83 



ing that the existing species originated in Mid-Cretaceous times, 

 since the latter can certainly for the most part be matched by 

 variants among the leaves of the existing species, which is a pre- 

 Darwinian conception of the fixity and immortality of species which 

 is as outmoded at the present time as is Noah's flood. The accom- 

 panying illustrations show two carpels of Liriodendron hesperia 

 from the Brickyard locality at Spokane (Figs. 1, 2), a fragmentary 

 leaf from the Vera locality that is referred to the same species (Fig. 

 3) and a similar leaf of Liriodendron tuHpijera from a spring shoot 

 of a sapling (Fig. 4). The resemblances between the last, both in 

 form and venation are obvious. 



There are two existing species of Liriodendron — one in central 

 China and our familiar tree of the southeastern United States — 

 those two regions that are botanically so similar and share so 

 many arborescent genera that are found only in these two regions 

 {e.g., Gordonia, Hicoria, Magnolia, Sassafras, Nyssa, etc.). 



In this connection attention should be called to two Lirioden- 

 dron leaves recorded (6) by Endo in 1934 from the Neogene of Japan 

 under the name of Liriodendron honsynensis, and also to carpels 

 which Endo has described (7) from the Miocene of Korea (Tyosen) 

 as Liriodendron meisenensis and which he states dififer from those 

 of L. hesperia. These discoveries in Japan and Washington state 

 help to bridge the gap of about 160° of longitude between our 

 existing Liriodendron tiilipifera Linne and L. chinensis Sargent. 



