Journal of Agricultural Research 
Vol. XIV, No. 4 
170 
low yields of seeds, extending along the left margin of the diagram, has a 
corresponding area of low top yields. It is to be noted, however, that 
two cultures, R2C2 and R4C3, which produced high dry weights of tops, 
lie within the areas of low yields of seeds. 
It thus appears that the yields of tops and of seeds do not vary in the 
same way with reference ter the variations in the proportions of the three 
salts in the solutions. It is entirely possible, of course, that this lack of 
correlation between the growth of tops and the production of seed is 
related to other factors than the physiological properties of the media 
in which the plants were grown. Several of the more important of these 
factors may here be mentioned. The plants of a series of cultures like 
those here employed must vary greatly in the degree of productiveness, 
which is certainly much higher in some plants than in others, even when 
grown under exactly similar conditions with respect to the medium and 
the aerial surroundings. This is apparently an hereditary quality and 
one certainly not easily to be controlled in experiments of this character. 
Furthermore, these series of cultures were conducted during a season of 
the year when insect pollination was not possible, and pollination by 
artificial means was perhaps imperfectly accomplished, or it may have 
been less perfectly accomplished with some of the cultures than with 
others. Either one or both of these factors may have exerted an influ¬ 
ence upon the manner in which the yields of seeds varied throughout the 
series. Whatever influence these factors may have had upon seed pro¬ 
duction, it is certain that an abundance of seed was produced by nearly 
all the cultures, as will be seen from an inspection of the last three col¬ 
umns of Table I, giving the numerical data of seed yields. From the 
last column of this table showing the ratios of the yields of tops to the 
yields of seeds, it will be observed that the ratio values for eight cultures 
lie between 2.0 and 3.0, while the ratios for 26 cultures have values between 
1.0 and 2.0. The average of these ratio values for the entire series is 
1.94. This indicates that the average yield of tops for the entire series 
is less than double the corresponding yield of seeds. Nearly all the seeds 
obtained from these cultures were large, fairly uniform in size, and well 
filled. Very few small or imperfectly formed seeds were present. 
II.—TRANSPIRATION AND WATER REQUIREMENTS 
As previously stated, the total amount of water lost by transpiration 
during the growth period was determined for each culture by summing 
the losses recorded for the partial periods between each two successive 
changes of solutions. From the total water loss for each Culture con¬ 
sidered in connection with the dry weights of tops, of roots, and of seeds, 
have been derived the ratios representing the amount of water lost by 
transpiration for each single gram of dry plant substance produced. 
These ratios of transpiration to yields represent the water requirements 
of the plants. Table V presents the data of transpiration for the series 
