118 
ME. W. K. PAEKEE ON THE STEUCTUEE AND 
primary “ superorbital bar” (Plate I. figs. 1, 2, 3, 4, 6, 7, s.ob .) ; it is much more than 
the seat of that frequent additional bone which overshades the eye in Fishes ; in the 
fifth stage I shall describe the cartilage which develops in it as a boundary to the “ great 
fontanelle.” 
The auditory sac is only partially invested with cartilage, and at this stage it is an 
irregular and fenestrate shield (figs. 2 & 5, au.), imperfect within, above, and below 
(Plate I. figs. 2 & 5, and Plate II. figs. 2 & 4). As far as I can see from various views, 
the great blastodermic involution which forms the ear-sac shuts off the labyrinth before 
the outer opening begins to close, although this gaping space itself is very temporary. 
The sac itself seems to be formed from the epiblastic layer, which has developed a roof 
to the labyrinth over the rudiments of the three semicircular canals, thus shutting off the 
new sense-organ from the outer skin and its primordial opening. But the mesoblastic 
layer has developed immediately around the new racemose labyrinth (Plate I. figs. 2 & 5) 
a partial investment of young cartilage ; this cartilage has not crossed over below the 
primary opening, nor has it finished the floor (see Second Stage, Plate II. fig. 4), a 
deficiency being left there for several months ; nor does it ever in this Telostean wall-in 
the labyrinth, craniad. The finishing of the upper and lower faces of the periotic cartilage 
is remarkable and instructive; and whilst the primordial opening of the sac is closing-in 
the semicircular canals are getting more roof, which grows from before backwards, and 
which is not completed in newly hatched fry (see Fourth Stage, Plate III. fig. 6, ep.). 
No such gradual or even partial growth of cartilage is seen in the Frog; in that type 
the epiderm, at least , is closed over the sac before hatching, when the head and tail of 
the embryo pitch very little beyond the yelk-mass. Then, indeed, there is a primordial 
“fenestra ovalis” (“Frog’s Skull,” Plate in. fig. 3, and Plate iv. fig. 1), but otherwise 
the cartilaginous ball is complete, except where it allows the “ portio mollis ” to enter 
on the opposite side. Soon (that is in Tadpoles less than half an inch in length) this 
fenestra is filled in, again to reopen by the segmentation out of its own substance of a 
stopper, the “ stapes ” (op. cit. Plate iv. fig. 7, and Plate v. figs. 1, 2, & 4). 
We shall see the last of the primordial fenestra ovalis of the Salmon in the seventh, 
or first summer stage, but I have already shown that there is no reopening. 
The auditory cartilage is entirely free at first, but as my first stage is passing into the 
second it coalesces with the “ investing mass ” by two connective bands that enclose the 
primordial “ fenestra ovalis ” (Plate II. figs. 2 & 4) ; this takes place a little before the 
fusion of the trabeculoe with the same plates. The foregoing description will be now 
supplemented, and in some degree recapitulated, by a description of the transverse sections. 
Two of these sections (Plate I. fig. 10, and Plate II. fig. 1) are made through the eye- 
balls, and two through the ear-sacs (Plate I. fig. 11, and Plate II. fig. 2). The first of 
these sections (Plate I. fig. 10) is of a very immature but unusually symmetrical embryo, 
one answering to figs. 1, 2, & 8 as to development. At this part the “ umbilical vesicle” 
(u.v.) is quite free from the projecting head of the embryo; a clear watery space inter- 
venes between the two parts, and this space is larger on each side of the flat head. Two 
