360 HARVEY M. SMITH. 



4. Leaks were practically eliminated. It was possible to de- 

 velop a strong negative pressure in the jars, close all valves, and 

 allow the apparatus to stand for several hours with no appre- 

 ciable diminution of pressure. 



5. The negative pressure in the respiration jars averaged 

 about 2 cm. of mercury. This factor was practically constant 

 for the entire course of the experiments. 



6. Measurements of the rate of flow with the air current 

 moving at as near the average rate as was possible to judge 

 yielded two minutes for each liter of air. This stream was 

 divided between a jar of four liters capacity (used for Necturus, 

 Rana catesbiana and Ampkiuma) and one of one liter capacity 

 (used for the smaller animals). Tests on each jar separately 

 showed that the air in the large jar was being changed every 

 eleven minutes, while that in the small jar was changed every 

 five minutes. This should be sufficient speed in each case to 

 keep the atmosphere around the animal relatively free of carbon 

 dioxide. Attempts to cut down the rate through the small jar 

 to more nearly equal that of the larger jar were unsuccessful 

 because it was found that, with too slow a rate, the holes in the 

 bubblers in the gas collecting jars became clogged with precipitate 

 of barium carbonate and the aeration stopped. 



7. Determinations of the dissolved carbon dioxide in the water 

 surrounding the animal at the end of the preliminary aeration 

 and again at the end of the final aeration yielded approximately 

 the same figure in each case. For instance, while a carbon 

 dioxide test was being made on a bull frog, no titratable CO 2 was 

 found in the water used in the jar, 5.5 mg. CO 2 per 100 cc. 

 were present in the water at the end of the preliminary aeration, 

 and the same figure at the end of the final aeration. 



8. Several blanks were run with the regular amount of water 

 in the respiration jars, but without animals, and one with a com- 

 pletely empty jar. The results of these trials are shown in 

 Table III. 



For all the above tests 250 cc. of N/io barium hydroxide were 

 measured into two gas-washing bottles, and the bottles filled to 

 capacity (300 cc.) with distilled water. This same procedure 

 was followed in filling the jars for the regular carbon dioxide 

 output tests. For numbers 7 and 8 solutions were used which 



