METHODS OF MEASUREMENT. 23 



accurate thermometer, which was read at the beginning and end of 

 each experiment. The average of these readings was used for the 

 slight correction of the volume of oxygen due to temperature changes 

 of the meter. The meter was calibrated 1 by passing weighed amounts 

 of oxygen through it at various rates, and the factors thus obtained 

 were used in all subsequent computations. Several calibrations were 

 made during the course of the investigation with approximately uni- 

 form results, which were but the fraction of 1 per cent high. This 

 meter proved very satisfactory, especially the integrating feature, 

 which eliminated the danger, ever-present with other types of meters, 

 of failure to record accurately the total number of revolutions of the 

 pointer. The oxygen used in the experiments was made and supplied 

 by the Linde Air Products Company. 



Barometer. In the 12-minute periods of this research the barometric 

 changes under ordinary atmospheric conditions were almost negligible. 

 For measuring such changes a barograph was used, checked by a re- 

 liable barometer, and read to 0.1 mm. These readings were made at 

 the beginning and end of each period, the average being used for 

 reducing the data for the oxygen consumption to standard conditions. 



Kymograph. The usual Porter kymograph 2 was employed, but in 

 place of records on smoked paper, pen-and-ink tracings on an unsmoked 

 glazed paper were obtained. The pen, which was fastened to the 

 counterpoise of the spirometer-bell, was a small glass capillary pen 

 with platinum tip, such as is employed on many forms of automatic 

 reading devices in large steam and electric plants. 3 This method was 

 found to be simpler and cleaner than working with smoked paper and 

 a varnishing solution. The speed of the kymograph-drum was 

 1 revolution in 15 minutes. A small pen, such as is used with the 

 ordinary barograph, was attached to a signal magnet connected with 

 the clock, and a record thus obtained in minutes on the glazed paper. 



Respiration counter. The number of respirations during the period 

 was at first counted from the tracings of the pen attached to the spiro- 

 meter-bell. Later a device (see fig. 3) was attached to the arm of the 

 spirometer (see A, fig. 3). The cord B, leading to the counterpoise 

 of the spirometer-bell, by rubbing on the fiber sleeve C caused the 

 platinum points E and E\ to dip into the mercury cups D and DI and 

 complete a circuit to a counting device known in the telephone trade 

 as a "p. b. x. message register." (See fig. 4.) This device proved a 

 great labor-saver in counting the respiration tracings on the kymo- 

 graph. 



Benedict, Phys. Review, 1906, 22, p. 294. 



'Harvard Apparatus Company, Dover, Massachusetts. 



3 The pen which gave the most satisfactory results was the Cochrane pen, supplied by the 

 Harrison Safety Boiler Company of Philadelphia, Pennsylvania. It was about 40 mm. long and 

 6 mm. in diameter. With ordinary care the pen withstood a remarkable amount of use before 

 it was worn out. 



