Aggregation tendency in small groups of Drosophila melanogaster

Aggregation tendency in small groups of Drosophila melanogaster

Eduardo del Solar Osses

Institute of Ecology and Evolution, Austral University of Chile, Casilla 567, Valdivia, Chile

ABSTRACT

A study was made of the spatial pattern of the progeny produced in boxes with 25 oviposition tubes, by groups of six Drosophila melanogaster females during 24, 72 and 216 h. Each female was fertilized by a male carrier of a dominant mutation in such a manner that the behavior of each female of the group as well as that of each group could be determined by a study of their progeny. The pattern of spatial distribution of the progeny was aggregated, reaching its highest value at 72 h. Gregarious behavior of individuals was defined by the number of sites among which they disseminated their progeny in a 24 h-cycle; while for groups it was estimated by the variance/mean ratio. The fertility, egg-adult development time and mortality rate were recorded. Comparisons between the tubes showed statistically significant differences in the number of individuals and of phenotypical composition. These gregarious distributions could contribute to local genetic differentiation.

INTRODUCTION

It is widely accepted that populations constitute discrete entities formed by groups of individuals of different sizes, disseminated in a heterogeneous space (Forney and Gilpin, 1989; Messina 1989; Milne, 1992). This non-random spatial distribution pattern has been described with gregarious or contagion statistics (Pielou, 1969; Poole, 1974). Laboratory and field work done with Drosophila has shown that aggregated distribution is genetically determined (del Solar, 1968; Ruiz and del Solar, 1986, 1992; del Solar and Ruiz, 1992) and can be modulated by environmental factors (reviewed by Grossfield, 1978).

The most widely studied expression of this gregarious behavior is selection of an oviposition site. In a recent paper (del Solar and Ruiz, 1992) (op. cit.) found that individual Drosophila melanogaster females can be classified into one of two mutually exclusive categories: i) those which concentrate all eggs oviposited during 24 h in one site and ii) those which disseminate their eggs among two or more different sites. These results for individual flies could be affected by interactions that occur among flies in groups. The purpose of the present study was to examine this behavior in small groups.

The use of small groups is a realistic approximation and theoretically simple since quantitative evaluations of population densities and sizes of different species of Drosophila in temperate zones show values of 0.07 up to 9.0 individuals per 100 m² and populations between 400 and 16,000 flies (Taylor and Powell, 1983).

MATERIAL AND METHODS

The experiments were carried out with a population of D. melanogaster from Valdivia collected four months before starting the study. They were kept in a mass culture in standard laboratory conditions at 21 ± 1^oC, 60% relative humidity, and 12-h light-darkness cycles.

The procedure was as follows: every hour virgin females were collected and distributed in groups of 10 individuals and deposited in 250-ml capacity bottles with 50-ml culture (Burdick, 1954). Ten males carrying a dominant mutation with an easily recognizable phenotypical effect were added to each group. Each group was kept for three days under the same laboratory conditions described above. The male phenotypical markers were Bar (B) and Ocellusless (Oce) on chromosome I, Curley (C) and Lobe (L) on chromosome II, and Wrinkled (W) and Kinked (Ki) on chromosomes III (Lindsley and Zimm, 1985, 1990).

On the fourth day, new groups of six females were formed by taking one female from each bottle so that each of them had been fecundated by a different marker male. Each of these groups was placed in a 40 x 40 x 10 cm transparent acrylic box, with 25 oviposition tubes geometrically arranged in a five by five block. Each tube was 2.5 cm height, 2.0 cm of diameter and separated each other for 2.5 cm. There were 15 population boxes in total. After 24 h, five boxes were picked at random and the number of tubes containing eggs, the number of eggs per tube, and the relative position of the tube in the box were registered. Each tube containing eggs was introduced into a sterilized bottle and incubated independently until the emergence of adults. These were classified according to the marker phenotype, registering the number and diversity per tube, and at the same time the egg-adult interval of development.

The same procedure was carried out with five other boxes at 72 h and the final five boxes were processed at 216 h. All the manipulations were carried out without anesthesia, using an aspirating tube. The whole procedure was repeated a second time to obtain 10 replications at 24, 72 and 216 h. The aggregation indicators were: i) the average number of tubes occupied by eggs in each cage; ii) the frequency of individuals emerged in the "most preferred tube" (the tube with the highest number of adults emerged among all the tubes from the boxes), and iii) an aggregation index estimated by the variance/mean ratio. At the individual level, the criterion for aggregation was the number of tubes used per female to disseminate her progeny among the 25 available tubes in each box.

The development time was calculated in days. Mortality was estimated as the difference between numbers of oviposited eggs and adults emerged from each tube.

RESULTS AND DISCUSSION

There was an increase of the number of adults and of occupied tubes and a decrease in the percentage use of the preferred tube with increased time in each cage (Table I). At the same time, a decrease of the aggregation rates was expected since a decrease in the number of areas available for oviposition would tend to make the distribution of eggs among the available tubes in each population cage more uniform. The value expected for a random distribution is a variance/mean equal to 1.0. The aggregation rate was high in all cases, with a maximum at 72 h, and the differences were statistically significant between 24 and 216 h. The mean index and its standard error were similar for 24 and 216 h, but the variation coefficient was smaller at 216 h than at 24 h.

Time in hour Adults emerged (X ± SE) Number of occupied tubes (X ± SE) % Progeny in the preferred tubes Aggregation index (X ± SE) 24 23.2 ± 5.4 5.4 ± 1.3 61.2 9.85 ± 3.8 72 89.1 ± 4.7 13.2 ± 1.1 46.6 22.70 ± 3.6 216 548.1 ± 38.0 24.9 ± 0.1 10.7 8.44 ± 1.6

Time in hour	Adults emerged (X ± SE)	Number of occupied tubes (X ± SE)	% Progeny in the preferred tubes	Aggregation index (X ± SE)
24	23.2 ± 5.4	5.4 ± 1.3	61.2	9.85 ± 3.8
72	89.1 ± 4.7	13.2 ± 1.1	46.6	22.70 ± 3.6
216	548.1 ± 38.0	24.9 ± 0.1	10.7	8.44 ± 1.6

Table I - Mean number of Drosophila melanogaster adults emerged, occupied tubes, frequency in the preferred tubes and eggs oviposition aggregation index of groups of six females maintained in oviposition cages for 24, 72 and 216 h.

The differences between the means of the number of progeny among the different experiments were large (see Column 2, Table I). The permanence times of the groups were in a 1:3:9 relation and as each was under a daily cycle (12-h of light/12-h of darkness) one would expect that the progeny means would be proportionally related to the number of permanence cycles in each population cage. Taking as a basis the averages obtained at 24 h, we would expect averages of around 70 individuals at 72 h and 209 individuals at 216 h. These values were exceeded considerably. The chi-square test for this comparison was highly significant (c² = 5.527 and 551.362; 1 d.f. and P < 0.05). A plausible explanation would be synergy of the ovipositor facilitating and stimulating processes, such as modifications of the texture of the medium produced by the locomotive activity of the adults, insertion of eggs and excavating activity of the larvae (del Solar and Palomino 1966; Sokolowsky, 1980; Atkinson, 1983; Godoy-Herrera, 1986), plus the effect produced on oviposition and larval development by adult and preadult metabolic residues (Weisbrot, 1966; Budnik and Brncic, 1975).

This increase in the number of progeny was not associated with a proportional increase of the mean of occupied tubes due to the fact that the above-mentioned factors stimulate female fertility but do not necessarily influence the distribution among the tubes. A distinction must be made between number of eggs and their distribution over discrete areas. In fact the same number of eggs can have very different spatial orderings, which shows the importance of animal behavior over the spatial distribution pattern exhibited by the group. This procedure also permits us to know the behavior of oviposition for each individual female within the group by the simple means of following up the tubes where her progeny was disseminated per cycle of daily activity. Thus it was possible to establish that in 24 h, 20 females used one tube, 11 females used two tubes, and there was one female which used up to seven different tubes for disseminating her progeny. Within the 72 h, 24 females used only one tube, and 5 females used between 10 and 14 different tubes. Within 216 h, 11 females used one tube and 18 females used between 20 and 23 different tubes to disseminate their progeny.

These results are consistent with those found previously by del Solar and Ruiz (1992) who also classified the gregarious behavior of single (isolated) females.

As the time of permanence of the flies was extended in the population cage, the number of tubes with two or more phenotypes increased (Table II), which is a consequence of the dynamics of occupation of the areas available in the cage and the following diminishing of the probabilities of finding empty places for oviposition.

No. of phenotypes 24 h 72 h 216 h N % N % N % 1 28 51.8 52 39.4 19 7.6 2 16 29.6 45 34.1 26 10.4 3 5 9.2 18 13.6 63 25.3 4 2 3.7 7 5.3 76 30.5 5 3 5.5 5 3.8 51 20.5 6 - - 5 3.8 14 5.6 Total 54 99.9 132 100.0 249 99.9

No. of phenotypes	24 h		72 h		216 h
	N	%	N	%	N	%
1	28	51.8	52	39.4	19	7.6
2	16	29.6	45	34.1	26	10.4
3	5	9.2	18	13.6	63	25.3
4	2	3.7	7	5.3	76	30.5
5	3	5.5	5	3.8	51	20.5
6	-	-	5	3.8	14	5.6
Total	54	99.9	132	100.0	249	99.9

TableII - Number of different phenotypes registered per oviposition tube after 24, 72, and 216 h.

The females with the B and L markers contributed with 31 and 27 per cent of the total progeny (Table III), while the females with C and Oce did it with 11.8 and 11.4 per cent, respectively, and those with W and Ki contributed with 9.7 and 7.9 per cent. These data give rise to at least two points of discussion: i) if these fertility differences among individuals can be considered "normal" for a given population, and ii) if there is any correspondence between fertility and the number of occupied tubes; that is to say, that the fertile females would be the ones that tend to disseminate their progeny in a larger number of places. From the early studies of Pearl et al. (1927), a vast quantity of information about D. melanogaster fecundity and fertility has been gathered from which it can be concluded that variations of order of magnitude among individuals from the same population are frequent and consequently the differences observed in these experiments can be considered as a more realistic approximation to the understanding of small group dynamics, especially when they are composed arbitrarily and the variations of females carrying the same genetic marker are added (Miller, 1964; David and van Herreswege, 1969; Lints, 1971).

Phenotypes 24 h 72 h 216 h Total No. No. No. Progeny Tubes Progeny Tubes Progeny Tubes Progeny Tubes B 105 26 386 96 1549 184 2040 306 C 33 15 102 44 642 184 777 243 Oce 25 14 74 23 651 126 750 163 L 36 17 146 49 1641 204 1823 270 W 16 14 81 26 539 135 636 175 Ki 17 10 102 42 459 71 518 123

Phenotypes	24 h		72 h		216 h		Total
	No.		No.		No.
	Progeny	Tubes	Progeny	Tubes	Progeny	Tubes	Progeny	Tubes
B	105	26	386	96	1549	184	2040	306
C	33	15	102	44	642	184	777	243
Oce	25	14	74	23	651	126	750	163
L	36	17	146	49	1641	204	1823	270
W	16	14	81	26	539	135	636	175
Ki	17	10	102	42	459	71	518	123

Table III - Number of progeny per female of Drosophila melanogaster and number of oviposition tubes utilized out of a total of 375 registered after 24, 72 and 216 h in the oviposition cages.

B, Bar; C, Curley; Oce, Ocellusless; L, Lobe; W, Wrinkled, and Ki, Kinked markers.

Regarding the second question, to test the relationship between fertility and the tendency to disseminate the progeny among a larger number of tubes, we estimated Spearman’s Rho coefficient for females with markers B and Ki in the 24-h experiment because they corresponded to the highest and lowest fertility, respectively. The 24-h observation was used because it is presumed that the effects of the culture environment are smaller. The results were for B: Rho = 0.268, 8 d.f. and P > 0.05 non-significant and for Ki: Rho = 0.741, 8 d.f. and P < 0.05. The critical values for the Spearman’s Rho coefficient were 0.643 for P < 0.05 and 0.833 for P < 0.01. Due to these contradicting results, the analysis was extended to the six classes of females of the 24-h experiment, obtaining a value of Rho = 0.458 non-significant. Finally the same calculations were made for the 72 and 216 h, obtaining values of Rho = 0.943, 8 d.f. and P < 0.01 for 72 h and Rho = 0.600, 8 d.f. and P > 0.05 for 216 h. The same inconclusive results were obtained for the follow-up of the marked females in different experiments which makes us think that the quantity of progeny produced by each individual is independent of the way she distributes it in space. In any case, specifically designed experiments are required to solve this problem.

Follow-up of the tubes also permitted us to obtain information about the velocity of egg-adult development. The variance analysis (see Table IV) shows highly significant statistical differences between the experiments which are graphically illustrated in Figure 1. The emergence percentages according to the different time and of the different experiments are shown in Figure 1A. The results are clearer when the accumulated percentages for the emergence times of 50% of the total of registered adults are displayed in a graph. In round numbers we can say that 50% emerge in the 24-h experiment on day 13, in the 72-h experiment on day 15 and in the 216-h experiment on day 17 (Figure 1B). This is a complex problem which is frequently related to the quantity of available resources for larval growth and the perturbations (disturbances) of the rate of development are frequently interpreted as effects of the competition for space and food (Burdick, 1954; Barker, 1983).

Source d.f. Mean square F Between treatments 2 23.425 256.331 Between replicates 5 0.111 1.217 Residuals 10 0.091 - Total 17 2.842

Source	d.f.	Mean square	F
Between treatments	2	23.425	256.331
Between replicates	5	0.111	1.217
Residuals	10	0.091	-
Total	17	2.842

Table IV - ANOVA for Drosophila melanogaster developmental time in the three experiments.

Figure 1 - Developmental time of Drosophila melanogaster in the experiments. A, Percentage of adult emergence vs. time in days. Experiment l (); Experiment II () and Experiment III (). B, Accumulated percentage up to 50% of adults emerged.

Nevertheless this effect should be understood as a consequence of the aggregated distribution. These experiments showed that the progeny was dispersed in discrete tubes, with the same quantity and quality of resources, but each contained a variable number of preadults. If the average progeny per tube is calculated, the following results are obtained: 24 h = 2.41 individuals; 72 h = 3.18 individuals; and 216 h = 6.06 individuals. But when one considers the ranks among tubes, we have for 24 h = 1-40; for 72 h = 1-56; and for 216 h = 1-75. That is to say, out of the total number of the tubes registered in the 24-h experiment, only one contained 40 eggs or larvae; and, assuming this is the critical density to produce competition effects, only one of the 96 tubes registered would have it. As, additionally, of the 232 individuals which were born in this experiment, only 40 were submitted to competition, this would mean that 82.7% would not have had any effect attributable to this interaction.

Another point that must be mentioned is that the most common form of competition observed in Drosophila experiments are produced by interference among larvae rather than by limitation of resources. This is because the different activities performed by the larvae produce mechanical interactions which disturb feeding.

From the comparisons made among the number of eggs and larvae registered in the tubes and the number of adults emerged from them, we can make a gross estimation of 17.3% preadult mortality in the three experiments. Adding up all the cases which show a range of 13.5 to 18.1%, there are no significant differences among the replications of each experiment and among experiments. The chi-square value was c² = 1.008; 17 d.f. and P < 0.05.

Finally it is important to mention that the follow-up of the progeny of the 1280 tubes examined showed that there were practically no group of tubes with identical phenotypical composition. This fact suggests that the spatial fragmentation produced by the gregarious behavior could additionally have some effect on the genetic differentiation among groups.

Series of 6 tubes were chosen at random and were submitted to a heterogeneity test. The results showed significant statistical differences in all cases (Table V). Nevertheless, as the variation in the number among tubes and fertility differences among females with different phenotypical markers could affect any statistical test, the data were grouped according to the number of individuals per tube in classes of ten and were corrected for the progeny differences of each female class.

Mutant Tube Fertility index c 2 1 2 3 4 5 6 B 20 23 17 26 2 17 0.311 1.043 C 5 6 5 15 14 0 0.118 0.022 Oce 1 17 27 1 0 15 0.114 8.052 L 13 15 22 0 23 37 0.274 0.395 W 7 12 2 7 3 0 0.097 0.722 Ki 4 2 0 5 10 0 0.079 2.449 c 2 3.365 16788 48.640 22.096 37.103 24.052 0.997 12.683

Mutant	Tube						Fertility index	c 2
	1	2	3	4	5	6
B	20	23	17	26	2	17	0.311	1.043
C	5	6	5	15	14	0	0.118	0.022
Oce	1	17	27	1	0	15	0.114	8.052
L	13	15	22	0	23	37	0.274	0.395
W	7	12	2	7	3	0	0.097	0.722
Ki	4	2	0	5	10	0	0.079	2.449
c 2	3.365	16788	48.640	22.096	37.103	24.052	0.997	12.683

Table V - Heterogeneity test for a random sample of six oviposition tubes with more than 40 adults of Drosophila melanogaster.

For abbreviations see legend to Table III.

Thus the number of B females was multiplied by 0.311 while C females were multiplied by 0.118, since these were the proportional progeny contributions for B and C, respectively. Even with these corrections, most exhibited highly significant heterogeneity values.

Even though this is not an appropriate experimental design to measure pheno or genotypical differences among groups, these results suggest that spatial ordering is not a trivial component to explain the differences in number of individuals between two or more points of the geographical distribution of a population, and may contribute to the genetic variability observed among sub-areas.

ACKNOWLEDGMENTS

I thank Prof. Oriana González for her comments and translation of this manuscript and N. Kohler for her technical assistance and an anonymous reviewer who critically reviewed the manuscript. This work was supported by project S-94-39, DID-UACH.

RESUMO

O padrão espacial da prole produzida em caixas com 25 tubos de desova por grupo de 6 fêmeas de Drosophila melanogaster durante 24, 72 e 216 h foi estudado. Cada fêmea foi fertilizada por um macho portador de uma mutação dominante, de tal forma que o comportamento de cada fêmea do grupo, assim como o comportamento de cada grupo, puderam ser tratados pelo acompanhamento de sua prole. O padrão de distribuição espacial da prole foi agregado, alcançando seu valor máximo em 72 h. Em indivíduos gregários, o comportamento foi definido pelo número de locais nos quais eles disseminaram sua prole em ciclos de 24 h, enquanto que, para grupos, foi estimado pela taxa média de variância. A fertilidade, o tempo de desenvolvimento ovo-adulto e a taxa de mortalidade foram registrados. As comparações entre os tubos mostraram diferenças estatisticamente significantes no número de indivíduos e de composição fenotípica com distribuições gregárias que poderiam contribuir para a diferenciação genética local.

REFERENCES

Atkinson, W.D. (1983). Gregarious oviposition in D. melanogaster is explained by surface texture. Aust. J. Zool 31: 925-929.

Barker, J.S.F. (1983). Interspecific competition. In: The Genetics and Biology of Drosophila (Ashburner, M., Carson, H.L. and Thomposon, J.N., eds.). Vol. 3C. Academic Press, New York, pp. 285-341.

Budnik, M. and Brncic, D. (1975). Response of Drosophila pavani, Drosophila gaucha and their hybrids to larval biotic residues. Experientia 31: 781-782.

Burdick, A.B. (1954). New medium of reproductive quality stable at room temperature. Dros. Inf. Serv. 28: 170.

David, J. and Van Herreswege, J. (1969). Action repulsive de la levure vivante sur l’ ovipositions de D. melanogaster. C. R. Acad. Sci. 268: 1778-1780.

del Solar, E. (1968). Selection for and against gregariousness in the choice of oviposition sites by D. pseudoobscura. Genetics 58: 275-282.

del Solar, E. and Palomino, H. (1966). Choice of oviposition sites in D. melanogaster. Am. Nat. 100: 127-133.

del Solar, E. and Ruiz, G. (1992). Behavioral analysis of the choice of oviposition site by single females of D. melanogaster (Diptera: Drosophilidae). Insect Behav. 5: 571-581.

Forney, K.A. and Gilpin, M.E. (1989). Spatial structure and population extinction: A study with Drosophila flies. Conserv. Biol. 3: 45-51.

Godoy-Herrera, R. (1986). The development and genetics of digging behavior in Drosophila larvae. Heredity 56: 33-41.

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

Lindsley, D. and Zimm, G. (1985). The genome of Drosophila melanogaster. Dros. Inf. Serv. 62: 1-24.

Lindsley, D. and Zimm, G. (1990). The genome of Drosophila melanogaster. Dros. Inf. Serv. 68: 1-382.

Lints, F.A. (1971). Life span in Drosophila. Gerentologia 17: 33-51.

Messina, F.J. (1989). Host-plant variables influencing the spatial distribution of a frugivorous fly, Rhagoletis indifferens. Entomol. Exp. Appl. 50: 287-294.

Miller, R.S. (1964). Interspecies competition in laboratory population of D. melanogaster and D. simulans. Am. Nat. 98: 221-238.

Milne, B.T. (1992). Spatial aggregation and neutral models in fractal landscapes. Am. Nat. 139: 32-57.

Pearl, R., Miner, J.R. and Parkers, S.L. (1927). Experimental studies on the duration of life XI. Density of populations and life duration in Drosophila. Am. Nat. 61: 289-318.

Pielou, E.C. (1969). Introduction to Mathematical Ecology. Wiley Interscience, New York, pp. 286.

Poole, R.W. (1974). An Introduction to Mathematical Ecology. John C. Wiley and Sons, New York, pp. 532.

Ruiz, G. and del Solar, E. (1986). Effect of selection on oviposition site preference in D. melanogaster. Austral. J. Biol. Sci. 39: 155-160.

Ruiz, G. and del Solar, E. (1992). A diallel analysis of the gregarious oviposition in D. melanogaster. Heredity 70: 281-284.

Sokolowski, M.B. (1980). Foraging strategies of Drosophila melanogaster. Behav. Genet. 10: 291-302.

Taylor, C.E. and Powell, J.R. (1983). Population structure of Drosophila: Genetics and Ecology. In: The Genetics and Biology of Drosophila (Ashburner, M., Carson, H.L. and Thompson, J.N., eds.). Vol. 3D. Academic Press, New York, pp. 29-54.

Weisbrot, D.R. (1966). Genotypic interaction among competing strains and species of Drosophila. Genetics 53: 427-435.

(Received June 21, 1995)

Publication Dates

• Publication in this collection
  06 Jan 1999
• Date of issue
  Mar 1998

History

• Received
  21 June 1995