58 BULLETIN 754, U. S. DEPARTMENT OF AGRICULTURE. 
the appearance of some seeds with the dominant character the “ fail- 
ure of dominance” is considered demonstrated. This is all very well, 
but it is also just what would be expected if segregation were not 
definite or complete. The finding of some seeds bearing the dominant 
character on plants grown from the suspected seeds does not: explain 
the discrepancies observed in the parent stock unless the proper pro- 
portion of plants show the dominant character. In most cases the 
actual number of plants necessary to determine whether this propor- 
tion is as expected is so large that investigators have been content 
when some of the plants exhibited the expected dominant character. 
COLORED X SELF. 
Twenty-eight ears were obtained from self-pollinating plants 
grown from the heterozygous colored seeds of ears Nos. 1110 and 
1134. Of these 28 ears, 17 were considered as approximations of the 
monohybrid ratio of 3 colored to 1 white. The total number of 
seeds secured from these 17 ears was 6,132, with 25 per cent white. 
The individual ears are also shown in Table XXY. The ears 
marked with a star are those considered as approximations of some 
percentage other than 25 per cent. Of the.17 ears, none deviated 
from the 25 per cent in excess of three times the probable error. The 
remaining 11 ears with two exceptions were close approximations to 
the dihybrid percentage of 43.75 per cent white. The total number 
of seeds obtained from the 11 ears was 5,102, with 48.2 per cent 
white. The deviation of 0.55 per cent below the expected just equals 
the probable error. With the exception of ears Nos. 1914 and 
1862 none of the 11 ears deviated from the 43.75 per cent white in 
excess of three times the probable error. The deviations noted on 
ears Nos. 1914 and 1862 practically balance each other. No. 1862 
being 5.2 times the probable error below the expected and No. 1914 
being 6.5 times above the expected ratio. 
Ear No. 1862 has a percentage of white seeds 11.3 per cent above 
the 25 per cent expected ratio for a monohybrid and 7.45 per cent 
below the 43.75 per cent expected on a dihybrid ratio. The probable 
error is +1.4 and the deviation is 5.3 times the probable error from 
the dihybrid percentage, which the ear more nearly approximates. 
The same plant which produced ear No. 1862 also produced ear 
No. 1861, which was the result of pollen from the plant which pro- 
duced ear No. 1871. These ears, Nos. 1861 and 1871, are reciprocals, 
and are very close approximations to the monohybrid percentage of 
25. If the plant which produced ear No. 1862 was heterozygous 
for two color factors, pollinating this plant with one homozygous 
for one color factor and heterozygous for the other would result in 
an ear with 25 per cent of its seeds white. This same percentage of 
white seeds would of course be the result if both the plants in 
