Behavioral Ecology Vol. 13 No. 6: 800-807
© 2002 International Society for Behavioral Ecology
Genetic analysis of the mating system and opportunity for sexual selection in northern water snakes (Nerodia sipedon)
a Department of Biology, McMaster University, Hamilton, Ontario L8S 4K1, Canada b Department of Biology, Carleton University, 1125 Colonel By Drive, Ottawa, Ontario K1S 5B6, Canada
Address correspondence to P.J. Weatherhead, who is now at the Program in Ecology and Evolutionary Biology, University of Illinois, Shelford Vivarium, 606 E. Healey, Champaign, IL 61820, USA. E-mail: pweather{at}uiuc.edu. H.L. Gibbs is now at the Department of Evolution, Ecology and Organismal Biology, Botany and Zoology Building, The Ohio State University, 1735 Neil Avenue, Columbus, OH 43210-1293, USA. G.P. Brown is now at the School of Biological Sciences, A08, University of Sydney, Sydney, NSW, Australia.
Received 2 January 2001; revised 26 October 2001; accepted 12 March 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We used data collected over 3 years at two study sites to quantify the rates and consequences of multiple paternity and to determine the opportunity for selection on male and female northern water snakes (Nerodia sipedon). We sampled litters from 45 females that gave birth to 811 offspring. Using eight microsatellite DNA loci (probability of exclusion of nonparental males > 0.99), we assigned paternity to 93% of neonates from one study population and 69% of neonates from the other population. Observations of participation in mating aggregations predicted individual reproductive success poorly for two reasons. First, males regularly courted nonreproductive females. Second, more than half of all sexually mature males obtained no reproductive success each year, despite the fact that many of them participated in mating aggregations. The number of sires per litter ranged from one to five, with 58% of all litters sired by more than one male. Multiple paternity increased with female size, apparently both because bigger females mated with more males and because the larger litters of big females provide paternity opportunities to more males. Multiple paternity was also more prevalent in years with shorter mating seasons. We detected no advantage to multiple paternity in reducing either the number of unfertilized ovules or stillborn young. Despite the majority of males siring no young each year, some males fathered young with as many as three different females in one year. Male reproductive success increased by more than 10 offspring for each additional mate, whereas female success increased by fewer than 2 offspring for each additional mate. The opportunity for sexual selection was more than five times higher in males than females.
Key words: mating system, microsatellites, multiple paternity, Nerodia sipedon, paternity analysis, sexual selection, snakes.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Snakes are poorly represented as subjects of sexual selection research (Duvall et al., 1993
;
Olsson and Madsen, 1998)
because their cryptic nature usually makes systematic observation of
individually marked animals impossible. Molecular methods to determine
parentage of offspring promise to greatly enhance the feasibility of studying
sexual selection in snakes, as they have for other taxa
(Gibbs and Weatherhead, 2001).
Snakes possess a unique suite of traits that suggests they will be valuable
model organisms for studying sexual selection. For example, multiple mating by
females and multiple paternity of litters appear to be common in snakes
(Olsson and Madsen, 1998).
Unlike for many other vertebrates in which these phenomena occur (e.g.,
birds), however, identifying the proximate and ultimate factors influencing
multiple mating in snakes should be simplified by the lack of social bonds
between mating individuals and by the lack of parental care by either sex.
Competition among male snakes for access to females is intense in many species
(e.g., Devine, 1977;
Madsen and Shine, 1993;
Weatherhead et al., 1995).
Unlike in most other taxa, however, strong competition among males is not
always associated with male-biased size dimorphism in snakes
(Olsson and Madsen, 1998), and
male snakes do not exhibit any elaborate ornamentation that is often
associated with strong sexual selection. As a first step toward using
DNA-based measures of paternity to investigate these phenomena in snakes, we
used microsatellite DNA markers to describe the mating system of northern
water snakes (Nerodia sipedon) and to quantify sexual selection
(Arnold and Wade, 1984). We
also to began exploring the causes and consequences of sexual selection in
this species.
Several attributes of the population of northern water snakes that we
studied made them particularly appropriate for our purposes. First, the snakes
we studied occurred in relatively small, isolated marsh populations
(Prosser et al., 1999;
Weatherhead et al., 1995),
which allowed us to sample most of the adult population. This increases the
potential success of determining parentage when using genetic techniques.
Second, previous observational studies of this population suggested that
sexual selection through competition between males was strong
(Brown and Weatherhead, 1999b;
Weatherhead et al., 1995). In
addition to guiding the development of specific hypotheses to be tested (see
below), this past research created the unique opportunity for us to determine
how well behavioral indices of mating success agree with genetically
determined measures of reproductive success, something not previously done for
snakes. Finally, a previous study using the same population demonstrated that
multiple paternity occurs frequently in this species
(Barry et al., 1992), so N.
sipedon was an appropriate subject for addressing questions about this
phenomenon.
Our first goal was to address the technical issue of the extent to which
the use of microsatellite DNA markers would alter previous estimates of
reproductive patterns from the same water snake population. For male water
snakes, previous estimates of reproductive success based on behavior could be
unreliable for several reasons (e.g., disrupted copulations, unobserved
copulations, sperm competition). Although female reproductive success can be
documented accurately without genetic methods, the previous study of multiple
paternity used allozyme-based genetic markers
(Barry et al., 1992), which
precluded determination of the exact number of sires per litter and the number
of young sired by each male.
Our second goal was to identify the factors that contribute to variation in
multiple paternity. Although there is no a priori reason to expect that the
proclivity to mate with multiple males should vary with female attributes such
as size, males might pursue larger females more vigorously because larger
females are more fecund and thus provide males greater potential for
reproductive success (Brown and
Weatherhead, 1997; Duvall et
al., 1993). If the number of males with which a female mates is
related to the number of males that court her, then larger females should more
often give birth to multiply sired litters. Alternatively, large females might
be more likely to produce multiply sired litters because, for a given number
of males mated, the females' larger litters provide opportunities for more of
those males to sire young.
At the population level, the occurrence of multiple mating, and thus
multiple paternity, might depend on the number of males encountered by
receptive females. We investigated three possible factors influencing
encounter rates. First, the operational sex ratio (OSR) is male biased in
northern water snakes but can vary substantially between years
(Brown and Weatherhead, 1999b;
Weatherhead et al., 1995). We
tested the prediction that, as the OSR becomes more male biased, the incidence
of multiple paternity increases. Second, we tested the prediction that
multiple paternity would be more prevalent in years when the mating season is
relatively protracted. Female water snakes can remain reproductively active
throughout the mating season (Weatherhead
et al., 1995), so we assumed that, with more time available,
females should encounter more potential mates. Third, we tested the prediction
that multiple paternity should be more common when the spatial distribution of
reproductive females is both patchy and predictable. Previous research has
shown that females are four times more clumped in one of our study populations
than in the other, which results in females being courted by more males
(Brown and Weatherhead,
1999b).
Our third goal was to determine, at least proximately, whether multiple
mating has any adaptive consequence for females. We considered just two of the
hypotheses that have been proposed to explain why females of many species mate
with multiple males (Andersson,
1994). First, if a single mating is not sufficient to fertilize
all a female's ova, then females that mate with more than one male should have
fewer unfertilized ovules than females that mate once
(Gromko et al., 1984). Second,
multiple mating may be a bet-hedging strategy to increase the genetic quality
of offspring. In an inbred population of European adders (Vipera
berus), Madsen et al.
(1992) found that multiple
mating reduced the proportion of young that were stillborn. They suggested
that multiple mating could promote intrauterine sperm competition, whereby
sperm of higher quality males are more likely to fertilize eggs. Although our
northern water snake populations are not inbred, there is reduced gene flow
between pond populations (Prosser et al.,
1999). Therefore, we examined whether multiple paternity reduces
the occurrence of stillborn young.
Our final goal was to use the data on reproductive success and mating
patterns to determine the mating system, and hence the potential for sexual
selection, in northern water snakes. Determination of the intensity of sexual
selection requires measures of the true reproductive success of all
individuals in the population (Duvall et
al., 1993). We generated these estimates by documenting the mean
number of mates for males and females and then determining how reproductive
success varies with the number of mates for both sexes (i.e., sexual selection
gradients; Arnold and Duvall,
1994). Because individual males have been observed courting as
many as eight different females in the same season
(Brown and Weatherhead, 1999b),
we predicted that the potential for sexual selection should be much greater in
males than in females.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sampling individuals
We studied water snakes from 1994 to 1996 in two spring-fed beaver ponds (Barbs Marsh and Beaver Marsh) located approximately 1 km apart and 10 km from the Queen's University Biological Station in eastern Ontario, Canada (45°37' N, 76°13' W). Brown and Weatherhead (1999a
) estimated mean
population sizes as 101 and 84 for Barbs and Beaver Marshes, respectively.
Beginning in late April when the snakes emerge from hibernation, and
continuing through May (the mating season) until early September, we searched
both marshes and attempted to capture all unmarked snakes (see below) by hand.
We conducted searches twice each day during the mating season and at least
twice a week thereafter.
We measured mass, total length and tail length, and determined the
snout-vent length (SVL) of each snake by subtracting tail length from total
length. We sexed each snake by probing its cloaca to detect the presence of
hemipenes. We collected a sample of approximately 100 µl of blood from the
caudal vein in the tail and stored the sample in 800-1000 µl of Queen's
lysis buffer (Seutin et al.,
1991). Snakes were permanently identified by implanting a passive
integrated transponder (PIT tag, Anitech) subcutaneously. We also marked
snakes captured for the first time each season with a unique three-color code
using spots of nontoxic acrylic paint, which allowed us to identify
individuals in the field with minimal disturbance
(Barry et al., 1992). We
occasionally recaptured marked individuals during the mating season to reapply
color marks that were fading.
In late July and early August we attempted to capture all adult females in
both marshes. Once these females gave birth or it became apparent that they
were not gravid, we released them where they had been captured. We housed
females individually or in pairs in fiberglass tanks with plastic carpeting
covering the floor, in a room with an ambient temperature of 25-33°C.
Heating rocks were rotated among the tanks so females could thermoregulate
regularly. Females always had access to water and were fed minnows ad libitum
twice a week. After parturition we collected approximately 50 µl of blood
from the caudal vein of each neonate. We removed the terminal 1 cm of the tail
from young that were stillborn. We stored blood and tissue samples in lysis
buffer (Seutin et al., 1991)
in a freezer. We released all live neonates at their mother's marsh.
Behavioral observations
While searching for snakes during the mating season, we also documented
reproductive behavior. In Barbs Marsh the substrate appropriate for basking
and mating occurs throughout the marsh, whereas the only appropriate substrate
in Beaver Marsh occurs around the periphery of the marsh. This difference
accounts for the much more clumped distribution of females in Beaver Marsh
during the mating season (Brown and
Weatherhead, 1999b). We modified our method of searching the two
marshes to maximize the likelihood of finding mating aggregations.
We considered a female and male(s) to be in a mating aggregation when we
observed at least one male courting the female. Courtship involves
stereotypical behaviors by the male including chin-tapping, body jerking and
alignment of the body to bring the male's cloaca into contact with the
female's cloaca (Mushinsky,
1979; Perry-Richardson et al.,
1990). We attempted to identify all the participants in
aggregations, but to avoid disturbing the snakes, we did not approach closer
than several meters. Because this made it impractical to determine whether
intromission was occurring, we simply scored the number of times each
individual was observed in mating aggregations, as has been done previously
(Brown and Weatherhead, 1999a;
Weatherhead et al., 1995).
Genetic analysis and paternity assignments
We used eight microsatellite DNA loci to assign paternity. Six of these
loci (Nsµ 2, 3, 4, 6, 9b, and 10) have been described previously
(Prosser et al., 1999). The
other two loci (Nsµ 110 and 119) are reported in Prosser
(1999). Locus-specific
probabilities of correctly excluding males that were not the father of a given
neonate (Westneat et al.,
1987) ranged from 0.21 to 0.86, with a combined probability across
all eight loci of >0.99 in both marshes
(Prosser, 1999;
Prosser et al., 1999).
Detailed genotyping protocols are given in Prosser et al.
(1999). For efficiency, we
initially genotyped all adult males, mothers that gave birth, and all neonates
at the four loci (Nsµ 2, 3, 9b, and 10) with the highest probabilities of
exclusion (all single locus PE 
0.52). We then
genotyped offspring that could not be assigned to a single father, their
mothers, and all other putative fathers (i.e., males not excluded in the
initial screening) at the other four loci, using each locus as necessary until
we excluded all but one male. However, for litters in which all but one or two
neonates were assigned to a single male using the first four loci, we
genotyped the entire litter using the remaining four loci to confirm
assignments. Note that on the basis of an observed heterozygote deficiency,
Prosser et al. (1999)
suggested that Nsµ 4 had a null allele(s), which was subsequently confirmed
by analyzing patterns of segregation between mothers and their young
(Prosser, 1999). Therefore, we
only used genotypes at this locus for exclusions when both the parents and
offspring being compared were identifiable heterozygotes and hence did not
have null alleles. When assigning paternity we only considered males sampled
within a study site as potential sires for neonates produced from that site
because movement of snakes between the two sites was highly improbable.
However, for neonates from a given year, we considered all males sampled at
the study site across all years as potential sires, even if we had not
observed some of those males that year.
Data analysis
For several reasons we decided that when pooling data across years, it was
appropriate to treat individuals present in more than one year as independent
observations. First, because water snakes continue to grow after sexual
maturity, most attributes of interest (e.g., body size, fecundity) of
individuals sampled twice change between samples. Second, our sample sizes
were rather large and the numbers of individuals sampled more than once were
relatively small (2 of 43 females, 55 of 150 males). Because the F
distribution is rapidly asymptotic, the analytical danger of pseudoreplication
(artificially increasing error degrees of freedom) is minimal (e.g., for
females, a change in error df from 43 to 41 corresponds to a change in
critical F value of approximately 0.03;
Leger and Didrichsons, 1994).
Third, among males, for which the proportion of resampled individuals is
highest, the variance in reproductive success was higher within than among
resampled individuals (F = 0.66, df = 54,67, p = .94),
suggesting that pooling data should not bias our results
(Leger and Didrichsons,
1994).
We inspected plots and distributions of residuals from all parametric analyses and detected no significant violations of the assumptions of normality and homogeneity of variance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Over the 3 years of the study we brought 81 females into captivity. One female died and one female gave birth to an entire litter of undeveloped follicles; we excluded both from subsequent analyses. Of the remaining 79 females, 45 (57%) gave birth to a total of 798 live (mean = 17.7, SD = 4.9, range 4-28) and 13 dead (i.e., stillborn or partially developed embryos) offspring (mean = 0.29, SD = 0.59, range = 0-2, Table 1). The overall mean litter size (i.e., live and dead combined) was 18.0 (SD = 4.7, range = 5-28) offspring. One additional female died in captivity before giving birth, and we dissected out 16 partially developed offspring from which we were able to extract DNA. We also found a gravid female dead in the field and were able to extract DNA from 19 of her 22 embryos. We used data from these two dead females only in our determination of levels of multiple paternity.
|
Litter sizes and numbers of live and dead offspring did not differ significantly among years in Barbs (all F < 0.53, df = 1,19, all p > .48) or Beaver Marsh (all F < 1.84, df = 1,20, all p > .19). Because only one female from Barbs Marsh gave birth in 1995, we excluded 1995 for the comparison of litter sizes across years. After pooling data for each marsh across years, we found no significant differences in litter sizes or numbers of live or dead offspring between the two sites (all F < 0.29, all p > .60). Therefore, for analyses other than those specifically examining differences in rates of multiple paternity in the two marshes, we pooled the data for all litters across years and sites. Using the first four loci (Nsµ 2, 3, 9b, and 10), we were able to exclude all but one male for 290 of the 414 (70.0%) neonates from Barbs Marsh and 207 of the 432 (47.9%) neonates from Beaver Marsh. With an additional four loci (Nsµ 4, 6, 100, and 119), we increased the percentage of neonates assigned to a single father to 92.8% in Barbs Marsh and 68.8% in Beaver Marsh. Less than 2% of these assignments involved cases in which two males matched all the paternal alleles and thus could not be distinguished. In these cases we assigned paternity to the male that had sired other offspring in the same litter, which always allowed us to exclude one of the two males. For all unassigned neonates we excluded all potential fathers we had sampled. For litters with unassigned neonates, we estimated the minimum number of sires for the litter from the number of paternal alleles at the locus with the most paternal alleles present.
Based on numbers of paternal alleles detected, we determined that at least five males from Barbs Marsh and 17 males from Beaver Marsh collectively sired the unassigned neonates. The mothers of those offspring provide clues regarding why those males may have avoided capture. Of the 10 females that produced offspring sired by unknown males at Beaver Marsh, 8 were captured for the first time the year they gave birth, and 7 of the 8 were captured at the end of, or after, the mating season. Thus, these females may have mated with males in one of the nearby wetlands before moving to Beaver Marsh. In contrast, at Barbs Marsh, which is more remote from other wetlands than Beaver Marsh, all three females that produced offspring with unknown sires were long-term residents of the marsh. The unknown sires for these three litters seem likely to represent cases where the males were present in the marsh but avoided capture.
We identified 46 males that sired offspring over the 3 years of the study.
The smallest reproductively successful male had an SVL of 43.3 cm, which is
similar to the estimated minimum size at sexual maturity of 42.5 cm based on
mating behavior (Weatherhead et al.,
1995). Therefore, we assumed that all males longer than 43.3 cm
SVL were sexually mature and thus capable of siring offspring. Based on these
criteria, the number of mature males in Barbs and Beaver Marshes from 1994 to
1996 ranged from 29 to 38 and from 33 to 47, respectively.
The 45 females who reproduced (Table
1) ranged in SVL from 63.3 to 83.1 cm. Our minimum size for sexual
maturity for females is substantially larger than the minimum size estimated
by Weatherhead et al. (1995)
from mating behavior (63.3 vs. 55 cm). In contrast to males, female water
snakes may not reproduce in consecutive years
(Brown and Weatherhead, 1999a;
Feaver, 1977). Of the 14
females that gave birth and that we brought into captivity the next year, none
gave birth in the second year. Therefore, we estimated the number of sexually
active females by the number of females seen mating in the field plus the
number of females not seen mating but that gave birth. The number of
reproductive females estimated this way ranged from 4 to 16 in Barbs Marsh and
from 16 to 17 in Beaver Marsh over the 3 years.
Behavioral vs. genetic estimates of reproductive success
Pooling data from the two populations, we estimated that 85, 70, and 62
males were sexually active in 1994, 1995, and 1996, respectively. Of these 217
males, we only observed 92 (42%) courting or mating with a female in a total
of 120 mating aggregations observed over the study. Of the 92 males observed
attempting to mate, only 30 (32%) sired offspring. In total, 46 (21%) males
achieved some reproductive success. Thus, 16 (35%) males that sired young were
never seen in a mating aggregation. Furthermore, nearly half the mature males
(48.8%) known to be present were neither seen mating nor reproduced.
Nonetheless, males observed in mating aggregations were more likely to sire
offspring than males not seen in aggregations (
2 = 9.11, df =
216, p < .0025), so behavioral data had some value in predicting
which males sired young. The number of offspring sired by individual males was
not significantly correlated with the number of times individuals were seen in
mating aggregations (F = 3.05, df = 1,211, r = -0.12,
p = .08) or with the number of different females that males were seen
with in aggregations (F = 0.13, df = 1,91, r = -.04,
p = .73).
Over the 3 years, 45 (57%) of 79 females brought into captivity gave birth
(excluding the female that died and the female that produced entirely
undeveloped follicles). Only 22 of these 45 reproductive females (49%) were
observed in mating aggregations, so 23 females that produced offspring were
never observed mating. Similarly, 15 females that we had seen in mating
aggregations did not give birth. Thus, unlike males, females observed in
mating aggregations were no more likely to give birth than females not seen in
aggregations (2 = 0.18, df = 78, p = .67). In
addition, the number of offspring a female produced was not significantly
correlated with either the number of times she was seen in mating aggregations
(F = 0.53, df = 1,43, r = .11, p = .47) or with the
number of different males seen courting her (F = 2.21, df = 1,43,
r = .22, p = .15).
Finally, we considered the identities of males and females seen together in mating aggregations and determined the frequency with which those individuals produced offspring together. We observed 156 dyads of known males and females in the field. Of these, only seven (4.5%) produced offspring together.
Patterns of multiple paternity
Over all years, 58% of litters were sired by more than one male. Although
one litter had five different sires, nearly 80% of all litters were sired by
either one or two males (Figure
1). The proportion of offspring sired by each male in multiply
sired litters ranged from 0.05 to 0.95.
|
We predicted that multiple paternity would be higher in litters of larger
females. Consistent with our prediction, the number of sires per litter
increased significantly with maternal size (F = 13.90, df = 1,43,
r = .49, p = .0006;
Figure 2). This pattern could
be a consequence of bigger females mating with more males. Alternatively,
because bigger females produce larger litters
(Weatherhead et al., 1999),
multiple paternity might increase with female size because larger litters
provide opportunities for more males to sire young. When both litter size and
maternal SVL were included in an ANOVA model, using data for all females
(N = 45), only maternal size explained significant variation in the
number of sires (SVL: F = 6.51, df = 1,42, p = .014; litter
size: F = 0.60, p = .44). When the same analysis was
restricted to the 26 females that produced multiply sired litters, however,
only litter size explained significant variation in the number of sires (SVL:
F = 2.56, p = .12; litter size: F = 5.78, df =
1,23, p = .028). These results suggest that, relative to smaller
females, bigger females copulated with more males, but thereafter, chance
tended to determine which of the males that mated sired offspring.
|
We predicted that multiple paternity should have been more prevalent in longer mating seasons and when the OSR was more male biased. The length of the mating season (the interval between the first and last day we observed mating aggregations) ranged from 11 to 29 days (Table 1). We estimated the OSR by dividing the number of sexually active females (females seen in mating aggregations and/or known to have reproduced) by the number of sexually active males (all males large enough to be sexually mature) in each marsh each year. OSR estimates ranged from 0.13 to 0.55. We found no correlation between the proportion of multiply sired litters and OSR (Spearman rank correlation, rs = .49, p = .32), but multiply sired litters appeared to have been more prevalent when the mating season was shorter (Spearman rank correlation, rs = -.81, p = .05). The latter relationship was affected by the very brief mating season for Barbs Marsh in 1995 (11 days), when only one female gave birth (to a multiply sired litter). When we excluded this value, the correlation between the proportion of multiply sired litters and length of the mating season remained negative but was no longer significant (Spearman rank correlation, rs = -.67, p = .22).
Finally, we predicted that multiple paternity should be higher in Beaver
Marsh than in Barbs Marsh because females in Beaver Marsh are more predictably
located by males and are courted by more males, on average, than females in
Barbs Marsh (Brown and Weatherhead,
1999b). In total, 16 of the 23 (70%) litters from Beaver Marsh and
10 of the 22 (46.7%) litters from Barbs Marsh were sired by more than one male
(2 = 2.70, df = 44, p = .10). In Beaver Marsh,
litters were sired by a mean of 2.04 (SD = 0.88) males compared to a mean of
1.77 (SD = 1.11) males in Barbs Marsh (t = -0.91, df = 43, p
= .37). Although the difference was in the predicted direction, it was not
close to significant.
Consequences of multiple paternity
Pooling data across marshes and years, there was no significant difference
in the proportion (or absolute number) of unfertilized ovules between singly
and multiply sired litters (both t < 1.13, both p >
.26). There was also no significant difference in the mean proportion of
stillborn young (including partially developed embryos) between singly and
multiply sired litters (t = 0.65, df = 43, p = .52).
Mating system and sexual selection
In quantifying the mating system, we assumed that no sexually mature female
failed to reproduce because she could not find a mate (i.e., failure to
reproduce is a life-history decision for females rather than a sexual
selection issue). We consider this a reasonable assumption for several
reasons. The mobility of northern water snakes would allow the snakes to
circumnavigate either study marsh in a matter of hours, so in a mating season
of several weeks a female should have no problem finding a male. The
male-biased OSR indicates that there are always excess males available, and in
both marshes, mating locations were often quite predictable, further
increasing a female's ability to find males. In contrast, we assume that
sexually mature males should not willingly forgo reproduction in any given
year. Indeed, males that sired offspring in one year were no less likely than
males unsuccessful in the same year to be successful (2 =
2.05, df = 66, p = .15) or to be seen in mating aggregations
(2 = 1.69, df = 66, p = .19) the following year.
Males sired young with a mean of 0.32 and 0.26 females at Barbs Marsh and
Beaver Marsh, respectively. However, male reproductive success was highly
skewed (Figure 3), with the
proportion of males achieving no reproductive success ranging from 59% to 97%
among marshes and years. Considering only successful males, in Barbs Marsh
males sired young with a mean of 1.46 females, whereas in Beaver Marsh males
sired young with a mean of 1.15 females. The maximum number of mates for any
male was three.
|
A quantitative method for defining mating systems involves comparing sexual
selection gradients (the relationships between fecundity and the number of
mates) for males and females (Arnold and
Duvall, 1994). Reproductive success of male water snakes increased
significantly with the number of females with which a male sired young
(F = 615.38, df = 1,215, r = .68, p < .0001;
Figure 4). We estimate that for
each additional female mated, male success increased by an average of 11.41
offspring. If we include males that achieved no reproductive success, then
success increased by 10.08 offspring for each additional mate. For female
water snakes, reproductive success also increased with the number of males
that sired young in their litters (F = 7.29, df = 1,43, r =
.38, p = .01). However, if we control for female size, the effect of
number of sires becomes nonsignificant (F = 0.59, df = 1,42,
p = .37). Thus, the apparent increase in litter size with number of
sires is a consequence of larger females producing larger litters that are
sired by more males (Figure
2).
|
The potential for sexual selection to act on males relative to females is
estimated from their respective relative variances in reproductive success
(Is = variance in success divided by the square of mean
success). Relative variance in success provides an upper limit to the strength
of directional selection and has been termed the "opportunity for
selection" (Wade, 1979;
Wade and Arnold, 1980).
Following the same reasoning as above, we included mature males that achieved
no reproductive success in our calculation of Im and
excluded females that failed to reproduce from our estimate of
If. Values of both Im and
If varied between years and marshes, with the former
ranging from 2 to 9 times higher than the latter. Overall, the relative
opportunity for selection (Im divided by
If) was 5.21 times higher for males than for females.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our results demonstrated the value of using genetic markers to study sexual selection in snakes. Observations of mating behavior did not reliably predict reproductive success for either males or females. Previous research on this population of water snakes assumed that males only courted females that were going to reproduce that year (Barry et al., 1992
). Males sometimes ignore sexually mature females that are
later seen in aggregations, suggesting that the females advertised their
receptivity (presumably using pheromones) and thus are courted only when
receptive. However, we found that males regularly courted females that
subsequently failed to reproduce. Females may have copulated with males but
stored the sperm; sperm storage between years has been documented in many
species of snakes (e.g., Fox,
1956; Gist and Jones,
1987; Halpert et al.,
1982; Schuett,
1982), although not for water snakes. If females stored sperm,
then some males present in one year but not the next should have sired
offspring that second year. However, none of the 46 males identified as sires
involved this circumstance. Thus, misdirected male behavior seems a more
likely interpretation and is an important reason that male mating behavior
poorly predicted male reproductive success.
Two other factors also seem likely to have contributed to the unreliability
of behavioral estimates of male reproductive success. First, the facts that
some females that gave birth were never seen mating and that some males that
sired young were never captured means that our mating observations and
population samples were not as thorough as we had assumed. Capture records for
mothers of offspring with unknown sires suggested that part of the sampling
problem involved movements of snakes between wetlands, particularly at Beaver
Marsh. Second, even though females in this population regularly mate with
multiple males (Barry et al.,
1992; this study), assuming that all males in a mating aggregation
achieve some success is erroneous. Documenting intromission may have improved
our behavioral observations, but because little is known about factors
affecting sperm competition in snakes (e.g., timing and order of mating), even
observations of intromission success may be imperfect predictors of
reproductive success.
Our estimate that 58% of all northern water snake litters were multiply
sired is substantially less than the 86% estimated by Barry et al.
(1992) for the same population.
We found that multiple paternity declined in longer breeding seasons. In 1991,
the year Barry et al. (1992)
collected their data, the mating season in Barbs Marsh was relatively long
(Weatherhead et al., 1995), so
it seems unlikely that differences in length of breeding seasons explain the
difference between studies. We also found that multiple paternity increased
with litter size, which also was the case in Barry et al.'s
(1992) study (the 2 litters out
of 14 that only had a single sire were the second and third smallest in their
sample). The mean litter size in our study was 18.0 compared 22.7 in Barry et
al.'s (1992) study. Thus, the
smaller litters in our study might have contributed to the lower prevalence of
multiple paternity we documented. Indirect evidence suggested that the
increase in multiple paternity with litter size that we documented was a
consequence of two factors. First, bigger females appeared to have copulated
with more males relative to smaller females. Second, larger litters provide
more opportunities for paternity, so multiple paternity also occurred more
often in larger litters by chance.
We predicted that longer mating seasons should provide more mating
opportunities for females, thereby increasing multiple paternity. Although our
analyses lacked power (involving only five or six values), the results
indicated that longer mating seasons did not promote multiple paternity and
may, in fact, even have acted to decrease its incidence. Multiple paternity
could be more common in shorter seasons if the time between matings is also
shorter and if temporal proximity of inseminations somehow makes the outcome
of sperm competition among males more equitable. However, evidence from sand
lizards (Lacerta agilis; Olsson
et al., 1996) and European adders (Stille et al.,
1986,
1987) suggests that longer
intervals between inseminations increases multiple paternity. Timing of
ovulation relative to insemination could influence how the interval between
inseminations affects sperm competition
(Olsson and Madsen, 1998).
However, both adders and water snakes ovulate approximately 1 month after
mating (Aldridge, 1982;
Bauman and Metter, 1977;
Nilson, 1981), so on that
basis the relationship between the timing of inseminations and sperm
competition should be similar in the two species. An experimental approach may
be the most productive way to investigate this phenomenon.
We addressed two possibilities for why female water snakes mate with
multiple males. We found no support for the hypothesis that a single male does
not produce adequate sperm to fertilize all of a female's litter
(Gromko et al., 1984);
unfertilized ovules were rare and no more prevalent in singly sired than in
multiply sired litters. We also found no support for the hypothesis that
females mate with multiple males to reduce the costs of inbreeding
(Madsen et al., 1992). The
small number of undeveloped follicles and stillbirths rendered these weak
tests of the hypotheses (power to detect a significant difference at 
=
0.05 given observed effect and sample size = 0.10-0.27). Given the small
differences we found between singly and multiply sired litters, much larger
sample sizes would have been required to find significant differences.
Nonetheless, because the incidence of unfertilized follicles and stillborn
young was so small, we can conclude that any potential benefit from multiple
mating in reducing this incidence would necessarily also have to be small.
Furthermore, the occurrence of stillborn young is a blunt tool for detecting
possible genetic benefits of multiple mating. Olsson et al.
(1994) showed that multiple
mating can produce offspring that are more viable after they are born, a
possibility we could not explore with our data. Thus, it would be premature to
conclude that female water snakes do not realize a genetic benefit from
multiple mating.
Genetic evidence confirmed the suggestion from research on mating behavior
of northern water snakes that competition among males is intense
(Barry et al., 1992;
Brown and Weatherhead, 1999b;
Weatherhead et al., 1995).
Standardized variances in reproductive success
(Wade, 1979;
Wade and Arnold, 1980)
indicated that the opportunity for selection was more than five times greater
for males than for females. More than half of all sexually mature males never
sired young, and males that were successful regularly sired young with more
than one female. Thus, multiple mating by females was far from sufficient to
allow most males any reproductive success.
Our estimates of competition among males might change if seasonal estimates
of success poorly reflect lifetime success. However, there are several reasons
to expect this not to be the case. Over the 3 years of our study, only two
males sired young in more than one year. This was a consequence of some males
that were successful one year being unsuccessful in the other years, as well
as many other males only being present for one year as a mature individual.
Year-to-year survival of male water snakes is less than 50%
(Brown and Weatherhead, 1999a),
so many males only have one or two potential breeding seasons. Thus, estimates
of lifetime reproductive success seem unlikely to change our general
conclusion that competition among males is intense.
Our estimate that the opportunity for sexual selection due to variance in
reproductive success was approximately five times greater for males than
females is similar to the value for European adders, the only other snake
species for which this has been estimated. Madsen and Shine
(1994) calculated an
Im of 0.96 and an If of 0.17, giving a
relative value of 5.65. These values were based on estimates of lifetime
reproductive success. However, the value for males has the potentially serious
shortcoming of relying on mating success rather than on reproductive success.
Nonetheless, the opportunity for sexual selection in their study appeared to
be much greater for males than for females. Demonstrating variance in
reproductive success is not adequate to imply that selection occurs on
phenotypic traits Fincke,
1988; McVey, 1988;
Sutherland, 1985). The
challenge now is to determine how the variance in true reproductive success of
male snakes differs among species, the ecological reasons for those
differences, and whether those differences explain variation among species in
traits such as sexual size dimorphism.
|
ACKNOWLEDGEMENTS |
|---|
We thank P. Comm and K. Kissner for help in the field; G. Alderson, F. Mohammed, and L. De Sousa for help in the lab; M. Daly, P. Dunn, R. King, T. Madsen, J. Quinn, and D. Westneat for comments on earlier versions of the manuscript; Queen's University for use of the Biological Station; and the Natural Sciences and Engineering Research Council of Canada and Parks Canada for providing financial support through grants to H.L.G. and P.J.W.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Aldridge RD, 1982. The ovarian cycle of the watersnake Nerodia sipedon, and the effects of hyophysectomy and gonadotropin administration. Hepetologica 38: 71-79.
Andersson M, 1994. Sexual selection. Princeton, New Jersey: Princeton University Press.
Arnold SJ, Duvall D, 1994. Animal mating systems: a synthesis based on selection theory. Am Nat 143: 317-348.
Arnold SJ, Wade JM, 1984. On the measurement of natural and sexual selection: theory. Evolution 38: 709-719.
Barry FE, Weatherhead PJ, Philipp DP, 1992. Multiple paternity in a wild population of northern water snakes, Nerodia sipedon. Behav Ecol Sociobiol 30: 193-199.
Bauman MA, Metter DE, 1977. Reproductive cycle of the northern watersnake, Natrix s. sipedon (reptilia, serpentes, colubridae). J Herpetol 11: 51-59.
Brown GP, Weatherhead PJ, 1997. Effects of reproduction on survival and growth of female northern water snakes, Nerodia sipedon. Can J Zool 75: 424-432.
Brown GP, Weatherhead PJ, 1999a. Demography and sexual size dimorphism in northern water snakes, Nerodia sipedon. Can J Zool 77: 1358-1366.
Brown GP, Weatherhead PJ, 1999b. Female distribution affects mate searching and sexual selection in male northern water snakes (Nerodia sipedon). Behav Ecol Sociobiol 47: 9-16.
Devine MC, 1977. Copulatory plugs, restricted mating opportunities and reproductive competition among male garter snakes. Nature 267: 345-346.[Medline]
Duvall D, Schuett GW, Arnold SJ, 1993. Ecology and evolution of snake mating systems. In: Snakes: ecology and behavior (Seigel RA, Collins JT, eds). New York: McGraw-Hill; 165-200.
Feaver PE, 1977. The demography of a Michigan population of Natrix sipedon with discussions of ophidian growth and reproduction (PhD thesis). Ann Arbor: University of Michigan.
Fincke OM, 1988. Sources of variation in lifetime reproductive success in a nonterritorial damselfly (Odonata: Coenagrionidae). In: Reproductive success: studies of individual variation in contrasting breeding systems (Clutton-Brock TH, ed). Chicago: University of Chicago Press; 24-43.
Fox W, 1956. Seminal receptacles of snakes. Anat Rec 124: 519-539.[Medline]
Gibbs HL, Weatherhead PJ, 2001. Insights into
population ecology and sexual selection in snakes through the application of
DNA-based genetic markers. J Hered 92:
173-179.
Gist DH, Jones JM, 1987. Storage of sperm in the reptilian oviduct. Scan Microsc 1: 1839-1849.
Gromko MH, Newport MEA, Kortier MG, 1984. Sperm dependence of female receptivity to remating in Drosophila melanogaster. Evolution 38: 1273-1282.
Halpert AP, Garstka WR, Crews D, 1982. Sperm transport and storage and its relation to the annual sexual cycle of the female red-sided garter snake, Thamnophis sirtalis parietalis. J Morphol 174: 149-159.
Leger DW, Didrichsons IA, 1994. An assessment of data pooling and some alternatives. Anim Behav 48: 823-832.
Madsen T, Shine R, 1993. Temporal variability in sexual selection acting on reproductive tactics and body size in male snakes. Am Nat 141: 167-171.
Madsen T, Shine R, 1994. Components of lifetime reproductive success in adders. Vipera berus. J Anim Ecol 63: 561-568.
Madsen T, Shine R, Loman J, Hakansson T, 1992. Why do female adders copulate so frequently? Nature 355: 440-441.
McVey ME, 1988. The opportunity for sexual selection in a territorial dragonfly, Erythemis simplicicollis. In: Reproductive success: studies of individual variation in contrasting breeding systems (Clutton-Brock TH, ed). Chicago: University of Chicago Press; 44-58.
Mushinsky HR, 1979. Mating behavior of the common water snake, Nerodia sipedon sipedon, in eastern Pennsylvania (Reptilia, Serpentes, Colubridae). J Herpetol 13: 127-129.
Nilson G, 1981. Ovarian cycle and reproductive dynamics in the female adder, Vipera berus (Reptilia, Viperidae). Amphib-Reptil 2: 63-82.
Olsson M, Gullberg A, Tegelström H, 1996. Mate guarding in male sand lizards (Lacerta agilis). Behaviour 133: 367-386.
Olsson M, Gullberg A, Tegelström H, Madsen T, Shine R, 1994. `Promiscuous' matings enhance maternal fitness in lizards. Nature 369: 528.
Olsson M, Madsen T, 1998. Sexual selection and sperm competition in reptiles. In: Sperm competition and sexual selection (Birkhead TR, Møller AP, eds). San Diego, California: Academic Press; 503-577.
Perry-Richardson JJ, Schofield CW, Ford NB, 1990. Courtship of the garter snake, Thamnophis marcianus, with a description of a female behavior for coitus interruption. J Herpetol 24: 76-78.
Prosser MR, 1999. Sexual selection in northern water snakes, Nerodia sipedon sipedon (PhD thesis). Hamilton, Ontario: McMaster University.
Prosser MR, Gibbs HL, Weatherhead PJ, 1999. Microgeographic population structure in the northern water snake, Nerodia sipedon, detected using microsatellite DNA loci. Mol Ecol 8: 329-333.[Medline]
Schuett GW, 1982. A copperhead (Agkistrodon contortrix) brood produced from autumn copulations. Copeia 1982: 700-702.
Seutin G, White BN, Boag PT, 1991. Preservation of avian blood and tissue samples for DNA analysis. Can J Zool 69: 82-90.
Stille B, Madsen T, Niklasson M, 1986. Multiple paternity in the adder, Vipera berus. Oikos 47: 173-175.
Stille B, Niklasson M, Madsen T, 1987. Within season multiple paternity in the adder, Vipera berus. Oikos 49: 232-233.
Sutherland WJ, 1985. Chance can produce a sex difference in variance in mating success and explain Bateman's data. Anim Behav 33: 1349-1352.
Wade MJ, 1979. Sexual selection and variance in reproductive success. Am Nat 114: 742-746.
Wade MJ, Arnold SJ, 1980. The intensity of sexual selection in relation to male behaviour, female choice, and sperm precedence. Anim Behav 28: 446-461.
Weatherhead PJ, Barry FE, Brown GP, Forbes MRL, 1995. Sex ratios, mating behavior and sexual size dimorphism of the northern water snake, Nerodia sipedon. Behav Ecol Sociobiol 36: 301-311.
Weatherhead PJ, Brown GP, Prosser ML, Kissner KJ, 1999. Factors affecting neonate size variation in northern water snakes, Nerodia sipedon. J Herpetol 33: 577-589.
Westneat DF, Frederick PC, Wiley RH, 1987. The use of genetic markers to estimate the frequency of successful alternative reproductive tactics. Behav Ecol Sociobiol 21: 35-45.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Ursenbacher, C. Erny, and L. Fumagalli Male Reproductive Success and Multiple Paternity in Wild, Low-Density Populations of the Adder (Vipera berus) J. Hered., May 1, 2009; 100(3): 365 - 370. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. D. Lorch Using Upper Limits of "Bateman Gradients" to Estimate the Opportunity for Sexual Selection Integr. Comp. Biol., November 1, 2005; 45(5): 924 - 930. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. I. Schulte-Hostedde, J. S. Millar, and H. L. Gibbs Sexual selection and mating patterns in a mammal with female-biased sexual size dimorphism Behav. Ecol., March 1, 2004; 15(2): 351 - 356. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Weatherhead, M. R. Prosser, H. L. Gibbs, and G. P. Brown Male reproductive success and sexual selection in northern water snakes determined by microsatellite DNA analysis Behav. Ecol., November 1, 2002; 13(6): 808 - 815. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||