140 
H. H. BARTLETT 
it is obvious that they cannot bring about the origin of any fundament- 
ally new type. We cannot, therefore, have any comprehensive 
knowledge of the genetic relationships of organisms until we know 
something about progressive mutation. Mendelian researches have 
shown us how to verify supposed relationships in the simplest cases, 
those in which characters have been lost. But can we always be sure 
that the recessive is the derived form? Is it not sometimes true that 
the dominant is the derived form, and that a new character has ap- 
peared? On this point the evidence is very unsatisfactory. Cases of 
progressive evolution have indeed been observed, but for the most part 
they have occurred out of the beaten track of the Mendelian and have 
been largely discounted, for Mendelism is the present fashion in 
genetics. In Primula sinensis, however, Professor Keeble has been 
fortunate enough to observe the origin of a Mendelian dominant. 
There are two types of gigantism in Primula. In one type the 
chromosome number is the same as in the typical forms, and the 
giant acts as a simple Mendelian dominant when crossed with the 
parent. The other type is of a very different nature, and will be 
referred to later. A number of giants of the simple Mendelian type 
have been known among the cultivated varieties of Primula, but no 
one knew how they had originated until the "Giant White Queen 
Star" appeared as a mutation in the third guarded generation of a 
pure line of the "White Queen Star" variety. It came true from seed 
and has remained uniform through five successive generations, which 
included several hundred individuals. Although we cannot doubt 
in this case that a new dominant has arisen, there is unfortunately 
one flaw in the evidence. The new form has proved sterile when 
crossed with the parent form and with all other non-giant forms, so 
that its dominance is only inferred from its perfect analogy with other 
giants of the same type. The latter have been found by Gregory to 
show simple dominance over the non-giant forms. Our argument is 
based upon evidence fitted together from two sources. Other examples 
might be brought forward, but they are not as striking or as well 
authenticated as this one. There is a type of non-Mendelian pro- 
gressive evolution, however, which has been far more satisfactorily 
studied. I refer to mutations in which new characters may be de- 
finitely associated' with mutative changes in the chromosome number. 
It appears that, in general, each species is characterized by a 
constant number of chromosomes. Within a genus, different species 
