184 



NATURAL SCIENCE 



[March 



regularity which, with 500 specimens, I think would be hard to sur- 

 pass. The recorded maxima are 34 and 36 — the calculated mean, 35. 

 The range of variation is very much less in the second series 

 than in the first. 



Series I. — Maximum 65, minimum 15, 

 Series II. — Maximum 52, minimum 17. 



Fig. 6 is the curve of the two series combined. It is a very fair 

 example of a continuous variation curve. The hump in fig. 4 at 

 2 5 still, however, makes its presence felt, and destroys the symmetry 

 of otherwise an exceedingly good curve. 



FIG. 3.— VARIATION IN STAMENS OF 1000 SPECIMENS OF RANUNCULUS REPKNS (L.) 



NUMBER OF STAMENS 



On examination of the appended charts I think it will be con- 

 ceded that there is here an undoubted case of undisturbed con- 

 tinuous variation. The addition to the first, of the second series, 

 simply rounds off and does not alter the curve in general. Any 

 irregularities in the resultant curves are only apparent, and are due 

 to the comparatively small number of specimens under notice. I 

 consider, from these results, that the concluding remark of my first 

 paper must be answered in the affirmative.^ 



' I suggested {NaHrcd Science, May 1897, p. 328) that the recorded irregularities in 

 the stamen and carpel curves were only due to the small number of specimens (500) 

 then under discussion, and that they would disappear if a larger nimiber were 

 examined. 



