232 
HARLEY HARRIS BARTLETT 
fair to state that he does not altogether agree with the writer as to the 
specific independence of the two forms. Thus, under date of September 
26, 1913, he writes: "... I must confess that I prefer to consider 
them as subspecies of Oe. cruciata Nutt. and would have published 
them as such if left to my own initiative. But of course if you are 
consistently describing everything which is genotypically distinct 
and morphologically distinguishable as species, these forms certainly 
deserve the same treatment." 
The so-called Oe. cruciata of recent collectors, of which numerous 
specimens have accumulated in our herbaria since 1890, seems to 
consist not of one species or group of genetically related species, but 
of the cruciate varieties of several independent broad-petaled species, 
together with a number of species in which the cruciate character has 
probably become specific. In other words, Oe. cruciata, sensu latiore, 
is a purely artificial group, the components of which must ultimately 
be distributed according to their true relationships. The writer 
believes that cruciate variations of any species of Onagra are likely to 
occur as bud sports or germinal mutations. The almost certain origin 
of the cruciate variety of true Oe. biennis on the sand dunes of Holland 
is a case in support of this belief. One solitary individual of this var- 
iety was found by Ernst de Vries in August, 1900, growing with the 
typical form on the dunes near Santpoort. No such form had hither- 
to been known. As de Vries^*^ pointed out, it could not have originated 
by hybridization with any other cruciate form, since there was no such 
form in Holland at the time except the rare Oe. cruciata var. varia of 
botanic gardens. Quite aside from the extreme improbability that 
two normally self-pollinated strains growing at a great distance should 
have been crossed, it will suffice to point out that Oe. cruciata var. 
varia de Vries, is characterized by a dark red coloration of the stems 
and foliage which is dominant in all its hybrids. In the cruciate 
variety of Oe. biennis there is no such coloration. The writer's 
cultures of 191 3 included fifty plants of it, from seeds sent by de Vries, 
descended from the original plant found by his son. They were a uni- 
form lot, and differed in no respect from Oe. biennis, sensu strictiore, 
other than that the petals were linear instead of obovate. The 
variety can, therefore, hardly be regarded as other than a mutation, 
the occurrence of which gives a tangible clue to the manner in which 
other cruciate forms have arisen. 
10 De Vries, H. Die Mutationstheorie, 2, 1903, p. 599. 
