Page images
PDF
EPUB

visible signal (in preference, I mean, to some electric signal), i.e., a light-message, which travels so quickly that we can wholly neglect the time it has taken, in comparison with that taken by the sound. Obviously, we have no such means of measuring the passage of light, for what is our signal to be if the velocity with which it is conveyed is so largely to exceed the velocity of light that we can neglect the time occupied in its transmission? But if we had even a satisfactory answer to this question (instead of having none whatever), the problem would yet be insoluble in this manner. Suppose the light were shown at a distance of 500 miles-about the limit at which any terrestrial light could be seen, even under the most favourable atmospheric conditions-yet the time occupied by the lightwaves in traversing this distance would be but about the 360th part of a second. What instrument or what observer could note such an interval-to say nothing of measuring it, which would yet be absolutely essential to the successful solution of the problem?

I shall not here enter into a full account of the means

by which Foucault and Fizeau solved a problem apparently so intractable, referring the reader to Pouillet's Physics' and other works in which the subject of light is dealt with. The general principle of the method employed by Fizeau may be thus presented. Suppose we see an object by light-rays which have been caused to traverse a long path by means of several reflections. Now conceive that the continuity of the long path is simultaneously broken at regular intervals at two

points, one near the beginning the other near the end of the path, the path being broken-re-made-broken -re-made, and so on. Then if light travelled with infinite velocity, the light-rays, which at any instant traversed the first part of the path at a time when the path was made there, would traverse the last part also, because at that same instant the path would be complete there also. But light not travelling with infinite velocity, the light-rays which pass the first part of the path may be stopped by the break in the second part, if the interval between the making and breaking be but short enough. Now, Fizeau had a revolving toothed wheel, and matters were so arranged that when a tooth of this wheel was opposite a certain small aperture, the path of light was broken both at its beginning and end, for the light had to pass through this aperture, and then, after pursuing a long course, to pass out again through the same aperture. Thus, when the revolution was moderately rapid, light-rays which passed through the aperture found the aperture open when they came back again to it; but by causing the revolution to be very rapid indeed, so that a very minute fraction of a second elapsed between the passage of successive teeth across the aperture, it was possible to cause the light-rays which had entered while the aperture was open to be prevented from passing out again by the interference of a tooth of the wheel. It is easily seen that when the wheel revolved at this particular rate there would be a total eclipse so far as light coming through the aperture was concerned. For light

which went in when any portion of the aperture was free, would return to the aperture when just that portion of the aperture was closed. Then, as Fizeau had the means of telling at what rate the wheel was revolving when total eclipse thus occurred, he could tell precisely what fraction of a second elapsed between the passage of tooth after tooth across the aperture; and knowing the length of the path traversed by the lightrays he could measure the velocity of light with a considerable degree of accuracy. Foucault adopted an arrangement in which a plane mirror was caused to rotate very rapidly, and the principle of his plan (which, to be fully explained, would require more space than is here available) depends on the duration of visual impressions. Fizeau's method, in some respects inferior to Foucault's, resulted in assigning to light a velocity of 194,600 miles per second; but Foucault's gave a velocity of only 185,300 miles per second, fall

* It may be thus exhibited :—An image of a certain wire is seen directly, and when a certain plane mirror which can be rapidly revolved is in a certain position, the image of the wire is caused to appear in coincidence with the wire seen directly. The mirror is so rotated as to take up once in each rotation the required position; and so long as the rotation is slow the reflected image makes successive appearances. With an increase in the velocity of rotation the image appears continuously in one place-owing to the continuance of visible impressions. Now, for moderate velocities of rotation, the position thus taken up by the image coincides with that of the wire seen directly. But with a very rapid rotation the position of the mirror suited for causing the image to be visible (after reflection), no longer accords appreciably with the position required when the mirror is at rest. Accordingly, the image appears appreciably separated from the wire seen directly; and the amount of separation, combined with the known velocity of rotation, supplies the means of estimating the velocity of light.

ing considerably short of the estimate of 192,000 miles above referred to. So satisfactory were Foucault's experiments, that this discrepancy was held gravely to affect the estimate of the Sun's distance, on which the latter value is based. It would follow, if Foucault's experiments were held to indicate truly the velocity of light in the interplanetary spaces (as well as in air), that the Sun's mean distance would be but 91,400,000 miles, his parallax 8′′-942.

The Astronomer Royal also suggested the application of a method already employed by Flamsted, and later by Bond, of America. He pointed out that instead of comparing the position of Mars on the sky (when the planet is near opposition) as seen from different stations, an even more satisfactory estimate of the planet's distance, and so of the Sun's, might be obtained by observing how far the diurnal rotation of the Earth, by shifting the place of any fixed station, affected the apparent position of the planet. It is clear that if the station E (fig. 9) is supposed to be carried by the Earth's rotation to E', the observer can as effectively compare the distances at which the observed places of Mars, m and m', lie from a fixed star s, as though there were two observers, one at E and the other at E' at the same instant of time; for astronomers know well how to take into due account the motion of Mars during the interval.

In 1862 this method was employed, as well as the former method of treating observations of Mars. The result was to confirm the impression that the Sun lies

nearer to us than had been so long imagined. Stone, of Greenwich, by combining the two methods, deducing the solar parallax first from observations of Mars made at Greenwich alone, then from observations made at Greenwich and Capetown, then from observations made at Greenwich and Williamstown, and combining all these results, deduced a solar parallax of 8"-943, with a probable error of 0"051. This corresponds to a distance of about 91,400,000 miles, with a probable error of about 500,000 miles. Winnecke, by combining observations of Mars made at Poulkowa and Capetown, deduced a parallax of 8"-964, corresponding to a distance of about 91,200,000 miles. Newcomb, from the observations of Mars in the same year, deduced a parallax of 8" 855, corresponding to a distance of about 92,300,000 miles.

Besides these, there were estimates by Leverrier, Stone, Pogson, and several others, founded on the re-examination of processes already referred to, or on observations enabling those processes to be applied anew with more or less chance of exactitude in the results.

By the year 1864 it had become abundantly clear that the accepted estimate of the Sun's distance was too great. All the new values clustered around a value of about 8"-9 for the parallax, corresponding to a distance of about 91,850,000 miles.

Thus astronomers were led to re-examine the observations of the transit of 1769 in order to see whether they could be so interpreted as to correspond with the

« PreviousContinue »