TRANSACTIONS OF THE SECTIONS. " 27 



standard barometer before begiuuino: our observations in order to ascertain its 

 index error, and if we then gradually increase tlie jjressiu'e, using- the above index 

 error, we shall find that the instrument lags behind, that is to say, reads too low 

 up to 24 inches, at which point its behaviour' appears to be reversed, and it there- 

 after rises too fast. 



2. The instrument is, however, tolerably accurate up to 25 inches, or through a 

 range of 6 inches. 



3" If we compare our aneroid -^-ith a standard at the end instead of the begin- 

 ning of the observations, we shall get much less reliable results. 



So much for the double experiment, in which the pressure is first lowered and 

 then raised. 



Now, if at the end of this experiment we compare our aneroid with a standard 

 once more at the ordinary pressure, we shall find that, on the whole, its indications 

 have fallen, or it reads too low, but graduallv, and in com-se of time, it recovers 

 itself. 



This is seen by the following instances : — 



2-inch aneroid. 

 Error before experiment + '47 

 Immediatelv after „ -\--li) 

 2:3 hours after „ + -34 



40 „ „ -t--37 



2 f -inch aneroid. 

 Error before experiment -|- •!! 



45-inch aneroid. 



EiTor before experiment -I- -04 



Immediately after „ — -OG 



1 hour after „ — ■0.'3 



18 hours after „ — -01 



3 days after „ -|- '01 



3 weeks after „ -|- -07 



Immediately after „ + '03 



18 hours after ,, -|- '10 



In the next place, I would remark that large aneroids are better than small ones, 

 as will be seen by the following Table denoting the average behaviour of small 

 and lai'ge instrimieuts for the down observation. 



Error Error 



Pressure of large. of small. 



30 -00 -00 



29 -I- 04 -I- -04 



28 -I- -04 +-02 



27 -f02 -00 



2G -H-Ol -'02 



25 -00 --06 



24 : .... --02 --07 



23 -'04 --11 



The experiments are not yet quite finished, but we may perhaps conclude — 



1. That if a good 4-inch aneroid be first of all comi)ared with a standard baro- 

 meter, and then gradually subjected to a decrease of pressure, it will give reliable 

 results through a range of 6 inches. 



2. That if a good 4-iuch aneroid be first compared with a standard barometer 

 at a low pres.?m-e, and then gradually subjected to an increase of pressure, it will 

 give reliable results through a range of 6 inches, starting from the low pressure. 



3. The results would probably be still better if the instrimient, before use, were 

 compared with ?. standard barometer after the manner I have now described. 



Storm-Warnings, their Importance and Practicahility. 

 By Colonel Stkes, M.R, F.E.S. 

 The author adduced the testimony of nmuerous men of scientific eminence, and 

 the Keports from the seaports to the great importance of the signals lately in prac- 

 tice at the Meteorological Department of the Board of Trade, both from humane 

 and commercial points of view, and then stated that out of 405 warnings given in 

 three years, the prognostications were correct for 305 times. No one could tell the 

 possible number of li\-es and amount of property which had thus been saved ; and 

 he asked if this did not sufliciently justify the continuance of these storm-v.'arn- 



