216 PROFESSOR FRANK J. COLE 
Since writing the above I have examined the curious structures in question much more 
closely by means of serial sections, injections of various fluids and masses, and by whole 
preparations, and I am now able to add considerably to the description just quoted. 
1. Merruops. 
It will readily be admitted that if the vascular papillae are in open communication 
with the surrounding spaces, so that blood may pass from the arteries into the lymph 
sinuses or pleural sacs by which the gills are enclosed, we have to do with a fact so 
unexpected that nothing short of complete demonstration will ensure its acceptance. 
Assuming such a connection to exist, it may seem at first sight a simple matter to 
establish it. It is, however, by no means easy to do so; and I had continuously investi- 
gated these structures for some time, and had actually written withdrawing my pre- 
liminary statement, when some final injections settled the matter in favour of its 
accuracy. Such a contradictory result is partly accounted for by the fact that whilst 
in some papille there is an undoubted communication between the artery and the 
sinus, in others, and perhaps the majority, the communication has been closed, and 
the papillee are vestigial structures. 
My preliminary statement was based on the study of serial sections ; and if in some 
of those sections (e.g. figs. 2-5) the existence of the communication seems beyond 
question, it is so easy to misinterpret the sections that such evidence in itself cannot 
be held to be convincing. 
Injections of preserved material are useful, but not decisive. I have tried a large 
number of such injections with all the ordinary injection fluids and masses, and have 
never once found the medium to pass right through the gills into the efferent arteries. 
Hence negative results obtained by this method prove nothing either way. An 
injection, to be completely successful, must be carried out on an animal as soon as 
possible after it has been removed from the sea, and immediately after death, with the 
heart still beating. One can, in fact, almost deduce the time the animal has been dead 
by the speed with which it may be injected. A syringe injection may be so rapid that 
some of my first attempts, put on one side as failures owing to apparent leaking, were 
afterwards found to be injected throughout the whole body, and the leakage due to the 
injection mass, having returned to the heart, escaping through the cut auricle. 
Myaxine are not only difficult to kill—like so many marine animals in captivity they 
will neither live nor die—but their reflexes remain functional for a long time. This spoils 
many an injection. After trying a number of methods, I find the best is to snip off 
the tail immediately behind the cloaca, and to immerse the animal in warm sea-water. 
The latter to a certain extent acts as an anesthetic, whilst the main blood-vessels are 
drained, and the tail can be ligatured as soon as the injection mass begins to escape. 
Gelatine masses are unsatisfactory for two reasons: (1) they must be thrown in 
hot, and the passage of a hot mass through the vessels of a cold-blooded animal affects 
the elasticity of their walls so as to almost vitiate the result where very fine channels 
