562 T. H. Morgan and N. M. Stevens. 



anterior to the tentacle anlagen. When they are seen within the 

 tentacle region they are by no means evenly distributed over the 

 surface, but occur in one or more irregular masses so slightly 

 attached to the wall that they are often washed away by the cur- 

 rent. These masses of pigment and other debris are sometimes 

 carried to the basal end or to a sharp bend in the stem, remain 

 there for a time, and later appear again at the distal end. Such 

 being the case, the pigment granules could with difficulty be 

 imagined to be a hydranth-forming substance. 



Third. A microscopic examination of the pigment in the wall 

 of the tube, and also in the circulating fluid, shows that it is com- 

 posed of irregular crystalline-like red and yellow particles of 

 various sizes and shapes. In addition there are spheres of an 

 oily appearance that may represent reserve food material. These 

 spheres also help to give a general yellowish tinge to the stem, 

 especially to old stems and stolons, but the main color in regions 

 of new growth is due to the red pigment. Where a new hydranth 

 is developing, more pigment begins to appear in the endoderm 

 and a proportionately larger number of small granules seems to 

 indicate that this pigment is manufactured by the cells themselves. 

 That the endoderm cells have this power there can be no question 

 from the facts to be mentioned presently. Why, then, since this 

 can be shown to be the case, imagine another wholly superfluous 

 process to go on, viz: an absorption of red pigment from the 

 circulating fluid t 



Fourth. If the amount of pigment in the entire circulation be 

 compared with the amount of new pigment in the developing 

 hydranth, it will be found that the latter is many times in excess 

 of the former and hence must have been formed in situ. This is 

 especially noticeable in very short pieces. 



Fifth. The amount of pigment set free by the degeneration of 

 the ridges in the hydranth-forming region is approximately the 

 same for short and for long pieces. Hence the relative amount 

 in the circulation is far greater in short pieces, yet these, if moder- 

 ately short, develop no faster than do longer pieces, and if very 

 short it takes a longer time for the hydranth to regenerate. In 

 the latter case, however, though not in the former, the problem 



