90 Mv. J. A. Broun on Terrestrial Magnetism. 



in high north latitudes in December ; as we proceed south, it 

 it will be found, when we approach the tropics, that December 

 is no longer the month of minimum, but the months of Novem- 

 ber and January or February ; further south, near the magnetic 

 equator, March and October are the months of least diurnal 

 range ; still further south, near the tropic of Capricorn, April or 

 May, and September or August ; and finally, June and July in 

 high south latitudes. This curious shift of the epoch of mini- 

 mum has been hitherto unknown. 



The months of maximum oscillation are before and after the 

 summer solstice in high north latitudes ; before and after the 

 winter solstice in high south latitudes ; and in August and Ja- 

 nuary near the magnetic equator at Trevandrum. 



The amount of the daily range therefore has no relation to 

 temperature, as indeed I have already shown from the observa- 

 tions at INIakerstoun alone. 



According, then, to the conclusion that I have shown might 

 be drawn from the facts previously known, if all things were 

 equal for both hemispheres, we might expect for some one posi- 

 tion of the sun and of the needle that the sums of the positive 

 and negative forces producing the lateral movements would be 

 zero. If, however, from any cause (such as the unequal and 

 unsymmetrical distribution of magnetism on thetwo hemispheres) 

 the two laws should not be exactly similar, the result should at 

 least be a line of minimum movement. The Trevandrum results 

 will confirm this supposition. The following are the ranges of 

 the mean diurnal variation for the months when the range is a 

 maximum and minimum : — 



Thus the movements for August and January differ little from 

 those of the /7<?e needle in Europe for the same months, while 

 those for March and October are only about one-fourth of the 

 movements of the free needle in Europe in the same months. 



It will be remarked that the small range of the mean diurnal 

 oscillation for March and October might be wholly due to the 

 movement changing sign in the course of the month, without 

 any diminution of the daily range. This, however, is not the 

 ease ; the range diminishes from January to February, and from 

 February to March. The range in some days of March, it is 

 true, is frequently above 2' ; but this is evidently due to disturb- 

 ance, which is a maximum in that month. The real diminution 

 of range, however, is shown quite distinctly when we take the 



