162 HAROLD H. SMITH 



in Smith and Bacon (1941). The fourth was received originally from the 

 director of the Tabak-Forschungsinstitut, Baden, Germany, under the name 

 of texana, a designation which we have retained. It is a small, early-maturing 

 type. The four parental varieties were of highly inbred stocks maintained by 

 the Division of Tobacco, Medicinal and Special Crops of the United States 

 Department of Agriculture. The earlier part of the breeding program was 

 carried out while the writer was associated with this organization. 



The advanced selection, designated Al, used in these experiments has a 

 complex genetic history of crossing, backcrossing, and inbreeding. This can 

 be briefly summarized by stating that its ultimate composition was, on an 

 average, 60 per cent 34753, 22 per cent Olson, 12 per cent tall type, and 6 

 per cent texana. About 82 per cent of the Al genotype was, on chance alone, 

 contributed by the two most vigorous parents, 34753 and Olson 68. This 

 calculation does not take into account any differential effect of selection on 

 changing the frequency of genes introduced from diverse parental origins. 

 Observation of the Al phenotype led us to believe that selection had further 

 increased the proportion of genes from the two most vigorous parents. 



In 1947 the four parents, the six possible F/s, the three double crosses, 

 and the F4 generation (preceded by three generations of inbreeding) of line 

 Al were grown in a randomized complete block design with fifteen plants in 

 each plot and replicated six times. In 1949 the two most vigorous varieties 

 (Olson 68 and 34753), the Fi, F2, backcrosses of the Fi to each of its parents, 

 and the Fe generation of line Al were grown in a randomized complete block 

 design with twenty plants in each plot and replicated eight times. 



Measurements were made on plant height, number of leaves or nodes, and 

 length of the largest leaf. In addition, data were taken on the width of the 

 largest leaf, number of days from planting to appearance of the first flower, 

 and total green weight of individual plants. 



Typical plants of Olson 68, 34753, the Fi between these two varieties, and 

 selection Al are illustrated in Figure 10.1. 



EXPERIMENTAL RESULTS 



Data obtained from the 1947 and 1949 plantings are summarized in 

 Tables 10.1 and 10.2, respectively. 



Phenotype-Genotype Relations 



Preceding further biometrical analysis of the data, tests for evidence of 

 differential environmental effects and genetic interactions were made. For 

 the former, the relation between genotype mean and non-heritable variabil- 

 ity was determined by comparing means and variances of the parents and Fj 

 (1949 data. Table 10.2). For the characters plant height and leaf length, the 

 variances were unrelated to the means and the parental variances were not 

 significantly different^from each other. For node number, however, the 



