TRA^-SACTIONS OF SECTIOX K. . 1011 



from one condition to the other, and how finally the growing point itself is influ- 

 enced by the external agency (apparently light), which leads to a change of the 

 leaf-arrangement. This seems to be the case in many other Phanerogamic plants. 



A particularly interesting account is also given of similar changes in Selaginella. 

 Some eight species are orthotropic, radial, and isophyllous. S. sangzdnolenta shows 

 a direct response to external conditions, being upright and isophyllous in bright 

 and dry situations, plagiotropic and anisophyllous in damp and shady situations. 

 The bulk of the genus are, however, either plagiotropic and anisophyllous through- 

 out, or some may have an early orthotropic stage. But he concludes that even in 

 ' liabitually ' anisophyllous Selaginellas we have to do with an adaptive character, 

 induced probably by light. 



We see then good evidence that in certain cases the dorsiventral shoot is a 

 result of adaptation, and the radial probably the primitive. Was this always so? 

 We need not discuss the case of the gametophyte, as the problem there is even 

 more varied and difficult, and does not at the moment engage our attention. 

 But the question whether in the sporophyte the radial was in all cases the primi- 

 tive type is clearly related to our theory of the strobilus. The sporo^onia of 

 Bryophytes are, with few exceptions, orthotropic, and almost uniformly radial ; 

 e.xceptions such as Diphyscium and Buxhaumia have been shown to have an inter- 

 esting relation to the incidence of light, and are readily recognised as derivative. 

 The distinctively strobiloid Pteridophytes mostly maintain this radial structure ; 

 this may be so both in strobilus and vegetative organs, as in Eguisetum, Isoetes, 

 in most species of Lycopodium, and in some Selaffinellas ; or the vegetative 

 region may be dorsiventral, and the strobilus return to the radial type, as in 

 some species of Lycopodium and most Selaginellas ; but in some Selaginellas even 

 the strobilus may be dorsiventral. 



In the Ferns the case is less obvious ; the large size of the leaves, combined 

 often with a dorsiventral structure of the shoot, makes a comparison with a radial 

 strobilus less easy. Goebel has pointed out that in many dorsiventral Ferns 

 the dorsiventrality is already defined in the punctum vegetationis, and does not de- 

 pend upon a subsequent shifting of the parts. But it should be remembered how 

 many Ferns are orthotropic and radial ; that almost all the large genera include 

 species with simple unbranched leaves. Further, the series of the Ophioglossacefe, 

 possibly a distinct phylum from the true Ferns, may be held to illustrate a progres- 

 sive elaboration of the leaf, from smaller-leaved forms which are orthotropic and 

 radial, to larger-leaved forms, which are sometimes orthotropic and radial (Botru- 

 chimn), sometimes plagiotropic, and dorsiventral {Helminthostachys). It is not, I 

 think, improbable that these, and also the true Ferns, are referable in origin to an 

 orthotropic strobiloid type, with radial structure. This opinion was in substance 

 suggested in 1894 at Oxford ; these recent observations of Goebel on the deriva- 

 tive nature of dorsiventral shoots strengthen the position then taken up, while 

 they supply us with fresh examples of homoplastic development. 



Conclusion. 



This discussion was entered on with a view to finding whither phylogeny as a 

 basis of morphology would lead us. However unprepared we may be to pursue it 

 with certainty into detail, or to apply a terminology to the sequences which we re- 

 cognise, we must, I think, accept phylogeny as the natural basis for morphology, I 

 do not think that any middle course between this and an artificial system is possible 

 or reasonable. But here we launch ourselves upon a sea of uncertainties on which 

 we must keep our course with care. Following it, we think we espy certain great 

 movements in Nature. We may recognise what we believe to be a true evolutionary 

 sequence, but who is to say whether it is a progressive or a retrograde sequence ? 

 It may even be one divergent from some middle point. Our best friend may read the 

 sequence in opposite order to ourselves and arrive at a diametrically opposite con- 

 clusion. There is no finality tothis judging of probabilities, a fact which should 

 be alwavs before the mind, especially in the warmer moments of discussion. 



It is interesting to trace the parallel between the progress of classification of 

 1898. 3 X 



