regular shrub, with many branches, terminated by large trusses of bright-yellow flowers. Asa Gray says this plant is "common northward" in wet ground, but I was not lucky enough to hit upon it during my visit to America. However, I have seen living specimens from Teesdale in England, and from them I perceive that, in general habit, the plant greatly approaches the rock-roses (Helianthemum), which grow in very similar situations. The leaflets of the shrubby potentilla, long, narrow, and silky beneath, resemble, at first glance, the leaves of the rock-roses, thus showing how similar conditions tend everywhere to produce similar results, even when starting from the most unlike organic forms to begin with.
One other potentilla, the goose-weed or silver-weed (P. anserina), I must needs mention for form's sake, though I have nothing special to say about it. It is a creeping species, growing close to the ground, with long pinnate and prostrate leaves, silvery white below, with silky down. Both in Europe and America it is very common as a road-side weed, and in moist ditches; but with us it is a weedier and scurvier plant than with you—evidently a sufferer from our long civilization. In America it grows mostly by river-banks and in brackish marshes; in Europe, it belongs rather to waste places and stony pastures than to streams or mud-banks. Few temperate plants, however, have a wider distribution. It is a circumpolar weed in both great continents, extending through Russia and Siberia to Alaska and British America, and it reappears once more, under like conditions, in the southern hemisphere. Nothing kills it out, and it will bear both inundation and trampling under foot to a greater degree than any other plant of equal importance.
The handsomest of your American potentillas, however, is the marsh five-finger (P. comarum or palustris), a very bold and elegant waterside plant, bluish-green in stem and leaves, and with loose corymbs of exceedingly pretty though dingy flowers. The calyx, inside, is lurid-red, and the large petals are tinged with a gloomy and peculiar purple. This fine ornamental plant loves cool northern bogs and marshes, being common in Canada and in the Scotch Highlands. But what gives it to me the deepest interest is its exact resemblance in hue and general aspect to a purple avens (Geum rivale), also common to either hemisphere. Both are plants of the cold swamps and peaty places; both depend for fertilization upon water-side insects; both have lurid-reddish calyxes, and both have large and dingy purplish petals. The inference seems to me irresistible that the color has been evolved in both cases by the special tastes of the upland water-creatures to whose aid both owe the impregnation of their ovules. Indeed, it is often easy thus to classify flowers functionally by their color and the tastes of the particular insects that habitually visit them. In Europe, at least, I believe the particular insect in this case to be Rhingia rostrata, which I have observed in great abundance upon both flowers. Amer-
