1. It is shown that results for the loss in weight of a soil on oven heating can be obtained to a very satisfactory degree of accuracy when a Hearson electrically controlled oven is used.

2. Results obtained by heating soils to temperatures ranging from 5° to 250° give smooth curves connecting loss in weight with rise in temperature; from which it is concluded that there is no sudden alteration in the structure of a soil when it is heated to 100°, and that the airdry moisture of a soil, as determined with sufficient accuracy by the usual methods, is a convenient empirical factor, but not a representation of any fundamental soil property.

3. An examination is made of the factors contributing to the observed total loss in weight when the soil is heated, and an explanation offered of the contributions made by the different types of soil water and by the soil colloids. It is found that the conclusions drawn from this discussion confirm views developed earlier with regard to the behaviour of the water in the soil, and the absence of any sharp dividing line between the different classes into which the soil water is usually divided.