220 



Prof. Erman — Magnetic Elements at Berlin, [April 28, 



and when brought under the form 



(A) T=179183-0-000048119{^-59-930f, 



identical with I., it shows that the horizontal intensity reached in 1859*930 

 the maximum of 1*79183. 



It ought to be observed, that equal probability has been attributed to the 

 error 



±1 in T, 



that is to say, equal errors to an intensity determined by each of the three 

 methods, — this supposition being at once the most simple and the most 

 conformable to my experience, by nearly contemporaneous repetitions of 

 each class of observation. 



All my determinations of absolute intensity have been obtained either by 

 one of two, or by^two magnetometers ; the first of which is a Gaussian of 

 large size, by Meyerstein, the second my declination- and transit-instrument 

 by Pistor, completed by the usual graduated holders for deflecting magnets, 

 and perfectly adapted to observations in the open air. 



2. Inclination, 



The values of inclination here employed are taken for 1806*0, 1832 '5, 

 and 1836*87, from the observations of Humboldt, Rudberg, and Encke; 

 for the ten other dates since 1825*0, they have been obtained by my own 

 applications of the methods exposed in my * Reise um die Erde,' Physikal. 

 Beob., tome ii. pp. 8-42, to two different instruments — viz. till 1850 to 

 a large and highly perfect one by Gambey, and since that time to a 

 smaller dip-circle by Robinson. The methods of observation leave no 

 room for any constant error in the resulting inclination, as long as no di- 

 rective magnetic force is exerted upon the needle by the instrument itself. 

 In order to free my results from any influence from this improbable (but 

 not impossible) source, I compared, in 1860, three full determinations by 

 the last-mentioned apparatus, with an equal number which I obtained 

 under identical circumstances with a most perfect copy of Weber's in- 

 ductive inclinometer. The result was an agreement of the two kinds 

 of determinations within the limits of accidental error of the first — that 

 is to say, far below one minute in the inclination ; I venture, therefore, to 

 say hat the following numbers must give the absolute value of the element 

 in question with no less certainty than the rate of its secular variation : 



