Seasonal decline in reproductive success of the Great Tit: variation in time or quality?

Ecology, Dec, 1995 by S. Verhulst, J.H. Van Balen, J.M. Tinbergen

INTRODUCTION

Reproductive success varies seasonally in animal species of diverse taxa, e.g., insects (Ohgushi 1991, Cushman et al. 1994); fish (Schultz 1993); and mammals (Clutton-Brock et al. 1982). Seasonal declines in reproductive success have been particularly well documented in birds (Perrins 1970, Daan et al. 1989). This seasonal decline in reproductive success could be causally related to the timing of breeding, but could also be caused by variation in quality between early and late breeders; i.e., if high-quality birds breed earlier than low-quality birds, this would result in a seasonal decline in reproductive success.

Differences in quality between breeding birds could result from differences in phenotypic quality, e.g., foraging ability, or from differences in environmental quality, e.g., food availability in the territory (see Ens et al. 1992). Similarly, effects of timing of breeding can be subdivided into separate categories. Absolute timing (calendar date) may be important when, for example, birds have to molt or migrate before winter. Timing relative to phenology of other organisms may be important when breeding is synchronized with seasonal fluctuations in the population of prey species (van Balen 1973). Timing relative to conspecifics may be important when age or prior occupancy play a role in territory settlement (Nilsson 1990, Sandell and Smith 1991), or, for colonial breeders, when synchronization with other members of the colony reduces predation rate (Parsons 1975, Hatchwell 1991).

To understand phenotypic variation in breeding dates in an evolutionary context requires knowledge of the fitness consequences of alternative breeding dates. In other words, it requires that we establish if there is a causal relationship between timing of reproduction and reproductive success (Daan et al. 1990, Hochachka 1990). To disentangle the effects of quality and the timing of breeding on reproductive success, experiments that manipulate the timing of breeding are needed. Such studies are relatively rare (Parsons 1975, Hunt and Hunt 1976, Dijkstra 1988, Nilsson 1990, Hatchwell 1991, Norris 1993, Brinkhof et al. 1993, Wiggins et al. 1994), and have focused exclusively on fitness components of the manipulated clutch, while ignoring possible effects of the timing of breeding on future reproductive output of the parents. However, selection acts on inclusive fitness, which includes both current and future clutches, so that evaluation of the effect of a trait on fitness requires accounting for effects on parents as well as offspring.

In this paper, we report results from an experiment testing the hypothesis that there is a causal relationship between reproductive success and timing of breeding in Great Tits, Parus major. We manipulated timing of breeding during four years by removing first clutches to induce birds to lay replacement clutches, thereby creating a group of experimentally delayed breeding pairs. The 'timing' and 'quality' hypotheses yield unique predictions, which are schematically illustrated in Fig. 1. The comparison between control pairs and experimentally delayed pairs yields information on the effect of timing; i.e., if experimental birds have lower reproductive success than control birds, then there is evidence of a causal relationship between timing and success of reproduction. The comparison between experimentally delayed pairs and the natural seasonal trend yields information on the effect of quality. If experimental birds have higher reproductive success than control birds breeding at the same time, then we would conclude that there is a difference in quality between early and late breeders, which contributes to the seasonal decline in reproductive success. This natural seasonal trend in reproductive success is calculated using all unmanipulated clutches. Intermediate results are also possible because the two hypotheses are not mutually exclusive.

Timing of breeding cannot be manipulated without causing some form of experimental bias. In our study, delayed females had to lay an extra clutch, which could have affected their quality, e.g., their nutrient reserves. Although we acknowledge that such biases may exist, we found no evidence of negative effects associated with laying a replacement clutch (see Discussion).

Preliminary conclusions, based on the first year of this study only, have already been published (Verhulst and Tinbergen 1991). In that paper, we concluded that the seasonal decline in pre-fledging reproductive success was caused by differences in quality between early and late breeders, whereas the seasonal decline in post-fledging reproductive success was causally related to timing of breeding. In this paper, we analyze the effects of the experiment on pre-fledging reproductive success, post-fledging reproductive success, and future reproductive output of the parents during all four years of the study. We then assess the relative importance of timing and quality in the seasonal decline in reproductive success, by integrating the experimental effects on different fitness components into one measure of fitness (lifetime reproductive success).