36 WILLIAM H. ALEXANDER Vol. XXII, No. I 



types, will cause thunderstorms in Ohio only when an area of 

 high pressure prevails over the eastern Lake Region or New 

 England. See weather map of April 17, 1917, 7 A. M. As these 

 types are usually followed by a high pressure area from the 

 northwest of more or less intensity and therefore move with 

 considerable rapidity, the thunderstorms incident thereto are 

 apt to be of short duration and are seldom of a violent 

 character. 



But to the approach of the South Pacific and Texas types is 

 to be attributed by far the greater portion of the thunderstorms 

 in Ohio. These types prevail from early May into late October. 

 As a rule thunderstorms will set in over the western jDart of 

 the State when the center of the "low" reaches Missouri or 

 southern Illinois and will probably become general over the 

 State. These cyclonic types often bring thunderstorms of a 

 very violent nature. When the "low" passes over the north- 

 western corner of the State, thus forcing the isotherms far 

 northward of their normal position, and is followed by a "high " 

 of moderate intensity, hailstorms are likely to occur with the 

 shift of the wind — passage of the squall line — and subsequent 

 increase in pressure. See weather maps of March 10 and 11, 

 1917. The position of the Atlantic high does not seem to have 

 any material effect on the rain -producing characteristics of these 

 "lows." When the path of these cyclonic types suddenly 

 curves to the north and passes into the Lake Region from 

 northern Indiana or Illinois, thunderstorms are likely to occur 

 in Ohio both on the approach and the passage of these areas. 

 Normally, however, their passage just over or just south of 

 the State is followed by brisk westerly winds, clearing weather 

 and falling temperature. 



The East Gulf and South Atlantic types gave rise to no 

 thunderstorms in Ohio during the year 1917. 



(b) The data seem to show^ certain centers of maximum 

 activity and storm-frequency. The southwestern part of the 

 State is certainly the most favorable portion for the develop- 

 ment of the tornado as all tornadoes of consec|uence in the his- 

 tory of the Bureau have occurred in that section. 



(c) Thunderstorms were reported on 169 days, midnight 

 to midnight. Of these 169 days, thunderstorms occurred in the 

 forenoon only on 22 days, on the afternoon only on 80 days, on 



