174 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. 
bubbles can be seen coming off freely from the inner surface 
of the uninjured wall. Similar experiments on the right coronary 
vein gave only negative results, and from neither the right nor the 
left vein could bubbling be produced from the inner surface of the 
ventricle. 
As this experiment can be successfully carried out with only a 
slight pressure of air, there is no reason to suppose that vessels 
were ruptured, and we believe we are justified in concluding that 
the left coronary veins have connections which open freely on the 
inner surface of the left auricular wall. These connections can be 
none other than the vessels of Thebesius. Attempts to blow from 
the inside of the auricle through to the left coronary vein always 
failed, doubtless because of the impediment offered by the spongy 
nature of the auricular wall. 
Experiments of a similar character carried out on the coronary 
arteries resulted in the production of small bubbles on the inside of 
the left auricular wall. This, however, was accomplished only after 
very vigorous blowing and consequently demonstrates that the con¬ 
nections between the coronary arteries and the veins of Thebesius 
are much more restricted than those between the coronary veins 
and these vessels, a condition already observed by Pratt (’98) in 
mammals. 
On inflating either the right or the left coronary vein of Raja 
erinacea with air, bubbling could also be demonstrated from the 
uninjured inner surface of the auricle, but no bubbling was ever 
observed from the inner surface of the ventricle. 
If the single opening of the coronary veins in Amia calva be 
inflated, bubbling takes place from the inner surface of the ven¬ 
tricle as well as of the auricle. This fish was the most satisfactory 
of the three species for the demonstration of the vessels of Thebe¬ 
sius. 
These experiments, in our opinion, show that the hearts of fishes 
possess veins of Thebesius which open into the ventricles as well 
as into the auricles and which connect more freely with the coro¬ 
nary veins than with the coronary arteries. 
Conclusions. 
When the blood vessels of the heart in fishes are compared with 
those in mammals, the most noteworthy feature is the striking simi- 
