PROBLEMS OF TERRESTRIAL MAGNETISM 753 



On May 8, 1902, as you will recall, a great catastrophe overwhelmed 

 and annihilated the town of St. Pierre on the Island of Martinique. 

 The destructive agency was the products from an eruption of the 

 neighboring volcano, Mt. Pele". All reports agree that this eruption 

 occurred shortly before eight A. M., St. Pierre time, and you may 

 remember that the hands of the clock on the town hospital were 

 found stopped, according to Heilprin, at 7 h. 52 m. A. M. No distant 

 earthquake effects or barometric fluctuations were observed in con- 

 nection with this eruption, such as were recorded resulting from the 

 mighty eruption of Krakatoa in 1883. The Mt. Pel6 eruption left, 

 therefore, no record behind on any seismograph or barograph. 



However, coincident with this eruption a magnetic disturbance 

 set in simultaneously around the entire globe. On the diagram 

 exhibited, the disturbance, as recorded by the horizontal intensity 

 magnetographs, is shown for twenty stations encircling the entire 

 earth, some of them situated in the southern hemisphere, the 

 majority being in the northern hemisphere. It is noticed that 

 shortly before eight A. M., St. Pierre time, a sudden rise in all the 

 curves occurs, resulting in an increased intensity, on the average, 

 of about one fifteen-hundredth part of the usual value. For about 

 one and one half to two hours after the first impulse, the curves 

 progress fairly smoothly, when all at once they are broken up into 

 a system of most interesting and characteristic waves, whose cor- 

 responding features can be traced from station to station. 



- If now we determine the absolute time of beginning of the magnetic 

 disturbance at each station, we shall find that the times differ from 

 each other by quantities on the order of the error of the time deter- 

 mination, and that, hence, the magnetic disturbance traveled over the 

 whole earth with such great velocity as to make the times of beginning 

 practically the same over the whole globe. Thus, by comparing the 

 times of beginning of the magnetic observatories closest to Mt. Pele* 

 with the times obtained at the magnetic observatories halfway 

 round the globe, we shall find that they agree within one minute. The 

 mean of all the times, considering the disturbance in the three mag- 

 netic elements declination, horizontal and vertical intensity was 

 7:54.1 A.M. St. Pierre time, or practically the same as given by the 

 town clock. Since we have no means of knowing how accurately the 

 town clock kept local mean time, it is possible that the most accurate 

 determination of the time of the eruption of Mt. Pete on May 8, 1902, 

 was afforded us by this unique magnetic disturbance. 



I have called this disturbance unique for several reasons. First, it is 

 the only case at present known in which the occurrence was so 

 sharp and decisive as to lead several magneticians to suggest, inde- 

 pendently of each other, a causal connection with the volcanic erup- 

 tion. While it is quite possible that upon research it may be found 



