168 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1923 



lief erring again to "the butterfly diagram," the shift of the 

 habitat of spot groups during the progress of the solar cycle shown 

 therein would appear to be a sort of surface ripple on the sun, mov- 

 ing in both hemispheres from high latitudes to low, as each cycle 

 proceeds from its beginning to its close. The case of the Cepheid 

 variables has made us familiar in recent years with the idea of stars 

 which change their volume — stars with pulsating photospheres. In 

 the sun we have, as yet, no proof of a general pulsation of the photo- 

 sphere; only of this ripple of disturbance on its surface. But, since 

 the rotation period, as indicated by sunspots, is shorter as the 

 equator is approached, it follows that the mean rotation period of the 

 sun is quicker before minimum than after. 



But the above diagram puts the fact of the difference in the rota- 

 tion periods of different latitudes in much too crude a form. A 

 paper on " The solar rotation period from Greenwich sun-spot meas- 

 ures, 1879-1901," appeared in the Monthly Notices, 65, and on page 

 817 a table was given (Table II) which was intended to show how 

 wide are the differences of rotation period, as derived from different 

 spot groups even in the same latitude. Every group has a proper 

 motion of its own; even within the same group there are internal 

 movements. The leader spot in the early days of a group tends to 

 rush forward over the surface ; in the group's later history to slacken 

 speed and return on its track. 2 Also the groups which live the 

 longest, returning two or more times to the visible hemisphere of the 

 sun, move more slowly on the whole than short-lived spots, and yield 

 different rotation periods during their successive apparitions. 



One remarkable peculiarity is common to both the cycles, 1879- 

 1891 and 1891-1901, and since it was brought out by Carrington's 

 inquiry two cycles earlier it is probable that it expresses a real pecu- 

 liarity of the solar rotation. In spite of the great irregularity in 

 the rotation periods given by the spots in any particular zone, there 

 does appear to be a distinct tendency for the shortest mean period to 

 be given, not at the equator, but slightly to the north of it. The curve 

 given by the different rotation periods is not precisely symmetrical 

 with respect to the equator, and, on the ivhole, there appeal's to be a 

 tendency for the periods in the northern latitudes to lengthen more 

 rapidly with distance from the equator than with those of the southern. 

 A similar feature was suggested in " the butterfly diagram," wherein 

 the southern hemisphere appeared to encroach slightty on the northern 

 at the equator. The equator of rotation would appear not to coincide 

 with the equator of figure, but to lie slightly to the north of it. 3 



2 " Notes on some of the spot groups measured at the Royal Observatory, Greenwich, on 

 photographs of the sun taken in the year 1915," Monthly Notices, 79, 451. 



8 See a paper communicated to the British Astronomical Association on " Rotation 

 periods of the sun as determined from flocculi and from sunspots " (Journal of the 

 British Astronomical Association, 32, 104-107) : " We thus infer a state of strain not 

 only between the southern and northern solar hemispheres, but also between the higher 

 and lower strata in those hemispheres " ; p. 106. 



