van der Hut R.M.G. (1986) Habitat choice and temporal differentiation in reed passerines of a Dutch marsh. ARDEA 74 (2): 159-176

Habitat choice, territory size, territorial overlap, arrival pattern and breeding Phenology of reed passerines are studied in the Jagersveld (Fig. 1), a marsh c. 15 kms north of Amsterdam. The species studied are Locustella naevia (Ln), L. luscinioides (Ll), Acrocephalus scirpaceus (As), A. palustris (Ap), A. schoenobaenus (Asch) and Emberiza schoeniclus (Es). A census of breeding birds was carried out during 1982 and 1983, employing the mapping method. Territorial boundaries were determined by mapping territorial activities. Territory size and territorial overlap were measured using maps on a scale of 1: 1,000. Habitat structure variables examined are water level, vegetation density, half density height, vegetation height density of rigid vertical elements and percentage of tree cover. Discriminant analysis was employed to test discriminative power of these variables. The reed passerine assembly comprised 215-216 territories, constituting a density of 9.0-9.4 territories per ha (Table 1). Territory size differs significantly between several species, but variances are large (Table 2). Intraspecific differences in territory size between different males are probably related to differences in vegetation structure and consequently in the extent to which a territory is used as a feeding site. Seasonal variation could have been brought about by the varying level of intra- and interspecific aggression and by changing vegetation structure in the course of the season. Shifting of territories and change of territory size within a season probably reflect a shift in function of the territory, viz. the cease of advertising. Territorial overlap is extensive. The area occupied covers only 38-39% of the sum of the areas of all territories (Fig. 2). Interspecific overlap is large in the congeners As and Asch and between heterogeners, resulting in two overlap groups: Ln-Asch-Es and Lt-Asch-Es. Arrival pattern shows four groups: Es, Ln-Lt-Asch, As and Ap (Fig. 3). Differences in arrival time are related to migration strategy and correlate with time of availability of suitable nesting sites on the breeding grounds. These phenological differences are reduced in the latter half of the breeding season (Fig. 4). Ln, Lt, Asch and Es frequently raised second broods (Fig. 3). The species' habitats can be arranged according to a zonation from wet to dry considering water level, soil type and floristic composition (Table 4). Three types of vegetation density profile can be discerned (Fig. 5): low vegetations with dense ground layers occupied by Ln, tall vegetations with sparse ground vegetation occupied by As and Ap, and vegetations of medium height with dense ground layer occupied by Lt, Asch and Es. Discriminant analysis reveals density of rigid elements, vegetation height, and half density height as good separating habitat variables (Fig. 6, 7). Congeners appeared to be separated clearly, although obvious overlap remains between Acrocephalus species, and heterogeners overlapped extensively. The reality of this overlap is confirmed by territorial overlap figures. However, Locustella and Acrocephalus are known to differ in vertical habitat selection, Locustella species being walking ground dwellers and Acrocephalus species being clingers at higher vegetation levels. This could result in different foraging strategies also. The Locustella species, separated well by habitat selection, are quite similar in phenology. The Acrocephalus species differ partly both in habitat selection and phenology, but also show competition, as witnessed by frequent aggressive interactions. Es is able to coexist with the other species being a generalist in horizontal and vertical habitat selection and feeding strategy.