The EU-project GABRIEL (FP6, LSH-2004-1.2.5-1) focuses on genetic and environmental causes for the development of asthma and allergies in childhood. The „study of the origin of asthma and allergies within children in rural regions of Germany, Switzerland, Austria and Poland” deals with the effects of microbiological load, rural surroundings and farm life on the origin of the above mentioned illnesses. Therefore, the children were divided into 3 strata: (1) children who live on a farm, (2) children who have regular contact to farms, but do not live there, and (3) children who have no contact to farming environment at all. These 3 strata were subdivided into additional groups: (i) children who have asthma, (ii) children who are atopic, and (iii) children who have no asthma and are not atopic. 895 children out of the above mentioned groups were selected for further studies on the preventive effect of farm milk consumption against asthma and allergy development which has been observed during former studies. Therefore, milk samples were taken at the households by fieldworkers, and were analysed for different parameters. The aim of this master thesis was the determination of the microbiological composition of 222 selected milk samples. Thus, the following groups of microorganisms have been analysed: total viable count, Enterobacteriaceae, Lactic Acid Bacteria, Micrococci/Staphylococci, psychrotrophic count, Pseudomonas, Bacilli/Endospores and Yeasts & Moulds. The milk samples have been dived into 5 groups: (1) raw farm milk, (2) farm milk warmed up below 80C, (3) farm milk heated above 80C, (4) pasteurized milk, and (5) UHT-milk. The result showed that all groups had the highest geometrical average in the raw and warmed up farm milk samples. The geometric mean of the total viable count in the raw farm milk samples is 11790,9, in the warmed up samples one 5992,9, whereas in the heated up milk samples the averages are considerably lower (farm milk heated: 176,4; pasteurized milk : 306,9; UHT-milk: 34,4). Also the other bacterial counts have, in terms of their geometrical average, a similar distribution. The highest values could be detected in the raw farm milk and in the warmed up one. The only difference is the level of the geometric means, as they are much lower than the total viable count.