66 PROCEEDINGS OF THE AMERICAN ACADEMY 



as a western group of areas, and, on the other hand, eastern and south- 

 eastern Massachusetts and Rhode Island,* as an eastern group, the 

 hours of maximum- percentages come out as follows: — 



This clearly implies that distance from some at present unknown 

 starting-point exerts a decided influence on the hour of maximum 

 storm frequency ; or, in other words, that most of our New England 

 storms come to us ready made from a source where they begin at a 

 rather uniform hour of the day. Another season's work may perhaps 

 determine where our storms begin. The importance of this question 

 in the attempt to predict thunder-storms is evident. In Bavaria, Von 

 Bezold finds a second faint maximum in the early morning hours, and 

 attributes it to the arrival of storms from a source more distant than 

 that which furnishes the more numerous ones of the afternoon. Our 

 faint secondary maximum, at five to six o'clock in the morning, may 

 have a similar meaning ; but the question cannot l)e decided from the 

 observations of a single season, when one or two storms would exert a 

 strong effect on the total percentage. 



A brief examination of the " general account " suflfices to show that, 

 while the summer storms there recorded are much more frequent than 

 they are in the winter months, still they are not at all uniformly distrib- 

 uted through the summer season : they appear in greater number and 

 size for a few days, and then are almost or quite absent for a time. The 

 cause of this seems to be found in their dependence on the larger dis- 

 turbances in the atmospheric circulation, for which I should wish to use 

 the name of " cyclonic storms" (following Piddington, Redfield, Ferrel, 

 and many European writers), but which are perhaps more generally 

 known by the phrase " areas of low pressure," as used in the Signal 

 Service publications, or as " barometric minima" in Germany. It is as 

 a rule only when we stand in a certain attitude with respect to the cen- 

 tre of these low pressure areas that large and well-developed thunder- 

 storms appear. It may therefore be concluded that the development 

 of such storms depends, not only on the heat of the summer afternoons, 

 but also on the equilibrium of the atmosphere as determined by the 



* Fourteen reports of the squall of July 21, falling between 12'i and 1S\ are 

 omitted in this count. 



