Milk urea content as an indicator of nutritional status may be a useful tool if major sources of variation are considered. Blood and milk samples were collected frequently during 16 to 19 h from four Holstein cows to study diurnal variation of urea content. Corn silage, alfalfa hay, and concentrates were fed. Rumen ammonia, VFA, and pH were measured in three of the cows. A clear serum urea peak, 70 to 85% higher than the lowest concentration, was observed in the higher yielding cows. The serum urea peak occurred 1.5 to 2.0 h after the rumen ammonia peak. Urea in milk equilibrated with serum with a time lag of 1 to 2 h when the rate of change in serum was .5 to 1.0 mM/h. At this rate, the average difference between serum and milk urea content was .8 mM. Urea in total milk tended to be more closely correlated to serum than samples from the gland cistern, but deviations were minor. Our results indicate a relatively rapid equilibration between blood serum and milk urea, also in the gland cistern. Equilibration may be explained by diffusion of urea along the mammary ducts and through the mucosa in the alveoli. If urea is to be used as an indicator of nutritional status, diurnal variations of serum and milk urea should be considered; time of sampling versus time of feeding is crucial. A small milk sample from a healthy quarter may give information on urea that is as good as that of a sample from regular milking.