4 NINTH REPORT — 1839. 



appears to me superior in delicacy of representation, conveying 

 by shading (which is by no meaiis so good in Schumacher's 

 print) an idea of the relative intensity of the different parts of 

 the spectrum. Both preserve admirably the varied characters 

 of the several groups of lines, and present a faithful picture of 

 the actual object. All this, however, is almost entirely lost in 

 the plate in the Edinburgh Encyclopaedia, the execution of 

 which is coarse, and the characters of the different bands ill 

 preserved, especially at G ; while two small groups of lines be- 

 tween and below the bands at H, are made so conspicuous as to 

 be mistaken for them. In the plate in the Edinburgh Philoso- 

 phical Journal (No. 18.), H is distinctly marked at the point 

 midway between the two bands, instead of being opposite the 

 lower. 



In the plates in the Munich Transactions, and in Schu- 

 macher's Joui-nal, the appearance of the numerous lines about 

 G is beautifully given -, and I have closely compared these re- 

 presentations with the actual object, both as seen in the small 

 telescope of my apparatus, and also in one with a power of 20. 

 With this power all the smaller lines are seen as in Fraunhofer's 

 plate, but it is insufficient to resolve the two broad bands. In 

 that plate, however, they are represented as formed of masses of 

 very fine lines close together; and in the less refrangible group, 

 as nearly as possible at its centre, there is one line a little 

 stronger than the rest, opposite to which G is marked. 



Thus, the middle of the lower band, in my observations, ap- 

 pears correctly taken for the exact position of G. As some 

 guide to the appearance and position of the lines, I have an- 

 nexed a map laid down from my own observations, and giving 

 their general character as exhibited in the small telescope of my 

 apparatus (see Plate II.). 



With regard to the accuracy of the observations, and the de- 

 gree of accordance between one set and another, it should be 

 borne in mind, that the liquid prism is necessarily exposed to 

 the heating power of the sun's rays during the whole time of 

 observation. Hence, the refractive power will be liable to con- 

 tinual small changes ; for which evil no remedy seems applica- 

 ble, but that which may be supplied in multiplying observations, 

 from which, it may be presumed, the resulting mean values will 

 furnish determinations on which increasing reliance may be 

 placed. In general, the observations at higher temperatures 

 are less to be depended on. 



No comparison can be made between the refractive powers of 

 different media, until some means can be found of reducing them 

 to a common temperature. I find a proportional diminution 



