[Habitat selection and metapopulation structure: a multi-year study of distribution of the Hodgson's pipit, Anthus Hodgsoni Richm. (Aves, Passeriformes)].

Numbers and distribution of the Hodgson's pipit (Anthus hodgson Richm.: Aves, Passeriformes) were studied in the Yenissei middle taiga region over an area of about 450 sq km. Distribution of breeding pairs was mapped on fixed study plots (up to 450 ha in total) annually during 15 years. Habitat properties were described and measured in detail on 53 4-ha homogeneous plots within the area. It is shown by means of multiple regression, an average bird abundance over these plots depends on five habitat features (R2 = 0.74) including development and accessibility of moss cover and absence of a potential competitor, the tree pipit (A. trivialis L.). Correlation with these factors ruterated every year, therefore the average bird abundance was used as an index of habitat favourability. Density deviations from the mean in years of high and low numbers appeared to be closely related to habitat favourability; in particular, the relationship can alter the sign depending on the spatial scale of population structures. Changes in abundance reversely related to favourability in adjacent habitats, according to the prediction of despotic distribution hypothesis (Fretwell, Lucas, 1970), strictly indicating dominance behaviour during selection of a breeding territory. The numbers in larger population groups occupying a patchy habitat complex changed synchronously and proportionally to their average habitat quality. This assumes another mechanism governing the distribution of individuals, requiring no local knowledge and no dominance relationships. Dynamics and distribution of individuals among population groups of different hierarchic ranks agree with investigations on establishing of individual site fixation in birds and allow splitting the process into four consecutive steps. 1. During the juvenile dispersal, birds spread around quite evenly, disregarding of habitat quality. This maintains entirety of the metapopulation and occupation of isolated habitat spots. 2. The juvenile dispersal ends with switching to a search for a nearest site containing potential breeding territories. Such a site gets imprinted as the site for next breeding. According to our data, it covers an area of a few square kilometers. 3. During the pre-breeding period, an individual chooses a territory regarding to the favourability, but dominance of old residents forces it to search around for an empty patch within the imprinted area. As a result, dominance relationships affect both breeding density in best habitats and availability of empty patches. 4. The next years, an individual keeps connection with the breeding site fixed due to site tenacity. Rising of the social status allows the individual to occupy a better territory in future. Thus, a population group proportionate to individually imprinted area, with a concentration of favourable patches in the central part, comprises an elementary structural and functional unit of a metapopulation. Its individual members share the same well-known imprinted area and the social structure in common. Its numbers are regulated by density-dependent dominance relationships. Individuals spread over such groups in proportion to their carrying capacity in density-independent manner. Groups with best habitats can be more profitable at population lows, and less densely populated ones can offer more profit at population tops. Despite this, lack of information restricts profitable movements between them. Hence each group offers the same average fitness to its members.