148 BULLETIN OF THE UNITED STATES FISH COMMISSION. 



tube, for such it is from the fact that a furrow appears ou the outer sur- 

 face of the hhistoderm kuown as the medulhiry groove which exteuds 

 from the head end of the embryouic area of the blastoderm to the rim 

 and which causes the blastoderm to be pushed dowu before it. This 

 groove first closes at the head of the embryo, while it remaius open for 

 a considerable time at the tail end. The cells of its walls form the em- 

 bryonic spinal canal which afterwards becomes the spinal cord, brain, 

 and retina of the more advanced embryo. 



Fig. S represents an older embryo Spanish mackerel, eleven hours 

 after development began. It is seen with the head towards the ob- 

 server, and behind or beyond the head on the opposite side of the egg 

 the rim r of the blastoderm is seen through the transparent vitellus, 

 which will close over the latter entirely in the course of another hour. 

 The segmentation cavity sg is shown in optic section between the epi- 

 blast ep and the hyj^oblast hy, extending all round the egg except the 

 portion taken up by the embryo above, and that part not yet covered 

 by the blastodermic rim r on the opposite side of the egg. 



Fig. 9 represents an egg of the twelfth hour of development with the 

 caudal pole turned towards the observer. The small area y is all of the 

 vitellus which now remaius uncovered, and the blastodermic rim is con- 

 tracting and will soon close at the end of the tail of the embryo, when 

 the formation of the blastoderm may be said to be completed. The seg- 

 mentation cavity sg has of course been somewhat extended on account 

 'of the approximate closure of the blastodermic rim. In Fig. 11 a section 

 is carried transversely from left to right through the opening y, which 

 still remains behind the tail of the embryo represented in Fig. 9, to show 

 how the segmentation cavity is finally limited at the caudal end of the 

 embryo by the blastodermic rim, which after its closure takes a large 

 share in the formation of the tail of the j'oung fish, as i>ointed out by 

 Kuj)ffer and His. It may be considered the true tail-swelling, as it 

 thickens into a round, knob-like prominence immediately after the in- 

 clusion of the yelk is accomplished. In consequence of the remarkable 

 extent of the segmentation cavity in fish embryos, as amply x>roved by 

 our studies upon a number of widely-separated genera, the yelk becomes 

 inclosed in a shut sack derived from and including the greater part of 

 the innermost embryonic layer or hypoblast, which is surrounded by a 

 film of fluid, and which is agaiu inclosed by an external epiblastic sack. 

 Besides the demonstration of this structure in the living egg I have used 

 the following method with success: Immerse a living embryo in a one- 

 half \)er cent, solution of osmic acid for a few minutes, or till the outer 

 sack becomes light brownish, then place it under a Fol's compressor 

 and gradually bring pressure to bear on the vitellus while it is under 

 observation in the microscope, when the outer covering, now made brit- 

 tle by the osmic acid, will rupture and expose the vitelline or hypoblastic 

 sack; the epiblastic covering will frequently open, wrinkle, and slide 

 back off of the vitellus like a cowl when pushed back off of the head. 



