Abortion in dairy cows in well-managed dairies is not common but differences have been reported probably due to variation in animal health, nutrition and management, as well as difficulties in observing the aborted material. A 38-month study of 507 large Danish dairy herds revealed 3354 late-term abortions and 224,419 calvings were recorded. During the study period, a total of 3717 submissions were made to the Danish Institute for Food and Veterinary Research (DFVF). A broad spectrum of abortive agents was isolated but none were found to be statistically associated with abortions. The number of abortions in a month on a dairy was significantly (P<0.001) associated with the number of cows in the third trimester but explained only 11% (R(adj)(2)=0.114) of the variability of the reported abortions. A total of 23 herds (4.5%) reported 531 abortions (15.8%). Although a marginally significant (P=0.11) risk of increased abortions was found to exist in bovine viral diarrhea virus- (BVDV) infected herds, it could be at least partially explained by additional calvings in those herds. Temporal correlation between inseminations and abortions was statistically significant (P<0.001) with the highest correlation (r=0.47-0.51) after lagging abortions on insemination by 6-8 months. No indication of spatial clustering was detected for either specific-abortogenic pathogens or high aborting dairies using either Cuzick-Edwards' (P>0.17) or spatial scan tests (P>0.23). Ederer-Myers-Mantel test was applied to 3 years of data on the highest aborting dairies and showed that July had nearly double the expected number of maximum monthly abortions (P<0.001). These findings provide further insight into the reported abortion pattern in Danish dairies and may facilitate planning future control programs.