The conception of ecomorphes as ecological groups of living organisms has been developed by A. L. Belgard (1950) applicable for species of the highest plants of the southeast of a steppe zone of Ukraine. Conceptually close system of vital forms-biomorphs of animals and plants has been created by Mikhail Pavlovich Akimov (Akimov, 1955). A key task of an ecomorphic approach is the ecological analysis of ecosystems structure. The ecomorphic approach has been applied to various groups of animals: entomological fauna of wood plants (Apostolov, 1981), complexes of land arthropods (Barsov, 1996), communities of birds (Ponomarenko, 2002), soil mesofauna (Zhukov, 1996), coleoptera communities agrocoenosis (Sumarokov, 2007). To identify animal species as ecomorphes the expert approach was used: the expert in taxonomy group relying on the experience and knowledge of object identifies it ecomorphes. The algorithm of ecomorphes allocation of soil animals has been offered by Zhukov (Zhukov, 2009). With some changes this algorithm has been applied to allocation of spider species of the Dnepropetrovsk region (Prokopenko et al., 2011). A lack of the specified algorithm is that it yields satisfactory results only for abundant and frequent species in regional fauna. The shortcoming reason – it relies on parametrical statistics for which compliance of experimental data to the normal law of distribution is essentially important that actually can be established only for limited number of species. As result, for a number rather rare in regional fauna of species of spiders incorrect conclusions have been drawn on their ecological status which is known on data from spiders of regions more studied from the point of view of fauna. In our work nonparametric procedure of a multidimensional scaling which is tolerant to a statistical property of distribution of an abundance of types has been taken for a basis of ecomorphic classification of herpetobiont spiders. It including has allowed to integrate the data collected by various authors in a wide time and spatial span for the general analysis. Faunistic collection also differed and by a technique: the sizes of the traps, fixing liquid, number of traps, an exposition time. It is necessary to consider these circumstances objective as ecological classification of regional fauna should be based on considerable on coverage in time and space a material which cannot be collected on completely uniform procedure. The multidimensional scaling represents adaptive ordination procedure which assumes a choice of the final decision proceeding, first of all, ecological criteria, instead of especially mathematical. Such adaptability is reached by comparison of ordination decisions with markers of an ecological situation which are received at the biogeocoenosis description of places of sampling. These descriptions are presented in terms of typology of biogeocoenosis of Belgard (1950, 1971): their coenotic status, and also assessment of a mode of a fertility and humidity. The key material is received within the Prisamarsky biospheric station of the Dnepropetrovsk national university where within a monitoring profile reference types of wood biogeocenoses of a steppe zone of Ukraine are presented. Primary data of ecological-faunistic researches are represented in the form of matrixes (tables) where columns are presented by a species, and lines – sampling points. Sampling points may be ecologically processed on the basis of biogeocoenosis descriptions. If to find nature of compliance between sampling points and species, it is possible to make interpretation of the ecological status of species, i.e. to reveal key ecological groups (ecomorphes) and to establish belonging of species to them. The multidimensional scaling allows to estimate within one metric space an arrangement as species, and sampling points. Co-ordinates of sampling points in dimention of a multidimensional scaling are used as predictor of ecological characteristics of the environment in these points. Applying the obtained regression models it is possible to estimates of optimum conditions for species if in these models to use as predictor co-ordinates of species in those dimention of a multidimensional scaling. Species which are characterised by similar optimum values of ecological factors form ecological groups, or ecomorphes. Respectively for spiders we allocate coenomorphes, hygromorphes, trophocoenomorphes.