evidence offered by one side was generally refuted by testimony from the other. The Supreme Court divided upon the case, and the seven thousand printed pages of evidence in the suit seem rather to prove the fallibility of human testimony than anything else. See article on Daniel Drawbaugh, by H. C. Merwin, Atlantic Monthly, Sept., 1888.”*

Resulting

common

places.

Since memory is always unreliable, almost the first thing for one who is doing scientific work of any sort, no matter how humble, to learn is the importance of keeping full and clear records of every detail of his experiments or observations that may have the slightest bearing on the question at issue. If a detail is to be preserved at all, it must be taken down at once; it is usually almost as easy to take down a point of doubtful value as to neglect it, and the best and most accurate of experimenters are only too liable to find their work less valuable than it might otherwise have been because there was some small detail of which they did not make a note. Most beginners need to be warned, too, not to keep notes on loose scraps of paper, not to use unfamiliar abbreviations without writing down their meaning, and to make their writing very legible. If they themselves are to be sure of its meaning, the record should be so clear and unambiguous that it could be easily understood by any one. In other words, it should be a true record, and not a mere series of suggestions for the memory. It is important, too, to number the pages (unless the record is in a book) and to leave plenty of blank space on each of them. There is a strong tendency to get things crowded some time or other before the results are finally computed, and a crowded record is very confusing. Finally, the notes should be indexed and put away in such order that they can be found at a moment's notice for years afterwards. To observe these simple precautions is to save much time and annoyance for everybody concerned; and what is said here

* W. H. Burnham, loc. cit.

about scientific work is just as true, mutatis mutandis, of a farmer trying to remember what his fields have done each year, or of business, school-teaching, housekeeping, or anything else where it is worth while to remember transactions accurately.

CHAPTER XXXIV.

THE DISCOVERY OF PAST AND FUTURE EVENTS IN GENERAL.

WHEN we wish to ascertain some specific fact that we have not been able to observe for ourselves there are only two ways of doing it. One is to depend upon the testimony of others; and the other is to draw an inference from what we know of the general laws of nature and the specific facts that we or others have observed. The latter method can be described very briefly, and so we shall speak of it first. This is the method used in tracing the history of the solar system from, or rather back to, the vapor and the star-dust from which the planets were made and in prophesying the condition of cold and darkness and lifelessness to which they may be destined. It is the method pursued by geology in tracing the changes which have taken place on the earth's crust and prophesying those which will take place. And it is the method pursued by evolutionary biology in tracing the history of life in the world as it has developed from one form to another. In all these sciences the starting-point is the present, and the question is always this: Granting the truth of the general laws assumed or ascertained by various sciences, what is the only concrete state of affairs that could have preceded the one which we observe to exist at the present, and what is the only concrete state of affairs that can succeed it?

The first thing to notice about this method is that we

In telling about history of

always start from the present. any sort we may often begin at the beginning'. The startIn investigating it we never can. Moreover, if ing-point. we are mistaken about some of the general laws or if we are not quite accurate about some of the concrete facts with which we start, the consequences of our error will affect all our history and all our prophecy; and since there is a chance of overlooking some essential fact or making some miscalculation at each stage of our regress into the past or progress into the future, the chances are that the farther we go the less accurate our account of things can be. The possibility of such accumulation of errors will always make a definite and detailed description of the world more doubtful the farther the described state of affairs is removed from the present data with which we have to start. But indeed what we know about the laws of nature and present concrete conditions is so slight in comparison with what we do not know, and even those things that we do know are so enormously complex, that no one really attempts to work out the problem in all its details, and the most that any scientist attempts to tell about either the distant past or the distant future is the broad outlines of things, which would remain substantially the same no matter what were true about any one of countless smaller details. A geologist can tell with perfect confidence. that where there is now a certain group of hills there was once a fairly level plateau, and he can tell that the change from one to the other was due in the main to the action of water running down to the valley below, but he would never attempt to tell the exact location of every stream or the amount of earth that one of them carried down on some particular day ten thousand years ago.

Even what a geologist does tell about the past and the future is not based upon the most ultimate laws of matter known. If it were, he would have deduced the history of the world from the known laws of chemistry and molecular physics; and such deduction is impossible because the situ

ations that these sciences deal with are exceedingly simple, and from the relations that are found to exist in these simple situations no one could possibly calculate what would happen under the vastly more complex conditions that are dealt with in geology. The geologist starts rather with empirical laws' which are much less precise (so far as particular molecules are concerned) than the laws of molecular physics, but which give a much better idea of what happens when things are arranged as he supposes them to be in the large. He sees, for example, that streams actually do wear away earth and rock from their beds and carry the débris away, and he determines by actual measurement the amount of earth of a given kind that a stream of a given size and swiftness carries off in a given length of time; and then he applies the 'empirical law' which he derives from such measurements directly to the problem in hand. He knows, of course, that the facts in the case are consistent with molecular physics, but he knows also that his data are much too crude and complex to be dealt with by that science; and so he works away with his empirical laws' in comparative oblivion of it. Almost all of our history of the world and our prophecy of its future is based upon such empirical laws' as these, derived from a view of things in the large; and of course any history or prophecy which is based altogether upon such broad rough laws cannot attempt to describe small details. Another thing to notice about scientific history and prophecy is that there is nothing in the laws and concrete facts upon which they are based to tell the scientist when the whole world-process began or when it will come to an end. For all we know or ever could know, God may annihilate the whole world to-morrow; but our prophecy based on laws and concrete facts of the present looks forward without limit towards a whole eternity. In the same way we trace what we suppose to be the broad outlines of history further and further into the past, and we never reach and never can reach a point at which we can say: Here everything

The limit.