258 On the Heights of some of the Cotswold Hills. 



height of each of which is the result of a single observation with 

 the aneroid, the agreement of its indications with those of the Ord- 

 nance determinations is very remarkable, considering the errors 

 to which the present construction of that instrument render it 

 liable. I must observe, however, that they are brought nearer to 

 the trigonometrical measurements by my having rejected some 

 of my first attempts, in which I am almost certain that I made 

 mistakes, and by subsequently adopting the mean of two or three 

 observations, a process which always reduces the extremes of 

 error. Thus for Painswick Hill I had three observations — 



One giving it . . . 919 feet. 



Another 934*8 „ 



The third .... 954 „ 



giving a mean result of 935*9 feet, which differs only 6*9 feet 

 from Capt. Yolland's figures. 



I am sorry that I have not had time to try more of our heights ; 

 but I thought it better to repeat the observations on the same hills 

 in order to obtain mean results, and thus to sift my own probable 

 errors, than to persevere in them undetected. 



Throwing out of consideration, then, some of my first trials, 

 before I was quite up to the use of the instrument and its tables, 

 the results which I have just given are highly satisfactory. But 

 on the other hand I tried it against the published sections of the 

 Cheltenham and Great Western Railway with less success, as the 

 following comparison will show : — 



By Company's By Aneroid. Error, 

 sections. 



feet. feet. feet. 



Stroud station above Gloucester station 116*3 12475 +8'45 



Summit-level at top of Saperton tun- \ ^kq-o 41 3-7 4-fil -S 



nel above Gloucester J . ' ' } 



Now these were the means of two trials; in the latter case 

 the discrepancy is greater than I can easily explain, unless the 

 oscillations of the railway carriage have any effect on the instru- 

 ment, which I can hardly suspect ; for in all other cases, however 

 carefully carried, it must have been exposed to rough shaking*. 

 On the whole therefore I must suspend my opinion as to the 

 merits of the aneroid for measuring heights till after further ex- 

 periments, and at any rate would recommend the improvements 

 in the construction, to which I have before alluded, to be effected, 

 viz. the decimal graduation to be adopted, and the index to be 

 placed closer to the face of the instrument. 



P.S. Since the compilation of the above paper I have been 



* The error may be this — that the Company's sections were published 

 before the completion of their line, which was eventually carried at a rather 

 higher level than these sections show. 



