Journal of the American Control Association, 14(3):273-276,1998 Copyright O 1998 by the American Mosquito Control Association, Inc.

AEDES (DICEROMYIA) FURCIFER (EDWARDS) AND AEDES (DICEROMYIA) CORDELLIERI HUANG IN SOUTHERN AFRICA: DISTRIBUTION AND MORPHOLOGICAL DIFFERENTIATION

P G. JUPP

Special Pathogens lJnit, National Institute for Virology, Department of Virology, University of the Witwatersrand, Private Bag X4, Sandingham 2131,

ABSTRACT. Two species in the Aedes (Diceromyia) furciftr complex, Ae. furciftr s.s. and Ae. cordellieri, occur in southern Africa. They occur either allopatrically or sympatrically in lowland wooded savanna in tropical or subtropical regions in the Northern Province, Mpumalanga Province, and northern KwaZulu-Natal in South Africa and in Zmbabwe. A map of the distribution of these mosquitoes is presented including 2 new locality records where Ae. cordellieri is allopatric. A morphological study showed that immatures and adult females of the species are indistinguishable. This included a detailed study of the female terga. Differences in the male gonocoxite remain the only means to distinguish the 2 taxa.

KEY WORDS: Aedes furcifer, Aedes cordellieri, distribution, southem Africa, morphological differentiation

INTRODUCTION ed if females of the 2 speciescould be identified. Jupp et al. (1993) reported initial observations on In a preliminary publication, Jupp et al. (1993) morphology for females of the 2 species.The pres- reported that a reexamination of all the male spec- ent paper reports the results of a more detailed imens of the Aedes furcifer group housed in the study designed to determine whether females and museum collection of the National Institute for Vi- immatures can be identified. A map is also pre- rology (University of the Witwatersrand,Sandring- sentedshowing the known distribution of the 2 spe- ham, South Africa) indicated the occurrence of only cies. Ae. furcifer s.s. (Edwards) and Aedes cordellieri Huang in southern Africa. Whether Aedes taylori Edwards, the 3rd member of the group, is com- MATERIALS AND METHODS pletely absent from the region or not remains to be As mentioned in the earlier paper (Jupp et al. seen. It is quite possible that mosquito collecting 1993), unless population densitiesof these mosqui- over a wider area could reveal its presence n Zim- toes are high, bait traps and biting catches fail to babwe and/or northern Mozambique. In the same sample them so the only practical method remain- paper, the localities at which either or both Ae. fur- ing is to collect eggs depositedin bamboo pots ex- cifer and Ae, cordellieri occurred were listed and posed in wooded areasas ovitraps. Such pots were the importance of the Ae. furcifer group as arboviral used at Pafuri, Shingwedzi, and Mica (Northern vectors was emphasized. The group includes vec- Province), at Skukuza (Mpumalanga Province), and tors of dengue, , and vi- at Ndumu (KwaZulu-Natal) during at least 5 sum- ruses in West Africa and chikungunya virus in mers. The aedine eggs deposited in each pot were southern Africa. Each of the 2 species has been reared and the resulting adults allowed to emerge encountered either allopatrically or sympatrically in into a different cage representing each pot. The the tropical and subtropical wooded savanna of the genitalia of each Ae. furcifer group male emerging Northern Province and Mpumalanga Province (pre- were checked and the long golden setal tuft on the viously called Transvaal),northern KwaZulu-Natal, apex of the gonocoxite used to identify male Ae. and the Zimbabwe lowlands. Collecting remains to furctfer; this tuft is entirely absentin male Ae. cor- be done in Mozambique, but at least Ae. furciftr dellieri. The adult mosquitoeswere then allowed to and Ae. cordellieri are expected to be found there. engorge on hamsters and separate families subse- Observationson the ecology of Ae. furcifer have quently were reared from individual gravid females. been made in South Africa. It is a tree hole breeder For each family, slide mounts were prepared of the that feeds on nonhuman primates (vervet monkeys genitalia from several males to confirm their iden- and baboons) at ground level but to a greater extent tity according to gonocoxite morphology (Huang in the tree canopy or high rocky outcrops (koppies) 1986). In some of thesefamilies, siblings were pre- (Mclntosh et al. 1977). This speciesalso feeds on served for detailed taxonomic study; adult males when they venture into this habitat. Aedes and females were pinned and larval and pupal ex- cordellieri appearsto have a similar ecology. Such uviae were preservedin an ethanol-glycerine mix- ecological studies as well as the collection of fe- ture. Many families were thus identified from each male mosquitoes for vector competence experi- locality but no Ae. taylori specimens were encoun- ments with the 3 viruses would be greatly facilitat- tered; only Ae. furciftr andAe. cordelliei were col-

zt3 274 Jounrver or tss AvsnrcaN Mosquno CoNrnol AssocrerroN Vol. 14, No.3

in the search for possible differences between the 2 species.A few adult femaleAe. furcifur andAe. cordellieri were also examined with the scanning electron microscope at 7O-:7Wx magnification to :basomedian investigate whether any structural differences were white band detectablein the antennaeand maxillary palps. :basomedian whiteDatch RESULTS

white basolateral Distribution patch The distribution already published for southern yellow apicolateral Africa (Jupp et al. 1993) can now be updated with patch 2 additional localities where Ae. cordellieri is ap- : scatteredyellorv parently allopatric, namely Pafuri (22'27'5, dorsomedianscales 3l'2L'F.) and Shingwedzi (23'5O'S, 31'26'E). These are both located in the Northern Province in the northern part of the Kruger National Park and only Ae. cordellieri was collected there during 5 successive summers. Figure 2 maps the latest known distribution of both species. Fig. 1. Diagrammatic representationof abdominal ter- ga showing charactersA to D examined on female mos- quitoes. Morphology Most of the principal features on the adult female lected and identifled. The presence of both species mosquitoes were examined critically to see whether in a pot was unusual during the particular season a way of differentiating the 2 species could be when the families were being reared and preserved found. These included scutal ornamentation, max- for this morphological study. illary palps and antennae, vertices, wing venation, Figure 1 shows the 4 characters examined on basal white bands on tarsomeres, and tarsal claws each tergum I-VII. A character was regarded as on all legs. Light microscopic examination revealed present even if the character was only faintly visi- that these characters were unsuitable for differen- ble as a smaller number of scales. The terga were tiation, so the antennae and maxillary palps were examined on female specimens that had been examined further with the scanning electron micro- pinned laterally to protect the dorsal scale patterns. scope, also without detecting any differences. Be- Other parts of these specimens were also examined cause Huang (1986) had found she could distin-

ffi rroolet

a Ae. furcifer O Ae. @rdellied O Both species ..r.. 180mean midwints(July) isotherm

Fig. 2. Map showing known distributions of Aedes furcifer and Aedes cordellieri Arntce' 275 SepreN{ept 1998 AE. FURCIFERLNo Aa CqRDEr'I'IERIIN SourssnN

Ae. cordelliei Ae. furciter and shows a full analysis of the 4 characters on II-V[ in 8 siblings of Ae. cordellieri from SIBS SIBS terga Shingwedzi and 8 siblings of Ae. furcifer ftom 12345678 12345678 Ndumu. A character on a particular tergum varied lx x x x x x x xl A by being both absent and present within the sibling ll lxxxxxxxxl B presentor absent x x x r . r .l group or when it was consistently l. c 1...... 1 n in one species,either the same applied to the other speciesor there was variation in the other species. The absenceor presence of each character on terga lxxxxxxxxl A l...xx..xl II-VII for the whole sample of Ae. furcifer andAe. ,,,lxxxxxxxxl B lx.xxxxxxl cordellieri, respectively,is given in Table 1, which "'lrul shows the numbers and percentagesof mosquitoes 3 li:::::l:ldisplaying the character in each sample. Consider- able similarity occrrred between the 2 species and A |ffil none of the characters,not even that on a single B lx x x x x x x xl tergum, would serve to separatethe species'Simi- x . x x x x .l c lx larly, the immature stages could not be distin- D guished. At the cortmencement of the study, the number of denticles on the larval pecten and the lffirl l.-ffi':l number of branches on head seta 9 and the dimen- B lx x x x x x x xl sions of the trumpet in the pupa looked promising. V c x . x x x x xl lx However, examination of a series of mounted im- 1..I:IIIIIIil x . . ..1 D mature exuviae showed that overlap occurred in both larvae and pupae of both species,rendering A tffi] these charactersunsuitable for diagnosis. B lx x x x x x x xl c D lxxxxxxxxl'''' ' 'l l'' DISCUSSION The map of the known distributions of the 2 taxa A (Fig. 2) shows that each species occurs either al- B ; ; lopatrically or sympatrically. The map also indi- c l;i : : ; ;l cates that Ae. cordellieri usually occurs within or D close to the tropical region (east of the 18"C mean Fig. 3. Variation of the 4 different morphological midwinter isotherm), whereasAe. furcifer usually characters (A-D) on terga II-VII in a sibling group of 8 occurs in the subtropical regions (west of this iso- Aedes cordellieri and a sibling group of I Aedes furcifer, therm). In Fig. 2, the boundary of the tropical re- respectively. gion in South Africa and Zimbabwe can be seento follow the 18'C mean midwinter month (July) sur- face isotherm, which separatesit from the adjacent guish the speciesfrom tergal characters,these char- subtropical region (Poynton 1964, Mclntosh 1980). acters were given close attention under the light Therefore, temperature probably limits the distri- microscope, particularly as the lst observations bution patterns of the 2 taxa. Further distributional looked promising. Unfortunately, considerable in- records in southern Mozambique across the South traspecific variation was found in tergal morphol- African border and in Zimbabwe .ue needed to con- ogy within a family. Figure 3 is an example of this, firm this pattern.

Table 1. An analysis of the number and percentage (in parentheses) of female mosquitoes displaying each character (A-D) on terga II-VII, respectively.

Aedesfurciferl Aedes cordellieri2 Tergum C il 73 (99) 7r (96) 55 (74) 4 (s) 44 (98) 44 (98) 34 (76) 0 III 24 (32) 68 (e2) 62 (84) s (7) 42 (e3) 33 (73) 0 IV 4l (5s) 70 (es) 6s (88) s (7) 33 (73) 4r (er) 37 (82) | (2) 34 (46) 69 (e3) 66 (8e) 13(r8) 30 (67) 37 (82) 36 (80) 8 (r8) VI 34 (46) 65 (88) 66 (8e) 24 (32) 30 (67) 35 (78) 33 (74) e (20) VII 29 (4O)3 6 (8). 58 (80)r 30 (41) 27 (60) 23 (sr) 23(sr) 5 (11)

t Seventy-fou specimens from 9 families were examined. 'Forty-five specimens from 12 families were examined. 3 Tergum VII was damaged in one specimen, hence these percentages are based on 73 specimens. 276 JoURNAL oF THE AMERTCIN Mosqurro Colmnol AssocrATroN Vor-. 14, No. 3

The absence of morphological distinctness be_ may possibly be used to differentiate females of the tween adult females of Ae. furcifer andAe. cordel_ 2 taxa. This would be of some practical value in lieri makesecological observations on the 2 species the research program, although ttre disadvantage of difficult to cary out, as all identifications must be this method is having to kill or damage the iros- confirmed from the male genitalia, unless repeated quito in order to identify it. collections over an extended period show th- pres_ ence of only one species. This has happenedin the ACKNOWLEDGMENTS case of Pafuri and Shingwedzi in the Northern I wish to thank Province where only Ae. cordellieri has been taken the Directors of the National Parks Board and the KwaZulu in ovitraps over 5 successivesummers. However, if Bureau of Natural Resourcesfor allowing me female specimens are collected for vector compe- to collect mosquitoesin the Kruger National Park and Ndumu tence experiments from Skukuza, Mica. and Ndu- Game Re- serve, respectively. Special thanks are due to mu, where both species occur side by side, the Leo Braak, Scott Ronaldson, and Louis Olivier of the progeny from each individual female must be Kruger Park for their kind assistance. My col- reared and a male identified before live specimens league, Alan Kemp, helped with the collection and can be pooled to assemblea homogenouscollection rearing of mosquitoes, for which I am grateful. I of one species for an experiment or for starting a am also indebted to the Board of the Poliomyelitis laboratory colony. An outside chance exists that Research Foundation for a grant towards the pro- some of the Ae. furctfer andAe. cordellieri at these ject. 3 localities may have cross-mated when first reared from ovitraps in the laboratory. However, this is REF.ERENCES CITED considered very unlikely because of the rarity of Huang, Y. M. 1986. (Diceromyia) Ae. cordellierl during the particular summer when Notes on the Aedes group with a description of a new species (Dip- specimensof Ae. were furcifer furcifer extracted from the tera: Culicidae). Proc. Entomol. Soc. Wash. 88:634- pots for the present taxonomic study. No Ae. cor- 649. dellieri were identified from Ndumu and only 2 and Jupp, P G., B. M. Mclntosh and A. Kemp. 1993. The 8Vo were identified as such from Skukuza and Aedes (Diceromyia) furcifer group (Diptera: Culicidae) Mica, respectively. Five of the 9 Ae. furcifer farn- in southern Africa: identification and occurrence ofAe. ilies studied came from Ndumu, 3 came from Sku- furcder (Edwards) and Ae. cordellieri Huang. Afr. En- tomol. 1:126-128. kuza, and only 1 came from Mica where the highest Mclntosh, B. M. 1980. The epidemiology proportion of - of A. cordellierd occurred. Furthermore, borne viruses in southern Africa. D.Sc. thesis. Univer- male genitalic structure of each species has re- sity of Pretoria, Pretoria, South Africa. mained constant over the 30-year period that the Mclntosh, B. M., P G. Jupp and I. Dos Santos. 1977. Ae. furcifer group has been studied in my labora- Rural epidemic of chikungunya in South Africa with tory and no suggestion of intermediates has oc- involvement of Aedes (Diceromyia) furcifur (Bdwards) curred. Either the long curled golden-yellow gon- and baboons. S. Afr. J. Sci.73:267--269. Poynton, I. C. 1964. The present (in biotic divisions of southern Af- ocoxite setal tuft has been Ae. furctfur) rica, as shown by the amphibia, pp. 2O6-218. In: D. H. or it has been completely absent(inAe. cordellieri). S. Davis (ed.). Ecological studies in southern Africa. Application of ribosomal DNA sequenceanalysis Junk, The Hague, The Netherlands.