228 BULLETIN OF THE 



The difficulty is increased by the fact that by gradual modification the 

 leaves pass to evidently difFei-ent generic types. The genus Liriophyl- 

 lum, for example, represents leaves of Liriodendron nearly round in out- 

 line, but split at the top and to below the middle in two lobes joined in 

 a narrow sinus upon the medial nerve, a peculiar division and facies 

 which have no analogy in any species of plants of our epoch. This new 

 genus is by its leaves intermediate between Liriodendron and Populus. 



The local distribution of the leaves may be relied upon to give some 

 directions for the separation of species. For of course if the analogous 

 forms are found in separate and distant localities, the marked differences 

 are raoi'e likely to be specific. On this subject the specimens collected 

 by Mr. Sternberg afford good opportunity for examining the question. 

 Omitting details, it suffices to mention what is known of the distribution 

 of the leaves of Liriodendron. 



L. Meekii is described from specimens found in Nebraska and Minne- 

 sota only. The specimen of L. primavum is labelled Blackbird Hills, 

 Nebraska. This form is allied only to L. semi-alatum of Kansas. L. in- 

 termedium is also from Nebraska, represented in one specimen only, and 

 has no affinity with any of the Kansas species. L. gigantetim was fii'st 

 described from a fragment found near Fort Harker, Kansas, while the 

 specimen of the Museum is from two and a half miles from Glascoe, in 

 another county, where also were found L. acuminatum and L. pinna- 

 tifidum, whose leaves have little affinity of characters between them. 

 L. semi-alatum was found at a different locality seven miles distant from 

 Glascoe, and L. cruciforme at Elkhorn Creek, twelve miles northwest of 

 Ellsworth. 



From this kind of distribution it seems legitimate to conclude, not 

 only for Liriodendron, but also for all the other groups of this flora 

 whose leaves present the same degree of affinity or of difference, that, if 

 the forms are derived from synthetic types, and if they are found some- 

 what modified in characters at distant localities, it is a proof that the 

 modifications are already fixed, have become local characters, and that 

 they may be considered as specific. 



On another question, that of the derivation of ^ the vegetable remains 

 found in the strata of the Dakota group, and of their distribution, either 

 in place, from trees grown there, or as transported by water from dis- 

 tant localities, a question examined already in the Cretaceous Flora, the 

 collection of Mr. Sternberg affords the same degree of evidence as for 

 the preceding. 



It has been observed in the Flora, Vol. VI., loc. cit, that the Creta- 



