Several infectious diseases (Brucellosis, Enzootic Bovine Leucosis, BHV1) are controlled by Bulk-Milk-Serology (BMS). BMS has proved to be highly accurate and effective thus it could be a valuable tool for the control of Mycobacterium avium subspecies paratuberculosis (MAP) infections, too. In the present study a registered ELISA (Svanovir, Uppsala) was analysed for BMS.Totally 2,677 pairs of blood and individual milk samples were collected from 118 herds. Aliquots of milk samples per herd were pooled (n = 104 pools). Sera were tested in three ELISAs (Svanovir, Uppsala; Idexx, Wörrstadt; Pourquier, Montpellier). Individual milk-samples and pooled milk samples were tested in Svanovir-ELISA.Individual milk testing was less sensitive as compared to serum testing in the same test. However, 83% to 92% of animals which sera were positive in absorbed tests (Idexx and/or Pourquier) were recorded positive by milk testing.Testing pooled milk only two herds were negative (OD% <5%), 31 and 69 herds were classified as low (5%<OD%<15%) and high (OD%>15%) prevalent, respectively. Alternatively optical densities (OD%) were grouped as follows: <10%, 11-20%, 21-30% and >30%. For each group the median, first and third quartile of herd prevalence as determined by serum testing with Svanovir-ELISA was calculated. Although the median of the serological herd prevalence was positively correlated with pooled milk OD%, first and third quartiles of herd seroprevalences overlapped significantly. Herds were grouped according to serology and bacterial culture in high risk and low risk herds. BMT was not able to distinguish between high and low risk herds.Currently BMS is not recommended for diagnosis of MAP, this does not necessarily exclude BMS for epidemiological research.