438 BROOKLYN BOTANIC GARDEN MEMOIRS 
14 :8:8 instead of 19: 3:7. A nearer agreement might well have 
been expected even with this small number of families. The dis- 
crepancy is probably due in part to the small size of some of the fami- 
lies. Thus among the families which were of necessity included in 
the first section of the table there are six families, the largest of which 
had no more than 33 individuals and if these families belonged properly 
to the second group of the table, only 1 or 2 individuals of the recessive 
type were to have been expected. That some of these small families 
which contained no recessives, had them nevertheless potentially 
present may be taken as a certainty. Only 5 such families need be 
shifted from the first section of the table to the second section, to 
bring about as perfect agreement with the expected ratio as is the- 
oretically possible among 29 F3 families. 
A more positive demonstration of the duplication of the B factor 
may be expected from the F; families derived from the Bremen bio- 
types in which the suppression of the dominant rosette characters is 
less extensive. Further experiments involving these biotypes are in 
progress. 
DISCUSSION 
The discovery of a second case of duplication of determiners in 
Bursa bursa-pastoris owes its chief interest to the facts pointed out 
in one of my former papers (Shull, 1914), (a) that cases of actual 
duplication of genes appears to be rare, and (0) that there is some 
likelihood that the duplication of factors may involve ‘“‘a series of 
special genotypic phenomena.” 
It will be recalled that only in the red pericarp of wheat, yellow 
endosperm of maize, the occurrence of a ligule in oats, and of triangular 
capsules in shepherd’s-purse, was the demonstration of dimery con- 
sidered adequate. Several new instances have been brought to light 
more recently, and at least one of these must be admitted as fully 
demonstrated (Ikeno, 1917). This relates to the quantity and dis- 
tribution of chlorophyll in Plantago major, in which two seemingly 
identical factors, G and H, determine independently the normal full 
green pigmentation of the plant. Either of these two factors may be 
entirely wanting, without modifying in any way the self-green color, 
but when both are absent, the plants are conspicuously variegated 
with white or pale yellowish blotches and stripes, the total quantity 
of chlorophyll being considerably reduced. 
Hallqvist (1916) has found the pinnatifid leaves of Brassica Napus 
apparently produced independently by two factors, the recessive 
undivided type reappearing in the F; in only one plant in 16. As the 
lobed plants showed considerable variation the results in F3 will be 
necessary to confirm the author’s conclusion in this case. 
