204 Dr. Bostock's Observations on [March, 



and from which I transcribed the account of the barometer for 

 the months of January and February, 1807. These I have com- 

 pared with the observations taken at the Royal Society's apart- 

 ments during the same period, and likewise with a register of 

 the barometer published by Mr. Bent, of Paternoster-row. 

 The result of this comparison, which is inserted in No. 5, is 

 similar to the former ; although the instruments must have been 

 nearly on a level, the absolute heights of the mercury are con- 

 siderably different, and the proportion between the heights of 

 the different instruments is not uniform. 



From the above statements I think we cannot hesitate to 

 admit that the barometer, as it is commonly constructed, is not 

 an instrument which can be depended upon for making perfectly 

 accurate observations on the weight of the atmosphere, even 

 when it is in the hands of those the best qualified to judge of 

 its condition, and where it has remained stationary, and has not 

 been exposed to any circumstances likely to injure or derange 

 it. And it is the more objectionable, because the imperfection 

 of the instrument is not visible to the eye, and can in fact only 

 be discovered by comparing it with another instrument which is 

 supposed to be more perfect. 



I shall not now attempt to point out all the sources of inac- 

 curacy which may attach to the barometer as it is usually 

 constructed ; but the following may be enumerated as among 

 those which are probably of the most frequent occurrence : 

 1. The barometers being exposed to various degrees of tempera- 

 ture, and no notice being taken of this circumstance, or no 

 correction being made for it. 2. The index not being accu- 

 rately adapted to the height of the column of mercury : this 

 appears to have been the case in almost every one of the baro- 

 meters of which the register has been given above ; for in no 

 instance does the average height of any of them correspond 

 with the absolute elevation at which it is placed, and in some 

 cases they are in direct opposition to it.* 3. The reservoir in 

 which the tube terminates not being regularly cylindrical, or of 

 the same diameter in its different parts ; so that there is not the 

 same ratio preserved between the reservoir and the tube in the 

 different states of the instrument. These two latter imperfec- 

 tions are noticed by the Rev. Mr. Wollaston in his" account of 

 the thermometrical barometer, as occurring in the instruments 

 upon which he made his comparative experiments, and being 

 detected in them by means of his newly-invented apparatus .f 



* Although the altitude of some of the barometers is given with apparent accu- 

 racy, yet it is not always easy to ascertain their exact height. The barometer at 

 Somerset-house is stated to be placed at "81 feet above the level of the low water 

 spring tides." Mr. Banks informs us that his barometer is " 104 feet above the 

 level of the sea," and Mr. Bent's is said to be hung " 51 feet higher than the bed 

 of the River Thames." I conceive that all these expressions are rather vague, and 

 that it will not be easv to determine the relation which they bear to each other. 



+ Phil. Trans. 1817. 



