1920] SNOW— DIAPHRAGMS 313 



possible that the greater diaphragm distance found in the experi- 

 mental culms may have been due only indirectly to decreased 

 pressure, through its effect upon growth. In every instance 

 (except region 95-135, experiment 3) an increase in the diaphragm 

 distance was correlated with decreased growth rate. This agrees 

 with the results reported earlier in this paper. 



Number of air chambers. — In reviewing the literature one 

 finds a rather general opinion that air spaces increase with the 

 amount of water in the substratum (2, 3, 7, 10, 12, 16). Recent 

 work by Folsom (6) on Ranunculus, however, has shown that 

 while the aerenchyma of the stem varied directly with the amount 

 of water in the soil, that of the root showed no such constant 

 relation, and in some cases even varied inversely. Various func- 

 tions have been attributed to the air-containing tissue by different 

 authors. Some consider it as "floating tissue' ' (2, 9, 10), while 

 others consider it a means of oxygen supply, giving a lack of oxygen 

 as its direct cause (7, 10, 12). On the other hand, Wieler thinks 

 it has no function, and attributes its formation to the direct stimulus 

 of water contact. Devaux (5) thinks that the hypertrophy of 

 lenticels found in water and moist air is due to a checking of the 

 transpiration, a factor which apparently has not been tested in * 

 connection with the formation of air spaces, although suggested 

 some time ago by Cowles (4). 



If low oxygen pressure is the cause of increased air spaces, it is 

 rather strange that in Folsom's experiments the roots, which 

 under any condition are farther from the oxygen supply than the 

 stem, should show either inconstant increase or a positive decrease 

 in aerenchyma with increase in water content of the soil. In 

 experiment 3 there was a small increase in the total number of 

 air chambers in the experimental culm, at the same time that the 

 control remained the same. On the other hand, experiments 1 

 and 2 give evidence of a greater variation in total number of spaces 

 and in their uniformity in size between the culms under like condi- 

 tions than between the experiment and the control. Contrary- to 

 expectation, therefore, low atmospheric pressure appears to have 

 no effect upon these two characters in Scirpus. 



