194 BOTANY. 



to the European G. ciliata, also a perennial species, which used to be claimed 

 as an annual; barbettata, however, has seeds similar to those of serrata, though 

 much smaller, while ciliata has the winged seeds of simplex, and has an indefi- 

 nite number of leaves. I have since had the opportunity of studying barbel- 

 lata in the mountains of Colorado, and found that it possesses a creeping, 

 filiform rhizoma, 2-3 inches below the surface, from which at intervals filiform 

 stems arise. These bear, at their thickened upper end, where they reach 

 the surface of the soil, an undeveloped terminal bud of indefinite growth, 

 and lateral annual flowering stems, the scars of which, enveloped by 

 withered leaf-bases, can be traced sometimes for five or six years back. 

 The vegetation of the plant is accomplished in the following manner. — Each 

 season the terminal bud developes two pairs of basal leaves ; from the axil 

 of one of the outer leaves, the single flowering branch originates. Inside of 

 the two leaf-pairs just mentioned, we find a third and a fourth pair unde- 

 veloped, about half an inch long, which are to grow into the basal leaves in 

 the following season; and within these the four leaves of the next succeeding 

 season, now only half a line long, are already preformed. The flowering 

 branch, usually 3 or 4 inches high, normally bears one pair of leaves in 

 the middle, and a second involucral pair just below the almost sessile 

 flower ; the four sepals are opposite these four leaves, and the four corolla- 

 lobes alternate with the sepals, and so on. In the axil of one of the third 

 pair of basal leaves preparing for next year, usually alternating with, or 

 sometimes opposite to, the present flowering branch, the bud of next year's 

 flowering apparatus is already four lines long ; it shows plainly the two 

 pairs of leaves and the calyx, and, in a very rudimentary state, also the 

 corolla. Thus each year's vegetation exhibits at the flowering period 

 (August and September), on the primary axis, two pairs of leaves for the 

 present, two pairs for the next, and two for the third year, a secondary 

 axis with two leaf-pairs and the flower, and another preformed secondary 

 axis with the rudiments of the same organs for the next year. No other 

 Gentian has, as far as I know, such a typical growth, with the regular 

 preformation of all the organs, but we find the same among other plants in 

 other families, a striking example of which is furnished by our Nelumbium. 

 The regularity in our Gentian is not as absolute as in Nelumbium; for occa- 



