FRANK M. Si;iiKACR. — TIIK FiRSCLT OF SFl.KCTIM, FIJ arATINC VAIilAÏKXNS. ti", 



1. Of tliP hventy-four ears upod lo starl lliis exporinieni in ISIIO onK (>ii(lil 

 aro represonled hy progpiiy in llic IhinI ^enoralion; 



2. Al iJic ciiil nrilic sixili g-encr;ili(in ail l)iil lliree ol llio on'i^iiial liiirs liaxc 

 heeii (irop|)P(l ; 



.". 'l'Iiesp tliree Unes are maintainod lo tlie end of tlie experimenl. Tlic 

 luinihi'i' of selected ears in the laler g-enoration are not, howevor, oqiially divi- 

 ded l)el\veen thoso llii-ee Unes. Line No. I I lis l)y far (lie niosl |)rolilic pindncoi- 

 of ears with liigli oil renient; 



i. Again rcrlain apparenlly cnirial points in Ihe hislory of thèse lines are 

 noted. Tlius in Line No. III, ear No. 600 produced an exceptionally large 

 nnmber of ears witli a liigli oil content. Similarily in line No. 1 18, ear No. 710 

 produced fonr good ears while seven olher ears from llic same line in tlie same 

 génération faili-d lo pnidnce any whicli were good enoiigh to be selecled. 



Lnw OU Plol!<. 



As sliown in table (i, twelve ears were selecled l'or starting Ihe experimenl 

 in decreasing the oil content. From this table we noie : 



1. Thaï only tiireeol'lhe oi-iginal lines arc represenled in the scvcnlli gêne- 

 rai ion : 



2. Of thèse tliree only two, viz : IIMi and 110, are represenled in Ihe cle- 

 venth génération. Line No. lOX is mainlained np lo Ihe elevcnlh génération 

 and thcn dropped ; 



3. Of thèse two lines. No. 100 is by far the better, and contrilintcd sixtcen 

 of the twenty-four cars to the last génération ; 



4. Again il is noted that certain ears havc a niarUed tendency to produce a 

 large nuniber of ears with a low oil content. Examples of Ibis are seen in the 

 seventli génération in ears No. 722. 72(t. and 710. 



Discussions nml ('nnclusion. 



Taking info considération the resnlls of ail four experiments as displayed 

 hère it is clearly seen tbat one of Ihe most striking efl'ecfs of the sélection prac- 

 ticed bas been to reduce the nnmber of lines. Two of flie experiments (Tables 

 ô and 5) were started with twenty-four ears eadi. AI the end of tlve years the 

 high-protein corn shows progeny from only tliree of the original ears, and the 

 high-oil from fonr of the original ears. At the end of ten years the higli-pro- 

 lein corn showed progeny from only one of the original ears and the liigh-oil 

 from tliree of the original ears. So far the resuifs are in accord with the géno- 

 type conception of Johannsen as applied to non-seif-ferlilized organisms. 

 However, if nothingbuf the isolation of favoralile génotypes had laken place 

 the extrêmes of prolein production would not bave been greatly cbanged. 

 This, however, did fake place bere. In the later years we find many indivi- 

 dual ears with a per cent, of protein far beyond anyfbing which occurred in 

 the earlier years. In this respect thèse results parallel the classical case of de 

 Vries's sélection of buttercups. In this latter case the extrême was moved far 

 beyond what it was before the sélection. Unfortunately in Ihe case of de 

 Vries's work pedigree data is not available. 



Table 7 shows for each year the maximum and minimum per cent, of pro- 

 tein in the ears which were selected for planfing Ihe high-protein plots. 



