58 PROCEEDINGS OF THE AMERICAN ACADEMY 



American Journal of Science,* has led me to the conclusion that a very 

 large number of the more faint lines of the solar spectrum, hitherto 

 known simply as air lines, are due solely to the aqueous vapor of our 

 air, and hence that the absorption of the luminous solar rays by the 

 atmosphere js at least chiefly owing to the aqueous vapor which it 

 contains. 



The appearance of the Fraunhofer's line D, seen under precisely 

 the same conditions, but with increasing quantities of aqueous vapor in 

 the atmosphere, is shown in Figures 1, 2, 3, and 4. The D line is 

 selected, because, being a favorite test object for the spectroscope, its 

 general appearance is well known to all observers. But even more 

 marked changes than those here illustrated have been noticed in other, 

 although chiefly in contiguous, portions of the solar spectrum. 



These changes attracted my attention from my earliest observations 

 with the spectroscope ; but with my first instrument, and the bisulphide 

 of carbon prisms then employed, it was almost impossible to eliminate 

 the effects which might be caused by the variations in the condition of 

 the instrument itself; and as these were known to be very great, it 

 was possible that they might account for all the variations observed. 

 With the improved instrument, however, just referred to, absolute 

 constancy of action is obtained, and all merely instrumental varia- 

 tions avoided. 



A peculiar condition of the atmosphere gave the first clew as to the 

 cause of the changes under consideration. The weather on the 17th 

 of November, 1865, at Cambridge, Massachusetts, was very unusual 

 even for that peculiar season known in New England as the Indian 

 Summer. At noon the temperature on the east side of my laboratory 

 was 70° F., while the wet-bulb thermometer indicated 66°, showing an 

 amount of moisture in the atmosphere equal to 6.57 grains per cubic 

 foot. At the same time the atmosphere was beautifully clear, and the 

 sun shone with its full splendor. I have never seen the aqueous lines 

 of the spectrum more strongly defined than they were on this day ; and 

 the total number of lines visible in the yellow portion of the spectrum 

 was at least ten times as great as are ordinarily seen. The appear- 

 ance of the D line on that day is shown in Fig. 4. Between the two 

 familiar broad lines D x and D 2 there were eight sharply defined lines of 

 unequal intensity, which is only very imperfectly represented by the 

 woodcut. In addition to these, on the more refrangible side of the space 



* American Journal of Science and Arts, Vol. XL, November, 1865. 



