130 



They float for a long period. Of five carpels put to soak on 

 March 24, 1902, one sank on June 6, and the others were still 

 floating on February 27, 1903, when the experiment was discon- 

 tinued, a total period of 340 days. In an ocean current traveling 

 but one half mile an hour they would be carried over 4,000 miles 

 in 340 days. At the rate of flow of the Gulf Stream off New- 

 foundland (about one and two fifths miles per hour) they would 

 be carried 10,744 miles, so that even if there was no land bridge 

 to Europe by way of Greenland and Iceland the spread of the 

 ancient species would not seem remarkably difficult. 



A reverted Leaf-form. — Fig. i represents a leaf of Lirioden- 

 dron Tidipifera L. from a grown tree at Passaic, N. J., showing 

 the basal lobing which represents the first stage in the formation 

 of stipules. It is very similar to the leaf figured in the Bull. 

 Torrey Club, 28 : pi. 41. f. 1. 190T, in both form and venation, 

 representing however a stage not quite so far advanced as the latter. 



Anomalous Leaf of Magnolia {fig. j). — The close relation 

 of the genera Magnolia and Liriodendron is further emphasized 

 by the retuse leaf of Magnolia Virginiana L. here figured. The 

 similarity of the modern bud-leaves and stipules has been pointed 

 . out by Meehan and the writer. * The contemporaneous appear- 

 ance of Magnolia and Liriodendron in the Raritan formation of 

 New Jersey, their parallel development in geological time, and 

 their similar geographical distribution at the present day, as well 

 as their dispersal in ancient times all point to a common ancestor. 

 This leaf, occurring next to the blossom (localized stage of Pro- 

 fessor Jackson) also serves to corroborate the theory of the devel- 

 opment of the Liriodendron leaf from a lanceolate-leaved ancestor. 



With the exception of being smaller, our leaf is the counterpart 

 of the Liriodendron leaf figured in Bull. Torrey Club, 28 : pi. 41. 

 f. 6. The Liriodendron leaves figured in Torreya, 2 : pi. i.f. /, 

 j. 1902, arc about the same size and are exactly similar in out- 

 line except for being slightly more retuse. Among fossil leaves 

 it greatly resembles various arctic leaves referred by Heer to 

 Colutea and to Liriodendron Meekii, and except for being only 

 half as large it is a counterpart of the leaf of Liriodendron 



* Compare Jig. 2 of Michelia fuscata ( Andr. ) Hance with Bull. Torrey Club, 28 : 

 pi. 42. f. to. 11. 1901 ; and Torreya, 1 : //. /, 2. 1901. 



