16 PROCEEDINGS OP THE AMERICAN ACADEMY 



even times occur with natural frequency, such as wind-squall 16.37, 

 t:ain 16.39, heaviest rain 16.45 to 16.58 (this being an actual record 

 from one of our best observers), bears strong internal testimony to its 

 accuracy. In many storms, an inaccuracy of seven minutes in timing 

 the beginning of rain would cause an error of almost six miles in 

 charting the rain-front. An accurate timepiece, carefully recorded, is 

 therefore essential to the best success. 



Reduction of Observations. — On receiving records from our observ- 

 ers, the forms were classified, first by dates, next by storms under the 

 same date if possible, and last by States. Next, the time of rain be- 

 ginning was charted, on a large map (six miles to the inch) if the 

 records are numerous ; on an intermediate scale (seventeen miles to, an 

 inch) for many dates ; on a small map (thirty-five miles to an inch) for 

 days when the records were scattering. The greatest part of this work 

 has been done by Mr. Oswell. When reduced to this form, the rain 

 times are examined to discover evidence of the progression of the rain 

 area across the country, and, if this appear clearly enough, lines indi- 

 cating the probable position of the rain-front are drawn for the even 

 hours,* or for smaller intervals, and from these the direction and ve- 

 locity of the storm's advance are measured. Only a small portion of the 

 work thus performed can be presented in this report. Having obtained 

 a general idea of the stoi'm, the records are examined again in order 

 of rain times, and the questions of wind directions, cloud motions, and 

 temperature changes are considered. 



Our insufficient number of stations has been a continual regret dur- 

 ing this work ; the storms have in several cases seemed to choose 

 districts freest from observers, so that in one sense the labor of this 

 first season's investigation must be regarded as having its best result in 

 exciting more general interest and better work another year. The 

 storms of New England are also generally so moderate that the detec- 

 tion of their circulation will be probably more difficult than if these 

 studies were carried on in the Western Stajtes. Nevertheless, it is be- 

 lieved that the collection of facts here briefly summarized will prove 

 of value in the study of local storms in this country, and that it will 

 serve well as the basis for more extended study in New England during 

 the present year. 



Daily Maximrim Temperatures. — The occurrence of thunder-storms 

 is in many cases so closely connected with the change from high to low 



* Throughout this report, the hours are counted from midnight to midnight, 

 80 that afternoon and evening hours fall between 12 and 24. 



