no JOHN T. BUCHHOLZ 
broad primordia in a state of division, since in this species the fusion occurs 
before the cotyledons begin to elongate and is therefore much more rapid 
than in Pinus Banksiana. In the latter case, where the period of fusion 
usually lasts longer, this error of a reversed interpretation could not so eas- 
ily be made. Furthermore, the variations that occur in the cotyledon 
number make it more difficult to recognize either a fusion tendency or the 
opposite condition. 
To prove that these are cotyledonary fusions in Picea, it is necessary to 
rely in part on the statistical method. If, on an average, the younger em- 
bryos can be shown to have a larger number of cotyledon primordia than 
Picea Mariana. Figs. 4-6 are from lot B showing double cotyledon primordia that 
are undergoing fusion. Figs. 7-8, embryos from lot C with no fusing primordia. Figs. 
9-10, embryos from lot D with young cotyledons which have developed well beyond the 
fusion stages. X 32. 
the average number of cotyledons in the older embryos, then there is un- 
questionably a reduction in their number, and the double primordia are in 
the act of fusion. If, on the other hand, the older embryos have a greater 
number of cotyledons than the number of primordia, or, counting these 
double primordia as two, if the resulting cotyledons are of the same average 
number as the primordia, then these double primordia must be cases of 
splitting cotyledons. 
Sufficient variation in size was found between the youngest and the old- 
est embryos to make possible such a study, though they were all secured in 
one collection and dissected out within an interval of only five days, during 
which the material was kept living, though probably not growing as vig- 
orously as if still attached to the tree. Unfortunately, other collections 
from the same tree were not possible after an interval of several weeks, for 
this would very much have simplified the task of studying them by this 
method. However, the facts were well shown even though based on this one 
collection of material, because there was considerable difference in size be- 
tween the youngest and oldest embryos that were obtained. 
By studying the embryos in a watchglass with a binociilar dissecting 
microscope, examining them one at a time from all sides, they were divided 
into four lots. The embryos of lot A included all that were smaller than 
any of the figures shown ; embryos too small to be considered because their 
primordia had either not appeared or were not distinct enough to be counted 
