19 1 q] SCHNEIDER— AMERICAN WILLOWS 43 



Folia distinctius petiolata; amenta vulgo ultra 15 mm. longa, 

 crassiora vel antherae magis ellipsoidales oblongiores et folia 

 superne pl.m. stomatifera (specimina mascula specierum se- 

 quentium sub anthesi sine foliis perfecte evolutis accurate 

 discernere saepe impossibile videtur; confer etiam 28. 5. lin- 

 gulatam speciem incertam alaskanam). 



Folia superne pl.m. (interdum tantum sparse secundum 

 nervos) stomatifera, vulgo oblonga, 2|-4plo longiora quam 

 lata, sed vix ultra 4:1.8 cm. magna. 



Amenta pedunculis vix ad i cm. longis suffulta, 8-15:7 

 mm. magna, densifiora; folia pedunculorum (saltem sub- 

 tus) dense breviter sericea vel villosula (confer etiam 27. 

 S. anamesam e Groenlandia) . ... 23. 5. pseiidolap ponum 

 Amenta saepe longius pedunculata, i . 5-3 .5:0. 8-0 . 9 cm. 

 magna et basim versus laxiflora vel folia pedunculorum 



etiam subtus glabra vel subglabra 24. 5. desertorum 



Folia superne baud stomatifera, majora, latiora vel longiora. 

 Amenta deflorata ad fere 3 . 5 cm. longa, basi vulgo pi. m. 

 laxiflora vel folia pedunculorum satis oblonga, circ. 2^-3^ 



plo longiora quam lata 25. S. glaucae varietates 



Amenta deflorata vulgo baud ultra 2 (-2. 5) cm. longa, 

 etiam basi densiflora vel folia pedunculorum latiora 

 brevioraque vix ad 2|plo longiora quam lata 



26. S. cordifolia 



Enumeratio sectionum specierumque 



I have omitted from the keys and the following enumeration the 

 well known and easily recognizable S. Candida Fliigge which Ball 

 (1909) includes in his section Arcticae, because I attribute to it a 

 different systematic position. 



Sect. I. Reticulatae^ Fries in Sylloge PI. Nov. Soc. Ratisb. 

 2:38 (Consp. Disp. Sahc. Suec). 1828, quoad S. reticulata; for 



* There is the older name Chamaetia given by Dumortier in Bijdr. Natuurk. 

 Wetensch. 1:56 (Verh. Gesl. Wilgen 15) 1835 to a group, including S. rctusa, S. her- 

 bacea, and S. reticulata. Unfortunately neither the International Rules nor the 

 Philadelphia Code contains a precise rule in regard to the application of names of 

 sections or similar groups. I do not accept Dumortier's name because in his paper 

 he proposes two very different arrangements, and he does not in my opinion make a 

 definite statement. 



