16 



Description of an Automatic 



being disturbed. It was set with the standard barometer 

 (Fastre), at 4 p. m., April 24th. 

 1865. 



Date. Temperature. Standard Barometer Printing Barometer Difference. 

 Centigrade, reduced to deg. Cent, reduced to deg. Cent. 

 Apr. 24, 4h. 8.5° 29.947 in. 29.947 in. +0.000 in. 



« 25, 22 10.0 29.934 29.929 +0.005 



« 26, 2} 13.0 29.789 29.784 +0.005 



" 27, 21i i3.o 29.862 29.859 +0.003 



« 30, 12.0 29.755 29.750 +0.005 



During this period the barometer reached a maximum 

 of 30.114, and a minimum of 29.435. 



On the 30th, the machine was taken down to modify the 

 printing apparatus. 



A comparison of these results show that even with in- 

 ferior and comparatively rough apparatus, the individual 

 results are equal to a reading obtained from the best stand- 

 ard barometer. For when we consider that the error of 

 the temperature constant, for the new apparatus, may in 

 this case, even for a difference, of 5° amount to 0.005 in., 

 it would show that the errors introduced by the mechan- 

 ism are almost inappreciable. At this time the registering 

 barometer was not compensated for temperature; the 

 reductions being made from a short table computed for 

 this purpose. 



The reduction of the barometrical results to 0° Centri- 

 gracle or 32° Fahrenheit, can be accomplished by support- 

 ing the barometer tube by means of a combination of 

 brass rods ; the office of these rods being to carry the 

 whole barometer tube in an opposite direction over a space 

 equal to the motion of the mercurial column from the 

 effect of temperature. An apparatus of this kind has 

 been used in the construction of the photographic barome- 

 ter at the Eadcliffe Observatory. In the place of zinc rods 

 there used, we propose to employ brass tubes. 



The amount of compensation will be determined from 

 direct experiment, or comparison of the printed results, 



