Jos~ M. Serrano, 1 Laureano Castro, 1 Miguel A. Toro, and Carlos. L6pez-Fanjul 2,3. Received 21 November 1990--Final 4 June 1991. The rate of homosexual ...
Behavior Genetics, Vol. 21, No. 6, 1991
The Genetic Properties of Homosexual Copulation Behavior in Tribolium c a s t a n e u m : Diallel Analysis Jos~ M. Serrano, 1 Laureano Castro, 1 Miguel A. Toro, ~ and Carlos L6pez-Fanjul 2,3
Received 21 November 1990--Final 4 June 1991
The rate of homosexual copulation has been defined as the ratio between the number of homosexual mountings and the total number of mountings (homo and heterosexual) performed by a Tribolium castaneum male during a period of 30 min. In a laboratory population, the average rate when a number of males (m) and females (k • m) are tested together has been estimated in each of the six situations defined by m = 2 and 10 and k = O.5, 1, and 2, k being the sex ratio among scored individuals. Good agreement was found between the observed rates of homosexual copulation and those expected assuming random contacts between pairs of individuals totally indiscriminate with respect to sex. The genetic properties of the trait have been investigated by means of a diallel analysis of six highly inbred lines derived from the same population and their F1 crosses. Significant general and specific combining ability effects were detected. When noninbred females were used for testing, the rate of homosexual copulation is expected to be higher for inbred than for noninbred males. This prediction, implying the existence of inbreeding depression for the trait, also was confirmed by the data. KEY WORDS" homosexual copulation; diallel analysis;
Tribolium castaneum.
x Departamento de Producci6n Animal, Instituto Nacional de Investigaciones Agrarias, Apartado 8111, 28080 Madrid, Spain. 2 Departarnento de Gen6tica, Facultad de Ciencias Biol6gicas, Universidad Complutense, 28040 Madrid, Spain. 3 To whom correspondence should be addressed. 547 o001-8244/91/110o-0547506.50/0 9 1991 Plenum Publishing Corporation
548
Serrano, Castro, Toro, and I~pez-Fanju|
INTRODUCTION Male insects often show a low threshold of sexual discrimination, copulating responses being easily induced even by exposure to artificial models mimicking conspecific females very crudely. Homosexual attempts at mating have been frequently reported for males of many species (for a review see Thornhill and Alcock, 1983). Usually, these have been simply considered as perception errors. However, the existence of some degree of discrimination generally has been assumed rather than directly i n v e s ~ t e d , even in those cases in which the alternative hypothesis appeared plausible. For example (Aiken, 1981), Palmacorixa nana males tend to mount individuals larger than themselves, thus allegedly increasing their chance of mounting a female. Nevertheless, about half of the total number of mountings scored by Aiken were of the homosexual type, this being the expected frequency with complete indiscrimination and a large number of individuals observed (see model below). In other instances, homosexual interactions have been interpreted as a by-product of male aggression or attributed to the absence of females (Mathieu, 1967). Quantitative genetic studies of homosexual copulation in populatior~s la;ave not been undertakefi. Current work is restricted to estimating the effects of allelic substitutions on the capacity of immature Drosophila melanogaster males to elicit a sexual response from mature males (Galley et aL, 1985). We attempt to establish the causes underlying homosexual copulation behavior in the red flour beetle Tribolium castaneurn based on a simple probabilistic model. We also have investigated the genetic properties of this trait in a population by means of a diallel analysis. Tribolium has the advantage of showing a comparatively simple sexual behavior and can be handled easily in the laboratory. Moreover, it has been the subject of considerable quantitative genetics research, including some aspects ~of mating behavior (Boake, 1986; Graur and Wool, 1982). Since explanations of insect homosexual copulation are likely to be species dependent, some background information on the sexual behavior of 7". ~astaneum is given hereafter [a comprehensive review is given by Sokoloff (1974)]. T. castaneurn is a cosmopolitan species that occurs primarily as a pest wherever cereal products are stored. This habitat makes visual or contact perception of a mate difficult, even though population densities can be relatively high. Moreover, the species has no obviously specialized tympanal organs and no evidence of sound communication has been reported. Olfaction appears then as practically the only means of establishing a relationship. There is a male-secreted aggregation pheromone
Homosexual Copulation in Tribolium
549
identified as 4R, 8R-(-)-dimethyldecanal (Levinson and Mori, 1983), strongly attractive to both sexes and perceived from the first day after emergence (Faustini et al., 1981). Populations of different geographical origin have been shown to differ in their sensitivity to this pheromone, suggesting the existence of genetic variation for this trait (Boake and Wade, 1984). Behavioral observations in the closely related species T. confusum have been interpreted tentatively as evidence for a femalesecreted sex pheromone, acting as a short-range attractant to the male only (Ryan and O'CeaUachain, 1976). However, the secretion has not been isolated so far, and therefore, its existence is purely speculative. The external adult appearance of the two sexes is nearly identical, although females are, on average, larger and heavier than males. Males can be identified only by (Sokoloff, 1974) (1) a prothoracie femoral setiferous patch, (2) a sclerotized projection at the posterior end of the seventh abdominal sternite (intromission by the copulating male is possible only if the female relaxes this sternite), and (3) differences in the strial arrangement in the elytra. In addition, no preliminary cour~hip has been described. These observations suggest that sex recognition ~ the males will not be easy, and apparently, they are sexually indiscriminate, attempting copulation even with objects vaguely resembling a beetle. Thus, the frequency Of homosexual mountings was calculated as nearly one-third in laboratory populations (Wool, 1967) and oscillated from 66 to 88% in three stocks marked with different melanic mutations (Rich, 1972, 1989). MODEL
A simple model will provide the framework within which our data can be discussed. It takes into account not only the behavio~ of the mounting male but also possible differential reactions of the individuals which are being mounted, traditionally regarded as passive partners. Consider in males and f = km females within an enclosure, k being the sex ratio among scored individuals. Assume that contacts take place between only two individuals at a time and that they occur at random with respect to sex. Thus, the number of possible contacts will be (~) between males and mfbetween a male and a female. These contacts will result in mounting with probabilities x and y, respectively. As females do not mount, the probability of homosexual mountings given that a mounting takes place COn) can be expressed as
550
Serrano, Castro, Toro, and L6pez-Fanjul
where ~x stands for the probability ratio y/x. Therefore, Pm will vary depending both on the experimental conditions under which observations have been made (specified by the values of m and k) and on the particular properties of the population being tested(defined by the value of the ratio a). When rn is large and k = 1, Pm tends to 1/(1 + 2 a ) . Theoretically, Pm can vary from 0, with strict heterosexuality (x = 0), to values approaching 1. Once the experimental conditions have been defined, p,, is only a function of a. Ideally, in a contact between two males both can mount and be mounted, but in a heterosexual contact only the male can mount. Therefore, a = 0.5 if both sexes are totally indiscriminate and oppose being mounted equally (x = 2y). However, if males resist being mounted more effectively than females, x will be expected to be smaller than 2y (i.e., a > 0.5.). The reverse will be found (or < 0.5) if, experimentally, normal females are exposed to physically weak (inbred) males. Obviously, Pm can be estimated from data, and from this, the corresponding c~ can be calculated. The sampling errors (SE) of these two parameters are related by the approximate expression SE (or) = (rn - 1) SE (pm)/2km. MATERIALS AND M E T H O D S
The Consejo population was captured near Madrid and has been maintained in a cage in our laboratory since 1964 with no artificial selection and with minimum inbreeding. All beetles in this experiment were kept in the same dark incubator at 70% relative humidity and 33~ The culture medium consisted of 95% whole wheat flour and 5% dried brewer's yeast by weight, powdered. Pupae were sexed by examination of the genital lobes. Male pupae were always kept individually and female pupae stored at low densities, until tested (12-19 days after adult emergence). Handling was performed at room temperature under constant artificial illumination. Estimation of the Rate of Homosexual Copulation
The structure of the data corresponds to a complete block design for two values of the number of males tested (rn = 2 and 10), three values of the sex ratio among tested individuals (k = 0.5, 1, and 2), and 40 males scored per cell. Mating cannot be watched within the culture medium as beetles show a strong photonegative reaction. In this situation, our observations were made as follows: For rn = 2, two males (identified by a white or a yellow spot of
Homosexual Copulation in Tribolium
551
paint on the thorax) were placed in a glass vial (3-cm diameter) with a filter-paper floor, together with one (k = 0.5), two (k = 1), or four (k = 2) virgin females. For each male, the number of homosexual and heterosexual mountings was recorded for a period of 30 min on each of 4 consecutive days, the females being substituted by new ones each day. Fifteen vials, five corresponding to each k value, were observed simultaneously. For m = 10, petri dishes (7-cm diameter) were used instead of vials to maintain a similar density. Half the males within a dish were identified with a white and the other half with a yellow spot of paint. Three dishes, each corresponding to a different, k value (0.5, 1, and 2), were simultaneously observed for a period of 30 min on each of 4 consecutive days. Again, the females were replaced each day by new ones. In this case, only the total number of mountings and their type were recorded. The whole experiment was repeated four times. Therefore, 40 males were scored for each of the six combinations between rn and k values. Only those homosexual matings in which the mounting male showed symptoms of being sexually aroused (extension of aedagus) were recorded as such. However, heterosexual mountings were not distinguished from intromissions. Thus, the rate of homosexual copulation can be calculated for each male as the ratio between the number of homosexual mountings and the total number of mountings recorded (homo- and heterosexual). All females used in the tests were sampled from the Consejo base population.
Diallel Analysis Six Consejo lines inbred by regular brother • sister mating during 49 generations were crossed in all possible combinations (but no reciprocals). The average rate of homosexual copulation of all lines and F1 crosses was estimated in the (m = 2, k = 1) situation. A diallel analysis was performed on the data, according to the following model (Griffing, 1956):
xijk = u + gi + gy + sij + ( ~ eijk)/h, where u is the overall mean, gi and gj are the general combining ability effects for the ith and jth parents, sij is the special combining ability effect for the cross between the ith and the jth parents (s~ = syi is assumed), e~k is the error effect associated to the ijkth observation, and h is the harmonic mean of the number of observations per cell (7.6). Our
Serrano, Castro, Toro, and L 6 p e z - F a n j u l
552
inbred lines were the only survivors of an initial sample of 60 lines started from the Consejo population, and therefore, they can hardly b e considered a random sample. Thus, a fixed model analysis was used, where the significance of general and specific combining ability effects were both tested against the error mean square. RESULTS Rate of Homosexual Copulation The average rate of homosexual copulation estimated in each of the six experimental situations studied, pooled over replicates and observation periods, is shown in Table I, together with their pertinent expected values for complete sexual indiscrimination and both sexes opposing being mounted equally (a = 0.5). No significant differences between replicates within treatments were found. Significant differences between observed and predicted rates were detected in a single instance, corresponding to the largest values of m and k considered. Experimentally, this case proved to be that in which departures from the model assumptions were greatest. In petri dishes, as opposed to vials, the males' tendency was to aggregate in some central position, and the females' to disperse centrifugally. Therefore, contacts between pairs of males were more common than expected from random movements, and conversely, heterosexual contacts were less frequent. This is in agreement with observed rates of homosexual copulation being generally greater than expected, discrepancies being higher for increasing values of m and k. The average rate of homosexual copulation of the 40 males observed in each experimental situation is shown in Table II for each of the four observation periods. Significant between-period heterogeneity was found
Table I. Observed (0) Rates of Homosexual Copulation (%) and Their Expected (E) Value Under Total Sexual Indiscrimination" rn
k
0
E
X21
M
2
0.5 1 2 0.5 1 2
48.4 __-2.3 3 1 . 6 • 1.9 21.8.4-1.7 66.4__- 1.8 50.6 --- 1.8 38.2 • 1.4
50 33.3 20 64.3 47.4 31
0.5 0.8 1.2 1.3 3.4 28.2*
3.07 3.68 3.64 4.11 5.04 7.24
10
rn males and km females tested together. M ---- average number of mountings (homo and heterosexual) performed b y a male during 30 rain, Means pooled over tests, * p < .01.
Homosexual Copulation in Tribolium
553
T a b l e II. Rates of Homosexual Copulation (%) Estimated on Each of 4 Consecutive Days a Day m 2 10
k
1
2
3
4
X23
0.5 1 2 0.5 1 2
48.6 35.7 25.3 67.1 47.2 45.5
48.8 37.7 20.1 70.2 56.0 34.6
48.3 25.5 19.7 63.3 46.3 35.2
47.7 27.1 21.9 63.6 53.6 35.0
0.1 7.5 1.7 2.2 3.8 11.8"
m males and/on females tested together. * p < .01.
only once, and it could be attributed to a single value, widely departing from the remaining three, the later being very close both to each other and to the expectation.
Sexual Activity The average number of mountings ( h o m o - a n d heterosexual) performed by a male during 30 min, pooled over tests, is presented in Table I for all situations. For a given number of males, the average number of mountings per unit of time generally rose with the number of females present. This number also increased for m = 2 to m = 10. These results can bc ascribed to an increasing number of interactions between pairs of individuals, not necessarily related to their sex. A priori, the assumption of independence of the sexual behavior of the males tested together may be suspect, particularly so when their number is small (m = 2). Nevertheless, this hypothesis has been empirically supported by the data. Correlations (Table III) between the scores of the two males tested together in the same vial have been calculated for (a) number of homosexual mountings, (b) number of heterosexual mountings, (c) total number of mountings, and (d) rate of homosexual copulation. Likewise, the correlations between (e) the number of heterosexual mountings by a male and that of homosexual mountings by the other and between (f) the total number of mountings by a male and the rate of homosexual copulation of the other also have been estimated. In general, these correlations between aspects of sexual behavior of the two males tested together in the same vial were small and nonsignificant. Only in the extreme case in which a single female was available (k = 0.5) were the correlations between the number of heterosexual mountings
554
Serrano, Castro, Toro, and L6pez-Fanjul
Table HI. PhenotypicCorrelationsBetweenthe Performanceof Two Males Tested
Together with 2k Femalesa k
Ho:Ho
H~:Ho
T:T
p2:p2
Ho:Ho
T:p2
0.5 1 2
0.06 -- 0.22 0.31
- 0.50* 0.33 --0.06
- 0.44* O.17 --0.16
0.03 0.09 0.13
- 0.22 -- 0.02 -0.17
0.25 - O.15 --0.01
Ho, numberof homosexualmountingsper male; H~, number of heterosexualmountings per male; T, total number of mountings per male; P2, rate of homosexualcopulation per male. * p
