THE colorimetric method for the estimation of sugar in the blood as originally introduced by Lewis and Benedict [1915] has recently been modified by Benedict [1918], and considerably simplified. Addis and Shevky [1918], writing previous to the publication of Benedict's paper, have pointed out that the colour produced by the picric-sugar reaction is not strictly proportional in intensity to the amount of sugar present. Lewis and Benedict in their original paper state that in estimating pure sugar solutions absolutely exact results were obtained, using different amounts of sugar against the same standard, but the work of Addis and Shevky would appear to negative this conclusion. That the latter observers are correct in their statement admits of no doubt and this difference was observed by me before I knew of their paper. Addis and Shevky ascertained that the nearest approach to an exact proportion between intensity of colour produced and amount of sugar present was obtained when an amount of sodium carbonate sufficient to give a concentration of 10 % was present, and the heating was continued for 45 minutes at 1000. Suclh conditions are not realised in Benedict's modification and in estimations of pure sugar solutions by this method a considerable error is in fact found. In the estimation of sugar in normal bloods, however, the direct error would be of no great magnitude since the sugar content of such bloods approaches closely to that of the standard.The second possible source of error, the presence of interfering substances in the blood, was also dealt with by Addis and Shevky. They prepared large amounts of filtrates from hyperglycaemic bloods, and, comparing the curves obtained from these under different conditions of alkali concentration and temperature with those obtained from dextrose solutions, came to the conclusion that no interfering substances were piesent. In these experiments, however, the heating appears to have been prolonged for 45 minutes, no readings having been carried out during the very early stages of the heating.