Chromosomes in Pollen Mother-cells . 315 
Because of this fact, these strands suggest the nature of chromatin. They 
do not, however, hold the stain as strongly as the chromatin masses, and 
there can be no doubt that we have a differentiation between chromatin 
and linin, although what the exact difference is cannot be stated. It has 
been observed in a number of preparations that, at the stage represented in 
Fig. 15, the chromatin masses show a tendency to group themselves closely 
about the nucleoli, and the linin strands to radiate from the collection. 
This condition gives a very noticeable appearance even with low powers of 
the microscope. What significance, if any, such a grouping may have, the 
writer is not prepared to say, neither is it known to be a regular and 
perfectly normal phenomenon ; although the preparations show no evidence 
of abnormal fixing. In former publications the writer expressed the 
opinion that the nucleolus may contribute material for the formation of 
chromatin. By this he does not mean, as some observers have interpreted 
the statement, that the nucleolus is reserve chromatin, for he has no well- 
defined idea of what reserve chromatin may be, or that such a substance 
exists. 
With further development of the anther, the pollen mother-cell and its 
nucleus increase in size. The chromatin masses increase through growth 
and become more regularly distributed within the nuclear cavity, but more, 
especially along the periphery. The nucleoli lie within the nuclear cavity, 
which is, as a rule, freer from the network, and their connexion with the 
chromatin is less intimate, or they are entirely free (Figs. 16, 17). Here, as 
in Podophyllum the chromatic masses, which are always variable in size 
and in shape, are connected into a reticulum by means of linin threads. 
These threads are sometimes slender, showing very fine granules ; some- 
times they are thicker and retain to some extent the chromatin stain. The 
same phenomena are manifested, whether the stain used was the triple or 
the iron-haematoxylin ; but for the differentiation into chromatin and linin 
the triple stain seems to be superior. It may be remarked also that the 
number of these masses always far exceeds the somatic number of chromo- 
somes. In preparations of this and later stages (Figs. 15, 16, 17) which 
are not too densely stained, it is to be seen that each chromatin mass is 
composed of many smaller granules (Fig. 18). Fig. 18 was taken from the 
same anther loculus as Fig. 16. In this the stain differentiated in such 
a manner as to leave the linin very much paler or almost colourless. 
A similar condition is figured by Allen for Lilium canadensis. In the 
period of development in L. Mart agon between the stage shown in 
Figs. 16 and 17 and synapsis, there seems to be undoubtedly a very 
remarkable and striking variation. After a slight increase in size, nuclei, 
presenting the structure of Fig. 17, may undergo the synaptic contraction 
(Figs. 22, 23). On the other hand, the chromatin masses may undergo 
a further reticulation into finer granules, so that the nucleus reveals a linin 
