268 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 51 



That the superheating of the lowest layer of air and conse- 

 quently the accompanying unstable equilibrium within it, is a 

 regular precursor of heat thunderstorms, can certainly be con- 

 sidered as a fact, even now, although the further establishment of 

 its certainty is very desirable by means of observations such as can 

 be furnished in necessary completeness only by self-registering 

 instruments. 



Thus from the observations at Freiburg in Bavaria and on the 

 Hochenschwand, 719 meters higher, Sohnke 17 has shown that in 

 general the difference of temperature between these two stations 

 before the thunderstorm breaks, exceeds its normal value, while in 

 three cases (1881, June 3 and June 26, and 1882, July 22) it passed 

 the limit of unstable equilibrium; indeed on June 3, 1881, it exceeded 

 this limit by a very considerable amount, since on that day the 

 temperature diminished at the rate of 1.53 C. per 100 meters. 18 



Similarly, in the memoir by Assmann 19 on The Thunderstorms in 

 Central Germany, we find on p. 68 a collection of observations of 

 temperature on the Inselsberg and at various lower stations made at 

 special moments before the outbreak of thunderstorms, including 

 a series of cases in which the diminution of temperature with altitude 

 exceeds i°C. per 100 meters and where consequently the limit of 

 unstable equilibrium is exceeded. 



An excessive superheating of the lower strata of air was found in 

 the case of the thunderstorm of March 29, 1888, minutely studied 

 by Assmann, 20 on which day there were observed gradients as high 

 as 2.26 C. per 100 meters. 



17 Sohnke (Der Ursprung, etc.) : The origin of the electricity of thunder- 

 storms, etc. Jena, 1885, pp. 69 et.seq. 



18 But I cannot agree with Sohnke when he thinks that from these observa- 

 tions he is warranted in drawing the conclusion that on such days as these 

 the isothermal surface of o° C. lies especially low in the atmosphere. In the 

 majority of the cases quoted by him the temperature is above its normal value 

 even in the higher levels of the atmosphere, and although in the lower levels 

 the departure from normal is still greater than in the upper, yet this does not 

 warrant us in applying the corresponding temperature gradient to higher 

 levels. We are no more justified in doing this than we are in arguing as to 

 the temperature at very high levels from the vertical gradients observed in 

 the anticyclones of the winter season. Moreover I don't understand why 

 Sohnke attaches such great importance to the special low position of this 

 surface before a thunderstorm, since in support of his theory we need only 

 to consider where this surface is within the thunder cloud itself, a matter that 

 we can only determine at present by computation based on the assumption 

 of adiabatic expansion. 



19 Assmann: Die Gewitter in Mittel Deutschland. Halle a. S. 1885. 



20 See von Bezold (Ergebnisse, etc.). Results of Met. Obs. in Prussia for 

 the year 1888, p. lvii. Berlin, 1891. 



