Data were analysed, where possible, using ANOVA. Prior to using ANOVA,
Cochran’s test ŽWiner, 1971. was used to test for heterogeneity of variance. Multiple
comparisons of levels within significant factors were made using Student Newman
Keuls ŽSNK. tests.
For limpet growth rates, data from each individual patch or plot were used to
construct a Ford–Walford plot of initial limpet length against the growth increment over
1 year. This enabled an estimate to be made of the growth increment for a standard
limpet of initial length 35 mm for each patch or plot. These data were then analysed
using ANOVA. In one patch on each sheltered shore, there were insufficient limpets
Ž-10. at the end of the experimental period to accurately estimate growth. Thus, the
number of replicate plots on the exposed shores was randomly reduced to five to match
the five replicates on the sheltered shores.