lines thick; whole plant even to the fruit bluish- 

 nd perfectly smooth, the flowers white or purple; 

 rays and petioles round and thick when g-reen, rath- 



nto leaf-she-iths below ; leaves 



to oblong-cordate, 3-5 lobed and lobes pinnate and fan-shaped 

 and imbricated, these closely deltoid-toothed and every third 

 tooth cut deeper, all teeth apiculate, when dry surface is 

 rouo^hened by shrinkage of the very leathery pulp, but per- 

 fectly smooth when fresh: rays 6-8 lines long, of variable 

 thickness in the same umbel, widely spreading; without in- 

 volucres ; involucels of about 8 oval to triangular, hyaline mar- 

 gined, lacerate, caudate, stiff bracts about 2 lines long, which 

 pre united below; flowers sessile, about 10; fruit from the very 

 first broader than long with wings wider than body, and 

 emarginate at both ends, the most mature fruit is apparently 

 not over half developed and is about 2 lines long, apparently 

 like that of C. Rosei (Aulospermum Rosei Jones C. & R. Umb. 

 179) ; wings 3. the central one variable, the later ones broad 

 nnd well developed, the margin being rounded and as thick as 

 the other part or thicker and not thinner as in the former, the 

 base is thickest as in that species then becomes thinnest and 

 then a trifle thicker from the middle of the wing to the edge as 

 is the case with that species except at the very edge ; oil tubes 



closely allied to C. Rosei, yet differing very widely; it has the 

 same tuberous root at the end of a very long underground 

 stem which branches at the top, the same crowns covered with 

 eld and new leaf sheaths, but the crowns are thick and the 

 stems thick, w^hile in the other they are small and the stems 

 almost filiform, the leaf dissection is very different, though the 

 color is much the same on both sides, though in Rosei the 

 glaucous color is scarcely evident, the purple voung fruit is the 

 same, and color of the flowers, the peduncles are about the 

 F-^me, but the rays very much stouter and raylets none. 

 C. Rosei shows marked relationship in habit and leaves to the 

 C. purpureus group, though also growing in the soil frequented 

 by the C. montanus and C. corrugatus groups. This snecies, 

 however, grows in the crevices of ncr.r'y bare lafallic rocks oi 



