DECEiMBtk y, iy::oj 



NATURE 



409 



the world, the results of which will be published in 

 due time. 



The tirst and most striking result is that the solar 

 radiation exerts a cumulative effect on the atmosphere 

 so that prolonged periods of high or low solar radia- 

 tion have a much greater response in atmospheric 

 action than shorter periods of greater intensity. 

 Thus a deviation of 5 per cent, from the average 

 radiation of short duration might not produce so 

 great an effect as a long period, with a mean devia- 

 tion of 2 per cent. 



The second result is that as the sun changes from 

 one hemisphere to another the effect on the weather 

 changes, so that in the hemisphere where the sun is 

 nearly vertical the pressure falls over the land surface 

 and rises over the water surface, while the opposite 

 effect is found in the other hemisphere. This effect 

 1 take to be a proof that a considerable part of the 

 increased solar radiation reaches the earth's surface 

 and intensifies the normal effect of absorption of solar 

 radiation by the land. There are, however, clear 

 indications that a considerable part of the increased 

 solar radiation is absorbed by the upper air and gives 

 origin to atmospheric waves which, if they are not 

 the determining cause, at least are very influential in 

 determining the discontinuities referred to recently in 

 Nature by Prof. Bjerknes. 



These atmospheric waves are of a complex nature, 

 and for this reason the correlations with solar changes 

 are best set-n by separating the solar and weather 

 changes into shorter and longer (or slower) waves. 



By taking averages of ten days we found for 19 16 

 correlations as high as 84 per cent, with certain 

 .\rgentinc stations, and correlations exceeding' 80 per 

 rent, in later years. The shorter waves do not give 

 such high correlations, apparently for several reasons, 

 the most important, perhaps, being that there are 

 still some errors in the solar values due to the diffi- 

 eulties of eliminating the effect of changing Irans- 

 [larencv of the air while the solar observations are 

 iK-ing made. 



This difTiculty has undoubtedly been greatly 

 diminished during recent months by the use of the 

 pvranometer, which enables the observers to get 

 measurements of the solar heat from a single holo- 

 graph. Ccrtainlv since this method was begun the 

 correlations of the solar variations with atmospheric 

 waves have inrrease<l. .Another <linricultv appears to 

 be that while the waves start from definite centres of 

 action, these centres shift position to some extent, 

 so that the time interval of the effects following solar 

 changes is variable, and this variability is of greater 

 importance in the shorter waves. 



The prolonged effects following unusually high or 

 low values of radiation shown by Dr. .-Vbbot's curves, 

 reproduced in Nature, I believe to result from the 

 normal sequence of events on the surface of the sun. 

 .After a marked excess or defect of radiation there is 

 likely to follow a similar deviation from the normal 

 after an interval of ten tt> thirteen days. 



This result I believe to arise from the fact that 

 when there is a marked outbreak of heated solar 

 gases on the edge of the sun. where absorption is 

 normally great, the total radiation is more intensified 

 than when the outbreak Is near the centre. If the 

 outbreak is near lh<- edge it will l>e carried by rotation 

 to the npp<i-.ile edge in from about eleven to nineteen 

 days, according as to whether the nonition is on the 

 east or west t^lge. 



In the same way cooled gaws pro<luro their greatest 

 absorbing effects when near the edge of the sun, and 

 there is the same lernlenrv to retHvil. The-ip effe.ts 

 also tend to rejwat themselves after 



NO. 2667, VOL. 106] 



but in that case the interval is so long that marked 

 changes have usually occurred. However, these 

 repetitions are sufficiently numerous to make it prac- 

 tically certain from the length of the rotation that 

 the conditions causing the marked deviation from 

 normal are in the region of the prominences, and not 

 in that of the sun-spots, which have a period of 

 about twenty-seven days. 



These studies of the relations between solar changes 

 and the weather have been in progress in the .Argen- 

 tine .Meteorological Office for several years, so that 

 when a station of the Solar Physics Observatory of 

 the Smithsonian Institution ' was opened in Chile, 

 arrangements were made between Prof. G. O. 

 W'iggin, director of the .Argentine Weather Service, 

 and Dr. Walcott, director of the Smithsonian Institu- 

 tion, for the transmission of the solar radiation 

 1 measurements by cable to Buenos Aires and for 

 I their use so far as possible in weather forecasting. 

 To this work Prof. VViggin and the assistant chief 

 have given enthusiastic support, and through their 

 efforts, and as a result of a growing interest in the 

 subject in Argentina, the Legislature has recently 

 appropriated some 25,000 pesos for the installation 

 of a solar observatory in Argentina. 



The forecasts are made for a week in advance bv 

 publishing estimated temperatures for each day, and 

 also forecasts of expected rains. With increased 

 experience there have been steady improvements and 

 an increasing demand for the forecasts by com- 

 mercial interests. 



There are many complexities which remain to be 

 solved, and perhaps some which can never be solved, 

 but I believe that a distinct advance has been made 

 in forecasting, and the progress is too far advanced 

 to be turned backward. One of the greatest difficul- 

 ties is the inversion of the solar relation with the 

 season, and occasional inversions for other reasons 

 not yet clearly understood. 



Heretofore, most of the studies by others in regard 

 to the relation of solar changes and atmospheric 

 phenomena have been in reference to the sun-spots. 

 Dr. .Abbot's observations indicate clearlv that there 

 is a change of solar radiation corresponding to the 

 eleven-year sun-spot period. The radiation increases 

 with the increase in spottedness, but this change is 

 of relatively small importance compared with shorter 

 changes of greater intensity. 



Sir Norman Lcwkyer suggested some years ago 

 that there were weather changes due to solar changes 

 other than the eleven-year period, and brought evi- 

 dence to show that a period of about four vears in 

 the outbreak of prominences caused atmospheric 

 changes having the same periotl. But solar radiation 

 shows a vari.-ibility far greater than is indicate<l by 

 the observations of prominences, and the high value 

 of the correlation between these and the weather 

 which we Itave found for the Argentine and for other 

 regions of the world leads mo to believe that these 

 changes are the chief, if not the only, cause of. 

 weather changes as distinct from the well-kitpwH' 

 diurnal .mil .innual periods. 



H. H. ri.AVTON. 



Chief of Forec.-tst Division. 

 Oficina Metcorol6gira .Argentina. 



The Physical Meaning of Spherical Aberration. 



The nature of the distribution of light rouml the 

 axial focus of a lens is a problem which has frequently 

 been attacked mathematically. It has been quite sac- 

 rcrssfully solved for points in the i>rincipal focal plane 

 of an " aplanatic " lens when the intcnxily is found to 



