CRUCIATE-FLOWERED OENOTHERAS OF SUBGENUS ONAGRA 237 
to point out (i) that it occurred in the fourth guarded generation of the 
strain, (2) that the species is strictly self-polUnating because the sta- 
mens surround the stigmas and because the flowers generally open only 
imperfectly or not at all, (3) that any accidental crossing with a broad- 
petaled species would have been evident in the next generation, (4) 
that until the year when the mutation was found there had been no 
other cruciate strain in the garden or vicinity with which it could have 
accidentally crossed, (5) that the densely hairy petals constituted an 
absolutely discontinuous variation, for in all the other plants of the 
strain they were glabrous except under microscopic examination, and 
(6) that the new character could not have been introduced into the 
strain by hybridization, inasmuch as no other Oenothera with which it 
could have crossed has hairy petals. The new mutation may be 
called Oe. stenopetala mut. lasiopetala.^^ 
1^ De Vries (Bull. Torr. Bot. Club 30: 75-82. 1903) reports considerable varia- 
tion in the degree of pubescence of the petals of Oe. cruciata v. varia. This form is of 
garden origin and is known only in European botanic gardens. In America there 
seems to be no Onagra of which the petals appear other than glabrous to the naked 
eye. 
The writer proposes that a trinomial system of nomenclature be used for 
mutations of garden origin, in order to set them clearly apart from forms of which 
cognizance must be taken in floras. Gates has made the suggestion that mutations 
be indicated by placing the abbreviation "mut." between the terms of the binomial. 
According to his practice, for example, the name of Oe. lata, the well-known mutation 
from Oe. Lamarckiana, would be written Oe. mut. lata. The use of a trinomial 
system, however, would appear to have decided advantages over a binomial system. 
Gates, for instance, has reported the discovery of a mutation from Oe. biennis parallel 
to Oe. lata, which he has called Oe. biennis mut. lata in order that the name might 
show the parallelism of the two variations. Certainly it is very helpful to be able 
to use the same mutational designation in such cases, and a trinomial system makes 
it possible to so do consistently. The use of the trinomial for all mutations would 
not imply anything about the degree of differentiation of any particular mutation; 
in other words, the category mutatio would indicate manner of origin rather than 
kind of variation. Oe. gigas de Vries is a mutation of specific rank; probably most 
botanists would consider Oe. brevistylis to be of varietal rank. If it were considered 
desirable, the abbreviations "mut. sp." and "mut. v." {mutatio specifica; mutatio 
varietalis) might be applied as follows: Oe. Lamarckiana mut. sp. g'^as and Oe. 
Lamarckiana mut. v. brevistylis. The latter variation is found wild in Holland; if 
it were necessary to include it in a flora the name would be Oe. Lamarckiana var. 
brevistylis, The former has been introduced into this country; if it spreads and 
must be taken account of by systematists the name would stand as Oe. gigas. 
