Excessive presence of wild ungulates can produce negative effects on herbaceous crops or woody species, and to face this problem, habitat improvements are often performed to recreate suitable environments for a given animal species and to attract animals far from cultivated crops. A common example of these interventions is represented by grassland restoration and to evaluate the real animal preferences on restored forage resources a proper trial was established in a hilly area of Tuscany (central Italy), inside the historical Park of Pratolino, near Florence. The trial compared six different forage species or mixtures sown in plots: vegetal material was represented by two pure stands (Onobrychis viciifolia and Medicago sativa) and four mixtures differing in number and kind of used species. Plots were utilised only by wild animals occurring in the area. Data collection consisted of botanical samples in each plot in different periods to obtain the percent presence of each species. At the same time, a visual estimation of animal intake on all occurring species was performed to obtain the browsing ratio of single species and overall defoliation rate for each species/mixture. Moreover, six camera traps were placed on the boundary of the experimental site to record videos of wild animals browsing in the area for identification of animals actually occurring on different plots and for comparison of these results with botanical data. Vegetation surveys permitted a proper evaluation of animals intake and of their feeding preferences. In general, sown species performed a major role in animal browsing, even if in some periods also a few native species (such as Plantago lanceolata or Cichorium intybus) were utilised in a strong way, depending on vegetation context and existing biomass. Camera traps results permitted the identification of browsing animal species (mainly represented by roe deer) and plots frequentation resulted to be highly related to animal intake found by botanical relevés.

There are many methods to estimate wild ungulate populations. One of these is represented by observations from vantage points usually employed in medium-low covered forest areas to estimate cervids. This method is subject to some limitations, such as the risk of double counting and the necessity to a high number of operators due to the small size of the observable areas. Such limitations could be reduced by integrating the "vantage points method" with Remote Piloted Aircraft System (R.P.A.S.) surveys. This study reports some experiences carried out mainly to check the suitability of R.P.A.S. use in the specific case of Roe deer (Capreolus capreolus) and Wild boar (Sus scrofa). The investigation was conducted in various areas with a high density of Roe deer and in the absence of disturbing elements. The first experimental phase concerned the definition of technical parameters useful to characterize the research system. For this purpose, the acoustic emission of the aircraft (dB) and shots of the underlying ground were recorded at intervals of 10 m up to 70 m above the ground. In the second phase, operating flights were carried out on an area of around 90,000 m 2 consisting of seven contiguous grazing areas separated by hedges and woody plants. The ability to fly over the natural visual barriers (land orography, trees, hedges, etc.) allowed to conduct a census over all the study area, despite its size and its characteristics. Only two operators can do all the flights in an average time of about 10-15 minutes with a R.P.A.S. speed of approximately 3 m•s-1. The analysis of the aerial pics taken at dusk from 50 to 60 meters has identified many Roe deer grazing in different areas and has verified the absence of any animal alarm reactions to the passage of the aircraft. As for Wild boar, animal species with more pronounced nocturnal activity, the tests were conducted using a thermal imaging camera installed on board the R.P.A.S.