A demographic model for an open population with space-limited recruitment has been previously proposed and tested using the barnacle Balanus glandula on a single shore in California. The model makes a number of qualitative predictions about population structure. At small scales, areas of high settlement should have on average less free space and fewer cohorts. At a larger unspecified scale (> 34.6 cm2) cohort structure should be more variable in areas of high settlement. We attempted to test these qualitative predictions on a number of different spatial scales using the barnacle Semibalanus balanoides. A recruitment gradient was defined for S. balanoides around the coast of Anglesey, Wales, with the north having lower recruitment than the south regardless of year. Given this recruitment gradient, a nested ANOVA was used to examine the qualitative predictions of the demographic model at different scales. The qualitative predictions of the demographic model were not generally supported by an analysis of data from Anglesey. Free space and small-scale population variance did not vary between areas of different recruitment density. Barnacle size variation was greater on high recruitment shores at an intermediate scale (100 m) as predicted by the model. However, this difference was not statistically significant. Even at the scale of individual 5 x 5 cm photographs there was no evidence of a positive relationship between barnacle size variance and free space. Possible reasons for the mismatch of the demographic model predictions and empirical observations are discussed. There may be insufficient spatial sampling, insufficient recruitment difference, large-scale variation in mortality and growth rates, or the model assumptions may be incorrect. One important difference between the 2 levels of recruitment is that low recruitment sites had a constant patch size below 25 cm2, whereas the high recruitment sites did not. If patch scale is variable, the predictions of the demographic model will be irrelevant as there is no single scale at which they can be observed. Use of the nested ANOVA technique with logarithmic spatial scales also allows hypotheses to be made about the important spatial scales for future demographic studies using manipulative experiments. For population variance, free space and opercular length the important scales were small (< 1 m), meso- (200 m), and large scales (>10 km).
- Nested ANOVA
- Patch and spatial scales
- Qualitative model predictions
- Semibalanus balanoides