however, as it was observed that the stars retained their relative places; that the times of their rising and setting varied with the seasons; that sun, moon, and planets moved among them in a plane, and the belt of the Zodiac was marked out and divided,—then a new order of things began. Traces of the earlier stage remained in the names of the signs and constellations, just as the Scandinavian mythology survives now in the names of the days of the week; but, for all that, the understanding was now at work on the thing; science had begun, and the first triumph of it was the power of foretelling the future. Eclipses were perceived to recur in cycles of nineteen years, and philosophers were able to say when an eclipse was to be looked for. The periods of the planets were determined. Theories were invented to account for their eccentricities; and, false as those theories might be, the position of the planets could be calculated with moderate certainty by them. The very first result of the science, in its most imperfect stage, was a power of foresight; and this was possible before any one true astronomical law had been discovered.

We should not therefore question the possibility of a science of history because the explanations of its phenomena were rudimentary or imperfect: that they might be, and long continue to be, and yet enough might be done to show that there was such a thing, and that it was not entirely without use. But how was it that in those rude days, with small knowledge of mathematics, and with no better instruments than flat walls and dial-plates, those first astronomers made progress so considerable? Because, I suppose, the phenomena which they were observing recurred, for the most part, within moderate intervals; so that they could collect large experience within the compass of their natural lives; because days and months and years were measurable periods, and within them the more simple phenomena perpetually repeated themselves.

But how would it have been if, instead of turning on its axis once in twenty-four hours, the earth had taken a year about it; if the year had been nearly four hundred years; if man's life had been no longer than it is, and for the initial steps of astronomy there had been nothing to depend upon except observations recorded in history? How many ages would have passed, had this been our condition, before it would have occurred to any one, that, in what they saw night after night, there was any kind of order at all?

We can see to some extent how it would have been, by the present state of those parts of the science which in fact depend on remote recorded observations. The movements of the comets are still extremely uncertain. The times of their return can be calculated only with the greatest vagueness.

And yet such a hypothesis as I have suggested would but inadequately express the position in which we are in fact placed toward history. There the phenomena never repeat themselves. There we are dependent wholly on the record of things said to have happened once, but which never happen or can happen a second time. There no experiment is possible; we can watch for no recurring fact to test the worth of our conjectures. It has been suggested fancifully, that, if we consider the universe to be infinite, time is the same as eternity, and the past is perpetually present. Light takes nine years to come to us from Sirius: those rays which we may see to-night, when we leave this place, left Sirius nine years ago; and could the inhabitants of Sirius see the earth at this moment, they would see the English army in the trenches before Sebastopol, Florence Nightingale watching at Scutari over the wounded at Inkermann, and the peace of England undisturbed by "Essays and Reviews."

As the stars recede into distance, so time recedes with them; and there may be, and probably are, stars from which Noah might be seen stepping into the ark, Eve listening to the temptation of the serpent, or that older race, eating the oysters and leaving the shell-heaps behind them, when the Baltic was an open sea.

Could we but compare notes, something might be done; but of this there is no present hope, and without it there will be no science of history. Eclipses, recorded in ancient books, can be verified by calculations, and lost dates can be recovered by them; and we can foresee, by the laws which they follow, when there will be eclipses again. Will a time ever be when the lost secret of the foundation of Rome can be recovered by historic laws? If not, where is our science? It may be said that this is a particular fact, that we can deal satisfactorily with general phenomena affecting eras and cycles. Well, then, let us take some general phenomenon; Mahometanism, for instance, or Buddhism. Those are large enough. Can you imagine a science which would have[1] foretold such movements as those? The state of things out of which they rose is obscure; but, suppose it not obscure, can you conceive that, with any amount of historical insight into the old Oriental beliefs, you could have seen that they were about to transform themselves into those particular forms and no other?

It is not enough to say, that, after the fact, you can understand partially how Mahometanism came to be. All historians worth the name have told us something about that. But when we talk of science, we mean something with more ambitious pretences, we mean something which can foresee as well as explain; and, thus looked at, to state the problem is to show its absurdity. As little could the wisest man have foreseen this mighty revolution, as thirty years ago such a thing as Mormonism could have been anticipated in America; as little as it could have been foreseen that table-turning and spirit-rapping would have been an outcome of the scientific culture of England in the nineteenth century.

The greatest of Roman thinkers, gazing mournfully at the seething mass of moral putrefaction round him, detected and deigned to notice among its elements a certain detestable superstition, so he called it, rising up amidst the offscouring of the Jews, which was named Christianity. Could Tacitus have looked forward nine centuries to the Rome of Gregory VII, could he have beheld the representative of the majesty of the Cæsars holding the stirrup of the Pontiff of that vile and execrated sect, the spectacle would scarcely have appeared to him the fulfilment of a national expectation, or an intelligible result of the causes in operation round him. Tacitus, indeed, was born before the science of history; but would M. Comte have seen any more clearly?

Nor is the case much better if we are less hard upon our philosophy; if we content ourselves with the past, and require only a scientific explanation of that.

First, for the facts themselves. They come to us through the minds of those who recorded them, neither machines nor angels, but fallible creatures, with human passions and prejudices. Tacitus and Thucydides were perhaps the ablest men who ever gave themselves to writing history; the ablest, and also the most incapable of conscious falsehood. Yet even now, after all these centuries, the truth of what they relate is called in question. Good reasons can be given to show that neither of them can be confidently trusted. If we doubt with these, whom are we to believe?

Or, again, let the facts be granted. To revert to my simile of the box of letters, you have but to select such facts as suit you, you have but to leave alone those which do not suit you, and,