1. Results from a uniformity trial with fifteen pairs each of one-egg and two-egg twins and pairs of half-sibs are described.

2. Intra-pair variance in age at the first observed heat was less for the one-egg than for the other pairs but not significantly so. The slowest growing pairs showed delay in onset of heat.

3. 71·8% of the intervals between heats shown by unserved heifers were 17–27 days in length. No genetic variation and no individuality could be detected in either this character or the number of services per conception.

4. The variance in the gestation period was less within the one-egg than within the other pairs but not significantly so.

The preparation of crystalline rennin previously described (1) was somewhat long and inconvenient, but the knowledge gained about the properties of the crystals suggested that a simpler method might be successful. Since then Møgensen (2) has shown that the mere precipitation of commercial rennet with saturated sodium chloride gives a product with proteolytic properties resembling those of the crystalline material, and with enhanced milk-clotting activity per unit of dry weight. Some years previously, Hankinson (3) had also achieved considerable purification by these means. The results of an attempt to discover a simple method of crystallization are described below.

1. Iodinated casein or L-thyroxine, under the conditions of this experiment, had no adverse effect on the general health of a dual purpose (Red Poll) or a dairy breed (Ayrshire) of cattle.

2. The treatment did not increase the total milk or butter-fat production.

3. The treated animals did significantly better than the controls in one lactation only and in the following variables only: (a) daily milk yield adjusted for pre-treatment differences—Red Polls, lact. 1; (b) milk yield for duration of treatment adjusted for pre-treatment differences—Red Polls, lact. 1.

The various theories commonly used to explain the increasing fat content of cow's milk with stage of milking are examined and found inadequate. It is suggested that the trends in fat content as a cow is being milked may be explained by the clustering of the fat globules and the consequent partial filtering out of the clusters as the milk flows from the gland. Evidence for clustering of fat in cow's milk at body temperature is presented and the application of the theory to the sow discussed.

In the determination of aldehyde values of slightly oxidized butterfat samples by Schibsted's method, the colour of the butterfat prevents accurate matching with the colour standard. Interference by butterfat colour is eliminated, and more accurate results are obtained if the colour is measured in a photoelectric colorimeter at a wavelength of 510 mµ. A small constant correction for the absorption by butterfat at this wave-length may be made.