BALDI Ricardo
congresos y reuniones científicas
Spatial distribution of mara (Dolichotis patagonum) warrens tested by refined nearest neighbour analysis.
Mendoza, Argentina
Congreso; 10th International Mammalogical Congress; 2009
The mara (Dolichotis patagonum) is a large cavid rodent endemic of Argentina; it is monogamous but breeds communally in settlements which can congregate dozens of individuals. Although previous studies have described the settlements as clusters of warrens, such a pattern has not been tested so far by means of any specific quantitative methodology. With the objective of characterizing the spatial distribution of mara warrens, we interrogated a data base of warrens’ GPS positions collected at four 2000-ha areas within Península Valdés (Argentinean Patagonia). The study sites represent two contrasting landscapes (shrub-grass mosaic and grassland) which could influence the distribution of warrens. The warrens were localized scanning the areas walking along 20 transects separated 200m. The data base included 88 records (14, 17, 18 and 39 data points for the 4 different areas). We tested for departures from a complete spatial randomness (CSR) hypothesis by applying a refined nearest neighbour analysis using the R software. We compared the average distance from a warren to its nearest neighbour with the expected distribution of this distance calculated from 10000 realizations of CSR using the same plot size and shape. The observed average distance was 390 and 387.82m in the shrub-grass areas and 385.16 and 248.22m in the grassland areas. We found that the distance between warrens was significantly shorter (P < 0.01) than expected under a CSR hypothesis for all areas, which is indicative of underdispersion thus supporting previous descriptions of the warrens’ spatial pattern. This result suggests that social processes may have a stronger influence than habitat selection in defining the spatial distribution of mara warrens. Further studies will be needed in order to identify 1) the specific nature of the departure from a CSR pattern, and 2) the processes associated with the non-random patterns and their possible relationships with habitat variables.