118 EXPERIMENT STATION RECORD. [Vol.43 



that, if the relation of j-ainfall to road condition has been previously studied, 

 recent reports of rainfall may ))e useful in indicating road condition in the 

 absence of actual reports on the subject. 



Average free-air conditions as observed by means of kites at Drexel 

 Aerological Station, Nebr., during the period November, 1915, to Decem- 

 ber, 1818, inclusive, W. R. Gregg {U. 8. Mo. Weather Rev., 48 {1920), No. 1, 

 pp. 1-11, figs. 8). — This article presents the results of the free-air observations 

 at Drexel, Nebr., in brief and convenient form for the information and use of 

 artillery and aviation services. " Tables and figures give mean monthly, sea- 

 sonal, and annual values of the different elements at various levels up to 5 km. 

 The data are compared with similar data for Mount Weatlier, Blue Hill, and 

 elsewhere, and a separate table contains comparative values of air density, as 

 determined by different investigators for various parts of the \vorl<l." One of 

 the points of special interest discussed is the veering of winds with altitude. 



A list of 12 references to literature on the subject is given. 



Relation of winds to temperature in central Ohio, H. H. Martin (TJ. S. 

 Mo. Weather Rev., 7/8 (1920), No. 2, pp. 85, 86, figs. 2).— "This paper presents v 

 by table and graph the existing relation between the direction of the wind and 

 the existing or current temperature, as well as the subsequent 24-hour tem- 

 perature change. The data are taken from the records of the Columbus (Ohio) 

 office. 1909-1918. . . . 



" Under ordinary conditions, at Columbus, Ohio, it seems .«afe to consider 

 the north, northeast, east, and southeast winds as prognostics of warmer 

 weather 24 hours later, except in sxmimer, when the east wind has no prog- 

 nostic value ; and the northwest, west, and south winds as prognostics of cooler 

 conditions 24 hours later, except in sununer, when the northwest wind is 

 usually followed by higher temperatiires. However, excepting the case of 

 the east wind in winter, the true prognostic values are low and do not justify 

 nnu-h reliance. Theii- value would be appreciable only when used in conjunc- 

 tion with other indications." 



Smoke formations in air drainage, C. Hallenbeck . ( [7. S. Mo. Weather 

 Rev., 48 (1920), No. 1, pp. 24. 25, fig. 1; al)S. in Bui. Amer. Met. Soc, 1 (1920), 

 No. 3, p. 35). — In connection with a study of temperatures and orchard heating 

 in the Roswell, N. Mex., fruit district, it was observed that during cold nights 

 the tops of fruit trees may be frozen while the bottoms are uninjuretl. This 

 was found to happen when air, cooled over the dry, open plains during a clear 

 night, slid in a thin sheet out over the air of the valley. The movement of 

 smoke eai-ly in the morning of December 9, 1919, showing this process in op- 

 eration, is described and illustrated. 



Spring frosts (U. S. Dept. Agr., Natl. Weather and Crop Bui, No. 17 (1920), 

 pp. 2, 3, figs. 2). — ^A chart is given which "shows for the country east of the 

 Rocky Mountains the dates after which killing frost is likely to. occur only Ij^; 

 year in 10 on the avei'age. After April 1 the chance of killing frost alon: 

 the south Atlantic coast and in the southern portions of the Gulf States ii 

 only 10 per cent. The line for May 1 extends southwestward through centra 

 Virginia and westei-n North Carolina, and then bends northwestward througl 

 eastern Tennessee and western Kentucky, and thence westward through t 

 southern portions of Indiana, Illinois, Missouri, and extreme southern Kansai 

 In nmch of North Dakota and in the northern portions of IMinnesota. Wii 

 cousin, and Michigan, as well as in parts of New York and in northern Ne 

 England, killing frost may be expected at least 1 year in 10 after June 1. T 

 is the case also in the central and western Rocky Mountain Plateau Stati 

 except in a few favored localities." 



