The work of the Danish botanist, Johannsen, has given us the most carefully analyzed case of selection that has ever been obtained. There are, moreover, special reasons why the material that he used is better suited to give definite information than any other so far studied. Johannsen worked with the common bean, weighing the seeds or else measuring them. These beans if taken from many plants at random give the typical curve of probability (fig. 74). The plant multiplies by self-fertilization. Taking advantage of this fact Johannsen kept the seeds of each plant separate from the others, and raised from them a new generation. When curves were made from these new groups it was found that some of them had different modes from that of the original general population (fig. 81 A-E, bottom group). They are shown in the upper groups (A, B, C, D, E). But do not understand me to say that the offspring of each bean gave a different mode. [Illustration: FIG. 81. Pure lines of beans. The lower figure gives the general population, the other figures give the pure lines within the population. (After Johannsen.)] On the contrary, some of the lines would be the same. The result means that the general population is made up of definite kinds of individuals that may have been sorted out. That his conclusion is correct is shown by rearing a new generation from any plant or indeed from several plants of any one of these lines. Each line repeats the same modal class. There is no further breaking up into groups. Within the line it does not matter at all whether one chooses a big bean or a little one--they will give the same result. In a word, the germ plasm in each of these lines is pure, or homozygous, as we say. The differences that we find between the weights (or sizes) of the individual beans are due to external conditions to which they have been subjected. In a word, Johannsen's work shows that the frequency distribution of a pure line is due to factors that are extrinsic to the germ plasm. It does not matter then which individuals in a pure line are used to breed from, for they all carry the same germ plasm. We can now understand more clearly how selection acting on a general population brings about results in the direction of selection. An individual is picked out from the population in order to get a particular kind of germ plasm. Although the different classes of individuals may overlap, so that one can not always judge an individual from its appearance, nevertheless on the whole chance favors the picking out of the kind of germ plasm sought.