MUSEUM OF COMPARATIVE ZOOLOGY. 35 



or root end. The spaces between the stalks are much greater than those 

 between the heads ; while the latter sometimes remain — even after the 

 prolonged action of water — in a continuous layer, the stalks often 

 appear to stand individually isolated. It is more common, however, to 

 find, as the result of the swelling, that both the heads and the stalks are 

 arranged in groups or patches, — better shown in tangential sections. 

 Even when the heads are not thus separated, the stalks may be gathered 

 into clusters which leave in radial sections broad lenticular spaces be- 

 tween them (Plate III. Fig. 1). 



The stalk gradually diminishes in size to near its zonal end, w T here it 

 enlarges rather promptly into a sort of conical foot, w^hich exhibits dark 

 longitudinal or radiating markings continuous with the dark outer ends 

 of corresponding pore-canals in the zona. In some cases the foot is 

 split into two or three strands, between which there is then left a space 

 that in radial sections is triangular, with its base resting on the zona 

 and its more acute angle rising into the stalk. The roots proper embrace 

 only the portions of the villi still occupying the pore-canals of the zona. 

 In some cases they are to be recognized as occupying every pore-canal, 

 in others some of the canals appear to be destitute of villous contents. 

 The roots are highly refractive, like the stalk, and seem to stain even 

 more deeply than the latter. They are always broadest at the outer 

 end, and taper until they are exceedingly fine threads. They seldom 

 reach more than a tenth or an eighth of the way through the zona, 

 although longer and larger roots are met with at intervals. They 

 always appear more tortuous — zigzag, or spiral — than the pore-canals 

 which do not contain roots, and are at times so irregular in form as to 

 have caused great distortions in the canals (Plate IX. Fig. 2). Their 

 finest tips, however, always appear continuous with the much more 

 faintly marked pore-canals. I cannot doubt, therefore, that they are 

 accommodated by simple enlargements of the pore-canals. The great 

 regularity in their distribution, too, allows no other interpretation than 

 that the position of the roots is practically determined by that of the 

 pore-canals. 



Tangential sections of stained eggs (Plate III. Figs. 2-5) afford the most 

 satisfactory evidence of the shape and grouping of both heads and stalks, 

 and is the only safe means of controlling the views of the foot region 

 gained by radial sections. The heads are at first close set, leaving only 

 the finest narrow lines, with here and there an irregular opening where 

 the prisms incompletely match (Plate III. Fig. 2) ; their cross sections are 

 angular and range from variously proportioned triangles to six- or seven- 



