INTEODUOTION. xi 



Desmobrya. I believe the idea of using this character as one 

 upon which to found a primary division of the annulate ferns 

 originated with myself, (see Phytol. ii. 373) ; but Mr. Smith was 

 the first to apply the idea, and to name divisions founded on the 

 differences pointed out. It must, however, be observed, that 

 Mr. Smith, in his primary divisions, lays great stress on a cha- 

 racter which now appears to me of somewhat secondary impor- 

 tance : I allude to the articulation of the stipes to the rhizome. 

 My own conclusion, from a careful examination of the species 

 within my reach, is, that the grand distinctive characters are 

 these : — First, that the rhizome of the Rhizophyllacese, and 

 the caudex of the Cormophyllacese, are not the same ol'gan : 

 that the rhizome is a root ; the caudex a stem : that the rhi- 

 zome never terminates in a frond ; that the caudex always 

 does : indeed, that its apex is constituted of fronds undeve- 

 loped ; its trunk, of the bases of fronds that have decayed. 

 Secondly : that the growing apex of a rhizome is always in 

 advance of the fronds ; that the fronds are always in advance 

 of the growing apex of a caudex. There are two other and 

 possibly less constant diagnostics : the rhizome of the Rhi- 

 zophyllaceje is scaly, the stipes naked ; the caudex of the 

 Cormophyllaceae is naked, the stipes densely paleaceous : the 

 formation of the fruit of the Ehizophyllacese follows the deve- 

 lopment of the frond ; in the CormophyllaceEe it precedes it. 

 In some Cormophyllacese there is a tendency to approach the 

 RhizophyUaceae : this is strikingly the case in Dryopteris, Phe- 

 gopteris, and Thelypteris ; but it is only necessary to examine 

 the growing apex of the rhizomatiform caudex of these well- 

 known ferns, in order to ascertain that it is always composed of 

 undeveloped fronds. There is a plant familiar to every one who 

 has a garden, that affords an illustration of the two modes of 

 growth, — the common Pyrus japonica. The branches of this 

 beautiful shrub always terminate in a bud, composed of unde- 

 veloped leaves ; such branches, therefore, are analogous to the 



