APPENDIX. 181 
in the genus Campylanthus, the seeds of which have a perisphe- 
rically-curved embryo, a character that by itself would place it 
in Atropacee ; but that distinguished botanist fixes its position 
among Scrophulariacee, on account of the form of its corolla and 
of its anthers, notwithstanding, as he observes, that it bears little 
analogy with any other genus contiguous to it. The principal 
reason however that appears to me to give the Verbascee the pre- 
ference of a place among the Scrophulariacee is the truly axillary 
origin of the floral peduncles, a character that in all such doubt- 
ful cases may be employed as a decisive line of demarcation be- 
tween that order and the Atropacee. The position of the Ver- 
bascee should then appear at the head of the Scrophulariacee, 
occupying the place of a suborder in the manner of the Salpi- 
glossidee of Bentham (DeCand. Prodr. x. p. 190), where they 
would serve as a connecting link of the closest affinity between 
these two families. 
Rerz1a.—This anomalous genus* has never yet found a cer- 
tain or satisfactory place in the system, and its position must 
remain problematical until the structure of its fruit and seed be 
more accurately investigated. By many botanists it has been 
placed in Convolvulacee ; others have indicated its relation. to 
Apocynee ; some have again referred it to Polemomacee, with 
which it certainly offers no affinity ; and Bartling proposed for it 
a new natural order, under the name of Retziacee, but this stands 
upon too insufficient grounds. Endlicher places Retziacee as a 
doubtfal order after Solanacee, and Dr. Lindley arranges the 
genus Retzia among Solanacee, after Sessea. Tn the form of its 
calyx and of its corolla, the number and position of its stamens, 
its bilocular ovarium with placente attached to the dissepiment, 
the structure of its capsule and of its seeds as far as they are 
known, offer characters strictly conformable with those of Sola- 
nacee ; but it would now rather fall among the Atropacea, on ac- — 
count of the estivation of its corolla, which is said by Endlicher 
and Lindley to be imbricate, and not valvate : the form of its 
embryo, which on the authority of Brown (Prodr. 482) is terete 
and straight, necessarily, if it were admitted into this family, 
would point to its situation as a, tribe near the Solandree. 
Thus far every feature appears 10 conformity with such an ar- 
* Rerzia, Thunb. Calye tubular, 5-fid, lobes lanceolate, somewhat unequal. 
Corolla tubular, eidaigated., straight, tube in no degree plicated, pote of 
5 short equal lobes imbricated in estivation. Stamens 5, equal, near hae’ 
sile in mouth of tube, alternate with its lobes, filaments extremely = ort, 
Anthers oblong, cordate, 2-celled, cells parallel, bursting longitudinally in 
front. Ovarium oblong, seated on a fleshy gland, 2-locular, ovules Ase 
placentz adnate to the dissepiment? Style longer than corolla, fi oes 
Stigma very short, bifid, with divaricate linear segments. Capsule bi 
cate, bil . bivalved. Seeds several. Embryo ee 
