80 BULLETIN 754, U. S. DEPARTMENT OF AGRICULTURE. 
of the characters, resulting in the distance separating them being 
widened from 25 to 40 units. To explain this gametic ratio on the 
reduplication theory requires but a simple alteration of the pe of 
cell division, assumed to account for 3-1-1-3 couplings. 
The 3-1- 1-3 gametic ratio is assumed to be the result of the type 
of cell division eens in figure 12, and the 3-2-2-3 coupling merely 
requires that the cells Ad id aB divide once while the cells AB, ab 
are undergoing two divisions. (Fig. 13.) 
The fact that both of these ears had approximately the expected 
percentage of both white and waxy seeds demonstrates beyond doubt 
that the distributions observed are not due to a failure to properly 
classify the material or to the presence of other factors that would 
influence the expected proportions. If these correlations are to be 
a&B 
Fig. 12.—Diagram showing the type of cell lic. 13.—Diagram showing a possible modi- 
division proposed by Punnett to explain a fication of figure 12 to account for a 
3—1-1-3 gametic coupling. 38—2—2-3 coupling. 
looked upon as the result of reduplication the only explanation is to 
admit other series than those proposed. ‘The remaining two ears of 
the three self-pollinated horny plants are sufficiently close approxima- 
tions of the expected 0.766. 
Five ears were secured from plants grown from the white waxy 
seeds of pedigree No. 1129 that were the result of pollinating these 
plants with plants grown from the colored horny seeds of the same 
ear. Three ears were also obtained where the parentage was re- 
versed. These eight ears are shown in Table XX-XIX. 
The first group had a correlation 0.806+0.017 between colored and 
horny, the expected being 0.8. Of the individual ears only one de- 
viated in excess of four times the probable error, that one—No. 1764— 
deviating by eight times the probable error. The correlation of 
0.91, observed on this ear, is intermediate between the 0.8 expected 
