That the comparison of the forenoon means should give in general a greater decrease than the comparison of those of the afternoon, is only a consequence of what has been stated above, viz. that the diurnal movements in the first year are smaller than in the second, and those of the second smaller than in the third. That difference must, therefore, not be considered as a real one, but merely as accidental; and we may expect, by a longer continuation of the observations, a difference in the opposite direction. As then there is no sufficient reason for preferring one of the results to the other, we can only take the mean of the two. The mean is, in the first year, 2′ 36″·5; in the second, 4′ 55″·9. We might regard this as a proof of an increasing rate of diminution in the declination: but this would be nothing more than a bad reason for a thing which in itself is true. It is known that the declination, which increased through all Europe during the last century, attained its maximum in the present century, and is now decreasing afresh. This transition must necessarily produce at first an imperceptible, and gradually a greater decrease. But although, from want of earlier observations, the precise year cannot be fixed in which the transition took place at Göttingen, yet it must be inferred from the observations made at other places, that it must have been at an earlier period than would follow from those two numbers, if we were to consider them as pure effects of the slow movement which we term secular. All other experience shows that 2′ 19″·4 is too great to be looked upon as a regular increase for a year. We, therefore, regard this difference as being for the most part accidental; so that, for the present, and until we have further experience to guide us, we must consider the mean value, 3′ 46″·2, as the annual decrease of the declination from 1834 to 1837.
As the difference between the declination in the forenoon.and afternoon is subject to an inequality, evidently varying with the season of the year, the question arises whether the change depending on the period of the year affects one of these declinations only, or one more than the other, or both equally, and what are the existing laws in this respect. A longer series of years, it is true, will be required to find out this law, than is needed to determine the mere difference between the declinations; nevertheless, it will be desirable to see what light the present observations throw on the subject.
