336 PROCEEDINGS OF THE AMERICAN ACADEMY 



XXI. 



ON THE METHODS OF STUDY OF THUNDER-STORMS. 

 By W. M. Davis. 



Communicated February 10, 1886. 



During the past summer the New England Meteorological Society 

 undertook a special series of observations on thunder-storms, and while 

 preparing the blanks and studying the records there has been occasion 

 to examine similar investigations elsewhere, from which it appears 

 that there is some diversity as to matter of observation and method of 

 discussion. In order to make the contrast among the several systems 

 better understood, I may first state briefly the peculiar features of 

 thunder-storms, and then explain why a special service of numerous 

 volunteer observers is needed to detect their structure and mechanism. 



General Phenomena of Thunder-storms. — In the first place, many of 

 the thunder-storms that pass over us are so small that they easily slip 

 between the Signal Service stations, without even being heard there. 

 Such a one was the violent thunder-squall of July 21, 1885, that came 

 early from New York, and traversed New England at the rate of 

 nearly fifty miles an hour ; it crossed the Hudson about eight o'clock 

 in the morning, quite unobserved from New York City or Albany ; 

 it ran along eastward on the northern boundary of Connecticut, pass- 

 ing north of New Haven and New London, and south of Boston, 

 probably in sight from these stations, but too near their horizon to 

 attract attention ; finally, it crossed from Plymouth to Provincetown, 

 Massachusetts, and went out to sea about half-past one o'clock, and was 

 heard of no more. It would have been practically unknown, had not 

 our observers determined its path by means of their fifty odd reports. 



Larger thunder-storms stretch in long, narrow belts over a con- 

 siderable distance, advancing sideways or obliquely across the country. 

 These often enough pass over the Signal Service stations, but, even if 

 fully observed there, many of their most characteristic features would 

 remain undetermined. Besides, the observers at the signal offices 

 have much routine work to do, and cannot well be called upon to 

 perform extra and irregular duties. The considerable amount of non- 



