74 



Cooperative Investigations on Plauts 



It would seem accordingly that individuality lies in the plant ratlier than 

 in the fruit, i.e. that an otfspring capsule only takes after the parent capsule, 

 because that cap.sule is gi-own on the parent plant, and tliat the individuality 

 of the capsule in the plant does not influence the seed it bears*. 



The seed of the above crops wa.s all of the same kind, naniely "capsule 

 parentage 14" consisted of all the seed Contents of many 1-i-capsules on a variety 

 of Hampden plants. W. R. Macdonell in a crop of 1901 modified this experiment 

 to some extcnt by selecting the seed of 57 capsules, of which 44 gave rise to 239 

 plants with 566 capsules — a 'starveling' crop. Thus while the plants at Hampden 

 had an average of 25'2 capsules per plant and this gave a wide variety of capsules 

 on the plant, the Enfield I had only 2'3 capsules ou the plant and the Enfield II, 

 the above 239, only 2'4 capsules per plant. The individuality of the plant in 

 Enfield II was thus possibly more likely to be represented by a selection of its 

 few capsules than by a selection from the many capsules on a Hampden plant. 

 The results obtained by Macdonell were : 



The regression here is more than double that found in tho earlier series of 

 experimeiits, bot it is still considerably less than the average value of the regres- 

 sion of otfspring plant on parent plant, -3422, determined on p. 73. Macdonell 

 further correlated the capsides of the Enfield II crop with the grandparental 

 Hampden plants, — grandparental means and offspring capsules. He fouiid for 

 the weighted grandparental mean of the 566 offspring 13'11 and for the grand- 

 parental .s. D. '707. This shows that niuch selection of grandparents had taken 

 place for the nu'an anil s. D. of tho Hampden parents were 12-68 and 1-219. The 

 correlation of grandparent plant and offspring capsule was -OOGC and the regression 

 of the former un the lattcr 0190. Both of those are really insignificant. But it 

 is difficult to appreciate in this case the allowance which lias to be inade (i) for 

 selecting in the intermodiate stage not a parunt plant, but an arbitrary capsule off 

 the parent plant, and (ii) for comparing offspiing capsule with grandparental 

 plant. The result as far as it goes confirms of cour.se the previous statement, that 

 the closer degree of hereditary relatiouship is between plant and jjlant and not 

 between capsule and plant ancestry. 



That a slight Variation in the material dealt with makes a considerable 

 difference in plants, which are peculiarly susceptible to environmental and seasonal 



* This conclusion haa been further conßrmed by the oase of NijjeUa Hispanica. There was sensibly 

 no relationship in tliis plant between the se),'mentation of tlie individunl capsule on a plant und thut of 

 the capsule from which the seed which gave riaa to that plant was taken. K. P. 



