Duchesne introduced the word to designate those cases in which species revert spontaneously to what are presumably long-lost characters. Atavism and reversion are used by most authors in the same sense.

I. The term atavism is employed to express the reappearance of characters, physical or psychical, in the individual, or in the race, which are supposed to have been possessed at one time by remote ancestors. Very often these suddenly reappearing characters are of the monstrous type, e.g. the three-toed horse. The appearance of such a monster is looked upon as a harking back to Tertiary times, when the ancestor of the modern horse possessed three toes. The threetoed condition of the monstrous horse is spoken of as atavistic. The employment of the term in connection with teratology is often abused; for many cases of so-called atavistic monstrosities have little to do with lost characters, e.g. the possession by man of supernumerary fingers and toes.

II. Atavism is also used to express the tendency to revert to one of the parent varieties or species in the case of a hybrid; this is the atavism of breeders. Crossed breeds of sheep, for example, show a constant tendency to reversion to either one of the original breeds from which the cross was formed. De Vries distinguishes this kind of atavism as vicinism (Lat. vicinus, neighbour), and says that it "indicates the sporting of a variety under the influence of others in the vicinity."

III. Atavism is employed by a certain school of evolutionistic psychologists to express traits in the individual, especially the child, that are assumed to be, as it were, reminiscences of past conditions of the human race or its progenitors. A child by its untruthfulness simply gives expression to a state that long since was normal to mankind. Also in the child's fondness for splashing about in water is exhibited a recrudescence of a habit that was quite natural to its aquatic ancestors; this latter is called water-atavism. Many such atavisms are distinguished, but it hardly needs to be said that they are in many instances highly fantastic. Atavism is commonly supposed to be a proof of the evolutlon of plants and animals, including man. Characters that were normal to some remote ancestor after having latent for thousands of generations suddenly reappear, thus give a clue to those sources to which the present living forms are to be traced back. That a character may lie dormant for several generations and then reappear, admits of no doubt; even ordinary observation tell us that a grandchild may resemble its grandparent more than either of its immediate parents. But the sudden appearance of a tailed man, for instance, cannot be said to prove the descent of man from tailed forms. Granting that man has descended from such ancestors, the phenomenon is more intelligible than it would be were no such connection admitted. But the proving force of atavism is not direct, because teratological phenomena are so difficult to interpret, and admit of several explanations. Darwin, pointing to the large canine teeth possessed by some men as a case of atavism, remarks: "He who rejects with scorn the belief that the shape of his canines, and their occasional great development in other men, are due to our early forefathers having been provided with these formidable weapons, will probably reveal, by sneering, the line of his own descent".

Atavism is appealed to by modern criminologists to explain certain moral abberations, that are looked upon as having been at one time normal to the race. Accepting the doctrine that man has by low progress, come up to his present civilized state from brute conditions, all that is brutish in the conduct of criminals (also of the insane), is explained by atavism. According to this theory degeneracy is a case of atavism. The explanation offered for the sudden reappearnace of remote ancestral characters is so intimately connected with the whole system of heredity that it is impossible to do more than indicate that most writers on heredity seek this explanation in the transmission from generation to generation of unmodified heredity-bearing parts, gemmules (Darwin); pangenes (De Vries); determinants (Weisman). (See Heredity|HEREDITY.)

Jos. C. Herrick