KEVISION or THE GENUS XAJAS. 383 



dichotomy of an axillary branch. Second, the archesporium is produced in the tip of 

 the floral axis. Third, the homology of the envelopes is difficult to explain on the 

 ordinary terms of definition of the parts of a flower. 



In course of development the cup-like wall surrounding the ovule and the outer 

 envelope of the anther are homologous, while the inner envelope, which remains closely 

 adherent to the anther, corresponds with the integuments of the ovule. 



Magnus proposes a very ingenious explanation. He regards the carpel-like structure 

 in the female flower as the homologue of the cup-like envelope round the group of 

 carpels in the allied genus Zannichellia. This arises similarly as an annular outgrowth 

 of the floral axis, hut surrounds more than one rudiment, each of which ultimately 

 develops into an ovule surrounded by a stigma-bearing ovary-wall. Magnus suggests 

 that the arrangement in Najas may have arisen from such a type by the suppression of 

 all the ovules but one, and all the ovaries, leaving a naked ovule surrounded by a cup- 

 shaped envelope (or j)erianth), which then developed stigmatic appendages. Najas, on 

 this view, becomes a gymnosperm, the female flowers of which are provided with one, 

 more rarely two, sac-like perianth-envelopes, corresponding with the two in the male. 



K. Schumann (Mart. !F1. Bras. iii. pt. 3, p. 720) opposes this vicAV, insisting on the 

 homology of the stigma-bearing envelope (or ovary) in the female and the two-lipped 

 inner envelope in the male. He regards the female flowers as always consisting of a 

 naked ovary, and the male as enveloped with a single envelope which may be regarded as 

 perigone or involucre at pleasure. He denies the existence of the similar envelope which 

 Magnus had described in the female flower of N. indica and others. He was, however, 

 occupied almost exclusively with South- American species, in which it does not occur. 

 Investigation shows this outer envelope to be more generally present. 



Campbell, in the memoir already referred to, draws attention to the primitive 

 character of the flowers, and shows also that in Zannichellia we are dealing with a 

 group of monocarpellary Howers, l. e. an inflorescence, not a single pluricarpellary 

 flower. 



It seems to me that the key to the whole arrangement lies in this primitive 

 simplicity of the flower. We have in both male and female an axial structure con- 

 taining sporogenous tissue which develojis respectively into structures obviously com- 

 parable with normal anther and nucellus, the latter becoming surrounded by integuments 

 forming a normal ovule. The ovule is surrounded by a cup-like outgrowth, w^hich 

 recalls the development of the ovary in Folygomim or Bumex (see Payer, ' Traite d'Orga- 

 nogenie,' pis. 64, 65), and which has the appearance and structure of an ovarv. The 

 inner envelope in the male flower I have regarded as a perianth. It is a lateral out- 

 growth of the floral axis l)elow the androecium, which it protects, and has therefore the 

 characteristics of a perianth which has arisen rather late in the history of the flower- 

 development. The outer sac which characterizes the male, and is occasionally present 

 in the female, I have called a spathe, remembering, however, that it is simply an out- 

 growth of the axis which ends in the flower below that flower. It is, I think, comparable 

 with the spathe so characteristic of submerged monocotyledonous water-plants, which 

 may have, moreover, a very similar appearance (e. g. Layaroslphon, Hydrilla, &c.). It 



