48 Papers from the Marine Biological Laboratory at Tortugas. 
ence of the postsynaptic chromosomes (sometimes still partially arranged in 
a thread). Thus the chromosomes have remained throughout the growth 
period in small individual bulk (always retaining their morphological iden- 
tity) and in a compact mass. Their proximity to the nucleolus is to be 
explained in connection with the maturation phenomena. 
Occasionally I have seen very close to the nuclear wall a small dumb-bell- 
shaped body (figs, 40, a, 49) (sometimes the body consists of three or even 
four globes)—somewhat larger than the ordinary bivalent chromosomes. 
Mathews (1895) describes very definitely the origin of the centrosome from 
within the nuclear membrane. I have tried to identify this problematical 
body with the centrosome of Mathews. This I am unable to do for several 
reasons. Bryce (1903) also describes very similar bodies in Echinus escu- 
lentus, but says that he has not been able to convince himself “ that they are 
more than accidents of staining and fixing” (p. 491). Hartmann (1902) 
makes no definite mention of such structure in Asterias glacialis, probably 
including this as well as the above-mentioned masses of granules under the 
“clumps of chromatin” and “ accessory nucleoli.” However, eggs in which 
such distribution of chromatin occurs Hartmann classifies as “ abnormal,” 
stating further that in normal eggs all “ genuine chromatin and plastin were 
combined in a single nucleolus.” Nor does Guenther (1903) mention this 
body in Psammechinus microtuberculatus or Holothuria tubulosa. I tried 
also to identify it with portion of the chromosome mass, but unsatisfactorily. 
Though I can make no positive statement concerning it, inclining to the 
opinion that it represents several or perhaps a single large bivalent chro- 
mosome, since it is not invariably present with certainty, I am convinced 
that it can not be the centrosome. For reasons soon to be given, I hold to 
an extra-nuclear origin of the centrosome. Furthermore, this problematical 
body is always at least double, while the centrosome arises as a single 
structure. Again, its size is several times larger than the largest centro- 
some I have seen during the maturation process. I have not been able to 
follow satisfactorily the fate of these bodies amid the general mingling 
and concentration of chromosomes and nuclear fragments and their exit 
from the nucleus at one point during the initial stages of maturation. 
MATURATION. 
NUCLEAR AND CYTOPLASMIC ALTERATIONS. 
Study of sectioned material confirms in every respect my observations on 
the living eggs. In batches of eggs fixed from 5 to 10 minutes after depo- 
sition in sea-water the initial stages of maturation are already visible. The 
first indication is a puckering of the nuclear wall on the side nearest the 
periphery of the egg (figs. 38, 49, 50). Dissolution or rupture of the wall 
occurs at this point after an interval of from 15 to 20 minutes (time always 
reckoned from time that eggs are placed in sea-water) allowing an inter- 
