unable to perceive how the definition is improved by this emendation. There is not, and cannot possibly be, such a thing as _unconditionally_ invariable sequence, as, indeed, Mr. Mill himself virtually admits by expressly assuming as an indispensable condition of all causation that 'the present constitution of things endure.' But if, notwithstanding the presence of this indispensable condition, it be permissible to call any sequence unconditionally invariable, then the sequences of night upon day and of day upon night are such sequences, and day and night continue consequently entitled to be styled each other's causes as much under the amended as under the original definition. For as long as 'the present constitution of things endures,' that is, as long as the earth continues to revolve on its axis, and the sun continues to shine, and no opaque substance intervenes between earth and sun, day and night will continue to be as invariably and unconditionally each other's antecedents as sunlight will continue to be the antecedent or concomitant of day. True, Mr. Mill denies that the earth's diurnal motion is part of the present constitution of things, because, according to him, 'nothing can be so called which might possibly be terminated or altered by natural causes:' but, if so, then neither ought sunlight to be so called, for it too quite possibly may, nay, in the opinion of many philosophers, most certainly will, be extinguished eventually by natural causes. If day ought not to be called the unconditionally invariable consequent of night merely because it would cease to be so if the earth were to cease turning on its axis, then neither ought it to be called the unconditionally invariable consequent of the unshrouded proximity of the sun, inasmuch as it would cease to be thereupon consequent if the sun were to become burnt out. If night be not, and if sunlight be, the cause of day, the reason is not that sunlight always hitherto has been, and, on one indispensable condition, always will be followed by day, for so equally has hitherto been, and on the same condition will hereafter be, night. The real reason is, that sunlight not only always has been and will be, but also always _must_ be followed by day; that unless the constitution of nature very materially change, wherever is sunlight day _must_ be; whereas not only might day be, although night had never preceded, but unless night had preceded, day must have been from the beginning. In short, to constitute cause, invariability, however unconditional, will not suffice. Another quality must be added, and that quality I contend to be obligatoriness. A cause, I maintain, would not be a cause unless its effect not only do or will, but _must_ necessarily follow it. In common with the great unphilosophic mass of mankind, I hold that between cause and effect there is a binding power which constrains the one to follow the other. If asked whence we suppose that