In observations on domestic animals, or with a group of men, extending over a considerable period of time, small changes in the actual amount of carbon-dioxide residual in the chamber at any given time are practically without influence upon the final results. When, however, the experimental period is shortened to such an extent that variations in the residual amount inside the chamber may become a measurable proportion of the total amount withdrawn during the period, the necessity of a careful determination of the residual amounts is obvious. This feature of the apparatus, which plays no r61e in experimenting with domesticated animals when long experimental periods are employed, requires the use of a gas-analysis apparatus for short experimental periods with humans. Practically all of our work was carried out by using a Sonden-Pettersson gas-analysis apparatus,1 which made it possible to determine the carbon dioxide to approximately 0.001 per cent. Since the total volume of the chamber was not far from 44,000 liters, it will be seen that each one thousandth of 1 per cent carbon dioxide corresponded, roughly speaking, to 0.9 gram of carbon dioxide; hence the accuracy was all that could be desired.

In the latter part of our experimenting during the season of 1917-18 we were much attracted by the accuracy of the small Haldane apparatus for carbon dioxide alone.2 This gives an accuracy for carbon-dioxide measurements that for the great majority of experiments is perfectly satisfactory. The apparatus is simple, relatively easy to manipulate, very rapid, sufficiently compact to be portable and has none of the fragile parts so essential to the Sonden apparatus. Our experience with the various types of Pettersson apparatus which have been put upon the market has been upon the whole rather unsatisfactory.

While in the large majority of our observations on Squad A, corrections for the residual air were not only unnecessary but at times, owing to errors of gas analysis, positively disadvantageous, we have in practically all instances corrected our periods for changes in the residual analysis, the analyses being for the most part made with the Sonden apparatus. Similarly, in check tests, in which the carbon dioxide was admitted to the large chamber from a steel cylinder and the carbon dioxide was determined in short periods, the residual analysis played a very important r61e, for the introduction of the carbon dioxide into the chamber was extremely irregular. On the other hand, with a group of normal men or women pursuing a certain definite procedure such as reading or walking, the carbon-dioxide production becomes very regular after a very short time. In certain instances the accuracy of the Sonden apparatus for finding the residual carbon dioxide was determined by having no ventilation in the chamber, analyzing the air at the start, introducing a certain amount of carbon dioxide, analyzing the air at the end, and comparing the carbon-dioxide content of the air at the beginning and the end of the test. This type of check also gave most gratifying results whenever used.

1 Benedict, Carnegie Inst. Wash. Pub. No. 166, 1912. 2 Haldane, Methods of air analysis, 1912, p. 62.

The humidity and temperature of the air inside the chamber were determined by the wet and dry bulb thermometer, easily read through the glass window. The method of calculation is indicated in detail for a single experimental period in table 4, and the summary of results of an entire night experiment in table 5.

Table 4. - Typical Calculation Of A Period With The Group Respiration Chamber (12h32m - 1h02m A.m.Oct. 7, 1917)

Calculation of residual carbon dioxide in the chamber*.

Calculation of carbon dioxide produced during the period.

* After Nov. 11,1917. it was assumed that each 0.001 per cent change in the residual corresponded to 0.8 gram of CO2 and calculations were no longer made. ** 2.54 equals the percentage of the total outgoing air that actually passed through one set of the absorption system, i. e.. the factor for the 60 mm. disk.

Table 5. - Summary Of Carbon-Dioxide Measurements With The Group Respiration Chamber On Night Of Oct. 6-7, 1917. - Squad B

1 2.54 equals percentage of total outgoing air actually passing through one set of the absorption system.

2 Periods from 1h.02m to 3h.02m a. m. selected as minimum periods. The average of these is 292 gms. CO2 per hour. Total body-surface of squad equals 21.8 square meters. Assuming 3.025 calories as heat equivalent per gram of CO2 at a respiratory quotient 0.81, the heat per square meter would be found by the following calculation:

292 X3.0254-21.8 - 40.5 calories per square meter per hour.

Arrangements For Sleeping

Night experiments alone were made in our use of this apparatus in the diet research. Twelve beds were provided in 3 sections of 4 beds each, as shown in figure 5, page 92. Good springs with suitable bedding made comfortable sleeping quarters. Glass jars for night urine were hung in wire frames at the foot of each bed.

Technique For Determining Effect Of Muscular Work

The men in Squads A and B were all engaged in various forms of muscular activity, ranging from the severe exercise of leading gymnasium classes for several hours a day to that of the activity necessary for moving about the campus from building to building. Even our most inactive man showed a considerable amount of muscular activity according to his pedometer and physical activity records. It was important, therefore, to determine the effect of the reduced diet on muscular activity and the physiological phenomena accompanying it.

Measurement Of Work Of Bicycle Riding

For this purpose we fortunately obtained the cooperation of Professor A. G. Johnson, of the faculty of the International Young Men's Christian Association College. As a part of an extended study upon the influence of muscular activity upon the heart rate, which had been carried out for a number of years at the college under the direction of Professor J. H. McCurdy and Professor Elmer Berry, a study was made of the length of time required for the heart rate to return to normal after a definite amount of muscular work. Use was made of a bicycle ergometer belonging to the Nutrition Laboratory, which has been described in detail in an earlier publication.1 This was shipped to Springfield, there connected with a storage battery consisting of fifteen 1.5 volt Edison cells, and a mil-ammeter and sliding resistance placed in circuit. By controlling the current passing through the fields of the ergometer, holding it constant at 1.35 amperes, and adjusting the rate of revolution of the pedals by means of a metronome beating 80 per minute, the subjects performed a definite amount of work for a period of 5 minutes. After mounting the ergometer, the man rode for an exact period of 5 minutes at the rate of 80 pedal revolutions per minute. Upon the completion of the 5-minute period, he lay down upon a bench and the pulse-rate was counted for the first 15 seconds of every minute until the normal resting pulse for the day was reached.