*

between the two D lines, there was a faint but broad nebulous band, barely resolvable into lines of still smaller magnitude. It is impossible to represent this band accurately with a woodcut; and the shaded broad band marked on the right-hand side of Fig. 4 only serves to indicate its position and approximate breadth.

The 26th of December was also a warm day for the season, with a brilliant sun. At one o'clock, P. M., the dry-bulb thermometer marked 55°, the wet-bulb 50°, and hence the amount of moisture in the atmosphere was 3.76 grains per cubic foot. The appearance of the D line at this time is shown in Fig. 3. Two of the lines, 7 and 8, and the nebulous band x, seen on the 17th of November, were invisible, and moreover the group of three lines on the left-hand side of the figure were only just within the limits of visibility.

On the 25th of December only two lines were visible within the D line, marked a and 8, in Fig. 2, and the last of these was quite faint. The temperature at the time of observation was 46°; the wet-bulb thermometer indicated 40°, and the amount of moisture in the air was 2.42 grains per cubic foot. The sky was clear and the sun brilliant. Lastly, on January 5th, 1866, one of the clear cold days which are so common in our climate during the winter, only the single line a was visible within the D line, as is shown in Fig. 1. At the time of observation, near noon, the dry-bulb thermometer marked 10°, the wetbulb 9°, and hence the amount of moisture in the atmosphere was only 0.81 of a grain per cubic foot. The sun, however, was as brilliant as in either of the previous cases. The D line also appeared as in Fig. 1 on the 8th of January, 1866, when the thermometer at noon stood at 10° below zero Farenheit, and when the barometer attained the unexampled height of 31 inches.

The above figures have been drawn so as to show, as nearly as possible, the relative intensity of the different lines under different atmospheric conditions. As no accurate means of making the comparison are yet known, I was obliged to depend upon my eye alone, and small differences at different times of observation may easily have escaped my notice. Indeed, I should have been liable to great error, were it not for the fact that one of the lines within the D line, marked a in all the figures, does not vary in intensity, and served as a constant standard in making the observations. This is the only line which is given

We use this word in the same sense in which it is used by astronomers with reference to the fixed stars.

by Kirchoff in his chart of the solar spectrum between the two D lines, and it is referred by him to the Nickel vapor, as the D lines themselves are to the Sodium vapor, in the sun's atmosphere. It is an undoubted solar line, and has been drawn with the same strength in all the figures in order to show that it is invariable.

With a very dry atmosphere the line a is the only one which appears within the D lines, as shown in Fig. 1. With a slightly greater amount of vapor the line ẞ makes its appearance. As the amount of vapor continues to increase, this line becomes more and more prominent, until at last, as shown in Fig. 4, it is even more intense than the line a. A careful comparison of these two lines might indeed serve as an approximate measure of the amount of vapor in the atmosphere; and a series of comparisons made under the same conditions at different heights would give data for determining the law according to which the amount of vapor decreases with the elevation above the sea level.

All the aqueous lines change in intensity like the line 8. They are first seen very faintly when the amount of vapor in the air reaches a definite point, varying for the different lines, and gradually gain in intensity as the amount of vapor increases. Thus the group of three lines de do not appear in Fig. 2, are barely visible in Fig. 3, but become very marked in Fig. 4.* The lines 7 and 6 and the nebulous band κ do not appear until the air is very moist; and even when it contains 6.57 grains of vapor per cubic foot, they are still very faint. Under yet more unusual atmospheric conditions they will undoubtedly become more intense, and we shall then probably be able to completely resolve the nebulous band and count the lines of which it consists.

η

K

It is hardly necessary to repeat, that the examples here given are selected from a large number of observations. During the cold dry weather of winter the appearance of the D line is uniformly as shown in Fig. 1, the line 8 only occasionally appearing when the atmosphere becomes more moist. During the warm weather of summer, when the absolute amount of moisture in the air is in almost all cases greater than in winter, the appearance of the D line is as uniformly that shown in Fig. 3. It is only very rarely in the dry climate of New England, even during the summer, that all the lines shown in Fig. 4

* With an increasing quantity of vapor in the atmosphere the line y of Fig. 3 γ is seen before the group of lines de C, and an intermediate figure between 2 and 3 might be given showing only the lines D1 a y ẞ D2.

are visible; and, as already stated, I never before saw them so sharply defined as on the 17th of November last.

Several conditions must evidently concur in order that the aqueous lines should be developed in their greatest intensity. In the first place, the air must be charged with vapor not only near the surface of the earth, but also through a great height of the atmosphere. Local causes might greatly increase the amount of moisture in the lower strata of the atmosphere, and affect powerfully the hygrometer, which would not, to the same extent at least, influence the indications of the spectroscope. In the second place, other things being equal, the intensity of the aqueous lines must be strengthened by increasing the length of the path of the sun's rays through the atmosphere, and this is the longer the lower the altitude of the sun. But then, again, the intensity of the light has such an important influence on the definition of the lines, and the slightest haze in the atmosphere so greatly impairs their distinctness, that I have generally found that the aqueous lines are seen best when the sun is near the meridian. Hence, with an equal amount of moisture in the atmosphere, the late autumn may be a more favorable season for seeing the aqueous lines than the summer; for then not only must the solar rays, when most brilliant at noon, traverse a greater extent of air, but, moreover, the atmosphere at this time is usually clearer, and the reflected beam of light which enters the spectroscope is at times more brilliant than when the sun attains a higher elevation and the light is reflected under less favorable conditions.

In the examples cited above, the comparisons were made under as nearly as possible the same conditions, so as to eliminate all causes of variation except the one under consideration. Days were selected when the atmosphere was perfectly clear, and the sun's light, so far as I could judge, equally brilliant. Moreover, the position of the spectroscope and mirror remained unchanged during the whole time. This mirror, which is used for reflecting the sun's light upon the slit of the spectroscope, is so arranged that it can be turned into any position by the observer while his eye is at the eye-piece of the spectroscope, and it was always carefully adjusted at each observation to the position of best definition. The manipulation of the mirror is fully as important in the use of the spectroscope as it is in microscopy.

It will be of course understood that the power of developing these faint aqueous lines depends very greatly on the optical capabilities of the spectroscope, and that the figures here given are relative to the

*

instrument used in the observations. This instrument. has been fully described in the article already cited. It is sufficient for the present purpose to state that it is provided with nine flint-glass prisms of 45° refracting angle, which bend the rays of light corresponding to the D line through an angle of 267° 37' 50", and that corresponding to the H1 line through an angle of 280° 42′ 20′′, when each passes through the prisms at the angle of minimum deviation. The dispersive power of the instrument for these two rays is therefore equal to 13° 4′ 30′′, and the rays corresponding to the two D lines are separated 1' 10". The object-glasses of the two telescopes of this spectroscope are 2 inches in diameter, and have a focal length of 15 inches, and lastly the size of the prisms, and of the various parts of the instrument, is adapted to these dimensions. With a more powerful instrument a larger number of aqueous lines would be seen under the same atmospheric conditions. The Cambridge instrument has a set of sulphide of carbon prisms which disperse the light nearly twice as much as the flint prisms. The sulphide of carbon prisms are very variable in their action; but, under the best conditions, they might show the D line as in Fig. 3, when with the flint prisms it would appear as in Fig. 2.

The facts stated in this paper fully account for the discrepancies in the representations which different observers have given of the D line. Some time since, Mr. Gassiot, of London, gave in the Chemical News a representation of the D line as seen with his instrument, showing several lines in addition to those seen by myself and other observers. On visiting the Kew observatory, in the summer of 1864, I was surprised to find that this instrument was less powerful than the one I was then using; and I also learned that these lines were only seen on a single occasion. The moist climate of England is the evident explanation of the additional lines.

As I stated at the first of this paper, the D line has been selected simply to illustrate a general truth. The development of aqueous lines in contiguous portions of the spectrum is even more marked than in the exceedingly limited portion here represented. Indeed, as has been already intimated, the number of these lines seen in the yellow region of the spectrum, on the 17th of November, was at least ten

*These prisms were furnished by the American Academy from the income of the Rumford Fund, appropriated for investigations on light and heat. See Proceedings of the American Academy, Annual Meeting, May 24th, 1864.