Finally, M. LeRoy cites the rotation of the earth; it has been answered: but this is not a fact, and he has replied: it was one for Galileo, who affirmed it, as for the inquisitor, who denied it. It always remains that this is not a fact in the same sense as those just spoken of and that to give them the same name is to expose one's self to many confusions. Here then are four degrees: 1º. It grows dark, says the clown. 2º. The eclipse happened at nine o'clock, says the astronomer. 3º. The eclipse happened at the time deducible from the tables constructed according to Newton's law, says he again. 4º. That results from the earth's turning around the sun, says Galileo finally. Where then is the boundary between the fact in the rough and the scientific fact? To read M. LeRoy one would believe that it is between the first and the second stage, but who does not see that there is a greater distance from the second to the third, and still more from the third to the fourth. Allow me to cite two examples which perhaps will enlighten us a little. I observe the deviation of a galvanometer by the aid of a movable mirror which projects a luminous image or spot on a divided scale. The crude fact is this: I see the spot displace itself on the scale, and the scientific fact is this: a current passes in the circuit. Or again: when I make an experiment I should subject the result to certain corrections, because I know I must have made errors. These errors are of two kinds, some are accidental and these I shall correct by taking the mean; the others are systematic and I shall be able to correct those only by a thorough study of their causes. The first result obtained is then the fact in the rough, while the scientific fact is the final result after the finished corrections. Reflecting on this latter example, we are led to subdivide our second stage, and in place of saying: 2. The eclipse happened at nine o'clock, we shall say: 2_a_. The eclipse happened when my clock pointed to nine, and 2_b_. My clock being ten minutes slow, the eclipse happened at ten minutes past nine. And this is not all: the first stage also should be subdivided, and not between these two subdivisions will be the least distance; it is necessary to distinguish between the impression of obscurity felt by one witnessing an eclipse, and the affirmation: It grows dark, which this impression extorts from him. In a sense it is the first which is the only true fact in the rough, and the second is already a sort of scientific fact. Now then our scale has six stages, and even though there is no reason for halting at this figure, there we shall stop.