82 



ANNUAL INEQUALITY. 



mean monthly values after an allowance has been made for the effects of secular change, assumed to 

 proceed iit a uniform rate throughout the year. This is due, at least in the case of the Declination, to the 

 fact that the regular annual inequality in temperate Europe if not absolutely nil is exceedingly small. 

 Table VIII gives the results obtained for the annual inequality at Winter Quarters. In obtaining the 

 Horizontal-Force results, the mean monthly values were taken from Table V, and the secular change 

 accepted was the mean given in Table VI. The Declination results under both (i) and (ii) are based on all 

 the days of registration. The results under (i) accept the mean monthly values given in Table IV, and the 

 mean value given in Table VI for the secular change, i.e., they answer to a variable base line as given by 

 the absolute observation. The results under (ii) assume the base-line value to be invariable, and the true 

 value of the secular change to be the mean given in Table VII. 



TABLE VIII. Annual Inequality. 



The range given for the annual inequality by any set of figures in Table VIII is simply enormous 

 compared to anything that exists in ordinary latitudes. The two sets of figures for the Declination are 

 far from similar. Those under (i) look at first sight the less improbable, as giving the smaller range. It 

 should, however, be remembered that at Winter Quarters 1' of arc in Declination answered to only about 

 1 - 9y in force, so that the range under (ii) when converted into force is only about 184y, or 40 per cent, of 

 the apparent range in Horizontal Force. Thus if a large range in an annual inequality is regarded as too 

 improbable a result to be accepted, the argument is not so strong against the Declination results under (ii) 

 as against the Horizontal-Force results. 



16. If we take a Midsummer mean from the months November, 1902, to February, 1903, and a 

 Midwinter mean from the months May to August of both years combined, both the epochs concerned 

 centre at January 1, 1903, so that the results are free from the uncertainty as to the real value of the 

 secular change. 



The results thus obtained are as follows : 



