310 
Part III. — Eighth Annual Report 
The ordinary laws of endosmosis and exosmosis are quite sufficient to ac- 
count for the presence of sea-water in the egg capsule, and to provide for its 
aeration. That the slits have some other function I feel quite certain, and 
possibly this has something to do with the preservation of the vital 
contents of the egg under different pressures (either natural or accidental). 
The yolk-absorbing function of these filaments has been maintained by 
Dohrn, who believed that he had detected a yolk-emulsion in the blood con- 
tained in the vessels of the filaments. At first, Dohrn says, he attached 
no importance to the appearance, but when he afterwards found the same 
yolk-emulsion in the roots and main trunks of the branchial arteries, 
he began to make further investigations. These taught him that not 
only were the gill-filaments full of this emulsion, but also the blood-cor- 
puscles themselves of the filaments were filled with it. 
As Dohrn never investigated this matter on living embryos, his researches 
remain incomplete. I have studied the gill-filaments of skate embryos in 
both live and preserved specimens. In none of those specimens and in 
none of my numerous preparations of other elasmobranch embryos of all 
ages have I been able to detect any trace of a yolk-emulsion. Dohrn also 
referred, in support of his argument, to the bladder-like widening of the end 
of the gill-filament of Raja and some other forms. I have figured a 
specimen showing this widening in fig. 18, which is taken from a filament 
of an embryo blue skate aged 23 weeks. One does not meet with this 
appearance in every gill-filament, and, for my part, I explain the widen- 
ing as due to a simple hypertrophy of the capillary vessel, which results 
in a slower flow of the blood, and gives opportunities for its increased 
aeration. 
There are many interesting morphological questions which arise in con- 
nection with the gill-filaments, but this is not the place for their discussion. 
The filaments arise very early as buds or processes on the hinder border 
of all the true branchial arches. Their development has been described 
by Dohrn, and it is figured by me in surface view in fig. 15 (b.f.). A 
much later stage, in which the filaments are functionally active, is shown 
in fig. 17 (b.f.). An embryo enclosed in a strong shell like that of the 
skate-purse cannot breathe after the manner usual to fishes, even when its 
muscles, etc. are well developed. It must either breathe by its yolk-sac 
circulation or by some other special apparatus. The yolk-sac circulation 
is increased in such forms as Mustelus, in which a sort of placenta is 
formed, but here the gill-filaments also help. Above the fishes, i.e., 
beginning with the reptiles, the allantois takes the place of external gills, 
and initiates the formation of the placenta. In the bony-fishes in which 
external gill-filaments, like those of elasmobranch s, are never formed, a very 
different sort of respiration may be established, if needed. I am now 
referring to the young of the viviparous Blenny (Zoarces viviparus), which 
grows for some eight or nine months within the mother. My friend, Mr 
J. C. Mitchell, drew my attention to numerous large intestinal villi in the 
intra-uterine young of this species. These are very vascular, and from the 
curious muscular and barrel-shaped structure of the intestine, I do not 
doubt, though as yet I cannot prove it, that the intestine in the intra- 
uterine life of this form functions as a respiratory organ. 
In the skate-embryo the filaments are said to disappear shortly before 
hatching. It may be expected that their atrophy commences when the 
purse ruptures sufficiently to allow of the passage of sea-water directly to, 
the embryo. Then the ordinary piscine mode of respiration would be 
initiated, and the external gills would disappear. 
