450 MINNESOTA BOTANICAL STUDIES 



a number of characters. The catkins are somewhat smaller than 

 in B. hitea. The lateral lobes of the bracts are slightly spreading, 

 whereas in B. lutea they are erect. The leaves are considerably 

 smaller than in the yellow birch. They have the texture of B. 

 pumila var. glandulifera and average six pairs of veins to four and 

 nine for B. pumila and B. lutea respectively. The samara is almost 

 identical with that of B. lutea which is characterized by having the 

 wmg narrower than the body of the nutlet and not projecting above 

 the base of the styles (Fig. 2, n,p,r,t.). In both these respects it 

 is strikingly different from the fruit of both B. papyrifera and B. 

 Sandbergi. The flavor of wintergreen in the bark of the hybrid 

 is an unmistakable evidence of its connection with B. lutea and is, 

 in the opinion of the writer, the character of most significance. 

 From these facts there can be no doubt that B. lutea is one of the 

 parents of this hybrid, but the parentage on the other side is less 

 easy to establish. 



As already stated, only three species of birch in addition to the 

 hybrid B. Sandbergi occur in the part of the state where this new 

 hybrid has been observed. They are found pretty well intermixed. 

 The crossing must therefore be between B. lutea and one of the three 

 other forms. From the evidence so far deducible which of these 

 appears the most likely? 



A cross between B. lutea and B. papyrifera would probably give 

 rise to a tree instead of a shrub, for in this region both of these spe- 

 cies are good-sized trees. The leaf would be expected to resemble 

 either one or the other of the parents or be intermediate between 

 them. In no respects does it comply with these expectations, for as 

 figure 1 (m,n,p,q,s) shows, it is quite distinct both in form and size. 

 and the texture is also different. The catkin, bract, and samara as 

 already shown resemble B. lutea very closely and indicate no ad- 

 mixture of B. papyrifera. That B. Sandbergi might be one of the 

 parents is not impossible, yet it appears rather doubtful, for the 

 reason that in one of the places where the shrub was found, only 

 B. lutea, B. pumila var. glandulifera and B. papyrifera occurred. 

 Furthermore, the spontaneous crossing between a species and a 

 somewhat rare hybrid seems less likely to take place than between 

 good and abundant species. On the whole, therefore, it appears 

 that B. pumila var. glandulifera must be the other parent. This 

 species is abundant in the swamps where the hybrid has been found 

 and as there is nothing in the form and appearance of its catkin. 



