CHRISTCHURCH TERM-DAY OBSERVATIONS. 



237 



the arithmetic mean of the I values, as given in the last columns of Tables VIIlA and IXx, we find the 

 following values for the ratio : 



In the case of D the value thus obtained for the year is less than that for Midwinter, and only a little 

 larger than that for Equinox ; but in the case of H the value for the year is larger than even that for 

 Midwinter. This can only be ascribed to variability in the phase of the diurnal inequality, and this 

 greater variability of phase no doubt is partly accountable for the larger excess of the R over the I values 

 in the case of H. 



The phenomena presented by 24 days from 12 consecutive months at a particular station may owe 

 a good deal to " accident." Thus it would be a mistake to make sweeping deductions from the 

 Christchurch data discussed here. I trust, however, that the sidelights that have been thrown on the 

 nature of the diurnal variation will at least suggest lines of profitable investigation. 



