A time-series bacteriological analysis has been carried out on milk collected on farms from 1997 to 2001 by a plant producing raw milk soft cheese, with the purpose of assessing the time course of the presence/absence of Listeria monocytogenes. A standard data collection procedure was used, in which farms were tested on a monthly or biweekly basis and 2-3 days after the detection of milk tanker contamination. This procedure yielded low figures for contamination frequencies. The average value and the median of the monthly prevalence of farms detected positive for L. monocytogenes were 2.4 and 0%, respectively. A seasonal effect (with peaks in winter) was observed. Between 1997 and 2001, there was no significant decrease of contamination rates, in spite of the efforts on the contaminated farms. Over the last year of the study (from March 2000 to February 2001), a new data collection procedure was implemented that allowed much better detection of sporadic occurrences. Milk samples were collected from the bulk tank of each participating farm just before pick-up, then stored and subsequently analysed whenever the milk tanker was found contaminated. The average value and the median of the monthly prevalence of positive farms were found equal to 7.7 and 0%, respectively (for a mean prevalence of L. monocytogenes in the milk tanker of 3.2%). These results confirm that farm milk contamination is, most often, a sporadic event In addition to this prevalence study, contamination levels were quantified by enumerating L. monocytogenes using direct plating of small volumes of farm milk previously tested positive. Most often, these levels were extremely low. A simple simulation model shows that, when milk tankers were found positive, contamination levels in the corresponding bulk-tank milk are themselves very low (typically, below 3 L. monocytogenes per millilitre with most probable concentration 0.1 Colony Forming Unit (CFU)/ml and median ranging from 5.10(-2) to 0.1 CFU/ml). Such low levels are very likely to be due to environmental contamination.