222 PROFESSOR FRANK J. COLE 
is further complicated by the overlapping of the gills, any gill, except of course the 
first, overlapping externally the one in front as far as the origin of its efferent gill duct. 
The result is that in the posterior gills those parts of the pleural sacs spreading over 
their efferent gill ducts course backwards side by side, and these portions of the sacs 
communicate freely with each other. It is hence not strictly correct to say that the 
gill sacs only communicate indirectly with each other through the medium of the 
longitudinal ventral and dorsal sinuses. 
The last gill sac communicates freely above with the sinus dorsal to the cesophagus, 
which itself ceases to exist behind the branchial region. Posteriorly, the last gill sac 
terminates blindly by a somewhat irregular border. 
In the mid-ventral line, and situated just above the inferior jugular vein, is a large 
longitudinal sinus, in which courses the cardiac aorta. Anteriorly, this sinus splits into 
two wide channels, which pass upwards and forwards in a curve to open into the first 
pair of pleural sacs. In the neighbourhood of the split another pair of channels are 
given off, which course straight up into the second pair of gill sacs. Similarly, four 
other pairs of channels arise from the median longitudinal sinus to open into the four 
posterior gill sacs. In the main channel and its branches are situated apparently the 
cardiac aorta and the afferent branchial arteries—connected with the walls of the sinus 
by means of numerous fine but strong threads. The exact relation of the arteries to 
the cavities in which they appear to lie can only be ascertained by a study of their 
development, but we can hardly be wrong in concluding, on @ prior: grounds, that the 
arteries are situated morphologically outside the cavities. 
The median sinus is prolonged backwards behind the exit of the last lateral channel 
on to the base of the ventricle, but here its cavity becomes much broken up by numer- 
ous attachments between its visceral and parietal wails. It has no communication 
with the pericardial cavity, nor is there any justification for regarding it as the peri- 
cardium itself. 
In Petromyzon, according to J. MUuusr, the gills lie in closed sacs, and there is no 
ventral longitudinal sinus. The afferent gill arteries pass to the gills between the two 
abutting walls of contiguous pleural sacs, thus differing from the Myxinoids in a striking 
manner. Although J. Mttier, RaruKke, and Rosin noticed blood in the so-called 
lymph spaces, it was LancERHANS* who first established any connection between these 
spaces, which he compared with those of Amphibia, and the blood vascular system. He 
injected the subcutaneous sinus, and found that the injection passed first of al) into 
other and internal lymph spaces, and finally reached the veins and heart. Ewarrt 
also states, independently of former writers, that the peribranchial spaces of the 
Lamprey normally contain blood. He did not succeed in finding any connection wa 
the arteries, but was able to fill the spaces by injection from the veins. He believes 
the veins concerned are the internal jugulars. At about the same time ScHNEIDER } 
* Verhand. nat. Ges. Freiburg, Bd. vi., 1873. + Jour. Anat, and Phys., vol. xii., 1877, 
t Beitr. z. vergleich. Anat. u. Entwick, d. Wirbelthiere, Berlin. 1879, 
