argus (Anura) in Natal: , biogeography and conservation

J.e. Poynton Department of Biological Sciences, University of Natal, Durban

A high degree of variation in the colour pattern of Hyperolius argus Peters (1854) is a particularly large-sized Hyperolius argus Peters, including sexual dichromatism and species of reedfrog, with - as the species is presently under­ a marked cline, has led to much taxonomic confusion. This stood - a large 'East African' range extending from Durban East African species extends down the Natal coastal to southern (Lanza 1978) along the coastal lowlands. lowlands as far south as Durban. It has been assigned to H. punctlculatus (Pfeffer) in Natal. In a discussion of It shows an unusually high degree of variation in the colour ' and Natal material, this assignation is shown pattern, including sexual dichromatism, which is further to be incorrect. The probable holotype of Cope's complicated by a marked north to south cline in the pat­ cinctiventrls is considered to be a synonym of argus. There terning. Not surprisingly, this has led to much taxonomic is some evidence that populations of argus intergrade with confusion. semidiscus Hewitt in Natal, and the occurrence of semidiscus on the periphery of the argus range is In this paper, fIrstly the taxonomy is briefly reviewed for discussed in relation to the 'central-marginal' model of the purpose of validating the use of the name argus Peters biogeographical patterning. It has not been possible to for Natal material, which at the time of the last revision of determine the relative vulnerability to habitat disturbance of Natal Hyperoiius was assigned to puncticuiatus Pfeffer (1893) . ) argus compared to any other species, but the will

0 (poynton 1964). Secondly, this paper discuss the bio­

1 conservation prospects of this species at the southern end

0 of its range currently seem extremely poor. geography and conservation of H. argus as a particular 2 instance of an 'East African' species that reaches the southern­

d S. Afr. J. Zool. 1986,21: 149-152 e

t most limit of its range in the highly disturbed environment a 'n Hoe graad van variasie in die kleurpatroon van Hyperolius of Durban. d ( argus Peters, insluitend seksuele dikrornatisme en 'n r e karaktergradient het gelei tot taksonomiese verwarring. Die h The name Hypero/ius argus s Oos-Afrikaanse spesie strek oor die Natalse kuslaaglande i l b so ver suid as Durban en is toegewys aan H. puncticulatus The high degree of polychromatism in both males and females u (Pfeffer) in Natal. 111 'n bespreking van die Mosambiekse en

P of H. argus was described and illustrated by Parker (1930), Natalse rnateriaal, word hierdie indeling as foutief bewys. e who studied the large series collected by Cott (1932) in the h Die rnoontlike holotipe van Cope se cinctiventris word t lower Zambezi, not far from the type locality of the species y beskou as 'n sinoniem van argus. Daar is bewys dat b

bevolkings van argus inskakel by semidiscus Hewitt in (Boror, approx. 1736 C). Parker showed that adult males d e Natal en die verspreiding van semidiscus op die periferie retain the juvenile colouration and markings, presenting a dark t n van die argus-verspreidingsgebied word bespreek in canthal line, darkening on the upper eyelid, and, occasionally, a r verhouding tot die 'sentraal-marginale'-model van die dark stippling over the dorsum. A light dorsolateral stripe is g

biogeografiese patroon. Dit was nie moontlik om die e usually evident, tending to pass forwards over the upper eyelid c relatiewe kwesbaarheid van argus jeens habitatversteuring n along the canthus rostralis. The upper surfaces were described

e in vergelyking met enige ander amfibiese spesie te bepaal c i nie, maar die bewaringsmoontlikheid van hierdie spesie aan as being green in life, and the dorsolateral stripe golden. The l

r die suidelike punt van sy verspreidingsgebied, blyk tans adult females were shown by Parker to develop a quite e d uiters swak te wees. different pattern, consisting of a broader light canthal band n S.·Afr. Tydskr. Dierle. 1986,21: 149-152 u which extends over the upper eyelid to form a pair of in­

y complete 'spectacles'. Light markings are present over the a

w dorsal surface as spots to irregular bands. The ground colour e t in life was stated to be brown to purple. a G

In Natal, while the adult male patterning described by t e Parker may occasionally occur (e.g. Passmore & Carruthers n i

b 1979: 246), the great majority of males have a brown ground a

S colour, sometimes with darker speckling, and with a usually

y heavy dark bordering to the light dorsolateral bands (poynton b

d 1964 fIg. 109, Passmore & Carruthers: 246). This pattern is e J .C. Poynton c not described in Parker's Zambezi material, and it strongly u Department of Biological Sciences, University of Natal, d Durban, 4001 Republic of recalls the pattern of puncticuiotus. Females collected in Natal o r hardly ever show the 'typical' light dorsal ocelli or spots of p e Received 5 August 1985; accepted 3 October 1985 argus: at most there are only dorsolateral bands or spots. R 150 S.·Afr. Tydskr. Dierk. 1986. 21(2)

This southern material was consequenuy assigned U) puncticu­ while. Southwards. the spots become increasingly yellow, latus (poynton 1964). reduced in size and number. and south of about the Save Material subsequently collected mainly by Broadley in River (21°S) the leg spotting seems to disappear altogether. solUhern Mozambique and the eastern lowlands of It therefore seems reasonable to treat aU this material as showed variation which appeared to include both Zambezi a single species with a cline extending aU the way along the and Natal patterns. Some of this material (then housed in the eastern coastal lowlands. Even in Natal, where males tend to Umtali Museum) was sene to Dr. A. Schi4Jtz for examination. resemble p1Dlcticulatus, the material cannOt correctly be called Two males from Marhumbini were identified as punclicuklius puncticulatus because sexual dichromatism is shown, which (Sch.i¢tz 1975), while the remainder were ret urned with un­ is not a feature of puncticu/atus. This was overlooked by certainty. as to whether they were puncticuialus or argus. The Poynton (1964) and Scbi4ltz (1975). Moreover, Natal females, range of variation among the males of this material includes even when lacking Light spots as in pUncliadatus, have no the argus pattern described by Parker, through specimens in extension of the canthal uppet~yelid band behind the eyes, which the dark speckling becomes denser above and below which is a characteristic of puncticufl11us. This re-evaluation the light dorsolateral bands, producing at the extreme the therefore leaves puncliaikJtus Pfeffer as a species with known dark-bordered 'puncticukllus' pattern which predominales in southern limits in upland areas of , and not in Natal material. The range of variation in females also includes Mozambique or Natal. the argus pattern described by Parker, with dorsal ocelli. An opportunity, in 1984. to examine Cope's Natal Hy~ through the specimens in which the light markings behind the !ius material in the Academy of NatUra! Sciences, Phila­ head become reduced and confIDed to dorsolateral bands or delphia, made it apparent that tbe specimen presumed by spots. to specimens only w1th 'spectacles' (Figure I). The Loveridge (1941) to be the type 0 f cincriventris Cope (1862) oceUated variant becomes increasingly rare to the south. (ANSP 11322) is a specimen of argus Peters (1854). Although faded, this specimen from Umvoti shows a distinct pair of dark lines arising between the nostrils, then each line·diverging to run along the canthus to a dark patch on the upper eyelid. This is a characteristic feature of immature argus: the mark­ ings on the presumed cinctivenlris type in fact agree exactly with Parker's somewhat faded British M~eum material of immature maJes and females. As Loveridge (1941) noted. there are indications of a somewhat diffuse light stripe running above the dark canthal line and continuing to two thirds of

.

) the way down the body: this is sirriilar to the patterning of 0

1 the green Richards Bay 'specimen of CJTg}JS illustrated by 0 2

Passmore & Carruthm; (1979: 246). d

e Loveridge (1941) placed H. semidi.scus Hewin (1927) in the t a synonymy of cinclivenlris, but semidiscus lacks [he cinCli­ d (

venlris/argus dark canthal tine, and !.he cinctivenlris type lacks r e the very conspicuous light dorsolateral band of semidiscus. h s i Loveridge did bowever note that the afftn.ities of semidiscus l b appear to lie with argus and pwu:tiaJJotus, and the relationship u P

between semi.discus and argus in Niual desetVes further study. e

h Wager (1965) found the eggs ai1d ladpoles of theSe two species t

y to be indistinguishable, but the cans are conspicuously dif­ b ferent (passmore & Carruthers 1979). H argus occurs on the d e

t coastal lowlands as far south.aS Durban. while sernidiscus n occupies a belt commencing slightly inland. Two Natal a r

g Museum specimens from Eshowe (NM 5744), a female with

e markings and a male with heavy dark bordering c semidiscus n to the dorsolateral more of than of e band, typical argus. c i l semidiscus. provide evidence that the two species could be r

e sympa1ric; but more material is needed to establish whether d tlgUn! 1 Variation in specimens from a collection of Hyperolius orgus intergrading does or does not oroJr there. Eight Na1aI n from Doodo, near Beira. Mozambique (maJerial in the NaluraJ History u

Museum males collected. from Mount Edgecombe (NM

y Museum, Bulawayo, co\Jecu:d December 1969). Narural sile. Top: males

a 4117 -4118,4137 - 4139,4140-4142), just north of Durban, (UM 2208! & 21985). Botlom: females (UM 21993, 21997 & 21998). w e Specimens 2!985 and 21998 represent the predominant male and female mostly show the typical Natal argus pattern, but in one there t a patterns in N3.laI. is light spotting on the legs •. a feature of semidiscus. and in G

another the dorsolateral bands are narrow and have a very t e ftne dark edging, which recalls semidiscus markings. Hybridi­ n i b The Natal material may therefore be seen as the extreme zation is suggested by this series, but., as is recorded in the a

S in a cline, with the southem establishment of a heavily next section of this paper,· this population is now extinct. y bordered dorsolateral band the male and loss of post­ It is worth noting that lhe reduction of variaJ:i.on in southern b in

d cephalic markings in the female. TIlls tendency is consistent populations of argus, and the presence of semidimJs at the e c with the cline previously noted in materia] occuning nOI1h periphery of the argus range, appear to fit the 'cenLral­ u d of the Zambe:zi (poynwn 1964; Schi¢tz 1975): Kenyan females marginal model', recently reviewed by Brussard (1984). The o r show intense spotting of the back and also the legs. These model holds that populations on the margin of a range are p e light markings are made even more conspicuous by being isolated, sparse, chromosomally monomorphic, and playa R s. Afr. J. Zoo!. 1986, 21 (2) 151

~ leading role in speciation events (Brussard 1984). The taxo­ males also call from lily pads; in fact antiphonal calling nomic situation regarding argus and the peripheral semidiscus between argus and pusillus males is not uncommon. Yet the seems worth studying in the light of this model. A particular small-sized pusil/us seems biogeographically more robust than ecological aspect of the model will be taken up in the next argus, judging from the much wider geographical range of section of this paper. pusillus both westwards and southwards, so its presence in a pond need not be an indicator that argus necessarily at one rhe status of H. argus in the Durban area time occurred there as well. Beyond this consideration, Brussard (1984) recently pointed out that the periphery of the though, the vulnerability of argus populations compared with geographical range of a species need not necessarily be those of pusillus or any other species is unknown. And the marginal with respect to ecological circumstances. It is, opportunity for the study particularly of any range-periphery nevertheless, generally assumed to be the rule that environ­ effects in argus at the southern extreme of its range now seems mental stress is maximum at the periphery of a range, to be lost. especially the range of a widespread species, and the distribu­ A disaster of this kind shows the necessity of coming to ( tion of H. argus is probably no exception. Habitat disturbance grips with an area of study that the zoological community accompanying human settlement could be expected in most has tended to treat with some aloofness, namely urban bio­ cases to compound the environmental stress at the periphery geography. Habitat destruction accompanying urban develop­ of a range. The results to date of a survey conducted to ment tends to proceed with uoimagined speed, yet there is investigate the situation regarding argus are, however, even evidence that much can still be saved by effective contact more negative than had been anticipated. between biologists and civic authorities (poyoton & Roberts The floating leaves of water lilies (Nymphaea) form the 1985). It is hoped that a herpetological survey of Durban, calling and retreating site of males. This lily-pond habitat is being carried out by the Department of Biological Sciences considerably at risk in any settled area. The southernmost of Natal University with much cooperation from the Du,rban known population of argus occurs in a typical lily pond in Corporation (poyoton 1985), will help to avert further the Stainbank Nature Reserve in southern Durban (29° 53'S), disasters of the H. argus type. In Durban the conservation controlled by the Natal Parks Board. No other populations situation is particularly critical owing to the city's location in are known in Durban. A population sampled in 1958 at the transition zone that exists between tropical and non­ Mount Edgecombe, just north of Durban (29°42'30"S), has tropical groups of and plants (poyoton 1961). disappeared, evidently owing to habitat destruction. This was the population showing features suggesting intergrading with Acknowledgements H. semidiscus, discussed earlier. In 1979, a large population Taxonomic work contributing to this study was done mainly

.

) of typical argus was flourishing 0°05' north of Mount Edge­ in the Umtali Museum (material now in the Natural History 0

1 combe in a lily pond in the grounds of a seaside estate, Museum, Bulawayo), the Natal Museum, Pietermaritzburg, 0

2 Shalimar Gardens; but following new ownership the ponds the British Museum (Natural History), London, and the

d in the estate have become overgroWn and filled in, and the Academy of Natural Sciences, Philadelphia. I am grateful for e t

a argus population is now probably at extinction point. The the provision of full facilities by these institutions, and for d

( funding by the University of Natal and the C.S.I.R. Fieldwork species has been searched for in the Tongaat-Compensation r

e area (as far north as 29°31'S), so far without success. reported here was carried out largely in assoc;iation with h s

i Water lilies occur in several ponds in the Durban-Com­ student projects undertaken by Mrs R. Royal and Mr M. l

b pensation area, apparently providing suitable habitat for argus. Burdon, whose diligent application to the work is valued. The u

P The absence of argus from such lily ponds is difficult to photographs were prepared by Mr D.M. Dlamini.

e interpret on account of the possibility of previous habitat h t disturbance in presently normal-looking ponds. An instructive References y

b BRUSSARD, P.F. 1984. Geographic patterns and environmental example of such a possibility is given by a pair of interlinked

d gradients: the central-marginal model in Drosophila revisited. e dams near Compensation (farm Redwoods). The dams at one t Ann. Rev. &01. Syst. 15: 25-64. . n time supported the southernmost recorded population of the

a COPE, E.D. 1862. Notes upon some reptiles of the Old World. r widespread East Mrican hyperoliid, Kassina macu/ata (Wager g Proc. Acad. nat. Sci. Phi/ad. 1862: 337 - 344.

e 1965). K. maculata also favours ponds with water lilies, and COTT, H.B. 1932. The Zoological Society's expedition to the c

n lilies are indeed currently present in the dams, but not K. Zambesi, 1927: no. 4. Proc. zool. Soc. Lond. 1932: 471 - 541. e

c HEWITT, J. 1927. Further descriptions of reptiles and batrachians i maculata, or H. argus. Investigation showed that some 25 l from South Africa. Rec. Albany Mus. 3: 371-415. r years ago, red-breasted bream (TIlapia melanopleura) was e JUBB, R.A. 1967. Freshwater fishes of southern Africa. Balkema, d introduced into the two dams. These fish almost totally

n Cape Town, South Africa.

u destroyed the water lilies, which led, among other things, to LANZA, B. 1978. On some new or interesting east African y the disappearance of jacarias (Actophilornis africanus). To a and reptiles. Monitore zool. ital. (N.S.) Suppl. 10:

w control the bream, the upper dam was poisoned and the lower 229-297. e t LOVERIDGE, A. 1941. South African of the genus

a dam drained. This treatment did not wholly eliminate the Hyperolius in the Museum of Comparative Zoology, G

bream, so bass (Micropterus salmoides) were introduced. t Cambridge, Massachusetts. Ann. Transv. Mus. 20: 283 - 291. e Being an 'avid predator' (Jubb 1967: 187), it is likely that the

n PARKER, H.W. 1930. A collection of frogs from Portuguese East i

b bass destroyed tadpoles of any K. maculata that had survived Africa. Proc. zool. Soc. Lond. 1930: 897 - 905. a

S the first treatment. The bass have however kept the bream PASSMORE, N.I. & CARRUTHERS, V.C. 1979. South African

y in check, so the aquatic vegetation is returning to what appears frogs. University of the Witwatersrand. Johannesburg, South b Africa.

d to be, at first sight, an undisturbed state.

e PETERS, W. 1854. Uebersicht der auf seiner Reise gesammelten c It may not safely be assumed that any body of water in

u Amphibien. Mber. k. Akad. Wiss. 1854: 614-629.

d the Durban area is undisturbed, even if there is presently a

o PFEFFER, G. 1893. Ostafrikanische Reptilien und Amphibien, r good stand of water lilies. Several such ponds exist, and most gesammelt van Herrn F. Stuhlmann im Jahre 1888 und

p Dr e carry populations of , a species whose 1889. Jb. hamb. wiss. Anst. 10: 71-105. l R S.-Afr. Tydskr. Dierk. 1986, 21(2) 152 -,. POYNTON, J.C. 1961. Biogeography of south-east Africa. Nature POYNTON, J.C. & ROBERTS, D.C. 1985. Urban open space 189: SOl - S03. planning in South Africa: a biogeographical perspective. S. Afr. POYNTON, J.C. 1964. The amphibia of southern Africa: a J. Sci. 81: 33-37. faunal study. Ann. Natal Mus. 17: 1-334. SCHHITZ, A. 1975. The treefrogs of eastern Africa. Steenstrupia, POYNTON, J.C. 1985. Durban's amphibians are interesting and Copenhagen, Denmark. varied. Parks, Recreation and Beaches Department of the WAGER, V.A. 1965. The frogs of South Africa. Purnell & Sons, Durban Municipality, Durban, South Africa. Cape Town, South Africa.

. ) 0 1 0 2 d e t a d ( r e h s i l b u P e h t y b d e t n a r g e c n e c i l r e d n u y a w e t a G t e n i b a S y b d e c u d o r p e R