Dicotyledon—Caltha palustris. 531 
are like small dumb-bells or may even have the form of a 
somewhat straight S. Their chromatic elements are so densely 
packed at this time that no structure can be made out. 
Although not able to follow it out with any certainty, I have 
always been strongly under the impression that it was the first 
nucleus that left the pollen-tube which fertilized the polar nuclei, 
for the polar fertilization is always much more advanced than 
that of the oosphere. I have read since that Professor Guignard 
finds this to be so in the Monocotyledons that he studied, and 
therefore it is extremely likely that it is the case in Caltha 
also. This may partly account for the fact that the process 
has been overlooked for so long. For when fertilization of the 
oosphere takes place, the polar fertilization may be too far ad¬ 
vanced to be noticeable by any one not on the lookout for it. 
By the time the vermiform generative nucleus has reached 
the middle of the sac, it has enlarged from an extremely 
small body to a nucleus of very considerable size, and its 
chromatic elements now show particularly clearly as small 
widely-scattered granules. The nucleus which fertilizes the 
oosphere, on the contrary, increases very little in size, and 
its chromatin still remains somewhat tightly packed, so that 
it is always very dark and shows merely a coarsely granular 
structure. 
The result of the two fusions taking place at a slightly 
different time is of course that one rarely gets very good 
instances of both in the same sac. Where the ovum and male 
nucleus are actually coalescing, as shown in Fig. 2, the polar 
fertilization is in a late stage, so that there is only one end 
of the vermiform nucleus clearly distinguishable from the 
general polar mass (Figs. 3 and 4). In the sac from which 
Figs. 8 and 9 are taken, the male nucleus has only just 
reached the oosphere, and therefore, as one would expect, the 
vermiform nucleus, although tightly wound round the definitive 
nucleus, and indeed coalescing with it in parts, yet shows 
several of its sharp turns. In P'igs. 6 and 7, however, we have 
a much better instance of the vermiform nucleus, for here it 
is only lightly applied to the definitive nucleus, and fusion has 
N n 2 
