534 
47. COMPOSITACEJE. 
gr., red or purple towards tlie subobtuse scarious tips. Ach. of 
outer row embraced or enwrapped by the conduplicate inner 
scales of inv. and about half their length, subcylindric, nearly 
smooth or even, with short cylindric beaks as thick as them- 
selves, abruptly truncate and crowned with a cup-like coronet 
of very short pale chaffs. Inner ach. all free slender fusiform 
dark chestnut-brown elegantly and distinctly echinulato-ribbed, 
gradually attenuated upwards into fine slender beaks or pappus- 
stalks as long as or a little longer (rarely in the outer ach. a 
little shorter) than themselves, the beaks becoming smoother 
upwards. Pappus pale brown chaffy-setose of about 10 or 12 
(8-15) bristles lanceolately dilated at the base, the fine seta- 
ceous upper part thinly plumose. Recept. convex alveolate 
naked, scarcely or very shortly velvety-fibrillose. 
Such is the universal everywhere abounding Mad. pi., exa- 
mined year after year a thousand times and never varying in 
r he least. It differs constantly from * T. hirta Roth as de- 
scribed by MM. Grenier and Godron in FI. de Fr. ii. 296 in 
having nothing about its root answering to u Souche courte, 
tronqvee et emettant surtout du collet de fortes fibres JUifonnes 
et nombreuses ,” in its extremely hispid-furry 1., and in the 
beak or pappus-stalk being in all the ach. except the outer row 
as long as the ach. the mselves i. e. twice as long as in T. hirta 
Roth. With this account agrees also in the main the figure of 
T. hirta Roth in EB. t. 555, nothing like which certainly has 
been noticed in Mad. No great dependence can be placed 
perhaps on the duration of the root in either sp. At least in 
the Mad. pi. it seems, though normally per., to be much in- 
fluenced by local accidents or circumstances. In all other re- 
spects, however, nothing can be more constant in its charac- 
ters than the present pi. as found both in Mad. and in the 
Canaries. 
I have never seen the beak of the ach. so long or so abrupt 
at its origin as in IlFG. 1. c. f. ii. 3, in which it is figured twice 
the length of the ach. themselves ; cf. T. maroccana Willk. et 
Lange 1. c. 
Ifj/Ofieris pij(jm<pa Ait. ! Ilort. Kew. ed. 1. iii. 130, supposed 
by Schultz in WB. 1. c. very naturally without examination of 
the original spec, to bo the present pi., was founded by l)ryan- 
