EVOLUTION 391 



of the Chilterns ; but this does not prove it to be, even if 

 one were found next year, an abnormal variation. Poppies 

 with a great number of stigmatic bands are very infrequent, 

 and very many seasons' poppies might have to be ex- 

 amined before such a poppy was found. We could reach 

 such a poppy probably far sooner by sowing the seed of 

 poppies with capsules of a high number of bands, and 

 again selecting from these and resowing. We should thus 

 shift our type towards many banded capsules, and the 

 deviations in the individual, which we have seen range on 

 either side of the type, would soon carry us beyond 20 

 bands,^ and produce a poppy rarely if ever found in the 

 field. Naturalists, whether botanists or zoologists, are in 

 the habit of preserving very infrequent normal variations ; 

 they get collected into museums, and labelled as abnormal 

 variations, because the connecting links have either not been 

 sought for, or if sought for are too infrequent to be easily 

 found ; thus the prevalence of abnormal variations becomes 

 a rather widespread idea. Further, another class of varia- 

 tions due to accidents of growth, to injury by insects or by 

 environment, should be excluded from abnormal variation. 

 The occasional but not very frequent distortion of the 

 stigmatic bands of the poppy capsule may often, I think, 

 be attributed to such accidents of growth. In the 

 examination of some 10,000 poppy capsules, wild and 

 garden, I have found, perhaps, only one fairly abnormal 

 variation, a case in which two capsules were attached at 

 their bases to one stalk. On the other hand, in examining 

 200 plants of Nigclla Hispanica with a view to counting 

 the segmentation of the seed capsules, I found fairly 

 frequent cases of abnormal variation, double and even 

 treble capsules in every state up to complete fusion 

 — capsules with tips of other capsules growing out of 

 them, capsules without tips, etc., in very considerable 

 variety occurred, thus rendering the counting of the 



1 By selection in this manner Professor de Vries soon obtained a buttercup 

 with more petals than had ever been observed in the field. Such a buttercup 

 may be looked upon as a new race, for its infrequency may exclude its 

 appearance in nature. 



