including a new Arrangement of Phanerogamous Plants. 195 



single carpel would occupy might be inferred ; yet from the con- 

 stancy of its position in certain families, this may be considered 

 as no longer a question, as in Chenopodiacese, Amaranthacese* 

 and Plurabaginese the ovule is with rare exceptions posterior or 

 lateral; but in Scleranthus annuus and Calytrix virgata always 

 anterior, in the latter instance two ovules being present. Scle- 

 ranthus therefore agrees with Tetragoniacese, and Calytrix with 

 Myrtacese, &c. ; and its variable position in Thesium tends rather 

 to confirm the evidence, as showing a correspondence with Au- 

 cuba, &c. in the variable position of the fertile carpel. 



The Position of the Raphe. 



From the remarkable regularity of the position of the raphe, 

 both in erect and pendulous anatropal ovules, and also of the 

 cotyledons and radicle in seeds produced from campylotropal 

 ovules, these characters may, there appears no reason to doubt, 

 be relied on as furnishing indications of the position of the car- 

 pel ; and also (the ovule being erect) of the fertile carpel, when 

 two or more being equally developed are united by their margins 

 and form a one-celled ovary. Thus, in Illecebrum and Atriplex, 

 where the ovary is dicarpous, the radicle of the embryo curves 

 down posteriorly, showing, in the latter instance at least, that 

 the placentation is anterior to the seed, as is the funiculus in 

 Beta and Rhagodia; while in Opercularia the position of the 

 raphe is variable, showing the position of the fertile carpel to be 

 variable, as in Cornacese and Caprifoliacese, and more frequently 

 posterior than in the latter f. 



General Characters of the Divisions. 



The division thus separated as Proterocarpous is natural, in 

 having no direct affinity either with Endogens or Rhizanths, with 

 both of which the second or Heterocarpous Division is so inti- 



* Gomphrena globosa, in which the funiculus is lateral, is further remark- 

 able for having the ovule almost always on its right side : may this be 

 owing to the direction of the spires of bracts ? In Cliffortia ilicifolia, 

 where the carpels are all lateral, more than two-thirds of them have their 

 bracts toward the ascending portion of the spire, whether the spire is from 

 right to left, or from left to right ; but so slight a variation is perhaps 

 scarcely deserving attention, and such instances should rather be referred 

 to their nearest affinities. 



t The relation of the raphe to the placenta in any given family must be 

 ascertamed before such an inference can be made, except it is in those 

 cases where the position of the raphe (in relation to the axis of the inflo- 

 rescence) is variable, as then the position of the fertile carpel must be va- 

 riable, of which Brunonia is an example ; and the same rule applies to the 

 position of campylotropal ovules. 



13* 



