Here's a worked example for comparing group averages with bootstrap confidence intervals and allowing for different subsample sizes by calling the strata argument within the bootstrap function.
The data is set up analogous to an before-after impact experiment conducted on plots across four different successional stages. Similarity was calculated for plot's composition before and after an impact. One hypothesis was that certain stages would show higher / smaller average similarities, that is, a higher / lower impact on composition. As plots within stages were situated within different subsites and the nr. of replicates was unbalanced, this had to be allowed for by use of the “strata” argument in the boot.ci call: