106 



larg^J detail, to which attention is especially directed, together 

 with a cross section of the petiole showing the fibrous margin 

 formed by the descending lowest primary vein. The lowermost 

 vein divides as it approaches the midrib, the upper branch join- 

 ing the latter, while the lower branch is directed downward and 

 passes along the side of the petiole, remaining distinct as a tiny 

 fibrous margin of the latter all the way to the point of insertion 

 of the stipular wings. 



Figs, i, 2, and 4 show specimens in which the leaf-blade 

 has only developed sufficiently to form very small, ovate leaves 

 which, both in shape and venation, are very similar to Lirioden- 

 dron cotyledons, or to what I consider the ancestral type of Lirio- 

 dendron leaf to have been. Their summits are crowned with a 

 longer or shorter length of the persistent awn-like tip of the mid- 

 rib (in Fig. 1 the latter is 5 mm. in length). 



Fig s - S> 9> an d 10 show what I considered after careful com- 

 parison and measurement to be anomalous flower bud-scales, 

 before I found them in position on the tree. Afterward I found 

 numerous specimens in position (Figs. 13 to 16). 



The forms figured at 10, which are quite common, have the 

 midrib developed for a considerable distance as a thread-like, 

 fibrous bundle with no trace of green tissue. In those forms 

 figured at 9, of which I have numerous specimens, the midrib is 

 much more extensively developed, being the normal length of a 

 true midrib, and bearing at its summit a thickened cylinder of 

 green tissue, evidently an abortive leaf-blade. 



In Fig. 5, this mass is expanded into a true leaf-blade, ovate- 

 lanceolate in shape, and of tiny dimensions, bearing at its summit 

 the extended midrib as an awn of 21.5 mm. in length. 



Fig. 10 minus the extended midrib shows the ordinary form 

 of the flower bud-scales which may be found in great numbers 

 rolled up on the ground beneath the trees as soon as the buds 

 have swollen sufficiently to cast them off. 



Figs. 5, 9, 10, 13, 14, 15, and 16 are especially interesting 

 inasmuch as they are practically identical with the foliate bud- 

 scales referred to in a previous paper (/. c.) as occurring in the 

 related genus Magnolia. 



