The enquiry concerning the nature and causes of volcanic action is far from being a completed one. It is true that many hypotheses upon the subject have been framed, but in too many instances these have not been based on accurate observations and careful generalisations, and can be regarded as little better than mere guesses. Indeed, the state of the enquiry at the present time would seem to be as follows. Although much remains to be done in the direction both of observation and experiment, the main facts of the case have been established upon irrefragable evidence. The classification and comparison of these facts have led to the recognition of certain laws, which seem to embrace all the known facts. To account for these facts and their demonstrated relations to one another, certain tentative hypotheses have been suggested; but in no case can it be truly said that these latter have so far stood the test of exact enquiry as to deserve to rank as demonstrated truths. A complete and consistent theory of volcanic action still remains to be discovered. [Sidenote: VALUE AND LIMITS OF HYPOTHESES.] In accordance with the plan which we have sketched out for ourselves at the commencement of this work, we shall aim at following what has been the order of investigation and discovery in our study of volcanic action; and in this concluding chapter we shall indicate the different hypotheses by which it has been proposed to account for the varied phenomena, which we have discussed in the preceding pages, and their remarkable relations to one another. We shall endeavour, in passing, to indicate how far these several hypotheses appear to be probable, as satisfying a larger or smaller number of those conditions of the problem which have been established by observation, experiment, and careful reasoning; but we shall at the same time carefully avoid such advocacy of any particular views as would tend to a prejudgment of the question. Hypothesis is, as we have seen, one of the legitimate and necessary operations in scientific investigation. It only becomes a dangerous and treacherous weapon when it is made to precede rather than to follow observation and experiment, or when being regarded with paternal indulgence, an attempt is made to shield it from the relentless logic of facts. Good and bad hypotheses must be allowed to 'grow together till the harvest;' such as are unable to accommodate themselves to the surrounding conditions imposed by newly-discovered facts and freshly-established laws will assuredly perish; and in this 'struggle for existence' the true hypothesis will in the end survive, while the false ones perish.