254 Dr Dickson on Diplosttmonous Flowers, dtc. 



lish. As to this, 1 would state that, although the accessory 

 or non-accessory nature of the younger stamens when in- 

 ternal may be very difficult to determine in some cases, 

 where the carpels, from multiplication, or reduction in num- 

 ber, fail to afford any indications, yet, when we consider 

 the relations of the parts in the flowers of Coriaria, Agro- 

 stemma, Cerastium, &c., where the gynoecium is isomerous 

 with the staminal whorls, and the carpels alternate with 

 the younger stamens, we can have no doubt as to such 

 flowers being truly diplostemonous, and therefore morpho- 

 logically distinct from those of Geranium, Erica, Mala- 

 chium, &c., where the younger stamens, in being external 

 to the older, occupy a position irreconcilable with the idea 

 of their forming a genuine whorl, and where the carpels 

 alternate with the older stamens. 



In the last place, we may consider certain anomalous and 

 somewhat perplexing pseudo-diplostemonous forms, occur- 

 ring in the Sapindacese and Polygalaceee. 



I have constructed, in accordance with Payer's observa- 

 tions, diagrams of the flowers of Polygala, Kcelreuteria, and 

 Cardiospermum. In these plants, the outer and younger 

 (accessory) whorl of stamens is incomplete. 



In Polygala (PI. III. fig. 9), the lower or anterior stamen of 

 the primary, and the upper or posterior stamen of the acces- 

 sory, whorl are absent.* It is worthy of remark, that in this 

 plant, while the disappearance of the anterior primary sta- 

 men appears to be the direct cause of a solution of continuity 

 of the staminal tube, the disappearance of the posterior 

 accessory stamen is unaccompanied by any such solution. 



In Koelreuteria (PI. III. fig. 11), the primary staminal 

 whorl is complete, while the accessory whorl is reduced to 

 three stamens alternate with sepals 1 and 4, 4 and 2, 6 and S.f 

 In Pavia (jEscuIus), Payer has described a similar arrange- 

 YXient, — only, the accessory stamen between sepals 1 and 4 

 (and sometimes also that between sepals 4 and 2) is absent, 

 and that between sepals 5 and 3 is occasionally resolved 

 into two, then resembling those in Peganum and Monsonia.X 



Now, it may seem an objection to my doctrine of acces - 

 sory stamens, that, in such plants, it requires us to admit 



^ Organogenie, p. 140. f Ibid, pp. 150-1. X I^^if^* P- 130. 



