Proc. Natl. Acad. Sci. USA Vol. 90, pp. 4689-4692, May 1993 Ecology Postcopulatory in an arctiid ( ornatrix)* (sperm precedence//parental investnent/polyandry/) CRAIG W. LAMUNYONt AND THOMAS EISNER Section of Neurobiology and Behavior, Cornell University, Ithaca, NY 14853-2702 Contributed by Thomas Eisner, March 1, 1993

ABSTRACT The offspring of twice-mated female Utethei- described (7). PA-free Utetheisa were reared on a pinto sa ornatrix show low incidence of mixed paternity. Most bean-based diet (PB diet), devoid of PA. PA-laden Utetheisa progeny are sired exclusively by one male, the larger one, were reared on another diet (CS diet), consisting of PB diet irrespective of parental age, male pyrrolizidine alkaloid con- supplemented with seeds of spectabilis, a natural tent, mating order, between-mating interval, or duration of food plant of Utetheisa, containing the PA monocrotaline. . Data are presented suggesting that the female A total of 53 virgin Utetheisa females, alkaloid-free (PB herself may exert control over the process by which one set of diet-raised) and of known age, were each mated with 2 virgin sperm is utilized at the expense of the other. Evidence for such males, also of known age. The matings were effected by postcopulatory female choice of sperm had not previously been pairing the sexes overnight in screened cylindrical containers obtained for an . Promiscuity provides the female with a (0.35-liter volume). Male age was recorded as the period in means for accruing nuptial gifts (nutrients, defensive alka- days since eclosion from the (range 1-8 days). Female loids). Sperm selection provides her with the option of utilizing age was the period from eclosion to second mating (range sperm from larger, potentially more fit, males. 2-11 days). The interval in days between the two matings was recorded for each female (range 1-4 days). These periods fall Female Utethesia ornatrix mate promiscuously. Spermato- within range of the first third of the life-span of Utetheisa phore counts indicate that they mate on average with four or (adults of both sexes live approximately 1 month in the five males in nature (2). Maximal spermatophore counts of laboratory). over 10 per female have been recorded both in nature (2, 3) With one subsample of females (Exp. 1; n = 21 females), and in our laboratory. By mating multiply, females accrue the pairings were respectively with one alkaloid-free male nutrient, transmitted to the female with the male's spermato- (PB diet-raised) and one alkaloid-laden male (CS diet-raised). phore. With each mating beyond the first, the female is able Sequence of mating was 10 females with alkaloid-free males to produce, on average, an additional 32 eggs (mean egg first and 11 females with alkaloid-laden males first. output ofonce-mated females ± SEM = 226 + 18; n = 46) (3). With the remaining females (Exp. 2; n = 32 females), both In Utetheisa the spermatophore also provides the female with male partners were alkaloid-laden (CS diet-raised), and a defensive pyrrolizidine alkaloid (PA), derived by the male record was kept of male mass and of the duration of copu- from its larval food plant (legumes of the genus Crotalaria). lation. The latter parameter was ascertained by visual check The female bestows alkaloidal gifts that she thus receives ofthe mating chambers, at 15-min intervals, from 1 hr before from males, together with intrinsic PA that she herself has dark until 3 hr after onset of the light phase, in a growth sequestered as a larva, on the eggs (3, 4). Eggs are protected chamber set at a 16 hr of light/8 hr of dark cycle (visual against predators as a result (4, 5). inspection during the dark phase was in red light). Male mass We here address questions concerning the paternity of Uteth- was determined by weighing within 15 min after termination eisa offspring. Specifically, we ask whether males, given that of copulation. they must on average "share" females with other males, have Paternity ofoffspring was assigned on the basis ofvariation assurance offathering at least some offspring with each mating. at the malate dehydrogenase and glucophosphate isomerase Is there sperm mixing in Utetheisa, and as a consequence allozyme loci. Allozyme phenotypes of adults and offspring multiple fathering of offspring? Or is there sperm precedence were determined electrophoretically (8-11). Eggs were col- [that is, exclusive or nearly exclusive siring by one inseminator lected from each female after the second mating and were at the expense of another (6)] with the result that males could allowed to hatch. When the emergent larvae had grown to at times "lose out" in matings? We mated individual female approximately 1 cm in length, they were frozen (-80°C) Utetheisa to two males, ascertained paternity of their progeny (mean ± SEM number oflarvae per female used in paternity by use ofenzymatic markers, and looked into whether specific assignment was 44 ± 3; range 7-113). parameters (male mass, male PA content, parental age, be- The 53 experimental females all met the following two tween-mating interval, duration of copulation) were determi- criteria: (i) they laid viable eggs in the interval between nants ofoffspringpaternity. We also experimented withfemales matings, and (ii) they were shown, upon postmortem dissec- in an attempt to shed light on the mechanism by which sperm tion of the bursa, to contain two spermatophores, of which achieve differential success in multiple matings. one (the more intact, presumed to be from the second male) was properly positioned (column aligned with passageway to pseudobursa) (12). Adherence to these criteria provided MATERIALS AND METHODS assurance that the first mating had been fertile and reasonable We reared Utetheisa in the laboratory (from stock collected near Lake Placid, Highlands County, FL), as previously Abbreviations: PA, pyrrolizidine alkaloid; PB diet, pinto bean-based diet; CS diet, Crotalaria spectabilis seed-supplemented diet. *This is paper no. 115 in the series Defense Mechanisms ofArthro- The publication costs of this article were defrayed in part by page charge pods; no. 114 is Eisner and Dalton, ref. 1. payment. This article must therefore be hereby marked "advertisement" tPresent address: Department of Molecular and Cell Biology, Uni- in accordance with 18 U.S.C. §1734 solely to indicate this fact. versity of Arizona, Tucson, AZ 85721.

4689 Downloaded by guest on October 2, 2021 4690 Ecology: LaMunyon and Eisner Proc. Natl. Acad Sci. USA 90 (1993) assurance that the second mating had also been fertile [we ofboth groups were promptly dissected, and presence know from laboratory observations that the spermatophore or absence of sperm (both apyrene and eupyrene) in the in Utetheisa is often misaligned in sterile matings, as noted spermatheca was visually assessed [apyrene and eupyrene also for other (13)]. Only 5 of our initial sample of 58 sperm are morphologically distinct sperm types, respectively twice-mated females had to be excluded for not meeting these anucleate and nucleate, characteristic of (13)]. criteria. Also noted was the state of motility (whether undulating or Moths were selected for pairing on the basis of their not) of the sperm, both in the spermatheca and (when there allozyme phenotypes, which were known beforehand, since was a sperm remnant) in the spermatophore. Four additional the moths were either from homozygous strains or had been females, treated like the experimentals, were given 2-3 hr for individually screened for allozyme makeup (analysis of left recovery from the anesthesia (CO2 = 2 females; N2 = 2 mesothoracic leg; the amputation had no noticeable effect on females) prior to dissection. the moths' subsequent mating performance). With 42 of the females, the two male partners were allozymically suffi- ciently different to allow for unambiguous assignment of RESULTS paternity to all progeny. With the other 11 females, the male It is clear from Fig. 1 that in twice-mated Utetheisa there is partners shared one allele, so that offspring bearing that allele sperm precedence: in all but 2 cases there was aprincipal sire, could not be unambiguously assigned to either sire. With a male that sired significantly more than 50%o of the offspring these 11 females, only offspring bearing the unshared allele [G statistic (16)] (the male siring less than 50% of offspring is were considered for estimation of paternity, since these henceforth called the nonprincipal sire). Sperm mixing was offspring could be definitely assigned to the sire (the het- minimal: in 36 of 53 females erozygote) with the allele. When offspring bearing the un- (nearly 70% of the sample) the shared allele were not produced by such females, the het- principal sire fathered fully 100% of the progeny. Notable is erozygous male was taken not to have sired any progeny. If, the fact that the principal sire was as often the first as the on the other hand, afraction ofthe progeny bore the unshared second male partner. This is unusual for Lepidoptera, where allele, the heterozygous sire was assigned twice that fraction it is often the sperm of the last male that gains precedence of the offspring, since only half his progeny (on average) (17). would have received that allele. The width of the 95% Results were analyzed to check into determinants of pa- confidence intervals for the estimated paternities ranged from ternity. To this end, the sample was reduced by subtraction 2% to 45% (based on the binomial distribution). of the two cases (both from Exp. 2) in which there had been To gain insight into the mechanism by which sperm are no principal sire. Alkaloid content was not a determinant: a transported in the female reproductive system, individual male was as likely to become a principal sire when he was females were anesthetized at a specific time (7.5 hr) after alkaloid-laden as when he was not (Table 1, row 1). Age also, initiation of mating. We knew from dissection of females whether of male or female, proved irrelevant: principal and killed at intervals after onset of copulation that at 7.5 hr nonprincipal sires were of comparable ages, and female age deposition of the male's spermatophore into the female's did not affect the outcome ofthe matings (Table 1, rows 2 and bursa is complete, and sperm begin appearing in the female's 3). Duration of copulation was also nondeterminant: mean sperm-storage chamber (spermatheca). For anesthesia, the copulation times were similar for principal and nonprincipal females were manually disengaged from the males and trans- sires (Table 1, row 4). Between-mating intervals also had no ferred for 2-3 hr to a small receptacle flushed with a stream effect on paternity outcome (Table 1, row 5). of CO2 (n = 8 females) or N2 (n = 8 females), treatments The one variable that proved determinant was male mass: which in other inseminated insects are known to inhibit principal sires were heavier on average than their rivals (Fig. muscle action without blocking the sperm's undulatory mo- 2 Left). The size advantage took effect only when the mass tions (14, 15). Another 16 females (controls) were identically discrepancy was greater than about 6 mg (Fig. 2 Right). At treated, but kept unanesthetized by transference to a recep- that level of difference, the larger male was the principal sire tacle flushed with air. Following the treatment interval, in 70% of cases (11 of 16). 20 * 1st m o 2ndr

15 J No p

E 01 0 z

5

0 0 1-10 11-20 21-50 51-79 80-89 90-99 100 Percent offspring sired by second male FIG. 1. Percent offspring sired by second male in twice-mated Utetheisa females. The principal sire is the one who fathered significantly more than 50o of the offspring [G statistic (16)]. Results are for 53 females, Exps. 1 and 2 combined. Downloaded by guest on October 2, 2021 Ecology: LaMunyon and Eisner Proc. Natl. Acad. Sci. USA 90 (1993) 4691 Table 1. Correlation of various parameters with siring success of male Principal Nonprincipal Parameter sire sire Statistics 1.* Number of alkaloid-laden males 12 9 G = 0.430, P > 0.5 2.t Mean male age at time of mating, days 3.4 ± 0.2 3.9 ± 0.3 Paired t = 1.385, P = 0.173 3It Mean female age at time of 2nd mating, days, when 2nd male was 5.7 ± 0.4 5.8 ± 0.5 t = 0.140, P > 0.5 4.* Mean copulation duration, hr 9.7 ± 0.3 9.3 ± 0.3 Paired t = 1.139, P = 0.266 5.t Mean between-mating interval, days, when 2nd male was 1.8 ± 0.1 1.9 ± 0.2 t = 0.271, p = 0.788 Means are given ±SEM. G statistic is from ref. 16. *Exp. 1(21 females). tExps. 1 and 2 combined (51 females). tExp. 2 (30 females).

The results of the anesthesia experiments are shown in In our judgement postcopulatory sperm sorting in Ute- Table 2. In females that were killed and dissected immedi- theisa is exercised by the female herself. Anestheisa in the ately after anesthesia, the sperm (actively undulatory) was inseminated female blocked the routing of sperm to the still lodged in most cases in the spermatophore. Such sperm spermatheca, without inhibiting sperm motility. We take this as was noted in the spermatheca ofthese females (four cases) to indicate that it is under action ofthe female's musculature, was exclusively or predominantly of the apyrene type (we rather than the sperm's intrinsic motion, that sperm are had learned from previous laboratory observations that apy- channeled to their destination. Indeed, the chambers and rene sperm are the first to reach the spermatheca after mating conduits of the female reproductive system (the bursa, in Utetheisa). In unanesthetized controls, both sperm types pseudobursa, spermatheca, and ducts linking these) (12) are had been routed in their entirety to the spermatheca in most invested by muscles, muscles which in fresh dissections pulse instances. Such routing had occurred also in all four females peristaltically and can be expected to be rendered quiescent that were allowed to recover from anesthesia before dissec- by anesthesia. While we lack direct evidence of how the tion. female might effect the favoring of one set of sperm over the other, it is not difficult, given the elaborate subcompartmen- talization of her reproductive system (12), to envision mech- DISCUSSION anisms by which the sperm might be selectively expelled or It seems established that larger males have a reproductive stored. By the same token, we cannot be absolutely certain advantage in Utetheisa. If a male is the larger (by 6 mg or that sperm are purely passive with respect to the conveyance more) ofthe first two mates of a female, he has a good chance mechanism that determines their fate. Anesthesia, while not of siring most or all of the offspring. The low incidence of inhibiting sperm motility, could nonetheless have affected the mixed paternity indicates that rival sperm are not merely sperm's ability to move in oriented fashion. vying numerically for access to the eggs, as they apparently One wonders to what extent our findings are applicable to do in some other insects (18-20). Instead, they appear to be female matings beyond the first two. Is siring success a function subject to a postcopulatory selective process, a process that of male size in even later matings? And is it always only results in virtual exclusion of the sperm of the smaller the last male and his immediate predecessor that vie for male. sperm precedence? Or does the sperm of a large Male Utetheisa appear to lack means for post- male retain affecting precedence for as long as only smaller males copulatory sperm selection. some mate subse- Unlike other Lepidoptera quently with the female? We suspect that male size may play (21), they do not leave sperm plugs after copulation: female a role in all matings except the first. Only as inseminators of Utetheisa are able to mate on consecutive nights. Further, virgins, and only with respect to offspring that the female males have no provision for removing a rival male's sperm: produces prior to her second mating, can males have assur- their aedeagus (penis), when positioned during copulation, is ance of paternity, irrespective of size. out of reach of the spermatheca. How do female Utetheisa, in exercising choice of sperm, assess the determinant parameter, male size? We suggest that they do so indirectly, by assessment of spermatophore size. Spermatophores in Utetheisa vary as a function ofmale body mass [data based on first spermatophore of males (3)]. One could thus envision females gauging spermatophores pro- prioceptively, as they are known to do in other Lepidoptera (22), on the basis of the degree of distention of the bursa. The present findings have a bearing on what is known about precopulatory sexual selection in Utetheisa. Male Table 2. Sperm content of spermatheca in experimental females (treated by postcopulatory anesthesia) and their controls No. with sperma- No. with sperm in spermatheca theca Apyrene Mostly Apyrene + Treatment empty only apyrene* eupyrene <6.02 mg >6.02 mg All males None (n = 16) 2 0 2 12 Males differing from Anesthesia (n = 16)t 12 3 1 0 rival male by Anesthesia + recovery = FIG. 2. Siring success ofmales, plotted as a function of male mass period (n 4)t 0 0 0 4 (30 females; Exp. 2). (Left) Results for all males. (Right) Siring *A small quantity of eupyrene sperm was also noted. success as a function of the difference in mass from rival male. tCombines C02- and N2-anesthetized females. Downloaded by guest on October 2, 2021 4692 Ecology: LaMunyon and Eisner Proc. Natl. Acad. Sci. USA 90 (1993) Utetheisa, prior to mating, "announce" to the female how or whether they are merely indiscriminately polygynous, much PA they hold in store. They do this by use of a remains unknown. , hydroxydanaidal, that they air during courtship One incidental observation reinforces the notion that males by everting a pair of brushes that are the glandular source of are chosen by size in the sexual selective scheme of Ute- the substance (5, 7). Hydroxydanaidal is derived from PA and theisa. Males, in lepidopteran species, are usually smaller is produced by the male in proportion both to his PA content than females (25). In Utetheisa, the reverse is the case (mean and to the amount of PA he transmits to the female at mating pupal mass ±SEM: male = 149.4 ± 3.5 mg; female = 141.7 with the spermatophore, for investment in the eggs (23). In ± 4.0 mg; n = 35 for both samples.) mating experiments, females have been shown to favor males We thank Scott S. Emerson (University of Arizona, Tucson) for containing PA and able to produce hydroxydanaidal (7). help with calculation of confidence intervals (estimated paternities), It appears, therefore, that there are two sexual selective G. Scoles and K. Amerman for technical assistance, and W. Mitchell processes operating in Utetheisa, a precopulatory process for Masters (Ohio State University, Columbus) and our colleagues selection of males bearing PA and a postcopulatory process Jerrold Meinwald and James F. Hare for helpful comments on the for selection of males of large size. Interestingly, large size manuscript. This study was supported by National Institute of and Allergy and Infectious Diseases Grant A102908 and National Insti- high alkaloid content correlate in Utetheisa adults in tute of Mental Health Grant 5 nature (Fig. 3), probably for dietary reasons. Moths that as Training T32 MN15793. larvae feed predominantly on the foliage of their food plant 1. Eisner, T. & Dalton, L. M. (1993) J. Chem Ecol., in press. grow up smaller and less alkaloid-endowed than larvae 2. Pease, R. W., Jr. (1968) J. Lepid. Soc. 22, 197-209. feeding on the more nutritious and more alkaloid-laden seeds 3. LaMunyon, C. W. (1992) Ph.D. dissertation (Cornell Univer- (24). We have suggested that female Utetheisa, by favoring sity, Ithaca, NY). courting males ofhigh PA content, may be selecting for males 4. Dussourd, D. E., Ubik, K., Harvis, C., Resch, J., Meinwald, J. & Eisner, T. (1988) Proc. Natl. Acad. Sci. USA 85, 5992- able both to bestow high alkaloidal gifts and to compete 5996. favorably for seeds as larvae, abilities that could share a 5. Eisner, T. & Meinwald, J. (1987) in Pheromone Biochemistry, genetic basis (5, 23). By favoring, in the postcopulatory eds. Prestwich, G. D. & Blomquist, G. J. (Academic, Orlando, context, males of large size, females are, in a sense, rein- FL), pp. 251-269. forcing the precopulatory selective process. But most impor- 6. Thornhill, R. & Alcock, J. (1983) The Evolution of Insect tant, by rating males after mating, females are able to Mating Systems (Harvard Univ. Press, Cambridge, MA). 7. Conner, W. E., Eisner, T., Vander Meer, R. K., Guerrero, A. compare them sequentially, favoring genetically only the & Meinwald, J. (1981) Behav. Ecol. Sociobiol. 9, 227-235. more fit, while accruing the nutritive and alkaloidal gifts of 8. May, B., Wright, J. E. & Stoneking, M. J. (1979) Fish. Res. all. Board Can. 36, 1114-1126. Males evidently stand a chance of being cuckolded by any 9. Harris, H. & Hopkinson, D. A. (1976) Handbook ofEnzyme competing inseminator oflarger size. They have assurance of Electrophoresis in Human Genetics (American Elsevier, New paternity only with respect to offspring produced by females York). prior to their second 10. Selander, R. K., Smith, M. H., Yang, S. Y., Johnson, W. E. & mating. Whether males themselves Gentry, J. B. (1971) Studies in Genetics IV (Univ. of Texas, exercise selection and favor virgins in their choice of mates, Austin), Publ. 7103. 11. Ridgeway, G. J., Sherburne, S. W. & Lewis, R. D. (1970) 1200 Trans. Am. Fish. Soc. 99, 147-151. 12. Hinton, H. E. (1981) Biology ofInsect Eggs (Pergamon, New CD:4.1000 York), Vol. 2. 13. Mann, T. (1984) Spermatophores (Springer, Berlin). co 14. Tschudi-Rein, K. & Benz, G. (1990) Ann. Entomol. Soc. Am. E 800- 83, 1158-1164. 15. Davey, K. G. (1958) J. 0. Exp. Biol. 35, 694-701. a) 600- 16. Sokal, R. R. & Rohlf, F. J. (1981) Biometry (Freeman, New C U York). 17. Drummond, B. A. (1984) in Sperm Competition and the Evo- 2 400 0 lution ofAnimal Mating Systems, ed. Smith, R. L. (Academic, Orlando, FL), pp. 291-370. 0 200 18. Dickinson, J. L. (1986) Behav. Ecol. Sociobiol. 18, 331-338. 19. Parker, G. A. (1970) Biol. Rev. 45, 525-567. 20. Lewis, S. M. & Austad, S. N. (1990) Am. Nat. 135, 351-359. 0 21. Dickinson, J. L. & Rutowski, R. L. (1989) Anim. Behav. 38, 20 40 60 80 100 120 154-162. Male body mass, mg 22. Sugawara, T. (1979) J. Comp. Physiol. 130, 191-199. 23. Dussourd, D. E., Harvis, C. A., Meinwald, J. & Eisner, T. FIG. 3. PA content of male Utetheisa, plotted as a function of (1991) Proc. Natl. Acad. Sci. USA 88, 9224-9227. male mass. Moths were raised on outdoor plots of Crotalaria 24. Conner, W. E., Roach, B., Benedict, E., Meinwald, J. & spectabilis, containing the PA monocrotaline. Data are from Conner Eisner, T. (1990) J. Chem. Ecol. 16, 543-552. et al. (24), where other correlations were presented. r = 0.78; P < 25. Opler, P. A. & Krizek, G. 0. (1984) Butterflies East of the 0.01. Great Plains (Johns Hopkins Univ., Baltimore), pp. 12-13. Downloaded by guest on October 2, 2021