A serious difficulty in lactational studies on small laboratory animals is the lack of a satisfactory means of milking them and thus directly measuring their milk yields. At present, lactational performance in rats can only be inferred from litter growth and survival data determined under carefully standardized conditions. Comparison of litter-growth curves (and survival data) for different groups of rats gives some idea of their relative lactational performances but it is not possible to calculate daily milk yields from such curves since the weight losses due to the excreta and insensible perspiration are unknown.Attempts to determine true milk yields by the complicated and not entirely unobjectionable (see discussion by Enzmann) procedures used by Enzmann [1933] for mice and by Brody & Nisbet [1938] for rats, which involve determination not only of the daily litter weight gains but also of the daily losses due to the excreta and perspiration, are clearly impracticable in experiments carried out on any considerable scale. Thus when, in the course of studies of the relation between the adrenal cortex and lactation [Cowie & Folley, 1946a, b], we wished to determine quantitatively the lactational responses of adrenalectomized rats to treatment with adrenal cortex hormones, we had to devise a quantitative criterion of lactation which was derived solely from the ordinary litter-growth curve.
THE EQUATION OF THE LITTER-GROWTH CURVEDaggs [1935] proposed a 'lactation index' for rats based on the claim (in which respect he followed Brody [1927]) that the points obtained by plotting the logarithm of the mean weight of the young against time fall along two straight lines which intersect at approximately í = 10 days. The numerical index used by Daggs has, however, little theoretical significance and cannot be used for quantitative comparisons even if the premises on which it is based are true.We have constructed a mean growth curve (Fig.