ment in case of disobedience. The sovereign of England, for example, may command his subject, the sovereign of Hanover, under pain of death, to collect taxes in his German dominions, and remit them to England. In his attempt to avoid this conclusion Austin concedes the very point at issue, and seems virtually to adopt the theory of sovereignty which has been suggested in the preceding pages; for, by distinguishing between the acts which the king, of Hanover performs as subject of England, and those which he performs as sovereign of a foreign country, and saying that the legislative power of England covers only the former, he admits that the British sovereign may have power to issue commands which relate to one class of acts, and at the same time may not have power to issue commands which relate to another. This is nothing less than an admission that the power of the sovereign is not always unlimited. He declares, moreover, that the question whether the legislative power of England extends to the acts of its subject performed as sovereign of Hanover is determined by the habitual obedience of the subject in that capacity. He considers, therefore, that, in this case at least, the extent of the sovereign's power is measured by the habitual obedience of the subject. The same or a similar difficulty is involved in Austin's statement (Lect. VI., p. 323, 2d ed.), that a person may be at the same time completely a member of one independent political society, and for certain limited purposes a member of another; but he makes no attempt to solve it.
If the extent of sovereign power is measured by the disposition to obedience on the part of the bulk of the society, it may be said that the power of no sovereign can be strictly unlimited, because commands can be imagined which no society would be disposed to obey. This may very well be true, and perhaps it would be proper to classify sovereigns, not according as their authority is absolute or not, but according as it is indefinite, or restrained within bounds more or less definitely fixed. Unless, indeed, the limits of power are tolerably well determined they tend to stretch farther and farther. Now, definite limits may be set to sovereign power in either one of two ways. They may result from the rivalry of two independent rulers, which settle by negotiation questions concerning the boundaries of their respective jurisdictions, and quarrel when they cannot agree; or they may be established by some formal declaration, which by sufficient precision enables the bulk of the society to have a general opinion about the extent of leg-
