194 Journal of Agricultural Research voi. iii, No. 3 



Compartment B was lined with galvanized iron. At the bottom the 

 horizontal coil I was connected through the partition with the vertical 

 coil H in compartment A. The ice rested directly upon coil I. To cool 

 compartment A, tap water was supplied to coil I and the ice-cold water 

 displaced in coil I passed into coil H. Automatic delivery of cold water 

 to coil H was accomplished b)- the operation of the sounder J, which in 

 turn was controlled by the thermostat (not shown) in compartment A. 

 The form of the thermostat has already been illustrated (13). The low- 

 tension circuit opened and closed by the thermostat operated a relay in 

 which a iio-volt D. C. circuit was opened and closed. In the iio-volt 

 D. C. circuit a 32-candlepower carbon-filament lamp and the sounder J 

 were placed in series. When the circuit was closed, the lever of the 

 sounder compressed a rubber tube through which a slow stream of tap 

 water was passing to waste. The water was then delivered to coil I. 

 The sounder was of a resistance of 4 ohms. 



The constancy of the apparatus is from 0.1° to '0.2° C, depending 

 upon a number of factors, of which the external temperature and the 

 rate of water flow are perhaps the most important. The humidity re- 

 mained high and constant. The apparatus is capable of operating con- 

 tinuously for weeks at a time in the neighborhood of 20° C. in hot weather, 

 with little or no attention, except that of being supplied with ice. The 

 cold-storage features of the humidity chamber, without which it would 

 be of little value for the purpose of operating at temperatures below that 

 of the room, were suggested by Mr. S. J. Dennis, of the Bureau of Plant 

 Industry. 



Four large bunches of average-sized green bananas, carefully trimmed 

 and sampled, were weighed and suspended in the humidity chamber and 

 the bananas allowed to ripen. When their color had just changed from 

 green to yellow, the tips of the fruit being still green, each lot was weighed, 

 sampled as before, again weighed, and allowed to ripen further. The 

 temperature was kept at 20° C. during ripening, and the air in the humi- 

 dor was constantly renewed. When the bunches had become very ripe, 

 most of the fruits showing slight superficial browning, the bunches were 

 weighed and samples again taken. The pulp of the bananas was now 

 soft, tender, and somewhat mealy. It was entirely free from translucent 

 semiliquid portions and from decay. On weighing the bunches at this 

 time many fruits broke off, and the experiment was discontinued. 



The records of the weights permit reference of the results to the basis 

 of the original bunches, assuming, as in the previously described studies, 

 that the stems lost weight at the same rate as the bananas themselves. 

 The analytical data are shown in Table III. In all cases nearly all of 

 the starch of the peel and the pulp gradually passed into sugar. For 

 completeness the ash, the alkalinity of the ash, and the protein were 

 determined in both pulp and peel. As in the previous study, the changes 



