FROM ORDINARY DAYS OF THE ELEVEN YEARS 1890 TO 1900. 
227 
in equinox, the amplitude and type at the beginning and end of the month must 
differ somewhat largely. The phenomena, for instance, near midnight must be 
considerably different in the beginning and end of March. What the March diagrams 
in figs. 8 and 9 show is a blend of different types. We want shorter periods than 
the calendar month if we are to trace the changes minutely. 
Fig. 9. Vector diagrams in horizontal plane. Details of variation near midnight (24h). 
(ON = OE = lOy.) 
Fig. 10 shows the horizontal plane vector diagram for the year in the case of the 
whole eleven years, as well as in the cases of the sunspot maximum and minimum 
groups of years. The diagrams for the sunspot maximum and minimum groups 
of years are drawn from a common origin, as serving best to bring out points 
of agreement and difference. Except as regards the amplitude of the vector, the 
most noticeable difference between the sunspot maximum and minimum diagrams is 
that the indentation extending from about 11 p.m. to 5 a.m. is much more prominent 
in the latter. This indentation represents an approach to retrograde movement, a 
characteristic, as we have seen, of the winter season. Also, while a bay is recognis¬ 
able in the afternoon hours in all three curves, it is deeper in the sunspot minimum 
diagram than in the other two. 
2 I 
VOL. CCXVI.-A. 
