beget like should now be clear. It is the continuity of the germ-plasm.

When one compares a number of members of the same species, whether men, hens, dogs, pansies, eels or elephants, he finds that they differ from ore arother. It is possible to measure these differences. These "observed differences" may be due to many things. Many of them may be involved with sex, and thus accounted for; some, with age; others may be due to the influence of surroundings in early plastic years, for example, the twisted twig and the bent limb. These last are changes in the bodies of plants and animals which are acquired; they are modifications, not inborn. When from the total observed differences, these peculiarities of sex, age, and modification are subtracted, a very interesting remainder is left, which we define as inborn or germinal variations. These variations are congenital, not made. They are often distinct at birth. They are in many cases, if not always, transmissible. They form what has been called the raw material of evolution.

In late years a quantity of facts bearing upon variations has been gathered. The study and organization. of this material has shown that there are probably two quite distinct types of variation: first, fluctuating or continuous variation, producing comparatively slight divergence from the parental character; and second, stable or discontinuous variation, producing generally great divergence from the parental type."

Fluctuating or continuous variation may be illustrated as follows: from the registration of variations that occur in the height of a large number of men taken at random, it was found that there was a proportion be5 Thomson & Geddes, op. cit., pp. 115-116. Metcalf, op. cit., p. 10.

tween the frequency of a particular variation and the amount of its deviation from the mean stature of the group. Among the measurements of 2,600 men, taken at ramdom (that is, as they come and without any conscious effort to select only the tall or the short), there are 1 of 4 ft. 8 in.; and 1 of 6 ft. 8 in.; 12 of 5 ft.; and about 12 of 6 ft. 4 in.; that is, equal numbers at equal distances from the mean of 5 ft. 8 in. This illustrates that when the frequency and the magnitude of the variations are registered, they show what is called the normal curve of frequency. This can be illustrated more clearly by reference to the following table of the heights in centimeters of 1,000 ten and one-half year old American school boys."

When this material is plotted in graphical form the distribution of stature is as represented in figure 1, letting the distance of each horizontal line from the base stand for the number of boys. Now if we were to draw a smooth curve through the tops of the columns we should have a bell shaped curve of the type shown in figure 2. This illustrates graphically what we meant

7 Thorndike, E. L.-Individuality, p. 8.

by the statement that there is a relation between the frequency of a particular variation and the amount of its deviation from the mean stature of the group.

The task of registering the variations that occur in any group of creatures may at first sight seem tedious and far removed from the warm pulsations of life, but

FIGURE 1. Distribution of Stature of American Boys 10 years old.

a little experience in the measurement of such things as length of rose petals, the length of bird wings, or of starfish arms, will convince the student that biometrics may lead him into the very heart of the matter. If the registration of the dimensions of a particular character be carried on year after year in similar material, and show a consistent increase in asymmetry or skewness of the curve (asymmetry or skewness means a curve in which the hump as in the figure, is not over the middle, but nearer one end, making the slope at that end more abrupt and at the other end more gradual) this must mean that the species is moving in a definite direction as regards the particular character measured. Similarly, the persistent occurrence of a well-substantiated double-humped

A

curve-not the result of modificational effects may vividly bring home the fact that the species is dividing into two sub-species. Thus, by means of statistics, which seems the dryest of all methods, we are able to see a species being born under our very eyes.

The point we have just made shows how a species might originate by the accumulation of extremely slight

FIGURE 2. Curve of Distribution.

variations. But evidence is at hand to show that organic structure may pass with seeming abruptness from one position of equilibrium to another. Changes of considerable amount sometimes occur at a single leap. These sudden jumps or changes are called "discontinuous variations," or sometimes, "sports," and, in certain cases, "mutations." Professor Hugo de Vries has made some very interesting and important experiments and observations on the origin of species in the plant kingdom. He found that species often arise from one another by discontinuous leaps and bounds as opposed to the continuous process. He therefore believes that

8 Thomson & Geddes, op. cit., pp. 121-122.

species (forms of life having a certain similarity and related by descent) appear "all at once" by mutations.

In support of this theory, Professor De Vries takes the case of a certain evening primrose which has shown sudden and repeated leaps with a remarkable subsequent constancy. The Chelidonium majus laciniatum appeared suddenly in the year 1590 in the garden of an apothecary at Heidelberg, and has remained constant ever since. These experiments and observations have lead to another theory of descent. It is now held by a school of biologists represented by De Vries, that species have arisen by this discontinuous process, in which each new unit, forming a fresh step in the process, sharply and completely separates the new form as an independent species from that from which it sprang. The new species originates from the parent species without any visible series of transitional forms. It can perhaps be made more clear by figure 3. The figure represents by A B the direct line of descent from which the parent B has sprung. Now with the usual fluctuating or continuous variation, the offspring of B would not be likely to have the same average (of any trait) as their own parents, but an average much nearer the average of the whole group to which the parents belonged. But in the case of the "sport," whose origin we are explaining, the offspring C of B will start a new and independent line of descent. That is, the offspring D of C, will not have an average nearer that of B than C was, but will have an average nearer that of their parents C. Thus the "sport" C, has established a new group type round which there will be fluctuating or continuous variation.

Galton has illustrated the process by analogy but from another point of view. The polyhedron can be