“…A welcome addition to the various classes of parametric distributions on the simplexthe additive logistic normal (Aitchison and Shen, 1980;Aitchison, 1986, p.113), the multiplicative logistic normal (Aitchison, 1986, p.130), partitioned classes (Aitchison, 1986, p.132) and the Dirichlet-embracing generalisation *Aitchison, 1985*Aitchison, , 1986) -is the multivariate logistic skew normal based on the multivariate skew normal class on R D introduced by Azzalini and Dalle Valle (1996) and further developed by Azzalini and Capitanio (1999). This allows for skewness in the logratio transformed data and promises to allow more extensive study of methods which depend on distributional form.…”