Biometry of sea anemone and corallimorpharian cnidae: statistical distribution and suitable tools for analysis

Abstract

Many studies have dealt with data on the sizes of cnidae within different groups of anthozoans, such as Actiniaria, Corallimorpharia, and Zoantharia. Statistical treatments of these data have been variable, according to the evaluation of the normality and hence the use of parametric or nonparametric tests. The normality of cnidocyst size data was assumed or proved by some authors who used parametric tests to make comparisons. Other authors carried out nonparametric tests, or even proposed alternative analytical methods, such as the use of generalized linear models. Despite controversy about the statistical distribution of cnidae sizes, there has never been an attempt to study the normality of cnidocyst size data involving a significant volume of samples, using several specimens from various different species and using the same statistical approach. The objective of this paper is to evaluate statistical adjustment to a normal distribution of cnidocyst length from four sea anemone and one corallimorpharian species. The cnidoms of all species are detailed and the hypothesis of no intraspecific variation of cnidae sizes tested as a study case. Normality was accepted in 36.42 % (SD = 17.91) of all data sets of all cnidocyst types analyzed from all studied species, while for the rest it was rejected. The evidence suggests that both normal and non-normal data sets are possible, although non-normality is slightly more frequent. Intraspecific variation of cnidocyst sizes is shown in 96.82 % of the analyzed data sets. This paper provides a simple and detailed methodology to perform comparisons of cnidae size data.