Seed rain mediated by frugivores is influenced by (1) the seed-deposition distances following fruit ingestion, (2) the disperser activity, as determined by its behaviour and habitat preferences, and (3) the structure of the habitat within the landscape. microhabitat (<100 m2). What are the potential effects of lizard habitat preferences on the seed rain? Do habitat preferences operating at different spatial scales show comparable effects on the seed rain? We addressed these questions using scenarios based on lizard habitat preferences, to simulate the resulting seed rain. It is important to note that the purpose of this habitat model is purely heuristic, instead of predictive - i.e., it is exclusively aimed at exploring the consequences of disperser behaviour patterns for seed rain, not at predicting their specific shape. Materials and Methods Study system (hereafter) is a dioecious evergreen shrub (up to 4 m in height) inhabiting the (semi)arid sclerophyllous shrublands of the Western Mediterranean basin and the Macaronesian region. In masting years, female plants can produce more than 10,000 fruits, which are available from June to September (unpubl. data). Arillated fruits (fruits hereafter) have red or yellow arils and bear only one seed (length: 4.870.11 mm; width: 2.100.06 mm; weight: 14.310.64 mg; n?=?233). 152121-30-7 supplier They are consumed by birds  and, in the Balearic Islands, by lizards . Owing to the low number of resident passerines during summer, Balearic lizards represent the almost-exclusive disperser of Ephedra seeds in some islets of the Balearic Islands (such as Dragonera Islet; unpubl. CRYAA data). Ephedra plants reproduce in episodic events of massive fruit production (i.e., masting years; pers. obs.) and recruits are extremely rare in natural populations (pers. obs.), which suggests that environmental conditions strongly limit its regeneration (e.g.,  for another Ephedra species). The Balearic lizard, (endemic to Eivissa and Formentera). Both species are important pollinators and seed dispersers of the native flora . Study site C description of habitat structure Field work took place at Dragonera, 152121-30-7 supplier a small (c. 300 ha) islet situated 1,300 m offshore Mallorca Island (Fig. S1), during an masting year (i.e., with high fruit availability). In masting years, fruits are a predominant component of the diet of Balearic lizards at this population (e.g., during a pilot study, c. 80% of lizard defecations had pulp and seeds, and this proportion was comparable for both males and females: 21?=?1.02, fruit production and with the period of maximum activity of Balearic lizards (JuneCAugust; pers. obs.). The study site was located in a stony, 152121-30-7 supplier steep slope facing southeast and located between the shoreline and 80 m a.s.l. The landscape is dominated by small soil pockets and rock outcrops that provide abundant refuges for lizards, interspersed with patches of open sclerophyllous shrubland dominated by and is the only frugivorous reptile inhabiting Dragonera. The spatial analysis of habitat characteristics and lizard preferences was based on an observation unit (grain) of 1212 m. The size of these units was chosen to match the spatial resolution of our lizard telemetry locations, which showed a median error of 11.60.3 m (Fig. S2; see 152121-30-7 supplier below). Hence, the study area was subdivided in a lattice of 110 grid-cells of 144 m2 (Fig. S1) and environmental variables were extracted, for each grid-cell, from a digital elevation model (DEM) and a habitat map. We obtained the DEM from 11000 cartography, while the habitat map was derived from the supervised classification of an aerial ortho-photograph, with four habitat types (rock, bare soil, sclerophillous shrub and shrub). For each grid-cell, we estimated two topographic variables (height and slope, derived from the DEM), two habitat variables (the proportions of rock and shrub; Fig. S1) and four variables describing landscape fragmentation (Text S1): (a) number of shrub patches, (b) mean size of shrub patches, (c) mean shape of shrub patches (patch perimeter divided by the perimeter of a circle of identical area) and (d) mean distance to the closest shrub patch. We estimated these fragmentation measures using the software FRAGSTATS (McGarigal 2002). Retention time and germination of seeds During the.