Subtle periodicity of pupation in rapidly developing mosquitos

Bull. Org. mond. Sante' 1970, 42, 951-955Bull. Wld Hlth Org. Subtle Periodicity of Pupation in Rapidly Developing Mosquitos With Particular Reference to Aedes vittatus and Aedes aegypti G. A. H. McCLELLAND 1 & C. A. GREEN ' Aedes vittatus and Ae. aegypti are important vectors ofyellowfever and other arbovirus diseases in Africa and they complete many generations, through all developmental stages, in a single year. Recent studies on Drosophila have offeredprecise techniquesfor analysing periodicities in populations of single developmental events in individuals. If the variability ofpupation times is low, periodicities may not be recognizable by an obviously polymodal pattern in a single synchronized population. Analysis of 8 synchronized populations of Ae. vittatus and Ae. aegypti, evenly spaced through 24 hours, provided a sensitive method of recognizing even subtle periodicities in rapidly developing mosquitos. By this means Ae. vittatus has been shown to exhibit a weak diurnal periodicity ofpupation while a strain of Ae. aegypti showed not the slightest periodicity ofpupation under the same light-dark cycle. The authors stress that periodicity of emergence or pupation in mosquitos can have important consequences for epidemiology and vector control and should be verified for each vector species for which control measures are envisaged. The expression of a circadian oscillation through a unique developmental event in the life of an indi- vidual animal has recently been interpreted by Skopik & Pittendrigh (1967) as a gating mechanism. The well-known example of Drosophila (Pittendrigh, 1954; Brett, 1955) illustrates how the frequency dis- tribution of individual eclosion times in a popula- tion is strikingly periodic and in phase with the diel light-dark cycle. During the period of larval growth, small differences in developmental rate accumulate and produce a very high variability in individual eclosion times which, in a light-dark (LD) cycle, are partitioned into many consecutive daily peaks. The variance was reduced by Skopik & Pittendrigh (1967) to nearly 24 h2 in constant light (LL) regimes by starting with synchronous pupae. In studying periodicity in pupation there are obviously no similar means of reducing developmental variance but, since the variance is to some extent proportional to the duration of development, a lower absolute value 1 Associate Professor of Entomology, University of Cali- fornia, Davis, Calif., USA, and Project Leader, WHO East Africa Aedes Research Unit, Dar es Salaam, Tanzania. I Medical Entomologist, De Beers Research Laboratory, Chiredzi, Southern Rhodesia. can be expected in more rapidly developing insects and at the temperature of fastest development. Periodicity of pupation has been most extensively analysed in the mosquito Aedes taeniorhynchus (Niel- sen & Haeger, 1954; Nayar, 1967; Provost & Lum, 1967; Lum, Nayar & Provost, 1968), where the lowest variance in pupation times was 72 h2 under LL regimes at 29.5°C with a mean larval duration of 100 h (Lum, Nayar & Provost, 1968). Even under the most favourable conditions, uni- modal pupation was never achieved under an LD cycle (Provost & Lum, 1967; Lum, Nayar & Provost, 1968). In contrast, the well-known yellow fever mosquito Aedes aegypti develops quite fast but was reported by Haddow, Gillett & Corbet (1959) and confirmed by Provost & Lum (1967) to show no pupational periodicity. Nevertheless, Rensing (1965) has sug- gested that this is due to comparatively small dif- ferences between maximum and minimum values, which implies either a wide gate or poor phase synchronization between individuals. Preliminary trials with a strain of Aedes vittatus from Southern Rhodesia which breeds in hot, exDosed. rock nools indicated Dossible periodicity 2536 951- G. A. H. MCCLELLAND & C. A. GREEN and showed that it had an extremely rapid rate of larval development with very low variance. At 36°C the fastest larvae pupated after 53 h and emerged as adult males in the 74th hour after the eggs hatched. Since this temperature is lethal for most other species, 31°C was used for the present comparison with Ae. aegypti. MATERIALS AND METHODS The colony of Aedes (Stegomyia) vittatus (Bigot) maintained in California and used in this study was started in January 1967 with eggs from adults col- lected as larvae in pools on the bed of the Chiredzi River, Southern Rhodesia (21°03' S, 31040' E). The strain of Aedes (Stegomyia) aegypti (L.) used was an F1 hybrid strain made by crossing females of the pale JD strain, originating from Saudi Arabia, to males of the dark YD strain originating from East Cameroon. Both parent strains were genetically marked and had been colonized for more than 7 years. The rationale for using such hybrid strains in research is well known (Bull. Wld Hlth Org., 1966) and the strain in question had already formed the basis of much prior work by the senior author. Larvae were reared in a set of 8 shallow enamel pans each containing 1800 ml of water maintained at 31'C±0.20C by partial immersion in a precisely THE DISTRIBUTION OF --H .. 12 is aA_ ^ _~~~~~~~~~21 RImEs controlled water-bath. In the light phase all pans received about 1200 lux. In the dark phase a dim red light (Wratten IA filter) gave less than 10 lux. Eggs, adhering to a paper strip, were put in the water at 0 h and removed at 3 h. An incubated infusion of 1 g homogenized Purina Mouse Breeder Chow was added at 0 h followed by 0.25 g of beef liver powder at about 3 h, 20 h and 40 h. Scum formation was prevented by gentle aeration, and evaporation was reduced by covering the pans with Plexiglass. In each experiment 8 consecutive populations were started at intervals of 3 h; these populations will be referred to by their relative hatch times 0, 3, 6 ..... 21 (part A of accompanying figure). Pupae were removed every hour and sexed on the basis of the shape of the ninth abdominal segment. Because the egg hatch was unpredictable and early instar larvae could not be counted under dim light, the number of larvae pupating per pan varied from 134 to 525, with 75%/ between 225 and 325. RESULTS AND DISCUSSION Aedes vittatus Reared under an LL regime, Ae. vittatus showed, as might be expected, no evidence of periodicity (A in fig.). The variance in pupation times was 25 h2 PUPATION TIMES IN SEQUENTIALLY HATCHED POPULATIONS OF MOSQUITO LARVAEa V .., l, ......... :x: >:-: >.: :.:. .: .;;s;B:::::::::::::::::::::,:, . :, :: .,::: ,:::::: Et.' ::B:: -:::: -: Saa Sw:::::::: E: :>:^.z.6- :-:.:.:6:6:6:.:x:::::x:-:^v ++ ,,,~~~~~~~~~~~~..., ,.E ..... ,~~~~~~S. a Each successively lower line represents a population hatched 3 hours later relative to the one above and reared at 31°C under the light regime indicated by shading. The hollow histograms represent the percentage of total males In each population pupating in a single hour; the solid bars similarly represent females. The oblique line cuts each base line at the same age for each population and represents the mean time of pupation of females in the 8 populations pooled (in B the value for A is used). A: Aedes vittatus LL B: A. vittatus LD (12:12). C: Aedes aegypti LD (12:12). --------------:.:Zdpdr-,.m a ... :;*;;:i :;:;;;: * : : _M90 .....__......... ,------- 952 SUBTLE PERIODICITY OF PUPATION IN RAPIDLY DEVELOPING MOSQUITOS for both sexes but only 8.2 h2 for the males alone. This contrasts with the lowest variance of 24.5 h2 in eclosion of females Drosophila under LL obtained by Skopik & Pittendrigh (1967). The mean larval durations ranged from 66.7 h to 68.7 h for males and from 71.9 h to 76.4 h for females. Because the distributions are clearly skew, little meaning can be attached to conventional sig- nificance measures, but it is clear that the sequence of mean times of pupation closely parallels the 3-hourly sequence of egg hatch and that there are no clear differences in the distributions of pupations among the 8 populations of either sex. The mean larval durations for the pooled populations were 67.3 h±0.10 h for males and 72.6 h±0.12 h for females. When the series was repeated, exactly as before except under an LD (12: 12) cycle, the mean larval durations for the pooled populations were 67.4 h± 0.09 h for males and 74.1 h±0.14 h for females. While this is virtually the same as the LL series, the distributions of the individual populations formed a strikingly different pattern (B in fig.). Note, however, that the distributions of pupations of both sexes in populations 0, 3, 18 and 21 were essentially unimodal. The mean larval durations are close to those under LL but the variances are obviously less. The onset of the dark period at 18.30 hours local time approximated to the natural evening, so that the four unimodal populations were all hatched in the period from 10.00 hours to 19.00 hours local time, which includes most of the working day. Casual observation of larvae from routinely hatched eggs in the laboratory would thus have failed to reveal any evidence of periodicity in pupation even if it had been suspected. The difference between the median larval durations of males and females would have been measured as 2.5 h to 3.5 h. The four populations, 6, 9, 12 and 15, hatched from 22.00 hours to 07.00 hours local time, presented a qualitatively different aspect. Pupation in one sex at least was clearly bimodal with the peaks separated by a little less than 12 h. The difference between the medium larval durations of males and females varied from 3.7 h in population 15 (hatched at 07.00 hours) to 11.6 h in population 9 (hatched at 01.00 hours). The pattern as a whole shows that pupation was minimal a little after the middle of the dark period and maximal a little before the middle of the light period. The males show 10 major peaks (peaks comprising at least 10% of the total males in the population) which form a series beginning with the second peak of population 12 (mean= 72.7 h), con- tinuing from the peak in population 21 to that in population 0 and ending with the first peak of popu- lation 15 (mean=63.0h). Two peaks of less than 10% in populations 9 and 18 show that the sequence probably extends further. The regression coefficient of mean larval duration on the relative time of hatching has a significant slope of -0.33+0.025. The females show a similar series of 10 major peaks beginning with the second in population 6 (mean= 79.7 h) and ending with the first in population 9 (mean = 67.4 h). The regression coefficient is -0.40±0.034 and is not significantly different from that of the males. The "jump" from the end of one sequence to the beginning of the next occurs in the populations showing double peaks whose means are separated by an average of 8.1 h for the males and 9.9 h for the females. The differences between the last and first means of each sequence are 6.7 h for the males, and 9.3 h for the females. The midpoints of these two periods occur 0.5 h after the midpoint of the dark period. The data in part B of the figure can clearly be interpreted as a subtle example of a gating pheno- menon in which the width of the gate, or permissive period for pupation, is about 17 h for males and 15 h for females in contrast to only 6 h in the case of both sexes of Drosophila (Skopik & Pittendrigh, 1967). Preliminary laboratory studies at constant tem- peratures show that emergence of adults occurs, like pupation, during the light phase. When the temperature was cycled between 27°C and 42°C emergence occurred at the temperature maximum during the light phase. In its rock habitat such an emergence pattern might be a protection from water surface predators such as lycosid spiders. It is clearly an adaptively different situation from that in Drosophila. Field data on periodicity in Ae. vittatus are, however, less clear. Water in the natural breeding- places in the bed of the Chiredzi River typically fluctuated daily between limits of 27°C and 42°C during a study period in 1968. Hourly counts of pupae showed good agreement with the diurnal gate found in the laboratory in several sets of observations. Emergence of adults in one rock pool on 19 March 1968 was almost entirely between 10.30 hours and 15.30 hours local time. During this period the water temperature rose from 34°C to 38°C and fell again to 34°C. In two other pools, 953 G. A. H. MCCLELLAND & C. A. GREEN on 24 April 1968, emergence was not synchronized and appeared to be aperiodic but unfortunately the temperature was not recorded. Aedes aegypti Eight populations were set up under an LD (12: 12) cycle exactly as in the previous experiment. The mean larval durations were 84.5 hI 0.16 h and 88.6 hI0.20 h and the variances 26.0 h2 and 32.9 h2 for males and females respectively. The pattern of distribution of pupation (C in fig.) shows no evidence of periodicity and closely resembles that of Ae. vit- tatus under LL (A in fig.). While this may not eliminate the possibility that some other non- hybrid strains of Ae. aegypti might show periodic pupation, it is at least clear that some degree of periodicity cannot be regarded as a general charac- teristic of pupation. GENERAL DISCUSSION AND CONCLUSIONS The importance of recognizing subtle periodicities can be illustrated by the failure of de Meillon, Sebastian & Khan (1967a) to find evidence of peri- odic emergence or pupation of the mosquito Culex pipiensfatigans in the laboratory despite their demon- stration (de Meillon, Sebastian & Khan, 1967b) of periodic emergence in the field. Their findings were based on a single synchronously hatched population and because the variance was low each sex could have exploited a single gate. Early reports of periodic pupation, summarized by Provost & Lum (1967), the work cited in the intro- duction, a recent study by Nayar (1968) and the pre- sent evidence show much quantitative and qualitative variation in this phenomenon between different species. The significance of periodicity of insect behaviour to ecology and, in the case of vectors, to epidemiology, is well known. Periodicity in emergence could be an important factor in determin- ing the optimal time of application or release of chemical and non-chemical control agents, including sterile or genetically altered males for autocidal control. It is important therefore that even subtle periodicities are recognized and that the absence of a typical polymodal pattern of pupation or emer- gence in the laboratory is not interpreted as an absence of periodicity. ACKNOWLEDGEMENTS We thank R. C. Witt and E. T. Schmidtmann for technical help. RESUME PtRIODICITI, DISCRtTE DE LA NYMPHOSE CHEZ DES MOUSTIQUES A DtVELOPPEMENT RAPIDE, tTUDItE EN PARTICULIER CHEZ AEDES VITTATUS ET AE. AEGYPTI Les techniques employees pour analyser la periodicite de l'eclosion imaginale chez Drosophila ont ete adaptees a l'etude de la nymphose chez Aedes vittatus et Ae. aegypti. Huit populations successives de ces deux especes ont ete constituees a intervalles de 3 heures, les larves etant maintenues 'a une temperature constante de 31°C. Les nymphes de chaque sexe ont &6 comptees toutes les heures. Chez Ae. vittatus plac6 dans des conditions constantes d'eclairement normal, la nymphose se produit dans chaque population suivant une distribution unimodale, l'ecart de 3 heures etant conserve. La duree du stade larvaire est de 67 heures pour les insectes males et de 73 heures pour les femelles. Lorsqu'on modifie les modalitds d'eclairement, avec alternance de 12 heures d'eclairement normal et de 12 heures d'obscurite, la dur6e du stade larvaire reste pratiquement inchangee, mais dans 4 populations sur 8 la nymphose s'opere chez l'un des sexes au moins suivant une distribution bimodale. Le passage au stade nymphal est tres limite pendant la periode de maintien des larves a l'obscurite; cependant le laps de temps oiu elle est possible atteint 17 heures pour les larves males et 15 heures pour les femelles, alors que chez Drosophila 1'eclosion imaginale ne peut se faire que dans un intervalle de 5 heures. Les observations faites sur le terrain sont plus difficiles a interpreter, en raison des 954 SUBTLE PERIODICITY OF PUPATION IN RAPIDLY DEVELOPING MOSQUITOS 955 variations de temperature, mais la nymphose et l'eclosion imaginale de Ae. vittatus ont lieu gen6ralement durant le jour. Une etude similaire portant sur une souche hybride d'Ae. aegypti n'a mis en evidence aucune periodicite de la nymphose. En ce qui concerne Culex pipiens fatigans, etudie par d'autres auteurs, il semble que l'echec des tentatives visant a d6montrer des faits du m8me ordre soit imputable a la technique utilisee. Les auteurs soulignent l'importance ecologique des phenomenes de periodicite dans le comportement des insectes et leur interet pour 1'epidemiologie et I'applica- tion des mesures de lutte dans le cas de vecteurs de maladies. REFERENCES Brett, W. J. (1955) Ann. ent. Soc. Amer., 48, 119-131 Bull. Wid Hith Org., 1966, 34, 460 Haddow, A. J., Gillett, J. D. & Corbet, P. S. (1959) Ann. trop. Med. Parasit., 53, 123-131 Lum, P. T. M., Nayar, J. K. & Provost, M. W. (1968) Ann. ent. Soc. Amer., 61, 889-899 Meillon, B. de, Sebastian, A. & Khan, Z. H. (1967a) Bull. Wld Hlth Org., 36, 7-14 Meillon, B. de, Sebastian, A. & Khan, Z. H. (1967b) Bull. Wld Hlth Org., 36, 163-167 Nayar, J. K. (1967) Ann. ent. Soc. Amer., 60, 946-971 Nayar, J. K. (1968) J. med. Ent., 5, 39-46 Nielsen, E. T. & Haeger, J. S. (1954) Bull. ent. Res., 45, 757-768 Pittendrigh, C. S. (1954) Proc. nat. Acad. Sci. (Wash.), 40, 1018-1029 Provost, M. W. & Lum, P. T. M. (1967) Ann. ent. Soc. Amer., 60, 138-149 Rensing, L. (1965) In: Aschoff, J., ed., Circadian clocks, Amsterdam, North-Holland, p. 406 Skopik, S. D. & Pittendrigh, C. S. (1967) Proc. nat. Acad. Sci. (Wash.), 58, 1862-1869