Some years ago my colleague, Mr E. T. Halnan, suggested the construction of the present instrument for the purpose of determining figures for domestic fowls corresponding to those of Kellner for ruminants.

Owing to the activity of the fowl, it is essential that an instrument for this purpose should be capable of recording accurately a heat output which may be varying with some rapidity.

From this point of view the advantages of the indirect method were considered but were eventually rejected for the following reasons: (a) the advantage accruing from our greater experience of the direct method in this laboratory, (b) the work of Adams & Poulton (1933, 1934, 1935) calling in question the fundamental bases of the indirect method, and (c) the very variable R.Q.'S obtained in earlier work by numerous investigators working with birds.

That the decision was wise soon became apparent when work on birds was begun. It is certain that no ordinary apparatus of that kind could possibly take account of the variations caused by the constant changes in activity.

The principle of a “compensatory” or “balanced” calorimeter more or less as developed by Noyons (1927), Benedict (1927) and Heyster (1933), was selected as the most likely to give promptness in registration, and this choice has been amply justified in the result. Such instruments can be very lightly constructed, although where living animals are concerned, since their movements cannot be controlled, as can those of a human being, a greater margin of strength has to be provided, which naturally increases the heat capacity and lag and reduces sensitivity to small changes in heat evolution.