Différences de fréquences d'occurrence entre espèces de micromycètes parasitant une même espèce de plante-hôte ; procédure d'appréciation statistique

Abstract

The differences of prevalence among species of micro-fungi that parasitize a same host-plant species ; how to evaluate their statistical significance, using a newly designed test. Beyond simply recording the list of parasitic microfungi that occur within a given region -which is a first necessity -some quantitative aspects regarding the distribution of species may also be of additional interest. Among these aspects, the differences between the occurring frequencies of different species of microfungi that parasitize a same host-plant species feature special interest. The subject, however, raises a specific issue regarding the testing for statistical significance of the differences between the recorded occurrence frequencies, since the common practice, while sampling parasitic micro-fungi, involves the census of fungi occurring on host but almost never the complementary census of the absences of the same fungi. Accordingly, enumerative census are available only and not the corresponding frequencies. As knowing those frequencies is mandatory for common statistical tests to be performed, some special adaptation of those tests has to be considered beforehand. Accordingly, a modified x2 test has then been designed (see the Appendice in English) to assess appropriately the statistical significance of the gaps between the frequencies of occurrence of different species of fungi parasitizing a same host-species. This relevant methodology is then applied to the data collected during a survey of the distributions of Erysiphales, Peronosporales, Ustilaginales and Uredinales in Southern Burgundy (mainly Saône-et-Loire department). The most obvious (statistically significant) differences between the levels of occurrence of different micro-fungi species parasitic on a same given host-plant species are thus disclosed and a particular attention is devoted to the comparison between the respective frequencies of autoecious and heteroecious Rust-species that parasitize a same host-plant species. Autoecious species are generally considered more evolved than heteroecious ones and, accordingly, are logically expected to take benefit from the easier completion of their simpler, one-host life-cycle. Surprisingly, this expected advantage does not reveal statistically significant when occurrence frequencies are compared between heteroecious and autoecious species occurring on a same host, at least on the basis of our recorded data. This trend, however, needs to be substantiate by further investigations based on additionnai similar works in the same and several other regions.