ment occurs whose cause is not looked for and generally found in the uniform operation of the simple law of gravity in its direct or reflected action upon the various members of the solar system. The singular fact of Encke's comet, which experiences a delay which has been attributed to a resisting medium, may be regarded as a solitary exception to the general truth. The laws of motion among the heavenly bodies are so few and clear, that the character of the disturbance will generally indicate something in regard to the cause which produces it. But the elements of the earth's magnetism are exposed to abrupt and violent fluctuations, which, so far as the circumstances are known, acknowledge no periods, and, although perhaps capable of being explained by many conceivable causes, which are in constant operation, and, therefore, at the disposal of the philosopher, they cannot be distinctly brought home to any single one, and are at present regarded as inexplicable. These magnetic hurricanes, as they have been fancifully called, are often exhibited during auroral appearances, but many of them, so far as has been observed, are not coincident in time with this or any other class of natural phenomena. Now, every observed position of the needle, for a given moment, is beset with all these regular and irregular variations; which must be carefully eliminated, by multiplying the number, and shifting the exposure of the observations, before we can be assured what is the exact and absolute value of the element for that time. And when we are seeking the amount of any particular order of perturbations, we must proceed as in the astronomical case, by selecting, as far as may be, times for observation, when the disturbance in question is a maximum, and all others are of minimum value. The practice of the observer will supply many artifices of this sort, for eluding or grappling with difficulties, which appear, at first sight, insurmountable. It is obvious, that when the object is to ascertain the steady and periodical variations of the meridian, we should exclude from the comparison those days whose serenity is affected by what Humboldt has denominated magnetic storms, just as we should pass over days of violent winds and tempests, in deducing the gradual rise and fall of temperature during the 24 hours. No attempt should be made to frame an hypothesis, or even to hazard a conjecture in regard even to the variations of the shortest period from one year's observations, however unremitted they may have been; but these observations may be of use in confirming a theory long entertained and well established by facts noticed in other places.

The observations of Graham, in 1772, which resulted in the detection of the diurnal variation of the magnetic meridian at London, have been repeated since in various parts of the world with increased delicacy and skill, and with the same general result which is briefly described. The magnetized bar, free to place itself in the magnetic meridian, does not remain in one fixed position during the day; but sometime in the morning, between six and eight o'clock, as

the average statement, it starts in a westerly direction, and moves that way till between one and three in the afternoon; then it begins to retrace its steps back to the east again. These points of maximum and minimum declination are formed in every diurnal curve, and at nearly the same hour. We shall hereafter see what the limits of the time are. There are two ways in which the bar regains its first position. In some places, as Paris for example, it arrives at its greatest eastern elongation again between eight and eleven o'clock in the evening, and then remains stationary till the time of morning excursion has come round once more. In other places, as at Cambridge, it travels eastward till evening, and then goes back to form a secondary point of maximum westerly deviation about three o'clock, A. M.; after which it passes eastward, and recovers at eight o'clock the place it occupied 24 hours before. In certain cases, especially in northern latitudes, even when the secondary maximum and minimum are not formed, the bar does not remain stationary during the night, but occupies nearly all the time from three P.M. to eight A.M. in returning through the space it has just passed over in seven hours. Again, the arc traversed by the bar in its daily excursions, varies perceptibly from one day to another; but the approximate law is, that in the six months from the vernal to the autumnal equinox, its value is between 13' and 15'; and, in the remaining six months, the mean of the daily arc is between 8 and 10. But there may be single days when it amounts to 25', and others when it is as small as 6'.

Gauss thinks that eight A. M. and 1 г. M., of mean solar time, are never far from the periods of daily minimum and maximum declination in Gottingen, and that part of the globe. It appears, from a report in regard to the magnetic state of the Russian empire, for 1837, that, at St. Petersburg, the greatest westerly position of the north end of the magnetic meridian is near two o'clock P. M., and the opposite position is at eight in the morning, with the exception of November, December, and January, when it occurs later. This is easily explained, by the high latitude of the place, when we come to consider the dependence of this daily motion on the sun. Since the declination is easterly in some parts of Russia, it follows that the maximum declination there is in the morning, and the minimum in the afternoon. As the report in question has been published with great care, we extract a table of the monthly

• Annuaire Magnetique et Météorologique du Corps de Ingénieurs des Mines de Russie. Année 1837. St. Petersburg, 1839.

means of the arc of daily excursion, to shew how near they correspond to the more extended means, which we have mentioned above.

We have no occasion, in this place, to remark on the cause of this great inequality of arc in the different months, or on the times when the maximum and minimum occur. We only wish it to be observed, that the daily curve, so perceptible in other places, and so marked by its general. uniformity of appearance, is clearly seen in the Cambridge observations. We have here referred to the plates, which represent some of these diurnal curves; but Table II., at the end, will display, in a condensed form, the results of 12 months' observation on this point. As the values for some of the months were deduced from scanty observations, they cannot be brought into a fair comparison with more comprehensive means. If it shall appear that the times of maximum and minimum declination are embraced within the limits of a few hours, it is extremely important that observations should be made during those periods every day in the year, to determine the precise moment when they occur each day, and the arc of excursion. Means drawn from such abundant data might lead to a satisfactory solution of the daily changes, and not leave them, as at present, to probable conjecture.

It appears, from observations on various parts of our planet, that periodical changes of days, months, years, and longer duration, are affecting its magnetic equilibrium; all of which come, with few exceptions, under the same general expressions, so much so as to leave no doubt that they have a common origin. Hence it might be expected, that if the corresponding curves of magnetic declination were drawn, for the same solar time, in different places, they would conform to each other, and exhibit a kind of parallelism. But in addition to the orderly variations, there are other perturbations of a sudden and irregular nature; and it becomes an object of extreme interest to inquire how far these are local and accidental, and to what extent they must be regarded as general, and proceeding from some grand central force. Such an examination was contemplated,

as among the good results of the simultaneous observations made in Germany, Russia, Italy, &c., an account of which may be found in Taylor's "Scientific Memoirs," and "The Russian Magnetic Annual," to which I have already referred; and a possible answer to an intricate question was held out as as inducement to engage in the magnetic crusade. The expected comparison exhibits surprising coincidences, in most of the irregular movements of the magnetometer, as if the cause were co-extensive with the range of magnetic posts. Hitherto our country, and, indeed, this whole western continent, has not been represented in this congress of nations for a scientific object; having been destitute of the means of contributing its portion to the general levy which has been made upon them. But the term-day observations at Toronto, U.C., Philadelphia, and Cambridge, will furnish materials for doing this now. Plate III. exhibits the diurnal curves of the October termday, for the magnetic observations of Toronto and Cambridge. The peclination of the meridian was very considerably deranged by derturbations during the first 12 hours of the magnetic day (which always begins at 10 P.M. of Gott. м. T.), and affords, therefore, a favourable opportunity of discovering the extent of extraordinary disturbances. Now a glance at the plate betrays wonderful concert in the motions of the two remote bars; almost every digression of the bar, at Cambridge, and every change of curvature in the magnetic curve of that place, has something corresponding to it in the curve of Toronto. There are a few singular exceptions; and, in a word, the bar at Toronto seems to have been more agitated, and to have made greater excursions. But these discrepancies throw no doubt on the subject; for they are just such as must be expected to occur in results which depend upon complex and multiform agents. A general, and not a mathematical agreement, is all that can be expected. The times which mark the limits of the eastern and western excursions, in this fluctuating motion of the bars, agree with great precision, except in two or three instances, where the Cambridge curve lags behind five minutes or less. The parallelism appears at once, from considering that the two curves, starting from points 12 minutes apart, and making several digressions, the same way, of greater amount than the original arc of separation, do not cross each other except once; and this solitary instance will hardly be regarded as a transgression of the rule, since it arose simply from one bar being more affected by a particular wave of the magnetic tide than the other bar. On Plate II. is described the diurnal curve for the May term-day. This, too, is singularly disturbed from 11 P. M., Gott. M. T., till 10 o'clock of the next morning. We have had an opportunity of comparing this with the similar curve observed by Professor Bache, at the Girard College, in Philadelphia, and here again the instances of agreement make a stronger impression than the rare cases of discrepance. The correspondences are frequent and imposing; while the points which exhibit so much coincidence

are few, and of less importance. The agreement as to time, is, in almost every place, precise; the disagreement is confined chiefly to the extent or the existence of the motion. At 12, 7, 9 o'clock, Gott. M. T., some excursions appear in the Philadelphia curve, which have little or no counterpart in the Cambridge curve. They are deserving of notice, because they are more considerable than any others of a local character which have been noticed, and yet they do not, any of them, exceed 5' of arc. But the remarkable excursions

in Plate II., at a, b, c, d, e, f, are all faithfully represented on the Philadelphia curve. Only the upper branches, which belong to the easterly motion, are exceeded by the corresponding simultaneous movement at Philadelphia, while the grand excursions, at the bottom of the plate, which are due to the westerly motion, or increase of declination, go beyond their parallel passages in the Philadelphia curve. For instance, the branches, b, c, d. e, f, extend respectively 10', 12', 11', 5', 6', farther in the Philadelphia than the Cambridge curve. The reverse happens at g, h, i, where the Cambridge branches go 13', 30, 14, farther west than the analagous ones at Philadelphia. The close coincidence of the times when the direction of any great motion changes, observed at Toronto, Philadelphia, and Cambridge, makes it probable that the longitudes of the three places are known to a close approximation, or, at least, that the difference of longitude between these places are not much in fault; for, if they were, it would, by its effect upon the regulation of the clocks, prevent a coincidence of magnetic disturbances which actually existed, and make the appearances different from what we observe; unless we can suppose, what seems extremely improbable, that the constant error of the time should by chance balance precisely the actual want of coincidence in the arrival of the magnetic impulse at each place so as to deceive us in the final result. This, however, is a subject which requires to be pursued longer. The observations contemplated in coming years will, it is believed, furnish the materials for a comprehensive study of this problem, and enable us to determine, with more confidence than can be reposed in the comparison of places not widely removed from each other, the extent and the laws of what are now classed among the irregular perturbations of the magnetic equilibrium. All attempts at induction now must be considered as subordinate to the final discussion; but they have their purpose in indicating from time to time the direction to which the attention of observers should be particularly turned. To this end an arrangement, supplementary to the large plan, was completed with Lieutenant Riddell, for making simultaneous observations on the declination magnetometer at Toronto and Cambridge, at intervals of two minutes from 0 45' to 1 45' P.M. Gott. M.T. Such observations were accordingly taken at Toronto for the assigned hour every day (Sundays excepted), from October 23d to November 19th, inclusive. Similar observations were made in Cambridge at the same time. Professor