98 THE AMERICAN MONTHLY [Apr., 



in a family of such meagre development, where the tri- 

 cliome root-fibers still take the place of specially organ- 

 ized roots, and which is just above the plane of evolution 

 where stems and leaves have become differentiated from 

 each other. The antheridia seem sometimes to be 

 formed by a direct prolongation of the axis itself, or else 

 of lateral shoots (branches) from it ; sometimes to repre- 

 sent leaves (as stamens do) ; and sometimes they appear 

 as trichomes (hairs) by their development from cells of 

 the epidermis in indefinite numbers and positions. 



In our object, as shown (X15) in figure 1, they might? 

 from their appearance and plosition, be either branch- 

 shoots, like the separate flowers in a head of daisy ; or 

 leaves, like the stamens of a buttercup ; or trichomes, 

 like the sporangia of the ferns. The last theory might 

 seem most plausible, from their own appearance and 

 that of the plant as compared with the ferns, but it 

 would involve the more or less improbable combination 

 of two different sorts of trichomes, these and the para- 

 physes, in the same cluster, and without the intermix- 

 ture of any abortive or intermediate forms ; the first 

 theory, likening them to the flowers in a composite head, 

 is also tempting, but this would suggest the paraphyses 

 (which seem plainly trichomes, and not leaves) as repre- 

 senting the leaves, like the scales or chaff in the flower- 

 ing heads of the Compositse, to which these antheridia 

 would be, however disguised, axillary branches ; the 

 second theory, likening them to ordinary stamens, 

 besides fitting well to themselves, accounts easily for 

 the paraphyses as hairs growing on the neighboring 

 portions of the stem, which seems on the whole the most 

 reasonable and probable. If the paraphyses be consid- 

 ered abortive leaves, then the antheridia would be con- 

 sidered branches. Further evidence of their real nature 

 should be sought at an earlier stage of growth. 

 The antheridia are often clustered, as here, on a dis- 



