BOTANICAL SUBJECTS. 35 



III.— VARIATION. 



Examining the flowers of JVeigelia rosea one day, I found on the 

 same branch, and all over the bush, flowers with three, with four, 

 and with five petals.* These three forms were pretty common, 

 although the majority were with five petals. There was 

 another curious form, which, from its frequency, can hardly be 

 considered accidental. The tube of the corolla was split up, 

 and displayed something of the form of the upper lip of the 

 honeysuckle flower. 



Curiously enough, in each of the three regular variations, 

 the stamens corresponded with the corolla divisions, and alternated 

 with them. Wherever I have seen this JVeigelia, the same 

 variations in the flower occurred. 



We cannot well suppose that the five-petalled flower turned into 

 the four-petalled by fusion of a pair of petals into one ; nor into 

 the three petalled by fusion of two pairs into two petals, for in 

 each case all the divisions of the corolla were perfectly symmetrical, 

 and of the same size. Moreover, it is hardly conceivable that 

 fusion of petals would reduce the stamens to an equal number, and 

 alternate them with the divisions. 



No, these are congenital variations, each tcith a phyllotaxis of 

 its own. Neither fusion in the one case, nor splitting up in the 

 other, if we reverse the origin, will explain the appearance of an 

 equal number of normally formed stamens, alternating with the 

 petals. 



Jackman's purple Clematis has four and also five-petalled 

 flowers on the same plant. 



The Syringa {Philadelphus) has sometimes also five-petalled 

 flowers among the usually four-petalled ones. 



On Chironia palnstris, I found flowers with six petals and six 

 stamens, and on Chironia peduncularis I found one with seven 



* " ScEpius 3-chotome divisas Ovarium 2-loculare, 



rarissime 3-loculare." (Bentham and Hooker, Gen. Plant.) 



^ p. 1724. C 2 



