246 METEOROLOGY AND ALLIED SUBJECTS. 



Hellmann having stated that in his opinion "the attempt recently 

 made to establish an international meteorology must be regarded as a 

 failure, since the prime and most elementary condition of uniformity as 

 regards hours of observation has been neglected," Hanu very justly 

 replies that the only object one can imagine likely to be attained by 

 uniformity in the hours of observation on local time is either du-ectly 

 comparable mean values for the various meteorological elements, or else 

 a convenience in tracing the differences in the diurnal changes of these 

 elements for various localities. But the study of actual observations 

 will speedily convince one that three daily observations will not give the 

 necessary data for coraijaring the peculiarities of climate in various 

 latitudes and altitudes, and continents and coasts. According as we 

 lay greater stress upon the pressure, the temperature, clouds, moisture, 

 winds, &c., we must choose different hours of observation. And it is 

 therefore necessary that in every land hourly observations should be 

 made, if we hope to attain tlie objects that Hellmann seems to have in 

 view. We are of the opinion that the present system of international 

 observations in Europe is a very useful one, and marks a great progress 

 in comparison with the condition of things before the Vienna Congress 

 of 1873. One can imagine better things still, but we "must not allow 

 the better to be the enemy of the good." {Z. 0. G. M., Yol. XIV, p. 2C3.) 



Dr. Galle, director of the observatory at Breslau, in deducing the an- 

 nual temperature curve for that observatory, from daily means based 

 on eighty-five years of observation, has adopted the following method, 

 which is shorter and more rational than that of Bloxam. If for a series 

 of consecutive days .... n — 2, n — 1, n, n-\-l, and w+ 2, we have 

 the mean observed temperatures ta—z, *n— i? ^n? &c., and we desire more 

 accurate determination of the temperature ta of the middle one of these 

 days ; then it is plain that if the mean daily temperatures are uniformly 

 increasing or decreasing, the true temperature ta will be given by the 

 formula : 



and a value t^ comparatively free from small errors of observation by tak- 

 ing the mean of five such values. But if the observed daily mean tem- 

 peratures should not have uniform weights in forming the desired mean, 

 because the temperature during these days has not changed uniformly, 

 then the observations must be combined according to the methods de- 

 fined by the law of probabilities, and this method is particularly con- 

 venient of application when we attempt to combine together seven 

 consecutive daily means. Thus if for brevity a = fn_3, l) = tj^_i, g = t^ + 2j 

 we then have either one of the following formulae: 



1. <^=-jL(3c + 4(Z + 3e) 



2. ta = ^{b + 3c+4:d+3e-^f) 



3. <a=4V(«+4& + 9c+12d + 9e+4/+5f) 



