50 



ICONES PLANTARUM. 



fascicnlata, |-| poll, longa, sessilia, teretia, snbacuta, sicca flava. 

 Spicm 2-4 poll, longa. Floras ^ poll, diam., fasciculati, sessiles; 

 bracteas floribus breviores, ovatse persistentes ; bracteolaa obtuse S-lobae. 

 PeriantJin segmenta 5, exteriora minora, snberecta, incurva. Stamina 

 4-5, situ subvaria, unico (ut videfcur semper) segmento exteriore peri- 

 anthii opposite, cetera segmentis alterna v. subalterna ; filamentis 

 perianthio longioribus filiformibus ; anthers parv^, extrorsum de- 

 hiseentes. Ovarium ovoideum, substipitatum, stigmatibus per paria 

 divaricatis. ^ Fruchcs subglobosus, paullo compressus, perianthii setr- 

 mentis longior, 8-costatus. Semen erectum, obovoideum, compressum ; 

 embrjo fere annularis, radicula infera. 



The genus LojMocarpus of Turczanimow (Wallinia, Moq. in T)G. 

 Prodr. xii. pt, ii. 143) has hitherto consisted of two species, both South 

 African,* of which the first known, i. polystachyus, has been collected 

 by Drege alone, and is described as having 5 perianth-segments and 

 as many stamens, characters repeated in the Genera Plantarunu In 

 preparing the plate of the second species for this work, the artist, Miss 

 Smith, pointed out to me that 4 was a prevalent number of its 

 stamens, which induced me to examine many flowers of both species, 

 with the result that 4 is a common number in both, and that they are 

 irregularly inserted in both ; as also that 4 unequal perianth-segments 

 with 3 stamens occur in some of the upper flowers, at any rate of 

 If. poly stachy lis. In both species, indeed, one segment la always smaller 

 (and external, I think) to the others. Then, as regards the positions 

 of the stamens, as a rule they are not symmetrically placed, one being 

 always opposite the small perianth-segment, while the other three or 

 four are approximately alternate with the other segments. This strong 

 tendency to alternation of the staminal with the perianthial whorls, 

 raises the question whether LopUocarpns should not be transferred to 

 Plnjtolaccacem, an order of which only two species {Phytolacca stricta, 

 Hottm., and P. ahyssinica) have been found in South Africa. If it 

 can be pointed out that there is any genus of that order to which it 

 was at all allied, or even a tribe into which it could be placed, I should 

 recommend its transference. As it is, however, it differs in most 

 important characters from all those tribes -from Pivinece in the em- 

 bryo; from Euphytolaccem in the solitary carpel with free stigmas; and 

 trom the Austrahao Gyrostemonece in the bisexual flowers and perianth. 

 Under these circumstances Lophiocarpus will perhaps continue to be 

 regarded as an anomalous member of the tribe Chenopodiece. of Cheno- 

 podiacem, with which it has most characters in common whilst differing 



these 



^^^1 ^r '»!° f "^ f'''"'*^" ^^■'^'^t stigmas and tubercled fruits without ribs. It 

 may be thus characterised : 



L. tenuissimns, Hook f. (nov. sp.) ■ annua, foliis filiformibus, spicis graciUimis, 

 S nC r;!; l™i?" segjnentis v-aMe in.qualibus. stign.atibus'nnnu1is eracti., 



Tab. 1463, fig. 10. Jl. 



Hab. The Transvaal, at Pretoria, Dr. A. Uchmann (ITerb. No. 401S). 



