40 



81. The theory given above takes account only of the influence of 

 temperature on the mercury and glass tube. The effects that result 

 from changes in the mechanisms described later for transmitting and 

 inscribing the record, and for holding the glass barometer tube itself, 

 all require consideration, but fortunately these are in the main so 

 small, especiall}?^ when considered in relation to the highly magnified 

 scale on which the record is inscribed, that they may be neglected. 

 In any case they can be incorporated with the mercury effect so that 

 by adding or removing small amounts a certain total volume, Vq, of 

 mercury at temperature to may be employed, and thus all uniform 

 effects of temperature on the whole apparatus will be automatically 

 compensated. 



If the siphon is not compensated, then the volume of mercury at 

 temperature to = Vi, which in general will be greater than Fo, but 

 may be less, and a small correction will be required, the amount of 

 which will be simply the apparent expansion of the excess of mer- 

 cury occupying the bend and short leg of the siphon. This expan- 

 sion may be imagined simply to lift the whole column of mercury a 

 small amount, Ah. 



The volumetric expansion will be 



(F -Fo) (m-g) (t-to), 

 and the rise of mercury, Ah, is given by the expression 



^(dnel)Ah=(V^-Vo){m-g) (t-to). 



In general, d^ and c?2 will be made sensibly equal, and, in fact, ^d. 

 Hence : 



ttCL 



Let y=the amount by which the mercury in the open leg of the 

 siphon stands higher than required for compensation. Then, since 

 an equal excess of mercury occupies the opposite branch of the U, 

 we have : 



Ah = y {m-g) (t-fo), 



or, for ordinary glass : 



AA = . 0000865?/ (t-to). 



That is, the temperature correction required is simply the apparent 

 expansion of the short column y of excess of mercury. Now, the 

 temperature variations affecting a barograph throughout the period 

 of a single record sheet, or rather during intervals when the record 

 may be checked by eye observations of a standard barometer, will 

 rarely exceed 10° or 20° F. If y^l inch, for example, then Ah for 

 20° = 0.00173, a quantity which, in general, may be neglected. 



