Let us apply these principles to the meteor to which the above observations refer. The time of appearance at New Haven was only approximate, and the other observations indicate 1h 16m 20s as the correct New Haven time. We saw it pass down nearly vertically 2° north of Jupiter.

At Haverford it appeared, as the diagram shows, above and probably to the east of Polaris, and passed a little below (say 3 or 4 below), x and 8 Cassiopeia. The general direction of its observed paths, the character and colors of the train, and the small number of the unconformable meteors, mark this as an undoubted member of the November group, and authorize us to assume that its direction was from the radiant in Leo. This assumption, and the distance above named from 8 and Cassiopeia, imply that the path passed 10° or 12° above Polaris.

The differences of latitude and longitude of New Haven and Haverford are 77"-95, and 142"-83. The latter number multiplied by the cosine of the middle latitude gives the departure, which, divided by the difference of latitude, gives the tangent of the azimuth of the line joining Haverford and New Haven upon the horizon of the point midway between the two places. The secant of this angle, multiplied by the difference of latitude, gives the distance in geographical miles. This, multiplied by 115, gives the distance, 153.5 statute miles, with a direction S. 54° 16' W.

Upon the diagram in plate II, I have marked by a and c the points 54° 16' from the south and north points on the horizon. Through a and c and the point b of the heavens (2° north of Jupiter), let there pass a great circle cutting the Haverford track in d. If we imagine a triangle, whose angular points are the observers at New Haven and Haverford, and that point of the track which was directly opposite Jupiter, as seen from New Haven, and produce the sides of this triangle, then will the three arcs, ab, bd, and de, severally measure the three angles of the triangle. We have, then, the angles and one side of a triangle to find the distances of this point of the meteor's path from the two observers.

The arcs of the semicircumference ac can be measured on the globe by a movable quadrant, or by bringing the circle abde to coincide with the horizon of the globe. I have, however, found a narrow tape line, graduated for the purpose, very convenient for such measuring.

Knowing the distance of either observer from the particular point of the meteor's path, we measure on the globe the zenith distance of the corresponding observed point, b or d, and hence readily compute the actual altitude of the meteor, at that in

stant, above the earth's surface. An allowance for the curvature of the earth is to be added to the height above the horizon of the observer. The azimuth of b, or d, when measured, gives the means of locating the place over which this point of the meteor's path was vertical.

The arcs, ab and bd, measured upon the globe, are about 281 and 57°, which give the distances of the track from New Haven and Haverford, 182, and 87-3 miles, respectively. But, owing to the fact that the arc, abd, cuts the Haverford track at a sharp angle, we seek a closer determination of the path by another method.

The position of Jupiter with reference to the track and cloud makes it worth while to compute its altitude and azimuth from New Haven at 1h 16m 20s, and 1h 44m. These are found to be S. 75° 16′ W., and S. 80° 8' W., for the azimuth, and 16° 48', and 11° 43' for the altitudes.

Now, the observed path should be in the arc of a great circle which passes through the radiant. By bringing the radiant and the point 2° N. of Jupiter to the horizon of the celestial globe, the zenith of New Haven will be found to be 5° above the horizon. Hence, the meteor passed (though then unseen) 5° S. of our zenith.

Again, the point in the direction S. 77° 16' W., 182 miles from New Haven on a line elevated 16° 48', is over latitude 40° 42', longitude 76° 9', and 56 miles high. The elevation of the center of the radiant area above the horizon of this point was about 20°. Hence we may carry back the meteor's track with an elevation of 20° till it cuts the meridian of New Haven. We see, thus, that the meteor passed at an elevation of 120 miles and 10.5 miles south of us.

This gives us a good determination of the vertical plane in which the meteor was moving, the accuracy of the determination being due to the nearly vertical path of the meteor as seen at New Haven, and the easy reference of it to Jupiter. But the New Haven observations are not suited to the determination of the altitude of the track in the plane. This may also be said of all the observations at Williamstown, Palisades, and Poughkeepsie.

The distance from Haverford to this vertical plane, measured on the chart in the direction of Wilkesbarre, or N. 20° W., is 55 miles. The observed altitude in the same direction is 48, which gives an altitude of 61 miles for the meteor at that point. The distance from Wilkesbarre to the plane is 35 miles in the direction S. 20° E. An altitude of 61 miles represents an angular altitude of 60°. The star Orionis happened to be in this same azimuth, and had an altitude of 57°.

This satisfies well enough the observations, for it is reasonable to assume that near λ Orionis" means a little above it, since, if the meteor had gone below that star, the reference would have probably been to the brighter stars of the constellation. We may, then, fairly conclude that the meteor passed the vertical plane through Haverford and Wilkesbarre at an altitude of 61 miles. This is 5 miles lower than the previous determination, for the former path is 66 miles high at this point.

The beginning and end of the apparent path of the meteor was not well observed. The middle point of the cloud, in its initial position, as seen at New Haven, was nearly opposite to Jupiter. The line to this point, and the track of the meteor, make an angle of 39°, which is determined by measuring upon the globe the distance from 20° N. of Jupiter to the radiant, and taking the supplement of the result. These lines meet in long. 76° Ö', and lat. 40° 43', at an elevation of 54 miles, and at a distance from New Haven of 175 miles.

Assuming that the apparent length was 6° at New Haven, we have an actual length of 175 sin 6° cosec 39°, or 29 miles. This represents an apparent length at Haverford of over 20°, instead of 15° as reported. If these were the only means of determining the length, the Haverford observation should be regarded as the best. But at Washington, the original length must have been appreciably more than 7°, since the train is said to have decreased in length to 7°. Each degree at Washington represents 3.2 miles on the track of the meteor. A length of only 9° for the observed track at Washington would be represented by the 29 miles of the actual path. We shall assume, therefore, that the length of the cloud was, at first, 30 miles, which implies that its eastern end was 59 miles, and its western end 49 miles high.

At what altitude the meteor was first seen, we cannot say. It passed the meridian of Haverford at a height of about 68 miles, but may have been visible before it reached that point. There is no reason to suppose that it passed beyond the western end of the cloud.

The train had, an appreciable breadth at once upon the disappearance of the meteor. Probably 20' would not be a large estimate for the breadth, as seen at New Haven, which implies a real diameter of one mile, and a volume of a dozen or a score of cubic miles. This volume was steadily and rapidly increasing through the whole time that the train remained visible.

Compare now this path with the observations at Washington, Palisades, Poughkeepsie, and Williamstown.

At Washington.-I have assumed that the meteor described by Prof. Eastman was unquestionably the same body as that ob

served at New Haven and other places. An error of a half hour in recording the time of appearance would make the two coincident in time. The Washington meteor was "the most remarkable meteor of the shower," its train exhibited one of the two remarkable apertures described by Prof. Gummere, its position was very near that in which the one we have described would have been seen, and its train endured for a like unexampled period. Such a meteor as that described by Prof. Eastman would have certainly attracted attention at other stations. Prof. Eastman says that it appeared "near & Ursa Majoris, and moved west across 8 Ursa Minoris." This path to be conformable to the radiant in Leo needs a change of direction. Looking from Washington at the path determined above from the Haverford and New Haven observations, the path would be seen crossing the line joining the two stars just mentioned. Only so much change of the path observed is needed as is necessary to make it conformable to the Leo radiant. The assigned path satisfies reasonably the Washington observations.

At Palisades.-The path of the meteor as above determined passes very nearly vertically over Palisades at an elevation of 95 miles. The lower end of the cloud was 22° high and 5° or 6° north of Jupiter. In one of his notes Mr. Gilman says that the track ended near a Piscium, an error, probably for 41 Piscium. The assigned length of path 50° or 60°, would imply a first altitude of 85, or 90 miles. This estimate, however, is not to be taken strictly.

α

At Poughkeepsie.-The beginning and end of the cloud as before determined would be viewed from Poughkeepsie in the direction S. 56° W., alt. 26°, and S. 59° W., alt. 19°. These are about 18° and 15° south of Jupiter, and a little higher from the horizon. The track would if produced backward pass 25° south of the zenith of Poughkeepsie. A sketch of the position of the track with reference to Jupiter, sent by Miss Mitchell, would imply that the meteor's path was nearer to the planet, than 15° or 18°. But the observations at other places forbid any essential diminution of this distance.

At Williamstown.-As seen from Williamstown the direction and altitude of the two ends of the cloud are S. 43° W., 16°, and S. 46° W., 12°, which makes the path about 30° south of Jupiter, much farther than would be implied by the sketch of the track. The meteor when due south would have been (if visible) 45° high. No essential transfer of the track toward Jupiter can be made, however, without rejecting either the New Haven or the Haverford observations altogether.

The position and direction of the cloud cannot therefore, I think, be reasonably changed from the previous determination.

Its central point may be regarded as 54 miles high, over N. lat. 40° 43', and W. lon. 76°, and its course S. 78° W., with an angle of depression of 20° upon the horizon of the places beneath it. The heights of its eastern and western ends were 59 and 49 miles or 95 and 79 kilometers.

Motion of the train of the meteor.

The cloud left by the meteor began to coil up, and float away, shortly after the disappearance of the body. The upper part must have moved northerly, and the lower portion southerly, and the whole either westward, or downward, to satisfy the appearances at New Haven. The successive positions with respect to Jupiter show this. Jupiter moved north 10' per minute, and downward 11' per minute during the time of observation. The motion of the lower part must have been in the main toward Washington, as after its first shortening the cloud retained its dimensions nearly constant. The motion was too nearly in the plain of Haverford and New Haven for us to entirely separate the downward and the westward motions, especially in the absence of determinate positions at specified times. But I think there was a descent of five or ten miles in the body of the cloud. This was certainly true of the eastern or upper end, and even a greater downward motion is prob

able.

The upper end was borne westward so that as seen from Washington the cloud was foreshortened. The central and lower parts were seen projected upon this upper part. Hence the bifurcation and the cavity described by Prof. Eastman. The elliptic form and the oval cavities seen at Haverford are also, in part at least, the effect of perspective.

The southward motion of the lower end at Haverford, 80 miles from the cloud, was 25°, and at New Haven, 175 miles distant, was 8° or 10° during two thirds of the period of visibility. These imply a motion of at least 40 miles, or about a mile per minute.

We must therefore assume that just below an elevation of about 50 miles, there was a rapid north wind, which swept the lower portion of this cloud with it southward. The wind may have come from one or two points east of north. Above this there was a south wind, (or S. S. E.,) whose velocity may have equalled that of the lower north wind, though it may also have been much less.

By its downward motion, the eastern end of the cloud was carried from the upper current into the lower, and strewn along into the horizontal cloud, seen at New Haven, in the latter part of the period of visibility.