The Geographical Distribution of the Frankeniaceae etc. 407 
Toichogonia cosmopolita. 
Section Toichogonia cosmopolita attains a very marked developement 
in Chili, corresponding to that of Australia, and Sect. Eufrankenia of the 
Mediterranean region. Reiche and Jonow (Flora de Chile I, 1896) enumerate 
six species with three varieties which would fall in this section: F. chilensis 
Presl, F. erecta Gay, F. glabrata Phil., F. Nicoletiana Gay and vars, florida 
and aspera Phil., F. micrantha Gay and var. Berteroana, and F. campestris 
Schau. Most of these are shrubby perennials very closely allied to F. pauci- 
flora of Australia and F. nothria of Capland. But at least two, F. micrantha 
and F. Berteroana, are F. grandifolia Ch. & Schl. The most familiar, and most 
widely distributed species, F. Berteroana, long known from its use as a 
source of common salt, is pure F. grandifolia. F. micrantha Gay, should be 
F. grandiflora var. micrantha Gay, corresponding to the inland North Ameri- 
can F. grandifolia var. campestris Gay. Frankenia grandifolia Ch. et Schl. 
was originally described as a Californian coast plant. It not only occurs 
in abundance along the whole of the Californian coast from San Francisco 
to Santiago, but also in the moist alkali soil of central and southern Calif., 
extending thence eastward to Southern Arizona, New Mexico, and South 
to Sonora and Coahuila, Mex. There can be little doubt, however, that the 
plant originally comes from Chili, where at present it not only possesses a 
like extensive distribution with that of North America, but stands in the 
circle of forms of which it is a member, whose developement is peculiar 
to Chili. 
Comparison of sections Eufrankenia and Toichogonia cosmopolita. 
If now, we consider the three species F. grandifolia, F. serpylli- 
folia and F. pulverulenta together, we find a most striking agreement of 
characters. Each is an annual with mostly plane elliptical leaves, each 
occurs extensively along the coast and is distributed over the entire floral 
region to which it belongs. Put together, the three would form a thoroughly 
cosmopolitan species. Yet one would scarcely say that F. pulverulenta with 
its small flowers and minute yellow seeds, is the same as F. grandifolia 
with its very much larger flowers and especially the large smooth reddish 
seeds, 8—40 times the volume of those of F. pulverulenta, or that either 
were F. serpyllifolia with its gray, hairy seeds. Nevertheless the species 
are very closely related. Such a striking agreement is not aecidental, and 
the capacity of each for wide distribution also points to a near relationship. 
Yet in what way are they related? If we should consider F. pulverulenta 
as the parent form we are met by some difficult questions. Above all 
F. pulverulenta, wherever we find it — North Africa, Capland, Songarei, — 
is the same unvariable plant. Specimens taken from these widely separated 
