In an ad libitum feeding experiment the greater growth of a group of animals may be due either to superior availability or to superior palatability of their diet. To distinguish between these two factors it is necessary to eliminate the effect of superior palatability by equalizing the food intake of corresponding animals. If the quantity offered is so small that it is always consumed entirely, the range of application of the experiment is restricted to conditions of substarvation, and at the same time the possibility of obtaining from the results any evidence of difference of palatability is sacrificed. The ration plan must therefore be designed to be as liberal as possible subject to the conditions that the aggregate food intakes are equalized in pairs (or larger groups), while at the same time the animal with the smallest appetite of each pair (or group) is always on the border-line of refusing a little of its portion. In the first set of experiments described in the preceding paper (1) a system of this kind is employed.