• Jones, Malcolm Timothy
  University of Maine
  

I examined avian abundance and distribution patterns for New England using data from the U.S.G.S. Biological Resources Division's Breeding Bird Survey (BBS). I identified three distinct avifaunal regions within New England, which formed the sample units for all further analyses. For each of these regions I tested the predictions of Hanski's core-satellite hypothesis, Brown's ecological specialization hypothesis and Wright's sampling hypothesis to determine which best explained the observed patterns.

Mean abundance (birds/route) was correlated with spatial incidence (i.e., proportion of sites occupied) for all 26 years examined in each region. The frequency distribution of spatial incidence was extremely bimodal for all but three years in any region, as predicted by the core-satellite species hypothesis (i.e., most species occurred on either a few or most sites). Also, this is the first known study to document the occurrence of core-satellite switching, with three species switching from satellite to core species and two species switching from core to satellite species by the end of the study.

No evidence was found to support Brown's ecological specialization hypothesis, which predicts that ecological generalists should have higher densities than specialists. Rather, I found that specialists had higher local densities than generalists. I suggest that the failure of other studies to show a similar relationship is the result of a failure to control for the confounding effect of spatial incidence. In all regions, Wright's sampling model explained 56-62% of the variation in spatial incidence.

I also examined species' spatial and temporal dynamics using Taylor's Power Law (TPL). Approximately 50% of the species had statistically significant temporal or spatial exponents. Classification and regression tree analyses indicated no correlation between the spatial exponents and natural history correlates; only 15-25% of the mean deviance of the temporal exponents was explained by life history correlates. I showed that this index was probably extremely biased, especially with species that had an overall mean abundance less than two.