136 
CARL D. LA RUE AND H. H. BARTLETT 
agreement between semialta X semialta and semialta X typica, and, 
conversely, the unexpected closeness of the modes for semialtaX typica 
and typica X semialta. It is believed, however, that the departures 
from the expected are all explained by environmental non-uniformity. 
The garden itself was relatively uniform, but the cultures were set 
out at different times, so that some of the plants were subjected to 
hot dry weather much sooner than others. The most rapid growth 
took place early in the season, with the result that the last plants to 
be transplanted were markedly the weakest. Without exception, 
the greater deviations from the measurements which would have been 
expected in matroclinic inheritance were correlated with the greater 
intervals between time of transplanting, and, conversely, the best 
agreements with expectation were found in the cases of cultures set 
out on the same day. 
A valuable evidence of shifting of the mode attributable to dif- 
ference in date of transplanting was quite accidentally obtained in the 
case of the cross hilonga X semialta. The culture had been partly 
transplanted at the close of the day, and the next morning the remainder 
was overlooked. It was set out a couple of weeks later. Leaves 
from the two lots were collected separately, and the data are given 
separately in Tables III and IV. 
Considering the unfavorable experimental conditions, the measure- 
ments provide as good a demonstration of matroclinic inheritance as 
could reasonably be demanded. Moreover, it would be unfair not to 
emphasize the fact that in the one or two cases where the measurements 
might appear ambiguous, the plants were in general aspect true to the 
expected type. Thus the cross semialta X typica had the whole ap- 
pearance of self-pollinated semialta, from which it differed only in 
being much more robust. Although we are of the opinion that en- 
vironmental non-uniformity explains the difference, we shall carry 
out more carefully controlled experiments to see if cross-pollination 
produces any effect similar to the vigor of heterozygosis. Such an 
effect might conceivably be measurable if environmental non-uni- 
formity were eliminated, but if it exists it is obviously not large enough 
to obscure the underlying phenomenon of matroclinic inheritance. 
The data for capsule length in the mutation crosses are incomplete 
because of the fact that some of the cultures of 1916 were just coming 
into flower at the time of the first heavy frost and were destroyed. 
Before examining the data for the crosses, the reader should glance at 
