PRESIDENTIAL ADDRESS. 727 



of plant-remains which we have thus come to recognise as belonging to the earlier 

 Spermophytes. 



For the position of these plants had remained in suspense. The elaborate 

 anatomical investigation which their vegetative organs had received at the hands 

 of Williamson, Scott, Solms-Laubach, and others showed them to occupy a 

 transitional position between the Ferns and Cycads. In certain respects they 

 showed an advance in the cycadian direction, whilst in others they were wholly 

 fern-like. Their fructifications were unknown, and their nature remained an open 

 question. It was for this group, or series of transitional groups, that Potoni^ 

 proposed the appropriate name of Cyeadofilices. 



We know now that-the Lyginodendreas and Medullosese bore seeds attached 

 to their fronds. The seeds have been found attached in some cases to reduced 

 fronds consisting of a branching rachis, in others to fronds of the normal 

 filicineau type. Indeed, so far as habit is concerned, these plants may rightly be 

 described as seed-bearing Ferns. 



As such, indeed, most people will be content to regard them — as forms, that 

 is, having close filicinean relationship in which the reproductive method has been 

 profoundly modified, the internal anatomy to a less extent, and the habit hardly 

 at all. Had these Pteridosperms come to light during the lifetime of Hofmeister, 

 that master of morphology must have pounced upon them as furnishing an im- 

 portant link in his chain. These fossils and the spermatozoa which the Japanese 

 botanists discovered in the seeds of Ci/cas and Ginkgo, indeed, afford the most con- 

 vincing direct evidence of the soundness of the Hofmeisterian scheme that it is 

 possible to conceive. Nor is that all. For by confirming the indications first 

 revealed by the earlier investigation of the vegetative anatomy, the Pteridosperms 

 have allbrded us a striking object-lesson of the value of the anatomical method — 

 of the significance of purely anatomical characters too long ignored by the 

 systematist. 



Not so long ago, when new examples of these Pteridosperms were turning up 

 on every hand, some pessimists were inclined to wonder whether, after all, any 

 groups of real Ferus existed in the Palaeozoic rocks. Such sporangia as were 

 known might well be the pollen-sacs of seed-bearing plants. All doubts on this 

 score are happily set at rest by the detection of germinating Fern-spores in con 

 temporary beds. Nor can I think of any more fitting tailpiece to the investiga- 

 tions which lead the way to the Pteridosperms than the discovery, by the same 

 investigator, of the antidote to these rather disturbing views. However, it is 

 needless to dwell further on these matters now, in view of Dr. Scott's address 

 to-morrow upon the Present State of Palaeozoic Botany. 



But to return to the history of the seed. In the absence of direct evidence, 

 one can only conjecture that some old generalised type of sporangium formed its 

 prototype, something substantial, on the lines of a Botryopferis or Zt/gopteris, 

 perhaps. The heterospory that was the precursor of the seed-like condition must 

 have been a transient phase, or else it is lost in the pre-Carboniferous obscurity. 

 Be that as it may, the passage from the dehiscent to the indehiscent monosporal 

 megasporangium finds its analogy in every group of plants. Where there is extreme 

 numerical reduction of the contained structures — be they spores or seeds — a multi- 

 tude of cases in the Fungi, in the Algae, and the angiospermic flowering plants 

 show that dehiscence tends to become obsolete. The failure to dehisce does not 

 appear to be directly correlated with any mechanical difficulty in ejaculation. It 

 is more probably one of those obscure cases of interdependence of phenomena in 

 which the vegetable kingdom abounds. A special investigation directed to the 

 elucidation of this point might be expected to yield interesting results. 



We now come to the consideration of a most characteristic organ of the seed — 

 the pollen-chamber. This cavity arises at the apex of the megasporangium, above 

 the big megaspore, and is found in all the Palaeozoic seeds, with the sole excep- 

 tion, so far as I am aware, of the ' seed-like ' structures in Lepidocarpon and 

 Miadesmia. The utility of the pollen-chamber is manifest, but its antecedents are 

 quite unknown. Upon such a structure as this may have depended the success of 

 the seed-method at a critical stage in its evolution. In the viviparous Selaginellaa, 



