338 PROCEEDINGS OF THE AMERICAN ACADEMY 



better, more devoted observers, so as to lead them on to their own 

 improvement as well as to ours ; while we have offered ouly the very 

 simplest form of record to those who are willing to do something, but 

 who are unable to do much, or who would be discouraged at the sight 

 of a page or more of printed explanations. So far as I know, no other 

 services thus classify their observers : this cannot be because they have 

 enough of the highest class to serve all their needs, for their observa- 

 tions are generally very elementary as to material, and in many cases 

 ai'e insufficient in number. Doubtless it would be best to have all 

 observers of one uniform and high standard ; but this is quite out of 

 the question in a volunteer service. Our observers were therefore 

 divided into three classes last summer, and, in attempting the work 

 again this year, it seems best to separate the classes still further ; not 

 by having more than three classes, but by reducing the demands on 

 Class A, and by asking for more from those whom last summer's work 

 has shown to be painstaking and close observers. 



Before comparing the observations on special phenomena desired by 

 different services, another point may be mentioned which is, I believe, 

 peculiar to the instructions of our New England Society. This is the 

 attempt to obtain material for synchronous maps of storms, without 

 announcing beforehand the hours for observation ; and although it did 

 not fully succeed, from lack of observers, it may be regarded as a 

 distinct improvement, and one well worth retaining in a second year's 

 work. Observers of Classes B and C were asked to take record of tem- 

 perature, wind, &c., every even half or quarter hour during the storm. 

 This is a heavy task, but in return it gives us data for the construction 

 of numerous synchronous maps at short intervals, showing not simply 

 the inferred advance of the storm-belt, but the actual attitude of all 

 the accompanying phenomena for the several times of record. The 

 advantage of taking hours thus marked out by the storm over any 

 hours appointed before hand is apparent at once ; and the importance 

 of the method may be estimated by considering the great advance that 

 was accomplished when it was introduced in the study of cyclonic 

 storms. 



In deciding what elements of tlie storm shall be observed most fully, 

 it must be first agreed by what feature the advance of the storm shall 

 be defined. In Italy, Ferrari places the storm by the attitude of its 

 "maximum phase" where the thunder is loudest. In Bavaria, Von 

 Bezold takes the time of the first thunder for his guide, and draws 

 " isobrontal " lines through points where it is heard at the same time. 

 The first rain has thus far been used in our reductions, and with very 



