G. R. Wieland — American Fossil Cycads. 433 



between the sporangia of even the living Marattiacean genera, 

 as is seen on passing from the free sporangia of Angiopteris 

 to the fused sporangia of Marattia, that in considering rela- 

 tionships, forms with free sporangia or with sporangia arranged 

 circularly, must not be neglected as in the wider sense unre- 

 lated — a fact which it is well to bear in mind since it must now 

 necessarily be extended to the cycadaceous forms as well. 



One other observation in this connection I shall also make. 

 As fully stated by Bower in his indispensable work just cited, 

 while the evidence of development is in favor of the Marat- 

 tian or Synangial type, developing from a single sporogenous 

 mass, as being older than the Angiopteris type with free spo- 

 rangia — an idea opposed to that of Stur (cited above), there 

 has been hitherto no direct paleontological evidence either 

 way. But now should the Bennettitacew and the living cycads 

 not be polyphyletic, the latter have suffered a breaking up of 

 the synangium which may well have been paralleled by a 

 similar separation in the Marattiacese. In this category of 

 pollen sac separation and reduction the Gingkoacese may appar- 

 ently be likewise included. The occasional occurrence of 

 pollen sacs in groups of three in the living Gingho, as sug- 

 gested by Thibaut (Recherches sur l'appareil male des Gym- 

 nospermes), may represent a primitive condition. And Haiera 

 Munsteriana from the Rhat with asteriate groups of seven 

 pollen sacs may certainly be regarded as still more primitive. 

 There would therefore appear to have been, following an 

 original soral septation, a more or less progressive separation 

 of sporangia and reduction in their number in the Marat- 

 tiacese, cycads, and Gingkoacese. Certainly if the facts observed 

 here cannot be construed as offering inferential testimony from 

 the paleontological side in favor of Bower's supposition, they 

 do not interpose any facts against it. In accordance with this 

 view I incline to believe that the Bennettitean synangium is 

 slightly more archaic in form than that of any of the living 

 Marattian genera except perhaps Dance, with which, as 

 pointed out above, the comparison is not quite direct. 



~Now further as to forms presenting analogous synangia or 

 sori. From the foregoing description it is clear that Uma- 

 topteris, Crossotheca, and especially Calymmatotheca may be 

 considered as distinctly related by reason of their synangial 

 forms. But these are all dimorphic ferns as hitherto viewed, 

 with reduced fertile pinnules suggesting those above described, 

 and at once carrying the relationship of these Mesozoic cycads, 

 themselves dimorphic as well, back to the Cycadofilices. 

 Hence from the variations in form seen in the Marittiacean 

 sorus, and no doubt also paralleled in the Cycads, at may be at 

 once stated that so far as we may infer from isolated fructifica- 



