All the phenomena of correlation show increase of growth corresponding to increase, and decrease corresponding to decrease. Now, the antithesis to correlation is compensation or balancement of growth. This alleged law, as applied to species under nature, was propounded by Goethe and Geoffroy St. Hilaire. It implies that the development of any one part is attended with the reduction or starvation of some other part. Not a little diversity of opinion exists respecting the validity of this law. Darwin inclines to believe that compensation occasionally occurs, but conceives that its importance has been overestimated.
We, however, are of opinion that there is really no such law. That correlation obtains, there is not the slightest doubt. The instances of correlation are innumerable; and every one of them is a disproof of the doctrine of compensation of growth. For the law of correlation is totally incompatible with the law of economy of growth. The latter, according to the hypothesis, makes decrease correspond to increase, and increase to decrease. The former entails the reverse. Both laws, then, cannot stand. One must, of necessity, fall. One must negative the other. Unquestionably, the stronger law is correlation. This law none can invalidate. It follows thence that there is no such law as that of compensation of growth.
The reader is now naturally desirous to know how we explain away the alleged cases of economy of growth. The explanation is, that they are merely manifestations of correlation. The reduction of the given parts is consequent, not, as alleged, upon the building up of some other parts, but upon the suppression or reduction of correlated parts. Strong confirmation of this view is given by the fact that seeming compensation of growth is more observable under nature than under domestication. As development under nature is slow and occasional, we would expect to find, upon the theory of Goethe and St. Hilaire, very few instances of apparent balancement of growth. On the contrary, the instances are most numerous; which fact is strictly in accordance with our hypothesis. For where we find the conditions entailing the reduction of many parts, there must we also find the reduction of other parts, induced by correlation. These parts, then, being in close proximity with characters which neither the conditions nor correlation have affected, their suppression is naturally referred to compensation of growth. Under domestication, however, development is carried on rapidly and to a great extent. A very large number of characters is selected and developed. Here, then, we should look for the most striking manifestations of compensation of growth. But it is a fact, of which the significance is at once apparent, that, instead of meeting with the fulfilment of our expectations, the converse thrusts itself most obtrusively upon our attention. Nature here is most prodigal; giving growth for growth, and meeting the development of one feature with the corresponding development of another. The cases illustrating apparent balancement of growth are here exceptional. They bear a very insignificant proportion to those under nature. Hence we conclude that the law of compensation of growth never obtains, that its apparent manifestations are really due to the operation of the law of correlation.
But there are two classes of cases of which correlation is not an interpretation. The first is the instances in which the tie of correlation is in a measure broken by man's selection of one part, and by his systematic suppression of another. Darwin refers to these when he declares it "scarcely possible in most cases to distinguish between the supposed effects of such compensation of growth, and the effects of long-continued selection, which may at the same time lead to the augmentation of one part and the diminution of another."
The following is an example of the second class of cases: The Polish fowl is distinguished by the possession of a crest of feathers on the head. In consequence of its development, there arises a protuberance on the skull. This is due to correlation. But in the cock, the skull is so perforated with small holes that at any point a pin may be sunk to the brain. This is adduced as an instance of compensation of growth. But a rational explanation may readily be assigned. Darwin has shown that the crest of feathers is abnormal in the male, that it normally belongs to the female. The feature has been gained by the male by the somewhat mysterious law of the transmission of secondary sexual characters. The economy of growth may then be considered as abnormal, and may reasonably be attributed to the character not completely harmonizing with its fellows.
The facts of correlation meet with an exhaustive treatment at the hands of Darwin. Herbert Spencer, however, almost totally ignores them. Although they are seemingly most striking exemplifications of evolution, he passes with only an occasional incidental notice. What we conceive to be Mr. Spencer's reason for thus ignoring them, we will venture to give further on. But, while Darwin extends to the facts of correlation a full recognition, he is by no means over-desirous to ascertain their cause. Correlation is another of those laws which it pleases Darwin to consider as ultimate.
Now, the supposition that the correlated part has arisen by evolution, involves the absurd conclusion that a centre of growth normally preëxists without a relative arrangement of parts. And on the evolution hypothesis, we are forced to believe that an evolved part is correlated to another part not yet in existence; that all the parts of the organism anticipate, as it were, the birth of the new feature, and so adjust themselves as to become immediately susceptible to its influence; and that, while the previous coördination of parts is destroyed, owing to the influence of the new-born feature ramifying throughout the whole organization, the organism is capable of immediately effecting a re-coördination. To assume for any organism such powers as these, is virtual hylozoism. The only escape for him who admits the evolution of variations, is to adopt the explanation furnished by the Duke of Argyll—that correlations are the direct manifestations of design.
This interpretation of the teleologist precludes all further argument. We, of course, concur in design. But we do not deem ourselves therefore bound to take for granted the validity of every argument adduced in proof thereof. We conceive that design can be proved by incontrovertible evidence, and that it can be shown to manifest itself in conformity to laws not merely empirical.
As for the ultra-evolutionist, if he were to cease regarding correlation as an ultimate fact, and if he were to employ himself in placing an interpretation upon it, he would perceive that the tie of correlation is strongly suggestive of reversion, and that its phenomena completely negative the hypothesis of evolution.
On the hypothesis of reversion, correlation is perfectly explicable. The supposition of reversion necessarily involves the conclusion that all the features of the species coexisted in each individual, saving, of course, the characters peculiar to the opposite sex. The perfect organism, then, is a balance of all the parts. The parts are correlated to each other with respect to centres, and these centres are correlated to each other with respect to the axis or the aggregate. All the parts are mutually dependent. When a part is reduced, it tends to involve the reduction of its corresponding part. The centre of the parts is then weakened, and this weakening entails the weakening of the other centres, to which this center is correlated. The loss or suppression of even one part, then, manifestly disturbs the physiological balance—destroys the coördination of the parts. Under nature, many parts have been lost or reduced, and these have entailed the loss or reduction of others. When, under domestication, characters develop, owing to selection and favorable conditions, they concur with the different centres of growth to effect a return to the balance, and, in consequence, the correlated parts arise and assume their primordial relations to their correlatives and to the aggregate. When all the parts are developed, by correlation and otherwise, there result an equilibrium and a consequent perfect coördination. Correlation is the inseparable concomitant of coördination. Each implies the other. And this is the reason, we apprehend, why correlation is barely noticed by Mr. Spencer. He feared, we surmise, that a lengthy philosophical treatment of the subject would suggest the conception that correlated growth necessarily implied previously imperfect coördination.