676 GENETICS: METZ AND BRIDGES 
of the problem, a third, but equally unsuccessful, effort was made by 
transferring such females as were still alive to fresh culture bottles. 
So many matings were attempted that it is impossible to regard their 
failure as accidental or as due to poor cultural conditions. It must be 
considered established that the fertility of the notch-facet hybrid was 
of a different order from that of either parent race. 
Notch has reappeared on at least three other occasions, and in one of 
these cases the facet test has been made. The new notch-facet test gave 
the peculiar notch-facet compound, but this compound was fully fertile. 
The original facet cross had been made to a notch which arose by a 
separate and independent mutation, and was probably not identical with 
the notch of the later test but was an allelomorph. The original notch 
seemed to be a somewhat more extreme type than the others. Further- 
more, a parallel case (in D. melanogaster) has been found, in which two 
allelomorphs that differ little in appearance differ markedly in their fer- 
tility relations. The mutation lethal 2 gave an aberrant linkage result^ 
which led to tests for deficiency by mating lethal 2 females to males of 
mutations whose loci were known to be close to that of lethal 2. In the 
test by the recessive 'club' the same apparent dominance was found as in 
the case of notch-facet; — that is, females heterozygous for lethal 2 
mated to club males gave half the daughters club and half normal, and 
it was easy to demonstrate that the ones that were club were also the 
ones that were lethal 2, for these lethal 2 -club hybrids were fertile. In 
other crosses not involving club very rarely a male having the lethal 2 
gene was able to hatch, and these males, which showed all the character- 
istics of club, were completely sterile. 
3. If we imagine such cases as these to occur in nature it is evident 
that they might be of evolutionary significance. For instance, suppose 
that the two mutants 'glazed' and 'rugose' of D. virilis appeared in the 
wild state. Their establishment would depend, of course, upon the via- 
bility of the mutants, and in this particular instance one of them (glazed) 
would probably be eliminated, although the other, judging from its 
vigor, would stand a good chance of surviving. But the consideration 
of viability need not affect the case as an illustration, for it must be 
assumed a priori that any variants, to be of evolutionary significance, 
must be viable. Supposing, then, that the two mutants in question 
appeared as viable forms in nature; the result would be a composite 
species consisting of three types, two of which were fertile with the third 
but infertile with one another. If for any reason the third type (in 
this case the normal) were to be eliminated, the first two would become 
distinct species, even though they differed in only one or two external 
characters to begin with. 
