Molecular markers are increasingly used for species identification and new taxa description. However, rules to determine frontiers between populations and species are not clear depending on taxa considered. For mites, few studies deal with molecular diagnoses, making rules for associated decision difficult. The present study focuses on a species of the predatory mite family Phytoseiidae (Phytoseius finitimus), considered for biological control of mites and small insect pests in fruit orchards and vineyards in the Mediterranean basin. This paper aims to elucidate the causes of great molecular variations and questions the occurrence of cryptic species. Molecular (12S rRNA, CytB mtDNA, ITSS) and morphological analyses were performed on four populations collected in Corsica and Italy in crops (vine and kiwi) and in an uncultivated environment (Viburnum lantana). Different methods for identifying species have been used (tree approaches, distances and ABGD algorithms). A reference database of distances within and between Phytoseiidae species has been elaborated to inform the present question and to assist with further diagnosis within Acari. Mitochondrial DNA analyses show that specimens from V. lantana were well separated from the three other populations with high genetic distances, suggesting the existence of a cryptic species. Molecular ITSS analyses coupled with morphological features show however that the four populations seem to belong to the same species. The great mitochondrial polymorphism is discussed in regards to: (i) genetic distances reported for Phytoseiidae species and (ii) potential biological differences between populations (cultivated versus uncultivated areas). This study clearly emphasizes the necessity of integrative taxonomy approaches for diagnosis decisions. Furthermore, based on the polymorphism herein detected, maximal intraspecific distances are proposed (9, 23 and 2.8 % for 12S rRNA, CytB mtDNA and ITSS) for diagnosis decisions within Phytoseiidae. Further statistical analyses are however clearly required to determine statistical error for general and reliable decision making.

No abstract