356 ON THE STRUCTURE AND 



supplement to the description of the plant which I have 

 given in the work referred to. 



Akenia membranacea, insecta parva alis conniventibus 

 quodammodo referentia, perianthio parum aucto stamini- 

 busque persistentibus cincta, iisque sesquilongiora, fere 

 distincta, ipsa basi, ubi receptaculo communi inserta, post 

 separationem intus aperta ibique e membrana simplici 

 crassiuscula imberbi nitente formata ; supra clausa et e 

 duplici membrana conflata ; harum exterior dense barbata, 

 pilis longis, strictis, acutis, deflexis, stylo persistenti brevi 

 arete reflexo rostrata : membrana seu lamella interior tenuis, 

 intus quandoque dehiscens. 



Semen unicum (rarissime duo), basi cavitatis membranae 

 interioris insertum, oblongo-ovale. teres, funiculo umbilicali 

 brevi juxta basin affixum. Integumentum duplex : Testa 

 membranacea laxiuscula, raphe tenui laterali et apice chalaza 

 parva insignita : Membrana interior tenuis separabilis. Albu- 

 men semini conforme, album, carnosum, subfriabile, e materia 

 oleosa cum granulis minutis mixta constans. 



Embryo parvus, in basi axeos albuminis, teretiusculus, 

 albus, rectus, albumine 4 — 5ies brevior. Cotyledones breves, 

 plano-convexse. Radicula teres, basin seminis attingens. 



Receptaculum commune fructus : tuberculum centrale, 

 parvum, brevissimum, subcylindraceum, cujus lateribus 

 bases apertse akeniorum adnata? sunt, apice convexiusculo 

 barbato. 



From this description, especially of the embryo, it is 

 evident that Cephalotus must be removed from Rosacea?, 

 to which it had been referred by M. Labillardiere ; and 

 also, though not with much confidence, in the account which 

 I published in 1814. M. de Jussieu, indeed, in 1818, 

 3i6] proposed to exclude it from Rosacea? and append it to 

 Crassulaceae ; and the structure of the seed, as well as of 

 the folliculi or akenia, and even their insertion on the minute 

 central receptacle or axis, may seem to confirm the correct- 

 ness of this approximation. 



Cephalotus, however, still appears to me sufficiently remote 

 from every natural order at present established, to entitle it 



