XXIII] COMPARISON 191 



(9) The spore-output is smaller (at most 64 against 128 or 256 of the 

 Schizaeaceae), while heterospory introduces entirely different and derivative 

 relations. 



(10, 11) The very reduced prothallus presents negative features, though 

 it may be remarked that a very simple prothallus, such as that of Schizaea, 

 would lend itself readily to reduction. The archegonia are standardised as 

 usual, but the antheridia present in the low number of the spermatocytes 

 a point of similarity with Schizaea and Anemia. 



(12) The embryology corresponds so exactly with that of Leptosporan- 

 giate Ferns that it confirms these otherwise peculiar plants in their Filical 

 relationship, while their general features point to a relatively primitive 

 position. 



The result of this comparative analysis is to render support to Campbell's 

 recognition of a real relationship of the Marsileaceae to the Schizaeaceae, 

 the similarity extending along the whole line of the comparison. Perhaps 

 the most divergent point is between the abaxially-flexed pinnae of Schizaea 

 and the (hypothetically) adaxially-flexed pinnules of the sporocarp of Mar- 

 silea. But on this discrepancy it is fair to remark (i) that the sporangia being 

 assumed to be marginal, as in fact they are in the Schizaeaceae, they would 

 be as easily displaced to the upper (adaxial) as to the lower (abaxial) 

 surface: (ii) that the hypothesis demands a folding of the pinnules of the 

 sporocarp in the same way as the pinnae of the sterile leaf of Marsilea, that 

 is adaxially: and (iii) that the indusial developments of the Schizaeaceae 

 are themselves so various that it is no great demand on probability that the 

 tissue filling in the sporocarp and protecting the sporangia in the Marsilea- 

 ceae should not comform to any one of the Schizaeaceous types. The 

 hypothesis is not definitely proved, but it can at least be claimed that there 

 is high probability of its truth. In any case the relationship to the 

 Schizaeaceae seems beyond doubt. The Marsileaceae may be held to be 

 a family of that same fundamental type, though specialised for aquatic life, 

 and heterosporous. But it does not seem possible to link them definitely 

 with any single genus of the Schizaeaceae: the similarities are, however, 

 closest to Schizaea and AneiniorrJiiza. In support of the comparison with 

 Schizaea it may be noted that ^, rupcstris is to be found growing in very 

 moist conditions, under dripping rocks, in the Blue Mountains, New South 

 Wales, as a veritable Hydropterid in habit. 



The early existence of the Marsileaceae in geological time has not been 

 satisfactorily demonstrated. Suggestive but not very convincing fossils have 

 been recorded from the Rhaetic, Jurassic, and Lower Cretaceous floras. But 

 Sagenopteris, which was formerly compared with the Marsileaceae, is now 

 believed to be the foliage of a primitive Angiosperm (H. H. Thomas, 

 "Caytoniales," Phil. Trans. Yo\. ccxiii, B, p. 199). 



