The MagnoliacecB. 225 



parts ; " a conclusion more applicable to the longifolia. A specimen im- 

 ported from France, and now in my ground, was ingrafted on the stock of 

 the purpurea. It forms a neat and healthy shrub of a dwarfish habit, and 

 yields many fine flowers. Another, on a stock of the acuminata, is rapidly 

 expanding into a tree, and furnishing blossoms in numbers proportioned to 

 its size. Neither specimen has ever matured a seed ; a defect arising, per- 

 haps, from its hybrid character. 



6. M. Tripetala (^Umbrella-tree). — This species is distinguished for the 

 size of its leaves and flowers. In this latitude, it is not inclined to send up 

 one main trunk like the acuminata, but usually rises from the roots in sev- 

 eral spreading branches. This habit can, however, be corrected by careful 

 training and pruning. Two trees thus managed stood in the grounds of 

 the late Dr. Hildreth, in Marietta, O., twenty years since. They were 

 straight-bodied, and symmetrical in form. When young, they were removed 

 from their native locality in Western Virginia. 



Another tendency is manifested by this species. The more central shoot 

 is prone to perish by the laterals robbing it of nutrition ; and they, in turn, 

 will attempt to supply its place as leading shoots. This tendency would 

 doubtless be corrected by propagation on the acuminata stock, and short- 

 ening from time to time all laterals assuming too rapid growth. 



The odor of its flowers is never pleasant, and is repulsive when much 

 concentrated. 



It requires a tenacious clay-soil, thoroughly underdrained, and enriched 

 with decayed vegetable matter. The cold of our severest winters it resists 

 with impunit)', but soon perishes under the impression of the sun in summer, 

 unless partially shaded, and unless the roots are preserved in a uniformly 

 moist condition. Careful attention to the latter precaution has, in some 

 instances, enabled it to flourish in open grounds. 



Seeds are occasionally matured, from which young plants can be raised. 



It was formerly found native near Grave Creek, in Western Virginia ; 

 which was probably its most northern locality. Michaux indicated the 

 " Western District " of New York as such ; but no other botanist has ever 

 found it in that region. 



In November, 1824, while travelling the Ridge Road, a few miles east of 

 Lewiston, on the Niagara frontier, several clumps of the M. acuminata and 



VOL. I. 39 



