36 BULLETIN OF THE 



sided polygons. After the prolonged action of water they become less 

 angular, and begin to separate along irregular lines, so as to leave the 

 heads arranged — as already indicated — in patches, which vary con- 

 siderably in size but are for the most part of a characteristic polygonal 

 outline, with borders which are necessarily jagged owing to the nature 

 of the lines of separation, for the latter never split a prism, but simply 

 separate adjacent ones. The heads may vary in diameter in the ratio 

 of one to two. 



The dark border already alluded to is best seen in thin tangential 

 sections (Plate III. Fig. 3), and is readily distinguishable on all the 

 heads when well stained and cut sufficiently thin. The line of separa- 

 tion between prisms is not always distinguishable, but whether this is 

 due to actual contact or not it is difficult to say, since the least obliquity 

 in the section is sufficient to obscure so faint a marking. 



The stalks also are found upon cross section to be prismatic, even 

 after the process of swelling has completely isolated them (Plate III. 

 Fig. 4). They are also arranged in groups which correspond fairly to 

 those of the heads, but the spaces between them are much greater. 

 Occasionally sections of stalks are to be seen, even from the middle of 

 the stalk-zone, the central part of which has not been stained (Plate 

 III. Fig. 4, a, a). Careful examination shows that such stalks are really 

 hollow, the boundary of the colorless area being sharply defined. I have 

 never seen vacuoles in the middle region of any of the stalks examined 

 in radial section ; besides, these cavities can often be traced continuously 

 on successive tangential sections toward the foot. They are, moreover, 

 increasingly frequent as one approaches the zonal attachment of the 

 stalks. The consideration of all these facts makes me quite sure that 

 many of the stalks, at least in their basal halves, are really hollow 

 prisms, although I have never been perfectly certain that I have seen 

 this condition in radial sections. One may, however, as before stated, 

 readily see on radial sections that the expanded foot of the stalk is often 

 apparently split into diverging roots, and that there is an intervening 

 unstained region. The prolongation of this space, which is triangular 

 in side view, forms, I believe, the cavity of the stalks in question. Al- 

 though the prisms appear sharply marked in cross sections, there is very 

 generally a trace of a filmy substance projecting here and there from 

 their edges in the form of faintly marked threads, which sometimes end in- 

 definitely in the inter-prismatic spaces, but at other times appear loosely 

 to connect neighboring stalks (Plate III. Fig. 4, /?). This substance 

 seems to stain less deeply than the stalks themselves, but it is exceed- 



