[5] The only point at which the general principle of the series fails in its application, is the subdivision of Physics; and there, as the subordination of the different branches scarcely exists, their order is of little consequence. Thermology, indeed, is altogether an exception to the principle of decreasing generality, heat, as Mr Spencer truly says being as universal as gravitation. But the place of Thermology is marked out, within certain narrow limits, by the ends of the classification, though not by its principle. The desideratum is, that every science should precede those which cannot be scientifically constitute or rationally studied until it is known. It is as a means to this end, that the arrangement of the phaenomena in the order of their dependence on one another is important. Now, though heat is as universal a phaenomenon as any which external nature presents, its laws do not affect, in any manner important to us, the phaenomena of Astronomy, and operate in the other branches of Physics only as slight modifying agencies, the consideration of which may be postponed to a rather advanced stage. But the phaenomena of Chemistry and Biology depend on them often for their very existence. The ends of the classification require therefore that Thermology should precede Chemistry and Biology, but do not demand that it should be thrown farther back. On the other hand, those same ends, in another point of view, require that it should be subsequent to Astronomy, for reasons not of doctrine but of method: Astronomy being the best school of the true art of interpreting Nature, by which Thermology profits like other sciences, but which it was ill adapted to originate.
[6] The philosophy of the subject is perhaps nowhere so well expressed as in the "Système de Politique Positive" (iii. 41). "Conçu logiquement, l'ordre suivant lequel nos principales théories accomplissent l'évolution fondamentale résulte nécessairement de leur dépendence mutuelle. Toutes les sciences peuvent, sans doute, être ébauchées à la fois: leur usage pratique exige même cette culture simultanée. Mais elle ne peut concerner que les inductions propres à chaque classe de spéculations. Or cet essor inductif ne saurait fournir des principes suffisants qu'envers les plus simples études. Partout ailleurs, ils ne peuvent être établis qu'en subordonnant chaque genre d'inductions scientifiques à l'ensemble des déductions emanées des domaines moins compliqués, et dès-lors moins dépendants. Ainsi nos diverses théories reposent dogmatiquement les unes sur les autres, suivant un ordre invariable, qui doit régler historiquement leur avénement décisif, les plus indépendantes ayant toujours dû se développer plus tôt."