Male secondary sexual characters can be quite distinct, striking, and elaborate in nature. Despite many advances in the field of sexual selection, much remains to be known regarding why some organisms evolve these features more than others. Since Darwin published On the Origin of Species in 1859, many studies have measured the strength of natural selection in the wild showing that it is often strong and rapid (Both and Visser 2001; Pelletier et al. 2007; Kinnison et al. 2008; reviews: Hendry and Kinnison 1999; Reznick and Ghalambor 2001; Stockwell et al. 2003; Strauss et al. 2008; reviews: Endler 1986; Kingsolver et al. 2001; Hairston Jr. et al. 2005). Despite all these examples, a recent review has shown that very few of the traits measured in these studies involve secondary selected traits (Svensson and Gosden 2007). This is unfortunate because secondary selected traits often represent the most complex, elaborate traits in a variety of taxa including plants (eg. Geber, Dawson, and Delph 1998), insects (ex. Stubblefield and Seger 1994), fish (eg. Basolo and Trainor 2002), birds (eg. Hill and McGraw 2006), and reptiles (eg. Schulte-Hostedde and Schank 2009). Moreover, the evolution of one these particular traits in nature, male guppy coloration, represents one of our best examples of rapid evolution (Endler 1980).

Adaptation requires both inheritance and selection, however most studies in rapid evolution either ignore heritability and concentrate on selective pressures or assume a particular mode of inheritance. Theoretical models have long established the importance of genetic architecture through sex linkage in sex-limited or sexually selected traits. However, empirical research in this topic is rare. In this thesis I present a comprehensive experimental study of how ecology and sex-linkage may interact to maintain the genetic variation and complex polymorphism in a sexually antagonistic, sex-limited trait in the wild.

First I use a standard multivariate animal model to evaluate the heritability of two sub-traits of male coloration known to be linked to male fitness; orange and black body coloration. I also partition phenotypic variance of two introduced populations of guppies Poecilia reticulata) into its environmental and genetic components. The genetic components are then further partitioned into Y-linked versus non Y-linked variance to test the idea that sexually selected male traits are generally linked to the Y-chromosome where evolution is presumed to be faster as established by theory. I also studied genetic correlations among the two color patterns, and use all findings to predict the future trend of evolutionary change in this novel introduction. Using a quantitative genetics approach in this manner can help extract the genetic parameters affecting evolutionary change, and to my knowledge is the first study that separates Y-linked from non Y-linked quantitative genetic variance using a wild pedigree. Results show high proportion of Y-linked to non-Y linked genetic variance and that overall variation in Y-linkage accounts for most of the phenotypic variation in both introduction sites. Both sub-traits are also highly heritable and so combined with the abrupt change in selection pressure with the introduction I predict evolutionary change to be rapid in these populations.

Second, following the results garnered in Chapter 1, here I track changes in adaptive divergence in both introduction sites bimonthly for one year post-introduction to see if our predictions were sound. My goal in this chapter was to investigate how variation in different selective pressures, such as predation and stream canopy cover, affect rates of divergence in a sexually selected polymorphic trait. Guppies were introduced from environments where they coexist with predators to two novel environments where there are no predators. In addition to the abrupt change in predation pressure, I also manipulated the canopy cover in one introduction site, hence doubling productivity in that environment. Results show rapid phenotypic and genotypic divergence in male coloration as expected, to date the fastest measure of change in wild guppies. Results also demonstrate that abrupt changes in habitat as well as predation-mediated mortality rates affect variation in rates of evolution of secondary sexual characters, an idea previously proposed but never formally tested.

In the third Chapter I test the idea that microgeographic variation in sex-linkage occurs in multiple high- versus low-predation guppy populations, the first step needed to test theory regarding interactions between sex-linkage and selection. I examine a hypothesis that high-predation guppies have mainly Y-linkage of color patterns whereas low-predation guppies have color patterns linked to both the X- and Y-chromosome. In my thesis chapter I examine multiple high- and low-predation natural population using hormone assays in female guppies (which normally do not show coloration, and do not have a Y-chromosome) to test for differences in X-/autosomal linkage. I presume that changes in the amount of non-Y linked inheritance are combined with changes in Y-linkage. I also examine three introduction populations (from high- introduced into low predation sites) to see if these differences in linkage relationship respond rapidly to selection pressure. Results show that indeed low-predation guppies show a significantly higher amount of non-Y linked color patterns compared to high-predation guppies, and that this variation in linkage relationship can evolve in a matter of few guppy generations.

There is remarkable diversity in the form and function of vertebrate reproductive mode, and adaptive explanations for the vast differences among species have fallen short. Instead, parent-offspring conflicts provide a parsimonious framework that describes why evolutionary transitions occur from one mode to another, and how the differences among species change the nature of sexual selection and speciation. My dissertation examines the effect of reproductive mode on vertebrate evolution by examining two topics – the evolution of matrotrophy following a transition from oviparity to viviparity and the effect differences in reproductive mode have the evolution of reproductive isolation and the rate of speciation. Cyprinodontiformes, an order of small mostly freshwater fish, are notable for exhibiting a wide range of reproductive phenotypes. I make use of the repeated transitions from oviparity to viviparity and from lecithotrophy to matrotrophy in Cyprinodontiformes to test hypotheses that parent-offspring conflicts have driven the evolution of reproductive mode.

In chapter one I demonstrate eggs from three oviparous species from Cyprinodontiformes and one from Atherinomorpha are capable of acquiring molecules from their surrounding environment via pinocytosis, a property that predisposes them to the evolution of matrotrophy following the transition to viviparity. In chapter two I find evidence that post-zygotic reproductive incompatibilities are evolving faster among populations within the placental species Poeciliopsis prolifica than within two closely related non-placental species of Poeciliopsis. In the placental species, offspring size decreased significantly as a function of increasing interpopulation distance, but offspring from non-placental species suffered no such fitness loss. In chapter three I demonstrate that interspecific post-zygotic reproductive isolation evolves at an accelerated rate among viviparous species relative to oviparous species, and that estimated levels of post-zygotic isolation are higher among matrotrophic species than among lecithotrophic species at all genetic distances. Similarly, I find diversification rates estimated from molecular phylogenies to be significantly higher for viviparous taxa than oviparous taxa, but marginally higher for lecithotrophic species than for matrotrophic species. As a whole, the results of this dissertation are consistent with hypotheses that parent-offspring conflicts have played a part in the evolution of vertebrate reproductive mode, and that variation in the nature of conflicts among taxa influence speciation.

Parent-offspring conflict arises because of the unequal patterns of relatedness among parents, offspring, and siblings. From an evolutionary perspective, parents optimize investment in each offspring to maximize their lifetime reproductive success. Females are equally related to all their offspring, so natural selection favors them allocating resources equally to each offspring. In contrast, natural selection will favor those offspring that acquire more resources, even at the expense of their mother and siblings. Recent thinking expands the potential influence of parent-offspring conflict from post-natal familial interactions to other facets of the organism’s biology. The evolution of a livebearing reproductive mode, particularly involving a placenta, for example, is predicted to cause pre-natal mother-offspring conflict. In the livebearing fish family Poeciliidae, a placenta-like organ has evolved independently at least nine times. In my dissertation, I examine the potential for reproductive barriers in placental (Heterandria formosa and Poeciliopsis prolifica) and non-placental (P. infans) poeciliid species. I perform both natural and artificial insemination crosses between populations within three separate species of fish in the family Poeciliidae. In my first chapter, I find that both placental and non-placental species exhibit reproductive incompatibility, but the location of the incompatibility is different. In my second and third chapters, I use artificial insemination to further investigate the effects of inter-population crosses on offspring size and number in two placental species, H. formosa and P. prolifica. In these species, natural differences in offspring size between populations of H. formosa are the source of conflict in offspring size. However, when females are mated to both their own and nonresident males, they produce intermediate offspring, discriminating against nonresident male’s embryos. In P. prolifica, while there are no natural offspring size differences in the two populations, we find a similar pattern of incompatibility present in crosses that is mitigated when females are provided with sperm from her own, and a genetically distinct population. These results provide a new path for studying reproductive incompatibility and conflict in placental species, as well as evidence for cryptic female choice and discrimination against genetically different males.

For decades there has been considerable interest in understanding how variation in dispersal arises in nature. However, studying the dispersal of natural populations in an evolutionary framework is logistically challenging, thus, much of what we know about dispersal evolution comes from studies performed in artificial environments or in species with simplified dispersal morphologies. The objective of this dissertation is to examine how divergent natural populations vary in their tendency to disperse. To do so, I examined dispersal in guppies (Poecilia reticulata) across the Northern Range Mountains of Trinidad. In Chapter 1, I used spatially-explicit, mark-release-recapture experiments to identify differences in dispersal patterns among natural populations adapted to high- and low-levels of predation. I found that across all comparisons, the high-predation (HP) individuals were more likely to disperse and dispersed further than their low-predation (LP) counterparts. In Chapter 2, I collected HP and LP guppy populations and tested for movement differences in multi-patch stream mesocosms. I found that HP guppies are more inclined to move and make more movements within the mesocosms. In Chapter 3, I collected and reared wild HP and LP populations to a F2 generation and used multi-patch stream mesocosms to compare their movements with wild populations. The results revealed that the movement differences are retained even after being reared in a common environment, which suggests that dispersal has a strong genetic component. In Chapter 4, I experimentally displaced fish and examined whether there is intraspecific variation in homing success that is related to predation regime or dispersal status. I found no association between homing success and dispersal at the individual level, but I found a strong effect of phenotype. HP individuals exhibited greater homing success compared to LP fish. Across all 4 chapters I have documented consistent and predictable differences in dispersal traits between HP and LP populations. These studies include 6 evolutionarily independent replicates, two of which were experimental introductions. Overall, this work highlights the ecological and evolutionary relevance of intraspecific, genotypic variation in dispersal tendency and demonstrates that dispersal evolves under selective regimes that also drive the rapid evolution of life history traits.

Across the tree of life there is a great deal of variation in life cycles, life histories, and reproductive strategies. Explaining this diversity in terms of selective pressures and ecological conditions is one of the grand challenges within the field of evolutionary biology. Studying the extremes, or the ends on this spectrum of life cycle variation, can provide a fuller understanding of how different strategies evolve and allow for population persistence. My dissertation research is united by two themes - the evolution of reproductive mode, and adaptation to ephemeral and variable environments. Killifish, small oviparous fishes within the Order Cyprinodontiformes, have evolved an annual life cycle and are adapted to life in seasonally ephemeral aquatic habitats. The most prominent adaptation of these short-lived killifish are embryos capable of undergoing diapause (halting development) at one or more of three different stages during embryology and remaining buried in the soil for much of the year. In this dissertation I combine a phylogenetic, comparative, and experimental approach to study the evolution of this life cycle and how through embryonic diapause these fish have adapted to ephemeral and variable aquatic habitat.

In chapter one I demonstrate convergent evolution of alternative developmental trajectories associated with diapause in African and South American killifish species. Adaptation to seasonal aquatic environments in annual killifish imposes strong selection during the embryo stage leading to marked diversification during a mid-embryogenesis period that is otherwise highly conserved during vertebrate development. In chapter two, I demonstrate that the embryos of an annual killifish Nothboranchius furzeri exhibit a combination of phenotypic plasticity and bet-hedging (a risk spreading strategy). Specifically, whether embryos enter diapause is influenced by environmental factors (temperature and light level) that vary seasonally but also exhibits a measure of intrinsic variability, even after controlling genetics and environment. In chapter three, I compile available evidence from the literature and my own comparative experiments and provide a plausible scenario for how an annual life cycle evolved through intermediate steps. Killifish are found in aquatic habitats that span a continuum from permanent and stable to seasonal and variable, thus providing a useful system in which to piece together the evolutionary history of this life cycle using natural comparative variation embedded in a phylogenetic context.

Convergent evolution of a trait can occur at interspecific or intraspecific levels. Traits that are convergent within species are predicted to have a similar genetic basis due to recently shared ancestry. However, it is unclear how likely it is to find similar genetics during cases of rapid evolution or evolution on historic timescales. My dissertation examines evidence for phenotypic and genetic convergence in guppies, which are found in recently-introduced and historically-existing sites that differ in predation regime. I use ornamental tail coloration as it varies highly among developmental stages within an individual, males and females, individuals within a site, and between sites.

In Chapter one, I demonstrate that color develops in a consistent order among males from different predation regimes and I propose that early maturing males are the ideal stage for future experiments. In Chapter two, I determine that black melanin-related genes are likely candidates for genetic differences among fish varying in the presence of ornamental coloration. In Chapter three, I find evidence that phenotypic color pattern differences during rapid evolution follow a mixture of unique and similar trajectories based on current selective environment experienced. The color gene differences exhibit partial molecular convergence, indicating that guppies maintain variation within and between sites. In Chapter four, I discover phenotypic color pattern similarity among independently derived drainages and predator communities with substantial genetic divergence. I find a lack of molecular convergence in the identity of differentially expressed color genes among geographic locations. Taken together, these results suggest that the scale at which convergence is examined is very important. Sites may exhibit an overall signal for convergence, but may also follow slightly different trajectories upon closer examination. Convergent evolution of a trait should not be considered as being entirely present or not present, but rather a mixture of contingent and similar evolutionary processes acting at the level considered.

Placental matrotrophy has evolved independently in many lineages, which is indicative of strong selection. Several ad-hoc hypotheses and one mathematical model have been proposed to explain its evolution, but little experimental data exists to support these claims. Recent studies have attempted to find correlates between placental matrotrophy and other life history traits, but the only common correlate found was a reduction in reproductive allocation (RA). Here I report on three studies. The first two focus on the Trexler-DeAngelis model for the evolution of matrotrophy. In chapter one, I use closely related placental and non-placental species from the northern clade of Poeciliopsis to test an assumption imperative to the model, that placental species abort a subset of developing offspring in low food conditions. The results show no evidence of abortion due to food level. Instead, placental species appear to be tethered to a brood once initiated, and sacrifice body condition to maintain reproduction when resources are restricted. However, an alternative explanation for these results is that the pattern of resource allocation is a function other life history traits, rather than placentation alone. Chapter two distinguishes between these alternatives, by performing a similar experiment on the southern clade of Poeciliopsis, which has the opposite relationship between life history traits and placentation seen in the northern clade. The results mirror those from the northern clade, indicating reproductive mode, rather than life history, dictates the pattern of resource allocation. This further adds to the difficulties of explaining the evolution of the placenta within the constraints of the Trexler-DeAngelis model. The third study characterizes locomotor consequences related to differences in reproductive allocation using the Trinidadian Guppy (Poecilia reticulata), because placentation is correlated with a reduction in reproductive allocation. Females with the highest RA were less streamlined, had lower escape velocities, and had to work harder during prolonged swimming. In low predation natural habitat, these same females were restricted to habitat with slower moving water, while females with lower RAs were able to use faster flows, which are thought to be preferred feeding grounds. All high predation fish used slow moving water, irrespective of RA.

Multiple, potentially interacting features of the environment and the phenotype shape the way organisms evolve through adaptation. Ecomorphology aims to understand such intricate relationships between and within environmental and phenotypic variables. In this dissertation I made an integrative analysis of the combined effects of current velocity and predation on the evolution of body shape and life history of fishes.

In the first chapter I made an interspecific analysis of body shape in Gambusia females. As predicted, they showed larger caudal areas in high-predation sites. In contrast, flow was not related to shape but an interaction between flow and predation was detected. Pregnancy reduced the magnitude of phenotypic divergence between females from contrasting predation and flow regimes. These results indicate that adaptation to predation, and to a lesser extent flow, is key to body shape evolution in the genus Gambusia.

In the second chapter I analyzed the macroevolutionary pattern of life history in female Gambusia and its relationship with predation, flow, and latitude. Gambusia females showed a large range of variation in reproductive allocation, number of embryos, and size of embryos whereas they varied little in their mode of maternal provisioning. Contrary to predictions, life history was not related with predation, flow, or latitude. This may indicate that life history traits are more labile than shape traits, and that other factors could be responsible for the variation observed.

In the third chapter I explored more in deep the relationship between flow regime, shape, and life history. I compared populations of Trinidarian guppies from sites with contrasting hydrologies but similar predation regime. Females from sites prone to floods had larger caudal peduncles, which could be useful to escape during a flood. Male shape was not related to flow regime. High-flood sites had females with more babies and higher reproductive allocation, and males maturing at larger sizes. This pattern of life history matches the predicted effect of an in increase in resource abundance under density dependence. In guppies, flood-prone sites may be environments with increased food abundance for the survivors, allowing them to grow larger and invest more in reproduction.

Food availability during the juvenile stage can a have profound impact on adult performance and fitness. Low food availability early in life generally leads to slower growth and development and maturation at an older age and smaller size, which can then lead to reduced survival and lifetime reproductive success. There is increasing evidence that organisms can respond to these early setbacks and in ways that might be adaptive. However, our current understanding of the diversity of compensatory mechanisms utilized, the costs and benefits of these responses, and the relative importance of these responses in the wild is still in a nascent stage. For this dissertation, I investigated how juvenile food availability influences adult reproductive decisions and success in Trinidadian guppies (Poecilia reticulata).

First, I examined whether juvenile compensatory growth has negative consequences for reproduction. I found that juvenile compensatory growth did not affect adult growth rates, litter production rates, or investment in offspring size, but had negative effects on litter size. Second, I investigated whether individuals during the adult stage can mitigate the negative effects of early setbacks caused by low food availability. I found that females reared on low food matured at a later age, a smaller size and with less energy reserves than females reared on high food. In response to this setback, they changed their investment in growth, reproduction and fat storage throughout the adult stage such that they were able to catch up in body size, increase their reproductive output and restore their energy reserves to levels comparable to those of females reared on high food. Finally, I explored the relative effects of juvenile versus adult environmental conditions on adult growth strategies in the wild. I found that events occurring during early and later parts of an individual's ontogeny both had important consequences for adult growth strategies, but the direction of their influence differed. Poor conditions during the juvenile stage had a positive effect on adult growth rates, while those same conditions in the recent past and in the present had a negative effect on adult growth.

Recent theoretical and empirical work argues that growth rate can evolve and be optimized, rather than always being maximized. Chronically low resource availability is predicted to favour the evolution of slow growth, whereas attaining a size-refuge from mortality risk is predicted to favor the evolution of rapid growth. Guppies (Poecilia reticulata) evolve difference in behaviour, morphology and life-history traits in response to predation, thus demonstrating that predators are potent agents of selection. Predators in low-predation environments prey preferentially on small guppies, but hose in high-predation environments appear to be non-selective. Because guppies can outgrow their main predator in low- but not high-predation localities, we predict that predation will select for higher growth rate in the low-predation environments. However, low-predation localities also tend to have lower productivity than high-predation localities, yielding the prediction that guppies from these sites should have slower growth rate. Here, we compare the growth rates of the second laboratory-born generation of guppies from paired high- and low-predation localities from four different drainages. In two out of four comparisons, guppies from high-predation sites grew significantly faster than their low-predation counterparts. We also compare laboratory born descendants from a field introduction experiment and show that guppies introduced to a low-predation environment evolved slower growth rates after 13 years, although this was evident only at the high food level. The weight of the evidence suggests that resource availability plays a more important role than predation in shaping the evolution of growth rates.