12 MEMOIRS FROM THE DEPARTMENT OF BOTANY OF COLUMBIA UNIVERSITY 



stamens, as, generally in Potentilla pentandm, Horkelia Gordonii, H. ShocMeyi, H. Baileyi, 

 H. lycopodioides, Gomarella muUifoliolata (Fig. 12), C. sabulosa and Sibbaldia procumbens. 

 Chamaerhodos is an exception to the rule (Fig. 11) in which the androecium consists of 

 the 5 antipetalous stamens. Stellariopsis santolmoides (Fig. 15) has always only 15 

 stamens and Potentilla 3Ionspelie7isis sometimes that number. In both the antipetalous 

 stamens are then lacking. In the former species they are represented by a thickening of 

 the tissue at the base of the parapetalous stamens. 



There is, however, within the tribe Potentilleae also another arrangement of stamens 

 widely different from the one described above. (See Figs. 1 and 13.) My attention was 

 called to it especially in 1896 while studying a new species, P. convallaria, nearly re- 

 related to P. arguta and P. glandidosa. It may be described as follows : The margin of 

 the disk is very thick ; this collar-like rim ( Fig. 1 ) is pentagonal in outline, with one 

 petal (p) fastened at each corner; in cross-section the collar is nearly semicircular and 

 bears on its upper surface 20 to 30 stamens ; these are not arranged in definite series, 

 but form concave arches (Fig. 13, fs) between the bases of the petals ; the number of sta- 

 mens in each is very variable, being 4, 5 or 6 in each arch in the same species or even 

 in the same flower. The stamens on each side of the bases of the petals are the oldest, 

 then the ones next to them, etc. Those opposite the sepals (as. Fig. 13) are, therefore, 

 the youngest, a condition entirely contrary to that found in the other Potentilleae. It 

 was this peculiar structure, although not well descriljed by him, that led Bigelow to 

 make PotentiUa arguta the type of a new genus, Bootia. One of our recent botanists, in 

 a paper published a few years since, slightly ridiculed Bigelow for so doing, saying : 

 " What a genus ! " It must be admitted that the above mentioned peculiarity alone 

 would scarcely warrant the exclusion of the species in question from Potentilla. There 

 are, however, other characters, more important, which I shall discuss later, that in my 

 opinion fully warrant such an exclusion. 



Alexander Dickson'^ gives an interesting discussion of a similar arrangement of the 

 stamens in P. fratkosa, in which the festoons generally contain 5 stamens ; he explains 

 the arrangement in the following words : " As I am unable to conceive of any possible 

 explanation of such a festooned arrangement of stamens, unless we view the androecium 

 here as consisting of five compound and confluent stamens, the terminal lobe of each 

 such stamen being developed as a petal, so-called. When there are five stamens in the 

 festoon, the central stamen must be regarded as an interstaminate lobe, analogous to the 

 interpetiolar stipules — to the intersepaline lobes in some species of Campamila, in Nemo- 

 phila, and in Potentilla itself ( the so-called epicalyx ), or the interpetaline lobe of the 

 corolla of Soldanella." 



'Seemann, Jonrn. Bot. 4: 273-281. 



