do with the formation of distinct species. The several islands of the Galapagos, each with its distinct species of land tortoises, lizards, etc., and the numerous species and varieties of the Achatinellidæ in the several valleys of the Sandwich Islands, are striking illustrations of this fact. If all the Galapagos Islands had been merged into one, and if the Sandwich Islands permitted easy migration of land snails from one valley to the next, it is very doubtful if there would be as many distinct, but closely allied, species on these islands as are now found. Dr. D. S. Jordan remarks:

In regions broken by barriers which isolate groups of individuals we find a great number of related species, though in most cases the same region contains a smaller number of genera and families. In other words, new species will be formed conditioned by isolation, though these same barriers may shut out altogether forms of life which would invade the open district.

Thus, through the eastern United States, unbroken by important barriers, there is but one true species of chipmunk, Tamias striatus, and one species of shorelark (Eremophila alpestris). In California, broken by many barriers of various sorts, there are a dozen or more different kinds of chipmunks, species and subspecies. But in the eastern states the fauna at large is much greater, because many types of birds and other animals have found entrance there, forms which are excluded from California by the barriers which surround that region.

Dr. Jordan summarizes the facts of distribution of related species by the following statement which has since been designated as Jordan's law:

Given any species in any region, the nearest related species is not likely to be found in the same region, nor in a remote region, but in a neighboring district separated from the first by a barrier of some sort, or at least by a belt of country which gives the effect of a barrier.

We sometimes find closely related species of plants in the same locality, especially if they are self-pollinating. In such cases the species are sexually isolated. Related forms may breed

at different times and thereby become prevented from intercrossing even if they live in the same area. Species of the same genus are often found side by side, but if they are not selffertilizing they are generally so distinct that they cannot cross, or else they produce progeny which are infertile.

Where closely related groups are not separated by barriers they frequently intergrade where their areas come into contact. Such intergrading groups are generally called subspecies, whereas groups that are clearly distinct, even though closely similar, are ranked as different species. It not infrequently happens that forms from different localities which were at first described as distinct species have been subsequently connected together by intergrading forms discovered in the intervening regions. An instructive instance of this is afforded by the American song sparrow, Melospiza melodia. Mr. Ridgway in his Birds of Middle and North America describes twenty subspecies of this species, each occupying exclusively a particular range. The typical subspecies is widely distributed over the Atlantic seacoast, the Mississippi valley, and the Great Plains of the western states. Farther west it is replaced by a number of subspecies of smaller areas of distribution. "In California, practically each distinct drainage area has its own peculiar form, one being strictly limited to the salt marshes fringing San Francisco Bay." Other subspecies range southward into Mexico, and northward into Alaska and the Aleutian Islands. Several of these were originally described as distinct species, but they were all merged into one after transitional forms were found in the intervening localities. The different subspecies of a species generally cluster in a common area, each subspecies being confined exclusively to its own subdivision of the continuous range. In the distribution of the wood rat Neotoma fuscipes, one subspecies, Neotoma fuscipes fuscipes, occupies an area in California along the coast north of San Francisco Bay. Around the upper end of the Sacramento Valley it intergrades with N. fuscipes streatori which ranges southward along the western slope of the Sierras to Kern County, where it intergrades with N. fuscipes simplex.

The latter in turn intergrades with N. fuscipes macrotis on its western boundary, and macrotis is connected by intermediate forms with N. fuscipes mohavensis to the north and east.

An analogous grouping of subspecies is repeated hundreds of times over among North American

birds and mammals. When Darwin wrote, few connecting links between existing species were known, and much was made of this circumstance by the critics of his theory. But since more attention has been paid to geographical variability numerous intermediate forms have been brought to light. Dr. C. Hart Merriam has stated that "among North American mammals and birds . . . there are more than 1000 species and subspecies, connected with other forms by intergrades." Had Darwin known that connecting links are as common as we now know them to be, one of the chief difficulties that worried him would have been removed.

If we assume that isolation has been an important factor in the formation of species how shall we conceive that it acts? There has been more or less controversy over this question, some writers, among whom is Mr. Gulick, maintaining that isolation per se is a cause of divergence independently of natural selection, while others, led by Wallace, contend

NEOTOMA FUSCIPES.

F. STREATORI.

FANNECTENS.

F. SIMPLEX.

"

F. MOHAVENSIS

F. MACROTIS.

FIG. 221-Distribution of the subspecies of the wood rat Neotoma fuscipes. (After Goldman.)

that isolated forms diverge, because natural selection has adapted them to live in slightly different environments. Whatever the influence of natural selection may be, it is doubtless true that if a

species were separated into different regions the members would become somewhat different as the result of mere isolation. Iso

FIG. 222-Mutants of beetles of the genus Leptinotarsa: 1, L. undecimlineata; 2, a mutant (angustovittata) of the preceding; 3-5, L. multitaniata (4), and its mutants rubicunda (3), and melanothorax (5); 6-9, L. decemlineata (8), and its mutants defecto punctata (6), pallida (7), and tortuosa (9); 10, 11, larvæ of L. multitaniata (11), and of rubicunda (10). (After Tower.)

lated groups, especially if small, tend to become homozygous through inbreeding, and two groups, even with identically the

same inheritance to start with, might happen to become homozygous for different characters.

Most systematists nowadays do not regard all the trifling characteristics which distinguish one species from another as necessarily useful. Many characteristics may arise as small mutations and be preserved because they are not sufficiently injurious to be eliminated by natural selection. It is quite probable that in different isolated groups, different mutations would make their appearance, and this in time might occasion a considerable amount of divergence.

Contrasted with the Darwinian doctrine of the slow formation of species, is the mutation theory which regards species as the product of sudden and stable mutations. The most prominent advocate of this view is the Dutch botanist, Hugo De Vries. Ordinary, or Linnæan, species, according to De Vries, are usually aggregates of smaller subdivisions called elementary species. Elementary species are considered by De Vries to be the product of a single mutation or, more rarely, a very few mutations. In the species Draba verna, for instance, more than 200 such forms have been described, a great many of them growing side by side. They are very similar but yet distinct, and the chief reason for their remaining distinct is the fact that they are self-fertilized.

It is with these smaller elementary species that the mutation theory of De Vries is mainly concerned. Originating suddenly, "they are not connected with the parent species by intermediates, and have no period of slow development before they reach the full display of their characters. They do not always arise, but only from time to time. A parent species may produce its offspring separately at intervals, or in larger numbers during distinct mutating periods. After this production, the old species is still the same as it was before, and it subsists in the midst of its children" (De Vries).

It must not be inferred that mutations are necessarily extensive changes; most of them are very small modifications which it takes an experienced eye to detect. Neither should it be in