584 REPORT — 1887. 



to the equator is made by Faye's formula, and a correction is introduced for the 

 motion of the earth in its orbit. 



This value corresponds to a period of 225- days, and confirms the observations of 

 Young and Vogel indicating a drift on the solar surface. Tbe vpriter hopes later 

 to determine the law according to which this drift varies with latitude and find 

 whether it differs from that deduced by Oarringtou from the motion of sun-spots. 



Throughout this work Professor Rowlands aid and suggestions have been in- 

 Taluable. 



8. On the Diffraction Bands near the Edge of the Shadow of an Obstacle. 

 By Professor G. F. Fitzgerald, F.E.S. 



9. Recent Determinations of Absolute Wave-lengths. By Louis Bell. 



The problem, left almost imtouched since the completion of Angstrom's great 

 paper, has been almost simultaneously attacked by four independent experimenters 

 within the past few years, and their publications have thrown some new light on 

 the subject. Angstrom knew a year or two before his death that the value 

 assigned to his standard of length was certainly too small ; but it was left for 

 Thalen to make the necessary correction, which was published about two 

 years ago. 



At that time the author was just attacking the subject with the object of con- 

 firming or correcting the results obtained by Mr. C. S. Pierce ; but long before the 

 work was completed the paper of Miiller and Kempf appeared, and during the 

 present summer another research has been added to the list in the admirable thesis 

 of Dr. Kurlbaum. 



For convenience the author tabulates the various results, reducing the values 

 given to the corresponding value of the D line for comparison with his own 

 value : — 



Thal6n (Angstrom corrected) . . . 5895-86 



Pierce 6896-26 



Muller and Kempf 5896-25 



Bell 5896-08 



Kurlbaum 5895-93 



Now it is quite evident that these values difier by quantities enormously greater 

 than can be due to pure experimental errors. Aside from these, the errors of a 

 wave-length determination may be due to errors in the assumed values of the 

 standards used, or to errors in the gratings. But, for instance, Miiller and Kurl- 

 baum used the same standards, so that there is an outstanding difference of about 

 one part in twenty thousand which must be ascribed to errors in the gratings. In 

 fact there was a discrepancy of more than tliat amount between the various 

 gratings used by Miiller alone. The nature of the error involved was discussed by 

 the author in his paper in the ' Philosophical Magazine ' for March last. In the 

 case of Pierce's gratings the error was approximateh' corrected by calibrating the 

 grating-spaces, and Pierce's result, as given above, was shown to be distinctly too 

 large. The same is certainly true of Miiller's result, and for the same reason. 

 The values of Thalen and Kurlbaum are also uncorrected for errors of ruling, and 

 probably would be somewhat increased if the proper corrections were applied. 



The standard of length used by the author has shown decided indications of 

 change, and consequently was taken to Berlin tbis summer and compared with the 

 standard used by Miiller and Kurlbaum, from which it appears that the author's 

 value may be too great by as much as one part in a hundred thousand. His 

 gratings will be re-measured at once and the correction for error of ruling recom- 

 puted by more than one method, which, it is hoped, will materially increase the 

 accuracy of his result. Meanwhile work will be continued with larger and better 

 gratings. 



To sum up. It is now quite certain that the wave-length of D does not differ 

 much from 5896-00, and consequently the numbers usually given for the wave- 



