340 MR. J. MIEllS ON SOME SOUTH-AMERICAK 



membranaceous endocarp, is unilocular, and contains about twenty- 

 seeds, irregular in shape, imbedded in a sweet pulpy substance 

 which fills the entire space : these seeds, subcompressed and 

 oblong, 12 lines long, 6 lines broad, have a thin, striated, very 

 white testa, surmounted by a laciniiform placenta, are imbedded 

 without any order in the pulp, and by the progress of growth 



have been transferred from their parietal place of origin into 

 the middle of the cell, enclosed in the pulp after the manner 



described in Crescentia (ante, p. 338). The exalbuminous embryo 



appears, from Aublet's figs. 4 and 5, to consist of two pear-shaped 



plano-convex cotyledons, turned back and suspended from one end 



of the terete radicle, of equal length. 



We find a repetition of nearly all these extraordinary charac- 

 ters figured in Delessert's 'Icones' 1 under Kigelea cethiopica, 

 Decaisne (Kigelia pinnata, DC. 2 ). There are, however, many 

 discrepancies in the descriptions of authors relative to the nature 

 of the embryo in Kigelia. DeCandolle quotes the authority of 

 Bojer, exemplified by his unpublished drawings, which show that 

 the exalbuminous seeds are oblong with rounded cotyledons, ex- 

 ternally plicated longitudinally 3 . This refers to Kigelia pinna ta> 

 DC. 4 , which appears, from the observations of Prof. A. DeCan- 

 dolle 5 , to be probably identical with the SpatJiodea campanulata, 

 Beauv. 6 , from Equinoctial Africa. It is not stated from what 

 quarter Delessert derived the information upon which he figured 

 his analysis of the seed given in his ' Icones.' This knowledge 

 was probably derived from Bojer's unpublished drawings alluded 

 to. According to his analysis, the embryo is seen folded in 

 fig. 10 ; it is expanded in fig. 11, where two orbicular foliaceous 

 cotyledons, applied together, are both folded at their middle, so as 

 to make their two halves turn towards and touch one another : 

 in this way they appear as if suspended from one extremity of an 

 inflected fleshy radicle of their own length — a transverse section 

 of this within the testa being shown in fig. 9, and another trans- 

 verse section of the same removed from the testa in fig. 12. All 

 this is sufficiently manifest. 



On the other hand, we have a very different version of the deve- 



1 Deless. Icon. v. tab. 95 B. * Prodr. ix. p. 247. 



3 In eodein loco, p. 247. 4 Prodr. ix. p. 247, in adnot. 



* Prodr. ix. p. 208.. e Benth. Niger Flor. p. 461. 



