Meeting Abstract

45.1  Saturday, Jan. 5  Mid-winter temperatures, not Spring temperatures, predict breeding phenology and fecundity in the European starling WILLIAMS, T.D.; Simon Fraser Uni. tdwillia@sfu.ca

The current working model for control of timing of breeding suggests that day length provides reliable ‘initial predictive information’ for general seasonal breeding but that ‘supplemental factors’, especially ambient temperature (Ta), fine-tune the actual timing of egg-laying. Furthermore, theory suggests that the greater the time-lag between an organism’s perception of a cue and the fitness consequences of their response, the less informative cues are likely to be. To date, most studies of timing of breeding in temperate avian species are fully consistent with this model: the time period where Ta is most highly correlated with onset of egg-laying is relatively close (< 1 month) to the population mean egg-laying date. Here I use an 11-year dataset on European starlings (Sturnus vulgaris) to show that mid-winter Ta, not Spring Ta, strongly predicts both breeding phenology and fecundity. Mean laying date was 10 April (range 5-13 April), and 80% of all nests were initiated over 2-8 days within-years. Despite this high degree of synchrony clutch size decreased significantly with date in most years. Sliding window analysis showed that a time period of 22 days from the 9 - 31 January provided the highest correlation between Ta and laying date (r = −0.87). Analysis of average monthly Ta data and mean laying date confirmed that laying date was independent of temperatures in February, March or April. However, January Ta was highly negatively correlated both with laying date and clutch size, i.e. when mean January Ta was colder starlings laid later, and laid larger clutches. Despite high synchrony of breeding starlings are exquisitely attuned to “date’, and they appear to use supplemental environmental cues, but in a very different way than predicted by our current models.