But this is just what is claimed to be unnecessary. Physics and chemistry are said to give us the key to everything. Eimer's great work is instructive in this respect. It is well known what persevering effort this biologist has devoted to demonstrating that transformation is brought about by the influence of the external on the internal, continuously exerted in the same direction, and not, as Darwin held, by accidental variations. His theory rests on observations of the highest interest, of which the starting-point was the study of the course followed by the color variation of the skin in certain lizards. Before this, the already old experiments of Dorfmeister had shown that the same chrysalis, according as it was submitted to cold or heat, gave rise to very different butterflies, which had long been regarded as independent species, _Vanessa levana_ and _Vanessa prorsa_: an intermediate temperature produces an intermediate form. We might class with these facts the important transformations observed in a little crustacean, _Artemia salina_, when the salt of the water it lives in is increased or diminished.[32] In these various experiments the external agent seems to act as a cause of transformation. But what does the word "cause" mean here? Without undertaking an exhaustive analysis of the idea of causality, we will merely remark that three very different meanings of this term are commonly confused. A cause may act by _impelling_, _releasing_, or _unwinding_. The billiard-ball, that strikes another, determines its movement by _impelling_. The spark that explodes the powder acts by _releasing_. The gradual relaxing of the spring, that makes the phonograph turn, _unwinds_ the melody inscribed on the cylinder: if the melody which is played be the effect, and the relaxing of the spring the cause, we must say that the cause acts by _unwinding_. What distinguishes these three cases from each other is the greater or less solidarity between the cause and the effect. In the first, the quantity and quality of the effect vary with the quantity and quality of the cause. In the second, neither quality nor quantity of the effect varies with quality and quantity of the cause: the effect is invariable. In the third, the quantity of the effect depends on the quantity of the cause, but the cause does not influence the quality of the effect: the longer the cylinder turns by the action of the spring, the more of the melody I shall hear, but the nature of the melody, or of the part heard, does not depend on the action of the spring. Only in the first case, really, does cause _explain_ effect; in the others the effect is more or less given in advance, and the antecedent invoked is--in different degrees, of course--its occasion rather than its cause. Now, in saying that the saltness of the water is the cause of the transformations of Artemia, or that the degree of temperature determines the color and