226 
AMERICAN JOURNAL OF BOTANY 
[Vol. io, 
species played the role of mother in the cross, but these hybrids were always 
so nearly sterile that in no case was more than one seed found in any capsule 
of one of the Fi plants, nor were more than two capsules, each with one 
seed, ever found on any one such plant, and only rarely did an Fi plant 
have even one seed. At one time two F 2 plants were secured from these 
rare seeds, but these were lost before their adult characters had been ob¬ 
served. The remarkable distinctness of the two species, bursa-pastoris and 
Viguieri, left no ground for surprise at this Fi sterility, and one would be 
inclined, perhaps, to say that Viguieri has deviated so far from the bursa- 
pastoris type, from which it doubtless took its origin, that it can no longer 
produce fertile progenies when back-crossed to the ancestral type. 
When crosses are made between B. bursa-pastoris and B. grandiflora, a 
somewhat different relationship is indicated. Nearly twenty crosses of 
this type have also been made, but only in one case have Fi hybrids (3) 
been secured, though the pollination usually results in the full development 
of the capsules and the ovules usually enlarge to nearly full size and then 
abort. Of three hybrids secured from one of these crosses, all were par¬ 
tially fertile, but fertility was slight in two, and in the third it was far from 
complete. From the most highly fertile of these an F 2 was grown which 
was so nearly sterile that it was not considered practicable to pursue the 
experiment farther. 
The complete failure to secure Fi hybrids from all but this one cross is 
in marked contrast with the easy production of Fi hybrids in the bursa- 
pastoris-Viguieri crosses. Here again it may seem a logical explanation to 
say that Bursa grandiflora has evolved so many (or such considerable) 
divergences from the bursa-pastoris type that it can only rarely produce 
hybrids when crossed to the latter, thus giving seemingly good ground for 
the acceptance of such hybrid sterility as a criterion of specific distinctness. 
But what becomes of the validity of this criterion when it is found that 
Bursa grandiflora and B. Viguieri , which give only sterile hybrids, or none 
at all, with the intermediately placed B. bursa-pastoris , produce fully fertile 
hybrids when crossed with each other? As seen from the table given above, 
the difference between B. grandiflora and B. Viguieri is essentially the sum 
of the differences of each from B. bursa-pastoris, so that we can not logically 
assume that they are specifically distinct from the latter species and only 
subspecifically distinct from each other. 
The non-validity of the sterility criterion is already fully recognized by 
every one in the case of the Orchidaceae, where fully fertile hybrids are 
produced, not only between very distinct species, but also between many 
different genera. 
Still another case in Bursa may be cited to show that two forms recog¬ 
nized as species by taxonomists may differ in only a relatively simple 
genotypic feature which does not affect capacity for interbreeding. Bursa 
Heegeri (Solms-Laubach), which was found by Professor Heeger on the 
market-place at Landau, Germany, in 1890, is so distinct from B. bursa- 
