Polygyny is the most common mating system in mammalian species (95%), yet our understanding of polygynous systems and microevolutionary processes is still limited. Pinniped mating systems range from extreme polygyny (e.g. elephant seals) to sequential female defence by males and hence have often been used as models for mating system studies. Parentage analysis has enabled the examination of mating success, the identification of pedigrees, and the elucidation of social organisation, greatly enhancing our understanding of mating systems (Chapter 1). However, such analyses are not without pitfalls, with erroneous assignments common in open systems (i.e. when parental and offspring samplings are incomplete). We investigated the effects of the user-defined parameters on the accuracy of parental assignment using two commonly used parental allocation programme, CERVUS and PASOS (Chapter 2). We showed that inaccurate user-defined parameters in CERVUS and PASOS can lead to highly biased output e.g. the assignment rate at 95% CL of offspring with a sampled known mother to sampled males decreased from 58% to 32% when the proportion of candidate males sampled in the parameter options decreasing 4-fold. We found that the use of both CERVUS and PASOS for parentage assignment can increase the likelihood of correctly allocating offspring to sampled parents to 97% in our study system. Incorrect parental assignment can bias estimates of various biological parameters, such as lifetime reproductive success and mate choice preference, and hence bias ecological and evolutionary interpretations. Here, we propose solutions to increase the power of parentage assignment and hence decrease the bias in biological parameter estimates.

In addition, we analysed the effects of the intrinsic bias in likelihood assignment approaches towards assigning higher probability of parentage on individuals with rare alleles and those with heightened offspring-parent matches, which increase with the number of homozygous loci (Chapter 3). We showed that, as a consequence of the algorithms employed in the programmes CERVUS and PASOS, heterozygote males with rare genotypes are assigned higher rates of parentage than males with common alleles. Consequently, where two males could both be biological fathers of a given offspring, parentage assignment will more often go to the male with the rarer alleles (most often in heterozygous loci). Thus, the commonly used parentage assignment methods may systematically bias the results of parentage analyses towards supporting the notion that females prefer more genetically unusual, most often heterozygous, males. Such a bias may sway investigators towards incorrectly supporting the concept that females choose genetically more unusual males for heterozygosity fitness benefits that underpin the good genes hypothesis, when in fact no such relationship may exist.

In polygynous mammals, successful males mate with multiple females by competing with and limiting the access of other males to females. When the status of many males (age, size, health, genetic etc.) prevents them from achieving the primary mating tactic, theory predicts selection for a diversification of male mating tactics. Recent studies in pinnipeds have shown that observed male mating success was correlated to male paternity success in some species (elephant-seals), but not in others (grey seals). The existence of alternative mating strategies can explain those discrepancies. Chapter 4 implemented the guidelines provided in Chapter 2 and 3 and focused on the polygynous New Zealand fur seal Arctocephalus forsteri, predicting that 1) competition for females is likely to cause a diversification of male mating tactics; and 2) that alternative tactics can yield reproductive success. Our results indicated three male behavioural profiles; one corresponded to large territorial males and two illustrated a continuum of alternative tactics employed by non-territorial subordinate males. Our study highlights that holding a territory is not a necessary condition for reproductive success in a population of otariids.

The degree of sexual size dimorphism in polygynous species is expected to increase with the degree of intra-sexual competition and in turn with the degree of polygyny. The life history of an individual is the pattern of resource allocations to growth, maintenance, and reproduction throughout its lifetime. Both females and males incur viability costs of mating and reproduction. However, male viability costs due to increase growth and male-male competition can be greater than female viability costs of mate choice and reproduction. Although an abundant literature on sexual dimorphism in morphology, physiology, and parasite infections is available, little is known on the intra-sexual differences in physiology and parasite infections associated to the reproductive success of different mating strategies in mammalian species. Chapter 5 examined the reproductive costs between territorial and subordinate males New Zealand fur seal related to their relative reproductive success using a multidisciplinary approach (behaviour, genetics, endocrinology, parasitology). We found that dominant New Zealand fur seal males endure higher reproductive costs due to the direct and indirect effects of high testosterone levels and parasite burdens. Our study highlights that holding a territory confers a higher reproductive success, but induces higher costs of reproduction that may impair survival.

Understanding microevolutionary processes associated to polygynous systems is fundamental in light of the ongoing anthropogenic alteration of the environment through climatic variations and habitat reduction which ultimately affect opportunity for sexual selection and shape the life history trade-offs.