Rvni'.EKG : Notes on Rosaceae 5.37 



0. anstralis resembles a small 0. opitlifoliiis, but it has the short 

 caruncle of the western 0. capitatus. Its carpels are only about 

 half the size of those of cither of the two species mentioned. 

 Spiraea caroliniana, mentioned in Marshall's Arbustum, may 

 belong here, but that species was evidently never published. To 

 0. anstralis belong, beside the type given, the following specimens : 



North Carolina: Craggy Mountain, July 8 and Sept. 8, 1897, 

 Bill more Plerbariuui 12S2 b. 



Virginia: Summit, Stony Man Mountain, Aug. 13, 1901, 

 E. S. & Mrs. Steele 170; Peaks of Otter, June 6, 1890, A. 

 Broivii, T. Hoij^g, &-c. 



In the herbarium of the New York Botanical Garden there is 

 a specimen, received from the herbarium of P. V. LeRoy, and ac- 

 cording to the label collected in Mexico by Vischer, in 1838. 

 This has even smaller flowers than 0. australis. As Opiilaster is 

 otherwise unknown from Mexico, there was probably a mistake in 

 labeling or a misplacement of labels, and the specimen probably 

 came from some other place. This specimen was therefore ignored 

 when the manuscript for the North American Flora was prepared. 



0. cordatns is most closely related to 0. intermediits, but the 

 two species differ in the form of the leaves and the ranges of the 

 two are widely separated. Besides those of the type collection, 

 given in the Flora, only the following specimens may be doubt- 

 fully referred to it : 



California: Portola, July, igo^,, Elmer 4804. 



O. alabamensis is somewhat intermediate between O. stellatus 

 and 0. interuieditis, but differs from both in the shape of the leaves, 

 especially those of the sterile shoots. To 0. alabamensis belong 

 the following specimens : 



Alabama: Auburn, Sept., 1900, Lloyd & Earle ; April 17, 

 1897, Earle & Baker ; July 11, 1896, Earle & Underwood. 



South Carolina : Sandy river bottom, Clemson College, May 

 20, 1906, H. D. House 2173 ; Six Miles Creek, May 19, 1907, 



3383- 



O. Hapemanii might be a hairy form of 0. monogymis, just as 



0. opnlifolius, 0. intermedins , 0. capitatns, and 0. malvaceus have 



strongly pubescent and almost glabrous forms. 0. Hapemajtii is, 



however, so unlike O. monogynus in habit that it was thought ad- 



