Gi K Gilbert—Special Processes of Research. 458 
methods by the way; and second, to describe a special system 
of procedure which, like the rule of three, stands ready to 
assist the student whenever a suitable occasion arises. 
The investigation chosen for my first purpose is that of 
thunderstorms, as undertaken by the New England Meteorologi- 
cal Society. It is planned by the society in a broad way and. 
has among its objects the classification of the storms (provided 
the discrimination of distinct types is possible), the determina- 
tion of their mechanism, the discovery of their causes, and the 
prediction of their occurrence. An account of the work, a 
summary of the first season’s observations, and an outline of 
the conclusions deduced therefrom have been published by 
Professor Davis. I have selected this research as-an illustra- 
tion of my theme, partly because it belongs to the living pres- 
ent and is somewhat familiar to most of my audience, and 
partly because the methods employed and the reasons for their 
adoption have been set forth by Professor Davis with excep- 
tional fullness and clearness.* . 
You are not to expect an abstract of his papers, but only of 
such passages as are necessary to an understanding of the gen- 
eral nature of the considerations determining, first, the formu- 
lation of the scheme of observations, and, second, the methods 
of classifying the observations after they had been made. My 
discussion of the scheme of observations will be introduced 
by a general account of the factors determining its principal 
features, and by an account of the features themselves, after 
which an analysis will be attempted of the logical relations of 
the two. ; 
The system of atmospheric phenomena which centers about 
a large area of low pressure is called a cyclonic storm. One of 
these storms usually has a diameter of several hundred miles, 
and moves across the country so slowly that an entire day is 
Spent in passing a single point. Their nature was discovered 
by comparing together simultaneous observations of atmos- 
pheric phenomena made at numerous stations scattered over a 
large area; and a knowledge of this character is utilized, as 
you all know, by the national weather bureau in the prediction 
of storms and other weather features from day to day. To 
obtain the data necessary for prediction, observers are stationed 
at a large number of points distributed at intervals throughout 
the United States; and it is arranged that they record a cer- 
tain set of aimospheric phenomena | at three concerted times in 
each 24 hours. Before the New England society began its 
* On the methods of study of Thunder-storms; by W. M. Davis. Proc. Am. 
Acad. Arts and Sci., vol. xxi, p. 336. 
Thunder-storms in New England in the summer of 1885; by William Morris 
Davis. Proc. Am. Acad. Arts and Sci, vol. xxii, p. 14. 
