NUCLEAR CHANGES IN RANA CLAMITANS. 125 



and corresponding gradations in size and depth of stain. In 

 cases of this sort there are often vacuoles or cytoplasm between 

 the nuclei as if the latter had shrunken, whereas in the normal 

 cord, the nuclei are so close together that no cytoplasm can be 

 seen between them. These facts indicate that normal nuclei 

 become deeply staining nuclei by contraction or by loss of achro- 

 matic material. 



This "contraction" of nuclei seems to be caused by contact 

 with the water or killing fluid, or the succession of the two, as 

 well as by direct injury from the scalpel, for other nuclei which 

 are in contact with the exterior only through the central canal 

 show this phenomenon. In some cases, the end of a nucleus 

 nearest the central canal is deeply stained and contracted while 

 the other part is normal (Fig. 4)- The question immediately 

 arises, why does not the water or other external factor enter the 

 open neural tube and cause the contraction of the inner parts of 

 practically all nuclei in the spinal cord? It is probably because 

 of the presence in the tube of some substance which prevents the 

 ready admission of external fluids, though capillarity would have 

 a similar effect. Since the sections show very little structure 

 within the central canal, this content must be liquid or semi- 

 liquid. However, in a number of sections there is a rather long 

 narrow band of cytoplasmic material which may be the more solid 

 part of a semi-liquid substance coagulated by the killing reagent. 

 There are other evidences of the presence of such a liquid. The 

 sections from two of the tadpoles killed one hour after the oper- 

 ation show a coagulation of the outer surface of the blood plasma 

 covering the wound, but over the spinal cord this coagulating 

 process is delayed. The most plausible explanation seems to be 

 that some cerebro-spinal fluid (compared by Barfurth to the 

 cerebrospinal fluid of mammals) exerts an outward pressure 

 which breaks through any slight hardening of the plasma at this 

 point. Perhaps transference of the animal to a medium of 

 different density, the kiUing fluid, aids the outburst. Sections 

 of another tadpole killed at one hour show the presence of this 

 coagulated plasma over the end of the spinal cord as well as over 

 other parts of the tail. 



The outward pressure of a fluid would tend to push out into 



