ERRATA 



IN 



VOLUME 79, SMITHSONIAN MISCELLANEOUS COLLECTIONS 



NOTE 



A list of errata to accompany volume 79 was printed by the Smithsonian 

 Institution in May 1929. 



Data for the period 1921-1930 previously printed in volume 79 have been 

 revised and reprinted in this volume, part I. 



NOTES AND EXPLANATIONS 



Page 



7 Bulawayo, Rhodesia. The height previously quoted (4440 ft.) was derived from 

 railway levels. Geodetic Survey has shown that this was 5 ft. too high. 

 22 Salisbury, Rhodesia. The heights previously quoted (Gaol 4835 ft., Meteoro- 

 logical Office 4860 ft.) were derived from railway levels. A resurvey of the 

 area has shown that these were incorrect. The correct heights are 4845 ft. 

 and 4890 ft. There has been no change since September 1921. 

 39 A close examination of the Darwin data has disclosed a progressive error extend- 

 ing over some years by which values which are too high have been assigned 

 to that station. The error has been masked by the fact that pressures have 

 actually been rather high in this region in recent years. A new instrument 

 was installed on August 1, 1931. 



The old barometer was a Kew Pattern, Adie No. 2397, and was one of a batch, 

 four of which have developed errors due to the etching of the inside of the 

 glass tube in the region traversed by the meniscus. To this etched surface 

 the mercury adhered in such a way as to cause little error in a rising 

 barometer. With a falling barometer the meniscus flattened, but this drop 

 was more than offset by the failure of the mercury adhering to the glass to 

 fall appreciably. The net result was a progressive decrease in the diurnal 

 range superimposed upon a slight rise of the 9 a. m. values. A study of the 

 curves suggested the necessity for a correction commencing in the year 1914, 

 just prior to which similar defects had been detected in the other barometers 

 under closer observation in southern districts. 



Formulae were derived for making the gradually increasing correction between 

 1914 and 1931. 



The newly derived data have satisfactorily passed a correlation with neighboring 

 stations and a scrutiny of the frequency distributions of values in the earlier 

 and later years and of the whole record. 



Corrected values are given in the tables which follow for page 430, for the years 

 1914 to 1930. The values for the years 1914 to 1924 should be substituted for 

 those given in World Weather Records, Smithsonian Misc. Coll., vol. 79, 

 p. 430. 



575 



