310 



SCIENCE 



[N. S. Vol. XXVII. No. 686 



earth, (2) the application of analytic methods 

 based on perfect periods for the solar varia- 

 tions which are only roughly periodic, (3) 

 the omission to consider the important effects 

 of atmospheric circulation upon the observed 

 temperatures. The temperatures of the 

 United States are very indirect functions of 

 the solar radiation, depending upon trans- 

 ported heat more than upon direct radiation. 

 Many years have been consumed in securing 

 homogeneous data, but we now possess tables 

 covering the interval from 1872 to 1905 for 

 the solar prominence frequencies, the Euro- 

 pean horizontal magnetic force, the tempera- 

 tures, vapor pressures and barometric pres- 

 sures of the United States, comprising more 

 than one hundred stations. The monthly 

 means were in all cases computed from the 

 daily values reduced to homogeneous condi- 

 tions for the interval. This paper refers only 

 to the variations of the annual means of the 

 several elements, and these are separated into 

 two portions by eliminating the short period 

 Grdinates of about three years from the long 

 period ordinates of about eleven years. The 

 result is approximately as follows : Referring 

 to an earlier paper, it was shown there that 

 the temperatures in the tropics vary directly, 

 but in the temperate zones inversely, with the 

 solar prominence frequencies, and that the 

 pressures of the eastern hemisphere vary 

 directly, but of the western hemisphere in- 

 versely, as a general rule. In this paper it 

 is shown that these rules hold good in the 

 United States, with certain local modifica- 

 tions due to the general circulation of the 

 atmosphere. In the Pacific states the syn- 

 chronism of the temperatures is direct, while 

 east of . the Eocky Mountains it is inverse. 

 The eleven-year period is well defined in the 

 Pacific States, but is nearly wiped out by the 

 circulation east of the mountains. The three- 

 year period persists in each district, but a 

 partial inversion and a lag of a few months is 

 indicated east of the Rocky Mountains. 

 These and other similar results are the direct 

 products of circulation, since the California 

 district is really a part of the system of the 

 tropics, because the high-pressure belt crosses 

 the United States from Florida to Oregon, 



and thence passes southwestward. These in- 

 versions are naturally connected with the 

 general circulation. An increase in solar 

 radiation, inferred from an increase in the 

 number of the solar prominences and the 

 strength of the magnetic field, causes in- 

 crease of temperature and circulation in the 

 tropic zone, but a decrease of the tempera- 

 ture in the middle latitudes because the 

 return currents from the polar zones bring 

 an excess of cold high areas into the United 

 States. The temperature for a given year in 

 a given locality is due to the transported heat 

 primarily, which rather remotely depends upon 

 the prevailing radiation. Aside from the 

 great difficulty of securing homogeneous tables 

 of the solar radiation by observations with 

 the pyrheliometer, on account of the difficulty 

 of eliminating the local absoi-ptions due to the 

 meteorological conditions of the atmosphere 

 above the station, it is yet more diiEcult to 

 make allowances for the effect of circulation. 

 The statistical method of recording the facts 

 must be pursued without interruption for 

 many years, and then this historical record 

 will gradually build up laws which will have 

 a practical value in seasonal forecasts. The 

 U. S. Weather Bureau now possesses a set of 

 homogeneous records covering one third of a 

 century, and all its observations are now made 

 so as to continue these synchronous data in a 

 perfectly automatic manner. 



E. L. Paris, 

 Secretary 



THE CHEMICAL SOCIETY OF WASHINGTON 



At the twenty-fourth annual meeting of the 

 Washington Section of the American Chem- 

 ical Society, held at the Cosmos Club, Thurs- 

 day, January 9, the following papers were 

 read : " The Detection of Thickeners in Ice- 

 cream," by Professor G. E. Patrick, and 

 "The Three Component System CuO, SO,, 

 H,0," by J. M. Bell and W. C. Tabor. 



The following officers were elected: 



President — Joseph S. Chamberlain. 



First Vice-president — P. H. Wall'Cer. 



Second Vice-president — G. H. Failyer. 



Secretary — J. A. Le.Clerc. 



Treasurer — F. P. Dewey. 



