EVOLUTIONARY STATUS OF POLYCOTYLEDONY 
117 
indefinite, and even when they become cycHc in the Ranales they are sub- 
ject to fluctuation. It is easy to find a buttercup with extra petals, but it 
is seldom that the flower of a Phlox or a bluebell will show variations from 
the regular floral formula. Therefore it is very reasonable to regard this 
variation in the cotyledon number as an earmark of the low genetic position 
of the pines among the Coniferales. 
This view falls in line with recent paleobotanical investigations which 
show that Pinus and the earlier form Prepinus are historically the most 
ancient conifers, also with the primitive position assigned to Pinus by the 
writer on the basis of its embryogeny. 
Again, turning to our comparison of the cotyledons with floral structures, 
it is noteworthy that when floral members are spiral, their number is subject 
to the widest variation. Cyclic parts do not fluctuate so much, but when 
the cyclic condition is only very recently established they are also subject 
to considerable variation. Similarly, if the cotyledons still retain traits 
that characterize the spiral arrangement, this would account for the varia- 
* tion in their number which may be found in dift'erent individuals of any spe- 
cies. The first simple leaves that appear above the cotyledons in the young 
seedlings are spiral, and this also suggests that the cotyledons, which are 
doubtless modified from the first of these simple leaves, were originally spiral 
but have become more or less cyclic. The appearance of primordia on one 
side of the embryo before they are visible on the other is therefore also a 
vestigial character and an evidence of the original spiral condition of the 
cotyledons. Figure 23 shows the condition of the cotyledon primordia of 
many embryos, which may be taken to suggest that they are essentially 
spiral in their origin. These three features, namely, the variation in the 
cotyledon number, the spiral arrangement of the first leaves, and the slight 
tendency for the cotyledon primordia to appear on one side earlier than on 
the other, constitute the basis for the conclusion that the cotyledons them- 
selves have become cyclic from a primitive spiral condition. To this might 
perhaps be added the evidence from the occasional displaced plumular 
leaves which Hill and DeFraine find added to the cotyledonary node. 
It is further interesting to call attention to the close parallel which seems 
to exist between the evolution of the floral members in angiosperms and 
that of the cotyledons of the embryo. In the flower, we pass from indefinite 
polypetaly to definite numbers in the floral members, then to sympetaly, 
which results in the corolla tube; this becomes two-lipped and finally ligulate, 
aside from its many other variations. A similar evolution has probably 
taken place in the history of the cotyledons. Originally, the cotyledons may 
have been spiral, but when the seed habit became established they soon 
became cyclic, the polycotyledonous condition today; these reduced their 
number by fusions, or occasionally by such methods as abortion of pri- 
mordia. In some forms fusions became more general, resulting in the 
cotyledonary tubes. 
