EMBRYO DEVELOPMENT AND POLYEMBRYONY 
129 
both of these characters disappeared from the ontogeny of the higher 
Abietineae. 
Rosette embryos (R-em) usually develop in Pinus, are somewhat less 
developed in Cedrus, while Abies has rosette embryos only rarely, about 
Fig. 9. Diagram to show the affinities of the genera of Abietineae, based on em- 
bryogeny. A, probable range of apical cell; Cl.p, range of, cleavage polyembryony ; R-em, 
range of rosette cells which may give rise to rosette embryos; R, range of abortive rosette 
cells. Two anatomical characters are included: Res.c, the range of resin canals in secondary 
wood, defining the Pinae of Jeffrey (19, 20); Sp.s, range of spur shoot, used by Engler and 
Gilg (16) and others in subdividing the Abietineae. 
as often as it has cleavage polyembryony. However, these two occasional 
characters of Abies are not necessarily associated with each other in their 
occurrence. 
In Picea, Larix, and Abies balsamea (normally) the rosette embryos are 
represented only by their initial cells (R). Tsuga also has only these 
aborting rosette cells, though its primary embryos separate as in Pinus. 
These abortive rosette cells are the last rudimentary structures that remain 
to indicate the origin of these embryos with simple polyembryony from the 
type of cleavage polyembryony found in Pinus. 
One of the most formidable obstacles to the opposite interpretation, 
namely, that the embryo of the Picea or Larix type has given rise to the 
pine type with its cleavage polyembryony, comes from the study of these 
rosette cells. It is impossible for these abortive structures as represented 
in Picea or Larix to have given rise to the active rosette embryos of Pinus 
and Cedrus. On the other hand, the view that rosette embryos and the 
abortive rosette cells are steps in the elimination of cleavage polyembryony 
offers a very satisfactory hypothesis that is consistent with the facts and 
that very definitely points out the direction of this polyembryonous evolu- 
tion. 
In Pseudolarix the rosette cells are present (29), but nothing has been 
described or figured to indicate whether cleavage polyembryony or rosette 
embryos develop at a later stage. The exact position of Pseudolarix is 
still doubtful, but from what is known of its embryo, and considering that 
it resembles Larix in its spur shoot, it is safest to place it in the diagram 
near Larix. 
