342 BOTANICAL GAZETTE [april 



Northwest Territories. The northern border line of the range of 

 S. longifolia and var. pedicellata is not yet exactly known. Approxi- 

 mately it seems to run in the west from Fairbanks in Alaska to Fort 

 Simpson in the Northwest Territories and through the Athabasca 

 Plains and central (or southern ?) Manitoba and southern Ontario 

 to the south of James Bay and to about Lake St. Johns in Quebec, 

 from where the eastern line turns southeast to western New 

 Brunswick (Woodstock, Pokiok) and then southward to New 

 Hampshire along the Connecticut River to Delaware and the 

 District of Columbia. 



The species apparently reaches its best development in the rich 

 river bottoms from Louisiana to Indiana, while in Oklahoma, 

 Kansas, Nebraska, and Iowa the form of the sand bars seems to 

 prevail, which has narrower, smaller leaves. In the region of the 

 Great Lakes and in the northeast, but also in other portions of the 

 range under similar ecological conditions, the following variety 

 seems to occur frequently: 



S. LONGIFOLIA var. Wheeleri, nov. comb. — S. interior var. 

 H^/?ee/m Rowlee in Bull. Torr. Bot. Club 27:253, />/. p,^g. 74. 1900. 

 — S. Wheeleri Rydberg in Britt., Man. ed. 2. 1061. 1905; Britt. and 

 Br., 111. Fl. ed. 2. 1:595. iQ^S- — S. longifolia (vel S. fluviatilis) 

 var. argyrophylla Auct. div. pro parte, non And. — I agree to a 

 certain extent with Schaffner (in Ohio Nat. 14:255. 1914), who 

 regards this variety as an ecological form, and I have already 

 pointed out that similar forms seem to occur in S. exigua (see var. 

 luteo-sericea) , S. melanopsis var. Bolanderiana, etc. Those forms 

 very often look quite distinct, especially in the herbarium. The 

 broad leaved forms of var. Wheeleri can easily be taken for a well 

 marked species if one does not have a very rich set of specimens 

 showing all the intermediates between such forms as we know from 

 Maine (Caribou) and New Brunswick and the narrow leaved forms 

 from Lake Champlain, Lake Superior, etc. It may be that the 

 easternmost forms are not quite identical with the typical var. 

 Wheeleri from the region of the Great Lakes, but to decide this 

 question we need a careful study of this form as it is observed in New 

 Brunswick, Maine, Connecticut, western Quebec, and eastern 

 Ontario. There is a male plant in cultivation in the Arnold 



