KOVEMBEE 3, 1911] 



SCIENCE 



607 



persions, the latter giving a maximum 

 displacement, at the solar equator, of ahout 

 0.090 mm. 



Considerable space is devoted to a discus- 

 sion of the possible sources of error and it is 

 evident that the greatest possible care was 

 taken to avoid all known causes of systematic 

 displacements of the lines and consequent 

 error in the velocity. In most astronomical 

 work systematic errors are much more to be 

 feared than accidental errors and this is espe- 

 cially true in this case where the line dis- 

 placements to be measured are small. The 

 greater relative importance of what might be 

 called plate errors over the accidental errors 

 of measurement is clearly shown by the re- 

 sults obtained in this investigation, where the 

 probable error of the mean value of 21 plates 

 is considerably less than the probable error 

 of a single plate as determined from the in- 

 ternal agreement of the 22 lines on the plate 

 — a ratio of plate errors to measurement 

 errors of more than five to one. The method 

 of measurement and reduction is concisely 

 and yet fully described and is followed by the 

 detailed measures of the plates obtained in 

 the two main and two supplementary series 

 which are then conveniently summarized. 



The discussion of these results is admir- 

 ably arranged so as to present in a convenient 

 form the conclusions reached, the most inter- 

 esting and important of which may be briefly 

 stated. 



1. So far as the period covered by these ob- 

 servations goes, the siin's rate of rotation is 

 constant, the slight difference found in the 

 two series being ascribed to the slightly less 

 satisfactory instrumental conditions in the 

 first series. 



2. The retardation of the rate in higher 

 latitudes is satisfactorily represented by an 

 equation of the Faye type taking the form 

 for these observations of 



f = ll°.04-h3°.50 cos= 



where | is the daily angular sidereal velocity 

 and ^ is the solar latitude. 



3. The lines of different elements in the re- 

 versing layer give different values of the ro- 



tational velocity, which, though small, are be- 

 lieved to be real, those known to lie at low 

 levels giving low values, and vice versa. This 

 is especially the case with Ha and Ca X 4227, 

 which move at a more rapid rate than the 

 general reversing layer and in which the re- 

 tardation towards the higher latitudes is very 

 much less. 



4. The comparison of Ha, A. 4227 and lines 

 in the reversing layer shows that the velocity 

 increases and the polar retardation decreases 

 with increasing distance outward, the cause 

 being assigned as probably due to the effects 

 of friction in the lower portion of the solar 

 atmosphere. 



The whole work sets an exceedingly high 

 standard of accuracy, which it will be diffi- 

 cult for other observers to equal. Taking for 

 example some of the probable errors of meas- 

 urement obtained, we have in the second series 

 the probable error of measurement of the dis- 

 placement of a single line r+r 0.009 km. per 

 sec, equivalent to a linear error of only about 

 0.0004 mm., less than half a micron. Those 

 who have had experience in measuring spec- 

 trum lines where a probable error of a micron 

 is considered good measuring will recognize 

 the remarkable accuracy obtained, several 

 times greater than previously secured in the 

 same problem. The corresponding error of a 

 plate is dr 0.002 km., the thousandth part of 

 the equatorial velocity. Notwithstanding what 

 was previously said concerning systematic 

 displacements the agreement among different 

 plates is also remarkably good, the probable 

 error of a single determination of the rota- 

 tional velocity being not much greater than 

 =: 0.01 km., giving the probable error of the 

 mean value of the velocity in the neighbor- 

 hood of ± 0.003 km. 



Professor Adams and iliss Lasby are to be 

 congratulated upon the very high accuracy 

 of this determination of the solar rotation, 

 upon the interesting and important conclu- 

 sions derived from their measures, and upon 

 the manner of presenting the formidable 

 amount of material on hand. Furthermore, 

 they, with the Carnegie Institution, are to be 



