The thermal inactivation curve for Mycobacterium paratuberculosis in milk at holder pasteurization temperature (63.5°C) has been shown to exhibit significant "tailing", resulting in survival of low numbers of the organism for extended periods. Results of investigations carried out to determine an explanation for these unusual inactivation kinetics suggested that "tailing" was due to some intrinsic characteristic of M. paratuberculosis, and not due to extrinsic factors such as the method of heating or composition of the suspending medium. Clumping of cells was thought to be the most likely explanation for "tailing". Therefore, a novel staining procedure was developed in order to monitor clumping and viability of M. paratuberculosis cells in milk of phosphate buffered salne during heating at 63.5°C. This consisted of 2-(p-iodophenyl)-3-(p-nitrophenyl)-5-phenyl tetrazolium chloride (INT) reduction as an indicator of viability followed by auramine O acid-fast staining to visualize all acid-fast cells present. This double staining procedure permitted simultaneous assessment of the extent of cell clumping, and also the proportion and location of viable M. paratuberculosis cells in relation to dead cells. It was observed that M. paratuberculosis cells that occurred singly were inactivated more readily during heating than cells present in clumps. It appeared that M. paratuberculosis cells within clumps were gaining extra protection from the lethal effects of heat from other cells surrounding them. At heating times corresponding to the "tail" region of the thermal inactivation curve, any viable cells observed were always located within clumps. It has been suggested that the "tail" will last until only one viable cell per clump remains. If it is assumed that each clump of cells, rather than each individual cell, produced a colony Herrold's medium then the above observations would explain why no apparent reduction in numbers was observed in the "tail" region of the thermal inactivation curve.