24 CIRCULAR NO. 120, BUREAU OF PLANT INDUSTRY. 
each combination at the side occurs four times in the corresponding 
horizontal row of squares. Thus each of the squares represents the 
result obtained by combining the gametes representing the horizontal 
and vertical rows that intersect at that point. In all cases where 
both S and XY occur together the seed should be horny, where only S 
occurs the seed should be sweet, when only XY occurs it should be waxy, 
and in ene square (No. 16), where neither S nor XY occurs, no predic- 
tion can be made, since this is presumably a new condition. 
Leaving the nature of the seed from square 16 out of considerate 
for the present, it can be seen that there are 9 squares in which both 
S and XY occur (horny), 3 in which X alone occurs (waxy), and 3 in 
which S alone occurs (sweet). With respect to the horny and waxy 
seeds, these numbers approximate the ratios that were actually ob- 
tained, there being roughly 9 horny seeds to 3 waxy, but the sweet 
seeds occur as 4 instead of 3 out of every 16. We must therefore 
assume that the new type, swaa, represented in square 16 and contain- 
ing neither the sweet nor the waxy factor is or resembles sweet. 
Careful scrutiny of the sweet seeds failed to show any consistent 
differences that would allow another class to be separated, but if the 
present method of looking at the cross is to be of use it should be 
possible to detect the differences between the ordinary sweet seeds and 
this new class by analytical breeding. 
One way to test the theory that this new class of apparently sweet 
seeds really lacks the element ordinarily concerned in the production 
of the sweet character is to cross it with pure waxy. The cross of 
sweet and waxy has heretofore always produced horny, but if this 
new class lacks the factor for sweet, the cross should have the gametic 
composition represented by squares 12 and 15 and should result in 
waxy instead of horny seeds. Another test could be applied by cross- 
ing the new class with ordinary horny varieties. The first genera- 
tion of this cross should be horny, but the formula of gametic com- 
position, S.’szv, would be the same as for the synthetic horny, and in 
the second generation 3 out of every 16 seeds should be waxy. 
Thus, if the present assumption regarding the nature of this class 
is correct we may expect two apparent anomales, not observed hith- 
erto, viz, waxy seeds as the immediate result of crossing sweet and 
waxy, and waxy seeds in the second generation of a cross between 
sweet and horny. 
COMPARATIVE WEIGHT OF HORNY, SWEET, AND WAXY SEEDS. 
The view that waxy as well as sweet endosperm may be compared 
to incomplete stages in the production of the horny endosperm 1s 
strengthened by the relative weights of the sweet, waxy, and horny 
seeds where all three classes occur on the same ear. The horny seeds 
[Cir. 120] 
