456 F. H. Bigelow — Solar and Terrestrial Magnetism. 



nent forces, it is only fair to bear in mind that we have no 

 better material than this to work with. The observations of 

 meteorology and magnetism are very hard to handle with suffi- 

 cient precision to give perfect laws, and critics should always 

 remember, (1) that laws which are even approximately made 

 out become very valuable for further advances, and (2) that 

 every criticism ought to make sure that the author's position is 

 not only understood, but also that the critic knows more about 

 the point in hand than the investigator himself. 



Taking up now the details of Bulletin No. 21, it will be seen 

 that an attempt has been made to overthrow two positions held 

 in terrestrial magnetism, (1) that the sun is not a magnetic body 

 because it is too hot, and (2) that the variations of the terres- 

 trial magnetic field can be accounted for by electric currents in 

 the cirrus cloud regions ; also a promise is made to examine the 

 meteorological problem in a forthcoming report on the Inter- 

 national cloud observations. Before recapitulating the argu- 

 ments used in sustaining these two propositions, it is to be 

 observed that the method of handling the observations is the 

 simplest possible, as explained in Bulletin No. 2, of the Weather 

 Bureau, 1892. It consists in assigning a normal field at a given 

 station, and then subtracting this from the observed field, 

 whereby a system of components is obtained which go to make 

 up the forces that deflect the actual field from its normal 

 position. Three rectangular elements are combined into the 

 corresponding vectors acting in space, at a given time, and the 

 collection of these together builds up the total vector system, 

 which can be studied by itself and on its own merits. This 

 plan has recently been advocated by von Bezold, Zur Theorie 

 des Erdmagnetismus, 1897; the coming Conference of Mag- 

 neticians to meet at Bristol, England, Sept. 7-14, 1898, will 

 discuss the same principle in different aspects; and indeed its 

 use is so obviously advantageous that what astronomers have 

 long been doing is now applied to terrestrial magnetism and 

 meteorology. By computing the vector systems belonging 

 respectively to the daily means and the hourly means, as 

 published, two different auxiliary fields have been disentangled, 

 which added to the normal field give the actual field as observed. 

 These two fields are the bases of our discussions and conclusions, 

 and they have characteristics which seem to indicate a simple 

 interpretation of the entire range of phenomena. 



The temperature argument. — It was easy to conclude that 

 because an ordinary magnet loses its magnetic force under 

 high temperature, therefore the sun must be non-magnetic ; 

 and it is so difficult to disprove this argument, that this opinion 

 is not only widespread, but tenaciously held as a canon of 

 modern science. Yet it has been treated too long a bogey to 



