Mav, 1921. 



SCIENTIFIC AGRICULTURE 



197 



at the top of the table do not apply to them. They were 

 iiH^hided in this table in order to make eomparison 

 ea.sier. They are placed in relation to the third genera- 

 tion families as they would be if grown in the same year, 

 Many more tliird !j;eiieration families were prrown than 

 are shown in the table. Additional ones are recorded 

 ill another papei- (4). The jiresent table jrives only a 

 few representative ones. 



In connection with this table the following points 

 should be noted : 



(1). Fj families are found in all po.ssible positions be- 

 tween the parental extremes and coinciding with the ex- 

 tremes. This may be due in i)art to environmental in- 

 fluences but the resemblance of the F4 families to their 

 F3 parent.s in regard to position makes it appear to be 

 the result of genetic constitution. The proportion of F, 

 families at any particular location in the genei'al range 

 of variation (a point which would be of importance for 

 Meudelian analysis) cannot be stated because the 

 amount of work invohed permitted the raising of only a 

 limited number of families and these were from parents 

 scattered over the whole range of F^ variation. 



(2). There is great difference in the variability of dif- 

 ferent families. For example number 12-4-16 varies 

 from the lower extreme of the early parent to the higher 

 extreme of the late parent, while the range of variation 

 of several other families covers only five days. In a 

 Mendelian interpretation the greater variability of some 

 families is due to the segregation of factors while the 

 slight variability of others is due to the homozygous con- 

 dition of the F. parent. It is to be noted that the very 

 vai'iable F., families gave rise to some very variable F, 

 families and to some slightly variable ones, while the 

 slightly variable F, families produced F^ of the same 

 kind. If only a small number of hereditary factors are 

 involved all the intermediate F, families should show a 

 wide variability. 



(:i). Most of the F.. families, like the Fo'show no heap- 

 ing up of individuals at any particular position. 



(4). In several F., families the earliness of the early 

 jiarent' has evidently been recovered in the homozygous 

 condition. This is confirmed by the F, results. Many 

 fiftli generation families raised in 1920 also were evid- 

 ently (juite as early as the early ]iarent and were homo- 

 /.\gous as well, though extended individual records could 

 not be taken. Some families are also as late as the late 

 parental variety. The proportion of F. plants which 

 gave rise to families coinciding with the parental con- 

 ditions was very small. The great ma.iority of F , plants 

 were intermediate as were the F. families. 



(•")). In one case (12-5-20) an F.. family seems to be 

 earlier than the early parent. One might consider this 

 to be due to environmental influences even though a 

 very large number of individuals of the early parent were 

 recf)rded in order to determine the complete range of 

 variation. But the F^ families descended from it bear 

 out the conclusion that it is really genetically earlier. If 

 so. there must have been on the ordinary interpretation 

 a recombination of factors by which the later parent, 

 even though itself later, contributed a factor for earli- 

 ness which the early one lacked. 



The results for a cross involving a greater difference 

 between the par-eiits ai'c shown in Table 5. The points 

 t<i which special attention has been directed in connec- 

 tion with the jii-evious table are well illustrated here and 

 this in itself is a point to be emphasized. Certain points 

 such as the great difference in variability in different 

 families are more clearly evident. The complete list of 

 F. fHiuilies jniblished in another paper (4) shows that 



all intermediate positions as well as the parental ones 

 are occu]iied. A much larger percentage of the F., fa- 

 milies are intermediate than in the previous cross. 

 Nevertheless some families are evidently homozygous 

 for the jiareiital conditions. 



Some results in the widest cross of all, which lias been 

 carried only to the third generation, are given in Table 

 (i. Though the condition of the late parent has been 

 recovered that of the early (me has not been in the com- 

 I)aratively small amount of material raised. Evidently' 

 it is much more difficult to recover the parental condi- 

 tions in crosses where the difference is very great. There 

 is then no simi)le Mendelian segregation of one or two 

 main factors. A factorial explanation must involve the 

 assumption of many small factors. 



The three crosses described represent three degrees of 

 difference between the parents. Other crosses liave been 

 made involving similar degrees of difference and still 

 others involving other degrees. In all cases the results 

 were similar to those .iust described. 



Discussion of Factorial Conditions. 



It is evident that in all the crosses mentioned we have 

 examples of typical blending inheritance, because most 

 of the F. individuals and F, families are intermediate. 

 Applying the multijile-factor explanation of blending 

 inheritance, the great variability in the F„ indicates 

 ;\Iende]ian segregation. The great difference in var- 

 iability among different F.. families would then be due 

 to the fact that some F, parents are homozygous while 

 others are heterozygous for earliness factors. 



In this interpretation of the results obtained there 

 must be many factors for earliness and lateness. Even 

 the varieties whose means are only a few days apart must 

 differ by several factors because the great ma.iority of 

 F,, individuals are intermediate — so that simple segre- 

 gation of one or two pairs of characters is not occurring 

 — and true breeding intermediate races are quickly estab- 

 lished. 



It is to be expected that many factors in the hereditary 

 constitution will affect the rapidity of maturing. In 

 addition to factors directly concerned with earliness, 

 many factors wliose jirimary effects are on entirely dif- 

 ferent characteristics may have secondary or indirect 

 effects on earliness. Tims a factor which causes the 

 production of a short stem may because of this fact put 

 forward the ripening period. 



While the results of each individual cross can thus Tie 

 exidaiued on the multiiile-factor hyjiotliesis there appears 

 to be difficulty in exijlaining the results as a whole. As 

 already pointed out each successive pair of parental 

 varieties must differ by a certain minimum number of 

 factors where there is blending inlieritance. Summing 

 up the factor differences between the individual pairs 

 we get the number of factors by which the latest differs 

 from the earliest. And this result can be checked by a 

 direct cross between the latest and earliest. The greater 

 the number of factors the le.ss the probability of re- 

 covering the parental types in F... For example where 

 the differences between the parents involve six pairs of 

 factors the smallest number of F„ plants which must be 

 grown in order to secure the parent type will be 4f)96 

 ("Mendel's 4" where \ = fi) ; for 7 pairs of factors lfi..'^S4: 

 for 8 pairs 65,536. Now the F, and F, results show that 

 in all crosses except one the parental conditions were 

 actually recovered among at most a few hundred indi- 

 viduals of the F,. In the case of A x E the early parental 

 type has not been recovered in the comparativelv small 

 number of plants raised, though the late parental con- 

 dition was recovered. The results ol)taiued on the mul- 



