CRITICAL DISCUSSION OF RESPIRATION APPARATUS. 231 



The leaks in the various portions of the apparatus can be easily con- 

 trolled; consequently the possibility of an error from this cause is not 

 a serious disadvantage. If the apparatus has not been properly con- 

 structed, leaks occasionally occur in weak portions; these may be due to 

 a defect in the rubber connections or to imperfections in the metal 

 parts. They may, however, be detected before an experiment by 

 setting the ventilating current in motion and noting the position of 

 the pointer on the spirometer bell. If the apparatus is air-tight the 

 pointer remains constant; if the apparatus leaks the position of the 

 pointer changes one way or the other. Should such a change occur, 

 it is only necessary to apply the usual tests to find what portion of the 

 apparatus is defective. 



The question of a leak in the connection between the subject and the 

 apparatus, as, for example, in the mouthpiece or the nosepieces, is of 

 much more significance and much more difficult to control, as it 

 depends so largely on the cooperation of the subject. The kymograph 

 records sometimes show when such a leak has actually taken place by 

 a break in the regularity of the respiration and a change in the level 

 of expiration. The leak may be so small, however, as to escape detec- 

 tion in this way, and, again, these irregularities in the respiration record 

 may not be due to leaks at all, but to actual irregularities in the 

 point to which the subject empties the lungs. The depth of expiration 

 may be controlled by using a pneumograph around the chest and 

 possibly another pneumograph around the abdomen. If these pneumo- 

 graphs are well adjusted and a sensitive tambour is used, it is nearly 

 always found that changes in the regularity of the pneumograph 

 record are accompanied by changes in the regularity of the respiration 

 record. 



In this connection a difficulty encountered in determining the oxygen 

 consumption may be considered. It is always assumed in the deter- 

 mination of the oxygen consumption that the volume of the apparatus 

 plus the volume of the respiratory tract of the subject is the same at the 

 beginning of an experimental period as at the end. This means that 

 when the valve is turned at the end of the period the subject has 

 expired to exactly the same point as when the valve was turned at the 

 beginning of the period. It can readily be seen that if there is a change 

 in the actual volume of the lungs the value for the consumption of the 

 oxygen will be seriously affected. Sometimes this change may be very 

 gradual and at other times abrupt. When it is abrupt the spirometer 

 record will show it definitely. It is, then, possible to calculate a cor- 

 rection for such change and apply it to the results. 



Roily has sought to overcome this difficulty by having the subject 

 expire as completely as possible and turning the valve at the end of the 

 forced expiration. This is objectionable for several reasons. In the 

 first place, it calls the attention of the subject to his respiration and to 



