ratios are nearly invariable from 10 A. M. to 6 P. M., and again from 8 P. M. to 8 A. M. The easterly maximum and westerly minimum at 8 P. M. form again a marked feature.

The law governing the disturbances during a solar day is clearly shown, and is of a systematic character. The diurnal variation caused by the disturbances, if superposed on the regular diurnal variation, would produce what is known by the naine "mean diurnal variation." If we plot the disturbance curve on the same scale, or actually superpose it on the curves of the regular diurnal variation, it would hardly show to the eye, and the compound curve of the mean variation would keep within the maximum distance of a dot from the regular curve in the diagram (D). The disturbance variation has but one maximum and one minimum. Its most prominent feature is the easterly deflection at 8 o'clock (+194m) P. M. (at Toronto it is at 9 P. M.); the maximum deflection amounts at that hour to 32" of arc, and to 45" at Toronto. The greatest westerly deflection occurs at 6h (191m) A. M. and amounts to but 14" (at Toronto the hour is 8 A. M. with 6", and from a five years series of observation, with 31" of deflection). The range of the disturbance variation equals 46". From 3 in the morning till 5 in the afternoon, the mean effect of the disturbances is to deflect the north end of the magnet to the west, and during the remaining hours (principally night hours) to the east. The westerly and easterly disturbance deflections, during a day, balance within 0'02.

The annual inequality in the amplitude of the diurnal dis turbance variation cannot satisfactorily be shown on account of the short and partly interrupted series of observations.

It is my intention to continue the discussion of the observations made at the Girard College Observatory.

After the above was written No. 1185 of the Astronomische Nachrichten came to hand, containing Prof. R. Wolf's interesting results on the close connection of the variation in the frequency of the solar spots, and the corresponding inequality in the amplitude of the diurnal variation of the declination. He deduces for Munich the simple formula, 8=6'273+0'051 a, where a represents a relative number expressive of the frequency of the solar spots, directly derived from observation, and the amplitude of the diurnal variation. He found a close correspondence between these phenomena, showing the observed and computed amplitude for the Munich observations between 1835 and 1850. The average length of the solar spot period is reaffirmed to be 11.11 years 0·04 years. For Philadelphia we obtain 8 7'080+0'039a, which formula represents the observations as follows:

The correspondence between the observed diurnal amplitude and the same computed from observations of the solar spots is further exhibited in the annexed diagram (E). The dotted curve is the approximate Toronto curve from observation specially introduced to show the agreement at the epoch of the maximum in 1848. By computation from the solar spot observ. ations, the amplitude at that time would amount to 11'00, the whole range of the inequality of the diurnal variation would therefore be 11'00-746-3'54.

It is much to be desired that this interesting branch of physical enquiry be further studied, as it forms one of the links connecting terrestrial with cosmical phenomena.

ART. V.-A Visual Method of effecting a Precise Automatic Comparison of Time between distant stations; by JONATHAN HOMER LANE. (With a plate.)

THE visual apparatus of which I here give a general description was invented several years ago, and is intended to supply the place, under certain circumstances, of the electric telegraph, in the determination of differences of longitude. Although the wires of the electric telegraph, when suspended in the air, appear to leave nothing to be desired, at least for distances of a few hundred miles, in situations where they are available, yet it has appeared to me that the visual method I propose may prove useful in many cases where the stations to be compared, particularly the astronomical stations in a trigonometrical survey, are removed to considerable distances from lines of electric telegraph. In those situations, also, where submarine or subterranean lines take the place of air lines, the visual method, on account of the comparatively slow velocity of the electric signal along the wires of such lines, and the open and irresolvable question whether the signal time might not be greater in one direction than the other, would be capable of furnishing a useful check upon the indications of the electric telegraph.

The general features of the method are the following:

First, an intense light shown at one station, A, and viewed at the other station, B, as a star.

Secondly, a uniform automatic movement at A, made to interrupt the light at regular minute periods of say one sixteenth of a second from the middle of one interruption to the middle of the

next.

Thirdly, a uniform automatic movement at B, by which the star of light is made optically to travel around a circle at the rate of one revolution in the same period of one sixteenth of a second, the consequence of which is that the periodical interruption becomes visible to the eye as a break in the luminous circle produced by the motion of the star, and according as such break is seen upon one part or another of the circumference of the luminous circle, the relation of the movement at A to that at B becomes, inside of the recurring period of one sixteenth of a second, known in like manner as if they had been connected by a coupling shaft extended from one to the other.

Fourthly, a supplementary flash of the light at A, occurring at each whole second and during the interval of one of the interruptions before mentioned, which supplementary flash finds itself subject, in the novement at B, to an action by which it would be carried optically around a circle once only in one whole second, and by the position at which it occurs on the circumference of that circle indicates the sixteenths of the whole second. Also, a still further signal, by which account may be taken of the whole seconds, in ways too obvious to require special notice, and which will complete the knowledge of the relation of the movements at A and B to each other, or of the quantity by which one may be in advance of the other in its motion.

Fifthly, any of the known methods of effecting automatic comparison of the movement at either station with the clock at that station, by which means the comparison of the clocks at the two stations will be made to the minutest fraction of a second. Or, a single clock may be used, say at A, and any observation at B can by known methods be automatically referred to the movement at that station, and thus compared at once with the clock at A.

Further, if the automatic movement at B, besides giving optical motion to the star of light shown from A, is simultaneously made to produce periodical interruption of another intense light shown at B and seen at a third station, C, provided with a movement like that at B, the comparison from A to B may be extended directly onward from B to C, and from C onward to a fourth station, and so on, and such is the degree of precision which seems, so far as we can judge without direct experiment, attainable in the comparison between contiguous stations, that the probable error of a single comparison between the extremes of a line of twenty stations may, I believe, be made smaller than the hundredth part of a second.

An arrangement suitable for carrying the above plan into effect is illustrated by a sketch in the accompanying plate. At the station A, the rays of the signal light, diverging from their source L (Plate II, fig. 1, a), are converged by a lens C so as to meet and cross each other in the focus of a telescope. Diverg ing from this point they traverse the object glass Ó, and issue from it in a nearly parallel beam directed to the station B. The light thus transmitted is periodically intercepted at the focus F by the projecting teeth of a rotating disc, or signal wheel, S, which is made to rotate uniformly at the rate of one revolution in about one second.

The telescope used at the station B for viewing the light shown at A, is furnished with a terrestrial eye-piece, and at that point where occurs the first image of the object glass, or crossing of all pencils of light that can pass through the telescope, is introduced a refracting glass prisin P, shown in section in Fig. 1, b. The pencil formed by the light from A, on traversing the refracting prism, is turned aside so that the star image, which otherwise would be formed at s, is formed at s', and the displacement is observed by the eye at E. Since this displacement is always in a plane at right angles to the edge of the refracting angle of the prism, if the latter be made to rotate on an axis zz, parallel to the axis of the telescope, the displaced star image s' will travel in a circle around s as a center. If the period of revolution be shorter than the duration of the luminous impression on the eye, and the light be unintermitting, the circle described by s' will appear to the eye as a continuous circle of light. The period of one sixteenth of a second may perhaps be taken as sufficiently small for continuity of luminous impression. Accordingly, if the prism be made to revolve about sixteen times per second, and precisely sixteen times during each revo lution of the signal wheel S, and if the primary division of the latter be made by sixteen equidistant slotted openings in its border, then the luminous circle, which but for the interposition of the signal wheel would appear continuous and entire, will be seen in part obliterated, as shown in Fig. 3, the luminous part SDS' having the same ratio to the obliterated part S'G S', that the width of one of the slotted openings of the signal wheel has to the width of a tooth. And as the luminous arc S'DS' appears on one part or another of the circumference, its angle of position, which may be observed by bringing the wire of a position micrometer into coincidence with the extremities of the arc, will determine the angle of position at which the prism arrives simultaneously with the arrival of the signal wheel at a given point of reference.

But for counting the whole sixteenths of a second it will be required to know also the simultaneous angle of position of a

wheel geared with the prism P so as to revolve only once in the time of one revolution of the signal wheel. The circular prismatic piece is of a diameter many times greater than that of the pencil of light, and the latter traverses the former at its border. Around the prismatic piece is fastened a toothed ring, into which gears the driving wheel H from which the prism takes its motion. Between this wheel H and the regulator from which it derives its uniform motion, is interposed a satellite wheel arrangement, by means of which the observer, without disturbing the invariable velocity of the regulator, can set the wheel H, and the parts to which it gives motion, forward or back in their course, and then allow them to proceed at once with the same correct velocity as before. In this way the observer will have absolute control of the angle of position of the luminous are S'DS', and it may be agreed that as this angle of position slowly changes in consequence of the want of perfect unison between the movements at the two stations, he shall, from time to time, bring the luminous arc back to near coincidence with a standard position, that for instance which is shown in Fig. 3, the exact angle of position to be measured, however, in the manner above mentioned. Provided, then, the arc be not allowed to stray far from its standard position, it will be obvious that one part of the border of the prismatic piece P will never be traversed by the light which passes the sixteen primary openings of the signal wheel and forms that arc. The part thus unused is made with parallel faces, as shown in the figure, and then any supplementary flash of light occurring midway between the primary ones, will pass through the parallel part of P unrefracted, and may be refracted by a second prism P', that moment interposed. This second prism is made to revolve once in the period of the sixteen revolutions of P, and in the best mode of construction that occurs to me is one of sixteen prismatic pieces P', P', P', &c., so cut out, and attached to the border of a wheel or disc K, made to revolve in that period, that the edges of the refracting angles of all of them shall be parallel to each other, the whole forming the equivalent of a single prism cut into a large toothed wheel. This wheel K, like P, takes its motion from H, and is so geared that during all the intervals of time in which a passing pencil would encounter the refracting part of P, it will have free passage through one of the spaces between the pieces P', P', P', &c., which, during the alternate intervals of time, will in their turn be interposed in the path of the pencil. Any flash of light, therefore, that escapes through any supplementary opening, as t, in the middle of one of the sixteen primary teeth of the signal wheel, will, in traversing the telescope at B, be refracted by one of the prisms P' alone, and not by P. And if it be recollected. that the several prisms P' are in effect parts of one prism, as dis