Dr Dickson on Diplostemonous Flowers, tc 243 



make their appearance in centrifugal succession upon an 

 annular cushion surrounding the centre of the floral axis, 

 and in this wise : — in the first place, a circle of five stamens, 

 superposed to the sepals, makes its appearance on the upper 

 part of the annular cushion ; later, alternate with and 

 below these, a second circle of five stamens is developed ; 

 still later and lower, ten stamens appear, one on either side of 

 each of the stamens of the second circle ; lastly, a great 

 number of stamens continue the centrifugal succession till 

 the annular cushion is completely covered. ''From this mode 

 of staminal development,'' he says, " may we not conclude 

 that the androecium of Cistus is composed of only two whorls: 

 the one superposed to the sepals, in which the stamens remain 

 simple and are the more internal ; the other superposed to the 

 petals, in which the stamens are grouped in five bundles, the 

 stamens in each bundle appearing from above downwards."* 

 Here, I cannot but think, Payer has introduced an un- 

 necessary complication into the subject. His interpretation 

 involves at least one serious improbability, — that in the same 

 flower there should be both simple and compound stamens. 

 What induced him to adopt this opinion was no doubt 

 the consideration that, if the five stamens of the first degree, 

 which are superposed to the sepals, were assumed to be the 

 apices of staminal groups, the stamens of the second degree, 

 which are superposed to the petals, must occupy neutral 

 territory between these groups. He seems not to have 

 taken notice of the fact which his figure! plainly indicates, 

 that the same difficulty occurs again lower down, where 

 there are stamens (apparently of the fifth degree) super- 

 posed to the petals, and therefore also on neutral terri- 

 tor}'. The same thing appears still more strikingly in his 

 figures of Capparis (the androecium of which he has 

 recognised as being similar to that of Cistus), where it 

 will be seen that in every second or third generation 

 of stamens there are some occupying neutral territory. 

 There seems, therefore, to be no more reason for con- 

 sidering the stamens of the second degree as the apices of 

 staminal groups than for so viewing any of the other sta- 



* Op. cit. pp. 16, 17, 



t Op. cit. plate 3, fig. 13. 



