A VERY extensive literature has grown up during the past half century, on the subject of the chemistry of normal human milk. So vast has this literature become, that the task of absorbing and correlating its conclusions into a coherent understanding of the subject as a whole has become one of extreme difficulty and laboriousness. On the other hand, some attempt to do so must form the basis of any new work. From a review of the main communications in the literature, a list of which is appended, certain facts stand out and can, we suggest, be made the basis of an attempt to explore the matter in greater detail. 1. We have reliable and accurate knowledge of the general nature of the main constituents of human milk, and adequate methods for their estimation. 2. We have on record a very large number of observations of milk at all stages of lactation, and are aware of the changes most usually found. If, however, we review the aggregate of facts presented by points 1 and 2, an admirable consideration of which appeared in a communication last year [Gardner and Fox, 1925], the main feature of the picture is the exceedingly wide range of values shown in normal milk by each constituent. These variations complicate our understanding of the composition of normal milk, and put a serious bar to any study of the conditions of abnormal milk. Furthermore, the great bulk of the work done hitherto on normal milk has not only been done abroad, but has also been achieved through the study of wet nurses-that is, of women accustomed to a specialised routine for the production and extraction of milk. As their conditions differ radically from those of ordinary nursing mothers, it cannot safely be assumed that the lactation results will necessarily be the same. The present paper is the report of an attempt on the basis of work already known to deal with both these positions, to investigate in ordinary normal mothers the factors determining the variations in the chemical composition of their milk.