ANALOGIES AND AFFINITIES. 297 
_ to the grasses and the sedges. From the aspect of the leaves 
I inferred the plant to be some lettuce wort either new to 
science, or at least to me unknown. Breaking off one of the 
stubbly stems, there exuded the expected milky juice, and I 
felt even more confident that I had before me some curious 
thing of the lettuce family, and this illusion was not dispelled 
until I drew near a rocky chasm where, out of reach of the 
ruminants, my new plant bloomed a most beautiful blue 
Lobelia.’ I had forgotten that, in these Santa Barbara hills, 
I was upon the native ground of one of the types of the 
entirely artificial genus, Palmerella. 
Experiences akin to this have not been rare with me ; and 
I have come to associate these two families of plants in my 
mind as related in a degree far more intimate than many 
herbarium botanists are likely to have thought of, or to be 
able to appreciate. But there is at least one good morpho- 
logical point of affinity between the two to which I would call 
renewed attention. It has long been known that while the 
pollen of all other tribes of Composite is globular or merely 
somewhat elongated, that of the Cichoriaces is distinctly and 
invariably dodeeahedral; a most remarkable and very funda- 
mental morphologieal difference ; such as no student of affi- 
nities can afford to overlook; and the same thing is more 
or less characteristic of at least a goodly number of the 
Lobeliaceæ. Many of the genera exhibit a decidedly angular 
pollen; few that I have examined, a globular grain. 
! LosEnrA Rornunockrr = Palmerella debilis var. serrata, Gray, Wheeler's 
Report, 367. T recognize this as quite specifically distinct from the type 
of Palmerella, which latter should be dedicated to its real discoverer, 
Geo. W. Dunn, under the name 
LoBErnrA Dunnit = Falmerella debilis, Gray, Proc. Am. Acad. xi. 81: 
Bot. Cal. i. 619. 
A third member of this group of Lobelias with long, slender and 
uncleft corolla tube may be called i 
LOBELIA PALMERI = Palmerelia tenera, Gray, Proc. Am. Acad. xxii. 433. 
All these western plants have, as regards the accidental peculiarity of 
their corolla, a counterpart in the Atlantie American Z. amæna, as M. 
Baillon has said. 
