284: F. H. Bigelow — Studies on the General Circulation 



zone, due to the relative action of the continents and the 

 oceans upon the general circulation in the surface levels where 

 the observations are generally made. Again, it is not possi- 

 ble to treat a continent, or even a portion of a continent, as a 

 unit. Thus, there is usually a temperature oscillation between 

 the Pacific States and the Central and the Eastern States in 

 North America, one being high while the other is low. Some- 

 times the interior stations of the United States show an excess 

 of temperature, while all the states bordering on the Atlantic 

 Ocean, the Gulf of Mexico and the Pacific Ocean have a simul- 

 taneous defect. The complications of the circulation are 

 excessive and difficult to classify in a simple system, but it is 

 important to emphasize the fact that stations from all parts of 

 the world cannot be massed together for short-period varia- 

 tions, and even for long-period oscillations a discriminating 

 knowledge and judgment is demanded. The selection of sta- 

 tions for any least square or harmonic analysis should have as 

 much regard for the fact of inversion-effects, proceeding from 

 the same external general cause, as for the other fact that 

 periodicity is not necessary for synchronism in physical pro- 

 cesses. 



Homogeneous direct observations. 



The means, departures, consecutive means, and the residuals, 

 depend upon original observations which must be homogene- 

 ous to produce valid variations of any sort. In the tempera- 

 tures, the observations are made at a few selected hours; in 

 the precipitation they are made occasionally, at the end of the 

 irregular rain and snow intervals ; in the solar prominences 

 they are made from a few stations as sights of the sun are per- 

 mitted by local cloudiness; and similarly with other phenom- 

 ena. In the United States the daily temperatures depend 

 upon observations made at 7 a. m., 3 p. m., 11 p. m., Washing- 

 ton meridian time, 1871-88 ; at 8 a. m., 8 p. m., 75th meridian 

 time, or maximum and minimum temperatures since July, 

 1888. Previous to the publication of Bulletin S, 1909, there 

 was no series of homogeneous temperatures in the United 

 States available for discussion. The hours of observation 

 differ widely in other countries, and in general the entire sub- 

 ject of temperature variations for the world is very chaotic 

 and unsatisfactory. The practical difficulty of securing 24 

 hourly observations every day, except by self-registering instru- 

 ments, is so great that different combinations of selected hours 

 have been used as substitutes, to the great detriment of scien- 

 tific meteorology. Since the selected hours of observation are 

 on Washington or 75th meridian mean time, it follows that the 



