ABSTRACT
Knowledge of the feeding ecology of ungulates in arid biomes offers an interesting model for understanding the drought resistance of large desert-adapted herbivores, a crucial issue in the face of increasing desertification due to climate change. To assess the feeding ecology of the endangered Cuvier's gazelle (Gazella cuvieri) in the Sahara desert, we used a multi-method approach combining faecal samples, direct observations, and the recording of indirect signs of feeding. We hypothesised that browser behaviour is the best foraging strategy for species living in hyper-arid environments, mainly due to long periods without grazing opportunities. Complementarily, we explored the effects of the main environmental descriptors (rainfalls and NDVI) on feeding patterns and diet quality. We found that Cuvier's diets are based mainly on acacias (Vachellia tortilis, V. flava) and occasionally on the annual forb Anastatica hierochuntica. In total, eighteen species (five trees, nine shrubs, three herbs, and one grass) belonging to fifteen families were recorded. Our result confirmed the browsers' characteristic of this species, reaffirming its ability to settle in a hostile environment. Acacias stand out as key species consumed at the southernmost limit of their range; hence, future conservation plans and strategies should take this into account for the survival of Cuvier's gazelle in desert environments.
ABSTRACT
Stomata distribution is an example of biological patterning. Formal methods used to study stomata patterning are generally based on point-pattern analysis, which assumes that stomata are points and ignores the constraints imposed by size on the placement of neighbors. The inclusion of size in the analysis requires the use of a null model based on finite-size object geometry. In this study, we compare the results obtained by analyzing samples from several species using point and disc null models. The results show that depending on the null model used, there was a 20% reduction in the number of samples classified as uniform; these results suggest that stomata patterning is not as general as currently reported. Some samples changed drastically from being classified as uniform to being classified as clustered. In samples of Arabidopsis thaliana, only the disc model identified clustering at high densities of stomata. This reinforces the importance of selecting an appropriate null model to avoid incorrect inferences about underlying biological mechanisms. Based on the results gathered here, we encourage researchers to abandon point-pattern analysis when studying stomata patterning; more realistic conclusions can be drawn from finite-size object analysis.
