CONTRIBUTIONS FROM THE GRAY HERBARIUM OF 
HARVARD UNIVERSITY — NEW SERIES, No. XLIX 
I. A REVISION OF THE NORTH AMERICAN 
SPECIES OF AMSINCKIA 
By J. Francis MAcBRIDE 
Tue genus Amsinckia reaches its greatest development in 
California, but several species are indigenous to Oregon and Wash- 
ington, one has been described from Mexico and one or more are 
known from western South America as far south as Patagonia. 
The range of one species apparently extends from northern Idaho 
and adjacent Washington to Arizona and California but most 
species are confined, at least as to their indigenous occurrence, to 
much smaller areas. Several, however, have greatly extended 
their natural ranges by spreading as weeds. Indeed the weedy 
character of Amsinckia is most pronounced. I have observed 
species in California and Idaho that, by reason of their aggressive- 
ness, have been able to take possession, as it were, of places laid 
waste of the indigenous vegetation. There Amsinckia, in the most 
severe colluctation with other native and, more especially, intro- 
duced plants that frequent areas where the natural equilibrium of 
species has been disturbed, has been able, not only to hold its own 
but often even to flourish at the evident expense or destruction of 
less aggressive types. One species is a frequent weed in the grain 
fields of eastern Washington (cf. Piper and Beattie, Fl. of 8. E. 
Wash. 209) and of Idaho, and one or more species have gained a 
foothold on the Atlantic coast (ef. Taylor, Fl. of New York and 
Vicinity, 522, and Macbride, Rhodora xviii. 23). But the distribu- 
tion of Amsinckia as a weed is not confined to the United States. 
I have seen specimens from Australia and as long ago as 1895 
Hansen in Bot. Tidss. xix. 302 recorded two species as adventive 
in Denmark. Moreover Dr. Thellung has written me that “ cer- 
taines espéces de ce genre se trouvent ¢a et la adventices en ON 
introduites probablement avec des blés américains; mais il m’a 
presque toujours été impossible, faute q un travail d’ensemble, de 
les determiner exactement et sdrement 
1 
