ENS-05674; No of Pages 9

Evolution and Human Behavior xx (2010) xxx – xxx

1 Original Article

2 Mating system change reduces the strength of sexual selection in an

3 American frontier population of the 19th century☆,☆☆
4 Jacob A. Moorad a,b,⁎, Daniel E.L. Promislow a , Ken R. Smithc , Michael J. Waded
a
Q1 5 Department of Genetics, University of Georgia, Athens, GA, USA
b
6 Department of Pathology, University of Michigan Medical School, Ann Arbor, MI, USA
c
7 Huntsman Cancer Institute, University of Utah, Salt Lake City, UT, USA
d
8 Department of Biology, Indiana University, Bloomington, IN, USA
9 Initial receipt 30 July 2010; final revision received 15 October 2010

10 Abstract

11 Sexual selection, or competition among members of one sex for reproductive access to the other, is one of the strongest and fastest
12 evolutionary processes. Comparative studies support the prediction that sexual selection is stronger in polygamous than in monogamous
13 species. We report the first study of the effect on sexual selection of a change in mating system, from polygyny to monogamy, within a
14 historical human population. Here we show that over the reproductive lifetimes of Utahns born between 1830 and 1894, socially
15 induced reductions in the rate and degree of polygamy correspond to a 58% reduction in the strength of sexual selection. Polygyny
16 conferred a strong advantage to male fitness as well as a weak disadvantage to female fitness. In contrast, mating with multiple males
17 provided little benefit to females in this population. Polygamy benefitted males by increasing reproductive rates and by lengthening
18 reproductive tenure. Each advantage contributed to roughly half of the increased total lifetime reproductive success. This study illustrates
19 both the potency of sexual selection in polygynous human populations and the dramatic influence that short-term societal changes can
20 have on evolutionary processes.
21 © 2010 Published by Elsevier Inc.
22
23 Keywords: LDS; Mormon; Utah; Bateman; Opportunity for selection; Polygamy; Polygyny

24

25 1. Introduction that sexual selection should be a much stronger evolutionary 30

force in polygamous than in monogamous mating systems 31
26 Sexual selection is the process identified by Darwin (Shuster & Wade, 2003; Wade, 1979; Wade & Shuster, 32
27 (1859) where the members of one sex, generally males, 2004b). Support derives primarily from extensive compa- 33
28 compete with one another for reproductive access to risons across taxa with different mating systems (Andersson, 34
Q2 29 members of the other sex (Morgan, 1991). Theory predicts 1994; Bateman, 1948; Jones & Avise, 2001; Lofredo & 35
Borgia, 1986; Shuster & Wade, 2003). We report the first 36
study of the effects on sexual selection of a change in mating 37
system, from polygyny to monogamy, within a population of 38
☆
Grants: This work was funded by National Science Foundation grant a single species, Homo sapiens. 39
DEB-0717234 to J.A.M. and D.EL.P., and NIH grant RO1GM065414-06 Polygynous marriage, where a man would simultaneously 40
and NSF grant DEB-0614086 to M.J.W and partially supported by National
Institute on Aging grant P30-AG013283. take on multiple wives, was a practice throughout much of 41
☆☆
Disclaimer: The study is solely our responsibility and does not the 19th century for some members of the Church of Jesus 42
necessarily represent the official views of the National Institute on Aging or Christ of Latter-day Saints (LDS or Mormons). In the Utah 43
the National Institutes of Health. Grant sponsor: National Institute of Aging Territory in the early part of the century, our estimates 44
(The Utah Study of Fertility, Longevity, and Aging); grant number:
from the Utah Population Database indicate that the 45
AG022095.
⁎ Corresponding author. Department of Biology, Duke University, frequency of polygamous men as a fraction of all married 46
Durham, NC 27708, USA. men reached a maximum of 17.8% among men born in 47
E-mail address: jacob.moorad@duke.edu (J.A. Moorad). 1833 (Fig. 1), a figure consistent with previous estimates 48

1090-5138/$ – see front matter © 2010 Published by Elsevier Inc.

doi:10.1016/j.evolhumbehav.2010.10.004
 2 J.A. Moorad et al. / Evolution and Human Behavior xx (2010) xxx–xxx

(e.g., the variances and covariances) between mating 81

success and reproductive success in both males and 82
females. In populations with an equal sex ratio of females 83
to males, the rates of polygyny are necessarily and 84
positively associated with the frequency of males who do 85
not mate (Wade, 1979; Wade & Shuster, 2004b). As a 86
result, polygyny increases the variation among males in 87
mate number (mating success or MS). Using results from 88
fruit fly experiments, Bateman (1948) was the first to argue 89
that variation among males in mate numbers was the most 90
important cause of a sex difference in the variance in 91
reproductive success (RS). Wade (1995) derived the 92
general relationship between the variance in male and 93
female RS, showing that the sex difference in σ2RS was a 94
function of male MS. Shuster and Wade (2003) reviewed 95
and extended this theory and applied it to several different 96
taxa (Andersson, 1994; Bjorklund, 1991; Lofredo & 97
Borgia, 1986; Shuster & Wade, 2003), including sex- 98
reversed species with polyandry among territory-holding 99

Fig. 1. Polygamous men in the UPDB as a frequency of all married men by females (Jones & Avise, 2001), to show how mating 100
birth year. system affected the strength of sexual selection. A recent 101
comparative study using 18 geographically diverse human 102
populations, a collection that included both monogamous 103
and polygamous societies (Brown et al., 2009), showed 104
49 (Smith & Kunz, 1976). Subsequently, changes in social that σ2RS and σ2MS tended to be higher in males than in 105
50 pressures over the next few decades reduced the frequency females. Following the methods described in the Methods 106
51 of such marriages to less than 1%. New strictures on the section, we examine these relationships within a single 107
52 practice of polygyny began with the Morrill Anti-Bigamy population during a period of change in the practice of 108
53 Act of 1862 outlawing plural marriage throughout the US plural marriage that affected mating success. 109
54 territories. The US Supreme Court upheld the ban in the
55 face of a freedom of religion challenge in 1878 and, in
56 1890, the Mormon Church issued the Woodruff Manifesto 2. Methods 110
57 which officially banned polygyny.
58 Data from this population afford a unique test of the role Our extract of the Utah Population Database (UPDB) 111
59 of polygyny in sexual selection in four ways. First, until now, included birth and death years of the 89,034 fathers and 112
60 the primary source of empirical support for sexual selection 96,407 mothers recorded in the UPDB born before 1900 113
61 theory has been comparative studies across taxa with (the majority of the UPDB reflecting Utah parents born 114
62 different mating systems (Andersson, 1994; Bateman, after 1900 was not available to us). Note that the Utah 115
63 1948; Jones & Avise, 2001; Lofredo & Borgia, 1986; population was not a strict closed society since potential 116
64 Shuster & Wade, 2003). Secondly, previous human studies wives could be recruited via religious conversion and 117
65 have compared relationships between social mating customs migration into Utah (Daynes, 2001). We used only those 118
66 and sexual selection across geographically and culturally individuals with known death year. There were very few 119
67 distinct populations (Brown, Laland & Mulder, 2009), individuals born before 1830 and these were excluded from 120
68 whereas this study uses temporally defined cohorts that our analysis. Fathers born after 1894 had missing data on 121
69 differ in mating system within the same population. Third, their wives who were often born after the cutoff year of 122
70 our sample size (180,082 parents) is 18 times larger than that 1900. These exclusion criteria meant we had a sample of 123
71 in all other human studies of sexual selection combined 72,812 fathers and 75,366 mothers born between 1830 and 124
72 (Brown et al., 2009), giving unprecedented power to 1894. Our extract also included the 322,551 sons and 125
73 describe sexual selection in a human population. Finally, 307,754 daughters who were born to parents represented in 126
74 these data span an important part of the demographic the UPDB extract. These parents were divided into 65 127
75 transition (Borgerhoff-Mulder, 1998) where natural fertility single birth-year cohorts from 1830 to 1894. Because the 128
76 conditions prevailed early but reproductive rates rapidly fell UPDB is derived from descendent genealogies, we could 129
77 over a few generations. not expect it to include reliable counts of nulliparous 130
78 To determine the strength of sexual selection that acts individuals. We used US census data and a few reasonable 131
79 on a population and to best understand how it is influenced assumptions (see Section 2.1 below) to estimate the size of 132
80 by mating system, we need to quantify the associations this class of individuals for each sex and birth years. The 133
 J.A. Moorad et al. / Evolution and Human Behavior xx (2010) xxx–xxx 3

134 sample sizes that reflect the total number of parents plus the nulliparous wives by estimating the true counts for non- 178
135 projected number of nulliparous individuals are illustrated reproductive women of age interval x and birth-year interval 179
W
136 in Fig. A1 in Appendix A. y as W Vxyð0Þ = 1 −WxyðxyUðÞ0Þ= Wxy + Sxy . The adjusted numbers of 180
137 We determined the total number of offspring (reproduc- W −W
mothers were W Vxyðm N 0Þ = 1 −WxyxyðU Þ xy=ð0WÞ xy , where mN0 indicates 181
138 tive success, ‘RS’) for each individual, the number of his or fertility. 182
139 her reproductive partners (mating success, ‘MS’) and the Always-single women (S) were assumed not to have had 183
140 number of years over which he or she reproduced children prior to the census year. However, women who 184
141 (reproductive tenure, ‘RT’). The birth year-specific means were single or married and nulliparous may have had first 185
142 of mating and reproductive success are shown in Fig. A2 in children after the census. This would cause us to overesti- 186
143 Appendix A. In addition, each reproductive female was also mate rates of nulliparity. For this reason, we considered only 187
144 assigned another value equal to the mating success of the those women who were 35 years or older at the time of the 188
145 partner that she most often reproduced with. In the cases census. Using reproductive records from the UPDB and 189
146 where females mated with more than one male, the mating pooled overall birth year cohorts of interest (1830–1894), we 190
147 success of these females' first and last mates was also noted. estimated that the probability that a mother had her first child 191
148 The choice of partner had no noticeable effect on the after 35 was δ35=0.0188. The average taken over the age 192
149 outcome of our analyses. interval x=35 to 39 (and weighted by age-cohort size at 193
1910) was δ̄ (35-39), (1871-1875)=0.0119. The corresponding 194

150 2.1. Estimating the number of nulliparous individuals probability for the next youngest age-cohort was δ̄ (30-34), 195
(1876-1880)=0.0406; the chance that missing these mothers 196

151 We need to know the frequency distributions of repro- may cause problems with our estimates seemed too high to 197
152 ductive success to estimate the strength of sexual selection. risk using these age cohorts. Women above 49 years of age 198
153 This includes those individuals that do not reproduce at all. were not considered owing to the exceedingly large age 199
154 When information comes from descendent genealogies, such window (25 years). Although the frequency of missed 200
155 as with the UPDB, these individuals can be undercounted categorized nulliparous women was gauged to be quite low 201
156 and omitting the zero values diminishes the variance in for the cohorts of interest (between 0.0002 and 0.0119), we 202
157 reproductive success and the strength of selection (Shuster & made the following adjustments to best account for this bias: 203
158 Wade, 2003). This section explains how we corrected for this W″xy (0) =W ′xy (0) (1−δ̄ xy) and W ″xy (mN0) =W ′xy (mN0) + δ̄ xy W′xy (0) . 204
159 bias and estimated the size of the nulliparous populations of Finally, we accounted for the probability that a women 205
160 males and females. did not survive from birth to the time that the census was 206
161 Our approach was to complement data on survival and taken. Using UPDB data, we estimated the probability L- 207
162 reproduction from the UPDB, which we expect to be biased (mN0) that a mother born at time y lived to age x. Using the 208

163 in terms of the relative number of reproductive men and set of individuals with no reproductive records, we 209
164 women, with information from US census data (Sixteenth estimated Lxy(0), which better reflected the cumulative 210
165 Census of the United States, 1943). These data include survival fraction of the nulliparous females. We reasoned 211
166 fertility and marital information about women living in Utah that if the census figures accounted for all females who 212
167 and reported in 1910 and 1940. Parity counts of women had died prior to the census and also correctly forecasted 213
168 relevant to our study period are reproduced in Table 1. These the frequency of future first-time mothers, then the total 214
169 include the age-specific counts of total number of women number of mothers and non-reproductive females would be 215
W Wxyðm N 0Þ W Wxyð0Þ
170 ever married (W), ever-married women who reported having Lxyðm N 0Þ and Lxyð0Þ , respectively, where x was the age at 216
171 no children (W0), ever-married women who did not report census and Y was the birth year. The relative frequencies 217
172 number of children (WU) and single women (S). Women of projected always-nulliparous females born in birth year 218
173 were grouped into age categories of 5-year intervals (e.g., 40 interval y, Py(F), as well as all parameter estimates from 219
174 to 44 years of age) until 50 years of age, at which point a the UPDB that were used in the necessary adjustments 220
175 25-year age window was considered. We assumed that described above, are shown in Table 2. 221
176 marriage and reproductive history did not influence reporting We regressed Py(F) onto the midpoint of the 5-year birth 222
177 and, in this way, we adjusted for the under-reporting of intervals in order to project nulliparous fractions on specific 223

t1:1 Table 1
t1:2 Fertility and marriage history of Utah women from US Census data
t1:3 Age Birth years Census Women ever married Single
class year women,
t1:4 Total reporting, W Zero children, W0 Not reporting, WU
S
t1:5 45–49 1891–1895 1940 12,480 1100 1320 520
t1:6 35–39 1871–1875 1910 8336 627 446 563
t1:7 40–44 1866–1870 1910 7204 371 372 405
t1:8 45–49 1861–1865 1910 6244 282 380 240
 4 J.A. Moorad et al. / Evolution and Human Behavior xx (2010) xxx–xxx

t2:1 Table 2
Cohort-specific factors estimated from the UPDB used to estimate the size of
t2:2 the nulliparous fraction of females
t2:3 Age class Birth years Census δ̄ X L̄XY(m=0) L̄XY(mN0) PY (F)
t2:4 45–49 1891–1895 1940 2.1510E−04 0.5259 0.8954 0.1897
t2:5 35–39 1871–1875 1910 1.1860E−02 0.4683 0.9326 0.2281
t2:6 40–44 1866–1870 1910 2.3040E−03 0.4322 0.9054 0.1876
t2:7 45–49 1861–1865 1910 2.8220E−04 0.3771 0.8679 0.1637
US Census data and UPDB data were binned into 5-year and 1-year
windows, respectively. To account for this, single-year values of δX,
LXY(m=0) and LXY(mN0) were averaged over 5 years (and weighted by cohort
t2:8 size) to arrive at the values above.

224 birth years, including those not included in the census study.
225 After weighting by the size of the informative census
226 sample (Wxy+Sxy), we estimated the least squares best linear
227 fit to be Py(F)=−.207+2.1370E−04×y, which predicted a
228 very gradual increase in the relative size of the ever-
229 nulliparous population from 0.184 in 1830 to 0.198 in 1894.
230 We applied these estimates to our counts of mothers from Fig. 2. Variation in lifetime reproductive success (solid line and left axis) and
mating success (dotted line and right axis) for males (blue) and females (red)
231 the UPDB to estimate the true size of the female nulliparous
by birth year.
232 population in the Utah population.
233 As no such census information was available for Utah
234 males, we had to infer the frequency of nulliparity from the
235 female results. We assumed that mean MS, adjusted for The variation in mating success, or number of mates, 260
236 unequal sex ratios at birth, was the same for both sexes born mirrors the pattern for RS (Fig. 2), although female 261
237 at the same time. This need not be absolutely true if some sex variation appears not to change much over the study 262
238 and birth year cohorts are simply more successful than period. Male σ2MS (0.687) was nearly three times higher 263
239 others, but because each new individual must have one father than that of females (0.237) in the early 1830s but only 264
240 and one mother, we expect that this is a fair assumption over 25.9% higher in the early 1890s (0.227 vs. 0.181). Male 265
241 the long-term. Following this argument, the nulliparous σ2MS dropped 66.9% from the first to last 5 years of the 266
242 fraction of males born at Y is study, while female σ2MS lost 23.5% over this time, 267
although this latter figure represents a very small absolute 268
MSY ðfemalesÞ change in the variance. As with reproductive success, the 269
PY ð M Þ = 1 − ; ð1Þ male decline in the variance of mating success was most 270
R × MSY ðfathersÞ
rapid leading up to the early 1860s. 271
Declines in the sex-related differences in σ2RS co-occur 272
244
243 where R is the sex ratio at birth, which we estimate to be
with change in the frequency of polygamy. Direct causality 273
245 1.048 based on the UPDB records.
cannot be ensured, however, because mean reproduction 274
declines over the study period. Crow's index for the oppor- 275
2
246 3. Results tunity for selection I=σ2RS/RS measures variance in relative 276
fitness and sets an upper limit on the rate of evolutionary 277
247 We observe pronounced sex differences in the variation in change (Crow, 1958; Wade, 1979; Wade & Shuster, 2004a). 278
248 RS among those born in the early 1830s (years indicate time Based upon Fisher's (1930) Fundamental Theorem of 279
249 of birth unless otherwise indicated) (Fig. 2). The average Natural Selection, Crow's index arises from the earliest 280
250 variance taken over the first 5 years (1830–1834) is nearly formulations of evolutionary genetics. Because it is stan- 281
251 2.5 times higher in males than in females [σ2RS (♂)=41.6, dardized by mean reproduction, this comparative measure is 282
252 σ2RS (♀)=16.8]. There is an overall decline in female more appropriate for identifying secular changes over 283
253 variation with increasing birth year [σ2RS(♀)=10.6 in the segments of the demographic transition, or times where 284
254 early 1890s, a decline of 37%]. Over this same 65-year mean reproduction changes rapidly, because changes in 285
255 period, male σ2RS declined by 75.6% (to 10.2), roughly mean fitness cannot, by themselves, change I. The variance 286
256 converging with that of females. Most of the decline in mate number among males weighted  by the  ratio of 287
ðfemalesÞ 2 2
257 occurred prior to 1860, in the period leading up to and squared sex-specific mean RS, Imates = RS RS ðmalesÞ
rMS , is a 288
258 coincident with the birth of males that reached reproductive useful comparative metric often called the opportunity for 289
259 maturity after the 1862 ban on polygyny. sexual selection (Shuster & Wade, 2003; Wade, 1995). The 290
 J.A. Moorad et al. / Evolution and Human Behavior xx (2010) xxx–xxx 5

291 strength of selection on males IM (which includes Imates)

292 was 1.160 in the early 1830s but fell rapidly to 0.699 in
293 the early 1850s and then increased slightly to 0.706 over
294 the next three decades. The strength of selection on
295 females (IF), which does not include sexual selection, was
296 less than half that of IM in the early 1830s (0.482, or
297 41.5% of IM), remained fairly constant into the late 1860s
298 and then increased slightly to 0.601 over the next three
299 decades. As a result of these secular changes, the
300 opportunity for sexual selection Imates was greatest in the
301 early 1830s (0.667), dropped by 42.4% in the ensuing
302 two decades (0.384 in the early 1850s) and converged
303 towards one-quarter at the end of the 19th century (0.279
304 in the early 1890s). Thus, the opportunity for sexual
305 selection, Imates, lost 58.2% of its strength coincident
306 with the change in mating system from polygyny to
307 monogamy (Fig. 3).
308 We can estimate the reproductive costs and benefits of
309 polygyny, for both males and females, by studying their
310 Bateman gradients. These are the marginal effects of Fig. 4. Bateman gradients by birth year (βM is blue, βF is red and βF⁎ is
311 mating system on reproductive success for each sex, black). Dotted lines and error bars correspond to estimated gradients
312 estimated from the year-specific slope of the linear when serial monogamists are excluded from the analysis. These are
313 regression of RS on MS (β=σMS,RS/σ2RS, conditioned on grouped by decade of birth. Of these, the upper (blue) line corresponds to
βM and the lower black line to βF⁎. Error bars represent the 95%
314 MS≥1). Differently put, these are the sensitivities of confidence intervals estimated by bootstrapping over parents (using the
315 parents' reproductive success to changes in mating success. adjusted bootstrap percentile method in R version 2.10.1 ‘boot’ package;
316 Until now there was insufficient evidence to characterize Canty & Ripley, 2009; Davison & Hinkley, 1997; R Development Core
317 the associations between reproductive and mating success Team, 2009) . In terms of fitness, polygyny strongly benefits males but
318 within either sex (Brown et al., 2009). The UPDB now weakly harms females.

gives us unprecedented power to describe these relation- 319

ships in a human population. In our analysis, we find that 320
the Bateman gradients change with birth year (Fig. 4). 321
Male gradients (βM) are highest in the 1830s, averaging 322
5.87 in the first 5 years (i.e., by taking on one additional 323
reproductive mate, the average father would have increased 324
his reproductive success by nearly six offspring). However, 325
βM diminishes linearly with birth year to 1.92 by the 326
early 1890s. 327
We can also describe the analogous relationship between 328
RS and MS for females (βF). It is weakly positive and 329
increases slightly over the 65 birth-year cohorts (from 5-year 330
means of 0.195 to 0.671; Fig. 4), indicating that increasing 331
the number of males with whom a female mates had a small, 332
but real, benefit on female reproductive fitness. This benefit 333
increased slightly over the study period. Multiple mating in 334
these females did not follow from polyandry; it is likely a 335
result of widowed or divorced mothers who remarried before 336
having more children. 337
In his classic 1969 article on the origin of polygyny, 338
Fig. 3. The opportunity for selection for males (IM, blue) and females Orians (1969) suggested that polygynous females face the 339
(IF, red) by birth year. The solid black line is the opportunity for sexual cost of having to share limiting resources with other females 340
selection Imates, and the black dotted line represents the proportion of all
selection explained by sexual selection. The strength of sexual selection
mated to the same male. They do so to their advantage, 341

decreases markedly over time, both in absolute terms and as a however, if the benefits (for example, mating with a male 342
proportion of all selection. that is of higher than average quality) outweigh that cost. To 343
 6 J.A. Moorad et al. / Evolution and Human Behavior xx (2010) xxx–xxx

344 assess the effects of polygyny on females, we calculated a the decade of birth instead of single years of birth (sample 400
345 third Bateman gradient, β ⁎F , which describes the effect of sizes are shown in Appendix A Fig. A1), ending with birth 401
346 male mating success on the reproductive success of his years 1870–1879. Otherwise, we re-analyzed Bateman 402
347 mates. A positive β ⁎F means that females benefit from gradients as before. 403
348 sharing the same mate with other females. A negative value One might expect social pressures late in the 19th 404
349 means that the relationships among females who share century to impose some additional cost to polygamist males 405
350 mates are antagonistic or competitive with respect to that would cause βM to decrease with time. In fact, we see 406
351 reproductive success (Wade, 1979). Between 1830 and no secular changes in βM, suggesting that the relative 407
352 1834, there was a small negative β ⁎F (the 5-year average benefit to males of polygamy remained high over the time 408
353 was −0.06). Over the next four decades, it became spanning this study. One likely explanation for this is that, 409
354 progressively more negative, reaching a low of −1.36 in after polygamy was outlawed in the Utah Territory, some 410
355 the late 1870s and then rising to −1.11 in the early 1890s individuals who opposed the ban continued to engage in 411
356 (Fig. 4). That is, for the average mated women born in the polygamy in Utah while others immigrated to other 412
357 late 19th century, sharing a mate with an additional female locations including Mexico and Canada where marriage 413
358 would have reduced her number of offspring by slightly practices were less likely to be regulated. This subset of 414
359 more than one. parents in the UPDB had an outsized contribution to βM in 415
360 Male mating success is not a perfect indicator of poly- the latter part of the 19th century as they were likely the 416
361 gamy in any species, including humans. Widowers with two major source of polygamists. Subject to reproductive 417
362 non-overlapping reproductive wives in sequence would constraints more closely resembling polygamists born in 418
363 have higher than average MS without being polygamists, the early part of the 19th century instead of their Utah- 419
364 for example. Nevertheless, these individuals and polyga- dwelling contemporaries, they married and reproduced 420
365 mists contribute to the Bateman gradients. The clear trend accordingly (Fig. 4). 421
366 showing that increased MS benefits men less with As with the male Bateman gradient, the gradient that 422
367 increased birth year may not necessarily indicate a secular measures the association between female RS and her mate's 423
368 reduction in the benefits of polygamy to males. Instead, MS, β ⁎F , changed little when monogamist males with 424
369 changes in βM may reflect a change in the relationship multiple mates were excluded from the analysis. Polygamy 425
370 between mating system and mating success. While the conferred a slight, but insignificant, cost to female repro- 426
371 frequency of polygamy declined over time in this duction in the first decade of births (the 1830s). These costs 427
372 population, monogamist males would still have remarried increased and became significant in the ensuing two decades. 428
373 to replace deceased or divorced reproductive partners. Over In the birth decades of the 1860s and 1870s, β ⁎F remained 429
374 time, these serial monogamists (men who married more negative but the 95% confidence intervals associated with 430
375 than once but never had more than one wife at a time) estimates were too wide to show significance. Loss of 431
376 comprised an ever-growing proportion of all males with significantly negative β ⁎F values may have followed from the 432
377 MSN1. Compared to males born early in the 19th century, dramatic reduction in the sample size of polygamists in these 433
378 MS in later-born males increased largely by mate later cohorts rather than a real change in the cost of polygamy 434
379 replacement (rather than mate accumulation), meaning to females. 435
380 that it was the older [and presumably less fertile — see Mating success and polygamy improve men's fitness. 436
381 Ford et al. (2000)] men from these age cohorts who gained Multiple wives can either (1) enhance male reproductive 437
382 high mating success. Similar alternative explanations may rates, increasing the fertilization opportunities at any one 438
383 account for the decline in β ⁎F with time. Women in multi- time, or (2) lengthen the reproductive tenure (RT) of the 439
384 wife households who were born in the late 19th century husband by offering him reproductive opportunities over a 440
385 may have suffered increased reproductive constraints greater span of time. To determine the relative importance 441
386 compared to their earlier counterparts. However, it is also of each, we calculated the partial covariance between 442
387 possible that later birth years brought a larger proportion of male MS and RS, holding RT constant. The ratio of the 443
388 women whose early deaths elevated their mate's MS by partial covariance to the total covariance is the fraction of 444
389 precipitating their remarriage. Alternatively, they may have the association between male MS and RS that is 445
390 themselves married widowers who are older on average and explained by changes in reproductive rates. We find that 446
391 may have reduced fertility. the ratio was initially positive (0.340 in the early 1830s), 447
392 To address these issues, we recalculated βM and β ⁎F using decreased until it became negative in the late 1860s 448
393 only those fathers (and their mates) who either (1) repro- (−0.167) and finally reached a low of −0.481 in the early 449
394 duced with only one woman (MS=1) or (2) were identified in 1890s (Fig. 5). This suggests that high male MS increased 450
395 the UPDB as being polygamists. This effectively removed both the reproductive rates and reproductive tenures (RT) 451
396 from the analysis those monogamist fathers with MSN1. of males born early in the 19th century. The effects 452
397 Polygamists with multiple reproductive partners became become antagonistic in males born later; increased MS 453
398 extremely rare towards the end of the 19th century (Fig. 1). decreased the per-year contributions to RS but increased a 454
399 For this reason, we grouped parents into cohorts based upon male's RT. 455
 J.A. Moorad et al. / Evolution and Human Behavior xx (2010) xxx–xxx 7

the span over which they mate and by increasing 480

reproductive rates. Both effects contribute to reproduction 481
in roughly equal measure. For females, the association 482
between female RS and MS (βF) is weak, especially so 483
early in the 19th century when the population presumably 484
more closely experienced natural fertility conditions. 485
Applied to recent humans at least, results from this 486
population seem to favor early views of female MS that 487
de-emphasize its value to fitness over the more recent 488
perspectives that stress its adaptive value (Hrdy, 1997; 489
Jennions & Petrie, 2000; Wolff & Macdonald, 2004). This 490
interpretation might not be applicable to all human 491
populations as evidence from other human datasets suggests 492
that the direction of the relationship between female MS 493
and RS may be culture specific (Brown et al., 2009). 494
Finally, we observe that, on average, a female's RS 495
declines as her mate's MS increases (β ⁎F b0), suggesting a 496
fitness cost of polygamy to females. This contrasts to 497
consistently high male benefits of polygamy and provides 498
key evidence in support of the polygyny threshold model 499
Fig. 5. Relative effect of MS on RS that is caused by increasing the (Orians, 1969). Earlier work on this historical Utah 500
reproductive rate of males. The lower black line uses all data grouped into
birth-year cohorts. The upper black line with dots and error bars (95%
population has shown a similar pattern of reduced fertility 501

confidence intervals — see Fig. 4 legend) excludes monogamous men with in females involved with plural marriages (Anderton & 502
MSN1. These are grouped into birth decades. Roughly half of the male Emigh, 1989; Bean & Mineau, 1986). 503
benefit of polygyny derives from increased male reproductive rates; the This work bears on the applicability of Alexander's 504
remainder of the fitness advantage follows from the increased time over (1979, 1987) influential ideas on the role of male choice to 505
which children are fathered.
the evolution of human mating systems. Alexander 506
proposed that cooperation between males is opposed by 507
high male reproductive variance and that this intense 508
456 As we observed with the Bateman gradients, changes in conflict suppresses the establishment of effective social 509
457 the relative contribution of reproductive rate vs. reproductive functions. Monogamy that is imposed on males by males is 510
458 tenure could be due to secular changes in the relationship viewed as one way by which within-group competition is 511
459 between mating system and mating success. As before, we lessened to the advantage of the competitive value of the 512
460 investigated this possibility by repeating the analysis after population as a whole. 513
461 eliminating monogamist males who reproduced with multi- Our findings indicate that the choice to forgo the oppor- 514
462 ple females. Polygamy increased reproductive success by tunity to be polygamous can be very expensive in terms of 515
463 lengthening reproductive tenure and increasing per-year fitness, at least for those males who have the capacity and 516
464 reproductive rates by roughly equal degrees. In the first two competitive means to have many females. Because the 517
465 birth decades (1830s and 1840s), 39.7% and 42.7% of the interests between the most powerful and influential males 518
466 benefit of MS to RS in males follow directly from increasing and the group as a whole are in conflict, the fitness 519
467 reproductive rates. The increase in rate effects on RS advantages gained by societies by banning the practice may 520
468 increased significantly in the following two birth decades also be large if cooperative restraints on polygamy are to be 521
469 (55.2% and 57.9%). This increase differs markedly from the favored by natural selection. In the case of members of the 522
470 negative effect observed from the more inclusive analysis LDS church in Utah, however, the imposition of monogamy 523
471 that includes serial monogamists. originated outside the group rather than from the social 524
organization of reproductively competing males within the 525
group. Further work should be done to test the feasibility of 526
472 4. Discussion Alexander's ideas, especially comparative analyses that 527
describe the relationships between rates and incidence of 528
473 Our study confirms the three predictions of sexual polygamy and population growth rates across many 529
474 selection theory (Bateman, 1948; Shuster & Wade, 2003; societies. Explicit multi-level selection models should be 530
475 Wade, 1979; Wade & Shuster, 2004b) in relation to mating developed to better understand what biological and social 531
476 system: (1) male σ2RS is greater than female σ2RS; (2) male conditions are necessary to allow the evolution of mono- 532
477 σ2MS is greater than female σ2MS; and (3) the association gamy by Alexander's mechanism. 533
478 between MS and RS is stronger in males than in females. Notably, the sex differences in reproductive and mating 534
479 For males, polygamy increases reproduction by lengthening success decreased as laws and religious customs changed 535
 8 J.A. Moorad et al. / Evolution and Human Behavior xx (2010) xxx–xxx

536 during the latter half of the 19th century, causing individuals
537 to become less likely to adopt a polygamous lifestyle. One
538 result is that the decrease in male σ2MS eliminated over 58%
539 of the opportunity for sexual selection (Fig. 3). Over this
540 time, the proportion of all selection caused by sexual
541 selection, pIm +pIðmates
1 − pÞIF [where p=R/(R+1) is the frequency of
542 males born], fell from a 5-year average of 41.2% (in 1830–
543 1834) to just 22.1% (1890–1894). Interestingly, the estimate
544 from early in the 19th century is congruent with the
545 estimated range of 28.4% to 77.7% from Y-chromosome
546 and mitochondrial sequence data from other human popula-
547 tions (Wade & Shuster 2004a).

548 Acknowledgments

549 We thank Wyatt Anderson, David Hall, Martie Haselton

550 and two anonymous reviewers for helpful comments and
551 criticism of this manuscript. Fig. A2. Mean reproductive and mating success. Males and female RS (blue 559
560
552 We thank the Pedigree and Population Resource and red, respectively) correspond to the left axis. Mean female mating 561
553 (funded by the Huntsman Cancer Foundation) for its success corresponds to the right axis. As we explain in the Methods section, 562
554 role in the ongoing collection, maintenance and support of we have assumed that mean male mating success is 4.8% higher. 563
555 the Utah Population Database (UPDB). We also acknowl-
556 edge Dr. Geraldine P. Mineau and Alison Fraser, MSPH,
References 564
557 for their careful management of and assistance with the
558 data used for this study. Alexander, R. D. (1979). Seattle: University of Washington Press, p. 317. 565 Q3
Alexander, R. D. (1987). Hawthorne, NY: A. de Gruyter, p. 301. 566
Andersson, M. B. (1994). Princeton, NJ: Princeton University Press, p. 624. 567
Anderton, D. L., & Emigh, R. J. (1989). Polygynous fertility — sexual 568
competition versus progeny. American Journal of Sociology, 94(4), 569
Appendix A 832–855. 570
Bateman, A. J. (1948). Intra-sexual selection in Drosophila. Heredity, 2(3), 571
349–368. 572
Bean, L. L., & Mineau, G. P. (1986). The polygyny-fertility hypothesis: a re- 573
evaluation. Population Studies, 40(1), 67–81. 574
Bjorklund, M. (1991). Evolution, phylogeny, sexual dimorphism and mating 575
system in the Grackles (Quiscalus Spp, Icterinae). Evolution, 45(3), 576
608–621. 577
Borgerhoff-Mulder, M. (1998). The demographic transition: are we any 578
closer to an evolutionary explanation? Trends in Ecology & Evolution, 579
13(7), 266–270. 580
Brown, G. R., Laland, K. N., & Mulder, M. B. (2009). Bateman's principles 581
and human sex roles. Trends in Ecology & Evolution, 24(6), 297–304. 582
Canty, A., & Ripley, B. (2009). boot: Bootstrap R (S-Plus) Functions. 583
R package version. Version 1.2-38; 2009. 584
Crow, J. F. (1958). Some possibilities for measuring selection intensities in 585
man. Human Biology, 30, 1–13. 586
Darwin, C. (1859). London: John Murray, p. 512. 587 Q4
Davison, A. C., & Hinkley, D. V. (1997). Cambridge: Cambridge University 588
Press, p. 582. 589
Daynes, K. M. (2001). Urbana: University of Illinois Press. x, 305 p. 590
Fisher, R. A. (1930). The genetical theory of natural selection. Oxford: 591
Clarendon Press; 1930. 592
Ford, W. C. L., North, K., Taylor, H., Farrow, A., Hull, M. G. R., & 593
Golding, J., et al (2000). Increasing paternal age is associated with 594
delayed conception in a large population of fertile couples: evidence for 595
declining fecundity in older men. Human Reproduction, 15(8), 596
Fig. A1. Sample sizes by birth year for males (blue) and females (red), 1703–1708. 597
including the projected nulliparous individuals. The black line represents Hrdy, S. B. (1997). Raising Darwin's consciousness — Female sexuality and 598
the sample size of polygamous men with MSN1. These are grouped into the prehominid origins of patriarchy. Human Nature-an Interdisciplinary 599
birth decades. Biosocial Perspective, 8(1), 1–49. 600
 J.A. Moorad et al. / Evolution and Human Behavior xx (2010) xxx–xxx 9

601 Jennions, M. D., & Petrie, M. (2000). Why do females mate multiply? Sixteenth Census of the United States (1943). Population. Differential 619
602 A review of the genetic benefits. Biological Reviews, 75(1), 21–64. fertility 1940 and 1910. Washington DC: Government Printing 620
603 Jones, A. G., & Avise, J. C. (2001). Mating systems and sexual Office. 621 Q6
604 selection in male-pregnant pipefishes and seahorses: insights from Smith, J. E., & Kunz, P. R. (1976). Polygyny and fertility in nineteenth- 622
605 microsatellite-based studies of maternity. Journal of Heredity, 92(2), century America. Population Studies, 30, 465–480. 623
606 150–158. Wade, M. J. (1979). Sexual selection and variance in reproductive success. 624
607 Lofredo, C. A., & Borgia, G. (1986). Sexual selection, mating systems, and American Naturalist, 114(5), 742–747. 625
608 the evolution of avian acoustical displays. American Naturalist, 128(6), Wade, M. J. (1995). The ecology of sexual selection — mean crowding of 626
609 773–794. females and resource-defense polygyny. Evolutionary Ecology, 9(1), 627
610 Morgan, S. P. (1991). Late nineteenth- and early twentieth-century 118–124. 628
611 childlessness. American Journal of Sociology, 97(3), 779–807. Wade, M. J., & Shuster, S. M. (2004a). Estimating the strength of sexual 629
612 Orians, G. H. (1969). On evolution of mating systems in birds and mammals. selection from Y-chromosome and mitochondrial DNA diversity. 630
613 American Naturalist, 103(934), 589–603. Evolution, 58(7), 1613–1616. 631
614 R Development Core Team. (2009). R: A language and environment for Wade, M. J., & Shuster, S. M. (2004b). Sexual selection: harem size and the 632
615 statistical computing. Vienna, Austria: The R Foundation for Statistical variance in male reproductive success. American Naturalist, 164(4), 633
616 Computing; 2009. E83–E89. 634
617 Shuster, S. M., & Wade, M. J. (2003). Princeton, NJ: Princeton University Wolff, J. O., & Macdonald, D. W. (2004). Promiscuous females protect 635
Q5 618 Press, p. 520. their offspring. Trends in Ecology & Evolution, 19(3), 127–134. 636
637
Our reference: ENS 5674 P-authorquery-v8

AUTHOR QUERY FORM

Journal: ENS Please e-mail or fax your responses and any corrections to:
Jill Shepherd
E-mail: J.Shepherd@Elsevier.com
Tel: 352-483-8113
Article Number: 5674 Fax: 352-483-3417

Dear Author,

Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in
the proof. Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-
screen annotation in the PDF file) or compile them in a separate list.

For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.

Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in
the proof. Click on the ‘Q’ link to go to the location in the proof.

Location Query / Remark: click on the Q link to go

in article Please insert your reply or correction at the corresponding line in the proof
Q1 No cities were provided in the affiliations. Please confirm whether the ones supplied are correct.
Q2 Morgan (1991) was not cited in the text and was temporarily inserted here. Please check.

Q3 Please provide editor(s) for Alexander, 1979, 1987 and Andersson, 1994.
Q4 Please provide editor(s) for Darwin, 1859, Davison and, Hinkley, 1997, and Daynes, 2001.

Q5 Please provide editor(s) here.

Q6 Please provide editor(s) and page range.

Thank you for your assistance.

Page 1 of 1