taken endeavor to make them parts of a general theory on the same scale in regard to details. A specialist, striving to generalize, nearly always goes more or less astray; but, instead of following him, even though with the idea of setting him right, it is best to take the special results he has obtained at the cost of much labor and research. If we do this with the work of all who have dealt specially with a subject, the chances are that we shall find we have gathered nearly enough to indicate a general theory, which shall include all these specially ascertained details. This the true theory, whatsoever it may be, should unquestionably do. My theory, at any rate, has been obtained in this way, and is intended in its wide generality to cover all the known facts.
So much premised, I note that my reasoning on the subject of comets and meteors starts from the idea which Professor Newton seems very properly to regard as almost necessarily true, that shooting-stars, fireballs, aerolites, and in fine all orders of meteoric bodies, belong to the same general class, differing only inter se in size, distribution, orbital motion, and the proportions in which the materials constituting them are distributed. It appears to me, as it does to him, that a theory which will account for such streams as supply the August and November displays, but not for the meteoric masses which fall sporadically, can not be the true general theory of meteors; nor can any theory be accepted as at once true and general which accounts for the holosiderites while it leaves unexplained the asiderites (so called, though in reality no meteors are free from iron). Again, noting that meteors have been associated with comets, we require for any theory which shall be accepted as generally true, that while it shall explain this connection between streams of bodies producing falling stars and certain special comets, it shall also be able to account for all comets as possibly associated with meteor-streams, and for all meteor-streams as possibly associated with comets. How much resides in this last condition, those only can guess who have put the matter to the test by striving to find a general theory of comets and meteors which shall not be found to be in conflict with some fact known about particular comets or some other fact known about particular meteor systems. Yet no general theory of comets and meteors can possibly be accepted which fails when thus tested, trying though the test may be. These trying tests are, indeed, particularly valuable for the seeker after truth, seeing that they serve to diminish his field of research by fencing out portions which can not usefully be dealt with. My own experience has convinced me that negative indications serve often to lead more directly to the truth than the most seemingly decisive evidence of the positive sort, though in reality it is by combining the two kinds of evidence, rejecting because of decisive negative evidence theory after theory from among those to which we are directed by decisive positive evidence, that we can alone hope to arrive at the true theory.
