1818.] the Imperfection of the Barometer. 205 



4. The reservoir being made so small (for the purpose of render- 

 ing the instrument as portable as possible), that, in any consider- 

 able diminution of the weight of the atmosphere, there is not 

 sufficient space to hold the mercury that ought to leave the tube, 

 so that it is retained in it above its proper elevation. This seems to 

 have been the case with the barometer which I employed near 

 Liverpool, which had not the power of descending much lower 

 than 29 inches, although, in other respects, it was a well-con- 

 structed and delicate instrument. 



I do not think it necessary to notice some of the more obvious 

 causes of inaccuracy in the barometer, such as the impurity of 

 the mercury, or the presence of air and moisture in the tube, as 

 these are not likely to exist in instruments that are made with a 

 proper degree of care, or, if they did, might be easily detected. 

 There maybe other causes, besides those which I have mentioned, 

 which may affect the action of the instrument, without any 

 change being produced in its appearance, which may indicate 

 the imperfection to the observer. In order to avoid the errors 

 that may arise from the different circumstances that have been 

 pointed out, there is a simple expedient which we may adopt, 

 and which will probably be effectual, at least sufficiently so for 

 making meteorological observations. The plan that I would 

 prefer is, that the observer be furnished with two barometers, 

 which may be fixed in the same frame, to which a thermometer 

 should be also attached. Whenever the observer examines the 

 barometer, he must notice the thermometer at the same time ; 

 and, by a table ready prepared for the purpose, he must make 

 the necessary corrections for the variations of temperature. If 

 the barometers do not exactly coincide, he must endeavour to 

 ascertain which of them is imperfect, either by referring to some 

 other standard, or by comparing together a number of observa- 

 tions made under different circumstances, by which the parti- 

 cular cause of the imperfection will be probably detected. Until 

 some plan of this kind has been adopted, it would be premature 

 to enter into any speculations concerning the causes which may 

 affect the barometer, or to form any hypotheses respecting the 

 variation in the weight of the atmosphere. 



After I had completed the above remarks, I received the 

 number of the Bibliotheque Universelle for October; in which is 

 contained a letter from Dr. M'Culloch to the editors of that 

 journal, with their remarks upon the same, on the subject of tin- 

 variations of barometers. Dr. M'Culloch had noticed the same 

 differences between them that are stated in the above paper, and 

 had given an account of his observations to the Geological 

 Society, of which an abstract was published in the Annuls of 

 Philosophy, and afterwards in the Bibliotheque Universelle fur 

 September, 1816. In his present communication, Dr. M'Cul- 

 loch enters into a short detail of his opinions, as well as of the 

 facts upon which they an; founded. By comparing the registers 



