8 i6 
Journal of Agricultural Research 
Vol. XXIV, No. lo 
(c) Any zygote derived from the fusion of a type gamete with a rogue gamete, 
whichever way the cross is made, produces a rogue; an intermediate gamete on 
union with a type gamete produces an intermediate or nonrogue form. 
(d) Rogues are apparently produced more often from the upper (roguelike) portion 
of the intermediate plants than from the lower (typelike) part. It is also found 
that pollen grains produce a greater ratio of rogue gametes to nonrogue gametes than 
do the ovules, and that the proportion of rogue gametes produced by the pollen grains 
increases more rapidly, in going from the lower to upper flowers, than the proportion 
of rogue gametes produced by ovules. 
(e) Intergrading intermediates of strains not throwing many rogues were found to 
produce rogues no more often from pods of the upper parts of the plant than from the 
lower nodes. 
(/) Rogues with narrow, pointed stipules and narrow^ curved pods, from whatever 
source, breed true. 
(g) The cross rogue X type, no matter how made, produces only ro^es in the and 
succeeding generations. In the seedling stage of the F, hybrid until the seventh or 
eighth node is developed, many of the plants are typelike in the character of the 
stipules and leaflets. These plants at maturity, with few exceptions, are all rogues. 
The plants in the Fa generation are recognizable as rogues at all stages of development. 
There is apparently no segregation in the Fj hybrids, at sporogenesis, of the rogue and 
type factors. 
(k) Intervarietal crosses of rogue with type plants of varieties producing rogues give 
results similar to crosses of rogue with the variety from which the rogue origmated as 
far as the rogue characters are concerned. Side by side with the apparently non- 
Mendelian behavior of the rogue characters is found the expected Mendelian segre¬ 
gation of such characters as pod shape, color of seed, shape of seed, and height of 
plant. 
Bateson and Pellew (i) have suggested in explanation of the anomalous 
genetic behavior of the rogue characters that they are due to somatic 
segregation of the type ‘‘elements” in the hybrid so that the type 
“elements” are missing from the germ plasm at the time of gamete 
formation. Intergrading intermediates are assumed to be mosaics of 
type and rogue tissue, rogue gametes being formed from rogue tissue 
and type gametes from type tissue. 
The present paper is in part a confirmation of the above results. 
However, a different interpretation of the results is advanced. In 
addition, data are submitted of the F^, Fj, and Fg generations of inter¬ 
varietal crosses of rogue and type plants with type plants of varieties 
that apparently do not produce rogues. These crosses have given 
entirely new results, from which important conclusions have been drawn. 
EXPERIMENTAL METHOD 
The experimental methods used in the work here reported were the 
same as those described in a previous paper (4). At that time the fact of 
cross-pollination of rogue flowers by bumble^es was noted. In growing 
large cultures of peas it is impossible to protect all the individual flowers 
with paper or cloth coverings. In 1921 an attempt was made to grow 
peas under wire mosquito netting. This method is too expensive to be 
used on a large scale, so that in Hie future it seems advisable to carry on 
as much as possible of the breeding work with rogues in the greenhouse 
during the winter. 
The results of 1919 and 1920 may be used to illustrate the amount of 
volunteer crossing that may be expected in the field in crosses be¬ 
tween Gradus and Rogue, The number of volunteer hybrids is suffi¬ 
ciently great to make necessary some sort of protection of the rogue 
plants. 
