As compared with the results of the first period, Steers 1 and 3 show a less digestibility of dry matter, organic matter and total energy, while in the case of No. 2 the result is the reverse. In all cases, the total nitrogen and the proteids were more completely digested than in the first period, possibly owing to the higher percentage of this ingredient contained in the hay. The results on Nos. 1 and 3 agree very closely while those with No. 2 are higher. As already stated, this was the case in most of these experiments, but the dinerence in this case is notably greater than in any other trial. Nothing appears in the records of this experiment, however, to suggest any abnormal conditions. The excretion from day to day was fairly regular, the percentage of dry matter in the feces of the three animals was nearly the same (19.41 for No. 1; 19.42 for No. 2 and 18.56 for No. 3) and the live weights fail to indicate any abnormally large or small

excretion.

PROTEID METABOLISM.

For the proteid metabolism, we have the following figures per day

The ration produced a slight gain of proteids in the case of No. 1 and a somewhat greater gain in the case of Nos. 2 and 3.

LIVE WEIGHTS.

The animals were weighed daily as in the previous experiment, with the following results:

These results are represented graphically in the right hand por tion of Plate I in the same manner as those of Experiment I. The plate shows a prompt and rather rapid increase in weight. following the change in feed. Apparently this increase was due simply to the greater amount of the new ration, since it was fol lowed after a longer or shorter time by a gradual and steady diminu tion of weight. During the first fifteen days of the new feeding, the average gain per day and head was:

Steer 1,
Steer 2,

Steer 3,

0.67 kgs.

1.27 kgs.
0.47 kgs.

In other words, if the effects of the feed had been judged of from the first fifteen days, the ration would have appeared to be sufficient to cause a very considerable production, while as a matter of fact the subsequent weighings showed that it was insufficient for mainte nance. These figures throw an interesting light upon conclusions drawn from changes in live weight during short periods, especially immediately after a change of feed.

Following the first increase in live weight, fluctuations occur which render an exact interpretation of the results difficult. On the whole, however, there appears to have been a slight falling off in the case of each animal, but the exact amount is not so easily determined. It has been more or less customary, in experiments of this character, to regard the difference between the final (average) live weight and that at the beginning of the experiment (usually exclusive of a preliminary period) as representing the gain or loss due to the feeding. In this case, however, it is evident that the first and last averages have no greater validity than those of a few days earlier or later, and a glance at the plate makes evident that the apparent loss per day may be very materially modified by a judicious choice of the averages to be compared.

It is believed, therefore, that a more accurate idea of the real rate of loss may be obtained by the graphic method. Accordingly the broken line in the chart has been drawn to represent the general course of the live weight, and the rate of loss computed from its indications. Of course, the location of this line represents to a degree the individual judgment of the experimenter, but it is a judgment in view of all the facts in the case and therefore seems likely to approximate the truth more closely than a mechanical comparison of two weights or averages which must be arbitrarily selected.

Comparing the actual average weights for certain days on which they correspond with those indicated by the assumed line, we find the following as the average loss per day:

Estimating the composition of this loss of weight as in the previous experiment, we have per day and head:

Using the same data as before, we obtain the following figures for

The same three steers which served for the experiments of the previous year were used in these experiments also. They were on mountain pasture during the summer and when brought in, in October, had increased somewhat in weight. No. 1 having gained about 30 kgs., No. 2. about 10 kgs. and No. 3 about 20 kgs.

*The same average live weights are taken as in the previous experiment, since the surface of the animals, upon which the computation is really based, cannot have materially changed.