294 HYDROID^. Part IV. 



in otHer instances, the projections ai-e but slightly developed {Fig. 9). The greatest 

 care is necessary in making such sections, in order not to break these partitions, 

 inasmuch as they are mostly very thin, oftentimes filmy, and brittle. A section 

 made by simply breaking the branch across, holding it very steadily in the fingers, 

 is much better than a cut by the section wheel. At the tips of the branches 

 the cells can hardly be recognized as such, but appear more like irregular depressions 

 of greater or less depth {Fig. 12, h cl e f). Between these the corallum is very 

 loose and spong}^, each cell communicating with the others through large, irreg- 

 ular channels, penetrating even to the centre {Fig. 12, a) of the branch. In 

 the specimen which we have figured (Fig. 12), the intercommunicating channels 

 are less numerous than in many cases ; for instance, a specimen now before me 

 is so thickly channelled, that the solid, calcareous deposit occupies much less room 

 than the open spaces. Passing down the branch, for half an inch, we come to 

 a j)oiut where the cells have a definite outline {Fig. 10) and the bottom {h) of 

 the cavity is clearly circumscribed. About the mouth («), or entrance, and between 

 it and that of the adjoining cells, the corallum is traversed by tortuous cavities 

 {i j), some like channels (/), and others like lacuna? (/), all of which communicate 

 freely with the cavity of the cell. Around the base {b) of the cell the corallum 

 [k) is more solid, and the intercommunicatiug channels {h) are smaller and fewer; 

 but around, and at the centre of the branch, we find, again, a sj)ongiform structure, 

 such as we have figured from a section lower down the branch {Fig. 11, a). Nor 

 is this absent at any age, even in the oldest part of the corallum ; at least we 

 have . found it at the centre of stems, from an inch to an inch and a half in 

 diameter. Sometimes, such is the irregularity in the rate of develojiment of the 

 branch, that we find the cells quite deep at the distance of half an inch from the 

 tip, and transversely divided into three or four superposed chambers {Fig. D). The 

 transverse partitions (c) which lie between these chambers are as thin and fragile 

 as the false partitions, but they are more regular, and seldom, if ever, perforated. 

 The same may be said for the oldest and deepest cells {Fig. 13). In fact there 

 is very little change in the structure of the cell after it has acquired three or 

 four transverse partitions; there are the same tortuous channels, both about the 

 youngest {Fig. 10, h i), the more advanced {Figs. 9, e / //, and 11), and the oldest 

 {Fig. 13); and beyond that, the corallum is, as we have described it in Fig. 10, k, 

 nearly solid, with only here and there a narrow channel, until we approach the 

 axis of the stem, where we always find a spongiform mass {Fig. 11, a). The 

 form of the cell, at all ages, is cylindrical {Figa. 9, 11, and 13), and the transverse 

 partitions are nearly uniformly arranged, at equal distances, one above the other, 

 and at such heights, that each included chamber is from one quarter to one third 

 broader than deep. The direction in which the cells trend is, more or less, along 



