Sun-spot Areas and Diurnal Temperature-Ranges. 295 



Tf, however, there be a 24-day Inequality, some of these values will 

 be larger and others smaller than 1000, and it is their differences 

 from 1000 that we finally record with plus and minus signs in order 

 to form our yearly series. It is, however, necessary here to remark 

 that in certain minimum years the observations which exhibit sun- 

 spots are sometimes too unfrequent to render the year tabulated as 

 above in rows of 24 a trustworthy representative of the Inequalities 

 we are in search of. In such cases we have taken in addition half a 

 year on each side of the minimum year, so that we have thus the 

 mean of two years instead of one, the central point of the series being, 

 however, as before, the middle of the minimum year. This treatment 

 has been applied to the years 1833, 1844, 1855, and 1856. 



The coarse of procedure described above may, we think, be justified 

 by the following considerations : 



We are engaged in applying to the sun-spot records a method of 

 analysis unquestionably founded on repetition. We are, therefore, in 

 as far as our analysis is concerned, trying to find if there is repetition 

 in solar records, that is to say, trying to discover if they are possessed 

 of periodicity, whether apparent or real. Then the question arises 

 how to prevent years of many sun-spots from dominating in our 

 results to the exclusion of years of few spots. Now it is nearly certain 

 that, assuming a connexion between the state of the sun and 

 variations in terrestrial meteorology, sun-spots are only a celestial 

 symptom of a. certain variable state of the sun's surface, perhaps of a 

 state of variable power, which is the true cause of terrestrial 

 variability. That is to say, it is highly probable that spots are not 

 themselves the causes- of these terrestrial changes, and it is quite 

 certain that the efficient (in contradistinction to the apparent sun- 

 spot) variation in the state of the sun, whatever it be, will not be 

 found to present ranges so great as those presented by sun-spots ; for 

 instance, spots may be entirely absent from the sun's surface, and yet 

 the sun will cause diurnal variations in magnetism and meteorology 

 of large amount. An efficient in contradistinction to a sun-spot 

 oscillation of short period (presuming this to exist), may peril aps 

 vary in magnitude from times of maximum to times of minimum sun- 

 spots of this we are ignorant all that we know is that the 

 variation in magnitude of this efficient oscillation is doubtless less than 

 that of the sun-spots which indicate it. The efficient oscillation may 

 possibly be represented by a periodic function of short periodicity 

 having a coefficient which varies from times of maximum to times of 

 minimum sun-spots. But if so we are at present concerned only with 

 the short period function, and we can see no other way of obtaining 

 this than by assigning what we imagine to be an average value to the 

 coefficient of the indicating sun-spot variation so as to give all years 

 an equal weight. This is virtually what we have done in our 



