CONFERVOID FILAMENTS OF MOSSES. 581 



two cilia, and their contents were granular (PI. LVIII. fig. 23 e). The smaller measured 

 about xwo by s^Vo^Ji of an inch. 



Thus far I have dealt with the confervoid filaments and their gonidia. I shall now 

 point out two other methods by which independent cells are set free from Mosses, pos- 

 sessing the power of segmentation, and doubtless of reproducing the parent Moss, 

 although I have not been able to trace them back again to the parent. 



It has been mentioned above, that in the thallogenous plants the power of producing 

 gonidia is diffused through almost any part of the thallus, but that in the Mosses it is 

 more localized. Yet it is not entirely confined to the confervoid filaments. 



I have frequently noticed that Gloeocapsa-lik^e cells are produced from the contents of 

 the cells of the older leaves, which, situated at the base of the stem, towards autumn and 

 during winter and spring have become brown. These leaves are not wholly dead. It 

 is their cell-walls only which, having become brown, give them the appearance of 

 dead tissue ; but the contents as yet retain their vitality and green colour within. 

 Their condition is precisely similar to that of the cells of the filaments described above, 

 whose walls have become dyed of a broAvn colour. After a time the old cell-wall 

 dissolves away, and then it becomes evident that the contents have assumed the form of, 

 or rather have become a Gloeocapsa, which certainly undergoes segmentation freely. I 

 have shown a portion of some of these leaves in this state in PL LVIII. fig. 24 ; a portion 

 of the margin has lost its dark cell-wall, and has already produced Gloeocapsa (PI. LVIII. 

 fig. 24 a). I have seen considerable masses of Gloeocapsa produced in this manner. 



Another very curious mode of forming free cells, capable of segmentation, is the 

 following. 



The axis springs up in the usual way, and proceeds to form leaves. The cells, how- 

 ever, which should in the ordinary way unite to form their lamina, in this case do not 

 cohere, but either run parallel to or branch away somewhat from each other (PI. LVIII. 

 fig. 19). The terminal cell of each of these pseudo-leaves possesses the power of 

 separating from the others (as the terminal cell of the filaments did, see PI. LVII. 

 fig. 10). In the instance from which I give this description, it might be noticed that 

 the freed cell was more or less of a pointed oval form, one side being, however, much 

 straighter than the other (PI. LVIII. fig. 19 c), and having somewhat the appearance of 

 the frustule of Isthmia nervosa. Each of these cells might be noticed to have already 

 begun to divide by binary division before their separation. 



Large numbers of these cells were shed, and disseminated readily. Many for the time 

 remained in contact with the structure from which they sprang, and gave the whole 

 the appearance of little yellow dots on the ground whence they arose. I cannot tell to 

 what Moss it belonged, but there were numerous examples of it. From the same stem 

 confervoid filaments arose, and some («) were giving rise to Gloeocapsa, as before men- 

 tioned. 



In PL LVIII. fig. 25, I have shown a condition of the axis of one of the Jioigermannke 

 from Jamaica, precisely like that of the Moss shown at PL LVIII. fig. 19. I observed 

 it in plenty in a hot-house, where the plant grew in abundance. The whole had very 

 much the appearance of an Alga, and was a very beautiful object, from the transparency 



