80 BOTANICAL GAZETTE [JULY 
on vegetative characters or places too much dependence upon the often 
fickle ‘“‘aspect.’’ Circumstances of this nature doubtless contributed 
largely to the treatment by NE son (Proc. Biol. Soc. Wash. 18:187. 
1905) of Phoenicaulis Menziesti as a species of Arabis, a disposition that 
was adopted later by NELSON and Macsripe (Bort. Gaz. 55:374. 1913). 
It must be admitted that the arguments in favor of this treatment are 
far from weak; on the other hand, the highly technical nature of the 
characters to be considered in the proper definition of groups in a natural 
family must be borne in mind, and PrRantt has used to advantage, in 
“keying”? Phoenicaulis and Arabis, the type of characters that furnish 
the best contrasts between Parrya and Phoenicaulis. The very possi- 
bility of considering P. Menziesit as an Arabis becomes, therefore, a 
strong argument for its retention as a genus distinct from both Parrya 
an 
We now come to a consideration of the plant which prompted these 
observations. This plant was described by Gray (Proc. Am. Acad. 
6:520. 1866) from meager material that was far past condition as Draba 
eurycarpa, and recently has been redescribed as Parrya Huddelliana A. 
Nels. (Bor. Gaz. 54:139. 1912). Here again we have an instance of 
the similarity of genera in this family, especially as regards vegetative 
characters. This plant would not seem at all out of place in Draba 
were aspect the only criterion we had to judge it by; and indeed the 
original ; consists only of two small plants which are so mature 
that the seeds have all fallen. But upon examination of complete 
material it becomes obvious that Gray’s species is allied to Parrya and 
_ Phoenicaulis. Itis not satisfactory, however, to refer it to either of these _ 
The branching pubescence, the inconspicuous white flowers, 
the subentire stigma, the broadly ovate-lanceolate pods, and the nearly 
membranous septum are some of the characters that forbid its reference 
to Parrya. The loose cellular testa about the seeds, the not at all tortu- 
ous areolae, and the i inconspicuous flowers are also characters in direct 
re 
contrast to Moreover, there is the unique habit 
which suggests. Draba rather than either of the genera to which it is 
most nearly related, but consideration of it as a Draba (to mention one 
outstanding feature) is out of the question because of the singular seed 
3 coat Although this is suggestive of the seed coat of Fore, it is of a 
different quality and is not winged. | 
Now in the proper generic allocation of these plants c “on: 
be given only to the value of 
