328 NATURAL SCIENCE. May. 1897. 



curve, in strong contrast to the last. The maximum is at 25 



and the calculated mean ^S'oy^. There is, however, a secondary 



maximum at 32, and a tertiary one at 38. If from the diagram of the 



siamens (Fig. 3) there are deducted those dots whose numbers have, in 



Fig. 4, made the curve imperfect, that is to say, the last or highest 21 



of the 25 carpels column, the last 2 of the 29 column, and the last 10 



of the 32 column, then the stamen curve remains practically the same. 



The deductions are fairly distributed on each side of the maximum 



and not localized at all, showing that here also the whorls vary 



independently. 



Analysis of Certain Specimens, 



Mean 50 78 443 293 



There being so few members of the calyx and corolla, I would 

 suggest that, if there were more, there would be greater variation ; 

 that is, given a greater range of variation, the chances would be 

 against such a high maximum occurring then as now. 



As regards the irregularities in the curves of the stamens and 

 carpels, is it probable, that, if a much greater number of specimens 

 (say 10,000) were examined, these curves would become more regular ; 

 or should these figures be taken as fairly representative ? 



John H. Pledge. 

 115, Richmond Road, Dalston, N.E. 



