OP ARTS AND SCIENCES. 



65 



generalizing the results obtained ; but a brief statement may be allowed 

 in review of the foregoing pages. 



Table V. gives the hours of " loudest thunder " for June, July, and 

 August, 1885, and their percentages; also the totals for the three 

 months. Owing to the growth in the number of observers through 

 June, and the omission of a number of reports not adapted to being 

 counted here, the table cannot be taken as giving a precise view of the 

 relative frequency of storms in the three summer months ; and, again, 

 on account of the smaller number of observers who record the storms 

 that occur at inconvenient hours, the percentages of night observations 

 are probably below their true figures. Nevertheless, the general dis- 

 tribution of storm occurrence is clearly made out. 



TABLE V. — Hours of "Loudest Thunder." 



The excess in the later afternoon hours appears distinctly enough. 

 But the time of greatest frequency does not depend simply on the hour 

 of the day. If we take, on the one hand, southern Vermont, southwest- 

 ern New Hampshire, western Massachusetts, and western Connecticut 



