Proc. Indian Acad. Sci. (Anim. Sci.), Vol. 95. No.2, April 1986, pp. 279-288. © Printed in .

Evolution and biogeography of the Indian genera of the family

K C JAYARAM and JR DHANZE Zoological Survey of India, 27, Jawaharlal Nehru Road, Calcutta 700016, India

MS received 18 April 1985; revised 31 October 1985

Abstract. of the family Ariidae are commercially important and are represented in Indian waters by 22 under 5 genera: Arius Valenciennes (18),Osteoqeneiosus Bleeker (I), Hemipimelodus Bleeker (I), Batrachocephalus Bleeker (1) and Ketenqus Bleeker (I). On the basis of certain osteological and morphological characters of the genera, the interrelationships of the Indian genera have been delineated. It is adduced that the Indian Ariidae have had two phyletic lines of evolution. Arius seems to have evolved from Osteoqeneiosus like ancestor whereas Batrachocephalus appears to be from Ketenqus like fishes. Arius is much more advanced than all other Indian genera and has been able to proliferate and reach wider areas than the others. An intermediate connecting link between Ketenqus and Osteoqeneiosus appears to be missing about which information is lacking at this stage.

Keywords. Ariidae; phylogeny; interrelationship of Indian genera.

1. Introduction

Catfishes of the family Ariidae are a commercially important group of marine fishes found in the warm tropical and temperate waters ofthe Indo-Pacific, the Atlantic and the Western Central Pacific regions. About 124 species are known the World over of which 32 are found in the Indo-Pacific, 67 in Central Atlantic and 25 in Western Central Pacific. A few enter freshwater and are confined to these waters. Of the 14 genera known Batrachocephalus and Osteogeneiosus are commonly found in all the fishing areas except Western Central Atlantic (fishing area 31).Arius is found in all the 4 fishing areas and is evidently represented by the largest number ofspecies (58)as compared to others. In respect of other genera, in most cases, the same species is found in the adjacent areas also. The family has not been recorded so far from the Mediterranean Sea (fishing area 37). The distribution of the family is presented in table 1.

2. Biogeography

Regan (1922) stated that the palaeontological evidence to the past history of Ostariophysi is very imperfect. In his review of fossil fishes Peyer (1928) thought that the ariids seem to have been widely distributed in the fresh and salt waters of Europe and elsewhere in Eocene times. Considering the fact that the Siluroids as a group are more or less known from Cretaceous period, it appears that the distribution of the ariids remained the same as they are today and possibly such a pattern came into being some period during pre-Eocene times. The available fossil evidence in respect of Ariidae is very fragmentary indeed. Most fossil identification ofSiluroids are based on fin spines or incomplete crania. Lydekker 279 280 K C Jayaram and J R Dhanze

Table 1. World distribution of genera of Ariidae.

Ocean and FAO fishing area

Western Western Eastern Western Central Central Indian Indian Pacific Atlantic 57 51 71 31

No. of No. of No. of No. of species species species species

Ariopsis x3 Bagre x2 Cathrops x9 Potamarius x2 Gaieichthys x3 Doiichthys x 1 x 1 Hemipimelodus xII x II Ketengus x I x I Nedystoma x 1 x 1 Tetranesodon x 1 x I Batrachocephalus x 1 x I x 1 Osteogeneiosus x 1 x 1 x 1 Arius x 58 x 18 x59 x9 Ancharius x

(1886) recorded Arius species from the Siwalik hills, India Newton (1889) described fossils ofA. bartoni from Eocene beds ofBarton, England. Woodward (1901)recorded A. crassus from the Middle Eocene of Bracklesham, England and A. iherinqi from the Tertiary lignite ofBrazil. Rao (1956)recorded A. kutchensis from the Lower Eocene of Kutch, Gujarat state, India More recently Khare (1975)recorded A. sahnii and Arius sp. from Middle Eocene Maroon argilaceous sandstones of Subathu, formation exposed near Beragua Coal Field, Jammu and Kashmir. From these evidences, it appears that the Ariidae seems to have existed as early as Eocene. Considering the fossil history of other siluroid families, it has already been stated that the Siluroids as a group can be traced only from Cretaceous era. The Bagridae are known as fossils from Pliocene and Pleiostocene.

3. Phylogeny of Ariidae

Gosline (1975) considered the intensive and extensive investigation of particular structure or structural systems as a possible approach to the unravelling of siluroid phylogeny. He considered some units like the caudal skeleton, pelvic girdle, pectoral girdle, Weberian apparatus and associated structures and otoliths as helpful tools. Chardon (1968) in fact utilised some of these in building up his classification of the Siluroids. However, not much information is available in respect of Ariidae. Regan (1911), Eigenmann (1921) and Berg (1940) considered Ariidae as the most primitive Evolution and biogeography of the Indian genera of the family Ariidae 281

next to Diplomystidae, whereas Bertin and Arambourg (1958)thought Bagridae as the most primitive member. Bhimachar (1933) on the basis of his studies on the osteocranium of Arius concluded that it is a more specialised member than other families.Gosline (1944)and Tilak (1965)also came to similar conclusion. Srinivasachar (1980) adduced evidence from chondrocranial studies of some siluroid genera, that Arius has branched quite early in a different direction from a primitive Plotosus like ancestor. Chardon (1968) in his work, opined that the generalised stock of Siluroidei gave rise to many suborders probably independently. The suborder Bagroidei was considered as having evolved as a result ofadaptive radiation. Ariidae was thought to have arisen from a Plotosid ancestor, in support of which it may be stated that Plotosidae are the only Catfishes to share with Ariidae an enormous utriculus.

4. Interrelationship of Indian genera

In the light of studies on the Indian genera of the family Ariidae certain data are presented towards this problem. Ariidae is represented by 22 species under 5 genera in Indian waters: Arius (18), Osteoqeneiosus (1),Hemipimelodus (1),Ketenqus (1),and Batrachocephalus (1).Ofthese Hemipimelodus is almost similar to Arius except in the fact that it has an edentate palate and as such is not considered. Many workers like Misra (1976) and others have misidentified some species of Arius as referable to Hemipimelodus and vice versa. Of the 4 Indian genera excluding Hemipimelodus a comparison of certain os­ teological and a few other morphological characters clearly indicate that Arius is the most specialised of all. The compact cranium, well developed ethmoid bones and epioccipitallamellae, reduced ectopterygoid and entopterygoid (or Mesopterygoid of Gosline, 1975) and the well developed metapterygoid are undoubtedly specialised features (figure lA). The dorsal fin is inserted near the occiput so that the spines form a protective defense triangle. The palate dentition in most species is wen defined than in other genera The extent of lateral fenestra varies from species to species, and the temporal fossa is present in a reduced manner in all species (absent in A. thalassinus). The palatine articulation is from a single point of the ecto, meso and metapterygoid which is also a noteworthy feature (figure IB). Considering the most primitive of the 4 Indian genera, Osteoqeneiosus qualifies for this position. The maxillary bone is produced in the form ofa long semiosseous barbel (figure 2B) with a round base. This may be viewed by some as a very specialised character but this feature is not known in any of the Indian siluroid genera. The ethmoid bones are very much reduced; The lateral ethmoid is peculiar in shape and structure. The dorsal surface ofthis bone is honey combed and the bone is not as solid as is normally found in other Ariid genera (figure 2A). The frontal bones are very much reduced and in the form of thin, rod like structure anteriorly. The metapterygoid is smaller (figure 2C) than in Arius. Perhaps to facilitate movement of the semiosseous maxillary barbels, the extensor tentaculi muscle are well developed only in this genus whereas in other genera it is much reduced and absent in Batrachocephalus. Moreover, in Osteoqeneiosus the palatines are well developed, stout and edentate. The extensor tentaculi, fills up the lateral fenestra and are also attached to the posterior half ofthe palatines. Both the meso and ectopterygoids are toothed and the metapterygoid is much smaller. It would seem that these modifications of the pterygoid bones are as a 282 K C Jayaram and J R Dhanze

ETH PMX--~fo;.''''iI'l LAC --C<7,~-;,;) NA --~~"',4H LETH-~"""" PLETH ---ll--_~ LF ---1'r--~ MF ----=::::::=~~,;:\' FR----\tF

PF ------f',\' SPH---P SOC---+... PT R--~~~,d'. PT 5T

A

MPT HM

PAL

';~ ENPT ~L. J"',';:.~~....

B lOP FIpre 1. Arius caelatus. A. Dorsal view ofskull. B. Hyomandibular suspensorium ofArtus caelatus. result ofthe change in the palata-maxillary mechanism discussed above.The dorsal and pectoral spines are well developed and the dorsal fin is inserted near the occiput as in Arius. Summing up it is evident that Osteoqeneiosus is certainly the most primitive of the 4 genera Batrachocephalus and Ketenqus form one group and share some characters in common as the large sized eyes (figure 3) wide arcuate mouth, reduced maxillaries, reduced palatines etc. However of the two genera, Ketenqus appears to be more primitive than Batrachocephalus. The lateral ethmoid is much reduced and is in the form of a paddle (figure 4A) without handle. Further, in Ketenqus alone, the premaxillaries are welldeveloped as a large plate (figure 4A) with a single row of teeth. This modification is again a unique feature found only in this genus of the Indian Ariidae. In Batrachocephalus the anterior end of the lateral ethmoid is squarish in shape, with the posterior arm interdigitating with the frontal (figure SA). The cranial cavity extends to the ethmoid region and thus the cranium is platybasic in Ketenqus a feature not present in the other 3 genera. The mesopterygoid bone is absent and its place seems to have been occupied by the enlarged metapterygoid (figure SB). Evolution and biogeography ofthe Indian genera ofthe family Ariidae 283

MX

-iI'H~-L F +!-+-4---MF

+-+---FR F SPH PTR MPT·----=~' SOC Q. ~--ST ~!lii.l.--PT TFO c EP A E PL :10----50 CP

Figure 2. Osteoqeneiosus militaris. A. Dorsal viewofskull. B. Extension ofmaxillary bone. C. Hyomandibular suspensorium.

8

Figure 3. Ketenqus typus. A. Lateral view [After Jayararn, 1972. Proc. Zoo/. Soc. Calcutta, 25(2):137, TF II. B. Dentition. 284 K C Jayaram and J R Dhanze

...----ETH --=::---::!=.... LETH PMX .1H<---- -A F "~p<>r----FR

I\if-~--PF MPT /"~-SPH r--....-~\ r-:c··.~_PTR ~-'-4,...... SOC PT ST EPL SOC

tern. I A

Figure 4. Ketenqus typus. A. Dorsal view of skull. B. Hyomandibular suspensorium.

,...----ETH

'f~--SPH 1Hf;~-S OC PTR . PT --. ,.". TF 0 .ST EP EPL SOC

A c

Figure 5. Batrachocephalus minQ A. Dorsal viewof skull. B. Hyomandibular suspensorium. C. Dentition.

From the evidences adduced above it would appear that the Indian Ariidae have had two phyletic lines of evolution. Arius seems to have evolved from Osteoqeneiosus like ancestors, whereas Batrachocephalus appears to be from Ketenqus like fishes. No doubt Arius is much more advanced than all the other genera. It seems likely that an intermediate connecting link between Ketengus and Osteogeneiosus is or was existing about which information is lacking at this stage. s~ Table 2. A comparison of certain osteological differences in 4-genera of Ariidae from Indian waters. E" s.... ::s Characters Arius Osteoqeneiosus Batrachocephalus Ketenqus I:l E. Maxillary Moderately developed. Barbel Produced in the form of a long, Much reduced. No barbel Very much reduced. Barbel present <::r present serniosseous barbel with a round ~c' base f':) ~o "'t Ethmoid Moderately developed, stout, inter­ Much reduced, thin, narrow and Welldeveloped very wide interdigi­ Much reduced, narrow, the pos­ I:l ~ digitating with the frontals pos­ short, interdigitating with the fron­ tating with the frontals posteriorly terior elongation being very thin ;:s- teriorly and the lateral ethmoids on tals posteriorly and the lateral et­ and the lateral ethmoid on either interdigitating with the frontals and '<: with the lateral ethmoid on either .s; either side hmoids on either side side ... ;:s­ side f':) Lateral ethmoid Well developed with a broad an­ Peculiar in shape and structure. The Anterior end squarish in shape, Much reduced, without any pos­ E. terior end interdigitating with the dorsal surface is honey-combed ex­ interdigitating with the ethmoid an­ terior elongation and is in the form - 6' ethmoid anteriorly and with the ternally. The anterior end is trian­ teriorly and with the orbitos­ of a paddle without an arm. ::s ICl orbitosphenoid postero-ventrally gular, interdigitating with the et­ phenoid posteroventrally Interdigitation is only with the et­ tI:l

hmoid anteriorly and with the or­ hmoid anteriorly ~ bitosphenoid posteroventrally I:l .s; Cranial cavity Not platybasic in the sense that it Not platybasic Not platybasic Platybasic in the sense that the does not extend to the ethmoid cranial cavity extends to the et­ ~ region hmoid region ~ 2. Fontaneles Three: anterior and posterior much Three: anterior smallest, middle one Single: posterior end not extending Two: anterior one very large ~ reduced largest, posterior extending to beyond the frontal bones supraoccipital ~

~

N v.00 IV 00 0\

Lateral Moderately large formed by the Very large, formed by the lateral Much reduced Absent ~ fenestra lateral ethmoid and the frontals ethmoid and the frontals (J Frontal bones Well developed Much reduced almost rod like Moderately developed Reduced ~ Temporal fossa Very small, in between supratem­ Same as in Arius Same as in Arius No temporal fenestra but a short t5 poral and pterotics and the post­ shallow, groove is present in be­ i?l temporal bones tween post-temporal, supratem­ ::i poral and pterotics t:o [ Epioccipital Well developed and prolonged Not much developed or elongated Not much developed or elongated Not much developed as in Arius '"~ Palatines Moderately developed and stout. Well developed and stout without Very thin, much reduced without Very much reduced, thin without I::l (>=Autopalatine Without teeth. Articulating with the teeth and overlapping endo- and teeth overlapping endo- and ecto­ teeth overlapping metapterygoid 5' of Gosline) ecto- endo- and metapterygoids ecto-pterygoids posteriorly pterygoids posteriorly and premaxillary ~ from a single point Premaxillary Normal with many rows of teeth Normal with many rows of teeth Normal with many rows of teeth Very broad with a single row of teeth Suspensorium: Ectopterygoid Reduced and shifted anteriorly. No Moderate, with teeth fused with Well developed, without teeth Absent teeth entopterygoid Mesopterygoid Much reduced and lies underneath Moderate with teeth Well developed, without teeth Absent the posterior end of the palatines. No teeth Metapterygoid Large and has occupied the position Small Small Large and perhaps has occupied the of ecto- and entopterygoids position of eeto- and ento­ pterygoids Evolution and biogeography of the Indian genera of the family Ariidae 287

FootDOte to table 2.

Arius Bacrachocephalus

Maxillary bone moderately de­ Maxillary barbel absent. Supra and veloped. Supra- and lateral ethmoid lateral ethmoid weU developed. developed. Cranium not platy basic. Cranium not platybasic. Lateral and Lateral fenesta moderately large. temporal fenestra reduced. Ecto­ Ecto and endopterygoid reduced, and endopterygoid present.

r WOOdy distributed Indian Ocean and t W. Central I Pacific

Osteoqeneiosus .....E------...... Ketenqus

Maxillary bone produced. Eth­ Maxillary barbel present. Supra­ moids, frontals reduced. Lateral and lateral ethmoid reduced. Platy­ Ethmoid honey combed. Meta­ basic cranium, Lateral fenestra ab­ pterygoid smaller. Extensor ten­ sent. No temporal fenestra. Ecto taculi weU developed [Indian Ocean and endopterygoid absent only]

Eastern Indian Ocean and W. 1__$_1 Central Pacific

Ariidae

Semi benthic marine fishes a few entering freshwater

Indian Ocean Central Pacific Central Atlantic

Probable phyletic line of Indian Ariid genera

5. Conclusion

The interrelationships of the Indian genera as discussed above indicate clearly the ancient nature ofthe family which has diversified in two different phyletic lines. It also establishes the fact that Arius is more successful of the 5 genera and has been able to proliferate and reach wider areas than its ancestors in view of its specialised characteristics and better adaptability.

Acknowledgements

The senior author gratefully acknowledges the kind deputation granted by the Department of Environment, New Delhi which enabled him to present this paper 288 K C Jayaram and J R Dhanze before the IVth European Congress of Ichthyologists, at Hamburg, 1983. Dr Michel Chardon, Institute de Zoologie, Liege (Belgium) reviewed the manuscript for which the authors are thankful to him.

Reference

Berg L S 1940Classification of fishes,both recent and fossil; Trao.Lnst. Zool. Acad. Sci. USSR S 517 (Russian and English texts. Also reprint, Ann Arbor, Michigan, USA, 1947) Bertin Land Arambourg C 1958 Super- des te1eosteens (Teleostei); Trait. Zool. Paris 13 2204-2500 Bhimachar B S 1933 On the morphology of skull of certain Indian Catfishes; Half-year J. Mysore Univ. 7 233-267 Chardon M 1968 Anatomie comparee de I' appareil de Weber et des structures connex chez les siluriformes; Ann. Mus. Roy. d'i Afr. Centrale, No. 169, p. n7 Eigenmann C 1921 The origin and distribution of the genera of fishes of South America, West of the Maraiaibo, Orinoco Amazon and, Titicaca basins; Proc, Am. Philos. Soc. Philadelphia 9 Gosline W A 1944 The problem ofthe derivation ofthe south American and African freshwater fish faunas; Ann. Acad. Bras. Cien. 16211-223 Gosline W A 1975 The palatine-maxillary mechanism in catfishes with comments on the evolution and zoogeography of modern siluroids; Occas. Pap. Calif. Acad. Sci. No 120, 1-31 Khare S K 1973 Eocene fishes and turtles from the Subathu formation. Beragua Coal mine, Jammu and Kashmir; J. Pal. Soc. India 18 36-43 Lydekker R 1886 The Indian tertiary and post-tertiary vertebrate-Tertiary fishes; Pal. Indica 3 241-264 Misra K S 1976 The Fauna of India and the adjacent countries. Pisces, 2nd edition, 3 pp. xxi + 367, 15 pis Newton E T 1889 A contribution to the history of Eocene siluroid fishes; Proc. Zool. Soc. London 201-207 Peyer B 1928 Die Weise des agyptischen Alttertiars nebst liner kritischen ubersicht uber aile fossilen Weise; Abh. Bayer Akad. Wiss. Math. Naturwiss. Kl. 3261 Rao V R 1956The skull of an Eocene siluroid fishes from Western Kutch, India; J. Pal. Soc. India 1131-185 Regan C T 1911 The classification of the teleostean fishes ofthe order Ostraiophysi 2. Siluroidea; Ann. Mag. Nat. Hist. 8 553-577 Regan C T 19n The distribution of the fishes of the order Ostariophysi; Bijdr Dierkd 22 203-207 Srinivasachar H R 1980 Evolution and of Catfishes of India; Proc. Indian Nat. Sci. Acad. B46 23-26 Tilak R 1965 The comparative morphology of the osteocranium and the Weberian apparatus of Tachysuridae (Pisces: Siluroidei); J. Zool. 146 150-174 Woodward A S 1901 Catalogue ofthe fossil fishes in the British Museum Natural History (London, British Museum) 4 330

Abbreviations used in the figures: A, Anguloarticular; AF, anterior fontanele; D, dentary; EP, epioccipital; ECPT, ectopterygoid; ENTP, mesopterygoid; EPL, epioccipital larnellae; ETH, ethmoid; FR, frontal; HM, hyomandibular; rop, inter-operculum; LAC, lachrymal; LF, lateral fenestra; LETH, lateral ethmoid; MAX, maxillary; MF, median fontanele; MPT, rnetapterygoid; NA, nasal; OP, operculum; PAL, palatine; PF, posterior fontanele; PLETH, posterior process oflateral ethmoid; PM X, premaxillary; POP, preoperculum; PT, supracleithrum; PTR, pterotic; Q, quadrate; SOl,2.3 suborbitals 1,2,3; SOC, supraoccipital; SPH, sphenotic; ST, post temporal; SOCP, supraoccipital process, TFO, temporal fossa.