58 BASIS FOR SYSTEMATIC TREATMENT OF FERNS [CH. 



detailed observations were extended by Bertrand, Poirault, Gwynne-Vaughan, 

 Jeffrey, Boodle, and others. The study of apical segmentation initiated by 

 Naegeli (1845) has led to the general view that segmentation of stem, leaf, 

 and root in Ferns runs parallel with that of their sporangia, and that the details 

 seen in them connote a general difference of more or of less robust cell-con- 

 stitution (34). The tendency has arisen to seriate the various types of Ferns 

 according to such varied characters as these, recognising most clearly those 

 types which appear to be extreme. The series would then extend roughly 

 between the delicate Leptosporangiatae and the more robust Eusporangiatae. 

 Thus the sequence already apparent in the arrangement of Mettenius seemed 

 to be confirmed. 



But here came in the insistent phyletic question which Prantl felt to be 

 so difficult: viz. which extreme type is the more primitive, and which the 

 derivative? It was first taken up on purely comparative grounds by Campbell 

 (1890). He recognised the massive Eusporangiatae as the more primitive, and 

 the relatively delicate Leptosporangiatae as the derivative types of Ferns (33). 

 If this conclusion be true it should stand the test of palaeontological enquiry. 

 It was shown in the following year that at least the great majority of the 

 Palaeozoic Ferns were Eusporangiate, while the Leptosporangiate types 

 become frequent in later periods, and that they are the characteristic Ferns 

 of the present day(34). Such considerations, supported as they now are by 

 greatly extended observations and comparisons, confirm that general seria- 

 tion of the main Families of the Filicales already laid down by Mettenius. 

 Subject to test by all the characters that can be used comparatively, that 

 sequence may be upheld as probably illustrating in a quite general sense a 

 progressive evolution. Inverting the order of the Families as given by 

 Mettenius it would run as follows: — 



I. Ophioglossaceae 



II. Marattiaceae 



III. Osmundaceae 



IV. Schizaeaceae 



V. Gleicheniaceae 



VI. Hymenophyllaceae 



VII. Cyatheaceae 



VIII. Polypodiaceae 



Relatively primitive. 



Relatively advanced. 



This is then the general position from which we may start upon a more 

 searching comparison of the Filicales, with a view to the arrangement of the 

 Class as nearly as may be according to their probable phylesis. The general 

 principle will be to widen to the utmost the range of the characters and com- 

 parisons that are to be used in attaining that end, and thus to base phylesis 

 tip on the sum of the available data. 



