Second chromosome polymorphism of Drosophila buzzatii in a natural population is not associated with gametic selection and does not affect mating pattern

In previous work we have demonstrated the adaptive significance of the chromosomal polymorphism of D buzzatii in a natural population from Argentina by means of selection components analysis. The experimental design employed for this purpose was based on 4 general assumptions. The results reported in the present study allow us to verify 3 of these assumptions: i) mating is at random, irrespective of the flies karyotype; ii) gametic selection may be considered absent; and iii) there is no differential attraction of flies to the collecting baits. The validity of these assumptions gives strong support to the conclusions reached in our previous selection components analysis. mating pattern / gametic selection / chromosomal polymorphism Résumé-Le polymorphisme du deuxième chromosome de Drosophila buzzatü dans une population naturelle n'est pas associé à une sélection gamétique et n'affecte pas le mode d'accouplement. Dans un travail précédent, nous avons démontré la signification adaptative du polymorphisme chromosomique de D buzzatii dans une population naturelle d'Argentine, à l'aide d'une analyse de la sélection en ses composantes. Le plan d'expérience utilisé dans ce but était basé sur quatre hypothèses. Les résultats présentés dans le présent travail permettent la vérification de trois d'entre elles: i) l'accouplement se produit au hasard, indépendamment du karyotype des mouches; ii) la sélection gamétique est absente; iii) il n'y a pas d'attraction différentielle des mouches vers les pièges de récolte. La * Correspondence and reprints vérification de ces hypothèses confirme les conclusions établies dans notre précédente analyse des composantes de la sélection. mode d'accouplement / sélection gamétique / polymorphisme chromosomique / drosophile INTRODUCTION

(Received 31 January 1991; accepted 13 January 1992) Summary -In previous work we have demonstrated the adaptive significance of the chromosomal polymorphism of D buzzatii in a natural population from Argentina by means of selection components analysis. The experimental design employed for this purpose was based on 4 general assumptions. The results reported in the present study allow us to verify 3 of these assumptions: i) mating is at random, irrespective of the flies karyotype; ii) gametic selection may be considered absent; and iii) there is no differential attraction of flies to the collecting baits. The validity of these assumptions gives strong support to the conclusions reached in our previous selection components analysis. mating pattern / gametic selection / chromosomal polymorphism Résumé -Le polymorphisme du deuxième chromosome de Drosophila buzzatü dans une population naturelle n'est pas associé à une sélection gamétique et n'affecte pas le mode d'accouplement. Dans un travail précédent, nous avons démontré la signification adaptative du polymorphisme chromosomique de D buzzatii dans une population naturelle d'Argentine, à l'aide d'une analyse de la sélection en ses composantes. Le plan d'expérience utilisé dans ce but était basé sur quatre hypothèses. Les résultats présentés dans le présent travail permettent la vérification de trois d'entre elles: i) l'accouplement se produit au hasard, indépendamment du karyotype des mouches; ii) la sélection gamétique est absente; iii) il n'y a pas d'attraction différentielle des mouches vers les pièges de récolte. La INTRODUCTION Several methodological approaches have been applied to the study of natural selection in nature (see Endler, 1986 for a review). One of them consists of dividing total selection into partial components throughout the life-cycle (selection components analysis) (Hedrick and Murray, 1983). This approach was employed to investigate the adaptive significance of chromosomal polymorphisms in wild populations of the cactophilic fly Drosophila buzzatii Hasson et al, 1991). These authors demonstrated not only the adaptive significance of the second chromosomal polymorphisms, but also that certain chromosomal arrangements are favoured during one selective episode and selected against in another, a pattern called endocyclic selection. Moreover, they suggested that this pattern might be involved in the maintenance of polymorphisms. However, their conclusions are based on the validity of 4 assumptions; i), random mating; ii), absence of gametic selection; iii), lack of differential attraction of flies to the collecting baits; and iv), lack of selection in the laboratory (see Ruiz et al, 1986).
In the present work we report the results of the analysis of the first 3 assumptions, with the aim of providing experimental support for our previous results.

MATERIAL AND METHODS
D buxzatii is a cactophilic species that breeds and feeds on the decaying cladodes of several Opuntia species. The present work was performed in the natural population of Arroyo Escobar, where D buzxatii breeds on the rotting cladodes of Opuntia vulgaris (for a description of this locality, see Hasson et al, 1991). Wild flies were caught in May 1987 by net sweeping on fermented banana baits. Flies were immediately sexed and females were maintained in individual vials with David's (1962) killed yeast culture medium at 25°C. The analyses of 8 larvae from the progeny of each wild inseminated female allowed us to assess the karyotype for the second chromosome of both parents assuming that the progeny of each female was fathered by a single male. Three chromosome arrangements were identified: st (standard), 2j and 2jz 3 . Observed mating frequencies were compared with those expected by random mating by means of x 2 tests.
Goodness of fit to Mendelian proportions was studied by means of X 2 tests for each mating class. Since the observed karyotypic proportions in some mating classes departed from the expected (see below), the segregation of 2j/st and 2j Z3 / st heterokaryotypes was studied in more detail. This analysis was performed with heterozygous individuals obtained by crossing flies from homozygous stocks.
These stocks were obtained through consanguineous crosses among the progeny of several wild females. Seventeen lines homozygous for 2j, 14 for 2jz 3 and 4 for 2st were isolated, and individually maintained until employed in further experiments. Heterokaxyotypes were produced through both reciprocal crosses by setting together 100 mature virgin females and 100 mature males of the corresponding homozygous stocks in an egg collecting chamber. Heterozygous eggs were seeded in culture bottles in uncrowded conditions (6-8 eggs per ml of culture medium).
One hundred heterokaryotypic 5-day-old flies were crossed with 100 flies of each of the 3 homokaryotypic stocks. Flies were introduced into an egg collecting chamber for a 48 h period. The collecting medium was changed every 24 h. In all cases, both reciprocal crosses were made.
Eggs were allowed to hatch and 500 first instar larvae were seeded in 5 culture bottles (100 larvae each) containing 40 cc of culture medium for each cross, in low density conditions and maintained at 25°C. Thirty-two third instar larvae from each bottle were sampled and their salivary gland chromosomes analyzed, according to Fontdevila et al (1981). Observed karyotypic proportions in the offspring of each cross were compared to those expected by means of goodness of fit X 2 tests. The remaining individuals in the bottles were maintained until pupation in order to estimate larval viability by counting all pupae.
Further collections were performed in A Escobar in March 1989 by means of banana baits and with Opuntia vulgaris rotting cladodes, in order to test the assumption that flies were not differentially attracted to the collecting baits.

RESULTS
The frequencies estimated for second chromosome arrangements were: standard (st) = 0.110; j = 0.575 and jz 3 = 0.315. The observed and random mating expected numbers of each mating class are shown in table I. Since the frequencies of certain mating classes were extremely low, goodness of fit tests were performed for each chromosomal arrangement. None of the X 2 values was significant (table II).
Additionally, Mendelian segregation was tested for the most common mating classes. The results shown in table III revealed significant departures in crosses j/st x j/j; i l st x j/jz 3 and jz 3 /st x j/j. From this analysis it may be concluded that a deficiency of 2st bearing individuals occurred in all cases. These results could suggest gametic selection against 2st. To test this hypothesis, flies heterozygous for 2st were backcrossed with individuals from the 3 homokaryotypic stocks. Significant departures from expected values were detected only in 2 independent crosses of 2j/st males with 2st/st females (table IV, fourth row) with homogeneous results (x 2 = 0.09, df = 1P > 0.75). This bias cannot be attributed to selection in the laboratory because there was no relationship between viability (table IV, fifth column) and segregation distortion for each cross. The results of the reciprocal cross did not depart from the expected values and when pooling the progeny that received 2st versus 2j in all crosses involving the 2st/j heterokaryotype (from table IV), the x 2 value was not significant (x2 = 3.13, df = 1, P > 0.05). Likewise, none of the crosses involving the 2st fj z 3 parents showed a significant distortion in the segregation pattern. However, all crosses involving heterozygous 2st/* parental males showed a homogeneous trend ( X 2 = 5.03, df = 5, P > 0.25) of 2st deficiency, and when the progeny of these crosses was pooled according to the arrangement received from their parental males: st(N = 413) and non st (N = 488) a significant distortion from 1:1 was observed ( X 2 = 6.24, df = 1, P < 0.025). This effect was absent in the reciprocal crosses (st/* parental females) (419:419) (table IV).
The validity of the assumption that flies were not differentially attracted to the collecting baits was tested in March 1989. Chromosomal frequencies estimated in samples of adult flies collected by means of banana baits and with rotting cladodes were not significantly different ( X 2 = 5.57, df = 2, P > 0.05, N = 189), pointing to the fact that flies were attracted to the baits irrespective of their karyotype.

DISCUSSION
Random mating is a general assumption in almost all models devised for population genetics purposes. In the D buzzatii natural population of Arroyo Escobar, observed mating frequencies of second chromosome karyotypes did not depart significantly from those expected, giving support to the assumption of random mating. This result is in agreement with those obtained by Ruiz et al (1991) in a Spanish natural population of D buzzatii.
However, our analysis is based on the assumption that the offspring of each wild inseminated female was fathered by a single male. Concerning this, female remating has often been observed in several species of Drosophila (Loukas et al, 1981;Markow, 1982). The frequency of multiple inseminations is dependent on environmental conditions such as population density (Turner and Anderson, 1983) and again, the assumption of absence of gametic selection seems to be acceptable.
Thus, the results reported herein give support to the conclusions reached in our previous paper. They demonstrate that the assumptions of lack of differential attraction of flies to the collecting baits and random mating can be accepted, and even the apparent weak gametic selection against 2st would not affect our previous results.