SUMMARY: Lengths of cnidae, sampled from fresh tissue (tentacles, acontia or column) from the sea anemones Metridium senile (Linnaeus), Cereus pedunculatus (Pennant), Sagartia elegans (Dalyell), Sagartia troglodytes (Price), Anthopleura thallia (Gosse) and Urticina eques (Gosse), were measured. The mean, mode, median, standard deviation, variance, coefficient of variation, minimum, maximum and range of each sample were calculated. Lengths of nematocysts (basitrichs and microbasic p-mastigophores) and spirocysts had normal (Gaussian) frequency distributions, and there was no important within-sample variability. A between-sample (same specimen) difference was detected between mean lengths of microbasic p-mastigophores from acontia in Sagartia elegans. Intraspecific (between-specimen) differences amongst mean lengths of basitrichs from column ectoderm were detected amongst four specimens of Urticina eques, and a difference occurred between the mean lengths of microbasic p-mastigophores from the acontia of two specimens of Sagartia elegans. However, for each of the species U. eques and S. elegans, there was no direct correlation between the sizes of the anemones examined and the mean lengths of their nematocysts. Overall, the results confirm that cnida size, when taken in isolation, is not a reliable taxonomic character for sea anemones. For other purposes, mean cnida measurements may be useful and a rapid sample range test is suggested for analysing differences between them. Methods are also presented for estimating means and standard deviations from data in publications that provide only ranges of cnida sizes and the number of cnidae measured. The present results and conclusions supplement those of Williams (1996, Scientia Marina, 60: 339-351).