INTRODUCTION TO VOLUME II 5 



assumed that progressions in respect of all the criteria will necessarily march 

 together. Not uncommonly a Fern may be advanced in respect of certain 

 features, and still retain one or more primitive characters : or Ferns essen- 

 tially primitive may show some single feature of advance. An example of 

 the latter is seen in Christensenia, in which the anatomical and soral features 

 are clearly Marattiaceous but the venation is reticulate. The like appears in 

 Anemia \Aneniidictyon, and in Lygodiuin '^Hydroglossum: h&re., while the 

 primitive characters of the Schizaeaceae are retained, the venation is reticu- 

 late, though in the rest of those genera it is open. It may be presumed that 

 in such examples there has been an advance in the type of venation while 

 other features retained their primitive state. Many species o{ Eu-Glcichenia 

 bear dermal scales, though the rhizome is protostelic, and the sorus primitive 

 in its structure and in its sporangia. A very notable example of discrepancy 

 is that of CJieiropleiiria, in which a protostelic axis, a singularly primitive 

 leaf-trace, and unbranched dermal hairs are associated with a reticulate vena- 

 tion and a mixed and Acrostichoid sorus. The fact that such examples are 

 exceptional makes them stand out prominently from the rest as witnesses 

 to the fact that the progressions in respect of the several characters march 

 as a rule parallel. 



The criteria thus recognised and evaluated were at the close of Volume I 

 abstracted, and the most primitive aspects of them combined into a verbal 

 specification of an ideal plant which should embody them all. The archetype 

 which was thus presented before the mind came out as a plant not unlike 

 the type of vegetation characteristic of the Devonian Rhynie Flora. This need 

 not in itself be held as evidence of actual ancestry. But at least the fact that 

 plants, comparable to this archetype extracted by analysis of Ferns at large, 

 did exist at a very early period, is in itself evidence that the method pursued 

 in Volume I has led to conclusions not in any way improbable. 



But a much more severe test will be the application of this method of ana- 

 lysis to the Ferns at large with a view to their seriation in time. If it appears 

 that the phyletic grouping thus arrived at for modern Ferns harmonises with 

 the appearance of the various families in successive geological ages, then it 

 may be confidently felt that we possess a morphological weapon that has 

 real value. This is the test that will be made in the succeeding Volumes. 

 The order of presentment will be roughly in accordance with probable 

 evolutionary sequence. But it is of course impossible in such arrangements 

 to do more than avoid the more gross misrepresentations : for actual affinities 

 are from the very first doubtful, and probably complex. As we progress to 

 the more modern types the sequences of probable affinity become clearer, 

 and the arrangement of the material will be such as to lead readily to their 

 recognition. But the most modern present a fresh difficulty, for owing to 

 parallel development in them the latest types become so standardised that 

 it becomes difficult to assign them severally to their phyletic source. 



