Nuclear Division in the Hepaticae. 505 
converging to this point. The radii themselves often extend 
quite out to the periphery of the cell in the direction of the 
polar axis, whilst those on all sides of this axis exhibit the 
well-known system of curves produced by holding a horseshoe 
magnet below a paper covered with iron filings. Now it is 
not practicable in these cells with their granular contents to 
demonstrate the permanence of the centrosomes during those 
periods of inactivity when the radiations are absent. Thus 
I find that they become indistinguishable through a consider- 
able part of the long interval which elapses whilst the nucleus 
remains in the aster stage, although they may often be quite 
plainly seen both before and after this period. Nor can they 
be recognized in the resting-cell. Even assuming them to be 
the permanent organs of the cell that Boveri 1 claims, they 
could scarcely be differentiated from the other granules 
scattered about in the protoplasm. Boveri and with him 
many other cytologists believe that they do actively direct 
and control the process of nuclear division, and these writers 
insist on their multiplication by fission alone — a necessary 
condition of the view of their morphological permanence. 
It may incidentally, however, be pointed out that their appa- 
rent formation de novo , which especially occurs during the 
maturation of the ova of certain animals, still demands a 
satisfactory explanation, regarded from this standpoint. 
And when we turn to the spore-mother-cells in the Jun- 
germannia series of Hepaticae, we are confronted with still 
other difficulties. Here we must assume that the centrosome 
either divides into four, simultaneously, or that the full number 
is reached as the result of a second bipartition, which might 
happen after the divarication of the first pair. And in either 
case it seems odd that the systems of radiations should defer 
their appearance until after the centrosomes have taken up 
their proper positions opposite each cell-lobe, seeing that the 
act of translation involves definite activity. This difficulty is 
also met with in a simpler form even in normal bipolar spindles, 
1 Boveri, Ueber das Verh. d. Centrosomen b. d. Befrucht. d. Seeigeleies. 
Verhandl. d. Physik.-Med. Gesellsch. z. Wurzburg, Bd. XXIX. 
