Selection 375 



types were T^Ti T2T2 tsts and tih tot-i T^T^. The plants of the 

 first strain averaged 46 cm in height; those of the second were 

 38 cm tall on the average. Both strains were homozygous and 

 therefore showed little variation. Let us simplify our problem 

 further, however, by assuming that there is no variation due to 

 the environment. If plants of these two strains are crossed, the 

 Fi would all be 42 cm tall and would all have the genotype 

 Tih ^2^2 ^3^3. When various Fi plants are selfed, F2 genera- 

 tions would be produced which in theory should average 42 cm 

 and should range from 54 cm at the one extreme to 30 at the 

 other. Let us assume, however, that because of lack of space 

 or other limitation our Fo was not sufficiently large to include 

 plants as tall or as short as is theoretically possible. We did, 

 however, have some 50-cm and some 34-cm plants, both of 

 which types were more extreme than either parent. Having 

 obtained this F2, we are not satisfied, but would like to build up 

 as tall a race and also as dwarf a race as we can. 



In order to obtain a tall strain, we self-fertilize some of the 

 50-cm plants. They are all T^h ToTo T^Ts, T^T^ T-.to^ TsTs, or 

 TiTi T2T2 Tsts. All three types will segregate into ^ 54 cm 

 tall : 3'2 50 cm tall : ^ 46 cm tall. So far, we have built up 

 our strain from 46 to 54 cm in three generations. Pleased with 

 our success, we then select the 54-cm plants and self-fertilize 

 those. To our dismay, however, their offspring are all only 54 

 cm tall. We then self-fertilize these, but again our offspring are 

 no taller than 54 cm. After several more attempts we are forced 

 to conclude that we cannot increase our height beyond 54 cm. 



We then turn to our smaller plants to attempt to create a 

 dwarf strain. Our smallest F2 plant is 24 cm tall and has one 

 of the following genotypes: ^1^1 ^2^2 ^^3^3, ^1^1 ^2^2 ^3^3, or Titi ^2^2 

 ^3^3. When one of these is selfed, we get F3 plants that are 38 

 cm tall (homozygous for one of the contributing genes), those 

 that are 34 cm tall (heterozygous for one locus), and some that 

 are 30 cm tall (^1^1 ^2^2 ^3^3). The last plants are shorte;r than 

 anything we encountered in any previous generation. When we 

 self-fertilize them, however, we find that they have apparently 

 reached their minimum size, for they produce nothing smaller 

 in any subsequent generation. 



In attempting to create a giant strain and a dwarf strain, we 

 systematically chose the tallest plants in each generation as the 



