STRUCTURE.] MORPHOLOGY OF MALLOW-WORTS. 73 



tion of the summit of their curve a constriction which ends 

 by forming a complete solution of continuity ; this, extend- 

 ing from above downwards, also divides the filament which 

 was at first simple, into two corresponding to the anthers 

 thus formed. This is a true duplication. 



This term would apply with less accuracy to the anterior 

 formations, from which the multiplication of the stamens has 

 resulted ; for we may say, that at each of these changes they 

 have doubled rather than multiplied. Be this as it may, we 

 have clearly five groups of organs alternating with the five 

 leaflets of the calyx, each comprising a petal and several 

 stamens, supported upon a base which is common and simul- 

 taneously developed. This is the whorl which is within and 

 alternate to the calyx, and which is ordinarily called the 

 corolla, with this difference, that here each petal is replaced 

 by a group or bundles of organs. 



One of the reporters on M. Duchartre's observations is of 

 opinion, that in those flowers which have stamens double in 

 number to the petals, whenever the stamens of the external 

 row are opposed to the petals (and this is most frequently the 

 case) they do not constitute a distinct whorl, but form a part 

 of that of the corolla. The development of the flower of the 

 Mallow-worts supports this opinion, exhibiting each of the 

 petals, opposed, not to a stamen, but to an entire bundle. It 

 is added, that such appears to be the most common symmetry 

 in polyadelphous polypetalous flowers, as is seen in so many 

 Myrtacese, Hypericacese, &c., where the bundles, which are 

 perfectly distinct, are opposite to the petals. 



What has become of the normal whorl of the stamens, 

 that which should alternate with the petals ? M. Duchartre 

 discovers this in the five terminal lobes of the staminal tube, 

 situated upon a plane anterior to that of the filaments, alter- 

 nating with their five groups, lobes which we observe in 

 many of the Malvaceae, although they are barely perceptible, 

 and even are entirely wanting in many others. MM. Dunal 

 and Moquin-Tandon recognised them, and considered them 

 as the border of a five-lobed disc. But the nature of the 

 disc is far from rigorously defined, and in many cases this 

 term exactly applies to abortive whorls, as may be seen in 



