There is a remarkable contradiction between the report of Katagiri, et al., that the liver of the cattle contains an enzyme which converts 6, 7-dimethylribolumazine into riboflavin, that of Suzuoki, et al. that the former was nearly ineffective in their curative tests on the riboflavin-deficient rats. The present authors duplicated the method of Katagiri, et al., confirmed that a phosphate buffer extract of the liver of the cattle transformed 6, 7-dimethylribolumazine into riboflavin in vitro. The same technique was applied to the liver of rats, but no increase of riboflavin was observed. Addition of diacetyl or sodium acetate as a carbon donor also gave the same result. These data are in good agreement with the result of Suzuoki, et al. in their tests on the whole body of the animal. Further the same experiments were attempted with the liver of rabbits the result was the same as in the case of the liver of rats. On the other hand, a rabbit was fasted for a definite time administered with 6, 7-dimethylribolumazine. The urine excreted before after the administration was collected separately subjected to paper partition chromatography. As a result, a distinct spot of riboflavin was observed in the foreurinebut it became faint in the after-urine, and a conspicuous spot of 6, 7-dimethylribolumazine appeared in its stead. Although this result is merely qualitative, it is consistent with the result of the experiment on the liver homogenate of animals.