaputra basin also.
44 species are now known from the N aga Hills. These species,
together with their distribution as now known, are listed below :-

Brahma-
Serial Chindwin Widely
Specific Name. putra Drainage
Number. Drainage distri-
System. System. buted.

BAGRIDAE.

1 MystU8 bleeke1'i (Day) ·. ·. X

SILURIDAE.

2 Ompok bimaculatus (Bloch) ·. ·. X

AMBLYOEPIDAE.

3 Amblyceps mangoi8 (Hamilton) ·. ·. X

SISORIDAE.

4 Erethistes hara (Hamilton) ·. ·. X

5 Erethistes elongata Day X ·. I ..

6 Exostoma labiatum (McClell.)
I X X ..
1 Barbu8 ticto was obtained by one of us from the sluggish waters of 1J:le Manipur
Valley, but it is recorded here from the Naga Hills. .
1935.] s. L. HORA & D. D. MUKERJI: Fish of the Naga Hills. 383

Brahma-
Serial putra Chindwin Widely
Number. Specifio Name. Drainage distri-
Drainage
System. System. buted.

CYPRINIDAE.
7 Danio (Danio) dangila (Hamilton)
·. ·. X
8 Danio (Danio) aequipinnaius (MoClen.) ·. ·. X
9 Danio (Danio) naganensis Chaudhuri ·. X ·.
10 Danio (Brachydanio) acutic4phala Hora X X ..
11 Rasbora rasbora (Hamilton) . ·. ·. X
12 Bariliu8 barila (Hamilton) ·. ·. X
13 Bariliu8 dogarsinghi Hora ·. X ·.
14 Barilius bendelesis var. chedra (Hamil-
ton).
xt ·. ·.
15 Barbus clavatus McClell. X X ·.
16 Barbus tor (Hamilton) ·. ·. X
17 Barbus hexagonolepis McClen. X X ·.
18 Barbus conchonius (Hamilton) ·. ·. X
19 Barbu8 oatesii Boulenger . ·. X ·.
20 ·.
Barbus ticto (Hamilton)
·. X
21 Oros8ochiluslatius (Hamilton) ·. ·. - X

22 Garra rupeculus (McClen.) X X ·.

23 Garra abhoyai Hora . ·. X ·.
24 Garra naganen8is Hora X X ·.
25 Garra kempi Hora X X ·.
26 Garra lissorhynchus (McClell.) X X ·.
27 Oreinus molesworthi Chaudhuri X X ·.
PSILORHYNOHIDAE.
28 P8-ilorhynchus sp. Hora X ·. ·.
29 P8ilorhynchus homaloptera, sp. nov. X ·. ...
COBITIDAE.
30 Lepidocephalichthys berdmorei (Blyth) ·. x* ·.
31 LepidocephaUchthys guntea (Hamilton) xt ·. ·.
32 AcanthophthalmU8 pangia (Hamilton) X x* ·.
33 N emachilu8 manipurensis Chaudhuri X X ·.
-
* denotes that the species is found in Burma also. .
t denotes that the species is found in other parts of IndIa also.
R2
384 Records of the Inii'1!n Mu~eum. [VOL. XXXVII

Brahma- Chindwin Widely

Serial putra distri-
Specific Name. Drainage Drainage
Number. System. bated.
Sysuem.

COBITIDAE-contd.

34 N emackilu8 zonalternans (Blyth) ·. x* ·.

35 N emachilu8 8ik maiensi8 Hora ·. X ·.
36 N emackilu8 kan,gjupkkulen8i8 Hora ·. x* ·.
37 N emackilu8 prashadi Hora. ·. X ·.
38 Nemackilu8 botia (Hamilton) ·. ·. X

39 N emackilu8 rupecola (McClell.) xt ·. ·.

40 N emachilu8 subJu8ca (MoClell.) X ·. ·.
MASTAOEMBELIDAE.

41 Rhynchobdella dkanasorii Hora X ·. ·.

NANDIDAE.

42 Badi8 badi8 (Hamilton) ·. ·. X

OpmOEPHALIDAE.

43 Ophicephalu8 gachua Hamilton ·. ·. X

44 OphieephaZu8 punetatu8 Blooh. ·. ·. X

* denotes that the speoies is found in Burma also.

t denotes that the speoies is fou~d in other parts of India also.

Of the 44 species listed above, 16 are widely distributed in India

and Burma; 11 are common to the Brahmaputra and the Chindwin
drainage basins, though Acanthopkthalmus pangia is found in. Burma
also; 8 species are found only in the Brahmaputra drainage of the
Naga Hills, but of these N emachilus rupecola, Lepidocephalichthys guntea
and Barilius bendelesis are also found in other parts of India. Similarly,
of the 9 species found only in the Chindwin drainage of the Naga Jlills,
N emachilus zonalternans, N. kangjupkhulensis, Lepidocephalichtltys
berdmorei and Barbus oatesii are found in Burmese waters also.
Another point of zoogeographical interest is the similarity between
the fish fauna of the Garo Hills, the Mishmi Hills, the Khasi Hills and
the Naga Hills inter se and of these hills with that of the Eastern Hima-
layas. This is particularly true of the forms adapted to live in torrential
streams, such as Exostorna labiatum, Erethistes elongata, Garra kempi,
G. lissorhynchus, Oreinus molesworthi, N elnachilus rupecola, N. subfusca,
etc. From a study of the Batrachian fauna of the Abor Hills, Annan-
dalel concluded that" All the evidence at present available, therefore,
supports the view that the fauna of the extreme ea3 tern part of the
1 Annandale} Ree. Ind. M'l~s~ VIII, p. 36 (1912).
1935.} S. L. HORA & D. D. MUKERJI: Fish of the Naga Hills. 385

Himalayan foot-hills is not, at any rate 80 far as the frogs and toads
are concerned, Himalayan in the sense in which'the term has hitherto
been understood, but allied rather to the fauna of Assam south of the
Brahmaputra or even to that of Burma." Our studies on the distri-
bution of fishes from these regions also lead us to similar conclusions.
The above list of species also shows that the fish fauna of the Naga
Hills contains a great preponderance of the Ostariophysi, 40 out of a
total of 44 species. Of these 40~8pecies, 6 belong to the Siluroidea (1
Bagridae, 1 Siluridae, 1 Amblycepidae and 3 Sisoridae) and the remain-
ing to the Cyprinoidea (21 Cyprinidae, 2 Psilorhynchidae and 11 Cobi-
tidae). The remaining 4 species are distributed among the families
MastacembeIidae, N andidae and Ophicephalidae.
The 44 species known from the Naga Hills belong to 19 genera, all
of which, with the possible exception of Psilorhynchus: are fairly widely
distributed in India and Burma. It is thus seen that the Naga Hills-
the watershed of the principal Burmese (Irrawaddy) and Assamese
(Brahmaputra) drainage systems-have no endemic element so far as
the genera of fishes are concerned. Among the species, however, there
are 5, viz., Danio naganensis, Barilius dogarsinghi, Garra abhoyai, N ema-
chilus sikmaiensis and N. prashadi, which are so far konwn only from
the Chindwin drainage of the Naga Hills and 3, Psilorhynchus sp., P.
hO'nlaloptera and Rhynchobdella dhanasorii, from the Brahmaputra
drainage. Danio (Brachydanio) acuticephala, Garra nanagensis and
N emachilus manipurensis are also endemic in the N aga Hills, though
they are found equally abundantly in both the watersheds. Thus
25 per cent. of the fish are ~ndemic in the N aga Hills.
We have to thank Dr. J. H. Hutton for the very careful way in
which his collection was preserved and labelled. To Drs. B. Prashad
and B. Chopra we are grateful for devoting a considerable part of their,
time in the Naga Hills to the collection of fishes. Babu R. C. Bagchi
has executed the drawmgs with his usual skill and care, and for this we
are thankful to him. '
SYSTEMATIC DESCRIPTION.

BAGRIDAE.
Mystus bleekeri (Day).
1921. Macrones bleekeri, Hora, Bee. Ind. Mus. XXII, p. 179.
1 specimen. Melori, Tizu River. J. H. Hutton, March 1927.
Mystus bleekeri.is represented in Dr. Hutton's collection by a young
specimen about 90 mm. in total length.

SILURIDAE.
Ompok bimaculatus (Bloch).
1921. Oallichrous bimaculatus, Hora, Ree. Ind. MU8. XXII, p. 178.
1 specimen. Melori, Tizu River. J. H. Huttom, March 1927.
The only specimen of Ompok bimaculatus in Dr. Hutton's collection
is about 142 mm. in total length.
 386 Recoras of tke Indian Museum. [ VOL. XXXVII,

AMBLYCEPIDAE.
Amblyceps mangois (Hamilton).
1933 • .Amblyeeps mangois, Hora, Bee. Ind. Mus. XXXV, p. 617.
1 specimen. Sakhai, Lizho stream. B. Prashad & B. Chopra, 23rd
February 1935.
17 specimens. Emilomi, Keleki stream. B. Prashad & B. Chopra, 26th
February 1935.
One of us (Hora, Ope cit., pp. 607-621) has recently discussed the
taxonomy and bionomics of the loach-like fishes of the genus Amblyceps.
In most of the specimens under report the caudal fin is almost truncate

'. . ,
'. I
? I
\

,,
I

---------- I

TEXT-FIG. I.-Outline sketch of caudal fin of a specimen of .Amblyeeps mangois (Hamil-

ton) from Lizho stream near Sakhai showing prolonged outer rays.
X 3.

or slightly emarginate. In the example from the Lizho stream, how-

ever, the end rays of the caudal fin are produced into thread-like processes
although the fin- is not deeply emarginate as is usually the case with the
.Siamese examples.
The largest specimen is about 120 mm. in total length.

SISORIDAE.
Exostoma1 Iabiatum (McClelland).
1923. Glyptosternum labiatum, Hora, Bee. Ind. Mus. XXV, p. 42, pI. i, fig. 2.
4 specimens. Purobami and Sahunyu, sources of the Tizu River. J. H.
Hutton, March 1927. .
4 specimens. Sakhai, Lizho stream. B. Prashad & B. Chopra, 23rd
February 1935.
2 specimens. Sakhai, Tizu River. B. Prashad & B. Chopra, 23rd Feb-
ruary 1935.
15 specimens. Sakhalu, Tizu River. B. Prashad & B. Chopra, 23rd Feb-
ruary 1935.
Exostoma labia tum has hitherto been known 'from the Mishmi and
the Abor Hills in the Brahmaput:r;a drainage basin. It is of interest
to note that all the specimens recorded here were obtained from the
Chindwin drainage basin.
The largest specimen is about 87 mm. in total length.
lOne of us (Hora, Bee. Ind. Mus. XXV, p. 8, 1923) considered Exostoma Blyth to
be a synonym of Glyptothorax Blyth as the type-speoies of the former-E. berdmorei
Blyth-is indistinguishable from species included under Glyptothorax. The name is,
however, used here in a loose but hitherto accepted sense, for the sake of oonvenience.
1935.] S. L. HORA & D. D. MUKERJI: 'Fish of the Naga Hills. 387

CYPRINIDAE.
~
Danio (Danio) aequipinnatus (McClelland).
1934. Danio (Danio) aequipinnatus, Hora & Mukerji, Ree. Ind. MU8. XXXVI,
p. 131.
1 specimen. Mohumi village, sources of the Rengma River. J. H. Hutton,
March 1927.
3 speoimens. Chakhabama, Sidzu River. B. Prashad & B. Chopra, 17th
February 1935. ,
Several specimens. Emilomi, Keleki stream. B. Prashad & B. Chopra,
26th February 1935.
Danio aequipinnatus is a widely distributed species, though in the
collections before us it is represented from 3 localities in the Brahma-
putra basin.
The largest specimen is about 102 mm. in total length.

Danio (Danio) naganensis Chaudhuri.

1912. Danio naganensis, Chaudhuri, Ree. Ind. Mus. VII, p. 441, pI. xl, figs. 1,
la, lb.
1 specimen. Purobami and Sahunyu, sources of the Tizu River. J. H.
Hutton, March 1927.
12 specimens. Melori, Tizu River; Leori, Phodung River, a tributary
of Tizu; Yisisu, Yazhiluwu River; a tributary of Tizu. J. H. Hutton, March
1927.
2 specimens. Laruri, Zhuzeti stream, a tributary of Titlo or Namtaleik.
J. H. Hutton, March 1927.
4 specimens. Khezhabama (Chizami), Chiteri stream. B. Prashad &
B. Chopra, 20th Fobruary 1935.
Numerous specimens. Sakhai, Lizho stream. B. Prashad & B. Chopra,
23rd February 1935.
Numerous specimens. Sakhai, Tizu River. B. Prashad & B. Chopra,
23rd February 1935.

Danio naganensis was described by Chaudhuri from the Lungting

River, Naga Hills, Manipur. The collections under report show that
the species is very common and widely distributed in the Chindwin
drainage of the Naga Hills. In most of the characters it is similar to
D. aequipinnatus, but can be distinguished from it by its less deep body,
one broad lateral band and eight divided rays in the dorsal fin. More-
over the eggs of D. naganensis are relatively much larger than those of
D. aequipinnatus. From the distribution of the two species, as judged
by the collections under report, it seems probable that naganensis re-
places aequipinnatus in the Chindwin drainage of the Naga Hills.
The largest specimen is about 88 mm. in total length.

nanio (Brachydanio) acuticephala Rora.

1921. Danio (Brachydanio) acuticephala, Hora, Roo. Ind. Mus. XXII, p. 193.
B specimens. Zhokami and Tekhubami, sources of the Dayang River. J.
H. Hutton, March 1927.
3 specimens. Purobami and Sahunyu, sources of the Tizu River. J. H
Hutton, March 1927.
Brachydanio acuticephala was described from a very large number
of specimens collected from the small streams and ponds of the Manipur
Valley. Dr. J. H. Hutton obtained specimens from springs forming
388 Records of the I rtdian Museum. [ VOL. XXXVII,
the sources of the Dayang (Brahmaputra basin) and Tizu (Chlndwin
basin) Rivers.
In ourl recent key to the species of the subgenus B1'achydanio, acuti-
cephala was included among the forms characterised by "Lateral line
short, not extending beyond pectoral fin." We find, however, that
the lateral line is totally absent in this species. The absence of barbels
distinguishes it from the other species in which the lateral line is totally
absent.
The largest specimen in Dr. Hutton's collection is about 35 mm.
in totallength.

Barilius barila (Hamilton).

1921. Ba'J'ilius barila, Hora, Ree. Ind. Mus. XXII, p. 190.
2 specimens. Purobami and Sahunyu, sources of the Tizu Uiver. J. H.
Hutton, March 1927.
1 specimen. Chakabama, Sidzu River. B. Prashad & B. Ohopra, 16th
February 1935.
4 specimens. Sakhai, Lizho stream B. Prashad & B. Ohopra, 23rd
February 1935.
6 specimens. Sakhai, Tizu River. B. Prashad & B. Ohopra, 24th February
1935.

The specimens of Ba'rilius barila in the collections under report

show the morphological characteristics referred to by one of us (Bora,
Ope cit.) as adaptations to life in hill streams.
The largest specimen is about 148 mm. in total length.

Barbus clavatus McClelland.

1921. Barb'us clavatus, Hora, Ree. Ind. Mus. XXII, p. 185, pI. ix, fig. 1.
4 specimens. Chipoketami, Tese-rii River. B. Prashad & B. Chopra,
22nd February 1935,
12 specimens. Sakhai, Lizho stream. B. Prashad & B. Chopra,
23rd September 1935.
7 specimens. Sakhai, Tizu River. B. Prashad & B. Chopra, 23rd September
1935.
One of us discussed the systematic position of Barbus clavatus in
1921 and re-described the species from specimens collected in the Sena-
pati stream (Barak River) at Kairong. The species had hitherto been
found only in the Brahmaputra drainage system, but all the specimens
under report were collected from the Tizu River or its tributary streams
which ultimately drain into the Chindwin River. The specimens are
well preserved and agree with the examples described from Kairong.
B. clavatus is readily distinguished by its large armed dorsal spine
and by the black markings on the membranes in between the dorsal
fin rays.
Attention may here be directed to the fact that one of us2 described
a variety of this species from the Myitkyina District, Upper Burma.
The Burmese form is distinguished by its longer snout, shorter dorsal
spine, lepidosis and colouration.
The largest specimen is about 200 mm. in total length.
1 Hora ~ Mukerji, Bee. Ind. MU8. XXXVI, pp. 130, 131 (1934).
2 MukerJl, Journ. Bombay Nat. Ht'st. Soc. XXXVII, pp. 64.67, text.figs. 10, II;
pI. iii, fig. 1 (1934).
1985.] S. L. RORA & D. D. MUKERJI: F'ish of tke Naga Hills. 389

Barbus bexagonolepis McClelland.

1839. Barb'lt.s ht:iCagonolepis, McClelland, As. Res. (Ind. Cyprinidae) XIX
pp. 270, 336, pI. xli, fig. 3.
1878. Barbus hexagonolep1's, Day, Fish. Ind., p. 564, pI. cxxxvii, fig. 4.
1921. Barbus hexastichus, Hora (nee McClleland), Rec. Ind. MU8. XXII, p, 186.
1924. Barbu8 hexastichu.s, Hora (nee McClelland), Rec. Ind. Mus. XXVI, p. 27.
Several specinlans (fry). Melori, Tizu River; Leori, Phodung River, a
tributary of Tizu; Yisisu, Yazhiluwu River; a tributary of Tizu. J. H.
Hutton, March 1927.
Several specimens (fry). Khazh.abama (Chizami), Chiteri stream. B.
Prashad & B. Chopra, 20th Febrqary UJ35.
5 specimens. Chipoketami, Tese~rii niver. B. Prashad & B. Chopra,
22nd February 1935.
7 specimens. Sakhai, Lizho stream. B. Prashad & B. Chopra, 24th Feb-
ruary 1935.
Several specimens (fry), Emilomi) Keleki strealn. B. Prashad & B.
Chopra, 26th February 1935.

Barbus hexagonolepis appears to be the commonest species in the

hills of Assam. According to Day" The character of the interrupted
groove behind the lower lip at once distinguishes this species from B.
hexastichus." It seems probable that the fry of both the ~pecies have
black spots on either side before the base of the caudal fin. This juvenile
character of the fry led one of us to refer specimens of B. hexagonolepis
from the Naga Hills and the Garo Hills to B. hexastichus. A thorough
examination of the specimens has shown that B. hexagonolepsis has a
more slender body form, 28:30 scales along the lateral line and an in-
terrupted groove behind the lower lip.
Barbus hexastichus is very closely related to B. to'l' (Hamilton). The
races and varieties of the latter are not yet properly understood. It
is difficult, therefore, to define the precise systematic position of B.
hexastichus.
The largest specimen is about 300 mm. in total length.

Barbus ticto (Hamilton).

1921. Ba'rb'lIS tieto, Bora, Bee. Ind. M'U8. XXII, p. 187.
1 specimen. Mohumi, sources of the Rengma. Rivor. J. H. Hutton,
March 1927.

The only specimen of Barbus tieto in Dr. Hutton's collection is about

67 mm. in total length. Besides the usual spots on the sides of the
tail, the distal half of the dorsal fin is grayish in colour.

Crosso~hilus latius (Hamilton).

1934. Gro8soehilu8 latius, Mukerji, Journ. Bombay Nat. Hist. Soc. XXXVII,
p. 49, fig. 6.
1 specimen. Melori, Tizu River. J. H. Hutton, March 1927.
The specimen of Orossochilus latius conforms to the Siamese and
Burmese form of the species as recently defined by one of us (Mukerji,
Opt cit.). This race differs from the forma typiea from Northern Bengal
in having 8 scales in a transverse series and 15 to 16 round the caudal
peduncle. The form is depressed and slender and the head is relatively
longer (4·3 to 4·8 times in the length of the body without the caudal
as agaiDs~ 5 to 5·2 times in the typical form).
The specimen is about 120 mm. in total length.
390 Records of the Indian Museum. [ VOL. XXXVII,
Garra lissorhync.hus (McClelland).
1921. Garra lissorhynchus, Hora, Ret. Ind. MU8. XXII, p. 662, pI. xxvi, figs.
2,2a.
1 specimen. Zhokami and Tekhubami, source of the Dayang River.
J. H. Hutton, March 1927.
2 specimens. Sakhai, Lizho stream B. Prashad & B. Chopra, 23rd
February 1935.
5 specimens. Emilomi, Keleki stream. B. Prashad & B. Chopra, 26th
}'ebruary 1935.

In all the 8 specimens the dorsal and the caudal fins bear the charac-
teristic markings of the female as described by one of us (Hora, Ope cit.).
In 5 examples the body cavity is full of ova in various stages of develop-
ment, while in the remaining three there are only empty sac-like struc-
tures.
G. lissorhynchus has so far been known from the Khasi and Jaintia
Hills which are situated in the Brahmaputra basin. The species is
here recorded from the Chindwin basin for the first time.
In colour markings the female specimens of G. lissorhynckus resemble
G. abhoyai Hora, but the two species differ in general facies and lepidosis.
The largest specimen is about 75 mm. in total length.

Garra kempi Hora.

1921. Garra kempi, Hora, Rec. Ind. Mus. XXII, p. 665, pI. xxvi, figs. 3, 3a.
8 specimens. Purobami and Sahunyu, sources of the Tizu River. J. H.
Hutton, March 1927.
15 specimens. Melori, Tizu River; Leori, Phodung R., triQutary of
Tizu; Yisisu, Yazhiluwu R., tributary of Tizu. J. H. Hutton, March 1927.
3 specimens. Laruri, Zhuzeti stream, a tributary of the Titlo or Namta-
leik. J. H. Hutton, March 1927.
Numerous specimens. Sakhai, Lizho stream. B. Prashad & B. Chopra,
23rd February 1935.

Garra kempi was originally described from a single specimen procured

by Dr. S. W. Kemp from the Siyom River, below Damda in the
Abor Hills. The waters from these hills drain into the Brahmaputra.
The large number of specimens under report come from several small
streams which join the Tizu Ri'ler, a tributary of the Chindwin River.
The species is readily distinguished by the large size of its mental
disc and by the fact that the anal opening is situated very far forwards,
i.e., almost midway between the commencement of the anal and that
of the ventral fins.
The largest example is about 110 mm. in total length.

Garra naganensis Hora.

1921. Garra naganensis, Hora, Rec. Ind . .Jlus. XXII, p. 667, pI. xxv, figs. 2,
2a.
!) specim.ens. Zhokami and Tekhubami, source., of the Dayang River.
J. H. Hutton, March 1927.
1 specimen. Purobami and Sahunyu, sources of the Tizu River. J. H.
Hutton, March 1927.
7 specimens. Melori, Tizu River; Leori, Phodung River, a tributary of
the Tizu; Yisisu, Yazhiluwu River, tributary of Tizu. J. H. Hutton, March
1927.
1 specimen. LarUl'i, Zhuzet stream, a tributary of the Titlo or N amtaleik.
J. H. Hutton, March 1927.
 1935.] S. L. HORA & D. D. MUKERJI: Fish of the Naga Hills. 391

2 speoimens. Stream at Mao. B. Prashad & B. Chopra, 14th February

1935.
17 speoimens, ~Sakhai, Tizu River. B. Prashad & B. Chopra, 23rd Feb-
ruary 1935.
Numerous specimens. Emilomi, Keleki stream. B. Prashad & B. Chopra,
26th February 1935.

Garra naganensis was described from a single specimen collected

at Kairong in the Senapati stream (Barak River) which drains into
the Brahmaputra. The collections under report contain a large number
of specimens of all sizes both from the Brahmaputra and the Chindwin
drainage systems.
The species is distinguished from the preceding form by its small
transversely oval mental disc and by the fact that the anal opening
is situated much nearer the commencement of the anal fin than that
of the ventral fins.
The largest specimen is about 130 mm. in total length.

Oreinus molesworthi Chaudhuri.

1913. Oreinu8 molesworthi, Chaudhuri, Ree. Ind. MU8. VIII, p. 247, pI. vii,
fig. 2, 2a, 2b.
7 speoimens. Sakhai, Lizho stream. B. Prashad & B. Chopra, 23rd Feb-
ruary 1935.
2 specimens. Sakhai, Tizu River. B. Prashad & B. Chopra, 24th Feb-
ruary 1935.
1 specimen. Emilomi, Keleki stream. B. Prashad & B. Chopra, 26th
February 1935.

Ore-inus molesworthi was described from a single specimen obtained

at Yembung in the Abor Hills. S~ce then it has been found in the
Eastern Himalayas below Darjeeling1 • The range of the species is here
extended to the Chindwin drainage system for the first time.
The largest specimen is about ,235 mm. in total length.

PSILORHYNCHIDAE.
PsilorhYDchus homaloptera, sp. nov.
(Plate VII, figs. 1-6).
17 specimens. Emilomi, Keleki stream. B. Prashad & B. Chopra, 26th
February, 1935.

D. 2/7; A. 2/5; P. 7-8/9; V 2/7; C. 18: L. 1. 42-44; L. trw 6-7

(3+3 or 3!+3!).
In its depressed and flattened qody, general facies and build Psilor-
hynchus homaloptera superficially resembles some of the highly specialised
Homalopterid fishes. The dorsal profile is evenly arched; it rises
gently from the tip of the snout to the insertion of the dorsal fin, beyond
which it slopes down gradually to the root of the caudal fin. The ventral
profile is more or less horizontal or very slightly arched. The ventral
surface of the head and the chest are flat and horizontal. The maximum
depth of the body lies in front and below the dorsal fin. The caudal
peduncle is short and fairly compressed; its least height is contained
from 1·4 to 1·7 times in its length.
1 Hora, Ree. Ind. MU8. XXII, p. 734 (1921).
 Records of the Indian M useu'»l. [VOL. XXXVII,.

The head is greatly depressed, small and sub-triangular. The length

of the head is contained from 5·1 to 5·6 times in the length of the body
excluding the caudal fin. The height of the head at the occiput is equal
to its length behind the nostrils. The snout is spatulate, smooth,
obtusely pointed and has ~renchant margins; it is considerably longer
than the post-orbital part of the head, its length being contained about
2 times in the length of the head. The eyes are rather small, globular
and dorso-Iateral; they are situated almost entirely in the posterior
half of the head. They are proyided with free orbital margin and are
not visible from the ventral surface. The diameter of the eye is contained
from 3·6 to 4 times in the length of the head. The interorbital space
is flat and much wider than the orbit. The gill openings are narrow;
they extend from the lateral line as far down as the termination of the
base of the pectoral fins. The nostrils are fairly large and situated
much nearer the anterior margins of the eyes than the tip of the snout;
they are separated by a conspicuous membranous flap. The internarial
distance is almost equal to the length of the head behind the middle
of the eyes.
The mouth is ventral and its opening is slightly arched; the width
of the gape is equal to the length of the head behind the middle of the
eyes. The upper jaw overhangs the vestibulum of the mouth. Both
the upper and the lower jaws are provided with sharp rasping horny
edges, bordered by thick, fleshy lips which are entire. The lips are not
continuous at the angles of the mouth. The upper lip is narrow, some-
what tough and partly covered with the rostral fold. The lower lip,
with its posterior superficial part, i~ in the form of a broad, thick flap
which is free from the jaw. A fairly deep, distinct lateral furrow passes
on either side from the post-labial groove to the sides of the snout,
and marks off the mouth parts and the thick rostral fold. The rostral
fold and the under surfaces of the chins are sparsely studded with minute
tubercles, while the skin behind the lower lip is somewhat papillated.
The dorsal fin is in advance of the ventrals and its commencement
is much nearer the tip of the snout than the root of the caudal; it is
almost equidistant between the anterior margin of the eye and the
commencement of the anal. The height of the fin is somewhat variable;
it may be equal to or a little higher than the depth of the body below it.
Its free margin is straight and oblique. The pectoral fins are pedun-
culate, broad, expanded and fan-shaped with a rounded free margin.
They are horizontally placed, considerably longer than the head and
are separated from the ventrals by a distance equal to one-third of their
own length. The ventral fins are like the pectorals; they are somewhat
shorter and have an oblique base, so that the fins on the two sides con-
verge when folded back. They have a more or less truncate free margin,
and they extend considerably beyond the anal opening which is situated
nearer the tip of the ventrals thari their origin. The anal fin is short
and, like the dorsal, has a straight and oblique outer margin. It is
about twice as long as its base and about one and half times longer than
the height of the body above it. The caudal :fin is lunate with unequal
and angulate lobes, the lower lobe being slightly longer than the upper.
It is considerably longer than the head and its own height.
1985.] S. L. HORA &; D. D. MUKER.JI: Fish of the Naga Hills. 393

The lateral line is feebly arched and runs along the middle of the
body to that of the root--cf the caudal fin. The scales are fairly large,
thin, and firmly adherent; they are arranged more or· less regularly
on the body. There are from 14 to 15 scales before the_ dorsal fin and
12 round the caudal peduncle. The number of-perforated scales along
the lateral line varies from 42 to 44. The number of scales in a transverse
series between the bases of the dorsal and the ventral fins is also some-
what variable. Usually there are 7 rows, 3i rows being arranged between
the base of the dorsal and the laterallin.e ana 3t rows between the lateral
line and the bases of the ventrals. In some specimens, there are only
6 rows, 3 rows above and 3 rows below the lateral line. The under
surface of the body from the chest to the origin of the velltrals is per-
fectly smooth and devoid of scales.
The air bladder though still of the normal Cyprinid type -has under-
gone considerable degeneration. In all essential features it is like that
of P. balitora and P. 8ucatio, but is very much reduced. The anterior
chamber is pea-shaped and somewhat flattened' dorso-ventrally. The
posterior chamber is very much smaller than the anterior chamber,
but is more or less similar in shape. It has shifted from its original
or normal position and is superimposed medially on the posterior part
of the anterior chamber where it is partly tucked in a shallow invagina-
tion. The walls of both the chambers are thick and fibrous. The
transverse processes of the second vertebra, on which the air-bladder
rests, are considerably modified (pI. vii, fig. 5) so as to provide a bony
covering for the anterior end of the bladder.
The basipterygium, like that of P. balito1'a, described and figured
by one of usl , has undergone complete ossification so as to form a shield-
like structure, similar to that found in most Homalopterid fishes. It
is slightly longer than broad, broadest at the zone of the ventral fin
rays, and is destitute of well differentiated antero-medial or lateral.
horns. Anteriorly the basipterygum. is broadly rounded. A fairly
elevated, oblique and narrow ridge passes on either side on the dorsal
surface of the basipterygium, almost parallel to the antero-Iateral border.
It extends a.s far as the anterior margin of the zone of the ventral fin
r~tys and externally cuts off a fairly wide, elongated strip which is COll-
Cltve throughout, while internally these two longitudinal ridges, together
'with the postero-medial ridges meeting in the centre, enclose an obtusely
triangular, broad space. This triangular enclosed area has a heart-
shaped central, convex surface. ~he latero-posterior border of the
bnsipterygium, i.e., the zone of ·the ventral fin rays, is more or less flat
al1d is provided with an antero-Iateral and a postero-lateral angle. Near
each antero-Iateral angle are present two 1at.eral foramina fOl' connecting
the basipterygium with the modified rib. This state of affair is very
characteristic of P. homaloptera. In P. balitora and P. sucatio lateral
foramina are altogether absent, while in Homalopterid fishes, ,vhen
present, there is only a single foramen. The postero-medial horns are
fairly developed and aloe widely separated by flat transverse' processes.
Above the base of each postero-lateral horn there is a short and angular
sub-vertical process (pI. vii, fig. 4).

J M:ukerji, Jount. Bumb(lllJ Nat, Hif>t. Soo. XXXVI, ~. 826, fig. 2 (1933).
394 Records of the Indian Museum. [ VOL. XXXVII,

The structures of the pharyngeal bone and teeth (pI. vii, fig. 6)
coo-espond almost entirely to those of P. sucatio described and figured
by one of usl . The pharyngeal bones are roughly triangular in outline,
a~d in a specimen, 80 mm. long, the length of their base is about 2 mm.
The teeth are set close together on a broad apex; they are slender,
almost of uniform length and thickness, hooked at the tip and arranged
in a. single row. Their approximate length is 0·6 rom.
The colouration in alcohol is uniformly pale olivaceous green,
except for the ventral surface which is somewhat lighter. The head
and the body are spattered all over with fine black dots. A fairly broad
dusky lateral band passes, along the lateral line, from the angle of the
operculum to the root of the caudal fin. In certain specimens a similar
band passes dorsally from the occiput to the tail, while in others this
band may break up, partly or entirely, into small irregular blotches.
All the fins are more or less dusky.
Type-specimen.-F. 11792/1, in the collection of the Zoological
Survey of India (Ind. Mus.), Oalcutta.

Measurements in millimetres.
Type.
Length of body without caudal 64·0 59·0 57·0 56'0
Height of body 12·0 11·0 9·5 9'0
Length of head 12·0 11'0 11·0 10'0
Breadth of head 10·0 10-0 10·0 10'0
Height of head at occiput 7·5 7·0 7·0 6·5
Length of snout 6·0 5·5 5·5 5'0
Diameter of eye 3·0 3·0 3·0 2·5
Interorbital width 5·0 5·0 5·0 4·5
Height of dorsal fin 12·5 11·0 10·0 10·5
Length of pectoral :fin 16·5 15·0 13·5 13'0
Height of anal fin 9·0 8·5 8·0 8'0
Length of caudal fin 14·0 13·0 12·5 12·5
Length of caudal peduncle 8·0 9·0 8·0 8·0
Least height of caudal peduncle 5·5 5·5 5·0 4°5

Remarks.-The systematic position of the genus Psilorhynchus has

been elucidated only in recent years 2 and from a study of the abundant
material of P. sucatio and P. balitora it has been shown that these fishes
are sufficiently remarkable to constitute a separate family. They are
distinguished by a peculiar type of mouth and lips; absence of barbels;
single row of pharyngeal teeth; free, though greatly reduced air-bladder
in the abdominal cavity; completely ossified, broad, plate-like basip-
terygium; and the presence of a number of unbranched rays in the
pair fins. Though the new species described here has assumed the form
of the most highly specialise¢! Homalopterid fishes, it exhibits all the
above characters in a marked degree. P. homaloptera, sp. nov. is not
only abundantly distinct in general facies and build, but can also be
distinguished by its small, dorso-Iateral eyes and the larger number of

Mukerji, Journ. Bombay Hat. Hist. Soc. XXXVI, p. 828, pI. i, fig. 5 (1933).
1
Hora, Bee. Ind. Mus., XXVII, pp. 457-460, text-fig. 3 (1925); Mukerji, Journ.
2
Bombay Nat. Hi8t. Soc., XXXVI, pp. 823-828 (1933).
1935.] S. L. HORA & D. D. MUKERJI: Fish of tke Naga Hills. 395

undivided rays in the pectoral fins. The absence of scales along the
entire ventral surface in b;ont of the anal fin is another distinguishing
feature of the species.

c.
TEXT-FIG. 2.-Lateral views of three Indian species of P8ilorhyncTtu8 McClelland.
a. Psilorhynchu8 sucatio (Hamilton). xli.
b. Psilorhynchu8 balitora (Hamilton). X ca. 1-1-.
c. P8ilorkyncku8 komaloptera, sp. nov. xlI.

Our field-notes concerning the ecology of the Indian species of Psi-

lorhynchus show that the differences in their form and structure can be
correlated with differences in their habitats. P. suca?io has a shovel-
like head, large lateral eyes, fairly complete lepidosis on the ventral
surface behind the bases of the pectorals, a fewer number (4) of Ull-
branched rays in the pectoral fins, and a spindle-shaped form. The
species was found by us in larger streams at Siliguri where the bed was
either muddy or where a large pool had formed. Specimens were also
collected from pebbly beds, but they were not so common in such situa-
tions. P. balitora has a Garra-like appearance, somewhat smaller,
dQrso-lateral eyes which are not visible from the ventral surface, irregular
lepidosis on the ventral surface behind the bases of the pectorals, larger
39{) [ VOL. XKX:VII,
number (6) of unbranched raYd in the peotoral fins and moierately
elevated bo:ly. It lives in p~bbly beds of small rapid-running streams
at the base of hills. We collectei a very large number of speoimsn~
from the Sevoke st'ream an:! the Mlhanlddi river at the base of the
Darjeeling Himalayas. The young sp~cimens reported by one of us
(Hora, Reo. Ind. Mu~., XIX, p. 211, 1920; ibid., XXII, p. 182, pI. ix:,
figs. 6, 6a, 1921) from the N aga Hills were collected from a pool in the
CGurse of a torrential stream at Piphim% and as they had not developed
all the characters of the adult yet" they cannot throw much light on
the adaptations undergone by the sp3~ies 1. P. hom'J,loptera has a
greatly depressed and flattened body, much smaller eyes, naked ventral
surface, greatly reduced bladder and large number (8) of undivided rays
in the pectoral fins, which are pe:lunculate. It was collected from the
Keleki stream at Emilomi, but unfortunately we have no particulars
of its precise habitat, though from an analogy with other torrential
fishes, it can be safely assumed that the specimens must have been
collected from rapids in the course of the strea m.
From a close study of the build and habitat of P. suoatio it seems
probable that the fish digd in sand or mud where it lives in burrows,
though it is also capable of adhering to rocks as was observed by one
of us by keeping the specimens in an aquarium. The modifications
for digging purposes are (i) greatly flattened, elongated, shovel-like
head and trenchant snout, (ii) small, ventral mouth, (iii) absence of
barbels, (iv) lateral eyes, (v) a number of undivided rays in paired fins,
(vi) position of the paired fins on the ventral surface, (vii) silvery peri-
toneal lining, and (viii) spindle-shaped body. The position of the
paired fins on the ventral surface is rather significant. In all hill-
stream fishes the paired fins are horizontal and spread outwards so as
to enable the fish to have a greater surface for adhesion; their position
on the ventral surface in P. sucatio, therefore, suggests that they are not
normally used for adhesion in this species. If the fins were to project
out at the sides, they will be a great hindrance at the time of burrow-
ing. The lateral eyes are useful because no mud or sand can adhere to
them. The undivided rays in the paired fins impart strength to the
structures which are presumably used for burrowing. The shape of
the head and body, the form and position of the mouth and the absence
of barbels are all useful characters for a burrowing fish. The silvery or
white peritoneal lining also suggests that the fish does not live in situa-
tions ,vhere it may be exposed to much light.
P. balitora shows modifications for life in rapid-iunning waters,
though as its name indicates it may also be a sand-digger (balitora in
Bengal means a sand-digger). Its form is more elevated, the snout
is not so elongated and depressed, the eyes are dorso-Iateral, the paired
fins are horizontal and project beyond the contour of the body at the
sides; the number of the undivided rays is increased from 4 to 6 and
the peritoneal lining is grayish. In P. hOlnaloptera the body is flattened
and .greatly depressed, the paired fins are pedunculate and spread

.1 It. is likely that the young specimens frC1Dl Piphima belong to our new species from
EmiloIDI, but we leave this question open till specimens of intermediate sizes are dif3-
covered. The Piphima specimens arc too young to warrant any conclusioQs.
193640] s. L. HORA & D. D. MUKERJI: Fisk oj the Naga Hills. 397

outwards and the peritoneal lining is black showing that the fish lives
on surfaces of rocks almost in direct sunlight. The eyes are small and
dorsolateral.
The mogifications undergone by these three species are of a progres-
sive nature and it is remarkable to note that though the original modifica.-
tions may have been induced by a burrowing habit, they have become
accentuated when the same fishes took to life in swift currents. Atten-
tion may also be directed to the fact that similar modifications some-
times result in apparently different .adaptati»ns, but when the nature of
modifications is analysed against the functions of the structures, then
the real value of adaptations becomes evident. For instance, a strong
pectoral :fi.rt. is as useful for digging as for adhering to rocks and in both
cases the result is achieved by the formation of a number of undivided
rays. Bottom life is as natural for a burrowing species as for a species
that adheres to rocks in a torrential stream, and the result is that in
both cases the air-bladder is 'greatly reduced. One of us 1 has .already
directed attention to such cases of adaptations. and explained their
significance.
So far as can be ascertained, Psilorhynchus is restricted to India and
Burma. The two species from China, Psilorhynchus sinensis Sauvage
& Dabry de Thiersant (=Hemimyzon sinensis) and P. fasciatus Sauvage
(=Pseudogastromyzon fasciatus), have been shown to be Homalopterid
fishes 2. The only other species described under this generic denomina-
tion is Psilorhynchus aymonieri Tirant 3 which has not yet been properly
studied. The descriptions and figures of the species recently reprinted
by Ohevey 4 do not show its Psilorhynchus affinities, but it is difficult
to assign any systematic position to it without examining the specimens.

C·OBITIDAE.

Nemacbilus botia (Hamil~n).

1921. Nemachilus botia, Hora, Ree. Ind. Mus., XXII, p. 199.
1 speoimen. Mekruohu stream, Dayang valley. J. H. Hutton, Maroh
1927.
N emachilus botia is represented by a single male specimen about 80
mm. in total length.

Nemachilu.s kangjupkhulensis Hora.

1921. Nemackilus kangjupkkulensis, Hora, Ree. Ind. Mus., XXII, p. 202,
pI. x, figs. 4, 4a.
1934. Nemackilu8 kangjupkhulensis, Mukerji, Journ. Bombay Nat. Hist. Soc.,
XXXVII, p.48.
1 speoimen. Purobami and Sahunyu, sources of the Tizu River. J. H.
Hutton, Maroh 1927.
15 specimens. Melori, Tizu River; Leori, Phodung River a tributary of
Tizu; Yisisu, Yazhiluwu River, a tributary of Tizu. J. H. Hutton, March
1935.
1 Hora, Phil. Trans. Roy. Soc. London (B) CCXVIII, p. 264 (1930) .
•2Hora, Mem. Ind. Mus. XII, pp. 299, 314 (1932).
3 Tirant, Bull. Soc. Etudes lndochinoises (1883).
, Chevey, Oeuvre lchtyologique de G. Tirant, p. 35 (Saigon: 1929.1934).
s
398 Records of the Indian Museum. [ VOL. XXXVII,
16 speoimens. Laruri, Zhuzeti stream, a tributary of Titlo or Namta.leik.
J. H. Hutton, March 1927.
17 specimens. Sakhai, Lizho stream. B. Prashad &' B. Chopra, 23rd
February 1935.
3 specimens. Sakhai, Tizu River. 'B. Prashad & B. Chopra, 24th Febru-
ary 1935.

N emachilus kangJupkhulensis was found to be widely distributed

in the hill-streams of the Manipur Valley whence the waters drain into
the Chindwin River. Mukerji recorded it from 'Burma and in the collec-
tions under report the species was collected from several streams of
the Chindwin drainage only.
The largest specimen is about 72 mm. in total length.

Nemachilus manipurensis, Chaudhuri.

1912. Nemaehilu8 manipurensis, Chaudhuri, Ree. Ind. j}!U8. VII, p. 443, pI. xl,
figs. 4, 4ti, 4b & pI. xli, figs. 1, la, lb.
1921. Nemach,ilu8 manipurensis, Hora, Ree. Ind. Mus. XII, p. 199.
33 specimens. Zhokami and Tekh~bami, sources of the Dayang River.
J. H. Hutton, March 1927.
Numerous specimens. Purobami and Sahunyu, sources of the Tizu River.
J. H. Hutton, March 1927.
3 specimens. Melori, Tizu River; Leori, Phodung River, a tributary of
Tizu; Yisisu, Yazhiluwu River, a tributary of Tizu. J. H. Hutton, March
1927.
1 specimen. Stream at Kohima. B. Prashad & B. Chopra, 14th Febru-
ary 1935.
4 specimens. Chakabama, Sidzu River. B. Prashad & B. Chopra, 16th
February 1935.
Numerous specimens. Paddy fields and irrigation channels, Kekrima.
B. Prashad & B. Chopra, 18th February 1935.
, 23 specimens. Sakhai, Lizho Stream. B. Prashad & B. Chopra, 23rd
February 1935.
2 speeimens. Sakhai, Tizu River. B. Prashad & B. Chopra, 23rd Febru-
ary 1935.
6 specimens. Emilomi, Keleki stream. B. Prashad & B. Chopra, 26th
February 1935.

N emachilus manipurensis was described by Chaudhuri from three

specimens collected in Manipur but later one of us reported on a large
number of specimens from the Manipur Valley and remarked that
"Except for slight variation in the colour of some specimens, they
agree with Chaudhuri's description of the species" The collections
now before us show that according to the colour variation the individuals
of this species can be divided into two groups, i.e. (i) banded forms and
(ii) mottled or plain forms. These differences refer to the colour of the
body only, for in all specimens the dorsal and the caudal fins are marked
with a large number of short, wavy bands. Correlated with the colour
differences, there are certain morphological differenees which are equally
well marked. In the banded specimens the head and body are not
greatly depressed and the ventral surface is only moderately flattened,
whereas in the mottled specimens .the head and body are greatly depress-
ed and the ventral surface is flat and horizontal. When these characters
are analysed against localities, it is seen that the banded and flattened
spe~imens were invariably collected from rapids with pebbly beds,
whereas the mottled specimens were collected in ponds, tanks or other
stationary or sluggish pieces of water. It is significant that, with a few
exceptions, the specimens collected in the Manipur Valley, where the
 1985.] s. t. HORA & D. D. MuxERir : Fisk of the Naga Hills. 399

streams are rather sluggish, are of the mottled or plain type whereas
those from the Naga Hills, where the streams are torrential, are mostly

TEXT-FIG. S.-Three colour variations of N emachilus manipurensis Chaudhuri.

a. A specimen from Sakhai, Naga Hills. X 2.
b. A specimen from. Purobami and Sahunyu, Naga Hills. X ca. 1~~.
c.' A specimen from Kekrima, Naga Hills. X ca.l1.
Specimens with no markings on body from the Manipur Valley have already been
figured by Chaudhuri (Bec. Ind. Mus. VII, pI. xl, fig. 4 and pI. xli, fig. 1, 1912).

banded. As the species is distributed over a considerable area and is

found in different types of habitat conditions, it has undergone remark-
able adaptative modifications.
Nemachilus manipurensis .is similar to N. multifasciatus Day 1 in
several respects, such as incomplete lateral line, barred dorsal and caudal
fins and a large number of bands on the body, but differs in the
form of the caudal fin (deeply emarginate in multifasciatus and truncate
or slightly lobed in manipurensis) and colour pattern on the body (mani-
purensis, when banded, has uniformly narrow bands throughout whereas
multifasciatus possesses few broad bands behind the dorsal and anteriorly
the bands break up into narrower bands). These differences do not
appear to be specific but until fresh material of Day's species becomes
available it is desirable to treat these species as distinct.
The largest specimen is about 80 mm. in total length.
1 Hora, Bee. Ind. Mus. XXXVII, p. 61 (1935).
T
400 Records of the Indian Museum. [VOL. XXXVII,

Nemachilus rupecola (McClelland).

1839. Schist'il1'a rupecula, McClelland Asiat. Re8. XIX (2) pp. 309, 441, pI. lvii,
fig. 3.
21 specimens. Emilomi, Keleki stream. B. Prashad & B. Chopra, 26th
February 1935.
In their general build the specimens of N emachilus rupecola from the
Naga. Hills agree with the form, N. rupecola var. inglisi, recently described
by one of us 1 from the Eastern Himalayas, but differ from it in the
absence of well-marked nasal appendages. It is probably an Assamese
race of the Himalayan species but as the material before us is not in a
good condition we refrain from giving it a separate name. In the typical

a. 6. c.
~rEX'l·.}'IG. 4.-0utline sketches of head of three geographical races of N emachil1.ts rupe.
cola (McClelland).
a. A specimen from Simla Hills. X ca. 2.
b. A specimen from Darjeeling Himalayas. X ca. 2.
c. A specimen from Naga Hills. X ca. 2.
form from the Western Himalayas the head, especially in the region of
the eyes, is not so broad as in the Darjeeling or Assamese forms. More-
over, the scales in the posterior region of the body in the typical form
are well marked. The broadening of the head and the reduction of the
scales in the case of the Darj eeling and Assamese forms show that the
fish is better adapted to withstand
the rush of torrents.
The jaws of N. rupecola are very
characteristic. The upper jaw is
produced into a beak-like process
with a broad, truncate apex, while
the lower jaw is deeply incised to
receive this process. This condition
is present in certain other species of
TEX'l'':FIG. 5.-Front view of jaws of a the genus also but the characters
specimen of N emachil~t8 '1'Upecolu
McClelland from Naga Hills. X 7!. are not well defined.
1= beak-like process of upper jaw; The largest specimen IS about
2= emergination of lower jaw. 102 mm. in total length.
1 Hora, Ree. Ind. Mus. XXXVII, p. 58 (1935).
1935.] s. L. HORA & D. D. MUKERJI: Fisk of the Naga Hills. 401

Nemaehilus suhEusea (McClelland).

1839. Schistura subjusca, McClelland, As. Res. (Ind. Cyprinidae) XIX, pp. 308,
443, pI. liii, fig. 5.
Several specimens. Mekruchu stream, etc., Dayang Valley. J. H.
Hutton, March 1927.
Several specimens. Emilomi, Keleki stream. B. Prashad & B. Chopra,
26th February 1935.
Nemachilus subfusca was described by McClelland from "Upper
Assam" and characterised as "Without spi~s; four cirri in front of
the mouth, various regular zones encircling the body; eyes approxi-
mating to each other on either side of a narrow trenchant ridge like
that of Cobitis phoxocheila; snout hard. Fin rays are,

n. 11 : P. 11 : V 7: A. 7: C. 17."
The above description is too meagre to establish the specific identity
of N. subfusca and for this reason its specific limits have remained un-
determined so far. One of us I , from a study of the Eastern Himalayan
species of the genus, regarded it as a synonym of N. scaturigina (Mc-
Clell.) but the abttndant material now before us shows that the Assamese
form has to be regarded as a separate species. In N. subfusca the lateral
line is complete or incomplete but it is fairly extensive (complete in N.
scaturigina). The scales are feebly developed and in the anterior region
of the body they are totally absent (scales well marked all over the body

TEXT-FIG. 6.-Two colour variations of Nemachilus subfusca (McClelland) from Naga.

Hills.
a. A young spooimen. X ca. 2. b. A mature specimen. X cn. Ii.

in N.scaturigina). In young specimens the colour-bands are more or

less similar to those figured by McClelland whereas in older specimens

1 Bora, Ree. lnd. Nus. XXXVII, p. 64 (1935).

402 Records of the I rulian Museum. [VOL. XXXVII,

the bands are broader and more numerous (incomplete bands in N.

scaturigina). Specimens similar to those before us are present in an
undetermined lot from the Khasi Hills and though McClelland has not
mentioned any definite locality for" his species, it is known that the
originals of several of his Assamese species came from the Mishmi and
the Khasi Hills. It is likely, therefore, that the specimens now before
us represent McClelland's species.
Ne'inachilus subfusca may now be characterised as follows:-

D. 2/8; A. 2/5; P. 1/10; V. 1/6; C. 18.

Nemachilus subfusca is a gracefully built and moderately elongated
species in which the upper profile is evenly arched up to the origin of the
dorsal fin, beyond which it is almost straight or slopes slightly to the
root of the caudal. The ventral profile is more or less horizontal in
front of the ventral fins. The head is long, narrow and somewhat
pointed anteriorly; its length is contained from 4·2 to 5 times in the
total length excluding the caudal and from 5 to 5·8 times including the
caudal. The head is slightly wider than deep; its width is contained
from 1·2 to 1·4 times and the height at occiput from 1·5 to 1·8 times in
its length. The snout is a little longer than the post-orbital part of the
head. The eyes are dorso-Iateral in position and are situated . almost
midway between the free margin of the operculum and the tip of the
snout. They are not visible from the ventral surface. The diameter
of the eye is contained from 4·5 to 6 times in the length of the head and
from 1·5 to 2·7 times in the length of the snout. The interorbital space
is modeJ;ately convex and slightly wider than the diameter of the eye.
The mouth is small and semicircular; its gape is equal to the length of
the snout in front of the nostrils. The lips are fleshy and plain; in
preserved specimens they appear to be feebly corrugated. The lower lip
is interrupted in the middle and deflected at the sides. The jaws are
unequal, the lower one being considerably overhung by the upper. The
lower jaw is sharp and shovel-like, being in some cases faintly incised
in the middle. The upper jaw is produced into a short vertical, beak-
like structure. The barbels are well-developed, all the three pairs are
longer than the orbital width, the outer rostrals being the longest.
The body is sub-cylindrical in shape and somewhat elongated; its
depth is contained from 5·5 to 6·9 times in the total length without the
caudal. The scales are feebly developed and are present only in the tail
region. They appear to be totally absent in the anterior part and on
the ventral surface of the body. Usually the lateral line is complete
running along the middle of the body; when incomplete, it is fairly
extensive and extends as far as the anal fin. The caudal peduncle is
considerably compressed, short and squarish; it is slightly longer than
high.
The dorsal fin commences in advance of the ventrals and its insertion
is somewhat nearer to the root of the caudal fin than to the tip of the
snout; its posterior margin is obliquely truncate. The longest ray
of the dorsal is in most specimens equal to or slightly shorter than the
1935.] S. L. HORA & D. D. MUKERJI: Fish of the Naga Hills. 403

depth of the body below it; only rarely it is longer. The paired fins are
short and horizontally placed; they have rounded margin. The pecto-
rals are nearly as long as the head behind the nostrils; they are separat-
ed from the origin of the ventrals by a distance almost equal to their own
length. The ventrals are shorter than the pectorals and do not extend
as far as the anal opening, which is situated midway bet}Veen the tip
of the ventrals and the commencement of the anal, or a little nearer the
latter. The anal fin is almost as high as the Qody above it; it has, like
the dorsal, a truncate outer margin: The cauttal fin is shorter than the
head; it is emarginate with tho lower lobe slightly more developed.
In spirit specimens the head and body are grayish to dusky above
and pale olivaceous below. In adult specimens there are from 12 to
18 broad, dark, vertical bands which descend from the dorsal surface
and extend almost to the ventral surface. These bands are very variable
in rega,rd to shape, size and position. They may be broader or narrower
than or equal to the alternating bands of the ground colour. They
may be straight, oblique or crescentic. Usually, these bands are more
regular and prominent in the posterior half of the body, while anteriorly
they may either break up into series of irregularly distributed blotches,
or they may be obliterated altogether. In certain specimens, some of
the bands are split into secondary bands. In the young and half
grown specimens, the bands are less variable, narrower and fewer in
number; they are more regularly arranged and seldom exceed 14 in
number. In all specimens the band at the root of the caudal fin is
much darker than the others. There is a prominent black spot at the
base of the anterior rays of the dorsal fin; all the fin rays of the dorsal
are dusky. The rest of the fins are diaphanous.

Measurements in millimetres.
Total1ength inoluding oaudal 82·0 70·0 64·0 51·0 45·0
Length of oaudal 12-0 9·0 9·0. 9·0 7'0
Length of head 14-0 12·0 12·0 10·0 9·0
Width of head 10·0 9·5 9·0 7'0 6·5
Height of head at oooiput 9·0 8·0 7·0 5·5 5·5
Greatest depth of body 12·0 9·0 10·0 7-5 5'0
Length of snout. 6·5 6-0 5·0 4·5 3·0
Diam.eter of eye_ 2·5 2·2 2·0 2-0 2·0
Interorbital distance 3·0 3·0 3'0 2·2 2-0
Length of oaudal pedunole 9·5 9·0 8·0 5·0 5-0
Least height of caudal peduncle- 8·2 8·0 7'0 4·5 4'0
Longest ray of dorsal 10·2 10·2 9·5 8·0 7'0
Length of peotora.l 11·0 9·5 9·5 7·0 7·0
Length of ventral 9·5 g·O 8'5 7·0 6·0
Longest ray of anal 9·0 9·0 8'0 6·5 6·0
Distance between tip of snout and
commenoement of dorsal 36·0 31'0 29'0 24-0 19·5
Distance between commencement
of peotoral and that of ventral 24-5 18'0 16·5 14·0 11-0
Distance between tip of snout and
anal opening 52'0 45'0 40'0 31·5 27'0
404 Records of the Indian M useU1/~. [ VOL. XXXVII,
OPHICEPHALIDAE.

Qphicephalus gachua Hamilton.

1934. Ophicepkalu8 gachua, Hora & Mukerji, Rec. Ind. 1JfU8. XXXVI, p. 135.
S specimens. Zhokami and -Tekhubami, sources of the Dayang River.
J. H. Hutton, March 1927.
6 specimens. Purobani and Sahunyu, sources of the Tizu River. J. H.
Hutton, March 1927.
7 specimens. Melori, Tizu River; Leori, Phodung River, a tributary of
Tizu River; Yisusu, Yazhiluwu River, a tributary of Tizu River. J. H.
Hutton, March 1927.
6 specimens. Kekrima. B. Prashad & B. Chopra, 17th February 1935.
4 specimens. Khazabama (Chizami), Chiteri stream. B. Prashad & B.
Chopra, 20th February 1935.
16 specimens. Phikrokezema. B. Prashad & B. Chopra, 21st February
1935.
1 specimen. Emilomi, Keleki stream. B. Prashad & B. Chopra, 26th
February 1935.
Ophicephalus gachua is a widely distributed species of the genus and
is represented in the collection from both the drainage basins of the
Naga Hills. One of us l had referred specimens of this species collected
in the Manipur Valley to Ophicephalus harcourt-butler-i but we (op. cit.)
ha ve recently shown that the two species are synonymous.
The largest specimen under report is about 150 mm. in total length.

1 Hora, Rec. Ind. 111u8. XXII, p. 208 (1921).

MGIPO-M-llI-8-11-20-9-35-510.