Abstracts

The behavior of sympatric Chilean populations of Drosophila larvae during pupation

Raúl Godoy-Herrera and José Luis Silva-Cuadra

Instituto de Ciências Biomédicas, Programa de Genética Humana, Facultad de Medicina, Universidad de Chile, Independencia 1027, Santiago-7, Chile. Fax: 56-2-7373158. Send correspondence to R.G.-H.

ABSTRACT

The behavioral mechanisms by which the larvae of sympatric Chilean populations of

Drosophila melanogaster, D. simulans, D. hydei and

D. busckii select pupation sites are described in terms of larval substrate preferences. These species share the same breeding sites in Central Valley of Chile. It is important to investigate larval substrate preferences to pupate in sympatric natural populations of species of

Drosophila because such behavior could contribute to the coexistence of the species in the wild.

D. busckii larvae preferred humid substrates with a smooth surface to pupate, whereas

D. simulans larvae selected humid substrates with a rough surface. Larvae of

D. melanogaster chose dry and humid substrates with a rough surface, whereas

D. hydei larvae occupied dry substrates with a smooth surface to form puparia.

D. melanogaster larvae dug deeper into dry than into humid sand, whereas

D. simulans larvae dug more into humid sand.

D. busckii larvae pupated in the upper layers of humid and dry sand, and

D. hydei larvae dug more into humid than into dry sand. Pupae of the four

Drosophila species showed aggregated distributions on the substrates. Larval prepupation behaviors of

D. melanogaster, D. simulans, D. hydei and

D. busckii could be important to their coexistence in the wild. INTRODUCTION In the present work we investigate larval prepupation behavior in four

Drosophila species collected in the same place and which use the same decaying fruits as breeding sites (Brncic, 1980, 1987). Little is known about the physical or possible social factors important to the larvae that are searching for pupation sites (Godoy-Herrera

et al., 1989). Most previous studies on pupation behavior of

Drosophila larvae have employed food vials, and pupation site preference has been measured by the height from the surface of the substrate to where the pupae are found (review in Singh and Pandey, 1993). This arrangement provides little insight into which features of the environment consider

Drosophila larvae to select habitats for pupating. Using other designs, Wong

et al. (1985) and Godoy-Herrera

et al. (1989) observed that

Drosophila melanogaster larvae react to humidity, light, and the texture and consistency of the substrates, and that all these behaviors have an influence on the selection of pupation sites.

Aspects of choice of pupation sites by Drosophila have a genetic basis (review in Grossfield, 1978). Pupation by D. melanogaster on dry substrates outside the rearing cup is dominant over pupation inside the cup, and there is also additive variation (Godoy-Herrera et al., 1989). Using shell vials, Singh and Pandey (1993) found that pupation height in D. ananassae is under polygenic control, whose variance seems to be mostly additive. The type of pupation site selected by larvae also affects pupal survival of D. melanogaster (Rodriguez et al., 1992). In this species the same larvae may dig into agar and pupate there; these larvae seem to be less frequently parasited by the hymenoptera Pechycrepoideus vindemniae than those that pupate on the surface of the substrate.

Here we compare larval prepupation behavior of Chilean sympatric populations of D. melanogaster, D. simulans, D. hydei and D. busckii in terms of larval substrate preferences. We also report on the depth at which these larvae pupate in different substrates. We used these four species because they coexist in the same decaying fruits in the wild. Besides, they represent the commonest Drosophila species in Central Valley of Chile (Brncic, 1987). If the behavior of larvae of different sympatric species of Drosophila during pupation leads them to pupate in different substrates, such behavior could have a direct effect on evolution of those species. Thus, the present investigation may provide evidence for niche subdivision in pupation site which may contribute to coexistence of the species.

MATERIAL AND METHODS

Stocks

Wild type flies of the sibling species

D. melanogaster and

D. simulans (subgenus

Sophophora, melanogaster group),

D. hydei (subgenus

Drosophila, repleta group) and

D. busckii (subgenus

Drosophila) were collected in the middle of April of 1994 (autumn in Chile) in Chillán, 420 km south of Santiago. In this season Chilean populations of

Drosophila reach their peaks of abundance (Brncic, 1980). The collections were made in an orchard of the University of Concepción (Faculty of Agronomy). This is a humid site with ornamental plants and bushes growing together with cherry, plum, medlar, apple and peach trees, tomato and grape plants, and native vegetation. Once these fruits are decaying on the ground, the species use them as breeding sites (Brncic, 1980, 1987).

Drosophila immigrans, D. pavani, D. repleta and

D. subobscura may also emerge from the fruits. About ten flies founded each base population. The stocks are kept by mass culture in the Department of Cellular Biology and Genetics, Faculty of Medicine, University of Chile. Not more than 4 generations in the laboratory had occurred when the flies were used in the present research.

The flies were all reared in a common environment under constant light at 24^oC in half-pint bottles containing about 50 cc of Burdick’s medium (1954).

Egg collection

Groups of 60-80 inseminated females were allowed to lay eggs for 3-4 h on plastic spoons filled with culture medium. Eggs laid were collected with the aid of a dissecting needle. One hundred eggs of each species were distributed on separate cups measuring 2.0 x 2.3 cm also filled with Burdick’s medium.

Behavioral experiments

We used transparent Perspex boxes as described by Godoy-Herrera

et al., 1989. The boxes were situated in the same site of the rearing chamber to ensure that no other factor such as variation in illumination and temperature fluctuations could be responsible for the differences observed (see Results). The species were examined separately; mixed species experiments were not ran.

Binary choices

In the first set of experiments (

Table I), larval pupation behavior of the four species was examined in dry and humid substrates. A food cup was placed in the center of each of four empty boxes (experiment 1a). These boxes served as controls for experiment 1b, c and d. Another four boxes were filled to a depth of 2 cm with 3% agar (experiment 1b) to provide a humid substrate. Four similar boxes were filled to a depth of 2 cm with dry yellow sand (experiment 1c). Finally, another four boxes were again filled with a 2-cm layer of sand moistened with 10 ml of distilled water (experiment 1d). The boxes were maintained at 24

^oC for 6 days (

D. melanogaster and

D. simulans) and 12 days (

D. hydei and

D. busckii). The larval period is longer for these two latter

Drosophila species. Then, the number of pupae found inside and outside the cup (the agar, sand and Perspex surfaces of the boxes) was recorded.

Ternary choices

In the next set of experiments, pupation substrate preferences of

D. melanogaster, D. simulans, D. hydei, and

D. busckii were tested, offering the larvae three different substrates simultaneously (

Table I, experiment 2a,b,c). For each of the four species, a set of twelve Perspex boxes was filled with 3% agar to a depth of 2 cm, then an area of agar measuring 9.0 x 4.5 cm was removed along one side to leave a dry surface. A food cup was placed at the bottom of four such boxes. Thus, larvae could choose to pupate either on agar, on the Perspex, or in the rearing cup (experiment 2a). The Perspex part of eight boxes was filled with the yellow sand (see above) forming a 2-cm thick layer. In four of the boxes the sand was dry (experiment 2b), and in the remaining ones the sand was moistened homogeneously with 10 ml of distilled water (experiment 2c). Again, a cup was placed at the bottom of each of the boxes, and they were maintained at 24

^oC for 6 days (

D. melanogaster and

D. simulans) and 12 days (

D. hydei and

D. busckii). The number of pupae detected outside (the Perspex, agar, dry and humid sand) and in the rearing cup was recorded prior to eclosion.

Digging

We recorded the number of pupae from the four species found on and underneath dry and humid sand (binary choices) (

Table II). In order to observe how deep the larvae had dug into the dry and humid sand we gently removed successive 5-mm thick sand layers. We scattered the sand of each layer on a white paper sheet counting and recording the number of pupae found. We also recorded the number of pupae observed inside the food cup. For each species we examined ten Perspex boxes filled with dry sand and another ten filled with humid sand. Each group of twenty boxes was set up and analyzed in the same way as described for the binary choices (experiments 1c,d,

Table I).

Table II - Percentage and S.E. of pupae of D. melanogaster, D. simulans, D. hydei and D. busckii found on and buried in dry and humid sand. Percentage of pupae detected in the rearing cup is also given.

Distance from the cup

When larvae of

D. melanogaster, D. simulans, D. hydei and

D. busckii pupated away from the rearing cup, the distance between the pupae found on the surface of the Perspex, agar, dry and humid sand, and the center of the cup (experiments 1a,b,c,d,

Table I) was recorded. We tested the normality of the frequency distributions of the data obtained. We presumed that if the data departed from normality it could indicate some type of clumped distribution of pupae. For each species, four boxes per substrate with their respective food cups sown with 100 eggs each were examined.

For another estimation of aggregation of pupae, the bottom of the boxes (see above) was divided into four equivalent quadrants by tracing two perpendicular lines which cut at the center of the cup. The number of pupae in each of the four quadrants was an estimation of aggregation (see below).

Statistical analysis

To test whether the number of pupae inside and the number of pupae outside the rearing cup (binary and ternary choices,

Table I) were significantly different we applied the G-test of Independence (Sokal and Rohlf, 1995). It was also applied to test whether the number of pupae on and underneath the dry and humid sand was significantly different. Before applying the test we carried out unplanned tests of the homogeneity for replicates made to each species in each experiment (the R X C test of Independence (Sokal and Rohlf, 1995)). Chi-square test was applied to test number of pupae per quadrant (see

Table III). The null hypothesis of no aggregation predicts a uniform distribution of pupae per quadrant.

₁) and kurtosis (g

₂) (Sokal and Rohlf, 1995). These statistical parameters are significant when data are clumped and the mean and medians do not coincide. Thus, skewness and kurtosis significances were other measures of aggregation of pupae. RESULTS The statistical analysis showed that there were no significant differences within a species and among replicates in each experiment and so we pooled the data as shown in

Figures 1 to 4.

Binary choices

The preferences of larvae to pupate on the Perspex, on agar, and inside the cup are shown in

Figure 1a,b.

D. melanogaster and

D. simulans pupae were preferentially found inside the cup when the Perspex surface was the only other place available to pupate (

Figure 1a). However, larvae of

D. hydei preferred the Perspex substrate to form puparia.

D. busckii pupae were found only in the rearing cup (

Figure 1a). Differences between

D. melanogaster and

D. simulans, and between these two species and

D. busckii were not significant (G-test of Independence; d.f. = 1, P > 0.05).

When agar was available to pupate (Figure 1b), we found that: i) larvae of D. melanogaster, D. simulans and D. hydei preferentially pupated outside the cup, and ii) D. busckii larvae pupated only outside the cup. The differences between D. melanogaster and D. simulans were not statistically significant (G-test of Independence, d.f. = 1, P > 0.05). However, these patterns differed with respect to those exhibited by D. hydei (G-test of Independence values in relation to D. melanogaster and D. simulans were 67.89 and 71.93, respectively; d.f. = 1, P < 0.001). Larval pupation behavior of D. busckii was also different from those of D. melanogaster and D. simulans (G-test = 129.92 and 130.02, respectively, d.f. = 1, P < 0.001) (Figure 1b).

When dry sand was offered as a substrate (Figure 2a), we found: i) most of D. melanogaster and D. hydei pupae outside the food cup, ii) D. simulans pupae only in the cup, iii) most of D. busckii larvae pupated in the cup. These differences in substrate preferences among the four species were all statistically significant (G-test of Independence values over 70.00, d.f. = 1, P < 0.001).

When humid sand was available as a substrate (Figure 2b), the results show that the species largely pupated outside the rearing cup. However, the species differed in the proportion of pupae found on the sand (G-test of Independence D. melanogaster versus D. simulans = 11.58, d.f. = 1, P < 0.05; D. simulans versus D. hydei, G = 13.58, d.f. = 1, P < 0.05; D. hydei versus D. busckii, G = 3.67, d.f. = 1, P > 0.05).

Ternary choices

When the rearing cup was surrounded by agar and Perspex (

Figure 3a), our results indicate that: i) most of

D. melanogaster, D. simulans, D. hydei and

D. busckii pupae were observed on agar, ii)

D. busckii larvae did not use the Perspex to pupate. However, the patterns of distributional pupae differed among species (

D. melanogaster versus

D. simulans, G = 17.22, d.f. = 2, P << 0.005;

D. simulans versus

D. hydei, G = 27.44, d.f. = 2, P << 0.0025;

D. busckii versus

D. simulans, G = 48.52, d.f. = 2, P << 0.0025).

When Perspex and dry sand surrounded the food cup (Figure 3b) we found that: i) D. melanogaster larvae occupied preferentially dry sand and the cup to form puparia, ii) D. simulans larvae pupated in the food cup, iii) most of D. hydei larvae formed puparia outside the cup on the Perspex and dry sand, iv) D. busckii pupae were preferentially seen inside the food cup, and dry sand was not used to pupate. All of the differences between the species were statistically significant (G-test of Independence yielded values over 80.00; d.f. = 2, P << 0.0025).

When agar and dry sand were available as substrates (Figure 4a), we found: i) most of D. melanogaster pupae on agar, and a few pupae in the food cup, ii) D. simulans pupae on agar, whereas other important group of pupae was detected on dry sand, iii) D. hydei pupae principally outside the cup on agar and dry sand, and iv) most of D. busckii pupae on agar. G-tests of Independence for substrate patterns of preferences of the Drosophila species were all significant (for example, in the case of D. simulans and D. hydei, which seemed more similar in larval prepupation behavior, G = 7.44, d.f. = 2, P < 0.05).

When agar and humid sand were offered as substrates (Figure 4b) the results indicate that the four species occupied preferentially the sand to form puparia. However, distributional patterns of pupae of D. melanogaster as compared to those of D. simulans and D. hydei were different (G = 15.56, d.f. = 2, P < 0.001; G = 15.22, d.f. = 2, P < 0.001, respectively). Distribution patterns of pupae of D. simulans and D. hydei were also statistically different (G = 13.55, d.f. = 2, P < 0.001). Larval pupation behavior of D. busckii was also different in respect to those of the other three species (see Figure 4b).

Digging

Table II shows the percentage of pupae of the respective species found on/underneath dry and humid sand and pupae in the food cup. A large proportion of

D. melanogaster pupae were found buried in dry sand. A number of these were in the deepest layers of the substrate.

D. simulans larvae pupated in the rearing cup. Almost 50% of

D. hydei pupae were detected outside the cup, most of these on the sand. In the case of

D. busckii, most of the pupae were detected on dry sand, whereas a few larvae dug into the upper layer. Differences in digging between

D. melanogaster, D. hydei and

D. busckii were statistically significant (G = 1419.99, d.f. = 4, P << 0.005).

D. melanogaster pupae were found largely on humid sand (Table II). An important number of D. simulans larvae pupated on humid sand. D. hydei pupae were found in the upper layer of the substrate, whereas D. busckii larvae preferred to pupate on humid sand. Differences in patterns of digging between D. melanogaster, D. simulans, D. hydei and D. busckii were statistically significant (G-test of Independence = 761.48, d.f. = 4, P << 0.0025).

Number of pupae per quadrant

Outside the rearing cup (the Perspex, agar, dry and humid sand) larvae of the species formed puparia near other larvae, as shown by the distributional patterns of pupae per quadrant (

Table III). The quadrant containing the greatest number of pupae varied among the replicates, suggesting that the preferred quadrant seems to be selected at random by the first larvae that leave the rearing cup (Chi-squares for the four species were all over critical value 7.82, d.f. = 3, P < 0.001).

Distance from the cup

Table IV indicates that the observed frequency distributions of the four species depart from normality, as measured by parameters g

₁ (skewness) and g

₂ (kurtosis). The observed frequency distribution patterns depended on the kind of substrate and the species. On Perspex,

D. melanogaster showed a leptokurtic distribution of pupae, that is, most of the larvae preferred to pupate near the center of the distribution. In the same environment,

D. simulans pupae showed a left skewed and platykurtic distribution (the pupae were clumped at the center and the tails of the distribution). Interestingly, on Perspex,

D. hydei pupae followed a platykurtic pattern of distribution (most of the pupae were clumped along the center and the tails of the distribution). No

D. busckii pupae were detected on Perspex.

*P < 0.05; **P < 0.001.

On agar,

D. melanogaster pupae distributed following a left skewed pattern distant from the cup (

Table IV). On this substrate a platykurtic pattern of distribution was exhibited by

D. simulans. Both

D. hydei and

D. busckii pupae showed a left skewed and platykurtic distribution pattern (

Table IV). Comparable results were obtained on dry and humid sand for the four species (

Table IV). DISCUSSION

Larval substrate preferences

Taken together the results show that sympatric Chilean populations of four cosmopolitan

Drosophila species differ in their larval prepupation behaviors. The behaviors seem to reflect their phylogenetic relationships. Thus,

D. melanogaster and

D. simulans (

melanogaster group) exhibited the greatest similarities in prepupation behavior as compared to larvae of

D. hydei (

repleta group) and

D. busckii (subgenus

Drosophila). The findings suggest that larvae of

D. melanogaster, D. simulans, D. hydei and

D. busckii also differ in their information-processing mechanisms relative to water content, compactness and texture of substrates.

Such an inference is in line with laboratory results reported by Wong et al. (1985) and Godoy-Herrera et al. (1989) who were also able to show that water content, light, and surface texture and consistency of substrates affect larval prepupation behavior of D. melanogaster. As we tested larval preferences for substrates on which to pupate in the four Drosophila species individually, we do not know whether the presence of another Drosophila species may also modify this larval prepupation behavior. We are planning to investigate this further.

D. busckii larvae seemed to be very sensitive to the humidity and texture of the substrates outside the cup. In contrast, D. hydei larvae in different environmental circumstances tried to leave the rearing cup occupying microhabitats available. The larvae leave residues in the culture medium as uric acid and urea that contaminate it (Brncic and Budnik, 1974). In the case of D. hydei larvae, perhaps waste larval residues present in the cup could affect their viability, and thus they leave the food cup and pupate outside.

The response of D. melanogaster and D. simulans larvae to humidity and texture of the substrate seemed to be intermediate between D. busckii and D. hydei. However, D. melanogaster larvae occupied dry sand to pupate, whereas D. simulans larvae did not use this substrate. Little is known about the cellular biology and physiology of the sensory organs and central nervous processes in Drosophila larvae. However, the embryonic development of larval sensilla of D. melanogaster has been described by Harstenstein (1988), and it is known that this organ is involved in larval olfaction (review in Cobb and Danne, 1994).

Digging

Our findings suggest that digging behavior contributes to pupation site selection in

D. melanogaster and

D. hydei larvae. In contrast, digging into the substrate seemed to be less important in

D. simulans and

D. busckii larvae. Besides, larvae of the four species responded to humidity of the substrate by digging into it. However, the information seemed to be processed in different way by each species. Thus,

D. melanogaster larvae dug more into dry sand and less into humid sand. In contrast, the opposite behavior was exhibited by

D. hydei larvae.

Aggregation

Pupae found outside the cup showed an aggregated distribution on the substrates. Thus, larvae of

D. melanogaster,

D. simulans, D. hydei and

D. busckii seemed to also consider the presence of other larvae to select pupation sites. We have not investigated the mechanisms by which the recognition may occur. However, Pruzan and Bush (1977) reported that

D. melanogaster larvae are able to recognize residues of conspecific larvae located at certain distance from them.

D. melanogaster larvae can also respond to substances such as acetates, aldehydes and alcohols (rewiev in Cobb and Danne, 1994). Another alternative hypothesis is that the substrate is patchy from the perspective of the larva and some patches could be preferred by the larvae resulting in an aggregated distribution of pupae. We are also planning to investigate this further.

In short, we would like to suggest that the larval behaviors here described could be a part of the complex of behavioral mechanisms underlying the coexistence of natural Chilean populations of D. melanogaster, D. simulans, D. hydei and D. busckii. We thought that if the behavior of larvae of different sympatric species during pupation leads them to pupate in different substrates, such behavior could be a key mechanism to prevent competition for pupation sites between the species. Clearly, mixed species experiments could illuminate our comprehension about this problem.

ACKNOWLEDGMENTS This work is dedicated to the memory of Rosita-Herrera Sepúlveda, Raúl Godoy-Osses, Elbita Herrera-Sepúlveda, Domingo Valenzuela-Moya, and Francisco Aceval-Cid. R.G.-H. also dedicates this paper to his wife Alejandra Rosas-Rosas. We are indebted to Dr. Susi Koref-Santibañez for reading and commenting the manuscript. The present study started at the Department of Psychology, Sheffield University, U.K. Thanks are due to The British Council for a grant-in-aid. R.G.-H. wishes also to thank Professor Kevin J. Connolly and Dr. Barrie Burnet for their warm hospitality. Research supported by FONDECYT No. 1960727-96 and D.T.I. Universidad de Chile No. B-2309-9155. RESUMO Os mecanismos comportamentais pelos quais as larvas de populações Chilenas simpátricas de

Drosophila melanogaster, D. simulans, D. hydei e

D. busckii selecionam os locais de pupação são descritos em termos de preferências larvais pelos substratos. Essas espécies compartilham os mesmos locais de procriação no Vale Central do Chile. É importante investigar em populações naturais simpátricas de espécies de

Drosophila os substratos preferenciais das larvas para pupar, porque este comportamento poderia contribuir para a coexistência da espécie na natureza. As larvas de

D. busckii preferiram pupar em substratos úmidos com superfície lisa, enquanto que as larvas de

D. simulans escolheram substratos úmidos com superfície áspera. As larvas de

D. melanogaster preferiram substratos com superfície áspera, tanto úmidos como secos, enquanto que as larvas de

D. hydei ocuparam substratos secos com superfície lisa. As larvas de

D. melanogaster cavaram mais profundamente na areia seca do que na úmida, enquanto que as larvas de

D. simulans cavaram mais na areia úmida. As larvas de

D. busckii puparam nas camadas superficiais de areia tanto úmida como seca e as larvas de

D. hydei cavaram mais na areia úmida do que na seca. As pupas das quatro espécies de

Drosophila apresentaram distribuições agregadas nos substratos. Os comportamentos pré-pupação de larvas de

D. melanogaster, D. simulans, D. hydei e

D. busckii poderiam ser importantes para sua coexistência na natureza.

REFERENCES

Brncic, D. (1980). Genética ecológica de especies colonizadoras de

Drosophila que viven en Chile.

Proceedings of the IV Latinoamerican Congress of Genetics, Vol. 2, pp. 181-190.

Brncic, D. (1987). Coexistencia de diferentes especies de Drosophila en frutas fermentadas naturalmente. Medio Ambiente 8: 3-9.

Brncic, D. and Budnik, M. (1974). Rate of development and viability of hybrids between Drosophila pavani and Drosophila gaucha under competitive conditions. Ecology 55: 662-666.

Burdick, A.B. (1954). New medium of reproductivity quality stable at room temperature. Drosoph. Inf. Ser. 28: 170.

Cobb, M. and Danne, F. (1994). Multiple genetic control of acetate-induced olfactory responses in Drosophila melanogaster larvae. Heredity 73: 444-455.

Godoy-Herrera, R. (1994). Biometrical analysis of larval digging behavior in Drosophila melanogaster. Behav. Genet. 24: 427-432.

Godoy-Herrera, R., Cifuentes, L., Díaz de Arcaya, M.F., Fernández, M., Fuentes, M., Reyes, I. and Valderrama, C. (1989). The behaviour of Drosophila melanogaster larvae during pupation. Anim. Behav. 37: 820-829.

Grossfield, J. (1978). Non-sexual behavior of Drosophila. In: The Genetics and Biology of Drosophila (Ashburner, M. and Wright, T.R.F., eds.). Academic Press, New York, pp. 1-126.

Harstenstein, V. (1988). Development of Drosophila larval sensory organs: spatiotemporal pattern of sensory neurones, peripheral axonal pathways and sensilla differentiation. Development 102: 869-886.

Pruzan, A. and Bush, G. (1977). Genotypic differences in larval olfactory discrimination in two Drosophila melanogaster strains. Behav. Genet. 7: 457-464.

Rodriguez, L., Sokolowski, M.B. and Shore, J.S. (1992). Habitat selection by Drosophila melanogaster larvae. J. Evol. Biol. 5: 61-70.

Singh, B.N. and Pandey, M.B. (1993). Selection for high and low pupation height in Drosophila ananassae. Behav. Genet. 23: 239-243.

Sokal, R.R. and Rohlf, F.J. (1995). Biometry. 3rd edn. W.H. Freeman and Company, New York.

Wong, J., Sokolowski, M.B. and Kent, C.F. (1985). Prepupation behavior in Drosophila: Embedding. Behav. Genet. 15: 155-164.

(Received February 3, 1997)

Publication Dates

History