10 INDUCTION, DEVELOPMENT, AND HERITABILITY OF FASCIATIONS. 
precedes a ring-fasciation, though as yet no ring-fasciation was apparent. 
In all wild O. diennis the stems were infested with larvee below the fasciations 
and the grooves full of callus, yet it was impossible to find intermediate 
conditions. A plant with a fresh larval trail up the side fasciated after a 
month of elongation from the rosette stage, but by the time the character 
of the tip was well determined the first effects were obscured by the later 
growth. Unequal formation of wood on the two sides at the base of fas- 
ciated stems may be taken as an indication of local inhibition. Transverse 
sections of the lower, round part of branches, which are flat above, usually 
reveal variations in the width of the woody ring. The difference may be 
slight or, in a few cases, as in plate v, fig. 16, very marked and accom- 
panied by callus formation. In the groove-fasciations (plate Iv, figs. 14, 
46) the width of the primary wood where it adjoins the groove, at av, is 
narrower than atv. This is also found to be true in sectioning the rosettes 
cut in late summer from the old stems (plate v, fig. 9). 
What has been said applies to plants out of doors. It seemed probable 
that a different state of things would hold in the greenhouse. Yet the 
fasciated rosettes in the greenhouse have in the stems circular meristems 
about brownish discolorations and a significant feature of their development 
is one-sidedness of growth and a forcing out of the axillary branches. 
Rosettes of O. cruciata planted June 16, 1906, and kept in the greenhouse 
during the summer, were subject to such conditions in the pith. These, 
like the rosettes of O. parviflora of 1905-1906, showed rough places on the 
petioles and midribs of the leaves, incurling of some of the leaves in the 
growing tip, and ruffling of the margins. The O. farviffora planted in 
the summer of 1905, in December, showed larger and longer leaves on one 
. side than on the other. There was then no sign of linear growth, but in 
April they began to fasciate, and in May all four plants were fasciated. 
Frequently the rosettes tip up, owing to the premature development of a 
lateral branch (plate 1, fig. 4), so that one side is higher than the other. 
This looks as if there were inhibition of growth on the concave surface. 
The result of further growth is often a complete torsion of the fasciated 
main axis with fasciation also in the side branches. In studying fasciation, 
species with compact symmetrical rosettes are much to be preferred. O. 
grandifiora is among the impracticable forms, for the side branches normally 
come out very early. A double rosette of Ratmannia odorata, a near relative 
of the cenotheras, the plant illustrated in text fig. 1 and in plate v, fig. 17, 
when sectioned was found to have been injured below the bifurcation, and 
at this point (++) there was inhibition in the formation of wood. Only the 
bifurcated fasciations can be detected at the start, and these are of com- 
paratively rare occurrence. It is evident, however, that the rosettes under 
cover are not exempt from outside injury, and insects may readily enter the 
greenhouse through the open ventilators, besides the many which habitually 
live there. 
