The conclusions that have been arrived at from the foregoing statistical examination of the available data on the physical properties of the Natal soils used by Coutts are:

(1) The base exchange capacity of the soil, as measured by Schofield's potassium phosphate buffer method, appears to be of predominant importance for predicting several of the soil properties, as for example the sticky point and the moisture content at 50 per cent. relative humidity. On the other hand, the clay content seems to be of quite minor importance in predicting these soil properties.

(2) The information from the data supplied by the Keen-Raczkowski box indicates that ω, the weight of water held per gram of soil in the box is very closely correlated with the base exchange capacity of the soil, while the swelling and pore-space parameters are more complex. The swelling υ seems to be dependent on the base exchange capacity of the soil and a term probably representing the structure of the soil, while the pore space p seems to be dependent on the clay content and a soil structure term. By choosing an appropriate definition of the pore space practically all the structure information given by the swelling parameters is contained in it.

Abnormal leg trouble or perosis occurred in chickens getting an excessive amount of meat and bone meal in their ration, and rarely occurred when the bone meal content was low or nil.

The use of the battery brooder was found to intensify bone deformities, when the ration was unsatisfactory in mineral balance.

The floor space allowed was not found to be a factor in perosis.

Under the conditions of the experiment, slipped tendons were not produced by feeding calcium carbonate (ground limestone) in the ration.

The protein content of the ration and the rate of growth did not materially affect the occurrence of perosis.