40 Papers from the Marine Biological Laboratory. 
winding of a thin sheet of material around its longitudinal axis. While 
the process of formation of the octopus spermatophore has not been 
observed, the structure of the spermatophoric organ indicates that, 
as in the squid, the forming spermatophore is kept rotating on its 
longitudinal axis while a thin sheet of secretion is supplied and wound 
on, like the successive layers of fabric in a rubber hose. During 
ejaculation the middle membrane is shown to be very pliable, evidently 
much more so than in the squid. 
Orally, the middle membrane is firmly united to the outer tunic, 
where this tunic is joined by the cap (fig. 2mm’). Toward the aboral ex- 
tremity of the ejaculatory apparatus the middle tunic becomes thinner 
and almost if not quite disappears before the extremity is reached. 
Inside the middle membrane is the inner membrane. This is so 
thin that in all the figures it has been shown as a single line, but it is 
always distinguishable under a moderately high-power lens. This is 
also firmly united to the outer tunic at its oral extremity (fig. 2 mm) 
and extends throughout the length of the ejaculatory apparatus. 
Inside the inner membrane is a spirally coiled filament that is 
present but hard to distinguish near the oral end of the ejaculatory 
apparatus and becomes very much more prominent toward the 
aboral end of the ejaculatory apparatus (fig. 1 sr). This filament 
seems to be united to the inner membrane and has the appearance of 
an ornament on the inner membrane. It is evidently very flexible and 
does not break up into small fragments during ejaculation, as in the 
squid spermatophore. While it possesses elasticity and probably aids 
in keeping the ejaculatory apparatus from collapsing, there is no 
evidence that it possesses any spring properties. The diameter of the 
coil differs greatly. Toward the aboral end of the ejaculatory appa- 
ratus, where the middle membrane thins and probably disappears, the 
diameter of the coil of the spiral filament is greatly increased. It 
then narrows and near its aboral end becomes thin and relatively 
weak. This arrangement is significant in the act of ejaculation. 
The lumen of the ejaculatory apparatus inside the spiral filament is 
filled with a viscid material that adheres to the spiral filament and 
inner membrane and becomes spread over the outside in an irregular 
manner during ejaculation (figs. 2,12,and 15 uc). It is not so liquid 
as in the squid spermatophore and does not form a definite rod-like 
plug like that described by Racovitza (1894 c) for Rossia. There is 
frequently evidence that the core of this secretion is much more liquid 
than the outer parts (fig. 2 Hc’). Extending into the cap of the sper- 
matophore from this region is an indefinite, hazy appearance evidently 
due to the escape of material from the lumen. This appearance never 
involves the entire width of the lumen but only the central part. The 
remainder of the material is evidently responsible for forming the 
papille-like ornaments over the outside of the evaginating tube that, 
