I 
Pollen Development in Lachtca . 369 
is clear. If it is semipermeable, and the clear area of karyolymph surround- 
ing the reticulum contains more strongly osmotic substances, then pro- 
gressive contraction of the network would take place as water is withdrawn 
into the clear area between this membrane and the nuclear membrane. The 
excellent fixation of the cytoplasm in these cells makes it very difficult 
to suppose that these phenomena result from the treatment. 
But it is evident that the presence of an osmotic membrane surround- 
ing the reticulum will not entirely account for the contraction, because 
numerous stages of contraction have been seen where no such membrane 
is present. In view of the statement made by Lawson ( 1911 ), that the 
apparent contraction is really due to the sudden growth of the nucleus 
and is not a true contraction, a series of measurements were made 
beginning with the resting stage and ending with the postsynaptic spireme. 
These measurements showed conclusively that, in Lactuca at least, there is 
a real contraction of the nuclear content. There is at the same time 
a steady increase in the nuclear volume, and the two phenomena appear to 
be more or less simultaneous, though the contraction may begin first, and 
the nuclear expansion usually continues after the reticular contraction 
is completed. The suggestion of Davis ( 1910 ), that contraction is due 
to shortening and thickening of the threads of the network previous to 
their transformation into a spireme, also deserves consideration. Com- 
parison of the reticula in Figs. 8-10 with the resting nucleus shows 
clearly that the threads in the former take up considerably less space, 
but it has been impossible to find clear evidence that the threads are 
individually noticeably thicker. A membrane surrounding the contract- 
ing reticulum might then function as an osmotic aid to contraction, but that 
contraction can take place in its absence is shown by such figures as 7. 
At times, as in Fig. 7, the transformation of the thread into a spireme 
occurs prior to the formation of the true synaptic knot. 
As the contraction proceeds it is often accompanied in its later stages 
by a rearrangement of the threads of the reticulum to form a more or less 
continuous spireme. It is impossible to determine the exact nature of this 
rearrangement of threads. It appears to be a process quite distinct from 
the contraction, and in many cases (Fig. 7) the rearrangement is already 
taking place before the contraction has proceeded very far. In such cases 
the nuclear membrane is frequently not intact, and there is no evidence of a 
precipitation membrane. Nuclei in this condition at first suggest artifact, 
but the cytoplasm is well fixed, and as contraction of the nuclear content 
takes place in different directions in adjacent cells there is no reason for 
attributing it to the fixing fluid. The stages represented by Figs. 8-10, 
compared with Fig. 7, indicate that the amount of contraction of the 
reticulum may perhaps be conditioned to some extent by the presence of a 
precipitation membrane. 
