258 ON A NEW CHINESE BIGNONIAD. 
phurei subbilabiatilobis 4 rotundatis crispulis superiore majore, stamini- 
us 4 didynamis sine quinti rudimento anticis corollee tubum equantibus 
filamentis rubris antherarum lüteolarum glabrarum loculis divaricatis, 
magnis nigris depresso-subrotundis cincta rostro ad ges obtuso gla- 
berrimo terminata subtomento glandulis parvis elevatis conspersa locu- 
testa multo minore pallide brunnescente embryoni conformi idque 
arcte involvente cotyledonibus didymis 41 lin. latis basi apiceque fere 
eequaliter conspicue emarginatis radicula hilum spectante extra cotyledones 
vix protrusa. 
Spathodea stipulata, Wall., which is apparently the nearest ally of the 
bed i i 
Hai-nan species, is described* as having 4 to 5 pairs of leaflets downy 
phragma,t to which De Candolle assigns a three-lobed calyx.t ; whilst 5. 
stipulata is the type of Seemann’s genus Markhamia.§ Of neither genus 
as understood by him had my late friend, up to the time of his lamen 
death, given, so far as I am aware, any character. In his “ Revision of 
the Natural Order Bignoniacez," published in the ‘ Annals and Magazine 
of Natural History,|| "Dr. gE Up a tribe Jacarandee, composed 
M eae! p 
whilst he distinguished his new group, to which he referred both Mark- 
hamia an ,Heterophragma, by its * Capsula marginicida, se tum valvis 
oppositum.’ I have not succeeded in finding the definition of marginicidal, 
whic 
ed, but De Candolle himself expressly defines it** as signifying “ de- 
Cand. Prod. ix. 205; Mi Tet £901. 
+ Journ. of Bot. VIII. 339. quel, Ann. Mus. Lugd.-Bat. 1 
|| I cann 
separately printed wi 
S Prod ilh s e 
tt Contrib. to Bot. ii. 52, Seo also page 47. 
