454 REPORT—1590. 
matter of fact the pneumatic and bile ducts do grow out, one a little to 
one side of the embryo and the other to the other, a fact which 
unfortunately renders it impossible to obtain satisfactory median longi- 
tudinal sections. Yet in this case inspection convinces us that the devia- 
tion has no morphological significance. We may, in fact, perhaps best 
consider the alimentary canal as slightly twisted in this region, and if this 
twist became permanent we should have the pneumatic duct opening 
laterally (supposing it to persist). 
For the purpose of studying the question of the air-bladder I have 
preserved all stages possible of fish of which I could obtain the eggs in 
sufficient quantities. Of pelagic eggs floating freely and separately I 
obtained eggs of Labrax lupus (spawned in the aquarium) at the end -of 
January, and of Coris Giofredi (by artificial fertilisation) at the end of 
May. Obviously suitable for obtaining a series of stages are those eggs 
which float about connected together by a transparent (albuminous ?) egg 
case, such as those of Scorpena and Fierasfer, obtained in June and July, 
the latter being apparently rather plentiful. Suitable alsoare those which 
are attached by fibrous web-like tissue to seaweeds, as those of Cristiceps, of 
which I obtained a few in April. 
Of eggs that are attached side by side to rocks I have obtained those 
of Blennius in July, and a practically unlimited supply of the eggs of 
certain species of Gobius, such as G. paganellus, and another with eggs 
resembling those of G. niger. CG. capito was also plentiful, and there was 
a fourth of which I had a few. Finally a rich and obviously convenient 
source of material has been the Syngnathide. In the spring months, 
February to May, there was a plentiful supply of Syngnathus (various 
species), Siphonostoma, and Nerophis, and in the summer months of June 
and July these were replaced by Hippocampus, also very numerous. 
I have already cut a large number of sections from this material for 
various purposes, but as yet only part of the above types have been so 
treated. It is not, therefore, surprising that I have no conclusion at 
present to state as regards the air-bladder, especially as what I have as 
yet seen of the development of the alimentary canal of Teleosteans and 
Elasmobranchs, together with what I have read, has strongly suggested to 
me a line of enquiry which is new, the value of which, therefore, requires 
to be tested by the study of other types. When I have received eggs in the 
earliest stage I have, of course, been glad to be able to follow the develop- 
ment throughout, and have, in some cases, preserved a series of stages 
from the beginning. I became rather interested in the early stages of 
segmentation. Most of the eggs of Syngnathus that I received of this age 
were either dead or died soon after removal from the pouch, but I was 
able to watch the early segmentation processes in Coris Giofredi and 
repeatedly in Gobius paganellus and another species of Gobius (perhaps 
G. niger, perhaps G. jozo) which I will call Gobius b., in accordance with 
the label in my series. This last species specially interested me. There 
is here a very small proportion of yolk for a Teleostean egg, less 
apparently than in G. paganellus. 
In the case of this Gobius b. the first four segmentations (7.e., the 
stages until the egg is divided more or less completely into sixteen) 
exactly correspond to the first four segmentations in the frog as usually 
described. Thus the first two segmentation planes are vertical, the third 
enero and the fourth vertical and bisecting the angle between the 
rst two. 
