548 
E. G. Anderson 
and PL To the distribution given on page 546 may be added the data 
from a duplicate planting by Dr. Emerson: 
Per cent of 
crossing- 
Pl Sm PI sm pi Sm pl sm over 
111-2 26 7 4 23 
From Emerson 25 9 4 38 
51 16 8 61 17.6 
Two other backcrosses were then made, which the following year gave 
the results: 
Per cent of 
crossing- 
Pl Sm PI sm pi Sm pi sm over 
238-9 76 24 20 93 20.3 
241 60 4 4 66 6.3 
Construction and results of three-point tests 
In the meantime, crosses were made to involve the Y factor for yellow 
endosperm in addition to PI and Sm, since Y and PI were known to be 
linked. To get a satisfactory three-point backcross test involved several 
difficulties, as follows: 
1. Yellow endosperm is not easily distinguished from white if brought 
in only by the pollen. This is assumed to be due to the dominant F's 
being represented only once in the triple-fusion endosperm nucleus. It is 
therefore desirable that the Fi plants should be used as female parents 
in the backcrosses. 
2. Brown silks are not readily separated from green. This difficulty 
can be avoided only by having red pericarp in each plant. But the 
presence of red pericarp obscures the color of the endosperm. So in order 
to make endosperm separations possible, the female parent of the back- 
cross must be free from red pericarp. 
3. Purple and dilute purple plants usually have some purplish pigment 
in the pericarp, which in some cases interferes with the classification of 
yellow endosperm. 
