188 
CONTRIBUTIONS TO BOTANY. 
fruit 12 lines long, gradually thickening upwards, and 5-sul- 
cate ; the calyx is 1^ line long, 2 lines in diameter, seceding 
from the summit of the pedicel and remaining strung upon it, 
in the form of a fimbriated annular disk ; the petals are 7-8 
lines long ; the capsule is 6 lines long and broad, pointed at 
each extremity, opening upon the shai-p, sutural, much undu- 
lated edges of the valves, which, cuiwing back horizontally, 
show the seeds attached to each side of the margin of the 
dissepiments ; the pyriform seeds are 2 lines long, l-^ line in 
diameter 
5 . Crinodendron. 
This name was given to the well-known Patagua of Chile 
by Molina, who described it so imperfectly that Ruiz and 
Pavon did not recognize it when they founded their genus 
Tricuspidaria upon the same plant. Molina gave Cavanilles 
a rough drawing, showing the flower and seed, made from 
memory, which the latter described and figured in his ^ Dis- 
sertationes,’ the characters there assigned to it being altoge- 
ther erroneous. Sir William Hooker, in 1833, described a 
plant from the island of Chiloe, collected by Cuming, which 
he supposed to be the same as that incon-ectly described by 
j\Iolina and Cavanilles, and accordingly named it Crinodendron 
Patagua. In giving an outline of its generic character, he 
wrongly described the flower as having no calyx, which had 
fallen away from Cuming’s specimens ; the inflection of the 
petals was not noticed; and the remarkable glands were not 
observed upon the disk, which was figured as being simply 
columnar. Gay, in his ‘ History of Chile,’ eiToneously de- 
scribes the calyx ; but he gives an account of the structure of 
the fruit, which was not known previously. Crinodendron 
cannot be said to have existed as a genus until Sir William 
Hooker first established it in his ‘ Botanical Miscellany ;’ and 
he, perceiving its near affinity to Tricuspidaria., placed it in 
the .£'feoca?y)ecB, notwithstanding the then apparently discordant 
characters of its floral envelopes. He CandoUe has not noticed 
the genus ; but Endlicher placed it in his tribe Tricuspidarice, 
in association with Vallea and Tricuspidaria. Bentham and 
Hooker, in their ‘ Nova Genera,’ have regarded it as a syno- 
nym of Tricuspidaria, evidently unaware of the characters 
which separate it from that genus. The following is an 
amended diagnosis, aecording to my own observations, as far 
as regards the floral structure ; not having seen the ripe fruit, 
* A figure of this species, with full structural details, is shown in 
Plate 82. 
