of the Aquilegiu vulgaris. 11 



a half (fig. 10), Ave took off its calyx in water and with very fine 

 needles. The central part then showed itself as a little sphere, 

 whereon the stamina, having just quitted their form of green 

 foliary gibbosities, now assumed that of two parallel protu- 

 berances (figs. 11 and 12). Upon these the connective is pro- 

 portionally more developed than at a later period (fig. 12) ; 

 the filament is dilated and very small ; the anther is pro- 

 portionally much larger, but it is still discoid, so that it is 

 easier to discover in it the form of the blade of a leaf. 



We were veiy curious to ascertain what the petals then were. 

 The specimen which we dissected was one which would have 

 had two rows of cornets. Now one of these rows (the exte- 

 rior one) was formed by small circular laminae, barely provided 

 with a support, but these laminae exhibited the same consti- 

 tution as the anthers of the stamens described above (fig. 14) : 

 in fact, two gibbosities, representing the anther-cells (a, c) ; 

 a veiy broad connective {h) ; and around all this a disc [d), of 

 which, moreover, the trace also exists on the anther of a sta- 

 men proceeding in its development as such. 



Here it is impossible to mistake the primitive staminal na- 

 ture of the organ which at a later period is to become a hood- 

 shaped petal, that is to say, a cornet-shaped nectary. It is 

 evident that, after the first condition of the flower, — that in 

 which all the parts were still cellular tubercles, similar to the 

 primitive condition of a leaf, — the nectarial petal, before be- 

 coming such, was similar to a stamen. This is what caused 

 us to say above, that the spur-sliaped nectaries of the Columbine 

 did not produce stamens by ascending metamorphosis, but that 

 they were, on the contrary, stamens modified by a descending 

 metamorphosis. In short, before being petals, they are rather 

 stamens, or at least anthers, than anything else. 



The row of small scales, which also become hood-shaped 

 petals, but placed higher, exhibits at this age of the flower a 

 more complete disappearance of the anther-like form. The 

 anther betrays itself there only by the dilatation of the blade 

 and its attenuated border (fig. 13 b), but there is but one com- 

 mon gibbosity in the place of the two parts which represented 

 the loculi. 



We took a bud three millimeters long (fig. 15), and stripped 

 it of its calycinal envelopes. The stamens in this were better 

 constituted, the filaments lengthened, the connective propor- 

 tionally more contracted, and the loculi very distinct (fig. 17). 

 The petaloid blades, on the contrary, were very broad, hardly 

 pedicelled ; but in the middle there is still the trace of the con- 

 nective (6, fig. 19), and on the sides two protuberances, not so 

 large, but prominent enough to discover in them the anther- 



