508 SCIENTIFIC RECORD FOR 1893. 



rapidly. Rain may come with any wind, but principally with those be 

 tween southwest and northwest; for France these are from the ocean, 

 but for our Atlantic States the rain winds are from northeast to south. 



Fogs frequently arise when a warm current succeeds a cold current 

 or the reverse. They last longer in the second case than in the first. 

 They always accompany the zones of high pressure when these zones 

 are narrow and elongated between two depressions. 



The heat increases or the cold is tempered every time the barometer 

 undergoes a notable depression, for at such times southerly winds pre- 

 vail; cold weather comes with high barometer and the attending north- 

 erly winds. 



Two kinds of hail are to be distinguished: (1) The hails of winter; 

 (2) the hails of autumn and spring. Disastrous frosts coincide with 

 short periods of fine spring and autumn weather, and are frequently 

 produced by terrestrial radiation alone. The frosts of autumn and spring 

 are easily foreseen two and even several days in advance. They are to 

 be feared, (1) after a depression has passed over England and France, 

 and (2) when a depression exists in the Mediterranean Sea. 



A remarkable case of nocturnal radiation occurred on July 29, 1881, 

 in the middle of a long period of oppressive heat; the temperature was 

 38° centigrade in the sun, and fell daring the night to 2°.3 centigrade. 

 This reduction during clear nights has led many to erroneously suppose 

 that the moon is the cause of the frosts, an error that is now happily 

 being rapidly dissipated. 



III. — APPARATUS AND METHODS. 



Hamberg describes a modification of Lamont's earth thermometer 

 that he uses in Sweden, and which for moderate depths may be recom- 

 mended elsewhere. The thermometer proper is enveloped in glass and 

 wood and muslin pads, both to protect it from injury and to cut off 

 rapid changes of temperature. It is let down through a copper tube 

 into the mercury contained in an iron cylinder fastened at a proper 

 depth, and is pulled out, read off, and returned whenever an observa- 

 tion is made. (Z. 0. G. M., xvn, 116.) 



Sprung having devised a simple form of self-recording balance ba- 

 rometer, described already in 1876, has lately published results of 

 actual records as a demonstration of the high degree of accuracy attained 

 by it. The formula of reduction for instrumental corrections as given 

 by Sprung is exceedingly simple. The accuracy claimed is expressed 

 by the "probable error" (or the error whose size is such that it is equally 

 easy to make a smaller or larger error, — or one whose probability is 

 0.5), which is ± 0.11 mm , as determined by comparative readings of a 

 standard barometer, but only ± 0.07 if determined by independent 

 comparative readings of the barometer belonging to the barograph 

 itself, a part of the difference being due to the sluggishness of the 

 latter. A peculiarity of Spruug's barograph is that the special appa- 



