Experience counts: lessons from studies of differential allocation

Abstract

One of the most fundamental decisions that female birds face after pairing is how much to invest in a particular reproductive attempt (Zhang et al., 1996). According to differential allocation (DA) theory, females are able to adjust their investment according to factors that affect the value of the breeding attempt, such as the attractiveness of their mate. A higher value is placed on matings with attractive males because of direct and/or indirect fitness benefits (Andersson, 1982; Hamilton and Zuk, 1982; Hoelzer, 1989; Norris, 1990), and therefore females should be willing to bear greater costs and invest more in offspring fathered by such males (Burley, 1988; Sheldon, 2000). However, in direct contrast is the compensation hypothesis, in which a female is predicted to compensate for reduced offspring viability when paired to a lower quality male, and the reverse allocation pattern occurs (Bluhm and Gowaty, 2004a,b; Michl et al., 2005; Saino et al., 2002a). Avian research on DA began with Burley’s now classic work on zebra finches (Taeniopygia guttata), in which females mated to more attractive males fed their offspring more frequently than females mated to unattractive males (Burley, 1988). Evidence in support of DA has since accumulated in a number of species (primarily in birds) and in a variety of different forms, with most recent work focusing on primary reproductive investment. For example, females have been found to alter their investment at egg laying by producing more, larger or higher quality eggs when mated to an attractive male (Cunningham and Russell, 2000; Gil et al., 1999; Petrie and Williams, 1993; Saino et al., 2002b). Fundamental to DA are three basic assumptions. First, that females do allocate differentially. Second, that in order to do so they are able to respond in a flexible manner to environmental and social cues. Third, that a trade-off exists between current and future reproductive effort, such that elevated investment during one reproductive episode will be traded-off against a future episode (Sheldon, 2000). This trade-off is fundamental to understanding why a female should withhold investment in a current breeding attempt if a potentially more valuable breeding attempt should later present itself. Experimental manipulations are necessary to provide convincing evidence of DA (Sheldon, 2000), and crossover designs are widely used in such experiments (e.g., Balzer and Williams, 1998; Cunningham and Russell, 2000; Gil et al., 1999; Petrie and Williams, 1993). Here, individuals are assigned randomly to one of two treatments in the first experimental round, and then treatments are swapped in the second round, thus using each individual as its own control (Ruxton and Colegrave, 2003). Crossover designs are particularly useful in studies of DA because as well as being statistically powerful they provide a test of individual flexibility, which is necessary for DA to have evolved. In this article we use examples from our own zebra finch research and other studies to discuss how crossover designs can often produce complex carryover effects. These complicate the interpretation of the results (a warning made by Ruxton and Colegrave, 2003) but may themselves prove informative with respect to issues not previously considered in DA. For example, further investigation of carryover effects and the effects of prior female experience in general on current reproductive decisions may reveal interesting adaptations and provide some insight into the subtleties behind DA.