Introduction to Mass Selection

When the American farmers of the 19th century went into their corn fields at harvest and selected the most desirable ears from the most desirable plants to provide the seed for the following spring’s planting, they were practicing mass selection. Specifically, mass selection is a breeding method where the genetic values of individual plants are estimated and then based on these estimates plants are selected to be the parents of the following generation.

The greatest attribute of mass selection is simplicity. However, it does have its limitations. One limitation is the high probability of not correctly identifying the best genotypes from the potential parents available. The ear weights of 10 plants of each of two single-cross hybrids have been measured and plotted (Figure 1a); (numbers 1-10 were randomly assigned to the ears of each hybrid). Based on mean performance, hybrid B (mean = 280 g) is superior to hybrid A (mean = 245 g). However, if the decision to select either hybrid A or B is based on the evaluation of a single plant of each hybrid, it is possible to conclude that hybrid A is better than hybrid B. In fact, if the plant of each hybrid that is evaluated is selected randomly, then the probability of making this error is approximately 1/3 because three plants of hybrid have a greater ear weight than the mean ear weight of hybrid B. Therefore, making a selection based on the performance of a single plant negatively impacts the probability of success.

Fig. 1a:  Graph of plotted ear weights with relatively low heritability for ear weight (image credit: W.K. Russell)

All plants of a single-cross hybrid are genetically identical. Therefore, all the variance for ear weight within the two hybrids shown in Figure 1a is environmental. The greater the environmental variance relative to the average difference between hybrids and B, the greater the likelihood under mass selection of choosing hybrid A. In the next example (Figure 1b), the mean performance of each hybrid is the same as in the prior example, but the environmental variability is much less. As a consequence, there is now no chance of selecting hybrid A over hybrid B, regardless of which plant of each hybrid is measured. These examples illustrate that when heritability is high (as in Figure 1b), mass selection will be effective because the best parents can be identified reliably on a single-plant basis. However, when heritability is low (as in Figure 1a), mass selection will be much less effective. 

Fig. 1b:  Graph of plotted ear weights with relatively high heritability for ear weight (image credit: W.K. Russell)