Morphological vs. molecular delineation of taxa across montane regions in Europe: the case study of Gammarus balcanicus Schaferna, 1922 (Crustacea: Amphipoda)

Mountainous areas are characterized by substantial biodiversity and endemicity due to their complex geological history and habitat fragmentation. Hence, it can be assumed that particularly high species richness can be found in organisms with limited dispersal capabilities that inhabit mountain streams. A number of scientific papers focus on molecular phylogeography or traditional taxonomy of species or species groups inhabiting such habitats. However, there is a lack of studies that integrate morphological and molecular data to identify and delineate cryptic species. For practical reasons, uncovering cryptic diversity is crucial in taxa used in biomonitoring. Distinct species, hard to separate based on morphology only, may have different tolerance ranges towards a variety of factors. Thus, our goal is to combine the two approaches to reveal potential patterns of diversification within a species widely distributed across European mountains: the amphipod crustacean Gammarus balcanicus. The data were obtained from 13 populations spread across the range of the species. Individuals were initially ascribed to G. balcanicus based on conventional fauna key morpho-anatomical diagnostic features and were further analysed for 23 additional features to explore any putative diversification. Morphometric data were analysed with use of the multiple correspondence analysis and ANOVA. Molecular distances were calculated for 551-bp-long COI sequences. Test for isolation by distance was performed for both morphological and molecular data. The morphometric studies showed that some of the analysed features differed significantly between populations, although there was only a weak correlation between the morphological divergence and the between-population geographical distances. Moreover, high morphological diversity was present within sites. A set of 42 COI haplotypes was identified among the 135 individuals sequenced. No haplotype was shared among populations. The molecular p-distances within the nine localities presenting more than one haplotype were either almost null (ca. <0.003 for 7 localities) or relatively low (ca. 0.01-0.02 for 2 localities). In opposite, the molecular p-distances between localities were mostly at a high level (94% of pairwise comparisons being >0.14), similar as between other well-defined species of the genus Gammarus. Surprisingly, G. balcanicus appears to be polyphyletic based on topology of the neighbour-joining tree. The level of genetic distance between localities was not correlated with their geographical proximity. Globally, combining spatial patterns of morphological versus molecular divergence indicates a high level of cryptic diversity within a species conventionally defined based upon fauna key morphological features. In this context, the name G. balcanicus should be applied only to the population from locus typicus, while the other populations represent a number of putative distinct species. We may expect that such phenomenon would apply also to other animal taxa with conserved morphology, which are widespread over different mountain ranges in Europe.