The Auk 127(3):495-502, 2010 ? The American Ornithologists' Union, 2010. Printed in USA. PASSIVE SPERM LOSS AND PATTERNS OF SPERM PRECEDENCE IN MUSCOVY DUCKS (CAIRINA MOSCHATA) SARAH M. STAI1 AND WILLIAM A. SEARCY2 Department of Biology, University of Miami, Coral Gables, Florida 33124, USA ABSTRACT.?We investigated the importance of passive sperm loss in the outcome of sperm competition in captive, wild-type Muscovy Ducks (Cairina moschata). In nature, Muscovy Ducks can be expected to experience more intense sperm competition than most other waterfowl because of their non-monogamous mating system. We estimated the instantaneous rate of sperm loss from the reproductive tracts of female Muscovy Ducks as 0.0235 + 0.0018 Ir1 [SE], a fairly typical rate in comparison to the few other species of birds in which passive sperm loss has been measured. We also measured sperm precedence in trials in which a captive female was allowed to mate with two males in succession, with either a 24-h or a 72-h lag between matings. Paternity was determined with microsatellite markers. The mean proportion of a female's eggs fertilized by the second male (P2) was 0.72 + 0.14 in trials with the 24-h lag and 0.42 + 0.13 in trials with the 72-h lag. The last-male precedence observed in the 24-h trials can be explained by a quantitative model in which passive sperm loss alone determines average success, but this model is not consistent with the outcome of the 72-h trials. Other factors, including perhaps postcopulatory female choice, must be acting in addition to passive sperm loss in the trials with the longer lag. Received 18 August 2009, accepted 20 January 2010. Key words: Cairina moschata, Muscovy Ducks, passive sperm loss, sperm competition, sperm precedence, waterfowl. Perdida Pasiva de Esperma y Patrones de Precedencia Espermatica en Cairina moschata RESUMEN.?Investigamos la importancia de la perdida pasiva de esperma en el resultado de la competencia espermatica en patos cautivos de tipo silvestre de la especie Cairina moschata. En la naturaleza, se espera que los individuos de esta especie experimenten una competencia espermatica mas intensa que la mayoria de otros anseriformes debido a su sistema de apareamiento no monogamo. Estimamos la tasa instantanea de perdida de esperma del tracto reproductive de hembras de esta especie en 0.0235 + 0.0018 hr1 [EE], un valor relativamente tipico en comparacion con las pocas especies de aves adicionales en las cuales se ha medido la perdida pasiva de esperma. Tambien medimos la precendecia espermatica en ensayos en los que a una hembra cautiva se le permito aparearse con dos machos consecutivamente, con un intervalo de 24 h o de 72 h entre apareamientos. La paternidad se determine mediante marcadores microsatelites. La media de la proporcion de los huevos de una hembra que fue fertilizada por el segundo macho (P2) fue 0.72 + 0.14 en ensayos con intervalo de 24 h y 0.42 + 0.13 en ensayos con intervalo de 72 h. La precedencia del ultimo macho observada en los ensayos con intervalo de 24 h podria ser explicada por un modelo cuantitativo en el que la perdida pasiva de esperma por si sola determina el exito promedio, pero este modelo no concuerda con el resultado de los ensayos de 72 h. Otros factores aparte de la perdida pasiva de esperma, incluyendo quizas la seleccion postcopulatoria por parte de las hembras, deben estar actuando en los ensayos con intervalo mayor. SPERM COMPETITION CAN occur whenever individual females sperm competition. In waterfowl (Anatidae), almost all species are mate with multiple males in the course of single breeding at- socially monogamous (Oring and Sayler 1992), so we can expect tempts. Multiple mating by females is common across animals, sperm competition to be concentrated in the subset of socially so it is not surprising that sperm competition acts as an impor- monogamous species with a high frequency of extrapair mating tant selective force in a wide range of animal groups (Parker 1970; (Afton 1985, Sorenson 1994, McKinney and Evarts 1998, Dunn et Ginsberg and Huck 1989; Birkhead and Moller 1992, 1998; Sim- al. 1999, Cunningham 2003) and in the small minority of species mons 2001). In birds, multiple mating can be quite common even with non-monogamous social systems (Oring and Sayler 1992, in socially monogamous species (Griffith et al. 2002, Westneat Coker et al. 2002). Here, we investigate sperm competition in one and Stewart 2003), but other social mating systems, notably pro- of the latter, the Muscovy Duck (Cairina moschata), concentrating miscuity, are especially conducive to multiple mating and, thus, to on the phenomenon of last-male sperm precedence. 'Present address: EcoSmith Consulting, 9749 Queen Road, Bloomington, Minnesota 55431, USA. 2Address correspondence to this author. E-mail: wsearcy@miami.edu The Auk, Vol. 127, Number 3, pages 495-502. ISSN 0004-8038, electronic ISSN 1938-4254. ? 2010 by The American Ornithologists' Union. All rights reserved. Please direct all requests for permission to photocopy or reproduce article content through the University of California Press's Rights and Permissions website, http://www.ucpressjournals. com/reprintlnfo.asp. DOI: 10.1525/auk.2010.09138 -495- 496 STAI AND SEARCY AUK, VOL. 127 Last-male sperm precedence is the disproportionate fertiliza- tion of eggs by sperm from the last male of a series of two or more to copulate with a female. Last-male precedence is widespread in insects (Ridley 1989) and is usually thought to be the norm in birds as well (Birkhead and Moller 1992, Birkhead 1998). Consid- erable effort has been put toward finding a mechanism to explain this pattern. In the case of birds, three principal mechanisms have been discussed (Birkhead 1998). "Stratification" proposes that suc- cessive ejaculates form layers within the sperm-storage tubules of the female and that the uppermost layer, derived from the last male to copulate, is more available to fertilize eggs than lower lay- ers (Compton et al. 1978, Cheng et al. 1983, McKinney et al. 1984). This mechanism is often described as "last in, first out." A second possibility is "displacement," whereby later ejaculates enter parts of the reproductive tract such as sperm-storage tubules and force out sperm that are already there (Lessells and Birkhead 1990). Third, "passive sperm loss" explains last-male precedence as re- sulting from the gradual loss of sperm from the female reproduc- tive tract over time; because of such loss, if two ejaculates of equal numbers of sperm are received by a female, fewer sperm of the first will remain to fertilize eggs by the time the second arrives (Lessells and Birkhead 1990). More complicated models can be obtained by combining these mechanisms, for example by pairing stratifica- tion with passive sperm loss (Birkhead and Biggins 1998). Stratification predicts that if a long enough series of eggs is fertilized, the sperm of the last male to copulate will eventually be depleted and the sperm of an earlier male will start to predomi- nate (Lessells and Birkhead 1990). Because little evidence of this pattern has been found in birds, stratification has lost favor as a hypothesis (Birkhead 1998, Birkhead and Biggins 1998). Displace- ment is a mechanism that has been well demonstrated in insects (Waage 1979, Ono et al. 1989, von Helversen and von Helversen 1991) but for which there is no substantial evidence in birds (Birk- head 1998). In contrast to these first two hypotheses, passive sperm loss has received important support in birds as a general explana- tion for last-male precedence. The basic assumption that sperm are lost from the female reproductive tract has been verified in several species (Wishart 1988, Birkhead et al. 1993, Cunningham and Cheng 1999, Michl et al. 2002). Moreover, when empirically measured rates of sperm loss have been incorporated into math- ematical models, the models have correctly predicted fertilization patterns in certain cases (Birkhead et al. 1995b, Colegrave et al. 1995, Birkhead and Biggins 1998). The passive-sperm-loss model (Birkhead et al. 1995b) is ex- pressed as loge(P2/P1) = d + uT - logy, where P2 is the proportion of eggs fertilized by the second male, P1 is the proportion fertil- ized by the first male, d is the differential fertilizing capacity of the second male in relation to the first, u, is the instantaneous rate of sperm loss from the female reproductive tract, 7Ts the time inter- val between inseminations, and / is the size of the first insemina- tion in relation to the second. The first term, loge(P2/P1), measures the success of the second male in relation to the first. This model predicts last-male sperm precedence, as long as the second male does not differ from the first in sperm number or fertilizing capac- ity per sperm. The model also predicts that the relative success of the second male will increase as the time interval between insemi- nations increases. This prediction has been supported in several studies (Birkhead and Biggins 1998). We investigated sperm competition in the Muscovy Duck, which is one of a small minority of waterfowl whose social mating system is considered non-monogamous (Oring and Sayler 1992). The social mating system of Muscovy Ducks has been variously labeled as promiscuous (Delacour 1959, Crawford 1990) or polyg- amous (Clayton 1984, Todd 1996). An intensive study of free- living, wild-type individuals within their natural range classified the species as promiscuous on the basis of evidence that associa- tions between males and females during the breeding season were generally brief and nonexclusive (Stai 2004). In studying sperm competition in Muscovy Ducks, our first goal was to measure the rate of passive sperm loss, for comparison with other species and as the crucial parameter in the passive-sperm-loss model. Second, we wanted to test for the occurrence of last-male precedence in this species and test the prediction that the magnitude of last- male precedence will increase with the time interval between in- seminations. Third, we wanted to determine the quantitative fit between the predictions of the passive-sperm-loss model and the proportion of eggs fertilized by the second male in actual matings. Fourth, we used our genetic data to check whether fertilization success was biased toward males that were less closely related to the female, as has been suggested for other species of birds (Pizzari et al. 2004, Thuman and Griffith 2005). METHODS Acquisition and maintenance of birds.?We acquired 12 male and 10 female Muscovy Ducks, all young of the year, from Northwest Wildfowl (Everett, Washington) in November 1999. They were transferred to the Smithsonian Institution's Conservation and Re- search Center (CRC) in Front Royal, Virginia, in May 2000. Two ducklings, one male and one female, hatched in September 2000 and were raised to adulthood, and two more yearling females were acquired from Northwest in March 2001. Thus, the total captive population size was 26 birds at its maximum. All subjects were wild-type birds, descended from a stock that was originally ac- quired from Paraguay in the early 1970s (Stai and Hughes 2003). Ducks were housed year-round in a set of covered outdoor pens on a south-facing slope at CRC. Individual pens were 12.2 x 18.3 m enclosures with 4.6-m2 ponds in the center. Abundant natural vegetation grew in the pens, providing cover and ground- nesting sites. Wooden nest boxes were added to individual nesting pens during the 2000 season. Ducks had free access to running water and commercial duck feed and were given occasional sup- plements of mealworms. Females were supplemented with oyster shells during egg laying. Shelters with straw bedding and heat lamps were provided during the winter months. During the nonbreeding season, all birds were held in a single flock that occupied multiple adjoining pens. From the onset of egg laying in February or March until August or September, males and females were separated into same-sex flocks. The flock of males occupied multiple adjoining pens and had visual and auditory access to females through the wire fencing. Females were held in a flock adjacent to the males until late spring, when they were distributed into individual pens in preparation for trials and nesting. Experiment 1: Measurement of passive sperm loss rate.?Ten pairs consisting of one male and one female Muscovy Duck were placed each in a separate nesting pen on 27 June 2000. Pairs were JULY 2010 SPERM COMPETITION IN MUSCOVY DUCKS 497 allowed to copulate ad libitum until the first egg was laid, follow- ing Cunningham and Cheng (1999). The male was then removed to the holding pen. Each egg was collected on the day it was laid and replaced with a dummy egg until the female stopped laying. The last clutch was completed on 27 July. Eggs were stored at 4?C until dissection. The number of sperm trapped in the egg's perivitelline layer declines with successive eggs in a clutch, reflecting the decline in sperm numbers in the sperm-storage tubules (Wishart 1987, Bril- lard and Antoine 1990, Brillard and Bakst 1990). Thus, the loge of sperm number found in the perivitelline layer, regressed against time, produces a slope equal to the rate of passive sperm loss (u). We prepared the perivitelline membrane of each egg for sperm counting according to Wishart (1987). The fluorescent DNA probe (4,6-diamidino-2-phenylindole [DAPI]) and the rinse solution (Ca2+Mg2+-free Dulbecco's phosphate-buffered saline [PBS]) were obtained from Sigma-Aldrich (St. Louis, Missouri). We used a 1% solution of DAPI in PBS to stain the membranes and Krystalon to seal the mount. Prepared slides were stored in darkness at 4?C. We followed Wishart's (1987) counting technique, using flu- orescence microscopy and a x20 objective, to make six scans per slide. Each scan was 10 mm long and 0.940 mm wide (i.e., the di- ameter of the field of view); thus, all sperm in a 56.4-mm2 area were counted for each egg. We performed a simple linear regres- sion of loge of sperm numbers against time to estimate the rate of sperm loss. Because complete uptake of sperm by the storage tubules did not always occur immediately after the male was re- moved (as reflected by a lag in the peak number of trapped sperm), the first data point in each regression was the egg with the maxi- mum number of trapped sperm within its own clutch (Birkhead and Petrie 1995). Experiment 2: Sperm competition trials.?We ran two sets of sperm competition trials, one set with a 24-h lag between matings and the other with a 72-h lag. Females were allowed no physical contact with any male for at least 29 days before participation in a trial so that they would have no or virtually no viable sperm re- maining from any earlier insemination. Female Muscovy Ducks will lay eggs without fertilization, as is true of many species of birds (Romanoff and Romanoff 1949). We conducted trials by in- troducing a male into the female's nesting pen between 1 and 7 h after egg laying, thus avoiding a period of low fertility between -4 to +1 h surrounding egg laying and putting all trials within the +1 to +7 h period of high fertility shown by Raud and Faure (1990) in domestic female Muscovy Ducks. If no egg was laid on a trial day, either the female was palpated before male introduction to rule out the presence of an egg in the oviduct (which would block sperm uptake from a new insemination), or the male was intro- duced after 1500 hours (i.e., at least 2 h beyond the latest time of day that females were observed laying eggs). The mating of male 1 of a trial took place on the day the female laid the second (infertile) egg of a new clutch, and the mating of male 2 occurred 24 h or 72 h afterward (while maintaining a minimum of 1 h after egg laying before male introduction). Males were assigned to trials according to the following crite- ria: (1) each individual was both first male and second male at least once in a set of trials; (2) at least 2 days had passed since the male's previous copulation; (3) a dyad of males was not used together in the same trial more than once in a set of trials; and (4) males were paired such that paternity of offspring would be distinguishable with the available microsatellite genotypes (see below). On occa- sions when the preassigned male showed no intent to initiate a copulation attempt (i.e., by approaching female, pecking at dorsal feathers) within 30 min of introduction, it was removed from the female's pen and replaced with another male (always maintain- ing criteria 2 and 4, above, and observing 1 and 3 to the extent possible). In 36 of the 42 trials (86%), a copulation was initiated within 3 min of male introduction to the pen; in the remaining 6 trials, the lag to copulation was 4-49 min. We videotaped all trials and reviewed the videotapes if direct observation was am- biguous to ensure that copulations were behaviorally complete. A copulation was labeled complete when the sequence of copulatory events (mount, grasp, tread, tail bend) ended with a single ejacu- latory thrust by the male (McKinney et al. 1983, Sorenson 1994). Only one copulation was allowed per trial; any male that initiated a second attempt was removed immediately. If no second copula- tion was attempted, the male was left in the female's pen for up to 11 min (mean = 4.5 min) before removal. Females proceeded to lay eggs until the clutch was complete. We collected eggs on the day they were laid and replaced each with a dummy egg. Eggs were incubated in a Petersime incubator at 37.5?C and 86% relative humidity until ~7 days of embryonic development. We briefly froze and then dissected eggs to collect tissue for genetic analysis (see below) or to confirm lack of develop- ment or embryonic death. Tissue samples were preserved in 75 mM NaCl/25mM EDTA/1% SOS. We report here on 21 successful trials, 8 of which were 24-h trials and 13 of which were 72-h. Trials were conducted during July-August 2000 (? = 1), June-August 2001 (? = 13), and May 2002 (n = 7). Trials involved 10 females (1-3 trials each; mean = 2.1) and 13 males (1-5 trials each; mean = 3.2). Genetic analysis and paternity assignment.?All adults were genotyped at four microsatellite loci (Stai and Hughes 2003) on an ABI 310 Genetic Analyzer (Applied Biosystems, Foster City, California). The mean number of alleles (?a) and mean heterozy- gosity (H) for these four loci were lower for individuals from the captive, wild-type stock used in the present study (?a = 3.8, H = 0.47) than for a free-living population in Brazil (?a = 13.0, H= 0.88; Stai and Hughes 2003). The level of genetic variation found in our captive stock resembled closely the level found in domestic Mus- covy Ducks in Brazil (Stai and Hughes 2003). To provide additional genetic information when needed, some adults were genotyped at an additional two loci (APH16 and APH17; Maak et al. 2003) us- ing acrylamide gels. Alleles resolved on the ABI were sized using GENOTYPER, version 2.1, and alleles visualized on acrylamide gels were scored by three independent observers. Offspring were similarly genotyped at the one or more loci for which the possi- ble combinations of parental genotypes allowed paternity assign- ment (but not at the loci that could not be informative given the genotypes of the possible parents). Because adult genotypes had not been determined when the earliest mating trials occurred, there were seven clutches for which criterion 4 (see above) could not be fully observed. Paternity in those clutches was indistin- guishable for 1-4 eggs, whereas 55-78% of the offspring were as- signable. The remaining 14 clutches were 100% assignable. Following Thuman and Griffith (2005), we calculated the ge- netic similarity between females and their mating partners using 498 STAI AND SEARCY AUK, VOL. 127 the R (total relatedness) parameter from the program Mer (Wang 2002). R was estimated using the four microsatellite loci at which all adults were genotyped. We calculated difference in relatedness as the R value for the female and male 2 in a given trial minus the R for the female and male 1. Data analysis.?The proportion of eggs fertilized by the first (P-) and second (P2) males was calculated only for eggs for which paternity was assignable and that were potentially fertilizable by either male (i.e., for those eggs laid on or after the day following mating by the second male). Mean P2 was calculated separately for the set of trials with a 24-h lag between matings and for the set with a 72-h lag. Although no two trials within a set used the same combination of individuals, some individuals were used more than once in the same position (e.g., twice as the second male with a different female and first male). To limit pseudoreplication, we recalculated mean P2 by grouping trials within a set by the identity of the second male, calculating the mean per male, and averaging those means. We calculated 95% confidence intervals (CI) for ob- served P2 and compared them with the P2 expected according to the PSL model based on sperm-loss data from experiment 1. The other variables in the PSL equation were controlled in a statistical sense: because the same males served in both the first and sec- ond roles within sets of trials and because individual males had no way of determining whether they were the first or second male within a trial, both fertilizing capacity and size of the insemina- tion should have been, on average, the same for first and second males for both time lags. Accordingly, we used a simplified version of the PSL model in which both d (the difference in fertilizing abil- ity) and logy (the log of the ratio of insemination sizes) equal zero, causing those terms to drop out of the equation for second-male advantage. The terms d and loge7 should approximate zero only over sets of trials, on average; for any single trial, nonzero values are possible, contributing to error in predicting second-male ad- vantage based on the rate of sperm loss alone. To test the prediction that the male less closely related to the female should dominate paternity, we calculated Spearman cor- relations (rs) between P2 and the difference in relatedness of the female to male 2 and male 1. Ethical considerations.?These experiments were conducted with the approval of the Institutional Animal Care and Use Com- mittees of the University of Miami and the Smithsonian Institu- tion Conservation and Research Center. RESULTS Experiment 1: Measurement of the rate of passive sperm loss.? Seven of 10 females laid fertile clutches with a mean clutch size of 11.9 eggs (range: 9-19 eggs). Sperm uptake was assumed to be complete for a given female on the day that the maximum num- ber of sperm was found on the perivitelline membrane for that female. Complete sperm uptake took from 1 to 4 days (mean = 2 days) after the day of male removal. After sperm uptake was com- plete, all 7 females showed a log-linear decline of sperm numbers with time (Fig. 1). The regression of log mean sperm numbers against number of days since complete uptake was highly signifi- cant for each female (mean r2 = 0.925 + 0.022; results are pre- sented as means + SE), and the overall mean instantaneous rate of sperm loss was 0.0235 + 0.0018 Ir1 (n = 7 females). From a mean of 507 + 223 on the day of complete uptake, mean sperm numbers in the sampled area of membrane decreased to 4.8 + 2.1 by the 10th day after male removal (n = 7 females), and to 0.3 + 0.1 for the last egg of clutches with >10 eggs (n = 3). The last fertilized egg of the three largest clutches was laid an average of 17 days after male removal (range: 13-23 days), which indicates that few sperm re- main viable in the reproductive tract of female Muscovy Ducks for more than ~3 weeks. Experiment 2: Sperm competition trials.?Average clutch size (n = 21) was 12.6 eggs (range: 7-15 eggs). An average of 64% 0 !/5 0) 0 0 o Number of days after male removal FIG. 1. Linear regressions of sperm numbers found on the sampled area of an egg's perivitelline membrane against number of days after copulation for 7 Muscovy ducks. Each symbol represents a separate female. JULY 2010 SPERM COMPETITION IN MUSCOVY DUCKS 499 A (/) 4- (D .c u . -^ 3- 0) -Q 1- lB-H T T "T T~ 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 P2 (proportion sired by second male) B 1- 6- (/) (I) 0.8; in the other two, P2 was 0.22 and 0 (Fig. 2A). In the 13 trials with a 72-h lag between matings, mean P2 was 0.42 + 0.13. Two males were used twice as the second male in this set; when mean P2 was recalculated after first averaging the trials for these two males, the overall mean P2 was 0.38 + 0.14. P2 was 0.0 in seven of the 13 trials and 1.0 in four, with only two intermediate values (Fig. 2B). By incorporating the mean rate of sperm loss (0.0235 h"1) es- timated in experiment 1, the model in which passive sperm loss alone determines fertilization success predicts a P2 value of 0.64 for 24-h trials and a P2 value of 0.84 for 72-h trials. The value of P2 predicted for the 24-h trials (0.64) was close to the observed mean value (0.72) and was well within the 95% CI around that mean (0.45-0.99). By contrast, the value of P2 predicted for the 72-h tri- als (0.84) was much higher than the observed mean (0.42) and was outside the 95% CI (0.16-0.68). The predicted mean was also out- side the 95% CI around the observed mean recalculated to control for second-male identity. There was no obvious within-male consistency in fertilization success rate. Nine of 12 males that copulated in more than one trial had mixed success, dominating fertilization in at least one trial but losing the majority of fertilizations in at least one other trial. Only one male was consistently successful (in four trials), whereas two males were consistently unsuccessful (in three trials each). If each male in a pair had a 50% chance of dominating, a run of four suc- cesses would occur one 16th of the time, so one such run among 12 males is not unexpected. Similarly, the probability of three fail- ures in a row is 1 in 8, not much different from the observed 2 in 12. Although sample size was small, there was also little indication of consistency within females who mated with the same male in dif- ferent trials (n = 5). In three cases, the male dominated paternity in one trial but not the other (twice as male 2 and once as male 1). In the fourth case, the male dominated both times (once in each position), and in the fifth case, the male dominated in neither case (male 1 in both). Again, these observed patterns are similar to ones that would be expected by chance. Latency to copulation (time between male release and cop- ulatory mount) averaged 3.1 min (range: 0.1-49.1 min), and the mean duration of copulation (time between mount and thrust) was 2.5 min (range: 0.7-4.6 min). We compared the within-trial, between-male differences in latency and duration to the differ- ence in proportion of paternity (arcsine transformed), but there was no significant relationship between either variable and fertil- ization success. Combining the 24-h and 72-h trials, fertilization was domi- nated by the male that was less closely related to the female in 10 cases and by the male that was more closely related to the female in 9. In two cases, there was no difference in relatedness between the female and the two males in a trial. Differences in the related- ness of the female to the second male and the first were not corre- lated with P2 for the 24-h trials (r, = -0.12, P = 0.75), the 72-h trials (r, = -0.30, P = 0.30), or both sets of trials combined (r, = -0.24, P=0.28). DISCUSSION Passive sperm loss is inevitable, because the alternative would re- quire immortal sperm; therefore, the real questions in investigat- ing passive sperm loss are the rate at which loss occurs and the importance of this rate in relation to other processes. For Mus- covy Ducks, we estimate that sperm are lost from the female re- productive tract at a rate of-0.024 h_1. This rate falls in the middle of the range estimated in other species of birds: considerably be- low the rate (0.053 Ir1) in Bearded Tits (Panurus biarmicus; Sax et al. 1998), considerably above the rate (0.003 Ir1) in Wild Turkeys (Meleagris gallopavo; Wishart 1988), and quite similar to the rate (0.026 IT1) in Zebra Finches (Taeniopygia guttata; Birkhead et al. 1993) and the rate (0.019 Ir1) in Collared Flycatchers (Ficedula albicollis; Michl et al. 2002). Complete sperm uptake took up to 4 days in our female Muscovy Ducks. Delayed uptake might lead to mixing of sperm from successive inseminations before they reach the sperm- storage tubules, thus diminishing the likelihood of stratification 500 STAI AND SEARCY AUK, VOL. 127 and increasing the degree of shared paternity within clutches. We had few cases in which paternity was shared by the two males in a sperm competition trial, however. Thus, delayed uptake does not seem to have played a large role in determining fertilization pat- terns in our study. Despite the fact that sperm loss definitely occurs in Muscovy Ducks, the pared-down version of the passive-sperm-loss model was only partially successful in explaining patterns of male pre- cedence in fertilization. The mean fertilization success of sec- ond males after a 24-h lag was quite close to that predicted by the model, but the mean success after a 72-h lag was much different from the observed and was outside the 95% CI. In addition, pas- sive sperm loss predicts that the degree of second-male advantage should be higher after a 72-h lag than after a 24-h lag, whereas we observed the opposite. Finally, the assumptions of the passive- sperm-loss model?that sperm from different inseminations mix within the female and are available in proportion to the numbers that remain alive at the time that an egg is fertilized?would lead one to expect more intermediate values of P2, clustered between 0.5 and 1.0, rather than the large number of the extreme values of 0 and 1 that we observed (Fig. 2). A bimodal pattern of fertilization success such as we observed has also been reported in other stud- ies of sperm precedence in birds, both with natural mating (War- ren and Kilpatrick 1929) and artificial insemination (Cunningham and Cheng 1999). Clearly, then, fertilization success in our study was strongly affected by one or more processes other than passive sperm loss. Some of the possible processes are male-driven. Both ejaculate size (the number of sperm per insemination) and ejaculate qual- ity (usually assessed by sperm motility) vary within and between male birds (Wishart and Palmer 1986, Birkhead et al. 1995a, Fro- man et al. 2002, Cornwallis and Birkhead 2007) and can have a large effect on fertilization success (Birkhead et al. 1999, Denk et al. 2005). Some of the within-male variation is attributable to a decline in ejaculate size and quality with repeated inseminations over short time intervals (Birkhead et al. 1995a); such short-term declines should not have been a factor in our experiments because we were careful to rest males for at least 2 days between matings. Nevertheless, ejaculate size and quality undoubtedly varied be- tween matings and could have played a considerable role in creat- ing variation in fertilization success. These male-driven processes, however, are unlikely to explain one of our more salient results: the lower success of second males with a 72-h lag than of those with a 24-h lag. Between-male differences in ejaculate size and quality cannot explain this trend because males alternated be- tween the first-male and second-male role for both mating inter- vals. Within-male variation is also unlikely to explain this trend, because males had no obvious means of assessing the length of the lag between matings and so should not have been able to adjust their ejaculate to the length of the lag. Females, by contrast, had every opportunity to assess the length of the lag between matings. Thus, female-driven processes may have caused the trend toward lower success after a longer lag, though at present we cannot explain why female preferences would vary as a function of mating interval. Female-driven pro- cesses may also have caused some or all of the wide variation in second-male success seen within trials with the same lag. One level at which females might influence success is through their behavior during copulation. In our study, a male released into a female's pen often approached the latter so rapidly that it was dif- ficult to discern during the event whether the female had moved into the receptive prone posture before being reached by the male. We were able to confirm female receptivity from video re- play, however, and in all trials the copulation appeared to be be- haviorally complete. Nevertheless, it is possible that females were able to affect insemination through changes in their behavior too subtle for us to discern. Another possibility is that females ex- ercise some form of postcopulatory choice. Much attention has been given to the possibility of postcopulatory choice in a variety of animal groups (Ward 2000, Bussiere et al. 2006, Rosengrave et al. 2008), including birds (Cunningham and Cheng 1999, Pizzari and Birkhead 2000, Denk et al. 2005, Birkhead and Brillard 2007). Postcopulatory choice has been controversial, in part because it is often difficult to see a mechanism by which postcopulatory choice could be exerted. In birds, one simple mechanism has been dem- onstrated: female domestic chickens are able to accomplish se- lective ejection of sperm from particular males after copulation (Pizzari and Birkhead 2000). Male waterfowl possess an intromit- tent organ (Coker et al. 2002), which may make sperm ejection more difficult for females (Denk et al. 2005); however, anatomi- cal specializations of the vagina found in some waterfowl species (Brennan et al. 2007) may prevent males from penetrating deeply into the female reproductive tract and thus maintain sperm ejec- tion as an option. Another possible mechanism for postcopulatory choice is through storage of the sperm of successive male part- ners in different regions of the female's reproductive tract. Har- vey and Parker (2000) showed through computer simulation that intraspecific variation in sperm precedence could arise through a lack of mixing of sperm during storage, and King et al. (2002) used labeled sperm to show that sperm from different inseminations were stored in different storage tubules in domestic chickens and turkeys. Segregation of ejaculates during storage might result in bimodal patterns of sperm precedence, in which either the first or the second male to mate dominates fertilization, even if females are unable to choose which male dominates. ^elatedness might provide a criterion for postcopulatory choice, with females biasing fertilization toward the sperm of less closely related males (Griffith and Immler 2009). We found no evi- dence of such a bias in our results, but given that our measure of re- latedness was based on a relatively small number of loci, we believe that the possibility of such a bias deserves further investigation. In conclusion, we have demonstrated that even though pas- sive sperm loss occurs in Muscovy Ducks at a rate typical for birds, last-male precedence is not always observed, which suggests that processes other than passive loss are also important. Male-driven variation in ejaculate size and quality undoubtedly has some im- portance in determining which male dominates fertilization, but the evidence also suggests a role for female-driven processes. Female-driven processes that might be at work in Muscovy Ducks include variation in female behavior during copulation, selective ejection of male sperm, and segregation of ejaculates in different storage areas in the female reproductive tract. Given the promis- cuous mating system of this species, these sperm competition processes may be especially important in this species in determin- ing patterns of mating success and, thus, the outcome of sexual selection. JULY 2010 SPERM COMPETITION IN MUSCOVY DUCKS 501 ACKNOWLEDGMENTS We are grateful to a number of people at the Smithsonian Conser- vation and Research Center (CRC) for their help. In particular, we thank S. Derrickson for sponsoring the work at the CRC; P. Dye and G. Tacheny for breeding and housing the birds; C. Emerick, W. Lynch, L. Ware, and G. Bolen for assistance with the experi- ments and bird care; and R. Spindler for help with microscopy. We thank C. Hughes and D. Williams of the University of Miami for help with the microsatellite analysis. Funding was provided by a Smithsonian Predoctoral Fellowship, a National Science Founda- tion Doctoral Dissertation Enhancement Grant, Sigma Xi Grants- in-Aid of Research, and the University of Miami. LITERATURE CITED AFTON, A. D. 1985. Forced copulation as a reproductive strategy of male Lesser Scaup: A field test of some predictions. Behaviour 92:146-167. BIRKHEAD, T. R. 1998. Sperm competition in birds: Mechanisms and function. Pages 579-622 in Sperm Competition and Sexual Selection (T. R. Birkhead and A. P. M0ller, Eds.). Academic Press, San Diego, California. BIRKHEAD, T. R., AND J. 0. BIGGINS. 1998. Sperm competition mech- anisms in birds: Models and data. Behavioral Ecology 9:255-260. BIRKHEAD, T. R., AND J. P. BRILLARD. 2007. Reproductive isolation in birds: Postcopulatory prezygotic barriers. Trends in Ecology and Evolution 22:266-272. BIRKHEAD, T. R., F. FLETCHER, E. J. PELLATT, AND A. STAPLES. 1995a. Ejaculate quality and the success of extra-pair copulations in the Zebra Finch. Nature 377:422-423. BIRKHEAD, T. R., J. G. MARTINEZ, T. BURKE, AND D. P. FROMAN. 1999. Sperm mobility determines the outcome of sperm compe- tition in the domestic fowl. Proceedings of the Royal Society of London, Series B 266:1759-1764. BIRKHEAD, T. R., AND A. P. MOLLER. 1992. Sperm Competition in Birds: Evolutionary Causes and Consequences. Academic Press, London. BIRKHEAD, T. R., AND A. P. MOLLER, EDS. 1998. Sperm Competition and Sexual Selection. Academic Press, San Diego. BIRKHEAD, T. R., E. J. PELLATT, AND F. FLETCHER. 1993. Selection and utilization of spermatozoa in the reproductive tract of the female Zebra Finch Taeniopygiaguttata. Journal of Reproduction and Fertilty 99:593-600. BIRKHEAD, T. R., AND M. PETRIE. 1995. Ejaculate features and sperm utilization in peafowl Pavo cristatus. Proceedings of the Royal Society of London, Series B 261:153-158. BIRKHEAD, T. R., G. J. WISHART, AND J. 0. BIGGINS. 1995b. Sperm precedence in the domestic fowl. Proceedings of the Royal Society of London, Series B 261:285-292. BRENNAN, P. L. R., R. O. PRUM, K. G. MCCRACKEN, M. D. SOREN- SON, R. E. WILSON, AND T. R. BIRKHEAD. 2007. Coevolution of male and female genital morphology in waterfowl. PLOS One 2:e418 BRILLARD, J. P., AND H. ANTOINE. 1990. Storage of sperm in the uterovaginal junction and its incidence on the numbers of sper- matozoa present in the perivitelline layer of hens' eggs. British Poultry Science 31:635-644. BRILLARD, J. P., AND M. R. BAKST. 1990. Quantification of sperma- tozoa in the sperm-storage tubules of turkey hens and the rela- tion to sperm numbers in the perivitelline layer of eggs. Biology of Reproduction 43:271-275. BUSSIERE, L. E, J. HUNT, M. D. JENNIONS, AND R. BROOKS. 2006. Sexual conflict and cryptic female choice in the black field cricket, Teleogryllus commodus. Evolution 60:792-800. CHENG, K. M., J. T. BURNS, AND F. MCKINNEY. 1983. Forced cop- ulation in captive Mallards. III. Sperm competition. Auk 100: 302-310. CLAYTON, G. A. 1984. Muscovy Duck. Pages 340-344 in Evolution of Domesticated Animals (I. L. Mason, Ed.). Longman, London. COKER, C. R., F. MCKINNEY, H. HAYS, S. V. BRIGGS, AND K. M. CHENG. 2002. Intromittent organ morphology and testis size in relation to mating system in waterfowl. Auk 119:403-413. COLEGRAVE, N., T. R. BIRKHEAD, AND C. M. LESSELLS. 1995. Sperm precedence in Zebra Finches does not require special mecha- nisms of sperm competition. Proceedings of the Royal Society of London, Series B 259:223-228. COMPTON, M. M., H. P. VAN KREY, AND P. B. SIEGEL. 1978. The fill- ing and emptying of the uterovaginal sperm-host glands in the domestic hen. Poultry Science 57:1696-1700. CORNWALLis, C.!(., AND T. R. BIRKHEAD. 2007. Changes in sperm quality and numbers in response to experimental manipulation of male social status and female attractiveness. American Natu- ralist 170:758-770. CRAWFORD, R. 0.1990. Origin and history of poultry species. Pages 1-41 in Poultry Breeding and Genetics (R. D. Crawford, Ed.). Elsevier, Amsterdam. CUNNINGHAM, E. J. A. 2003. Female mate preferences and subse- quent resistance to copulation in the Mallard. Behavioral Ecology 14:326-333. CUNNINGHAM, E. J. A., AND K. M. CHENG. 1999. Biases in sperm use in the Mallard: No evidence for selection by females based on sperm genotype. Proceedings of the Royal Society of London, Series B 266:905-910. DELACOUR, J. 1959. The Waterfowl of the World, vol. 3. Country Life, London. DENK, A. G., A. HOLZMANN, A. PETERS, E. L. M. VERMEIRSSEN, AND B. KEMPENAERS. 2005. Paternity in Mallards: Effects of sperm quality and female sperm selection for inbreeding avoid- ance. Behavioral Ecology 16:825-833. DUNN, P. O., A. D. AFTON, M. L. GLOUTNEY, AND R. T. ALISAU- SKAS. 1999. Forced copulation results in few extrapair fertil- izations in Ross's and Lesser snow geese. Animal Behaviour 57:1071-1081. FROMAN, D. P., T. PIZZARI, A. J. FELTMANN, H. CASTILLO-JUAREZ, AND T. R. BIRKHEAD. 2002. Sperm mobility: Mechanisms of fertilizing efficiency, genetic variation and phenotypic relation- ship with male status in the domestic fowl, Gallus gallus domes- ticus. Proceedings of the Royal Society of London, Series B 269: 607-612. GINSBERG, J. R., AND U. W HUCK. 1989. Sperm competition in mam- mals. Trends in Ecology and Evolution 4:74-79. GRIFFITH, S. C, AND S. IMMLER. 2009. Female infidelity and genetic compatibility in birds: The role of the genetically loaded raffle in understanding the function of extrapair paternity. Journal of Avian Biology 40:97-101. 502 STAI AND SEARCY AUK, VOL. 127 GRIFFITH, S. C, I. P. F. OWENS, AND K. A. THUMAN. 2002. Extra pair paternity in birds: A review of interspecific variation and adaptive function. Molecular Ecology 11:2195-2212. HARVEY, I. F., AND G. A. PARKER. 2000. 'Sloppy' sperm mixing and intraspecific variation in sperm precedence (P2) patterns. Proceedings of the Royal Society of London, Series B 267:2537- 2542. KING, L. M., J. P. BRILLARD, W. M. GARRETT, M. R. BAKST, AND A. M. DONOGHUE. 2002. Segregation of spermatozoa within sperm storage tubules of fowl and turkey hens. Reproduction 123: 79-86. LESSELLS, C. M., AND T. R. BIRKHEAD. 1990. Mechanisms of sperm competition in birds: Mathematical models. Behavioral Ecology and Sociobiology 27:325-337. MAAK, S., K. WIMMERS, S. WEIGEND, AND K. NEUMANN. 2003. Isolation and characterization of 18 microsatellites in the Peking Duck (Anas platyrhynchos) and their application in other water- fowl species. Molecular Ecology Notes 3:224-227. MCKINNEY, F., K. M. CHENG, AND D. J. BRUGGERS. 1984. Sperm competition in apparently monogamous birds. Pages 523-545 in Sperm Competition and the Evolution of Animal Mating Systems (R. L. Smith, Ed.). Academic Press, Orlando, Florida. MCKINNEY, E, S. R. DERRICKSON, AND P. MINEAU. 1983. Forced copulation in waterfowl. Behaviour 86:250-294. MCKINNEY, E, AND S. EVARTS. 1998. Sexual coercion in water- fowl and other birds. Pages 163-195 in Avian Reproductive Tac- tics: Female and Male Perspectives (P. G. Parker and N. T. Burley, Eds.). Ornithological Monographs, no. 49. MICHL, G., J. TOROK, S. C. GRIFFITH, AND B. C. SHELDON. 2002. Experimental analysis of sperm competition mechanisms in a wild bird population. Proceedings of the National Academy of Sciences USA 99:5466-5470. ONO, T., M. T. SIVA-JOTHY, AND A. KATO. 1989. Removal and sub- sequent ingestion of rival's semen during copulation in a tree cricket. Physiological Entomology 14:195-202. ORING, L. W., AND R. D. SAYLER. 1992. The mating systems of waterfowl. Pages 190-213 in Ecology and Management of Breed- ing Waterfowl (B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, Eds.). Uni- versity of Minnesota Press, St. Paul. PARKER, G. A. 1970. Sperm competition and its evolutionary conse- quences in the insects. Biological Review 45:525-567. PIZZARI, T, AND T. R. BIRKHEAD. 2000. Female feral fowl eject sperm of subdominant males. Nature 405:787-789. PIZZARI, T., H. LOVLIE, AND C. K. CORNWALLIS. 2004. Sex- specific, counteracting responses to inbreeding in a bird. Proceed- ings of the Royal Society of London, Series B 271:2115-2121. RAUD, H., AND J. M. FAURE. 1990. Rhythmic occurrence of sexual behaviour and egg laying activity of Muscovy Ducks. British Poul- try Science 31:23-32. RIDLEY, M. 1989. The incidence of sperm displacement in insects: Four conjectures, one corroboration. Biological Journal of the Linnean Society 38:349-367. ROMANOFF, A. L., AND A. J. ROMANOFF. 1949. The Avian Egg. Wiley, New York. ROSENGRAVE, P., N. J. GEMMELL, V. METCALF, K. MCBRIDE, AND R. MONTGOMERIE. 2008. A mechanism for cryptic female choice in chinook salmon. Behavioral Ecology 19:1179-1185. SAX, A., H. HOI, AND T. R. BIRKHEAD. 1998. Copulation rate and sperm use by female Bearded Tits, Panurus biarmicus. Animal Behaviour 56:1199-1204. SIMMONS, L. W. 2001. Sperm Competition and Its Evolutionary Consequences in the Insects. Princeton University Press, Princ- eton, New Jersey. SORENSON, L. G. 1994. Forced extra-pair copulation and mate guard- ing in the White-cheeked Pintail: Timing and trade-offs in an asynchronously breeding duck. Animal Behaviour 48:519-533. STAI, S. M. 2004. Promiscuity and sperm competition in Mus- covy Ducks, Cairina moschata. Ph.D. dissertation, University of Miami, Miami, Florida. STAI, S. M., AND C. R. HUGHES. 2003. Characterization of mic- rosatellite loci in wild and domestic Muscovy Ducks (Cairina moschata). Animal Genetics 34:387-389. THUMAN, K. A., AND S. C. GRIFFITH. 2005. Genetic similarity and the nonrandom distribution of paternity in a genetically highly polyandrous shorebird. Animal Behaviour 69:765-770. TODD, F. S. 1996. Natural History of the Waterfowl. Ibis, San Diego, California. VON HELVERSEN, 0., AND O. VON HELVERSEN. 1991. Pre-mating sperm removal in the bushcricket Metaplastes ornatus Ramme 1931 (Orthoptera, Tettigonoidea, Phaeneropteridae). Behavioral Ecology and Sociobiology 28:391-396. WAAGE, J. K. 1979. Dual function of the damselfly penis: Sperm removal and transfer. Science 203:916-918. WANG, J. 2002. An estimator for pairwise relatedness using molecu- lar markers. Genetics 160:1203-1215. WARD, P. I. 2000. Cryptic female choice in the yellow dung fly Scathophaga stercoraria (L.). Evolution 54:1680-1686. WARREN, D. C, AND L. KILPATRICK. 1929. Fertilization in the domestic fowl. Poultry Science 8:237-256. WESTNEAT, D. E, AND I. R. K. STEWART. 2003. Extra-pair paternity in birds: Causes, correlates, and conflict. Annual Review of Ecol- ogy and Systematics 34:365-396. WISHART, G. J. 1987. Regulation of the length of the fertile period in the domestic fowl by numbers of oviductal spermatozoa, as reflected by those trapped in laid eggs. Journal of Reproduction and Fertility 80:493-498. WISHART, G. J. 1988. Numbers of oviductal spermatozoa and the length of the fertile period in different avian species. Proceeding of the 11th International Congress of Animal Reproduction and Artificial Insemination 3:362-364. WISHART, G. J., AND F. H. PALMER. 1986. Correlation of the fertilis- ing ability of semen from individual male fowls with sperm motil- ity and ATP content. British Poultry Science 27:97-102. Associate Editor: R. D. Dawson