1902] NOTES ON THE PHYLOGENY OF LIRIODENDRON 47 



a series of leaves of Liriodendron Tulipifcra, who can say where Z. 

 simplex ends and Z. primacvum or Z. Meekii begins, or where L.pn- 

 macvum ends and Z. semialatum begins? The same variability and 

 interrelation is shown among the leaves variously referred to 



M 



Plata- 



nus, Sassafras, etc.; and, while it may be argued that we should 

 expect just such variable and intermediate leaves at a time when 

 the flora of the globe was becoming rapidly differentiated, the 

 argument loses force when we turn to a similar variability and 

 interrelation of the modern forms of the same genera, when their 

 leaves alone are considered. 



In considering the thirty or more species and varieties of 

 fossil forms ascribed to Liriodendron and its allies Lirioden- 

 dropsis and Liriophyllum, we may distinguish them as all good 

 species, or as three or four, or we may identif}' most of them 

 with the modern form, dependent upon our conception of the 

 term species. If the Cretaceous period, at the time the Dakota 

 strata were being laid down, was the time of maximum develop- 

 ment and diversification of the dicotyledons, it is difficult to 

 understand how so many of the tvpes are still persistent with 

 all their essential characters unchanged (Myrica, Betuja, Quer- 



cus, Populus, Viburnum, Salix, etc., and representatives of nearly 

 all of the natural orders) , even in some cases to specifically sim- 

 ilar forms (as Mag7iolia acuminata and M. tripetala). If there was 

 such a diversification of species at that time as we are led to 

 believe, it is strange that most of the types persist comparatively 

 unaltered through the later formations to the present time, 

 mstead of losing their character in further continued variation. 

 We have listed as follows a few of the reasons for consider- 

 mg some or all of the described species of Liriodendron invalid : 



1. Their remains are more or less fragmentary, and therefore 

 not determinable with certainty. 



2. The shape and venation of leaves are poor specific charac- 

 ters at best. 



3- A series of leaves of LiriodaidroTi Tulipifera can be found 

 which parallels the fossil species in both shape and size. 



