262 Bayliss and Starling, 



It is evident then that in using the cooled air in respiration, we are 

 exerting by some means, a cooling influence on the base of the heart. 



In the light of these experiments we can account for Waller's 

 results on excised hearts. These were mostly allowed to get cold 

 after extraction from the body, or when examined in situ they were 

 no longer kept warm by the blood current. Now it is evident that 

 the thin wall at the base of the ventricles would cool more quickly 

 than the dense mass of muscle forming the apex, and the final result 

 would be the same as using cooled air in the living animal. 



In all cases of which mention has been made above, the ventricle 

 is supposed to have been beating in normal sequence to the auricle; 

 where this is no longer the case the contraction may probably start 

 at any point in the ventricles, as Waller has pointed out, and gene- 

 rally from any spot that is warmer than the rest of the surface. 



We have found that it is quite possible to obtain a normal diphasic 

 variation from an excised heart. The heart must be beating well, 

 and to ensure this we found it advisable to inject sufficient morphia 

 to produce complete anaesthesia, and allow the animal to recover from 

 chloroform, which is very apt to cause complete stoppage of the auri- 

 cular beat; a moist chamber warmed to about 38" C. is prepared, 

 the heart cut out as rapidly as possible and placed on a glass plate 

 in the warmed chamber, electrodes applied to two points of the ven- 

 tricular surface, the tambour arranged to shew the beats, and photo- 

 graphs of the variation taken at once. The whole process from the 

 moment the chest is opened can be got through in not more than 

 30 seconds. We then get a diphasic variation shewing negativity starting 

 at the base and followed by negativity of the apex. In some cases 

 we found that while the right ventricle gave a „stepped curve" (see 

 below) indicating partial cooling at the base, we could obtain the nor- 

 mal curve from the left ventricle. This illustrates well the effect of 

 thickness of the ventricular wall on the rapidity of cooling. 



We have now to consider the effects of warming or cooling dif- 

 ferent parts of the heart on the time-course of the electrical change. 

 Sanderson and Page shewed that the effect of cooling any portion of the 

 contracting ventricle of the frog was to increase the duration of the 



