42(; 



TiiK JoiKNAi. oi" I ll■■.Kl■:l)lT^■ 



plctcly masked by fluctuations. In 

 crosses in\'olvinj^ such characters as 

 height, even the rca])]jearancc of the 

 characters in their orijijinal form can 

 not be verified, and the only evidence 

 that we are still dealing with Mendelian 

 phenomena hes in the fact that the 

 blend of the first cross docs not persist 

 and the extremes represented by the 

 original parents of the cross are attained 

 or even exceeded in the subsequent 

 progeny of the cross. It seems not 

 illogical, however, to assume that the 

 differences, though very minute, are 

 still inherited in the same discontini:ous 

 manner. That wc have passed the limit 

 of analysis is certainly no reason to 

 assume that we have passed outside the 

 realm of law. 



It has been shown, however, that the 

 Mendelian behavior of minute differ- 

 ences would ]jroduce results in accord 

 with continuous evolution and Gallon's 

 law of inheritance. The possibility of 

 reconciling Mendelian inheritance with 

 Galton's laws was earlv recognized bv 

 Yule.' 



While the inheritance of blended 

 characters, such as height, may be 

 brought into accord with Mendelian 

 inheritance by the assum]jtion of a 

 sufficiently large number of factors, 

 such inheritance can h)e held to be dis- 

 continuous onl\' in the sense that curves 

 are held to consist of a series of straight 

 lines. An array of variations ma}' fit 

 the normal curve of continuous varia- 

 tion and at the same time comply with 

 the Mendelian laws of discontinuous 

 inheritance, jjrovidcd only that the 

 differences between the units are suffi- 

 ciently minute. 



To return now to characters which 

 are inherited in a consjjicuously discon- 

 tinuous manner, is it not possible that 

 here the individual stages in the dc- 

 N'elojjment of the character arc also 

 numenjus and extremely small? There 

 are still many gross characters that 

 stand as simjjle Mendelian units, but 

 few. if any. of these occur in jjlants or 

 animals that have been subjected to 



extensive investigation. There is now 

 such a large number of characters which 

 at first behaved as units, but which 

 have since been broken up by crossing 

 with suitably selected material, that it 

 seems not vmreasonable to belie\'e that 

 the remaining cases await only the dis- 

 covery of the right strains with which to 

 hybridize them to bring about corre- 

 sponding results. Even if the strains 

 necessary to subdivide the characters 

 no longer exist, wc may still look upon 

 the residue of "unit characters" as 

 composed of an indefinite number of 

 "factors." which arose independently 

 but are now inseparably associated. 



One of the best examples of a single 

 Mendelian character is the horny endo- 

 sperm of the seeds of certain t>-pes of 

 maize, as contrasted with the wrinkled 

 endosperm of the sweet varieties. The 

 simple Mendelian ratio that follows the 

 crossing of the sweet x horny \-arieties 

 has been demonstrated with greater 

 exactness than for any other character 

 ]3air. Taking only the results of Lock. 

 East and Hayes, and Halsted. the 

 number of individuals classified reaches 

 nearh- 100,000. With numbers of this 

 magnitude, deviations from the exj^ccted 

 3 to 1 ratio as small as one-half of one 

 percent, could be detected with assur- 

 ance. Many different varieties have 

 been crossed and always with the same 

 result: the difi"erence between wrinkled 

 and homy endosperm behaved as a 

 simple Mendelian unit. When, however, 

 a variety of maize with an cndospenn 

 that was neither hom\' nor wrinkled 

 came to be crossed with sweet \arieties, 

 a homy endos]jerm was formed wliich 

 segregated in accordance with a dihybrid 

 ratio, showing that this character, 

 which had behaved with such wonderful 

 regularity, appeared as a unit simply 

 because it had not been resoh'cd into 

 the elements or factors of which it was 

 really comjjoscd. 



That we have not >'et found a \ariet\- 

 of maize that will still further sub- 

 divide these endos])emi characters is 

 slight reason for assuming thai the 



' New Phytologisl. Vol. 1, NCs. ') and 10, l'M)2. 



A inatbcMiiatical ikiiionstration of the dose aKrctnifiit bc-lw'trn the Mcnck-lian iK-liavior of 

 minulc factors aivl ('lalton's law ha.s Ix'tTi j^ivt-n hv Brownk-c, IQIO: Proc. Rov. Soc, E<iinl)iir>^h 

 30, I'. 473. an<l 1011, 31, !'. 2.S1. 



