STRUCTURE.] MORPHOLOGY OF MALLOW-WORTS. 75 



place which five normal carpels should occupy. That border 

 of the pentagon which is first united sends out a series 

 of rounded tubercles, which subsequently become slightly 

 swollen externally and inferiorly, so that each tubercle pre- 

 sents two enlargements ; one external and inferior, the future 

 ovary, another superior and internal, the future style. The 

 latter becomes elongated and raised in proportion as the 

 former increases in size; but as it elongates, the stylous 

 portions, remaining distinct at their summits, are confounded 

 at their base, at least all those which correspond to the 

 same angle of the common support of the carpels ; an angle 

 which becomes more and more marked as far as the point at 

 which the entire body is, as it were, cut into five oblique 

 lobes loaded with* ovules on every part of their surface. A 

 bundle of styles, equal in number, distinct superiorly and 

 united inferiorly, thus corresponds to each of these systems 

 of ovaries ; and each of these systems, in the general sym- 

 metry, plays an analogous part to that which we have found 

 assigned to each of the bundles of stamens, because it 

 occupies the place which a single carpel should occupy, and 

 which it consequently represents. How is the cavity of the 

 ovary formed ? 



M. Duchartre has not in this case found that the margins 

 of a folded leaflet approximate towards one another, then 

 touch and adhere ; but, at a certain period, dissection has 

 exhibited to him the cellular mass of the ovary excavated by 

 a slight fissure, which continues to enlarge, without any 

 manifest external appearance. A third category, and that 

 includes the greater part of the Malvaceae, exhibits the carpels 

 not in constant relation with the quinary number of the 

 other parts of the flower ; but they form a perfect circle, are 

 not grouped into five systems, and frequently their entire 

 number is no multiple of five. However, M. Duchartre is led 

 to believe that the same symmetry occurs here as in the pre- 

 ceding case. The ovaries and styles are developed in the 

 same manner, with this difference, that all the styles are 

 united inferiorly into a single cylinder. 



Finally, a fourth category seems to belong to the first by 

 the quinary number of the carpels ; but here we observe ten 



