TERMS belonging to a system are defined, not by the meaning of their radical words, but by their place in the system. That they should be appropriate in their signification, aids the processes of introducing and remembering them, and should therefore be carefully attended to by those who invent and establish them; but this once done, no objections founded upon their etymological import are of any material weight. We find no inconvenience in the circumstance that _geometry_ means the measuring of the earth, that the name _porphyry_ is applied to many rocks which have no fiery spots, as the word implies, and _oolite_ to strata which have no roelike structure. In like manner, if the term _pœcilite_ were already generally received, as the name of a certain group of strata, it would be no valid ground for quarrelling with it, that this group was not always variegated in colour, or that other groups were equally variegated: although undoubtedly in _introducing_ such a term, care should be taken to make it as distinctive as possible. It often happens, as we have seen, that by the natural progress of changes in language, a word is steadily confirmed in a sense quite different from its etymological import. But though {302} we may accept such instances, we must not wantonly attempt to imitate them. I say, not wantonly: for if the progress of scientific identification compel us to follow any class of objects into circumstances where the derivation of the term is inapplicable, we may still consider the term as an unmeaning sound, or rather an historical symbol, expressing a certain member of our system. Thus if, in following the course of the _mountain_ or _carboniferous_ limestone, we find that in Ireland it does not form mountains nor contain coal, we should act unwisely in breaking down the nomenclature in which our systematic relations are already expressed, in order to gain, in a particular case, a propriety of language which has no scientific value. All attempts to act upon the maxim opposite to this, and to make our scientific names properly descriptive of the objects, have failed and must fail. For the marks which really distinguish the natural classes of objects, are by no means obvious. The discovery of them is one of the most important steps in science; and when they are discovered, they are constantly liable to exceptions, because they do not contain the essential differences of the classes. The natural order _Umbellatæ_, in order to be a natural order, must contain some plants which have not umbels, as _Eryngium_[37\4]. 'In such cases,' said Linnæus, 'it is of small import what you call the order, if you take a proper series of plants, and give it some name which is clearly understood to apply to the plants you have associated.' 'I have,' he adds, 'followed the rule of borrowing the name _à fortiori_, from the principal feature.' [Note 37\4: See _Hist. Ind. Sc._ b. xvi. c. iv. sect. 5.]