292 Epidemics. 



from Dr. Short's work on the " Comparative History of the 

 Increase and Decrease of Mankind in England, with a Meteo- 

 rological Discourse," 1767, and many recent authorities to 

 show that changes in temperature are neither gradual nor 

 equal, year by year, for the same locality. Here also consult 

 the meteorological charts of Greenwich and Kew, or Paris 

 and Berlin, when it will be found that thermometrical varia- 

 tions of 30 degrees are not very infrequent within 48 hours, 

 for like times of the day or the night. 



It is therefore clear that the mere surface of the earth, 

 with its equal or nearly equal solar rays, cannot at all affect 

 the equilibrium of temperature, by any sudden and violent 

 changes within two or three days. Again, as all our storms 

 and rain-falls are equally regulated by temperature, it 

 follows that no surface change in electricity or magnetism can 

 possibly accomplish such sudden and important variations. 



Electric and magnetic variations, that are not sub-alluvial 

 in their origin, must be accomplished by reciprocity of 

 changes going on between lunar and solar rays, animal and 

 vegetable organisms, and the general surface or stratified 

 rock superficies of the earth ; but the equal, gradual, and 

 uniform nature of each, cateris paribus, is such as to totally 

 exclude the earth's surface as the great regulator of the 

 sudden magnetic or electro-thermal variations now known to 

 be perpetually occurring, and we must, as a sequence, admit 

 a central or internal source as a regulator of the magnetic 

 and thermo-electric conditions apparent on the surface of 

 the earth. 



These considerations lead to the inferences arrived at by 

 Dr. Gilbert, 1600, and Halley at a somewhat later period 

 the former of whom judged that, towards the earth's centre, 

 there was a large magnetic mass, and the latter that the 

 earth consisted of an inner and outer magnetic shell, 

 revolving one within the other, but at very nearly equal 



