THE GENERALITY OF FIELD EXPERIMENTS: INTERACTIONS BETWEEN LOCAL AND BROAD-SCALE PROCESSES

Linking the results of localized field experiments to generalities about the role of specific processes is essential in ecology. Comparative studies conducted at multiple locations enable the general importance of processes to be assessed. However, spatial or temporal variation in the strength of local ecological relationships frequently makes it difficult to draw general conclusions, as increasing the extent of a study is likely to increase the physical and biological heterogeneity. To unravel the influence of differences in wave climate on local ecological interactions among adult and juvenile sandflat bivalves, an experiment was conducted at three sites in each of three harbors in the North Island of New Zealand. The selected sites covered a range of wind-wave exposures but were all mid-intertidal sandflats with macrobenthic communities dominated by bivalves. Four treat- ments were used: additions of 0, 15, 45, and 110 adults of the tellinid bivalve Macomona liliana. At each site a DOBIE wave gauge was used to provide time-series data on wave orbital speed at the seabed over the 35-d duration of the experiment. Significant experiment x location interactions indicated site-dependent variation in the strength and direction of treatment effects. However, multiple regression models based on site environmental char- acteristics were very successful in explaining differences between sites in the strength of experimental treatment effects (66-99% of the variance explained). We used the cube of the orbital wave velocity at the seabed (U3) as an index of energy dissipation by waves. Both the site average and maximum U3 were important variables explaining location- dependent treatment effects. Average U} increased the strength of the negative effects of experimental additions of adult Macomona on juvenile bivalves, presumably by increasing the opportunity for juvenile bivalves to be resuspended by small waves and transported away from areas with high adult densities. Maximum U3 decreased the strength of the experimental effects, probably by increasing the purely passive transport of juveniles with sediment bedload and thus obliterating patterns in the distribution of juvenile bivalves relative to adult Macomona. Hence, different aspects of the wave climate influenced adult- juvenile interactions in different ways. Ambient density of adult Macomona around the experimental sites was also an important factor. This multisite experiment confirmed the importance of local adult-juvenile interactions and suggested ways in which these inter- actions are influenced by local environmental characteristics. Where the influence of such broad-scale variables can be identified, linking broad-scale information to small-scale stud- ies that identify mechanisms can increase the generality of ecological experiments.