Problems in sampling desert shrub populations: a comment and a correction