10 Litoyp: LEAF WATER IN GOSSYPIUM 
relative to the graph of constant water when the latter is assumed 
to be horizontal, as it would be if plotted relative to constant leaf 
area (column 6). By making the value of the 6.20 hour figure 
say 160, we obtain results which differ little from those in column 
6, TABLE Iti, in calculating which a variable due to a somewhat 
inconstant relation between leaf weight and area was admitted. 
It is apparent that a much larger increase in dry weight for 
the first four hours of the day than that assumed, together with 
far different march of events during the whole period, must have 
obtained in order to.obviate the conclusion that on August 24 
there was a net reduction in the amount of leaf water followed by 
a material recovery. 
In addition to graphs 1 and ta are to be found in FIG. 2 the 
graphs for the young leaves in the remaining tables. The general 
similarity of the lot further impels the conclusion that graphs I 
and ta are not wide of the truth. 
I have taken the above somewhat indirect method of drawing 
the conclusions desired with reference to the observations recorded 
in TABLE Ill, for the purpose of developing a comparison of the 
observations for the two days, 24 hours apart, at Tucson. We 
have already examined those of August 24, presented in TABLE II. 
Those for August 26 are to be found in TABLE v, in which the 
weights are relative to 100 sq. cm. of leaf blade. Although there 
was a marked fluctuation in the leaf water on August 26, the loss 
of water between the 6.20 and 10.20 hours being as great as that 
on August 24, it is to be noted that the recovery after the 10.20 
hour was more rapid on August 26 (graphs 2, 3), and that the 
absolute amount of leaf water was considerably greater, viz. 
about 25 per cent., on this day than on August 24. Since the 
leaves from which the samples were taken were of uniform texture 
and at similar stages of development, this difference cannot be 
attributed to variation in these. The difference is due quite 
probably to the circumstance that on August 25 the ground in 
which the plants were growing was irrigated. This, if the proper 
explanation, carries with it the suggestion that it would be well 
worth while to determine with what precision the soil water 
conditions may be registered in the condition of the leaves. It 
is quite possible that the needs of the plant for a water supply 
