The Inheritance of Quantitative Characters in Maize 41 
and moisture conditions are as uniform as is possible in field 
experiments and if these conditions are suitable for normal 
development, unavoidable changes in meteorological conditions 
do not change the ear development to any great extent. The 
greatest changes occur thru physiological correlation. Number 
of rows of seeds, and therefore total number of seeds, is in- 
versely correlated with ear length. Size of seed and possibly 
the character of the seed may also have some effect, tho just 
what this effect is we are not prepared to say. Size of plant is 
directly correlated with length of ear. 
How much of these correlations is gametic in the sense of 
true gametic couplings is unknown. It is apparently physio- 
logical in most cases, tho this physiological correlation is imposed 
by gametic constitution. In other words, a plant which is 
gametically large and a plant that is gametically small have differ- 
ent possibilities of ear development even when the same genes 
immediately affecting ear development are present. Expressed 
in genetic language, zygote (tall + ear length AABBCC) gives 
longer ears than zygote (short + ear length AABBCC) . 
Tables 13, 14, and 15 give the data on lengths of ear for the 
crosses (60-3x54), (60-5x54), and (60-8x54). These, it will be 
remembered, are three crosses made between individual plants of 
Tom Thumb pop (No. 60) and Black Mexican sweet (No. 54). 
Unfortunately the data on the F x generations were taken only 
for the cross (60-5x54) . It can be stated definitely that the range 
of variability in the other two families was no greater, but not 
that the percentage of variability was no greater. It is possible 
that a slight error is introduced, therefore, in comparing the F, 
generations of the crosses (60-3x54) and (60-8x54) with the F x 
data taken for cross (60-5x54). From what is known of the 
F 1 generations, however, it is thought that the error is very small. 
In cross (60-3x54), three F 2 families were grown from selfed 
ears of the F 1 generation. If 12.48 ± .72 per cent represents the 
correct variability of the F 3 generation (Table 13), then the 
variability of each F 2 generation is much greater, namely, 
16.70 ± .67 per cent, 15.87 ± .63 per cent, and 15.54 ± .90 per 
cent respectively. Perhaps it is wrong to compare the variability 
of the total population in F 2 with that of F a , for the F 1 ears 
followed to the F 2 generation may have had different gametic 
formulae. The figures 22.30 ± .74 per cent are simply given for 
what they are worth. There is no way of disregarding the fact 
that one F 2 family, (60-3x54) -1, had a greater range of variability 
than F ± by five classes. Individuals were obtained, therefore, 
that were exactly like typical individuals of both of the parental 
varieties. 
Let us see how these individuals behaved in the F s generation. 
