Manuscript Click here to download Manuscript: manuscript.doc Click here to view linked References 1 2 3 4 1 Authors: 5 6 2 Cunha M1, Berglund A2, Monteiro NM1, 3 7 8 9 3 10 11 4 Title: The intrinsically dynamic nature of mating patterns and sexual selection 12 13 14 5 15 16 1 6 Centro de Investigação em Biodiversidade e Recursos Genéticos, Rua Padre 17 18 19 7 Armando Quintas, 4485-661 Vairão, Portugal 20 2 21 8 Department of Ecology and Genetics/Animal Ecology, Uppsala University 22 23 3 9 CEBIMED, Faculdade de Ciências da Saúde, Universidade Fernando Pessoa, 24 25 26 10 Rua Carlos da Maia 296, 4200-150 Porto, Portugal 27 28 11 29 30 31 12 Corresponding author: Nuno Miguel Monteiro, Centro de Investigação em 32 33 13 Biodiversidade e Recursos Genéticos, Rua Padre Armando Quintas, 4485-661 34 35 36 14 Vairão, Portugal. Fax number: +351252661780, email:
[email protected] 37 38 15 39 40 16 41 42 43 17 44 45 18 46 47 48 19 49 50 20 51 52 53 21 54 55 22 56 57 23 58 59 60 24 61 62 63 1 64 65
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
25
Abstract
26 27
Selection processes are influenced by both biotic and abiotic variables, most of
28
which seasonally fluctuating. Therefore, selection may also vary temporally.
29
Specifically, sexual selection, an integral component of natural selection, will
30
inevitably exhibit temporal variation but the scale at which these changes occur are
31
still not well understood. In this study, performed on a wild population of the sex-role
32
reversed black striped pipefish Syngnathus abaster (Risso, 1827), we explore
33
differences between variables such as male reproductive success, mating success,
34
relative fitness, female investment, mate choice and operational sex ratio in two
35
very distinct periods aiming at representing the onset and end of the breeding
36
season. Sexual selection seemed stronger early in the breeding season when
37
reproduction is beginning. Male reproductive and mating success revealed that size
38
is a much more important trait during the onset of the breeding season as when
39
compared to the end. Moreover, we found that larger females reproduce mainly
40
during the onset while smaller females had increased chances of reproducing only
41
towards the end. As our sampling was performed in two consecutive years,
42
between-year variation might have taken place. Nevertheless, data from previous
43
samplings suggest that between-year variation in many demographic parameters is
44
low. Hence, we suggest that the observed variation in mating patterns, following our
45
theoretical predictions, is predominantly due to the inherent dynamics of selection
46
occurring within the breeding season. We argue that time should always be a key
47
variable when studying mating system evolution and sexual selection.
48 2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
49
Keywords:
mating system; female investment; mate choice; reproduction;
50
variation; selection.
51
Introduction
52 53
Many evolutionary and ecological studies try to understand the underlying dynamics
54
of selection processes (Siepielski et al. 2009) and its translation into adaptive
55
evolutionary change. Since selection is influenced by a wide array of biotic and
56
abiotic variables, most of which fluctuate within even very short time frames, the
57
direction and strength of selection should vary temporally as well. For instance,
58
mate monopolization by males may change over a breeding season if initially
59
synchronous females vary in the amount of time needed for subsequent breeding
60
events (Reichard et al. 2008). Such temporal variation in mating patterns may, in
61
turn, modulate the intensity of sexual selection. Besides mate monopolization, many
62
other biotic factors also prone to fluctuations, such as population density (Streatfeild
63
et al. 2011), nest aggregation (Saraiva et al. 2009) or mate quality and availability
64
(Cratsley and Lewis 2005), play an important role in shaping mating patterns.
65
Simultaneously, many abiotic variables acting within variable time scales, such as
66
temperature, rainfall or photoperiod are also effective modulators of mating patterns
67
(McAllan and Geiser 2006; Silva et al. 2007; Olsson et al. 2011; Robinson et al.
68
2012). Indeed, environmental variability was early expected to be a key factor in
69
mating systems evolution and sexual selection (Emlen and Oring 1977). However,
70
rather few studies explicitly address the variation in mating patterns either
71
geographically (Mobley and Jones 2007; Mobley and Jones 2009; Monteiro and
3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
72
Lyons 2012; Saraiva et al. 2012) or across time (Reichard et al. 2008; Milner et al.
73
2010; Olsson et al. 2011; Robinson et al. 2012). For instance, sex roles may be
74
quite flexible over a breeding season (Forsgren et al. 2004), so understanding
75
sexual selection requires us to explore not only spatial but also temporal variation in
76
selection regimes. In the case of the sex role reversed worm pipefish, Nerophis
77
lumbriciformis (Jenyns, 1835) the breeding season starts during cold-water periods
78
and continues, depending on location, while temperatures increase more than 10 °C
79
(Monteiro and Lyons 2012). However, only large individuals reproduce during the
80
entire breeding season whereas smaller pipefish breed during the first few months
81
when water temperatures are lower. Thus, mate availability and the operational sex
82
ratio change over the reproductive season, and consequently mating patterns and
83
the intensity of sexual selection do too.
84
We use the sex role reversed black striped pipefish Syngnathus abaster to test if
85
mating dynamics and sexual selection can indeed change even within the extent of
86
a breeding season. We already know that mating patterns are able to fluctuate
87
seasonally (e.g., Silva et al. 2007, 2009). For instance, this pipefish performs yearly
88
migrations to the estuaries where most males rapidly become pregnant right at the
89
onset of the breeding season, when reproduction is predominantly size-assortative
90
(Silva et al. 2008). Subsequently, as the breeding season progresses, several
91
reproduction-related variables change (e.g., we find fewer pregnant males, and
92
those that are pregnant have fewer eggs). Yet, even if these studies are highly
93
suggestive of a progressive relaxation of selection processes throughout the
94
breeding season, we need more direct and in depth evidence from natural
95
populations to test the prediction that mating patterns are indeed inherently dynamic 4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
96
and vary significantly as the breeding season progresses, thus influencing the
97
strength of sexual selection even within short time periods. Taking advantage of a
98
new set of recently developed microsatellites (Diekmann et al. 2009), we genotyped
99
a subsample of the sexually active pipefish population together with their progeny at
100
the beginning and end of the breeding season. This allows us to determine how
101
many females each captured male has mated with, and even identify actual mating
102
pairs. Through kinship analysis, besides characterizing the black striped pipefish
103
genetic mating system, we will test if female reproductive investment and mate
104
choice strategies in a wild pipefish population are largely fixed or, as we believe,
105
deemed to change towards the end of the breeding season.
106 107
Material and Methods
108 109
Male pipefish of the genus Syngnathus possess brood pouches (marsupium)
110
located on the tail, where females deposit oocytes. In the sampled population, sex
111
roles are reversed, albeit mildly (Silva et al. 2006), and females are larger and more
112
competitive than males (Silva et al. 2006; Silva et al. 2007, 2010). Reproductive
113
behaviors are affected by water temperature (Silva et al. 2007), and as
114
temperatures fluctuate during the breeding season (BS), changes in the intensity of
115
sexual selection are expected.
116 117
Sample collection
118
5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
119
Samples were collected within the Aveiro estuarine lagoon, a brackish water habitat
120
on the Portuguese west coast (40º39’56’’N, 8º40’6’’W), from a large water reservoir
121
that fish can enter during the high tides. We sampled the population shortly after the
122
onset of the BS (May-June 2011) and towards the end of the BS (August 2010),
123
using a small manually operated bottom trawl (3 mm mesh size). Since we were
124
sampling within a periodically enclosed area we wanted to avoid that fish caught at
125
the onset of the BS would affect catches at the end of the BS, thus causing data
126
dependence. Fish were therefore captured in two consecutive years. The
127
advantage of this capture strategy is the avoidance of within-season data
128
dependence, but a possible disadvantage results from introducing between-year
129
variation. To estimate such effects, we contrast several demographic and
130
reproductive variables (e.g., adult sex ratio, % pregnant males, egg numbers in
131
pregnant males, male and female size as well as pregnant and non-pregnant male
132
size) estimated from the same population (samples obtained during the 2005-2008
133
period and present sampling), suggesting negligible between-year effects. These
134
past samples include data from 197 pipefish captured during three years either near
135
the onset or end of the BS (seawater temperatures at the onset of the BS falling
136
within the 24±1°C interval, while at the end being comprised within 22 ±1°C, for past
137
and present sampling events). In the current sample, 392 pipefish were caught (181
138
females and 211 males) near the end of the BS, while in the onset of the BS 110
139
individuals were sampled (61 females and 49 males). The difference in number of
140
captured individuals was not due to changes in overall pipefish density but to a
141
deliberate adjustment of sampling effort: our preliminary genetic analysis revealed
142
that fewer pipefish sufficed to calculate the used variables. Although we sacrificed 6
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
143
several pipefish, we estimate that we did not affect the Aveiro’s estuarine
144
population. In water reservoirs used for rearing fish, such as those sampled, a
145
considerable number of pipefish invariably die due to desiccation when the
146
reservoirs are slowly emptied for maintenance at the end of summer (only
147
economically relevant species are harvested, but pipefish are ignored since they
148
have no commercial value).
149
All captured animals were individually photographed (Olympus μ1030 digital
150
camera) alongside a ruler, and subsequently measured using ImageJ software.
151
Pipefish were carried to the lab in isothermic containers, over-anesthetized with
152
clove oil (10 ppm) and immediately preserved in ethanol (96%) until DNA extraction.
153
154
DNA extraction and PCR
155 156
DNA was extracted from females, pregnant-males and developing embryos present
157
in the male’s brood pouch using the Genomed JETquick tissue DNA spin kit.
158
Samples were amplified for 5 highly polymorphic microsatellite markers (Sabas3,
159
Sabas5, Sabas7, Sabas8, and Sabas9; Diekmann et al. 2009). For Sabas5 and
160
Sabas9, reaction tubes (10 µl) contained approximately 1 μl DNA, 10X PCR Rxn
161
Buffer (Invitrogen), 3 mM MgCl 2, 0.2 mM dNTPs mix, 0.8 μM primer, 0.8 mM
162
fluorescently labeled tail, 0.1U Platinum Taq DNA Polymerase (Invitrogen). The only
163
modification for Sabas8 and Sabas7 was a reduction in MgCl2 (2 mM), while
164
Sabas3 was amplified using 5 µl MasterMix (Alphagene), 1 µl H 2O, 0.8 µM primer,
7
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
165
0.8 mM fluorescently labeled tail. For four microsatellites (Sabas5, Sabas7, Sabas8,
166
and Sabas9), the PCR parameters used were identical: initial denaturation at 94 ºC
167
for 15 min followed by 40 cycles of 94 °C denaturation for 30 s, annealing at 62 °C
168
for 30 s, and 72 °C extension for 30 s. Another set of 10 cycles followed, with a
169
different annealing temperature of 53 °C due to the fluorescent tail. Sabas3 has a
170
different PCR setup with an initial denaturation of 95 °C for 10 min, followed by 20
171
cycles with 95 °C for 30 s, 74 °C annealing for 30 s with touchdown to reduce
172
unspecific amplifications (0.5 °C decrease for each repeat) and 72 °C extension for
173
30 s. Then we ran 15 cycles with 95 °C for 30 s, 64 °C annealing for 30 s and 72 °C
174
for 30 s, followed by 15 cycles for the fluorescent tail as described above. Final
175
extension for Sabas3 was 72 °C for 90 min while for the other four microsatellites
176
was 72 °C for 60 min. All PCRs were conducted in a BIO-RAD MyCyclerTM Thermal
177
Cycler (Applied Biosystems).
178 179
Microsatellite analysis
180 181
Allele sizes were determined on an ABI 3100 capillary sequencer (Applied
182
Biosystems) using LIZ 75-450 as size standard (Applied Biosystems). The size of
183
each fragment was then determined in Peak Scanner Software v1.0 (Applied
184
Biosystems). Approximately 10% of the samples were genotyped twice and no
185
genotyping errors were found. GenAlEx v6.5 (Peakall and Smouse 2012) was used
186
to assess allele diversity, frequency, observed (H0) and expected (HE)
187
heterozygosity, probability of identity (Pi) and probability of exclusion (PE) with one
188
known parent. Since we sampled individuals in consecutive years, we calculated 8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
189
the FST and G’’st for all combined locus with GenoDive 2.0b23 (Meirmans and Van
190
Tienderen 2004), using only parental genotypes in order to be assured that we were
191
indeed sampling the same population. Cervus 3.0 (Kalinowski et al. 2007) was used
192
to check the frequency of null alleles for every locus.
193 194
Parentage analysis and mating system
195 196
To describe the genetic mating system of S. abaster, we genotyped (for all 5 loci)
197
61 females from the onset of the BS, 61 females from the end of the BS, 35
198
pregnant males from the BS onset and 37 from the BS end, together with
199
subsamples of the pregnant male’s embryos (totaling 365 embryos). All the
200
individuals were randomly selected from the original sample. From our pool of 72
201
pregnant males, 50 males were randomly selected (25 in each period) for extended
202
embryo genotyping (see below). Nevertheless, due to recurrent failure in embryo
203
DNA extraction, 5 were ultimately excluded (3 in the onset and 2 in the end). Thus,
204
we ended up performing extended embryo genotyping in 45 males. The remaining
205
males (10 for the onset and 12 for the end of BS) were used with the sole intention
206
of extending the probability of finding putative mothers from our pool of sampled
207
females. In these males, fewer embryos were genotyped per male (comparing with
208
the method described below), as these were primarily used to increase the
209
probability of finding genotypes compatible with females we sampled. For embryo
210
genotyping, the pregnant male’s pouch was dissected. We genotyped a subsample
211
of the embryos in the following way: for males carrying less than 20 embryos, the
212
pouch was divided into two identical areas and 5 embryos were genotyped, two 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
213
from one area and three from the other. We avoided selecting neighboring embryos,
214
thus ensuring a good coverage within each area. In males carrying 20 to 39
215
embryos, we divided the pouch into three areas and genotyped 2 embryos from
216
each area following the same criterion as above. When males had 39 to 50
217
embryos, the pouch was divided into 4 areas and 2 embryos per area were
218
genotyped. Above 50 embryos, we used 5 areas with two embryos genotyped from
219
each area. As S. abaster females lay on average 17 oocytes in each successful
220
mating event (Silva et al. 2006), while embryos from distinct mothers tend to be
221
clustered within the male’s brood pouch as also observed in other species of the
222
same genus (Jones et al. 1999; Jones and Avise 2001), we believe that this
223
procedure yields accurate maternity information. Even if individual females laid
224
smaller egg batches than those reported by Silva et al. (2006), we could detect up
225
to three different mothers by genotyping at least 5 embryos from males with less
226
than 20 embryos. Moreover, in males with a large number of embryos we were able
227
to detect up to 5 distinct mothers by genotyping 10 embryos, a number well above
228
the maximum observed mean number of females per brooding male in natural
229
populations of syngnathids (see Mobley et al. 2011b).
230
After determining the paternal and embryonic genotype, we inferred the potential
231
maternal genotype in order to contrast it with the pool of genotyped females (from
232
which we also had information on size, an important variable in syngnathid
233
reproduction; Berglund et al. 1986; Rosenqvist 1990; Silva et al. 2009). The
234
detection of putative mothers from within our samples was performed using
235
PARFEX v1.0 (Sekino and Kakehi 2012).
10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
236
After identifying the putative mothers, we estimated the number of additional
237
mothers (individual female genotypes inferred after male genotype subtraction) that
238
contributed to each brood, using GERUDv2.0 (Jones 2005). We then assessed the
239
number of different females that laid oocytes within a single male (multi-maternity),
240
thus estimating female reproductive investment. To this end, we divided the total
241
number of eggs in the marsupium by the number of genotyped embryos. We then
242
multiplied this value by the proportion of sampled embryos from a specific female to
243
estimate clutch size, as female egg batches tend to be grouped within the
244
marsupium. To exemplify, we genotyped 6 embryos in a male carrying 30
245
developing eggs, meaning that each sampled embryo represented 5 embryos [this
246
value is well below the average female egg batch described by Silva et al. (2006)].
247
Thus, if we found that 2 females were responsible for siring the brood, one with 4
248
sampled embryos and the other with the remaining two, we would attribute a total
249
number of 20 embryos to the first female and 10 to the second. We believe that our
250
estimate of the number of oocytes laid by a female is sufficiently good to detect
251
differences between the two sampled periods of the BS (our main goal). In total, we
252
estimated that 128 distinct females (60 at the onset and 68 at the end of the BS)
253
contributed to the analyzed broods.
254
We furthermore calculated an estimate of the male’s mating success (number of
255
distinct mates per male) and relative fitness (number of embryos per male divided
256
by the average number of embryos in all pregnant males). Since a proportion of
257
non-pregnant males were observed on both sampling periods, we included this
258
proportion for each period. As an example, if 8% of the males were non-pregnant
259
and our sample comprised 22 pregnant males, we randomly selected 2 non11
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
260
pregnant males to the analysis of mating success and relative fitness. We
261
calculated male reproductive success using all the sampled males, since we knew
262
their size as well as the number of eggs they were carrying.
263
Finally, we estimated the operational sex ratio (OSR - proportion of ready to mate
264
males to ready to mate individuals within the population; Emlen and Oring 1977) for
265
the two sampling periods. We followed a standard procedure for pipefish, including
266
all the sexually mature females and only the non-pregnant males in the OSR
267
(Mobley and Jones 2007; Mobley and Jones 2009; Monteiro and Lyons 2012). We
268
used all the females in our OSR calculations as even the smallest had a genotype
269
compatible with a sampled male brood. Since female reproduction depends
270
primarily on the production of fertilizable oocytes, we looked also for the presence of
271
mature oocytes (see Begovac and Wallace 1988) by dissecting 114 females (55
272
and 59 randomly selected females from the onset and end of the BS, respectively).
273
This data can potentially help interpret the results emerging from the calculation of
274
the OSR if it shows that female ovaries become depleted as the breeding season
275
progresses.
276 277
Statistical analysis
278 279
In order to detect if the mating pairs observed within each sampling period (onset
280
and end of the BS) deviated from a random pairing strategy, we conducted a Monte
281
Carlo simulation. We used all the captured males and females and calculated all the
282
mating pair combinations, independently for each sampling period. Then, we
283
randomly extracted the same number of pairs as those detected in our actual 12
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
284
sampling (11 pairs for each period) from the pool of all available combinations within
285
each period (onset = 2989 possible pairs; end = 38191 possible pairs) and
286
registered how often females were larger than males (from a minimum of zero to a
287
maximum of 11). This procedure was replicated 1000 times. Once we had our null
288
distribution for each sampling period, we proceeded to locate our observed results
289
and extracted the corresponding p-value. All remaining statistical analyses were
290
performed in STATISTICA v.7 (StatSoft) and ACTUS2 (Estabrook et al. 2002). All
291
parametric analyses were performed after testing for normality and equality of
292
variances. If assumptions were not met after data transformation, appropriate non-
293
parametric tests were applied. ANCOVA models were selected after testing for the
294
homogeneity of slopes.
295 296
Results
297 298
Between-year variation
299 300
Comparing the onset and the end of the BS from different years, as we do here,
301
introduces between-year variation that potentially confounds the intended
302
comparison from the onset and the end of the BS. However, both demographic and
303
reproductive variables measured did not vary between past (2005-2008) and
304
present sampling events. Specifically, using an analysis of contingency tables with
305
simulation (ACTUS2), we observed that neither the averages of female size
306
(χ2=0.001, DF=3, P=0.98), male size (χ2= 0.106; DF=3; P=0.72), pregnant male size
307
(χ2= 0.182; DF=3; P=0.64), non-pregnant male size (χ2= 0.148; DF=3; P=0.66), egg 13
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
308
number in pregnant males ( χ2= 0.405; DF=3; P=0.59), % pregnant males ( χ2=
309
0.484; DF=3; P=0.48) or adult sex ratio (χ2= 3.417; DF=3; P=0.09) differed between
310
years. Given that demographic and reproductive variables were similar between the
311
current and past sampling events, data presented from now will address the most
312
recent samples (unless explicitly stated) from which we also have individual genetic
313
data.
314 315
Summary statistics
316 317
Adult sex ratio (ASR - proportion of males to all adults) did not differ from equality in
318
neither BS periods (Proportions comparison, two-sided test, BS onset, P=0.37; BS
319
end, P=0.26). The few non-pregnant males at the onset of the BS were smaller than
320
the pregnant males (Mann Whitney: U=14, N pregnant=45, Nnon-pregnant=4, P=0.005).
321
During the end of the BS no such size difference was observed (Mann Whitney:
322
U=5370, Npregnant=118, Nnon-pregnant=93, P=0.83). Female and male size did not differ
323
in any period (Mann Whitney: Onset: U=1411, N males=49, Nfemales=61, P=0.72 and
324
End: U=18891, Nmales=211, Nfemales=181, P=0.85).
325 326
Microsatellite analysis
327 328
Four of the five loci assayed showed a high degree of polymorphism in both
329
sampled periods, ranging from 25 to 29 alleles during the onset of the BS and from
330
24 to 30 at the end (Table 1). The exception was Sabas 8 that had only 4 alleles
331
near the onset and 6 alleles near the end (Table 1). The probability of identity (Pi, 14
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
332
which represents the likelihood of having identical genotypes in our samples)
333
obtained for a 5 loci combination was very low (P i Onset=8.3x10-10; Pi End=6.8x10-10).
334
The probability of exclusion, with one parent known, (Pe, which represents the
335
likelihood of making a correct assignment) was extremely high (P e
336
End=0.99).
337
period, with the highest frequency recorded for Sabas 8 during the end of the BS
338
(5.69 x 10-2). Moreover, neither the expected ( χ2= 0.156; DF=9; P=0.99) and the
339
observed heterozygosity (χ2=0.082; DF=9; P=0.99). Hence, the low FST value
340
(0.004) we obtained, together with a non-significant estimate of population
341
differentiation (G’’ST=0.028; P=1), indicates that we were indeed sampling the same
342
population.
Onset
= Pe
Null allele frequency was low for every locus regardless of the sampling
343 344
The genetic mating system
345 346
On average, each male mated with 2.82 ± 0.85 females, N=45 (average ± SD).
347
The genetic data also provided direct evidence for multiple mating by females since
348
egg clutches from actually sampled females were detected in different males twice
349
during the onset of the BS sampling. Additionally, females (as seen from genotype-
350
inferred females) were able to deposit eggs in the same male more than once within
351
the same pregnancy event.
352
Near the onset of the BS, 9 captured females had a genotype compatible with
353
embryos present in 11 different males, meaning that 2 of the females mated with
354
distinct males. In 7 out of the 11 detected pairs, females were larger than males
15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
355
(Figure 1). Taking into consideration size differences among all hypothetical pairs
356
that could be formed if mating was purely random, we cannot exclude the possibility
357
that the observed pairs were the result of random mating (Monte Carlo simulation;
358
P=0.134). At the end of the BS, we also detected 11 pairs, but only in two instances
359
were females larger than males (Figure 1). During the end of the BS, again taking
360
into consideration size differences among all hypothetical pairs that could be formed
361
if mating was purely random, we concluded that mating patterns were most
362
probably non-random (Monte Carlo simulation; P=0.029).
363 364
Male reproductive and mating success
365 366
For males captured at the onset of the BS, the reproductive success (number of
367
embryos per captured male) was significantly higher (42±19 embryos, N=49) than
368
in the end of the BS (12±12 embryos, N=211) (ANCOVA, separate slopes model:
369
F1,255=9.061, P
