252 IV' CONFKIIKM'.K I.NTi;i!NAÏI(l,\AI.K llK r.KNKTlOlK. 



Ihird ypar Ihe protein content was only \ I .ill por cent, or a gain of. 51 percent, 

 in four years. However, tiie nexl year tlie prolein content jumped to \^1.7t'2 per 

 cent., or a gain of. Sfi per cent, in one year. New referring bark to table 7> it 

 is seen Ihat it is just at tlie end of the third years sélection lliat Ihere is a greal 

 réduction in tiie number of lines represenled. Thus only six of the original 

 twenty four lines ar<> represenled in tlie fifth génération. Also it is just hère 

 tliat line No. l'21 begins to show its prepotency. The next year (1901) there is 

 a very much larger increase in the average per cent, of protein, viz : 1.80 per 

 cent, for the year. Il has already been slaled thaï climatic conditions caused 

 an exceptionally high per cent of protein in Ihis year. The nexl year (1902) 

 Ihe prolein content fell back again and Ihen rose wilh some irregularities. 



The most inleresting feature hère, however, is thaï so long as a compara- 

 tivelj' large number of lines are represenled in the pedigree the average prolein 

 content remains almost stalionary. When the number of Ihe lines begins In 

 be diminished the average prolein content begins lo be increased much more 

 rapidly. This may of course, be a coincidence but il is a point which needs 

 more analylical study. 



Before discussing this question furlher it will be weli to examine very 

 liriefly the remaining throe experimenls. In no one of thèse are the resulls so 

 siriking as in the high prolein plols. However, in a gênerai way Ihey point lo 

 Ihe same conclusions. 



Lmn Protein l>hit<<. 



Table 4 shows Ihe varions lines of the low prolein plots arranged in Ihe 

 same way as the high protein pedigrees. Il will not be necessary lo enter inlo 

 Ihe détails of thèse laides. They présent many inleresting Ihings to anyone 

 who cares lo study Ihem. Only a few of the more important conclusions are 

 pointed oui hère. 



In Ihe firsl place, il is seen thaï starling with twelve original lines there are 

 only Iwo of thèse represenled in the last génération. Only three lines persist 

 beyond the sixih génération. Of Ihe Iwo lines which hâve progcny in the last 

 génération, one is much superior to the olher in its ability, to produce earslow 

 in protein. Tliis has not been true of ils whole hislory. In the firsl seven 

 years, line No. 106 produced only a moderate number of ears low enough to l)e 

 selected. After Ihe seventh génération, iiowever. this l'acully appears lo hâve 

 increased very greatly. 



Comparing this table with column 7> of table 1, il is seen thaï il was not 

 unlil lUOi, or thesevenlh génération, thalany very greal or permanent decrease 

 in the per cent, of prolein was broughl aboul. Hère again the mosi marked 

 improvemeni in the direction soughl occvn-s only after the number of lines has 

 been very malerially reduced. 



The resulls of the low prolein experimenl while mA ((uile so siriking as 

 those of high prolein plot are nevertheless along Ihe snmo gênerai lines and tend 

 to confirm the conclusions drawn from il. 



///■(/// OU P/o/.s'. 



Table 5 gives the pedigrees for the high oil ]ilols. From Ihis table one may 

 briefly noie the following points : 



