336 Prof. John Milne on Becent Advances in Seismology. [Feb. 12, 



station is crossed by a steej) barometric gradient, whether the local 

 barometer is high or low. Tremors may be marked during a calm, 

 whilst during a gale, when doors and windows rattle, a tromometer 

 may be at rest. They are frequently observed during a frost or thaw, 

 and they are generally frequent when the temperature is falling and 

 when it is low. Although waves beating on a coast may produce 

 fretillements upon a surface of mercury, such actions are apparently 

 unconnected with the swinging movements of tromometers. 



Because tremors are seldom observed in a very dry building or in 

 an instrument beneath a well-ventilated covering, I am inclined to 

 the opinion that many of these perplexing disturbances can be ex- 

 plained on the assumption that, from time to time, beneath cases 

 which are even air-tight a circulation of air is established. This is 

 brought about, as may be shown experimentally, either in consequence 

 of a difference in temperature in different parts of a case, or, as is 

 shown by the introduction of a desiccating agent like calcium 

 chloride, by the difference in the rate at which moisture is condensed, 

 absorbed or given off at different points within such a cover. 



Although a suggestion like this tends to destroy many of the 

 records of so-called earth tremors, and for years daily maps were 

 issued showing the microseismic activity of the Italian peninsula, we 

 are left confronted with phenomena which it is the interest of all who 

 work with instruments susceptible to these influences to understand 

 more clearly 



Most particularly we should like to know the reason of their 

 frequency at particular hours and seasons, but above all things, how 

 to avoid visitors which may accelerate or retard the swinging of a 

 pendulum, or cause inaccuracy in the weighings of the assayer. 



[J. M.] 



