112 



in the vicinity of Chavies P. O. (Sand Mountain, it should be ex- 

 plained, is a splendid example of a s)'nclinal plateau capped by 

 Carboniferous sandstones. In DeKalb and Jackson counties it 

 is about 200 times as wide as high, being at least twenty miles 

 wide and only a few hundred feet above the narrow valleys which 

 bound it ; so that when on its summit away from the edges it is 

 difficult to realize that one is on a mountain at all.) Near 

 Chavies a clear stream known as Town Creek runs lengthwise 

 of the mountain, from northeast to southwest, and on exploring 

 the banks of this creek a short distance I made some rather 

 startling discoveries. 



On rocky banks, probably within the reach of floods, Cliondro- 

 pJiora virgata (Nutt.) Britton was quite common. This is a new 

 station for it, though it had already been reported from this 

 general region. * Associated with it was Coreopsis verticillata L., 

 which Dr. Mohr found on Lookout Mountain in the same county. 

 In crevices of rocks a little lower down, in the edge of the water, 

 were some tufts of rush-like evergreen leaves, which at first sight 

 I would have unhesitatingly pronounced an /s'<?(!'Vifi', especially since 

 I knew I was in the only Alabama county from which an Isoetes 

 has been reported. But on pulling up a tuft I discovered that 

 the leaves were jointed in the manner of many /unci, and con- 

 tained no sporangia at their bases. The odor of the plant then 

 proclaimed it to be an umbellifer, and the characteristic double 

 curvature of the leaves (outward and then upward) enabled me 

 to recognize it as a species which of all others I would have least 

 expected there. For this species is not one of those included in 

 Mohr's Plant Life of Alabama, and moreover, it is not even con- 

 generic with anything which was known to exist at the time that 

 work was published. It is the sole known representative of a 

 genus which had been described by Dr. Rose less than two 

 months previously, ^; from two collections made in the coastal 

 plain of Georgia, in 1902 and 1904. A few minutes later I 

 secured enough remains of stems and inflorescence to establish 

 its identity beyond a doubt, and a new genus was thus added 



* For notes on its distribution, see Bull. Torrey Club 32 : 1 68. 1905. 

 f Harperia nodosa Rose, Proc. U. S. Nat. Mus. 29 : 441. pi. j. O 1905. 



