Whose Sperm Fertilize the Female's Eggs if She Mates More Than Once?

KAREN S. OBERHAUSER, RACHEL HAMPTON, BRENDA JENSEN
Department of Ecology, Evolution and Behavior
AND
SANFORD WEISBERG
Department of Applied Statistics
University of Minnesota, St. Paul MN 55108

Summary  |  Introduction  |  Methods  |  Results  |  Discussion   |  Acknowledgments  |  References   |  Karen's Research Questions

Summary

Sperm precedence, or nonrandom fertilization success among a female’s mates, is often an important determinant of male fitness. We used an allozyme marker to study how sperm precedence in monarch butterflies varies with male mating history and time after a female’s second mating. We found strong second-male precedence, with large males and previously-mated second males tending to fertilize more eggs. P₂ (the proportion of eggs fertilized by the second male) changed with time in a way that was affected by the first male’s mating history; second male precedence tended to erode or remain constant if the first male had not mated, but increased if the first male had mated previously. The advantage of previously-mated second males, which transfer fewer sperm than unmated males, could be due to differences in the competitive ability of sperm, or to ‘cryptic female choice’.

Introduction

Sperm precedence, or nonrandom differential fertilization success among males, can be an important component of sexual selection in species in which females mate more than once (Parker 1970). There are several potential mechanisms for sperm precedence. It could be due to a numerical advantage, if sperm competition is like a raffle (Parker 1970). It could also occur if there is differential fertilizing capacity resulting from variation in sperm motility, longevity, or ability to penetrate the egg (Lanier et al. 1979). Effects of sperm numbers and differential fertilizing capacity are analogous to male-male competition. Cryptic female choice, or biased sperm use from preferred males, could also affect sperm precedence (e.g. Walker 1980; Thornhill 1983; Simmons 1987; Eberhard 1991, 1996; Birkhead & Mf ller 1993; Eberhard & Cordero 1995). Studies of the effects of sperm numbers and male traits carried out over long periods of oviposition will help to distinguish these mechanisms. Here we report one such study in monarch butterflies.

Mating male monarchs transfer a spermatophore (see spermatophores) into the female’s bursa copulatrix. A sperm ampule is embedded in the distal end of the spermatophore. Sperm are transferred last, and shortly after mating ends they begin to move through the ductus seminalis, and into the spermatheca (see sperm movement). The lepidopteran spermatheca is elongate, with sperm entering and leaving through the same end; this makes it likely that sperm from the last male to mate dominate the ‘fertilization set’ of sperm most likely to fertilize eggs (Parker et al. 1990). As in other Lepidoptera, monarchs produce both eupyrene (nucleated) and apyrene (unnucleated) sperm (see sperm production). Apyrene sperm could be important in sperm competition, but their function is not clearly understood (Silberglied et al. 1984); they may displace previous sperm from the fertilization set as they move into the spermatheca.

Methods

Experimental butterflies were second generation offspring of wild females. We determined adult genotypes soon after eclosion (see below) and set up matings to obtain experimental individuals with genotypes that would allow us to determine paternity in doubly-mated females. Larvae were reared on common milkweed under ambient summertime photoperiod and temperature. They were weighed to the nearest 0.01 mg on the day after eclosion, and fed a 20% honey solution daily.

We assigned females to mating treatments in which the amount of sperm transferred by two males varied: matings to a) two mated males (M, M), b) a mated male then an unmated male (M, UM), c) an unmated male then a mated male (UM, M), or d) two unmated males (UM, UM). All mated males had mated the day before they were needed; these males transfer a smaller volume of sperm than unmated males (see below). Males were only used for one experimental mating, and were five to eleven days old. Experimental females mated for the first time in large outdoor flight cages at ages four to ten days. Differences between the masses of a female’s two mates ranged from 2 to 162 mg.

The day after their first mating, we put females into separate outdoor oviposition cages with potted host plants. We collected approximately 30 eggs from each female during the first day of oviposition to determine their fertility and genotype. Beginning on the third day of oviposition, one or two males of the assigned second male mating history were put into each cage until the female remated. We counted and removed eggs from plants daily, saving 40 eggs from each female per day for genetic analysis. We maintained females until they had laid no eggs for seven days, could not fly, or died.

Females that laid no eggs within four days after mating, laid infertile eggs, or did not remate within seven days were excluded. Our final sample size was 38 females, plus one that remated within a few hours of the termination of her first mating. She was not included in treatment comparisons because she laid no eggs between matings, but we did analyze her eggs.

We used a volumetric measurement to estimate relative sperm numbers from unmated and mated males. We dissected spermatophores from females mated to males with these mating histories, removed sperm ampules from spermatophores, and measured their width and length to the nearest 0.01 mm under a dissecting microscope. We calculated the volume of a sphere with a diameter equal to the mean of the length and width. If the ampule ruptured, or if sperm had already begun to leave it, we discarded the sample.

We assigned paternity using the allozyme marker phosphoglucose isomerase (Pgi) and the electrophoretic methods and recipes of Hebert & Beaton (1989). To sample adult genotypes, we removed scales from the dorsal side of the abdomen with tape. We then cut a small incision in the sixth abdominal segment, drew one to five microliters of hemolymph into a micropipet, and transferred this hemolymph to a plexiglass well plate. Incisions healed quickly, and butterflies behaved normally after this procedure. All individuals had known genotypes at the Pgi locus (ss, ff or sf), with all females being homozygous for the slow allele. We did not have enough ff males to assign each female two mates of opposite genotypes, so first and second mates had genotypes of ss and ff (N = 10), ff and ss (N = 8), ss and sf (N = 11) and sf and ss (N=9). These genotype treatments were distributed evenly among mating treatments.

We allowed eggs to develop for four days, at which point fertile eggs had either just hatched or were easily distinguishable, and then held them at -75^o C. We homogenized eggs in ten microliters of buffer, and applied this mixture to the well plate.

We determined P₂ values (proportion of eggs fertilized by the second male) daily (daily values) and for the entire time females laid eggs after the second mating (overall values). When one male was sf, we assumed that the number of sf eggs was half the total fertilized by the sf male. We tested how mating history and size affected male success in sperm competition, including female intermating intervals, lifetime fecundity, and male genotypes in the models.

We analyzed sperm precedence using logistic models. These are more appropriate than models that use P₂ values themselves as experimental observations because they incorporate sample sizes of each value, essentially making each egg an individual observation. Using P₂ values themselves gives equal weight to values with different levels of confidence. We could not use standard binomial models to analyze the data, because between-female values were too variable. Instead, we used a logistic regression model for over-dispersion that allows additional variation not captured by a binomial assumption (Collett 1991).

Results

(a) Relative sperm numbers

Average sperm ampule volumes from unmated males and males mated one day previously were 11.44 mm³ (n = 6, s.d. = 4.02) and 4.91 mm³ (n = 5, s.d. = 0.73), respectively (t-test with unequal variances, p = 0.010), suggesting that unmated males transferred over twice as many sperm as mated males. It is possible that relative amounts of sperm and other materials in the ampules of these male types vary, but differences were not apparent when we looked at their contents; all were essentially full of sperm. It is also possible that the relative numbers of apyrene and eupyrene sperm varied. These two sperm types are easily distinguishable; we counted the number of both in diluted samples and detected no differences.

(b) Overall P₂ Values

We analyzed an average of 301 eggs per female (range 26 to 616). There was second-male precedence; 79% of the 38 females had overall P₂ values over 50% (figure 1). However, there were only six cases of complete precedence of one male. We excluded three cases of complete first-male precedence (P₂ = 0) from statistical analyses because of the possibility that these second matings were unsuccessful. Since we had excluded unsuccessful first matings, it seemed most reasonable to exclude these cases. Cases of complete second-male precedence were included because we had determined the success of all first matings. We did, however, do the analyses including the three 0 values, and none of our conclusions were affected.

Figure 1. Overall P₂ values separated by experimental mating treatment. All values above the line are second male precedence. All values, including those of total first male precedence, are included. Numbers next to points indicate that more than one female had that value.

The over-dispersion model assumes that a female’s true P₂, given treatment and other factors, is sampled from a distribution with mean p₂ and variance f p₂(1-p₂). This allows two females with the same treatment and other values of the predictors to have different true values of P₂. Using the method of Williams (1982), we estimated f = 0.28, indicating substantial over-dispersion. Given f, we calculated the logistic regression as a linear function of treatment and other predictors. Approximate likelihood ratio tests were used to assess significance of predictors of P₂.

The difference between the mass of the two males significantly reduced model deviance, with larger males tending to do better in sperm competition (table 1a & b). A quadratic effect of mass was insignificant. Mating treatment, male genotypes, intermating interval (which ranged from three to seven days), and total fecundity did not affect P₂, although there was a tendency for females in the treatments in which the second male had already mated to have higher P₂ values (figure 1).

All eggs from the female that remated immediately after her first mating were fertilized by her second mate (N = 460).

Table 1a. Analysis of deviance likelihood ratio tests for effects on overall P₂.

(Predictors with {F} in front of them have separate factors, or levels, which were included in the model as dummy variables. M1 and M2 represent the mass of the first and second males, respectively.)

Table 1b. Summary of overall binomial regression model, weighted by extra variance.

(c) Daily P₂ Values

We analyzed changes in P₂ over time using day/female combinations as units of analysis. Figure 2 shows the mean values for all of the females in each mating treatment, with regression lines for each treatment.  Repeated observations on each female are likely to be correlated, and it is appropriate to use an over-dispersion model similar to that used on overall P₂ values (Collett 1991). To estimate the over-dispersion parameter, we fit a separate linear logistic regression of P₂ on day, day², mass difference, and mass difference² for each female. This gave an overall estimate of 0.43 for f, which we used to fit the logistic regression of P₂ on the above factors plus treatments, and treatment/time interactions (table 1c & d). Again, the difference between the masses of the two males significantly reduced model deviance. The significance of the quadratic of the mass effect suggests that the magnitude of the mass difference matters, not just which male is larger. There are treatment differences, and a significant treatment/day interaction, which means that changes in P₂ depend on treatment. Time trends are generally linear in the logit scale, since the quadratic terms in day are insignificant.

Figure 2. Daily P₂ Values. All of the eggs laid by all females in a given treatment are combined into single data points for ease of interpretation, but the analysis itself included separate values for each female. The curves on the graph show separate logistic regressions for each treatment.

Table 1c. Analysis of deviance likelihood ratio tests for effects on daily P₂.

Table 1d. Summary of daily binomial regression model, weighted by extra variance.

(Positive coefficients for treatments indicate second male precedence just after the second mating, and probabilities test whether intercepts are significantly different from 0.5, i.e. whether one male’s sperm are favored just after the second mating. Positive (or negative) coefficients for treatment/day interactions indicate that P₂ increases (or decreases) over time, and probabilities test whether the slopes shown in figure 3 are significantly different from 0. Coefficients followed by the same lowercase (treatments) or uppercase (treatment/day interactions) letters are not significantly different.)

Binomial regression coefficients (table 1d) show the direction and significance of these effects, and treatment by time effects are illustrated graphically in figure 2. Both treatments with a mated second male have relatively high P₂ values early, although initial values in the UM, UM treatment are not significantly lower than in these treatments. This suggests that mated second males fertilized a higher proportion of eggs for the first several days after mating, despite transferring fewer sperm. The slope coefficient for the time/treatment interaction is significantly negative for the UM, UM treatment, and positive for the M, M treatment. Slopes in the M, UM and UM, M treatments are not significantly different from zero, but their signs and magnitudes suggest that P₂ decreases or stays relatively constant when the first male is a unmated male, and increases when the first male is a mated male.

Complete P₂ histories for three females are shown in figure 3. The female in figure 3a had consistently high second-male precedence throughout her life (figure 3a). The female in figure 3b had an erratic pattern, with strong second-, then first-, then second-male precedence. Figure 3c shows gradually changing P₂ values over time, a common pattern in our experimental females.

Figure 3. Complete P₂ data from three experimental females, with 95% confidence intervals for daily P₂ values, showing a) relatively constant, high P₂ values; b) erratic P₂ values; and c) gradually increasing values. We calculated confidence intervals based on the binomial distribution. For values based on 30 or fewer eggs, we used intervals from Blyth and Still (1983). When the two males shared an allele, we determined the confidence interval for P₂/2, which is based on the actual number of eggs that contained the unique allele. We then doubled the endpoints of this interval to obtain the confidence interval for P₂ itself.

Discussion

We found strong second-male sperm precedence in monarchs, a common pattern in Lepidoptera (Gwynne 1984, Drummond 1984). P₂ values ranged from 0 to 1, and the over-dispersion parameters indicate that much of this variation is not explained by factors that we manipulated or measured. We did see an advantage to large males and to second mates that had mated one day before the experimental mating, although the mated male advantage is dependent on time and the first male’s mating history. Changes in P₂ values show that the proportional representation of each male in the fertilization set changes over time. Here we discuss these patterns and potential mechanisms.

(a) Effects of male size and mating history

The only pattern apparent in both overall and daily P₂ data was the effect of male size; males larger than their rivals tended to fertilize more eggs. This is common in insects (Lewis & Austad 1990, Simmons & Parker 1992, LaMunyon & Eisner 1993, Gwynne & Snedden 1995), and could be a numerical effect if larger males transfer more sperm. However, the advantage of previously-mated males suggests that the number of sperm transferred is not the only determinant of sperm precedence in monarchs.

Preferential utilization of sperm from large males could act separately from a numerical effect. For example, female dung flies store sperm from larger males where they are probably most likely to be used to fertilize eggs (Ward 1993). LaMunyon & Eisner (1994) suggest that Utethesia ornatrix moth females selectively use sperm from males that transfer larger spermatophores, and spermatophore size correlates with unmated male size. A large male advantage could also be due to differential fertilizing capacity (Lanier et al. 1979). A positive correlation between sperm length and the degree of polyandry in mammals (Gomendio & Roldan 1991) and butterflies (Gage 1994) suggests that sperm competition selects for longer sperm. In mammals, longer sperm swim faster (Gomendio & Roldan 1991), and faster sperm may be more likely to fertilize eggs (Birkhead et al. 1995). While we are not aware of studies that show within-species correlations between male size and sperm size, this correlation exists across butterfly species (Gage 1994).

Mated second males tended to enjoy higher early P₂ values, despite transferring fewer sperm. This effect was not statistically significant when overall P₂ values were analyzed, but was when the effects of time were considered (table 1c & d). A possible explanation is that mated males were not a random subset of males; they had already proven their ability to mate. Males that are successful at obtaining matings may produce sperm that are successful at obtaining fertilizations. A similar effect occurs in flour beetles; male olfactory attractiveness to females correlates with subsequent fertilization success (Lewis & Austad 1994).

Non-zero slopes on the graph of P₂ versus time indicate that sperm precedence patterns change over time. It appears that changes over time could be affected by sperm numbers; mated first males fertilized progressively fewer eggs, but when the first male was unmated, P₂ values either stayed relatively constant or decreased over time (figure 2). P₂ values in several other insects decrease over time, and the relative numbers of sperm from the two males probably affect the degree to which mixing occurs (Schlager 1960, McVey & Smittle 1984, Simmons 1987, Siva-Jothy & Tsubaki 1989, Martin et al. 1989).

There is no indication that females ran out of sperm from mated male males. We did not see steep changes in P₂ after long periods of oviposition in treatments that involved mated males, which is expected if sperm are limiting. Additionally, females mated to a single mated male lay fertile eggs throughout their lives (What factors affect the number of eggs that female monarchs lay?, Oberhauser 1989, 1997).

(b) Summary of Possible Mechanisms

Second male precedence suggests that sperm in the elongate spermatheca are displaced by incoming sperm, resulting in stratification of sperm from different males. However, this stratification is rarely complete; some sperm from the first male are used to fertilize eggs right after the second mating (e.g. figure 3a). Changes in P₂ over time suggest that the number of sperm transferred by the first male affects their proportional representation as mixing occurs. Sperm from the first male may mix with those of the second male in a diffusion process, causing slow changes in P₂ (e.g. figure 3c). However, changes from total precedence of one male to total precedence of the other in a single day do occur (e.g. figure 3b). This could result if sperm from individual males occur in clumps within the spermatheca. The advantages of unmated male first males over time, and possibly large males, suggest that relative sperm numbers play a role in determining sperm precedence patterns, especially as mixing occurs.

The advantage enjoyed by mated second males suggests that sperm from some males are more likely to fertilize eggs, regardless of their numerical representation. This could be due to differential fertilizing capacity if "good maters" produce more competitive sperm. Mechanisms that would result in differential fertilizing capacity are discussed elsewhere (e.g. Lanier et al. 1979, Sivinski 1984, Dewsbury 1984), and this study does not resolve their relative merit. At this point we can only say that our results are compatible with a mechanism of differential fertilizing capacity, and that sperm from males that are more likely to mate, and possibly large males, might be more competitive. Post-mating female choice may also have affected our results. Simmons (1987) suggested two potentially relevant mechanisms by which females could bias sperm use. 1) They could oviposit only after mating with a preferred male. When a female remates within a few hours of a previous mating, sperm from the first male are unlikely to have left the bursa. This probably occurred when one female in our study remated immediately after her first mating. 2) They may control sperm displacement (see also Villavaso 1975). Because of the long pathway from the site of sperm deposition to the lepidopteran sperm storage organ, females have substantial opportunity to influence sperm displacement. The reproductive tract undergoes rhythmic contractions as sperm enter the ductus seminalis, which probably propel the sperm (Drummond 1984). Females may also be able to sequester sperm from the first male after remating, thus keeping it in the fertilization set.

Differential fertilizing capacity and preferential movement of sperm by females are analogous to male-male competition and female choice, except that they act at the gametic rather than the organismal level. Distinguishing between them will be difficult (Willson 1990), but important. An understanding of sexual selection in many species requires knowledge of events that occur after mating, and continued study of intraspecific variation and time-dependent changes in P₂ will be useful in elucidating the mechanisms responsible for the species averages that were reported in much of the early literature on sperm competition.

Acknowledgements

We thank D. Cansler, A. Feitl and C. Jessup for assistance; and D. Alstad, C. LaMunyon, D. Gwynne, M. Gage and L. Simmons for comments. Research was supported by the NSF (DEB-9220829).

return to Karen's Research Questions

References

Birkhead, T.R., Fletcher, F., Pellatt, E.J. & Staples, A. 1995. Ejaculate quality and the success of extra-pair copulations in the zebra finch. Nature 377, 422-423.

Birkhead, T. R. & Mf ller, A. 1993. Female control of paternity. TREE 8, 100-104.

Blyth, C. R., and H. A. Still. 1983. Binomial confidence intervals. J. Amer. Stat. Ass. 78,108-116.

Collett, D. 1991. Modeling Binary Data. London: Chapman and Hall.

Dewsbury, D.A. 1984. Sperm competition in muriod rodents. In: Sperm Competition and the Evolution of Animal Mating Systems (ed. R.L. Smith), pp. 547-572. New York: Academic Press.

Drummond, B.A. 1984. Multiple mating and sperm competition in the Lepidoptera. In: Sperm Competition and the Evolution of Animal Mating Systems (ed. R.L. Smith), pp. 291-371. New York: Academic Press.

Eberhard, W.G. 1991. Copulatory courtship and cryptic female choice in insects. Biol. Review 66,1-31.

Eberhard, W.G. 1996. Female control: Sexual selection by cryptic female choice. Princeton University Press.

Eberhard W.G. & Cordero, C. 1995. Sexual selection by cryptic female choice on male seminal products - a new bridge between sexual selection and reproductive physiology. TREE 10, 493-496.

Gage, M.J.G. 1994. Associations between body size, mating pattern, testis size and sperm lengths across butterflies. Proc. R. Soc. Lond. B. 258, 247-254.

Gomendio, M., & Roldon, E.R.S. 1991. Sperm size and sperm competition in mammals. Proc. R. Soc. Lond. B. 243, 181-185.

Gwynne. D. T. 1984. Male mating effort, confidence of paternity, and insect sperm competition. In: Sperm Competition and the Evolution of Animal Mating Systems (ed. R.L. Smith), pp. 117-150. New York: Academic Press.

Gwynne, D.T. & Snedden, A.W. 1995. Paternity and female remating in Requena verticalis (Orthoptera: Tettigoniidae). Ecol. Entomol. 20, 191-194.

Hebert, P. D., & Beaton, M.J. 1989. Methodologies for allozyme analysis using cellulose acetate electrophoresis. Department of Biological Sciences, University of Windsor. Windsor, Ontario. Helena Laboratories.

LaMunyon, C. W.& Eisner, T. 1993. Postcopulatory sexual selection in an arctiid moth (Utethesia ornatrix). Proc. Nat. Acad. Science USA 90, 4689-4692.

LaMunyon, C. W.& Eisner, T. 1994. Spermatophore size as determinant of paternity in an arctiid moth (Utethesia ornatrix). Proc. Nat. Acad. Science USA 91, 7081-7084.

Lanier, D. R., Estep, D.Q. & Dewsbury. D.A. 1979. Role of prolonged copulatory behavior in facilitating reproductive success in a competitive mating situation in laboratory rats. J. Comp. Physiol. Psych. 93, 781-792.

Lewis, S. M., & Austad, S.N. 1990. Sources of intraspecific variation in sperm precedence in red flour beetles. Amer. Nat. 135, 351-359.

Lewis, S. M., & Austad, S.N. 1994. Sexual selection in flour beetles: the relationship between sperm precedence and male olfactory attractiveness. Behav. Ecol. 5, 219-224.

Martin, J. D., Pashley, D.P. & Mason, L.J. 1989. Sperm use patterns of individual fall armyworms (Lepidoptera: Noctuidae). Ann. Entom. Soc. Amer. 82, 177-180.

McVey, M. E., & Smittle, B.J. 1984. Sperm precedence in the dragonfly Erythemis simplicicollis. J. Ins. Physiol. 30, 619-628.

Oberhauser, K. S. 1989. Effects of spermatophores on male and female monarch butterfly reproductive success. Behav. Ecol. Sociobiol. 25, 237-246.

Oberhauser, K. S. 1997. Fecundity, lifespan and egg mass in butterflies: effects of male-derived nutrients and female size. Func. Ecol. 11(2): 166-175.

Parker, G. A. 1970. Sperm competition and its evolutionary consequences in the insects. Biol. Rev. 45, 525-567.

Parker, G.A., Simmons, L.W. & Kirk, H. 1990. Analyzing sperm competition data: simple models for predicting mechanisms. Behav. Ecol. Sociobiol. 27, 55-65.

Schlager, G. 1960. Sperm precedence in the fertilization of eggs in Tribolium castaneum. Ann. Entom. Soc. Amer. 53, 557-560.

Silberglied, R.E., Shepherd, J.G., & Dickinson, J.L. 1984. Eunuchs: the role of apyrene sperm in Lepidoptera. Amer. Nat. 123, 255-265.

Simmons, L.W. 1987. Sperm competition as a mechanism of female choice in the field cricket, Gryllus bimaculatus. Behav. Ecol. Sociobiol. 21, 197-202.

Simmons, L.W., & Parker, G.A. 1992. Individual variation in sperm competition success of yellow dung flies, Scatophaga stercoraria. Evolution 46, 366-375.

Siva-Jothy, M.T., & Tsubaki, Y. 1989. Variation in copulation duration in Mnais pruinosa pruinosa Selys (Odonata:Calopterygidae). 1. Alternative mate-securing tactics and sperm precedence. Behav. Ecol. Sociobiol. 24, 39-45.

Sivinski, J. 1984. Sperm in competition. In: Sperm Competition and the Evolution of Animal Mating Systems (ed. R.L. Smith), pp. 86-116. New York: Academic Press.

Smith, R.L. 1979. Repeated copulation and sperm precedence: paternity assurance for a male brooding water bug. Science 205, 1029-1031.

Thornhill, R. 1983. Cryptic female choice and its implications in the scorpionfly Harpobittacus negriceps. Amer. Nat. 122, 765-788.

Villavaso, E.J. 1975. Functions of the spermathecal muscle of the boll weevil, Anthonomous grandis. J. Ins. Physiol. 21, 1275-1278.

Walker, W.F. 1980. Sperm utilization strategies in nonsocial insects. Amer. Nat. 115, 780-799.

Ward, P. 1993. Females influence sperm storage and use in the yellow dung fly Scathophaga stercoraria (L.). Behav. Ecol. Sociobiol. 32, 313-319.

Williams, D.A. 1982. Extra-binomial variation in logistic linear models. Appl. Stat. 31, 144-148.

          Willson, M.F. 1990. Sexual selection in plants and animals. TREE 5, 210-214.

Return to top  |  Karen's Research Questions  |  Reproduction Home Page  |  Research Topics  |  Site Overview