219 
wholly made up of undifferentiated cells which could from time to time 
become specialized to form connective tissue, follicular cells, or primitive 
sex-cells, as the case might be. 
The reader will find that I come to a definite conclusion in the 
present paper which is quite at variance with my former views. The 
present work has shown with sufficient clearness certain possible 
sources of error and misinterpretation in my former investigation such 
as to cast some doubt upon the conclusions there expressed upon this 
question. A re-study of the preparations used in that work, fully 
substantiates my former observations. One might, however, doubt the 
conclusions to be drawn from the observation of apparent transition 
forms connecting the peritoneal cells with the sex-cells. This has been 
the key note to the argument in favor of that view. The problem 
certainly demands further very careful and thorough investigation in 
the Mammals. The difficulties involved in such a work are, however, 
very great. 
In my previous work, I followed LorseL in applying the term 
“serminative” to those cells in the seminiferous tubules which are 
derived from the peritoneal cells. This would be a manifest misnomer 
when one arrives at the interpretation of them given in this paper. 
For this reason, we shall adopt Brenpa’s term and call them “vege- 
tative cells”. 
The species studied in the present work is Chrysemys marginata. 
The material was gathered largely upon the shores of Lake Wingra 
during the months of June, July and August. It includes all stages 
of development from the freshly laid egg, up to sexual maturity. 
Pieces of adult testicles were also fixed at various periods of the 
year. Since Chrysemys is found in this locality in great abundance, 
an unlimited quantity of material was available. Mention will be made 
only of representative stages of development taken from my collection. 
The fixation found most useful was TeELLEYESNICZKyY’s bi-chromate 
acetic mixture. Wherever possible, save in very early stages of de- 
velopment, HEIDENHAIN’s iron-alum haematoxylin stain was employed. 
At the time the egg is laid, gastrulation has already taken place 
and the ventral floor of the archenteron has disappeared. Neither 
neural tube nor notochord have been formed, although the anlage of 
the latter is indicated by the inclination of the columnar hypoblast 
cells. At this stage, all the cells of the blastoderm contain yolk 
spherules. These are of varying size. In general, the ectoblast and 
the portion of the hypoblast which lies within the area pellucida 
contain relatively small spherules, from 3 «u to 9 u in diameter, the 
