A novel and more rapid culture protocol for isolation of Mycobacterium avium subsp. paratuberculosis (MAP) from bovine milk was developed and compared with the conventional agar slant method. Bulk tank milk from a herd repeatedly tested negative for Johne's disease was spiked with different concentration of MAP, from 10⁶ to 10 CFU/ml. Five milliliters of milk were incubated for 3 hours with a solution containing CB18, a zwiterrionic detergent, known to concentrate MAP. Samples were centrifuged and decontaminated with an enzyme solution and plated into Middlebrook 7H10 agar containing mycobactin J and PANTA. Plates were screened weekly under a microscope (at 40X) and the time of detection was recorded. The assay was repeated 8 times and compared with results obtained from 50ml of milk spiked with the same MAP concentrations but decontaminated with 0.75% HPC solution for 5 hours and inoculated into Herrold's egg yolk medium (HEYM) and conventionally screened for bacterial growth. The presence of MAP in spiked milk samples could be detected between 14 and 45 days (n=42, mean=22.7) using the CB18 and microscopic screening method and between 21 and 63 days (n=41, mean=31) using the HEYM conventional culture method. The time to detection also differed with the MAP concentration with the samples containing higher concentrations detected earlier than the samples with lower concentrations in both methods. The modified culture method has reduced the time to detect MAP in raw milk, therefore representing a potential tool for the improvement of MAP control programs.