1912] Wiegand,— Amelanchier in eastern North America — 119 
found where characters plainly appear to have been recombined. So 
frequent have these cases become that it seemed desirable to work 
out a treatment of the genus based on the idea that the so-called 
“intermediates” are hybrids. The present paper is a result of this 
endeavor. The first and most difficult problem was to determine 
what were the real species, and what the hybrids. At first too few 
species were recognized, thus making necessary the interpretation 
of some plants as hybrids in regions where at least one of the parents 
was not known to exist. 
After several trials, however, the present treatment was settled 
upon for the following reasons: because the groups of sheets repre- 
senting each species obtained in this way seemed to be natural groups 
the individuals in which differed not more from each other than do 
individuals in species of other genera; because the range of each 
species so obtained was a natural range, and not local; because most 
intermediate specimens could be interpreted as recombining characters 
of the true species; because all characters were accounted for, so 
that supposed hybrids simply recombined old characters, and no new 
characters appeared in these intermediates; and because the range 
obtained was such that hybrids were not accredited to regions in which 
the parents did not exist. Unfortunately the time required to grow 
plants of Amelanchier in order to test the hybrids, and the fact that 
the writer is not conveniently located for such work, has made it 
impossible to test the hybrids assumed to exist. This paper must, 
therefore, be regarded as a provisional treatment based upon the 
hypothesis that the products of hybridization within this genus are 
extremely common. Indeed, such products of hybridization seem so 
common that in every large herbarium thus far studied the number 
of sheets to be considered of hybrid origin is about one-third of the 
whole. This great number of hybrid forms may seem to some unduly 
large when it is considered that Mendel ! and others have shown that, 
in nature, there is a strong tendency for the gross mass of individuals 
representing species and hybrids to return to the two specific forms. 
It may seem that we are assuming unreasonably frequent crossing to 
occur. It is probable, however, that many forms which are the results 
of hybridization come true to seed. "That certain new forms thus 
produced may come true to seed is not opposed to the Mendelian 
1 Mendel: — See Mendel's Principles of Heredity, by Bateson, p. 58, 1902. 
