442 G. Li. \Y 'u-la ad — Williamsonian Tribe. 



and distantly related plants, but with conifers, the last of the 

 seed ferns, persistent Cordaitaleans, and yet other little known 

 relatives of these gymnospermous types. As every "collector 

 knows, the utmost circumspection barely suffices in drawing 

 inferences from association of isolated organs of fossil plants ; 

 when behind a thickness of a few centimeters in a sedimentary 

 plant-bearing rock, or a mere joint or separation plane, may lie 

 the hidden change of a river bed and cutting away of a bank 

 in a different plant community, or perchance a year of time 

 with its change of winds and uprooting storms. Indeed with 

 such factors added to the infrequency of joined stems just 

 noted, absolute proof of leaf, stem and fruit unity is so dif- 

 ficult to establish that the claim for a great extent of the Wil- 

 liamsonian alliance might even excite the derision of one who 

 has not seen the field and studied the recurrence of species, in 

 widely separated localities. Not to forget how but recently 

 all true Paleozoic ferns threatened to vanish before an onslaught 

 of Cycadofilicaleans. 



Yet the recent field work, especially on the Yorkshire coast 

 (21, 25) — and in the newly discovered Mixteca Alta localities 

 (22, 27) — has minimized these difficulties more than might have 

 been anticipated, and will still further do so. The argument 

 for a heavily preponderant Williamsonian element in the 

 cycadophytean foliage of the Mesozoic, with a lesser Cycade- 

 oidean series and a nearly negligible cycadacean representation, 

 is chiefly one of analogy and association. The two main points 

 are : 



Firstly : The only known Mesozoic cycad stems with leaves 

 attached are cycadeoidean and Williamsonian ; and now that it 

 is seen that the nodes of lesser scars in forms like Wielandiella 

 as well as Yatesia, and Bucldandia (fig. 4) do not necessarily 

 indicate a succession of scales and carjjellary leaves, scarcely a 

 single Mesozoic cycad stem is left free from the suspicion of 

 belonging to the Williamsonise. Just as all the silicified trunks 

 finally proved to belong to the Cycadeoidese, so now either that 

 family or the Williamsonian tribe threatens to absorb the entire 

 series of stem imprints and casts. In fact Mesozoic stems indis- 

 putably belonging to the Cycadacese are most difficult to cite, 

 it being on the contrary the generalized and far more varied 

 Williamsonian types along with the earlier Medullosan forms 

 that appear to be ubiquitous. 



Secondly: The recurrence and association of strobili and 

 fronds is strikingly the same on the Yorkshire coast, in India 

 and in Mexico. And while in each of these great William- 

 sonian fruit-yielding regions, the fossil plant series is a most 

 varied one, carpellary leaves and cones of Cycadaceae are 

 exceedingly rare or absent. Nor are they ever abundant in 



