76 REPORT OF THE SECRETARY. 



departure comparable with those noted in 1903. So far as there were notable 

 variations from the mean in the temperature of the North Temperate Zone, 

 these occurred in January and February, and in October and November of 1904. 

 On the whole, these departures are confirmatory of the indications of the spec- 

 tro-bolometric determinations of total radiation for the same periods, but these 

 results, owing to scarcity of good weather, are lamentably inadequate. Still, 

 a medium-weight observation of June indicates average temperature and is 

 followed by it ; a first-class observation of October 5 indicates high temperature 

 and is followed by it, and a medium-weight observation of October 22, indicating 

 low temperature, is followed by it. Prior to December there are only two 

 observations of solar transmission, of which the first, a high value, comes just 

 at the end of a period of high temperature, while the second, a low value, in 

 November, comes just preceding a short period of low temperature. In Decem- 

 ber higher values are found and higher temperatures follow. 



The temperature departures of tlie first half of 1905 also present no sucli 

 favorable opportunity for comparison with solar radiation phenomena as was 

 presented in 1903. The data on total solar radiation have not been fully re- 

 duced as yet, so that they appear still more meager than in former years, and 

 more so than they will after all returns are in. But the weather seldom per- 

 mitted satisfactory solar constant determinations at Washington in the first 

 half of 1905. Those observations here given fall in fairly well, as will be 

 seen, with the record of temperature departures. Observations of the solar 

 transmission have been made very frequently since December, 1904, and are as 

 a rule nearly average values and indicative of average temperatures, and thus 

 in accord with the facts for the North Temperate Zone. Increasing values in 

 December and January and diminishing ones in February and in April and 

 May have their counterparts in the temperature curve for the North Temperate 

 Zone. The temporarily lower values in June find no counterpart in the tem- 

 peratures of the North Temperate Zone, but a fall of temperature may possibly 

 be shown early in July. 



Taking all the facts together we find high values of solar radiation and solar 

 transmission preceding and accompanying high temperatures of the North 

 Temperate Zone; low values of solar radiation and solar ti'ansmission preced- 

 ing and accompanying low temperatures of the North Temperate Zone, and 

 intermediate values of these solar phenomena aecomiianying intermediate ter- 

 restrial temperatures. The evidence at hand is still too fragmentary to pro- 

 duce full conviction, but subject to later confirmation or rejection, as further 

 work shall prove, the results obtained thus far strongly indicate that the 

 transparency of the solar envelope varies at irregular intervals, with conse- 

 quent fluctuations of the solar radiation transmitted to the earth, and that 

 this in turn affects the mean temperature of the earth. 



Observations of 19U.'i-5. 



(a) Solar constant work. — Determinations of tlie radiation reaching the 

 earth's surface, both in toto and for separate wave lengths, have been made 

 at all times when conditions permitted, and have been reduced as in former 

 years to give the transmission of our atmosphere, and the total radiation 

 reaching its outer limit. As above stated the number of days when the sky 

 has continued satisfactorily uniform for this purpose have been few, and besides 

 owing to the press of observing and computing work it has not been possible 

 to compute all the results as yet, so that on both accounts the tables which 

 follow ai-e more than usually meager. Table 1 gives in continuation of similar 

 tables in preceding reports the transmission of the atmosphere at various wave 

 lengths for vertical rays. 



