•35G QEOT.OGtCAL SURVEY OF CANADA. 



wiites that all the specimens sent are Queretis Jacohi, R Brown, and 

 quite distinct fi-om the Oregon Q. Garryana. It is therefore probable 

 that the Vancouver Island Oak will become Q. Jacohi, instead of Q. 

 Garryana. 



(1993.) Q. stellata, Wang.; Gray, Man., ed. YI., 475. 

 Q. oblugiloha, Michx. ; Macoun, Cat, III., 440. 

 We are still doubtful of the occurrence of this tree within oiir limits. 



(1999.) Q. coccinea, Wang., var. ambigua, Gray, Man., ed. 

 VI., 478. 



Q. amlngua, IMichx. ; Pursh Fl. II., 630- 

 Q. rulira, Macoun, Cat., III., 442, in part. 



In the new edition of Gray's Manual this variety is said to extend 

 along the north-eastern borders of Lake Cbamplain and northward. 

 This being the case, the Nova Scotia, New Ih-unswiek, and most of the 

 Quebec Q. rubra becomes this variety. Collectors in these provinces 

 should collect a good series of fruiting .specimens to determine this. 



555. CASTANEA. 



(2002.) C. sativa, Mill., var. Americana, Graj'. Man.,ed. VI., 479. 



References under C. vulgaris, var. Americana, Macoun, Cat., III., 443, 

 belong here. 



557. 8ALIX. 



(2007.) S. Brownii, Bebb. ; Bot. Gaz., XIV., 444. 



S. arclicu, R. Brown, Bot. Koss- Voy., ed. II., 194; Hook., FL II., 1.52; 

 Macoun, tat III., 444, in part. 



" Omit synonym S. cordifolia, Hook. The localities given in Part III. 

 are all correct, excepting ' Kotzebue Sound and Ounalashka {Rothr., 

 Alaska.},' which almost certainly belong to S. arctica, Pallas (not R. 

 Br.)." {Bebb.) South Twin Island, James Bay, 1887 ; shore of Hudson 

 Bay, Lat. 55°-56=, 18SG. (./. M. Macoun.) Mountains between Peel 

 River and La Pierre's House, Ai-ctic Circle, 1888. (McConnell.) 



(2008.) S. argyocarpa, Anders.; Macoun, Cat. UL, 445. 



''Richardson's locality, • Fort Franklin, on the Mackenzie.' were 

 better omitted. There is no trace of this species among his collections 

 that I have seen, and no evidence, at least in American herberia, that 



