REPORT OF COMMITTEE ON OBSERVATORIES l6l 



photographs of the sun dailj' on a scale of i decimeter to the solar 

 radius and measuring one of these for the area and position of the 

 spots. So far I think we fulfill our purpose sufficiently well. A 

 larger scale is not necessary for positions of the accuracy we seek; 

 indeed, the scale of i decimeter to the solar diameter was suffi- 

 cient for that purpose and was less costly and the photographs were 

 more manageable in size. To attempt to push the work of measure- 

 ment and reduction to a further refinement would immensely in- 

 crease the cost, and I doubt whether it would repay the outlay and 

 trouble. 



But when we come to the question of the details of the spot forms 

 and of their changes, then this scale is certainly not adequate. I 

 greatly wish, and have done so for years, that we had here a second 

 telescope with which we could take comparatively small areas of the 

 solar surface on a scale of at least i meter to the solar radius. I 

 think this is needed to supplement the other. 



But the study of spot spectra is much more urgently needed. My 

 own slight experience of it with a most hopelessly inadequate instru- 

 ment was sufficient to make me feel that it was absolutelj^ one of the 

 most important lines of research. 



I have always felt it a great misfortune that Sir Norman lyockyer, 

 in the work which he has carried on at South Kensington for so 

 many years, should have devised the method of recording just the 

 ' ' twelve most widened lines. ' ' It seems to me in every way a badly 

 devised scheme. If a long series of observations are conducted by 

 one and the same observer, I should think it ran a great chance of 

 stereotyping more or less accidental impressions. If the observer is 

 often changed, we have no longer any means for comparing observa- 

 tions made at different times ; and, at best, if we assume the ob- 

 servations free from all personality and absolutely immaculate in 

 quality, they seem to me to tell us hardly anything at all. The 

 general character of any particular spectrum — not to speak of im- 

 portant details — is left absolutely without record in such a system. 



Sir Norman Lockyer's chief result, namely, that the most widened 

 lines change with the progress of the spot cycle, opens out a great 

 number of questions. First, the spots at maximum are not only 

 more numerous, but they run much larger than at minimum. It 

 would be most important to observe both at minimum and at maxi- 

 mum a series of spots of a definite size. 1 would suggest, as it is a 

 size sufficiently frequent even at minimum, a spot of 200 to 300 

 millionths in area. Clearly it is a great assumption, if we find a 



