MILK SAMPLING AT WEIGH TANKS—I

1. Samples were taken on three consecutive days from twenty‐two suppliers at three creameries. Six samples were taken at different levels and positions in the weigh tanks. 2. Details of two dippers for sampling at different depths in the tank are given. 3. Maximum differences between the six samples from each tank of milk ranged from 0·05 per cent to 0·61 per cent for fat and 0·07 per cent to 0·56 per cent for S.N.F., with means of 0·22 per cent and 0·21 per cent respectively. 4. Coefficients of variation have been calculated for each tank of milk together with mean figures for each supplier and each creamery. 5. Estimates have been made of the accuracy of the fat and S.N.F. content of a single sample taken from a tank of milk as a measure of the composition of the whole of the milk within the tank. 6. Maximum differences show that very large differences in composition can occur at different points in the weigh tank. 7. Comparison of the coefficients of variation for individual herds showed no correlation between variability within the tank with either the gallonage or the fat or S.N.F. content of the milk. A significant difference in variability was shown when the creameries were compared. This was attributed to the presence in one of the tanks of an obstacle in the shape of a foam breaker. 8. Single samples of milk taken from an unagitated weigh tank cannot be relied upon to give an accurate assessment of the quality of the milk within the tank. Some form of agitation must be practised if a reliable sample is to be taken. 9. The results of a pilot investigation carried out at one creamery in the winter months indicate very much greater variation in composition within unagitated tanks at this time of the year. The results of this investigation show that the variability of fat content can be considerably reduced by a reasonable plunging of the tank contents. The effect of plunging on S.N.F. variability was not so encouraging.