Condition dependence and fitness consequences of sexual traits in zebra finches

The focus of this thesis is on sexual selection within a
behavioural and evolutionary framework. Sexual selection is
concerned with the factors that determine reproductive success. I
have used a laboratory population of zebra finches to address
fundamental questions that remain unanswered despite decades of
research on this model organism in studies of sexual selection in
monogamous species. The aim of this thesis was to perform a
rigorous investigation of how male zebra finches achieve
reproductive success. To do this, I tried to evaluate the
signalling value of candidate traits and investigate the mechanisms
that lead to reproductive success through both intra- and
intersexual selection. The honesty of signals has had a special
place in my work. Current theory suggests that secondary sexual
signals should be costly to be able to function as honest
indicators of male quality in a mate choice context. If ornaments
are costly to develop or maintain, this would lead to condition
dependent expression of ornaments. I tested the condition
dependence of a broad range of traits so that I could compare the
condition dependence of sexual and non-sexual traits and between
traits in males and females (chapter 1). I found that zebra finches
are remarkably resilient to stressful conditions during early
development and can develop normally in most traits with very
little consequences for fitness or longevity. These results
indicate that not all traits that appear to be sexually selected
characters will exhibit condition dependence. To study fitness
consequences, I used an aviary breeding set-up. This allowed me to
study the process of sexual selection in a socially complex
environment where natural selection pressures had been relaxed.
Since only selection acting on the genetic component of a trait can
lead to evolutionary change, I used a quantitative genetic approach
to look at selection pressures. Contrary to the prevailing view, I
found that beak colour and courtship song rate and song complexity
were not subject to female choice and not relevant for reproductive
success (chapter 3). Beak colour was also not important in
male-male competition (chapter 2). These findings necessitate a
re-evaluation of the function of these traits. I suggest that
courtship song rate reflects male reproductive strategies, while
song structure is important for individual recognition in this
highly social species and the beak colour might function as a
signal of breeding status. The strongest candidate for a
condition-dependent trait with consequences for fitness was body
size (measured as tarsus length). Tarsus length was dependent on
early condition (chapter 1), played a role in intrasexual
competition (chapter 2) and was related to reproductive success in
the aviaries (chapter 3). However, the quantitative genetic
approach revealed that the selection was acting mainly on the
environmental, not the genetic component of tarsus length,
explaining why sexual size dimorphism has not evolved. Contrary to
the expectations of honest signalling theory, males with a higher
undirected song rate (i.e. non-courtship song) were less successful
in the aviaries (chapter 4). However, females invested more into
reproduction when paired to a high undirected song rate male,
indicating that this type of song might function in reproductive
stimulation of the partner and maintenance of the pair bond.
Furthermore, female reproductive investment can be influenced by
male quality. In contrast to previous studies, I found that females
invested more when paired to a less attractive male (chapter 5).
This investment pattern fits with the breeding system of this
species, since life-time monogamy and a short lifespan means that
the chances of breeding with a higher quality male in the future is
very low. Thus, females should do the ‘best of a bad job’ and
compensate by investing more when paired to a less attractive
partner. Collectively, this work has provided several unexpected
insights that call for a reinterpretation of several of the
classical views of this model organism. This illustrates how novel
experimental approaches can provide new insights into the process
of sexual selection, also in well-studied model systems.