SUMMARYTen ewes were fed for nine months on a cobalt‐deficient ration of wheaten hay chaff and gluten. Four of the animals were given cobalt throughout the whole period. The remaining six animals were given cobalt for the last two months only. Three of the ewes in the + cobalt group were pair fed with three in the ‐ cobalt group.Biopsy samples of liver were taken at the beginning and end of the nine months period and at three intermediate times. Vitamin B12was measured in the liver samples by theL. leichmanniiassay method. Pantothenic acid was measured by means of microbiological assays withL. plantarumandL. casei. Total pantothenic acid, which was taken as a measure of coenzyme A, was measured after splitting its bound forms with alkaline phosphatase and a pigeon liver enzyme preparation. Free pantothenic acid, measured in half the liver plugs, averaged 2·5 p.c. of the total pantothenic acid (range 0·6–4·3 p.c.).Deprivation of cobalt led to a marked lowering of liver vitamin B12(from a mean value of 1·05 μg./gm. to a mean value of 0·12 μg./gm. in 6 months. The mean concentration rose to 0·48 μg./gm. two months after administration of cobalt.In six months of cobalt deprivation the pantothenic acid concentration in the livers of all the deficient sheep rose, the mean increase being 70 p.c. This change was reversible; two months after giving cobalt the pantothenic acid concentration had fallen again.Statistical analysis of the data for the pair‐fed sheep showed that there was a highly significant difference between the pantothenic acid concentrations in the livers of the sheep in the positive and negative groups, when the latter had been cobalt‐deficient for six months and seven months. The elevated pantothenic acid concentration was associated with a vitamin B12, concentration of about 0·1 μg./gm. of liver.There was no great or consistent difference between pantothenic acid measurements made by the two microbiological assay methods.