After a great deal of preliminary experimenting, in which the Leeds & Northrup Company have most generously interpreted our specifications, they have furnished us with an apparatus which meets to a high degree of satisfaction the conditions imposed. The thermometers themselves have already been discussed. (See page 30.) The recording apparatus consists of three parts: (1) the galvanometer; (2) the creeper or automatic sliding-contact; (3) the clockwork for the forward movement of the roll of coordinate paper and to control the periodic movement of the creeper.

Under ordinary conditions with rest experiments in the chair calorimeter or bed calorimeter, the temperature differences run not far from 2 deg. to 4 deg. Thus, it is seen that if the apparatus is to meet the conditions of the specifications it must measure differences of 2 deg. C. to within 0.01 deg. C. Provision has also been made to extend the measurement of temperature differences with the apparatus so that a difference of 8 deg. can be measured with the same percentage accuracy.

FUNDAMENTAL PRINCIPLE OF THE APPARATUS.

The apparatus depends fundamentally upon the perfect balancing of the two sides of a differential electric circuit. A conventional diagram, fig. 19, gives a schematic outline of the connections. The two galvanometer coils, fl and fr, are wound differentially and both coils most carefully balanced so that the two windings have equal temperature coefficients. This is done by inserting a small shunt y, parallel with the coil fl, and thus the temperature coefficient of fl and fr are made absolutely equal. The two thermometers are indicated as T{1} and T{2} and are inserted in the ingoing and outgoing water respectively. A slide-wire resistance is indicated by J, and r is the resistance for the zero adjustment. Ba, Z, and Z{1} are the battery and its variable series resistances. If T{1} and T{2} are exactly of the same temperature, i. e., if the temperature difference of the ingoing and outcoming water is zero, the sliding contact q stands at 0 on the slide-wire and thus the resistance of the system from 0 through fl, r, and T{1} back to the point C is exactly the same as the resistance of the slide-wire J plus the coil fr plus T{2} back to the point C. A rise in temperature of T{2} gives an increase of resistance in the circuit and the sliding contact q moves along the slide-wire toward J maximum until a balance is obtained.



Provision is made for automatically moving the contact q by electrical means and thus the complete balance of the two differential circuits is maintained constant from second to second. As the contact q is moved, it carries with it a stylographic pen which travels in a straight line over a regularly moving roll of coordinate paper, thus producing a permanently recorded curve indicating the temperature differences. The slide-wire J is calibrated so that any inequalities in the temperature coefficient of the thermometer wires are equalized and also so that any unit-length on the slide-wire taken at any point along the temperature scale represents a resistance equal to the resistance change in the thermometer for that particular change in temperature. With the varying conditions to be met with in this apparatus, it is necessary that varying values should be assigned at times to J and to r. This necessitates the use of shunts, and the recording range of the instrument can be easily varied by simple shunting, i. e., by changing the resistance value of J and r, providing these resistances unshunted have a value which takes care of the highest obtained temperature variations.

Fig. 19 shows the differential circuit complete with all its shunts. S is a fixed shunt to obtain a range on J; S' is a variable shunt to permit very slight variations of J within the range to correct errors due to changing of the initial temperatures of the thermometers; y is a permanent shunt across the galvanometer coil fl, to make the temperature coefficients of fl and fr absolutely equal; Z is the variable resistance in the battery-circuit to keep the current constant; r is a permanent resistance to fix the zero on varying ranges; S'' plus S{1} constitutes a variable shunt to permit slight variations of r to finally adjust 0 after S' is fixed and t is a permanent shunt across the thermometer T{1} to make the temperature coefficient of T{1} equal to that of T{2}.

The apparatus can be used for measuring temperature differences from 0 deg. to 4 deg. or from 0 deg. to 8 deg. When on the 0 deg. to 8 deg. range, the shunt S is open-circuited and the shunt S' alone used. The value of S, then, is predetermined so as to affect the value of the wire J and thus halve its influence in maintaining the balance. Similarly, when the lower range, i. e., from 0 deg. to 4 deg., is used, the resistance r is employed, and when the higher range is used another value to r must be given by using a plug resistance-box, in the use of which the resistance r is doubled.

The resistance S'' and S_{1} are combined in a slide-wire resistance-box and are used to change the value of the whole apparatus when there are marked changes in the position of the thermometric scale. Thus, if the ingoing water is at 2 deg. C. and the outcoming water at 5 deg. C. in one instance, and in another instance the ingoing water is 13 deg. and the outgoing water is 15 deg., a slight alteration in the value of S_{1}, and also of S', is necessary in order to have the apparatus draw a curve to represent truly the temperature differences. These slight alterations are determined beforehand by careful tests and the exact value of the resistances in S' and in S_{1} are permanently recorded for subsequent use.

THE GALVANOMETER.

The galvanometer is of the Deprez-d'Arsonval type and has a particularly powerful magnetic field, in which a double coil swings suspended similar to the marine galvanometer coils. This coil is protected from vibrations by an anti-vibration tube A, fig. 20, and carries a pointer P which acts to select the direction of movement of the recording apparatus, the movable contact point q, fig. 19. In front of this galvanometer coil and inclosed in the same air-tight metal case is the plunger contact Pl, fig. 21. The galvanometer pointer P swings freely below the silver contacts S{1} and S{2}, just clearing the ivory insulator i. The magnet plunger makes a contact depending upon the adjustment of a clock at intervals of 2 seconds. So long as both galvanometer coils are influenced by exactly the same strength of current, the pointer will stand in line with and immediately below i and no current passes through the recording apparatus. Any disturbance of the electrical equilibrium causes the pointer P to swing either toward S{1} or S{2}, thus completing the circuit at either the right hand or the left hand, at intervals of 2 seconds. The movement of the pointer away from its normal position exactly beneath i to either S{1} on the left hand or S{2} on the right, results from an inequality in the current flowing through the two coils in the galvanometer. The difference in the two currents passing through these coils is caused by a change in temperatures of the two thermometers in the water circuit.



THE CREEPER.

The movement of the sliding-contact q, fig. 19, along the slide-wire J, is produced by means of a special device called a creeper, consisting of a piece of brass carefully fitted to a threaded steel rod some 30 centimeters long. The movement of this bar along this threaded rod accomplishes two things. The bar is in contact with the slide-wire J and therefore varies the position of the point q and it also carries with it a stylographic pen. The movements of this bar to the right or the left are produced by an auxiliary electric current, the contact of which is made by a plunger-plate forcing the pointer P against either S{1} or S{2}. P makes the contact between Pl and either S{1} or S{2} and sends a current through solenoids at either the right or the left of the creeper. At intervals of every 2 seconds the plunger rises and forces the pointer P against either S{1}, i, or S{2} above. The movement of this plunger is controlled by a current from a 110-volt circuit, the connections of which are shown in fig. 22. If the contact is made at T, the current passes through 2,600 ohms, directly across the 110-volt circuit, and consequently there is no effective current flowing through the plunger Pl. When the contact T is open, the current flows through the plunger in series with 2,600 ohms resistance. T is opened automatically at intervals of 2 seconds by the clock.



The movement of the contact arm along the threaded rod is produced by the action of either one of two solenoids, each of which has a core attached to a rack and pinion at either end of the rod. If the current is passed through the contact S_{1}, a current passes through the left-hand solenoid, the core moves down, the rack on the core moves the pinion on the rod through a definite fraction of a complete revolution and this movement forces the creeper in one direction. Conversely, the passing of the current through the solenoid at the other end of the threaded rod moves the creeper in the other direction. The distance which the iron rack on the end of the core is moved is determined carefully, so that the threaded rod is turned for each contact exactly the same fraction of a revolution. For actuating these solenoids, the 110-volt circuit is again used. The wire connections are shown in part in fig. 21, in which it is seen that the current passes through the plunger-contact and through the pointer P to the silver plate S_{1} and then along the line G_{1} through 350 ohms wound about the left-hand solenoid back through a 600-ohm resistance to the main line. The use of the 110-volt current under such circumstances would normally produce a notable sparking effect on the pointer P, and to reduce this to a minimum there is a high resistance, amounting to 10,000 ohms on each side, shunted between the main line and the creeper connections. This shunt is shown in diagram in fig. 22. Thus there is never a complete open circuit and sparking is prevented.

THE CLOCK.

The clock requires winding every week and is so geared as to move the paper forward at a rate of 3 inches per hour. The contact-point for opening the circuit T on fig. 22 is likewise connected with one of the smaller wheels of the clock. This contact is made by tripping a little lever by means of a toothed wheel of phosphor-bronze.

INSTALLATION OF THE APPARATUS.



The whole apparatus is permanently and substantially installed on the north wall of the calorimeter laboratory. A photograph showing the various parts and their installation is given in fig. 23. On the top shelf is seen the galvanometer and on the lower shelf the recorder with its glass door in front and the coordinate paper dropping into the box below. The curve drawn on the coordinate paper is clearly shown. Above the recorder are the resistance-boxes, three in number, the lower one at the left being the resistance S_{1}, the upper one at the left being the resistance S', and the upper one at the right being the resistance Z_{1}. Immediately above the resistance-box Z_{1} is shown the plug resistance-box which controls on the one hand the resistance _r_ and on the other hand the resistance S, both of which are substantially altered when changing the apparatus to register from the 0 deg. to 4 deg. scale to the 0 deg. to 8 deg. scale. A detailed wiring diagram is given in fig. 24.

TEMPERATURE CONTROL OF THE INGOING AIR.

[Illustration: FIG. 25.—Section of calorimeter walls and part of ventilating air-circuit, showing part of pipes for ingoing air and outgoing air. On the ingoing air-pipe at the right is the lamp for heating the ingoing air. Just above it, H is the quick-throw valve for shutting off the tension equalizer IJ. I is the copper portion of the tension equalizer, while J is the rubber diaphragm; K, the pet-cock for admitting oxygen; F, E, G, the lead pipe conducting the cold water for the ingoing air; and C, the hair-felt insulation. N, N are brass ferules soldered into the copper and zinc walls through which air-pipes pass; M, a rubber stopper for insulating the air-pipe from the calorimeter; O, the thermal junctions for indicating differences of temperature of ingoing and outgoing air and U, the connection to the outside; QQ, exits for the air-pipes from the box in which thermal junctions are placed; P, the dividing plate separating the ingoing and outgoing air; R, the section of piping conducting the air inside the calorimeter; S, a section of piping through which the air passes from the calorimeter; A, a section of the copper wall; Y, a bolt fastening the copper wall to the 2-1/2 inch angle W; B, a portion of zinc wall; C, hair-felt lining of asbestos wall D; T-J, a thermal junction in the walls.]

In passing the current of air through the calorimeter, temperature conditions may easily be such that the air entering is warmer than the outcoming air, in which case heat will be imparted to the calorimeter, or the reverse conditions may obtain and then heat will be brought away. To avoid this difficulty, arrangements are made for arbitrarily controlling the temperature of the air as it enters the calorimeter. This temperature control is based upon the fact that the air leaving the chamber is caused to pass over the ends of a series of thermal junctions shown as O in fig. 25. These thermal junctions have one terminal in the outgoing air and the other in the ingoing air, and consequently any difference in the temperature of the two air-currents is instantly detected by connecting the circuit with the galvanometer. Formerly the temperature control was made a varying one, by providing for either cooling or heating the ingoing air as the situation called for. The heating was done by passing the current through an electric lamp placed in the cross immediately below the tension equalizer J. Cooling was effected by means of a current of water through the lead pipe E closely wrapped around the air-pipe, water entering at F and leaving at G. This lead pipe is insulated by hair-felt pipe-covering, C. More recently, we have adopted the procedure of passing a continuous current of water, usually at a very slow rate, through the lead pipe E and always heating the air somewhat by means of the lamp, the exact temperature control being obtained by varying the heating effect of the lamp itself. This has been found much more satisfactory than by alternating from the cooling system to the heating system. In the case of the air-current, however, it is unnecessary to have the drop-sight feed-valve as used for the wall control, shown in fig. 13.

THE HEAT OF VAPORIZATION OF WATER.

During experiments with man not all the heat leaves the body by radiation and conduction, since a part is required to vaporize the water from the skin and lungs. An accurate measurement of the heat production by man therefore required a knowledge of the amount of heat thus vaporized. One of the great difficulties in the numerous forms of calorimeters that have been used heretofore with man is that only that portion of heat measured by direct radiation or conduction has been measured and the difficulties attending the determination of water vaporized have vitiated correspondingly the estimates of the heat production. Fortunately, with this apparatus the determinations of water are very exact, and since the amount of water vaporized inside the chamber is known it is possible to compute the heat required to vaporize this water by knowing the heat of vaporization of water.

Since the earlier reports describing the first form of calorimeters were written, there has appeared a research by one of our former associates, Dr. A. W. Smith[11] who, recognizing the importance of knowing exactly the heat of vaporization of water at 20 deg., has made this a special object of investigation. When connected with our laboratory a number of experiments were made by Doctors Smith and Benedict in an attempt to determine the heat of vaporization of water directly in a large calorimeter; but for lack of time and pressure of other experimental work it was impossible to complete the investigation. Subsequently Dr. Smith has carried out the experiments with the accuracy of exact physical measurements and has given us a very valuable series of observations.

Using the method of expressing the heat of vaporization in electrical units, Smith concludes that the heat of vaporization of water between 14 deg. and 40 deg. is given by the formula

L (in joules) = 2502.5 - 2.43T

and states that the "probable error" of values computed from this formula is 0.5 joule. The results are expressed in international joules, that is, in terms of the international ohm and 1.43400 for the E.M.F. of the Clark cell at 15 deg. C., and assuming that the mean calorie is equivalent to 4.1877 international joules,[12] the formula reads

L (in mean calories) = 597.44 - 0.580T

With this formula Smith calculates that at 15 deg. the heat of vaporization of water is equal to 588.73 calories; at 20 deg., 585.84 calories; at 25 deg., 582.93 calories; at 30 deg., 580.04 calories;[13] and at 35 deg., 577.12 calories. In all of the calculations in the researches herewith we have used the value found by Smith as 586 calories at 20 deg. Inasmuch as all of our records are in kilo-calories, we multiply the weight of water by the factor 0.586 to obtain the heat of vaporization.

THE BED CALORIMETER.

The chair calorimeter was designed for experiments to last not more than 6 to 8 hours, as a person can not remain comfortably seated in a chair much longer than this time. For longer experiments (experiments during the night and particularly for bed-ridden patients) a type of calorimeter which permits the introduction of a couch or bed has been devised. This calorimeter has been built, tested, and used in a number of experiments with men and women. The general shape of the chamber is given in fig. 26. The principles involved in the construction of the chair calorimeter are here applied, i. e., the use of a structural-steel framework, inner air-tight copper lining, outer zinc wall, hair-felt insulation, and outer asbestos panels. Inside of the chamber there is a heat-absorbing system suspended from the ceiling, and air thermometers and thermometers for the copper wall are installed at several points. The food-aperture is of the same general type and the furniture here consists simply of a sliding frame upon which is placed an air-mattress. The opening is at the front end of the calorimeter and is closed by two pieces of plate glass, each well sealed into place by wax after the subject has been placed inside of the chamber. Tubes through the wall opposite the food-aperture are used for the introduction of electrical connections, ingoing and outgoing water, the air-pipes, and connections for the stethoscope, pneumograph, and telephone.

The apparatus rests on four heavy iron legs. Two pieces of channel iron are attached to these legs and the structural framework of the calorimeter chamber rests upon these irons. The method of separating the asbestos outer panels is shown in the diagram. In order to provide light for the chamber, the outer wall in front of the glass windows is made of glass rather than asbestos. The front section of the outer casing can be removed easily for the introduction of a patient.

In this chamber it is impossible to weigh the bed and clothing, and hence this calorimeter can not be used for the accurate determination of the moisture vaporized from the lungs and skin of the subject, since here (as in almost every form of respiration chamber) it is absolutely impossible to distinguish between the amount of water vaporized from bed-clothing and that vaporized from the lungs and skin of the subject. With the chair calorimeter, the weighing arrangements make it possible to weigh the chair, clothing, etc., and thus apportion the total water vaporized between losses from the chair, furniture, and body of the man. In view of the fact that the water vaporized from the skin and lungs could not be determined, the whole interior of the chamber of the bed calorimeter has been coated with a white enamel paint, which gives it a bright appearance and makes it much more attractive to new patients. An incandescent light placed above the head at the front illuminates the chamber very well, and as a matter of fact the food-aperture is so placed that one can lie on the cot and actually look outdoors through one of the laboratory windows.



Special precaution was taken with this calorimeter to make it as comfortable and as attractive as possible to new and possibly apprehensive patients. The painting of the walls unquestionably results in a condensation of more or less moisture, for the paint certainly absorbs more moisture than does the metallic surface of the copper. The chief value of the determination of the water vaporized inside of the chamber during an experiment lies, however, not in a study of the vaporization of water as such, but in the fact that a certain amount of heat is required to vaporize the water and obviously an accurate measure of the heat production must involve a measure of the amount of water vaporized. So far as the measurement of heat is concerned, it is immaterial whether the water is vaporized from the lungs or skin of the subject or the clothing, bedding, or walls of the chamber; since for every gram of water vaporized inside of the chamber, from whatever source, 0.586 calorie of heat must have been absorbed.

The apparatus as perfected is very sensitive. The sojourn in the chamber is not uncomfortable; as a matter of fact, in an experiment made during January, 1909, the subject remained inside of the chamber for 30 hours. With male patients no difficulty is experienced in collecting the urine. No provision is made for defecation, and hence it is our custom in long experiments to empty the lower bowel with an enema and thus defer as long as possible the necessity for defecation. With none of the experiments thus far made have we experienced any difficulty in having to remove the patient because of necessity to defecate in the cramped quarters. It is highly probable that, with the majority of sick patients, experiments will not extend for more than 8 or 10 hours, and consequently the apparatus as designed should furnish most satisfactory results.

In testing the apparatus by the electrical-check method, it has been found to be extremely accurate. When the test has been made with burning alcohol, as described beyond, it has been found that the large amount of moisture apparently retained by the white enamel paint on the walls vitiates the determination of water for several hours after the experiment begins, and only after several hours of continuous ventilating is the moisture content of the air brought down to a low enough point to establish equilibrium between the moisture condensed on the surface and the moisture in the air and thus have the measured amount of moisture in the sulphuric acid vessels equal the amount of moisture formed by the burning of alcohol. Hence in practically all of the alcohol-check experiments, especially of short duration, with this calorimeter, the values for water are invariably somewhat too high. A comparison of the alcohol-check experiments made with the bed and chair calorimeters gives an interesting light upon the power of paint to absorb moisture and emphasizes again the necessity of avoiding the use of material of a hygroscopic nature in the interior of an apparatus in which accurate moisture determinations from the body are to be made.

The details of the bed calorimeter are better shown in fig. 4. The opening at the front is here removed and the wooden track upon which the frame, supporting the cot, slides is clearly shown. The tension equalizer (see page 71) partly distended is shown connected to the ingoing air-pipe, and on the top of the calorimeter connected to the tension equalizer is a Sonden manometer. On the floor at the right is seen the resistance coil used for electrical tests (see page 50). A number of connections inside the chamber at the left are made with electric wires or with rubber tubing. Of the five connections appearing through the opening, reading from left to right, we have, first, the rubber connection with the pneumograph, then the tubing for connection with the stethoscope, then the electric-resistance thermometer, the telephone, and finally a push button for bell call. The connections for the pneumograph and stethoscope are made with the instruments outside on the table at the left of the bed calorimeter.

MEASUREMENTS OF BODY-TEMPERATURE.

While it is possible to control arbitrarily the temperature of the calorimeter by increasing or decreasing the amount of heat brought away, and thus compensate exactly for the heat eliminated by the subject, the hydrothermal equivalent of the system itself being about 20 calories—on the other hand the body of the subject may undergo marked changes in temperature and thus influence the measurement of the heat production to a noticeable degree; for if heat is lost from the body by a fall of body-temperature or stored as indicated by a rise in temperature, obviously the heat produced during the given period will not equal that eliminated and measured by the water-current and by the latent heat of water vaporized. In order to make accurate measurements, therefore, of the heat-production as distinguished from the heat elimination, we should know with great accuracy the hydrothermal equivalent of the body and changes in body temperature. The most satisfactory method at present known of determining the hydrothermal equivalent of the body is to assume the specific heat of the body as 0.83.[14] This factor will of course vary considerably with the weight of body material and the proportion of fat, water, and muscular tissue present therein, but for general purposes nothing better can at present be employed. From the weight of the subject and this factor the hydrothermal equivalent of the body can be calculated. It remains to determine, then, with great exactness the body temperature.

Recognizing early the importance of securing accurate body-temperatures in researches of this kind, a number of investigations were made and published elsewhere[15] regarding the body-temperature in connection with the experiments with the respiration calorimeter. It was soon found that the ordinary mercurial clinical thermometer was not best suited for the most accurate observations of body-temperature and a special type of thermometer employing the electrical-resistance method was used. In many of the experiments, however, it is impracticable with new subjects to complicate the experiment by asking them to insert the electrical rectal thermometer, and hence we have been obliged to resort to the usual clinical thermometer with temperatures taken in the mouth, although in a few instances they have been taken in the axilla and the rectum. For the best results the electrical rectal thermometer is used. This apparatus permits a continuous measurement of body temperature, deep in the rectum, unknown to the subject and for an indefinite period of time, it being necessary to remove the thermometer only for defecation.

As a result of these observations it was soon found that the body temperature was not constant from hour to hour, but fluctuated considerably and underwent more or less regular rhythm with the minimum between 3 and 5 o'clock in the morning and the maximum about 5 o'clock in the afternoon. In a number of experiments where the mercurial thermometer was used under the tongue and observations thus taken compared with records with the resistance thermometer, it was found that with careful manipulation and avoiding muscular activity, mouth breathing, and the drinking of hot or cold liquid, a fairly uniform agreement between the two could be obtained. Such comparisons made on laboratory assistants can not be duplicated with the ordinary subject.

It is assumed that fluctuations in temperature measured by the rectal thermometer likewise hold true for the average temperature of the whole body, but evidence on this point is unfortunately not as complete as is desirable. In an earlier report of investigations of this nature, a few experiments on comparison of measurements of resistance thermometer deep in the rectum and in a well-closed axilla showed a distinct tendency for the curves to continue parallel. A research is very much needed at present on a topographical distribution of body temperature, and particularly on the course of the fluctuations in different parts of the body. A series of electric-resistance thermometers placed at different points in the colon, at different points in a stomach tube, in the well-closed axilla, possibly attached to the surface of the body, and in women in the vagina, should give a very accurate picture of the distribution of the body-temperature and likewise indicate the proportionality of the fluctuations in different parts of the body. Until such a research is completed, however, it is necessary to assume that fluctuations in body-temperature as measured by the electric rectal thermometer are a true measure of the average body-temperature of the whole body. Indeed it is upon this assumption that it is necessary for us to make corrections for heat lost from or stored in the body. It is our custom, therefore, to compute the hydrothermal equivalent by multiplying the body-weight by the specific heat of the body, commonly assumed as 0.83, and then to make allowance for fluctuations in body-temperature.

When it is considered that with a subject having a weight of 70 kilos a difference in temperature of 1 deg. C. will make a difference in the measurement of heat of some 60 calories, it is readily seen that the importance of knowing the exact body-temperature can not be overestimated; indeed, the whole problem of the comparison of the direct and indirect calorimetry hinges more or less upon this very point, and it is strongly to be hoped that ere long the much-needed observations on body-temperature can be made.

CONTROL EXPERIMENTS WITH THE CALORIMETER.

After providing a suitable apparatus for bringing away the heat generated inside the chamber and for preventing the loss of heat by maintaining the walls adiabatic, it is still necessary to demonstrate the ability of the calorimeter to measure known amounts of heat accurately. In order to do this we pass a current of electricity of known voltage through a resistance coil and thus develop heat inside the respiration chamber. While, undoubtedly, the use of a standard resistance and potentiometer is the most accurate method for measuring currents of this nature, thus far we have based our experiments upon the measurements made with extremely accurate Weston portable voltmeter and mil-ammeters. Thanks to the kindness of one of our former co-workers, Mr. S. C. Dinsmore, at present associated with the Weston Electrical Instrument Company, we have been able to obtain two especially exact instruments. The mil-ammeter is so adjusted as to give a maximum current of 1.5 amperes and the voltmeter reads from zero to 150 volts. The direct current furnished the building is caused to pass through a variable resistance for adjusting minor variations in voltage and then through the mil-ammeter into a manganin resistance-coil inside the chamber, having a resistance of 84.2 ohms. Two leads from the terminals of the manganin coil connect with the voltmeter outside the chamber, and hence the drop in potential can be measured very accurately and as frequently as is desired. The current furnished the building is remarkably steady, but for the more accurate experiments a small degree of hand regulation is necessary.

The advantage of the electrical method of controlling the apparatus is that the measurements can be made very accurately, rapidly, and in short periods. In making experiments of this nature it is our custom first to place the resistance-coil in the calorimeter and make the connections. The current is then passed through the coil, and simultaneously the water is started flowing through the heat-absorbing system and the whole calorimeter is adjusted in temperature equilibrium as soon as possible. When the temperature of the air and walls is constant and the thermal-junction system in equilibrium, the exact time is noted and the water-current deflected into the meter. At the end of one hour, the usual length of a period, the water-current is deflected from the meter, the meter is weighed, and the average temperature-difference of the water obtained by averaging the results of all the temperature differences noted during the hour. Usually during an experiment of this nature, records of the water-temperatures are made every 4 minutes; occasionally, when the fluctuations are somewhat greater than usual, records are made every 2 minutes.

The calculation of the heat developed in the apparatus is made by means of the formula C x E x t x 0.2385 = calories, in which C equals the current in amperes, E the electromotive force, and t the time in seconds. This gives the heat expressed in calories at 15 deg. C. This procedure we have followed as a result of the recommendation of Dr. E. B. Rosa, of the National Bureau of Standards. In order to convert the values to 20 deg., the unit commonly employed in calorimetric work, it has been necessary to multiply by the ratio of the specific heat of water at 15 deg. to that of water at 20 deg. Assuming the specific heat of water at 20 deg. to be 1, the specific heat at 15 deg. is 1.001.[16]

Of the many electrical check-tests made with this type of apparatus, but one need be given here, pending a special treatment of the method of control of the calorimeter in a forthcoming publication. An electrical check-experiment with the chair calorimeter was made on January 4, 1909, and continued 6 hours. The voltmeter and mil-ammeter were read every few minutes, the water collected in the water-meter, carefully weighed, and the temperature differences as measured on the two mercury thermometers were recorded every 4 minutes.

The heat developed during the experiment may be calculated from the data as follows: Average current = 1.293 amperes; average E. M. F. = 109.15 volts; time = 21,600 seconds; factor used to convert watt-seconds to calories = 0.2385. (1.293 x 109.15 x 21600 x 0.2385) x 1.001 = 727.8 calories produced.

During the 6 hours 237.63 kilograms of water passed through the absorbing system.

The average temperature rise was 3.04 deg. C., the total heat brought away was therefore (237.63 x 3.04) x 1.0024[17] = 724.1 calories.

Thus in 6 hours there were about 3.7 calories more heat developed inside the apparatus than were measured by the water-current, a discrepancy of about 0.5 per cent.

Under ideal conditions of manipulation, the withdrawal of heat from the calorimeter should be at just such a rate as to exactly compensate for the heat developed by the resistance-coil. Under these conditions, then, there would be no heat abstracted from nor stored by the calorimeter and its temperature should remain constant throughout the whole experiment. Practically this is very difficult to accomplish and there are minor fluctuations in temperature above and below the initial temperature during a long experiment and, indeed, during a short experimental period. If a certain amount of heat has been stored up in the calorimeter chamber or has been abstracted from it, there should be corrections made for the variations in the temperature of the chamber. Such corrections are impossible unless a proper determination of the hydrothermal equivalent has been made. A number of experiments to determine this hydrothermal equivalent have been made and the results are recorded beyond, together with a discussion of the nature of the experiments. As a result of these experiments it has been possible to make correction for the slight temperature changes in the calorimeter.

It is interesting to note that these fluctuations are small and there may therefore be a considerable error in the determination of the hydrothermal equivalent without particularly affecting the corrections applied in the ordinary electrical check-test. The greatest difficulty experienced with the calorimeter as a means of measuring heat has been to secure the average temperature of the ingoing water. The temperature difference between the mass of water flowing through the pipes and the outer wall of the pipe is at best considerable. The use of the vacuum-jacketed glass tubes has minimized the loss of heat through this tube considerably, but it is advisable that the bulb of the thermometer be placed exactly in the center of the water-tube, as otherwise too high a temperature-reading will be secured. When the proper precautions are taken to secure the correct temperature-reading, the results are most satisfactory.

In testing both calorimeters a large number of electrical check experiments have led to the conclusion that discrepancies in results were invariably due, not to the loss of heat through the walls of the calorimeter, but to erroneous measurement of the temperature of the water-current.

DETERMINATION OF THE HYDROTHERMAL EQUIVALENT OF THE CALORIMETER.

While the temperature control of the calorimeter is such that in general the average temperature varies but a few hundredths of a degree between the beginning and the end of an experimental period, in extremely accurate work it is necessary to know the amount of heat which is absorbed with any increase in temperature. In other words, the determination of the hydrothermal equivalent is essential.

The large majority of the methods for determining the hydrothermal equivalent of materials are at once eliminated when the nature of the calorimeter here used is taken into consideration. Obviously, in warming up the chamber there are two sources of heat: first, the heat inside of the chamber; second, the heat in the outer walls. As has been previously described, the zinc wall is arbitrarily heated so that its temperature fluctuations will follow exactly those of the inner wall, hence it is impossible to compute from the weight of the metal the hydrothermal equivalent. By means of the electrical check experiments, however, a method for determining the hydrothermal equivalent is at hand. The general scheme is as follows.

During an electrical check experiment, when thermal equilibrium has been thoroughly established and the heat brought away by the water-current exactly counterbalances the heat generated in the resistance-coil inside the chamber, the temperature of the calorimeter is allowed to rise slowly by raising the temperature of the ingoing water and thus bringing away less heat. At the same time the utmost pains are taken to maintain the adiabatic condition of the metal walls. Since the temperature is rising during this period, it is necessary to warm the air in the outer spaces by the electric current. By this method it is possible to raise the temperature of the calorimeter 1 degree or more in 2 hours and establish thermal equilibrium at the higher level. The experiment is then continued for 2 hours at this level, and the next 2 hours the temperature is gradually allowed to fall by lowering the temperature of the ingoing water so that more heat is brought away than is generated, care being taken likewise to keep the walls adiabatic. Under these conditions the heat brought away by the water-current during the period of rising temperature is considerably less than that actually developed by the electric current and the difference represents the amount of heat absorbed by the calorimeter in the period of the temperature rise. Conversely, during the period when the temperature is falling, there is a considerable increase in the amount of heat brought away by the water-current over that generated in the resistance-coil and the difference represents exactly the amount of heat given up by the calorimeter during the fall in temperature. It is thus possible to measure the capacity of the calorimeter for absorbing heat during a rise in temperature and the amount of heat lost by it during cooling. A number of such experiments have been made with both calorimeters and it has been found that the hydrothermal equivalent of the bed calorimeter is not far from 21 kilograms. For the chair calorimeter a somewhat lower figure has been found, i. e., 19.5 kilograms.

GENERAL DESCRIPTION OF RESPIRATION APPARATUS.

This apparatus is designed much after the principle of the Regnault-Reiset apparatus, in that there is a confined volume of air in which the subject lives and which is purified by its passage through vessels containing absorbents for water and carbon dioxide. Fresh oxygen is added to this current of air and it is then returned to the chamber to be respired. This principle, in order to be accurate for oxygen determinations, necessitates an absolutely air-tight system and consequently special precautions have been taken in the construction of the chamber and accessories.

TESTING THE CHAMBER FOR TIGHTNESS.

As already suggested, the walls are constructed of the largest possible sheets of copper with a minimum number of seams and opportunities for leakage. In testing the apparatus for leaks, the greatest precaution is taken. A small air-pressure is applied and the variations in height of a delicate manometer noted. In cases of apparent leakage, all possible sources of leak are gone over with soapsuds when there is a slight pressure on the chamber. As a last resort, which has ultimately proven to be the best method of testing, an assistant goes inside of the chamber, it is then hermetically sealed, and a slight diminished pressure is produced. Ether is then poured about the walls of the chamber and the odor of ether soon becomes apparent inside of the chamber if there is a leakage. Many leaks that could not be found by soapsuds can be readily detected by this method.

VENTILATION OF THE CHAMBER.

The special features of the respiration chamber are the ventilating-pipe system and openings for supplementary apparatus for absorption of water and carbon dioxide. The air entering the chamber is absolutely dry and is directed into the top of the chamber immediately above the head of the subject. The moisture given off from the lungs and skin and the expired gases all tend to mix readily with this dry air as it descends, and the final mixture of gases is withdrawn through an opening near the bottom of the chamber at the front. Under these conditions, therefore, we believe we have a maximum intermingling of the gases. However, even with this system of ventilation, we do not feel that there is theoretically the best mixture of gases, and an electric fan is used inside of the chamber. In experiments where there is considerable regularity in the carbon-dioxide production and oxygen consumption, the system very quickly attains a state of equilibrium, and while the analysis of the outcoming air does not necessarily represent fairly the actual composition of the air inside of the chamber, it evidently represents to the same degree from hour to hour the state of equilibrium that is usually maintained through the whole of a 6-hour experiment.

The interior of the chamber and all appliances are constructed of metal except the chair in which the subject sits. This is of hard wood, well shellacked, and consequently non-porous. With this calorimeter it is desired to make studies regarding the moisture elimination, and consequently it is necessary to avoid the use of all material of a hygroscopic nature. Although the chair can be weighed from time to time with great accuracy and its changes in weight obtained, it is obviously impossible, in any type of experiment thus far made, to differentiate between the water vaporized from the lungs and skin of the man and that from his clothes. Subsequent experiments with a metal chair, with minimum clothing, with cloth of different textures, without clothing, with an oiled skin, and various other modifications affecting the vaporization of water from the body of the man will doubtless throw more definite light upon the question of the water elimination through the skin. At present, however, we resort to the use of a wooden chair, relying upon its changes in weight as noted by the balance to aid us in apportioning the water vaporized between the man and his clothing and the chair.

The walls of the chamber are semi-rigid. Owing to the calorimetric features of this apparatus, it is impracticable to use heavy boiler-plate or heavy metal walls, as the sluggishness of the changes in temperature, the mass of metal, and its relatively large hydrothermal equivalent would interfere seriously with the sensitiveness of the apparatus as a calorimeter. Hence we use copper walls, with a fair degree of rigidity, attached to a substantial structural-steel support; and for all practical purposes the apparatus can be considered as of constant volume. Particularly is this the case when it is considered that the pressure inside of the chamber during an experiment never varies from the atmospheric pressure by more than a few millimeters of water. It is possible, therefore, from the measurements of this chamber, to compute with considerable accuracy the absolute volume. The apparent volume has been calculated to be 1,347 liters.

OPENINGS IN THE CHAMBER.

In order to communicate with the interior of the chamber, maintain a ventilating air-current, and provide for the passage of the current of water for the heat-absorber system and the large number of electrical connections, a number of openings through the walls of the chamber were necessary. The great importance of maintaining this chamber absolutely air-tight renders it necessary to minimize the number of these openings, to reduce their size as much as possible, and to take extra precaution in securing their closure during an experiment. The largest opening is obviously the trap-door at the top through which the subject enters, shown in dotted outline in fig. 7. While somewhat inconvenient to enter the chamber in this way, the entrance from above possesses many advantages. It is readily closed and sealed by hot wax and rarely is a leakage experienced. The trap-door is constructed on precisely the same plan as the rest of the calorimeter, having its double walls of copper and zinc, its thermal-junction system, its heating wires and connections, and its cooling pipes. When closed and sealed, and the connections made with the cooling pipes and heating wires, it presents an appearance not differing from any other portion of the calorimeter.

The next largest opening is the food-aperture, which is a large sheet-copper tube, somewhat flattened, thus giving a slightly oval form, closed with a port, such as is used on vessels. The door of the port consists of a heavy brass frame with a heavy glass window and it can be closed tightly by means of a rubber gasket and two thumbscrews. On the outside is used a similar port provided with a tube somewhat larger in diameter than that connected with the inner port. The annular space between these tubes is filled with a pneumatic gasket which can be inflated and thus a tight closure may be maintained. When one door is closed and the other opened, articles can be placed in and taken out of the chamber without the passage of a material amount of air from the chamber to the room outside or into the chamber from outside.

The air-pipes passing through the wall of the calorimeter are of standard 1-inch piping. The insulation from the copper wall is made by a rubber stopper through which this piping is passed, the stopper being crowded into a brass ferule which is stoutly soldered to the copper wall. This is shown in detail in fig. 25, in which N is the brass ferule and M the rubber stopper through which the air-pipe passes. The closure is absolutely air-tight and a minimum amount of heat is conducted out of the chamber, owing to the insulation of the rubber stopper M. The water-current enters and leaves the chamber through two pipes insulated in two similar brass ferules soldered to the copper and zinc walls. The insulation between the water-pipe and the brass ferule has been the subject of much experimenting and is discussed on page 24. The best insulation was secured by a vacuum-jacketed glass tube, although the special hard-rubber tubes surrounding the electric-resistance thermometers have proven very effective as insulators in the bed calorimeter.

A series of small brass tubes, from 10 to 15 millimeters in diameter, are soldered into the copper wall in the vicinity of the water-pipes. These are used for electrical connections and for connections with the manometer, stethoscope, and pneumograph. All of these openings are tested carefully and shown to be absolutely air-tight before being put in use.

In the dome of the calorimeter, and directly over the head of the subject, is the opening for the weighing apparatus. This consists of a hard-rubber tube, threaded at one end and screwed into a brass flange heavily soldered to the copper wall (fig. 9). When not in use, a solid rubber stopper on a brass rod is drawn into this opening, thus producing an air-tight closure. When in actual use during the process of weighing, a thin rubber diaphragm prevents leakage of air through this opening. The escape of heat through the weighing-tube is minimized by having this tube of hard rubber.

VENTILATING AIR-CURRENT.



The ventilating air-current is so adjusted that the air which leaves the chamber is caused to pass through purifiers, where the water-vapor and the carbon dioxide are removed, and then, after being replenished with fresh oxygen, it is returned to the chamber ready for use. The general scheme of the respiration apparatus is shown in fig. 27. The air leaving the chamber contains carbon dioxide and water-vapor and the original amount of nitrogen and is somewhat deficient in oxygen. In order to purify the air it must be passed through absorbents for carbonic acid and water-vapor and hence some pressure is necessary to force the gas through these purifying vessels. This pressure is obtained by a small positive rotary blower, which has been described previously in detail.[18] The air is thus forced successively through sulphuric acid, soda or potash-lime, and again sulphuric acid. Finally it is directed back to the respiration chamber free from carbon dioxide and water and deficient in oxygen. Pure oxygen is admitted to the chamber to make up the deficiency, and the air thus regenerated is breathed again by the subject.

BLOWER.

The rotary blower used in these experiments for maintaining the ventilating current of air has given the greatest satisfaction. It is a so-called positive blower and capable of producing at the outlet considerable pressure and at the inlet a vacuum of several inches of mercury. At a speed of 230 revolutions per minute it delivers the air at a pressure of 43 millimeters of mercury, forcing it through the purifying vessels at the rate of 75 liters per minute. This rate of ventilation has been established as being satisfactory for all experiments and is constant. Under the pressure of 43 millimeters of mercury there are possibilities of leakage of air from the blower connections and hence, to note this immediately, the blower system is immersed in a tank filled with heavy lubricating oil. The connections are so well made, however, that leakage rarely occurs, and, when it does, a slight tightening of the stuffing-box on the shaft makes the apparatus tight again.

ABSORBERS FOR WATER-VAPOR.

To absorb 25 to 40 grams of water-vapor in an hour from a current of air moving at the rate of 75 liters per minute and leaving the air essentially dry under these conditions has been met by the apparatus herewith described. The earlier attempts to secure this result involved the use of enameled-iron soup-stock pots, fitted with special enameled-iron covers and closed with rubber gaskets. For the preliminary experimenting and for a few experiments with man these proved satisfactory, but in spite of their resistance to the action of sulphuric acid, it was found that they were not as desirable as they should be for continued experimenting from year to year. Recourse was then had to a special form of chemical pottery, glazed, and a type that usually gives excellent satisfaction in manufacturing concerns was used.

This special form of absorbers presented many difficulties in construction, but the mechanical difficulties were overcome by the potter's skill and a number of such vessels were furnished by the Charles Graham Chemical Pottery Works. Here again these vessels served our purpose for several months, but unfortunately the glaze used did not suffice to cover them completely and there was a slight, though persistent, leakage of sulphuric acid through the porous walls. To overcome this difficulty the interior of the vessels was coated with hot paraffin after a long-continued washing to remove the acid and after they had been allowed to dry thoroughly. The paraffin-treated absorbers continued to give satisfaction, but it was soon seen that for permanent use something more satisfactory must be had. After innumerable trials with glazed vessels of different kinds of pottery and glass, arrangements were made with the Royal Berlin Porcelain Works to mold and make these absorbers out of their highly resistant porcelain. The result thus far leaves nothing to be desired as a vessel for this purpose. A number of such absorbers were made and have been constantly used for a year and are absolutely without criticism.

Fig. 28 shows the nature of the interior of the apparatus. The air enters through one opening at the top, passes down through a bent pipe, and enters a series of roses, consisting of inverted circular saucers with holes in the rims. The position of the holes is such that when the vessel is one-fourth to one-third full of sulphuric acid the air must pass through the acid three times. To prevent spattering, a small cup-shaped arrangement, provided with holes, is attached to the opening through which the air passes out of the absorber, and for filling the vessel with acid a small opening is made near one edge. The specifications required that the apparatus should be made absolutely air-tight to pressures of over 1 meter of water, and that there is no porosity in these vessels under these conditions is shown by the fact that such a pressure is held indefinitely. The inside and outside are both heavily glazed. There is no apparent action of sulphuric acid on the vessels and the slight increase in temperature resulting from the absorption of water-vapor as the air passes through does not appear to have any deleterious effect.



The vessels without filling and without rubber elbows weigh 11.5 kilograms; with the special elbows and couplings attached so as to enable them to be connected with the ventilating air-system, the empty absorbers weigh 13.4 kilograms; and filled with sulphuric acid they weigh 19 kilograms. Repeated tests have shown that 5.5 kilograms of sulphuric acid will remove the water-vapor from a current of air passing through the absorbers at the rate of 75 liters of air per minute, without letting any appreciable amount pass by until 500 grams of water have been absorbed. At this degree of saturation a small persistent amount of moisture escapes absorption in the acid and consequently a second absorber will begin to gain in weight. Experiments demonstrate that the first vessel can gain 1,500 grams of water before the second gains 5 grams. As a matter of fact, it has been found more advantageous to use but one absorber and have it refilled as soon as it has gained 400 grams, thus allowing a liberal factor of safety and no danger of loss of water.

POTASH-LIME CANS.

The problem of absorbing the water-vapor from so rapid a current of air is second only to that of absorbing the carbon dioxide from such a current. All experiments with potassium hydroxide in the form of sticks or in solution failed to give the desired results and the use of soda-lime has supplemented all other forms of carbon dioxide absorption. More recently we have been using potash-lime, substituting caustic potash for caustic soda in the formula, and the results thus obtained are, if anything, more satisfactory than with the soda-lime.

The potash-lime is made as follows: 1 kilogram of commercial potassium hydroxide, pulverized, is dissolved in 550 to 650 cubic centimeters of water and 1 kilogram of pulverized quicklime added slowly. The amount of water to be used varies with the moisture content of the potash. There is a variation in the moisture content of different kegs of potash, so when a keg is opened we determine experimentally the amount of water to be used. After a batch is made up in this way it should be allowed to cool before testing whether it has the right amount of water, and this is determined by feeling of it and noting how it pulverizes in the hand. It is not advisable to make a great quantity at once, because we have found that if a large quantity is made and broken into small particles and stored in a container it has a tendency to cake and thus interfere with its ready subsequent use.

A record was kept of the gains in weight of a can filled with potash-lime during a series of experiments where there were three silver-plated cans used. This can was put at the head of the system and when it began to lose weight it was removed. The records of gains of weight when added together amount to 400 grams. From experience with other cans where the loss of moisture was determined, it is highly probable that at least 200 grams of water were vaporized from the reagent and thus the total amount of carbon dioxide absorbed must have been not far from 600 grams. At present our method is not to allow the cans to gain a certain weight, but during 4-hour or 5-hour experiments, in which each can may be used 2 or 3 hours, it is the practice to put a new can on each side of the absorber system (see page 66) at the beginning of every experiment. This insures the same power of absorption on each side of the absorption system so that the residual amount of carbon dioxide in the chamber from period to period does not undergo very marked changes. This has been found the best method, because if one can is left on a day longer than the other there is apt to be alternately a rise and fall in the amount of residual carbon dioxide in the apparatus, owing to the unequal efficiency of the absorbers.

These cans are each day taken to the basement, where the first section[19] only is taken out and replaced with new potash-lime. Thus, three-quarters of the contents of the can is used over and over, while the first quarter is freshly renewed every day. Potash-lime has not been found practicable for the U-tubes because one can not, as in the case of soda-lime, see the whitening of the reagent where the carbon dioxide is absorbed.

The importance of having the soda-lime or potash-lime somewhat moist, to secure the highest efficiency for the absorption of the carbon dioxide, makes it necessary to absorb the moisture taken up by the dry air in passing through the potash-lime can. Consequently a second vessel containing sulphuric acid is placed in the system to receive the air immediately after it leaves the potash-lime can. Obviously the amount of water absorbed here is very much less than in the first acid absorber and hence the same absorber can be used for a greater number of experiments.

BALANCE FOR WEIGHING ABSORBERS.

The complete removal of water-vapor and carbon dioxide from a current of air moving at the rate of 75 liters per minute calls for large and somewhat unwieldy vessels in which is placed the absorbing material. This is particularly the case with the vessels containing the rather large amounts of sulphuric acid required to dry the air. In the course of an hour there is ordinarily removed from the chamber not far from 25 grams of water-vapor and 20 to 30 grams of carbon dioxide. This necessitates weighing the absorbers to within 0.25 gram if an accuracy of 1 per cent is desired. The sulphuric-acid absorbers weigh about 18 kilograms when filled with acid. In order to weigh this receptacle so as to measure accurately the increase in weight due to the absorption of water to within less than 1 per cent, we use the balance shown in fig. 29. This balance has been employed in a number of other manipulations in connection with the respiration calorimeter and accessory apparatus and the general type of balance leaves nothing to be desired as a balance capable of carrying a heavy load with remarkable sensitiveness.

The balance is rigidly mounted on a frame consisting of four upright structural-steel angle-irons, fastened at the top to a substantial wooden bed. Two heavy wooden pieces run the length of the table and furnish a substantial base to which the standard of the balance is bolted. The balance is surrounded by a glass case to prevent errors due to air-currents (see fig. 2). The pan of the balance is not large enough to permit the weighing of an absorber, hence provision is made for suspending it on a steel or brass rod from one of the hanger arms. This rod passes through a hole in the bottom of the balance case, and its lower end is provided with a piece of pipe having hooks at either end. Since the increase in weight rather than the absolute weight of the absorber is used, the greater part of the weight is taken up by lead counterpoises suspended above the pan on the right-hand arm of the balance. The remainder of the weight is made up with brass weights placed in the pan.



In order to suspend this heavy absorber, a small elevator has been constructed, so that the vessel may be raised by a compressed-air piston. This piston is placed in an upright position at the right of the elevator and is connected with the compressed-air service of the building. The pressure is about 25 pounds per square inch and the diameter of the cylinder is 2.5 inches, thus giving ample service for raising and lowering the elevator and its load. By turning a 3-way valve at the end of the compressed-air supply-pipe, so that the air rushes into the cylinder above the piston, the piston is pushed to the base of the cylinder and the elevator thereby raised. The pressure of the compressed air holds the elevator in this position while the hooks are being adjusted on the absorber. By turning the 3-way valve so as to open the exhaust leading to the upper part of the cylinder to the air, the weight of the elevator expels the air, and it soon settles into the position shown in the figure. The weighing can then be made as the absorber is swinging freely in the air. After the weighing has been made, the elevator is again lifted, the hooks are released, and by turning the valve the elevator and load are safely lowered.

The size of the openings of the pipes into the cylinder is so adjusted that the movement of the elevator is regular and moderate whether it is being raised or lowered, thus avoiding any sudden jars that might cause an accident to the absorbers. With this system it is possible to weigh these absorbers to within 0.1 gram and, were it necessary, probably the error could be diminished so that the weight could be taken to 0.05 gram. On a balance of this type described elsewhere,[20] weighings could be obtained to within 0.02 gram. For all practical purposes, however, we do not use the balance for weighing the absorbers closer than to within 0.10 gram. In attempting to secure accuracy no greater than this, it is unnecessary to lower the glass door to the balance case or, indeed, to close the two doors to the compartment in which the elevator is closed, as the slight air-currents do not affect the accuracy of the weighing when only 0.1 gram sensitiveness is required.

PURIFICATION OF THE AIR-CURRENT WITH SODIUM BICARBONATE.

As is to be expected, the passage of so large a volume of air through the sulphuric acid in such a relatively small space results in a slight acid odor in the air-current leaving this absorber. The amount of material thus leaving the absorber is not weighable, as has been shown by repeated tests, but nevertheless there is a sufficiently irritating acid odor to make the air very uncomfortable for subsequent respiration. It has been found that this odor can be wholly eliminated by passing the air through a can containing cotton wool and dry sodium bicarbonate. This can is not weighed, and indeed, after days of use, there is no appreciable change in its weight.

VALVES.

In order to subdivide experiments into periods as short as 1 or 2 hours, it is necessary to deflect the air-current at the end of each period from one set of purifiers to the other, in order to weigh the set used and to measure the quantity of carbon dioxide and water-vapor absorbed. The conditions under which these changes from one system to another are made, and which call for an absolutely gas-tight closure, have been discussed in detail elsewhere.[21] It is sufficient to state here that the very large majority of mechanical valves will not serve the purpose, since it is necessary to have a pressure of some 40 millimeters of mercury on one side of the valve at the entrance to the absorber system and on the other side atmospheric pressure. A valve with an internal diameter of not less than 25 millimeters must be used, and to secure a tight closure of this large area and permit frequent opening and shutting is difficult. After experimenting with a large number of valves, a valve of special construction employing a mechanical seal ultimately bathed in mercury was used for the earlier apparatus. The possibility of contamination of the air-current by mercury vapor was duly considered and pointed out in a description of this apparatus. It was not until two years later that difficulties began to be experienced and a number of men were severely poisoned while inside the chamber. A discussion of this point has been presented elsewhere.[22] At that time mercury valves were used both at the entrance and exit ends of the absorber system, although as a matter of fact, when the air leaves the last absorber and returns to the respiration chamber, the pressure is but a little above that of the atmosphere. Consequently, mechanical valves were substituted for mercurial valves at the exit and the toxic symptoms disappeared. In constructing the new calorimeters it seemed to be desirable to avoid all use of mercury, if possible. We were fortunate in finding a mechanical valve which suited this condition perfectly. These valves, which are very well constructed, have never failed to show complete tightness under all possible tests and are used at the exit and entrance end of the absorber system. Their workmanship is of the first order, and the valve is somewhat higher in price than ordinary mechanical valves. They have been in use on the apparatus for a year now and have invariably proved to be absolutely tight. They are easy to obtain and are much easier to manipulate and much less cumbersome than the mercury valves formerly used.

COUPLINGS.

Throughout the construction of the respiration apparatus and its various parts, it was constantly borne in mind that the slightest leak would be very disastrous for accurate oxygen determinations. At any point where there is a pressure greater or less than that of the atmosphere, special precaution must be taken. At no point in the whole apparatus is it necessary to be more careful than with the couplings which connect the various absorber systems with each other and with the valves; for these couplings are opened and closed once every hour or two and hence are subject to considerable strain at the different points. If they are not tight the experiment is a failure so far as the determination of oxygen is concerned. For the various parts of the absorber system we have relied upon the original type of couplings used in the earlier apparatus. A rubber gasket is placed between the male and female part of the coupling and the closure can be made very tight. In fact, after the absorbers are coupled in place they are invariably subjected to severe tests to prove tightness.

For connecting the piping between the calorimeter and the absorption system we use ordinary one-inch hose-couplings, firmly set up by means of a wrench and disturbed only when necessary to change from one calorimeter chamber to another.

ABSORBER TABLE.

The purifying apparatus for the air-current is compactly and conveniently placed on a solidly constructed table which can be moved about the laboratory at will. The special form of caster on the bottom of the posts of the table permits its movement about the laboratory at will and by screwing down the hand screws the table can be firmly fixed to the floor.

The details of the table are shown in fig. 30. (See also fig. 4, page 4.) The air coming from the calorimeter passes in the direction of the downward arrow through a 3/4-inch pipe into the blower, which is immersed in oil in an iron box F. The blower is driven by an electric motor fastened to a small shelf at the left of the table. The air leaving the blower ascends in the direction of the arrow to the valve system H, where it can be directed into one of the two parallel sets of purifiers; after it passes through these purifiers (sulphuric-acid vessel 2, potash-lime container K, and sulphuric-acid vessel 1) it goes through the sodium-bicarbonate can G to a duplicate valve system on top of the table. From there it passes through a pipe along the top of the table and rises in the vertical pipe to the hose connection which is coupled with the calorimeter chamber.

The electric motor is provided with a snap-switch on one of the posts of the table and a regulating rheostat which permits variations in the speed of the motor and consequently in the ventilation produced by the blower. The blower is well oiled, and as oil is gradually carried in with the air, a small pet-cock at the bottom of the T following the blower allows any accumulated oil to be drawn away from time to time. The air entering the valve system at H enters through a cross, two arms of which connect with two "white star" valves. The upper part of the cross is connected to a small rubber tubing and to the mercury manometer D, which also serves as a valve for passing a given amount of air through a series of U-tubes for analysis of the air from time to time. It is assumed that the air drawn at the point H is of substantially the same composition as that inside the chamber, an assumption that may not be strictly true, but doubtless the sample thus obtained is constantly proportional to the average composition, which fluctuates but slowly. Ordinarily the piping leading from the left-hand arm of the tube D is left open to the air and consequently the difference in the level of the mercury in the two arms of D indicates the pressure on the system. This is ordinarily not far from 40 to 50 millimeters of mercury.



The absorber table, with the U-tubes and meter for residual analyses, is shown in the foreground in fig. 2. The two white porcelain vessels with a silver-plated can between them are on the middle shelf. The sodium bicarbonate can, for removing traces of acid fumes, is connected in an upright position, while the motor, the controlling rheostat, and the blower are supported by the legs near the floor. The two rubber pipes leading from the table can be used to connect the apparatus either with the bed or chair calorimeter. In fig. 4 the apparatus is shown connected with the bed calorimeter, but just above the lowest point of the rubber tubing can be seen in the rear the coupling for one of the pipes leading from the chair calorimeter. The other is immediately below and to the left of it.

OXYGEN SUPPLY.

The residual air inside of the chamber amounts to some 1,300 liters and contains about 250 liters of oxygen. Consequently it can be seen that in an 8-hour experiment the subject could easily live during the entire time upon the amount of oxygen already present in the residual air. It has been repeatedly shown that until the per cent of oxygen falls to about 11, or about one-half normal, there is no disturbance in the respiratory exchange and therefore about 125 liters of oxygen would be available for respiration even if no oxygen were admitted. Inasmuch as the subject when at rest uses not far from 14 to 15 liters per hour, the amount originally present in the chamber would easily suffice for an 8-hour experiment. Moreover, the difficulties attending an accurate gas analysis and particularly the calculation of the total amount of oxygen are such that satisfactory determinations of oxygen consumption by this method would be impossible. Furthermore, from our previous experience with long-continued experiments of from 10 days to 2 weeks, it has been found that oxygen can be supplied to the system readily and the amount thus supplied determined accurately. Consequently, even in these short experiments, we adhere to the original practice of supplying oxygen to the air and noting the amount thus added.

The oxygen supply was formerly obtained from small steel cylinders of the highly compressed gas. This gas was made by the calcium-manganate method and represented a high degree of purity for commercial oxygen. More recently we have been using oxygen of great purity made from liquid air. Inasmuch as this oxygen is very pure and much less expensive than the chemically-prepared oxygen, extensive provisions have been made for its continued use. Instead of using small cylinders containing 10 cubic feet and attaching thereto purifying devices in the shape of soda-lime U-tubes and a sulphuric-acid drying-tube, we now use large cylinders and we have found that the oxygen from liquid air is practically free from carbon dioxide and water-vapor, the quantities present being wholly negligible in experiments such as these. Consequently, no purifying attachments are considered necessary and the oxygen is delivered directly from the cylinder. The cylinders, containing 100 cubic feet (2,830 liters), under a pressure of 120 atmospheres, are provided with well-closing valves and weigh when fully charged 57 kilograms.



It is highly desirable to determine the oxygen to within 0.1 gram, and we are fortunate in having a balance of the type used frequently in this laboratory which will enable us to weigh this cylinder accurately with a sensitiveness of less than 0.1 gram. Since 1 liter of oxygen weighs 1.43 grams, it can be seen that the amount of oxygen introduced into the chamber can be measured by this method within 70 cubic centimeters. Even in experiments of but an hour's duration, where the amount of oxygen admitted from the cylinder is but 25 to 30 grams, it can be seen that the error in the weighing of the oxygen is much less than 1 per cent.

The earlier forms of cylinders used were provided with valves which required some special control and a rubber bag was attached to provide for any sudden rush of gas. The construction of the valve and valve-stem was unfortunately such that the well-known reduction valves could not be attached without leakage under the high pressure of 120 atmospheres. With the type of cylinder at present in use, such leakage does not occur and therefore we simply attach to the oxygen cylinder a reduction-valve which reduces the pressure from 120 atmospheres to about 2 or 3 pounds to the square inch. The cylinder, together with the reduction valve, is suspended on one arm of the balance. The equipment of the arrangement is shown in fig. 31. (See also fig. 5, page 4.) The cylinder is supported by a clamp K hung from the balance arm, and the reduction-valve A is shown at the top. The counterpoise S consists of a piece of 7-inch pipe, with caps at each end. At a convenient height a wooden shelf with slightly raised rim is attached.

In spite of the rigid construction of this balance, it would be detrimental to allow this enormous weight to remain on the knife-edges permanently, so provision is made for raising the cylinders on a small elevator arrangement which consists of small boxes of wood, T, into which telescope other boxes, T'. A lever handle, R, when pressed forward, raises T' by means of a roller bearing U, and when the handle is raised the total weight of the cylinders is supported on the platforms.

The balance is attached to an upright I-beam which is anchored to the floor and ceiling of the calorimeter laboratory. Two large turnbuckle eye-bolts give still greater rigidity at the bottom. The whole apparatus is inclosed in a glass case, shown in fig. 5.

AUTOMATIC CONTROL OF OXYGEN SUPPLY.

The use of the reduction-valve has made the automatic control of the oxygen supply much simpler than in the apparatus formerly used. The details of the connections somewhat schematically outlined are given in fig. 32, in which D is the oxygen cylinder, K the supporting band, A the reduction-valve, and J the tension-equalizer attached to one of the calorimeters. Having reduced the pressure to about 2 pounds by means of the reduction-valve, the supply of oxygen can be shut off by putting a pinch-cock on a rubber pipe leading from the reduction-valve to the calorimeters. Instead of using the ordinary screw pinch-cock, this connection is closed by a spring clamp. The spring E draws on the rod which is connected at L and pinches the rubber tube tightly. The tension at E can be released by an electro-magnet F, which when magnetized exercises a pull on the iron rod, extends the spring E, and simultaneously releases the pressure on the rubber tube at L. To make the control perfectly automatic, the apparatus shown on the top of the tension-equalizer J is employed. A wire ring, with a wire support, is caused to pass up through a bearing fastened to the clamp above J. As the air inside of the whole system becomes diminished in volume and the rubber cap J sinks, there is a point at which a metal loop dips into two mercury cups C and C', thus closing the circuit, which causes a current of electricity to pass through F. This releases the pressure at L, oxygen rushes in, and the rubber bag J becomes distended. As it is distended, it lifts the metal loop out of the cups, C and C', and the circuit is broken. There is, therefore, an alternate opening and closing of this circuit with a corresponding admission of oxygen. The exact position of the rubber diaphragm can be read when desired from a pointer on a graduated scale attached to a support holding the terminals of the electric wires. More frequently, however, when the volume is required, instead of filling the bag to a definite point, as shown by the pointer, a delicate manometer is attached to the can by means of a pet-cock and the oxygen is admitted by operating the switch B until the desired tension is reached.

[Illustration: FIG. 32.—Part of the oxygen cylinder and connections to tension-equalizer. At the left is shown the upper half of the oxygen cylinder with a detail of the electro-magnet and reducing-valve. D is the cylinder; K, the band supporting the oxygen cylinder and electro-magnet arrangement; F, the electro-magnet; E, the tension spring; and L, the rubber tubing at a point where it is closed by the clamp. The tension-equalizer and the method of closing the circuit operating it are shown at the right. C and C' are two mercury cups into which the wire loop dips, thus closing the circuit. B is a lever used for short-circuiting for filling the diaphragm J. G is a sulphuric-acid container; H, the quick-throw valve for shutting off the tension equalizer J; M, part of the ingoing air-pipe; N, a plug connecting the electric circuit with the electro-magnet; and O, a storage battery.]

In order to provide for the maximum sensitiveness for weighing D and its appurtenances, the electric connection is broken at the cylinder by means of the plug N and the rubber tube is connected by a glass connector which can be disconnected during the process of weighing. Obviously, provision is also made that there be no leakage of air out of the system during the weighing. The current at F is obtained by means of a storage battery O. The apparatus has been in use for some time in the laboratory and has proved successful in the highest degree.

TENSION-EQUALIZER.

The rigid walls of the calorimeter and piping necessitate some provision for minor fluctuations in the absolute volume of air in the confined system. The apparatus was not constructed to withstand great fluctuations in pressure, and thin walls were used, but it is deemed inadvisable to submit it even to minor pressures, as thus there would be danger of leakage of air through any possible small opening. Furthermore, as the carbon dioxide and water-vapor are absorbed out of the air-current, there is a constant decrease in volume, which is ordinarily compensated by the admission of oxygen. It would be very difficult to adjust the admission of oxygen so as to exactly compensate for the contraction in volume caused by the absorption of water-vapor and carbon dioxide. Consequently it is necessary to adjust some portion of the circulating air-current so that there may be a contraction and expansion in the volume without producing a pressure on the system. This was done in a manner similar to that described in the earlier apparatus, but on a much simpler plan.

To the air-pipe just before it entered the calorimeter was attached a copper can with a rubber diaphragm top. This diaphragm, which is, as a matter of fact, a ladies' pure rubber bathing-cap, allows for an expansion or contraction of air in the system of 2 to 3 liters. The apparatus shown in position is to be seen in fig. 25, in which the tin can I is covered with the rubber diaphragm J. If there is any change in volume, therefore, the rubber diaphragm rises or falls with it and under ordinary conditions of an experiment this arrangement results in a pressure in the chamber approximately that of the atmosphere. It was found, however, that even the slight resistance of the piping from the tension-equalizer to the chamber, a pipe some 26 millimeters in diameter and 60 centimeters long, was sufficient to cause a slightly diminished pressure inside the calorimeter, inasmuch as the air was sucked out by the blower with a little greater speed than it was forced in by the pressure at the diaphragm. Accordingly the apparatus has been modified so that at present the tension-equalizer is attached directly to the wall of the calorimeter independent of the air-pipe.

In most of the experiments made thus far it has been our custom to conduct the supply of fresh oxygen through pet-cock K on the side of the tension-equalizer. This is shown more in detail in fig. 32, in which, also, is shown the interior construction of the can. Owing to the fact that the air inside of this can is much dryer than the room air, we have followed the custom with the earlier apparatus of placing a vessel containing sulphuric acid inside the tension-equalizer, so that any moisture absorbed by the dry air inside the diaphragm may be taken up by the acid and not be carried into the chamber. The air passing through the pipe to the calorimeter is, it must be remembered, absolutely dry and hence there are the best conditions for the passage of moisture from the outside air through the diaphragm to this dry air. Attaching the tension-equalizer directly to the calorimeter obviates the necessity for this drying process and hence the sulphuric-acid vessel has been discarded.

The valve H (fig. 25) is used to cut off the tension-equalizer completely from the rest of the system at the exact moment of the end of the experimental period. After the motor has been stopped and the slight amount of air partly compressed in the blower has leaked back into the system, and the whole system is momentarily at equal tension, a process occupying some 3 or 4 seconds, the gate-valve H is closed. Oxygen is then admitted from the pet-cock K until there is a definite volume in J as measured by the height to which the diaphragm can rise or a second pet-cock is connected to the can I and a delicate petroleum manometer attached in such a manner that the diaphragm can be filled to exactly the same tension each time. Under these conditions, therefore, the apparent volume of air in the system, exclusive of the tension-equalizer, is always the same, since it is confined by the rigid walls of the calorimeter and the piping. Furthermore, the apparent volume of air in the tension-equalizer is arbitrarily adjusted to be the same amount at the end of each period by closing the valve and introducing oxygen until the tension is the same.

BAROMETER.

Recognizing the importance of measuring very accurately the barometric pressure, or at least its fluctuations, we have installed an accurate barometer of the Fortin type, made by Henry J. Green. This is attached to the inner wall of the calorimeter laboratory, and since the calorimeter laboratory is held at a constant temperature, temperature corrections are unnecessary, for we have here to deal not so much with the accurate measurement of the actual pressure as with the accurate measurement of differences in pressure. For convenience in reading, the ivory needle at the base of the instrument and the meniscus are well illuminated with electric lamps behind a white screen, and a small lamp illuminates the vernier. The barometer can be read to 0.05 millimeter.

ANALYSIS OF RESIDUAL AIR.

The carbon-dioxide production, water-vapor elimination, and oxygen absorption of the subject during 1 or 2 hour periods are recorded in a general way by the amounts of carbon dioxide and water-vapor absorbed by the purifying vessels and the loss of weight of the oxygen cylinder; but, as a matter of fact, there may be considerable fluctuations in the amounts of carbon dioxide and water-vapor and particularly oxygen in the large volume of residual air inside the chamber. With carbon dioxide and water-vapor this is not as noticeable as with oxygen, for in the 1,300 liters of air in the chamber there are some 250 liters of oxygen, and slight changes in the composition of this air indicate considerable changes in the amount of oxygen. Great changes may also take place in the amounts of carbon dioxide and water-vapor under certain conditions. In some experiments, particularly where there are variations in muscular activity from period to period, there may be a considerable amount of carbon dioxide in the residual air and during the next period, when the muscular activity is decreased, for example, the percentage composition of the air may vary so much as to indicate a distinct fall in the amount of carbon dioxide present. Under ordinary conditions of ventilation during rest experiments the quantity of carbon dioxide present in the residual air is not far from 8 to 10 grams. There are usually present in the air not far from 6 to 9 grams of water-vapor, and hence this residual amount can undergo considerable fluctuations. When it is considered that an attempt is made to measure the total amount of carbon dioxide expired in one hour to the fraction of a gram, it is obvious that fluctuations in the composition of residual air must be taken into consideration.

It is extremely difficult to get a fair sample of air from the chamber. The air entering the chamber is free from water-vapor and carbon dioxide. In the immediate vicinity of the entering air-tube there is air which has a much lower percentage of carbon dioxide and water-vapor than the average, and on the other hand close to the nose and mouth of the subject there is air of a much higher percentage of carbon dioxide and water-vapor than the average. It has been assumed that the composition of the air leaving the chamber represents the average composition of the air in the chamber. This assumption is only in part true, but in rest experiments (and by far the largest number of experiments are rest experiments) the changes in the composition of the residual air are so slow and so small that this assumption is safe for all practical purposes.

Another difficulty presents itself in the matter of determining the amount of carbon dioxide and water-vapor; that is, to make a satisfactory analysis of air without withdrawing too great a volume from the chamber. The difficulty in analysis is almost wholly confined to the determination of water-vapor, for while there are a large number of methods for determining small amounts of carbon dioxide with great accuracy, the method for determining water-vapor to be accurate calls for the use of rather large quantities of air. From preliminary experiments with a sling psychrometer it was found that its use was precluded by the space required to successfully use this instrument, the addition of an unknown amount of water to the chamber from the wet bulb, and the difficulties of reading the instrument from without the chamber. Recourse was had to the determination of moisture by the absolute method, in that a definite amount of air is caused to pass over pumice-stone saturated with sulphuric acid. It is of interest here to record that at the moment of writing a series of experiments are in progress in which an attempt is being made to use a hair hygrometer for this purpose.

The method of determining the water-vapor and carbon dioxide in the residual air is extremely simple, in that a definite volume of air is caused to pass over sulphuric acid and soda-lime contained in U-tubes. In other words, a small amount of air is caused to pass through a small absorbing-system constructed of U-tubes rather than of porcelain vessels and silver-plated cans. Formerly a very elaborate apparatus was employed for aspirating the air from the chamber through U-tubes and then returning the aspirated air to the chamber. This involved the use of a suction-pump and called for a special installation for maintaining the pressure of water constant. More recently a much simpler device has been employed, in that we have taken advantage of the pressure in the ventilating air-system developed by the passage of air through the blower. After forcing a definite quantity of air through the reagents in the U-tubes, it is then conducted back to the system after having been measured in a gas-meter.

This procedure is best noted from fig. 30. The connected series of three U-tubes on the rack on the table is joined on one end by well-fitting rubber connections to the tube leading from the mercurial manometer and on the other end to the rubber tube A leading to the gas-meter. On lowering the mercury reservoir E, the mercury is drained out of the tube D and air passes through both arms of the tube and then through the three U-tubes. In the first of these it is deprived of moisture, and in the last two of carbon dioxide. The air then enters the meter, where it is measured and leaves the meter through the tube B, saturated with water-vapor at the room temperature. To remove this water-vapor the air is passed through a tower filled with pumice-stone drenched with sulphuric acid. It leaves the tower through the tube C and enters the ventilating air-pipe on its way to the calorimeter.