Interaction between the predator Toxorhynchites brevipalpis and its prey Aedes aegypti1

Bull. Org. mond. Santh 1973, 49, 359-365 Bull. Wld Hith Org. Interaction between the predator Toxorhynchites brevipalpis and its prey Aedes aegypti' MILAN TRPIS 2 In a circumscribed area in Tanzania where the predacious larvae of Toxorhynchites brevipalpis were particularly abundant, it was found that water-filled tires and tins- containing Toxorhynchites larvae hadfewer larvae ofAedes aegypti than those without the predator larvae. The peaks of infestation with Toxorhynchites larvae occurred almost a month later than the peaks ofA. aegypti infestation. Cannibalism was observed among the predator larvae in these containers. Measurement of the effect of predation by the predacious larvae of the mosquito Toxorhynchites on the larvae of Aedes aegypti is important for the understanding not only of the population dynamics of this yellow-fever vector but also of the results to be obtained when species of Toxorhynchites are employed as biological control agents. Predator-prey relationships in entomology may be studied by the use of mathematical models (3, 6), but their charac- terization in the field is much more difficult. Certain components of the predator-prey system comprising Toxorhynchites and A. aegypti have already been studied (7, 1, 2). The programme of the WHO East African Aedes Research Unit in Tanzania offered an opportunity to study in detail the predator-prey relationships of one important species of Toxorhyn- chites. Toxorhynchites brevipalpis is a treehole-breeding mosquito widely distributed in the tropical zone of Africa. Its larvae occur commonly in the axils of plants such as Strelitzia and Dracaena in some regions of South Africa (7) but not in East Africa. Breeding in man-made containers has also been recorded in the past (5). At present breeding in treeholes is still more extensive than breeding in artificial containers, although the author has found larvae in coconut shells and in Achatina snail shells very occasionally. The occurrence of T. brevipalpis in unusually high numbers in the Buguruni automobile 1 This study was supported by a Public Health Service research grant from the Center for Disease Control, Atlanta, Ga., USA, and by the World Health Organization. ' Entomologist/Ecologist, WHO East Africa Aedes Re- search Unit, Dar es Salaam, Tanzania. Present address: Department of Biology, University of Notre Dame, Notre Dame, Ind. 46556, USA. dump located in one of the Dar es Salaam suburbs, where the larvae of this species were breeding mostly in tires, provided a unique opportunity for study. This one-hectare habitat contained about 500 wrecked automobiles, 3 000 discarded motor tires, 190 tins, 160 coconut shells, and 380 Achatina fulica snail shells. The number of tires increased from 2 500 in April 1969 (when this study began) to 3 240 in August 1969 and to 4 000 in May 1970. The purpose of the present investigation was to demonstrate the predator-prey relationship of Toxo- rhynchites and to follow the oscillations in relative numbers of predator and prey. In the laboratory, the author has found that one T. brevipalpis larva destroys an average of 154 A. aegypti larvae when reared at 26°C, and as many as 359 A. aegypti larvae at 32°C (10). In the field, the effect of predation of T. brevipalpis on A. aegypti can be found (a) indirect- ly, from the frequency distribution of the numbers of prey larvae per container in the presence and absence of the predator larvae, or (b) directly, by a field experiment with known numbers of predator and prey larvae in marked breeding sites. METHODS Nine tires grouped according to size (3 small and capable of holding 3 litres of water, 3 medium-sized holding 6 litres, and 3 large holding 9 litres of water) were sampled weekly. A 15-mm hole was made in the centre of the tires for transferring all the T. brevipal- pis and A. aegypti larvae into a large (500 x 300 x 100 mm) plastic pan, for transport to the laboratory for identification and counting. The physical condi- tion of the tires (presence or absence of water) was recorded. Nine tins were also sampled, similarly 3127 - 359 M. TRPJS divided into groups of three, the large tins having a capacity of 2.5 litres, the medium-sized 750 ml, and the small 250 ml. The figure for the population density of the larvae of both predator and prey per one hectare automobile dump was derived from the mean number of larvae per container and the per- centage of containers with water. The data for rainfall were obtained from the Meteorological Station at Dar es Salaam Airport (about 5 km from the study area) from April 1969 to December 1969. In January 1970 a rain-gauge was placed in the study area and checked daily by the members of the East Africa Aedes Research Unit until the end of the study. RESULTS Effect of Toxorhynchites on Aedes aegypti Of the total of 381 assessments in which water was present in the tires sampled weekly, 159 showed that A. aegypti was present alone, and 222 that Toxorhyn- chites was also present (Table 1). When the assess- ments are grouped into infestation size-classes ac- cording to the number of A. aegypti found present, it is seen that the presence of Toxorhynchites altered the frequency distribution in the direction of the smaller size-classes. In the presence of the predator, no infestation exceeded 50 per tire, whereas 3.4% of samples exceeded 50 in its absence. The frequency of the class where no A. aegypti was present increased from 47% in the absence of Toxorhynchites to 83% in its presence. To demonstrate its predatory activity in the field, ten 4th-instar larvae of T. brevipalpis were left in each Table 1. Distribution of infestation size (number of larvae). Frequency of occurrence of larvae of Aedes aegypti in tires in the absence and in the presence of the predator Toxorhynchites brevipalpis larvae Number In absence of In presence of of A. aegypti predator predator larvae per tire Number % Number % 0 75 47.2 184 83.0 1-10 51 32.1 27 12.1 11-20 16 10.1 8 3.6 21-50 10 6.3 3 1.3 51-100 6 3.1 0 0.0 101-300 2 1.2 0 0.0 Total 159 100 222 100 medium-sized tire colonized by this species; then 200 4th-instar larvae of A. aegypti were introduced into the tires. The number of prey consumed was recorded and added daily. Three experiments were conducted simultaneously and the observation was continued for 7 days. During 24 hours, the water temperature varied between 22°C and 25°C. It was found that one 4th-instar larva of T. brevipalpis living in tires con- sumes on average 12.3 A. aegypti larvae per 24 hours. In laboratory experiments, it has been shown that one 4th-instar predator larva consumes on average 16 prey larvae per 24 hours at a temperature of 26°C (10). Intra-species effects and size ofbreeding sites In Toxorhynchites, as a treehole-breeding mos- quito, intra-species competition for food and space results in cannibalism. In large breeding sites such as tires containing leaves, tree branches, debris, etc., its larvae can find shelter that to some extent protects against cannibalism. Thus, a container as large as a tire can shelter as many as 22 larvae (Table 2). On the other hand, the maximum density of predator larvae found in tins was 10, and this density occurred with a frequency of only 1.8 %. Tins are smaller and offer less shelter than tires. One larva per container was found most frequently in both tires and tins. The survival of only one predator larva is most probably the result of cannibalism, because every time that 10-50 lst-instar predator larvae were placed in 100-250-ml beakers without prey, only one larva in the 3rd or 4th instar was found after 5-7 days. We have found that T. brevipalpis females readily lay their eggs in black jars in the laboratory. It was supposed that these containers could be used for detecting seasonal variation in the relative density of a larval population in both suburban and feral biotopes. Ten black jars (ovitraps) 80 mm in dia- meter and 125 mm deep were placed in the Buguruni and Pugu forest biotopes, and filled three-quarters full with water. They were examined weekly for mosquito larvae over a period of one year. Almost 72% of the ovitraps in the Buguruni biotope and 85% in Pugu forest were found to be negative for Toxorhynchites larvae (Table 2). This might indicate that there are perhaps more attractive sites for oviposition than black jars, such as motor tires in Buguruni or treeholes in the Pugu forest. The rela- tively low number of jars could hardly compete with thousands of tires or hundreds of treeholes. A single larva per jar was the most frequent finding in both these biotopes. However, the black jars were 360 T. BREVIPALPIS AND A. AEGYPTI Table 2. Frequency distribution of infestation size (number of larvae) of Toxorhynchites brevipalpis in various containers Infestation Tires Tins Oviposition jars size Buguruni Pugu Forest(no. per container) No. % No. % No. % No. % 0 99 28.8 137 83.0 377 73.9 170 85.4 1 104 30.4 13 7.9 101 19.8 21 10.6 2 59 17.2 8 4.8 23 4.5 5 2.5 3 35 10.2 4 2.5 3 0.6 3 1.5 4-5 24 7.0 3 1.8 6 1.2 - - 6-10 18 5.2 - - - - 11-15 3 0.9 - _ _ _ _ - 16-22 1 0.3 - - - - - - Total 343 100.0 165 100.0 510 100.0 199 100.0 not suitable for indicating the seasonal variation in density of a larval population of T. brevipalpis, and no distinct oscillations in predator-prey density could be detected with them. The small size of the jars and the absence of shelter within facilitated cannibalism. Only in the periods when the prey and predator density was very high did more than one larva occur per jar. Predator-prey oscillations between Toxorhynchites and A. aegypti The monthly averages for the weekly samples of predator and prey (Table 3) show that the density of the prey was high in April 1969, when these studies began. Four peaks and three depressions in density of prey were recorded during the period of one year extending from April 1969 to March 1970. As has been shown (9), the density of A. aegypti in this area depends on the pattern of rainfall. The main rainy season in Dar es Salaam occurs in April and early May. From mid-May to the end of August there is a dry cool season, and the prey population was found to decrease steadily from April to August (Fig. 1). From September to November, when the short rains occurred, the prey population increased again, only to fall in December. The third peak occurred in January, during a dry hot season punc- tuated by occasional short showers. The decrease in rainfall in February caused a decrease in the prey density. Around the middle of March the new main rainy season started, having the effect of increasing the numbers of prey. The density of the predator followed the pattern of the prey numbers. Larvae of T. brevipalpis showed a direct numerical relationship with the density of A. aegypti. However, the effect was not immediate; there was always a time lapse between the emergence of the predator and its active predation. This period includes the pupal stage, the time until the females are able to lay eggs, and the duration of embryonic development until the hatching of new larvae. The interaction between predator and prey can be demonstrated, e.g., in the oscillation during the period from August to December 1969, when the prey density rose from zero (in August) to 11 708 in October, 2 347 in November, and 272 in December (Fig. 2, A). Neither the larval nor the adult popula- tion of the predator fell to zero after a long dry period. However, the predator is not present in large enough numbers to control the prey at the beginning of the rainy season. Eggs of A. aegypti laid from the end of the main rainy season until the beginning of the short rainy season remain dormant in breeding sites (9). The hatching of A. aegypti at the onset of heavy rain causes a sudden eruption of the popula- tion which the predator is not able to reduce because of its low density (Fig. 2, A). In other invertebrate predator-prey systems, the predators die off rapidly when their prey is depleted. The Toxorhynchites predator-prey system is different in this respect. When the prey is depleted the Toxorhynchites 4th-instar larvae are capable of starv- ing for 6 months or longer. The drop in population of the predator occurs when there is a shortage of 361 M. TRPIS I~ *0. - 0.Oo~ ." FCl 00.c emm > C' )*- 06 M > 'C' 21 06> 0.6L 0.'u 06 0. Z 0 CL 0 0 C- N 0, 0 9 0 0 0 Co 0) rC M(00) co Co (0 N I- r- c0 C-i l 00 Lo r- (0 CD _ 0 0 lq(o CD o LO r- _ r- N 0 00 0 N- Co C- (0N1 Co C- co 0 N (0 to co (0 co Co 'L6 (0 M- 9 (0 r- rs C-(0 CD I- Co) Co(0 co Co4 (0 6 N- CD - - Coi Co4 (I Co CN 0 LO. Co 0. 0 0 0 0 0Co 0 10 CN C- CD Co 0 I- 0 6 0 0 0 0 co (0 0 r- 6 N 0 coCo 0 C 0(0o0 o N Co N ND CD Lo Co 0 .0 co C) 0 co (0 CNO. 0 C- 0 Co 'Li Co4 N C- Co Co Co (0 M 0 co co S- X o) .0 ,_ >- a >- 0 a °.. 0 0 0 a)(0 CY) 362 Co) 0 Cr- Co4 0 Co LO co (0 CN Lo M N- 91 C- C-) M Co Coto 0 M 0,a4 0, a) r- Co04 (0 Lo 04 C- I- I- Co le LO N co co 0 (0 (0 N1 M M) 2 0 Co (3 0 k. 0 CD 0) - Ca 0 C) 0.o co co c0 CLO 0 E Co-o Co C0. 0 .°E 0. *0 ) Co C0 C *_ -- U.4 ._ E 4.'i Co D,0)E Cno~ ._ COOC o Co o0 @ ._ ) C~ ._ _. Cox CoC- C- 0 M 0 Co N r- Co 0 CN (o 0) C-(0 0 C:n 0 Co co co co 0, (0 Co C 0 Co Co N co w- co (0 Co Co Co- M Co N co w '- co Coi v- M (0 M (o 1- w- co - co co a) Co C- a) CNCo 0 0 Co Co Co Co n cn .0M 0 0 Z co 0 Co 0 rv awC-4 CD co N T- (0 .0 E .00 0 a o ] CO (0 Ntc CO w N oo co r Co C O. N NDN N C CD 0 N o N Co N N o C co Co C- N N Co CN co C1 co co N M -° 04 C' D ~ oL 0C- a) T. BREVIPALPIS AND A. AEGYPTI 363 April May June July August Sept. Oct. Nov. Dec. Jan. Febr March Fig. 1. Oscillations in density showing the predator-prey relationship between mosquito larvae of Toxorhynchites brevipalpis (predator) and Aedes aegypti (prey) in a suburban habitat in Dar es Salaam. *12,000 L ~~~Preyao % * 8.000 Predtor CL 4000 '6 30 I* . 70 10 b ~ ~AS 0 N. D E 0 c -o 40 10203 1020 1020 01020 1020 JU1LY AUGUtST SEPTEMBlER OCTOBER NOVEMBERt Fig. 2. Growth of the larval population of Aedes aegypti and Toxorhynchites brevipalpis during the short rainy season in East Africa. A. One oscillation in the population of prey (A. aegypti) and predator (T. brevipalpis). B. Change in the composition of larval instars of the predator during one oscillation. C. Rainfall of the short rainy season in the coastal zone of Tanzania (Dar es Salaam) in 1969. M. TRPIS A. aegypti. The 4th-instar larvae of T. brevipalpis start eating the earlier larval instars of their own species (10). When the larval density of T. brevipalpis reaches its peak there is a high population of adults in the area. There is no shortage of predator eggs but all freshly hatched larvae are eaten by higher instars. Density of the lst- and 2nd-instar larvae of the predator increased after prey population increased (Fig. 2, B). Apparently the higher instars of predator larvae eat the prey larvae. On the descending part of the prey-predator curves cannibalism occurs more intensively (Fig. 2, A). The data presented in this paper show that the oscillations in the population densities of both prey and predator in the field are such that the reduction in A. aegypti larvae soon after the rains is accom- panied by an increase in the production of Toxorhyn- chites larvae. This undoubtedly contributes to the suppression of the prey almost to zero level. How- ever, the presence of water in some tires even during the dryest period of the year in the heavily shaded Buguruni biotope prevents extinction of the predator population. At the same time the prey population is mostly in the egg stage, comprising dormant em- bryos. At the end of the dry season T. brevipalpis (whose embryos are not able to resist desiccation) is at a low density level and cannot control the prey population when it erupts at the first heavy rain. Since the predator's life-cycle is almost 3 times as long as that of the prey, the predator begins to be effective as a control agent only when the prey is already at all developmental stages. This delay in control can be eliminated by the release of adult T. brevipalpis at the end of the dry season, as the females must be at least 6 days old to produce eggs in time for the first rains. Since T. brevipalpis has been colonized (11) and its mass production developed (Dr E. J. Gerberg, personal communication), it would be possible to effect such a release. Among mosquito predators the larvivorous fish Gambusia affinis has been studied most, and probably not enough attention has been given to the study of T. brevipalpis. Considering the quantity of prey consumed, it is at least as effective in the control of mosquito larvae as Gambusia affinis. Naturally the application of Toxorhynchites larvae is limited to a certain group of biotopes, as Gambusia is limited to others. However, at present Toxorhynchites larvae seem to be good potential predators that can be used for the biological control of some disease-carrying mosquitos in thick rain-forests or in peridomestic habitats. The release of laboratory-reared adults at the end of the dry season would ensure a good production of predator larvae by the beginning of the rainy season, when they are most needed. ACKNOWLEDGEMENTS I wish to express my thanks for the help of the Tanzanian mosquito scouts and the laboratory assistants, Messrs L. Mwitwa, L. Mahikwano, and C. Peter. R1SUMt INTERACTION ENTRE LE MOUSTIQUE PR1tDATEUR TOXORHYNCHITES BREVIPALPIS ET SA PROIE AEDES AEGYPTI L'etude de la distribution de frequence des taux d'infes- tation des gites larvaires par Aedes aegypti montre que la pr6sence de larves pr6datrices de Toxorhynchites brevi- palpis a pour effet de reduire fortement le nombre de larves d'A. aegypti partout oiu ce dernier se reproduit. Ces observations ont ete faites a Dar es-Salaam, Tanzanie. Le plus souvent, on ne decele qu'une larve de T. brevi- palpis par gite ce qui indique que les larves de cette espece se d6vorent aussi entre elles. Une larve de T. brevi- palpis au 4e stade consomme en moyenne 12,3 larves d'A. aegypti au 4e stade par 24 heures. L'emploi d'ovipieges est apparu comme une methode peu sensible pour mesurer les variations saisonnieres des populations de Toxorhynchites ou du rapport pr6dateur/ proie dans les biotopes suburbains ou sauvages. Le faible volume des r6cipients utilises favorise le cannibalisme des larves pr6datrices. Dans la plupart des cas, on n'a trouv6 qu'une larve de T. brevipalpis par piege. La densite des populations de T. brevipalpis dans les gites varie en fonction directe de la densit6 des moustiques proies. La pr6dation est plus active a la fin qu'au debut de la saison des pluies. La faible vitesse de croissance des larves pr6datrices freine l'action destructrice sur les larves proies, au developpement rapide. On pourrait y rem6dier 364 T. BREVIPALPIS AND A. AEGYPTI 365 en lachant des Toxorhynchites adultes, 6lev6s au labo- ratoire, peu de temps avant le d6but de la saison des pluies. Le systeme pr6dateur/proie oii T. brevipalpis est le pr6- dateur diff&e des autres systemes. Dans la plupart de ces derniers, le predateur meurt d'inanition lorsque la proie a ete completement detruite; en revanche, les larves de T. brevipalpis au 4e stade sont capables de supporter un jefkne prolong6 (6 mois ou plus). La baisse de la densit6 du pr6dateur, quand il n'y a plus de proie disponible, est due au cannibalisme. REFERENCES 1. CoRBET, P. S. Observations on Toxorhynchites brevi- palpis conradti Grub. (Diptera, Culicidae) in Uganda. Bull. ent. Res., 54: 9-17 (1963). 2. CORBET, P. S. & GRIurm, A. Observations on the aquatic stages of two species of Toxorhynchites (Diptera: Culicidae) in Uganda. Proc. roy. Soc. A, 38: 125-135 (1963). 3. HOLLING, C. S. Principles of insect predation. Ann. Rev. Entomol., 6: 163-182 (1961). 4. HOLLrNG, C. S. The functional response of inver- tebrate predators to prey density. Mem. Entomol. Soc. Can., 48: 1-86 (1966). 5. HOPKINS, G. H. E. Mosquitos of the Ethiopian region. I. Larval bionomics of mosquitos and taxonomy of culicine larvae, London, British Museum (Natural History), 1952, pp. 1-355. 6. MoRRis, R. F. The effect of predator age and prey defense on the functional response of Podisus maculi- ventris Say to the density of Hyphantria cunea Drury. Can. Entomol., 95: 1009-1020 (1963). 7. MusPRArr, J. The bionomics of African Megarhinus (Diptera, Culicidae) and its possible use in biological control. Bull. ent. Res., 42: 355-370 (1951). 8. TRPIs, M. Adult population estimate of Toxorhyn- chites brevipalpis. Bull. Wid Hlth Org., 48: 758-759 (1973). 9. TRIus, M. Seasonal changes in the larval populations of Aedes aegypti in two biotopes in Dar es Salaam, Tanzania. Bull. Wld Hlth Org., 47: 245-255 (1972). 10. TRPIS, M. Development and predatory behavior of Toxorhynchites brevipalpis (Diptera, Culicidae) in relation to temperature. Environ. EntomoL, 1: 537- 546 (1972). 11. TRPis, M. and GERBERG, E. J. Laboratory coloniza- tion of Toxorhynchites brevipalpis. Bull. Wld Hlth Org., 48: 637-638 (1973).