Aim: We studied compositional turnover in two ectoparasite taxa asking whether (a) the main driver of turnover (environmental factors, host composition, or distance) is the same in both taxa; (b) the turnover of rare and widespread fleas and mites is driven by different factors; and (c) the turnover of either fleas or mites or both respond to the same environmental gradients as the turnover of their hosts. Location: Northern and central Palaearctic. Time period: 1959–2004. Main taxa studied: Fleas, gamasid mites, mammals. Methods: We used data on fleas and mites collected from 30 and 20 regions, respectively, and applied a novel metric, zeta diversity, and a novel approach, multi-site generalized dissimilarity modelling (MS-GDM). Results: In fleas, variance explained by the models with host turnover as a predictor was higher than that of models with environment and distance or environment only. In mites, similar proportions of variance were explained by models with and without a host-associated predictor. Host turnover strongly affected the compositional turnover of fleas, whereas mite turnover was equally influenced by host turnover and dissimilarity in air temperature. When host turnover was removed from the models, temperature became the most important predictor of turnover, followed by precipitation (for both taxa) and distance (for fleas). The effects of host turnover and environment were stronger for turnover of rare than widespread species. Temperature was the most important predictor of the host turnover independently of whether distance was included in the model. Main conclusions: We conclude that the strength and manifestation of the effects of host compositional turnover as compared to environmental dissimilarity differ between ectoparasite taxa. Moreover, the application of MS-GDM allowed us to reveal patterns that were obscured or distorted in earlier studies.