PINACEAE. PINUS ARMANDI 21 



but the cones themselves appear different. The cones from Japan are from 4.5 

 to 11 cm. long and have a rounded apophysis with the umbo undeveloped and 

 with the outer edge of the cone-scale recurved. Except for a slight inferiority in 

 size the largest agree exactly with Hayata's excellent figure of his P. Armandi, 

 var. Mastersiana. This herbarium is rich in Chinese material of P. Armandi 

 and a careful examination shows that these cones are from 4.5 to 20 cm. long and 

 that in many the outer edge of the cone-scale is recurved; in a number I find the de- 

 velopment of the umbo arrested and in several the apophysis is broad and rounded. 

 I have not found in any of the variants perfect seeds, and on re-examining the 

 Japanese material I find the seeds empty. To make certain that the arrested de- 

 velopment of the umbo and apophysis was correlated with infertile seeds I ex- 

 amined cones from the Arboretum living plants of P. Armandi, and found some 

 indistinguishable from those from Japan in size and other characters mentioned 

 and others approximately so, but none absolutely identical with perfect cones 

 from China. In the perfect cone with fertile seeds the apophysis with fully de- 

 veloped umbo is obtuse and triangular in shape and the outer edge of the cone- 

 scale is sometimes slightly reflexed. Between this and cones with infertile seeds, 

 arrested umbo and short rounded apophysis as figured by Hayata I find every 

 possible gradation. This variation in size and shape of the apophysis and in degree 

 of development of the umbo in P. Armandi is interesting and probably occurs in 

 many species of Soft Pines, but it is by no means so marked as in the cones of the 

 Rocky Mountain P. fiexilis James, its most closely related species. 



As I did not see P. Armandi wild in Japan I can give no account of the 

 trees, though they are evidently rare. Mr. Miyoshi, in his letter accompanying 

 the specimens, says those from Yaku-shima "were collected from a wild tree in 

 Hirase Government forest, Kamiyaka-mura, on October 19, 1914." Those from 

 Tanega-shima "from a wild tree about 700 years old, 90 ft. tall, 20 ft. in perim- 

 eter of trunk, growing in Yakugawa Government forest, Kitatane-mura, at 

 about 600 ft. above sea-level." The material was sent to me in a box made of 

 the wood of this species. This wood is heavy, close-grained and darker brown 

 than I remember seeing it in China, but this is a matter of no moment, for it is 

 well known that soil, situation and climate strongly influence the character and 

 quality of timber. 



The discovery of this Pine wild in southern Japan adds remarkably to its dis- 

 tribution. It is indigenous on the mountains of Formosa and on those of the 

 Chinese provinces of Kweichou, Hupeh, Shensi, Kansu, Szech'uan and Yunnan 

 to the extreme western limits of China proper. Very probably it will yet be found 

 on the higher mountains in the more eastern provinces of the Flowery Kingdom. 



Plants of P. Armandi received from Monsieur Maurice L. de Vilmorin in 1902 

 and others raised from seeds sent from China on several occasions by myself are 

 growing in this Arboretum. They are perfectly hardy, grow rapidly and are very 

 ornamental; they produce cones at an early age, but unfortunately they suffer 

 from attacks of boring insects and do not now promise to develop into good trees 

 in this part of eastern North America. 



Pinus Armandi was discovered in January 1873 on the Tsinling Mountains 

 in Shensi province, China, by Pere A. David and was introduced to the arboretum 

 of Monsieur M. L. de Vilmorin at Les Barres, France, by Pere P. Farges, who sent 

 seeds from northeastern Szech'uan in 1895. Seeds of this Pine were sent to Kew 

 by Augustine Henry in 1897, who collected them "ten miles north of Mengtsze," 

 in Yunnan province. A portion of these seeds were received here in June 1897. 



