276 



SCIENCE 



[N. S. Vol. LIV. No. 1395 



tiou with the lowest bid of $6,000. This sum would 

 afford us the chance of buying a small house as 

 a refuge for our collections and library, which 

 are in constant danger of being burned out. 



It seems to me that all who appreciate the 

 importance of Mendel's contribution should 

 be actively interested in this message. Just 

 what can and should be done and how to go 

 about it are matters for discussion. I would 

 suggest that those interested express their 

 views in the columns of Science. However, 

 some may wish to communicate directly with 

 Dr. litis. His local address is Backergasse 

 10, Briinn, Tschechoslwakei. 



E. B. Babcock 



NOTES ON METEOROLOGY AND 

 CLIMATOLOGY 



DETERMINING THE TRUE MEAN TEMPERATURE 



What is the true mean temperature ? It is a 

 much easier task to define the true mean 

 temperature than to determine it, but the 

 prosecution of meteorological and climatolog- 

 ical work demands that this element be de- 

 termined. A very detailed and thorough dis- 

 cussion of the question has been published by 

 Mr. C. E. P. Brooks, of the British Meteor- 

 ological Office, in the Monthly Weather Ee- 

 view,''- and it is of interest to review the varied 

 nature of the problem and the solutions offered. 



The " true mean temperature " is the mean 

 height of a thermograph trace corrected for 

 any sources of instrumental error. In prac- 

 tise, however, the mean of twenty-four hourly 

 observations, or even the mean of observations 

 every two, three, or four hours, is sufficiently 

 close to yield the daily mean temperature. 

 But it is not always feasible to secure such 

 frequent observations, and the problem of 

 reconstructing the true mean from three ob- 

 servations daily and the maximum and mini- 

 mum faces the meteorologist. There are four 

 ways of accom'plishing this: 



First, the combination of the means of the 

 three daily observations at fixed hours, or the 

 maximum and minimum, in proportions that 

 have been found to be satisfactory at certain 

 standard stations. If observations at Y a.m., 

 1 or 2 P.M., and 9 p.m. be designated by I, II, 



1 April, 1921, pp. 226-229. 



and ///, respectively, and the true mean by T, 

 it is found that 



T=(I + 11 + 2 X///)/4 



gives the best results, in general; but, in 

 Greenland, where the morning observation oc- 

 curs at 8 A.M., the formula commonly used is 



T=[2{I + 11) +5 xni]/9. 



The author carried out an investigation of 

 this type of formula by the method of least 

 squares, for various groups of stations, such 

 as western Europe, subtropics, and tropics, 

 and found that for the first the usual one 

 (given above) gave the best results; for the 

 subtropics, the best combination is 



T=[I + II + III — %o(.II — III)y3; 



and for the tropics, this last can be used, al- 

 though it is not as accurate for the tropics as 

 for the subtropics. An alternative formula 

 which gave very satisfactory results for Ba- 

 tavia is 



2-= [2(1 4- 77) +Z x777]/7. 



These apply, of course, to the hours specified 

 above. It is further pointed out that the maxi- 

 mum and minimum can be combined into such 

 formute, instead of the three observations, 

 and examples are given for Hamburg, in 

 which the maximum and minimum are com- 

 bined with the morning and evening observa- 

 tions ; and for Tunis and Egyptian stations in 

 which the minimum only was combined with 

 the three daily observations. 



Second, "the calculation of appropriate ad- 

 ditive corrections for various combinations 

 of hours or for the mean of the maximum 

 and minimum at standard stations, and their 

 transference without modification to other sta- 

 tions in the vicinity." This method is only 

 applicable to those stations or regions where 

 the conditions are similar, such as are to be 

 found in Eussia and Siberia, or in the eastern 

 half of the United States. The method is to 

 plot the corrections for standard stations and 

 to read oS for the intermediate stations the 

 appropriate correction. In mountainous dis- 

 tricts, such as western United States, or the 

 mountains of India, this scheme is unsatis- 



