404 E. R. Downiss, 
disappeared. When the tail attains its full growth it is about three 
times the length of the head. 
The mature sperm. 
The mature sperm (Fig. 68), (when at last all vestige of the 
surrounding cell wall has disappeared) consists of a cylindrical head 
tapering to a point at the anterior end, a somewhat hemispherical 
middle piece at the posterior end of the head, and attached to this 
the long tapering tail. The head stains deeply with any nuclear 
stain. But the extreme tip bears an acrosome. The middle piece 
is clear, taking the stain slightly, even a plasma stain. The lash- 
like tail is about three times the length of the head and middle 
piece. Within the head may be seen six bodies each consisting of 
two more or less connected spherical masses. These apparently are 
the persistent chromosomes, each consisting of two bivalent chro- 
momeres. These are not easily detected, aud I think mark the 
early stage of the mature sperm, disappearing shortly after complete 
absorption of the enclosing spermatid. Near the posterior end of 
the middle piece, the centrosome can be detected. Through it the 
axial fibre extends into the tail and anteriorly toward the tip of 
the head. During the development of the spermatozoa the centro- 
some could not be detected after the stage represented in Fig. 61 
until the surrounding cell was nearly absorbed (Fig. 67). It reap- 
pears in about the same position as it disappeared. That it persists 
seems probable. 
What this droplet forming the middle piece really is can only 
be judged from analogy. In many of the higher animals the centro- 
some of the spermatid divides. One part forms or helps form the 
middle piece; the other gives rise to the axial thread. PAULMIER (110) 
has shown in Anasa that the axial thread grows out from the 
centrosome. HERMANN (100), BexpA (95) and Meves (106) have found 
in Salamandra that one of the centrosomes gives rise to the middle 
piece or part of it at least, and the other is connected with the 
axial filament. Similarly the results of Moore (108), Suzuxr (113), 
Korrr (101), LENHOSSEK (103) and others agree that the axial thread 
comes from one part of the divided spermatid centrosome, while the 
other often greatly enlarged, forms or helps form the middle piece. 
We find in the adult sperm of Hydra that the axial filament passes 
through the centrosome and likely originates as an outgrowth from 
it. Possibly the centrosome divides when the other part may form 
