352 ON THE ANATOMY AND DEVELOPMENT OF PYROSOMA 
only a clear fluid, as before, except that a few indistinct stria are 
visible towards its upper end. One would hardly know what to make 
of them, if it were not for the circumstance, that a bundle of minute 
filaments, a few of which would readily give rise to the striation in 
question, hangs from the mouth of the duct. The filaments are 
sticking in its atrial aperture by one end, while the remainder of 
their length protrudes. 
The filaments are exceedingly delicate, and so entangled that 
their individual dimensions cannot be estimated. The whole bundle, 
however, measures about ;3,th of an inch in length. The ends of 
the filaments inserted into the aperture are thickened, and more 
or less rod-like. Ina slightly larger ovisac (fig. 4) no such filaments 
are visible about the mouth of the duct, but its upper dilated end 
contains a conical plug, composed of precisely similar bodies, and a 
similar plug occupied a corresponding position in every other ovisac, 
in this stage, which I have examined. If the ovisac is not disposed 
in such a manner, that the plane of the constricted junction between 
the duct and the ovisac is perpendicular to the stage of the micro- 
scope, so as to afford a true profile view, the broad end of the plug 
will appear to be in direct contact with the vitellus, close to the 
germinal vesicle. But I have never met with any such absolute 
contact in a true profile view. On the contrary, in such a view, the 
end of the plug appears to be jammed in the upper aperture of the 
duct, and there is a small interval between it and the surface of the 
vitellus. 
But it must be remembered that (as I have already pointed out) 
in the fresh state, the vitellus, in all probability, occupies the whole 
cavity of the ovisac, and itself stops the upper aperture of the duct; 
and, if this be the case, it is exceedingly likely that the slight separa- 
tion between the yelk and the plug of filaments is a post-mortem 
change. At any rate no filaments are ever discoverable in the cavity 
of the ovisac, and as I have been able to find no complete diaphragm 
across the upper aperture of the duct, there seems to be no reason 
for their absence, unless we suppose that the vitellus itself bars their 
entrance. But in this case the plug and the vitellus must come into 
direct contact. 
This point is of great importance, because there can be no doubt 
that the filaments in question are spermatozoa. The ‘plug,’ and the 
contents of the vas deferens of the testis are precisely similar in ap- 
pearance. The plug is not visible before the atrial end of the duct 
is open, thus providing free access for spermatozoa floating in the 
atrium. As there are no cilia on the inner surface of the duct, it 
