Agricultural fertiliser (farmyard manure), covers partially the nutrient elements demand of the vegetation cover. However, its application is difficult as regards the unknown concentration of fertilising elements contained and their biological availability for the plants, the important volumetric masses to manage and the general uneven land application. An inaccurate management of farmyard manure may cause groundwater pollutions with sometimes serious consequences, especially in the case of karst aquifers. The objective of this research was to specify various aspects of the farmyard manure cycle. It consists of three major parts: a general study of the difficulties related to the farmyard manure management in mountainous regions with the aim to provide practical recommendations to reduce the risk of groundwater contaminations, the evaluation of an alternative method of pig farming in liberty the study of the impacts of agricultural activities on the surface- and groundwater quality. Based on the analysed farmyard manure managements, the following practical recommendations have been made: split up the land application, but concentrate the input of nutrient elements in spring, during the period of plant growth. For sanitary reasons, a delay of at least 1 month has to be respected between the last land application and the beginning of pasture; collect in adequate installations all agricultural effluents which potentially cause water pollution, except if they are spread out immediately. minimise the production of fertiliser by the following two possibilities: infiltrate low contaminated waste water (from the roof, spring, eventually from the farm) reduce the time the animals stay in the stable. An alternative method of pig farming in liberty has been tested in the field. 25 pigs have been raised on a field of ca. 2 ha, representing a density of 0.71 LiveStock Unit/ha/y, respectively 0.18 LSU/ha/m during 4 months. Amongst others, the study has considered the impacts of pig pasture on the environment (vegetation cover, soil) and the animals health. The most important impact is in fact the destruction of the vegetation cover, with its direct repercussions on the botanical diversity and the concentration of nutrients in the soil. The experiment has made evident the importance to keep the vegetation cover intact, in order to adsorb the nutrient elements in the manure. An advantageous method to reduce the impacts on the vegetation cover is to equip the pigs with nose-rings. It is furthermore necessary to altemate the pasture parks in order to enable the vegetation to recover (2 to 3 different parks). The important impacts observed around the pig feeding through can be avoided by two ways: first, by changing the throughs placement during the pasture period or second, by the construction of a concrete place and the collection of the manure in the slurry pit. For the pasture of pigs in liberty one has to chose grasslands with thick soils of good quality but avoid oligotrophic ones. Cattle sheds and feeding through have to be placed in a manner to favour a maximum dispersion of the animals. Two hydrogeological observation networks, composed of 6 springs, 7 rivers and one lake, have been installed in order to study the potential contamination by pathogenic bacteria (Listeria, Campylobacter, Yersinia, Salmonella). No salmonella have been found, but several other species have been identified in the water samples of the different study sites: Campylobacter jejuni, C. coli, C. fetus Listeria monocytogenes (sv 1 /2a, 1/2b, 4b), L. ivanovii ivanovii (sv 5) Yersinia intermedia, Y. frederiksenii, Y. enterocolitica, Y. kristensenii, Y. pseudotuberculosis. The most frequently identified species in the studied sites is C. jejuni. It has been isolated in 75% of the sites, with an average percentage of isolation of 11% of' the spring samples, 22% of the river and 56% of the lake samples. L. monocytogenes has been isolated several times in 9 of the 14 sites. It is less abundant than C. jejuni, except in the analysed springs. The presence of C. jejuni and L. monocytogenes 4b has shown to be closely related to the summer pasturing period. Other germs, like L. monocytogenes 112a and yersinias show a uniform annual distribution. The presence of bacteria in surface and groundwater is controlled by several factors, of which the presence of animals, precipitation and temperature have been identified to be the most conclusive ones. In case of source contamination, the factors controlling the presence or absence of pathogen bacteria are different for the case of surface and groundwater: In the first case, the influences of the hydroclimatological conditions are dominant, whereas in the second case it is controlled mainly by the local hydrogeological characteristics. In case of aquifers with interstitial porosity (Clos Ister, Confluent), the presence of pasturing anirnals is however the controlling factor. The persistence of a contamination is controlled by the storage capacity of the catchment basin and by the relation of seep water to the aquifer total volume. In the Brassus spring case, the presence of animals on the catchment area can not explain by itself the bacteriological findings. The presence of the found pathogen bacteria can however be explained by their transport by the run-off rain. The presence of pathogen germs and bacterial pollution indicators has been analysed in a comparative study. Result: 2.3% of the samples contained pathogen germs without containing bacteriological contamination indicator. The species found in the absence of bacterial pollution indicators, are rare species, showing a monotonous annual occurrence. They have been found only in low concentration and the danger when drinking this contaminated water was very small. That means the bacterial pollution indicators choice presently used for water bacteriological quality evaluation is accurate. However, they should be tested with high flow levels, when bacteriological contaminations are particularly important.