THE ANDR(ECIUM OF MENTZELIA. 215 
begin to appear considerably before the staminal evolution is completed ; 
but, from the minute size of the staanual mammillse here, it is difficult 
to determine the period very precisely. 
In Geum^ Payer describes the stamens as "very numerous and dis» 
posed in whorls which alternate with each other, and appear succes- 
sively from above downwards, on the inner smface of the receptacular 
cup, which is pretty deep. Each whorl is composed of ten stamens. 
In the whorl which appears first, and which in consequence has the 
highest place on the inner surface of the receptacular cup, the ten sta- 
mens are grouped in pairs in such a manner that there is one to right 
and left of each petal."* (Fig. 10.) 
In Hosa, Payer describes the andi'cecium as identical in its develop- 
ment with that in Geuniy consisting of numerous alternating whorls of 
stamens appearing ceutripetally, each composed often stamens. f This 
statement, however, seems scarcely to agree with his figure of Rosa 
alpina ('Organogenic,' pi. 100, fig. 26), where the stamens of the 
second degi'ee are superposed in pairs to the petals and singly to the 
sepals, making, in all, fifteen stamens in that range ; and where there 
is apparently, moreover, a similar number of stamens in each of the 
succeeding ranges, I have no doubt that Payer's figure is quite cor- 
rect, as the same arrangement of the stamens of the second degree 
appears to be constant in Rosa sj)mosissma. The diagram (Pig. 9) 
which I have given, may therefore be relied upon as an accurate repre- 
sentation of the staminal arrangement in Rom. 
In Fragaria^ Spirtea^ and Cofoneaste7\ Payer has shown that there 
is a first range of ten stamens, disposed, as in Geum and Rom, so that 
there is one on either side of each of the five petals ; then a second 
range of Jive stamens superposed to the sepals ; and lastly, a third 
range also oifive stamens superposed to the petals (Pig. 11). 
The andrcecium in Geum and Rosa is that which most closely re- 
sembles, in its development, that in the Mentzeliea. In looking at the 
yoimg flowers when the andrcecium first appears, it is striking to ob- 
serve the great analogy which exists between the primary stamen in 
Mentzelia accompanied by two secondary ones, and the petal in Geum 
or Rosa accompanied by two stamens of the first range. In Bartonia 
the resemblance is still further heightened by the primary stamens 
being actually developed as petals. If we recall what I have mentioned 
* < Organogenies pp. 501-2. t Ibid. pp. 502-3. 
