278 DR. CHARLES CHREE: SOME PHENOMENA OF 



6. If we compare the results from the first and second days of a sequence of 

 disturbed days, we find a truly remarkable difference. There is an unmistakable 

 tendency to an oscillation, represented by a fall in all three elements during the first 

 day and a rise during the second. There are, it is true, a considerable minority of 

 occasions when the change was a rise on the first day and a fall on the second, and a 

 very few when the change on both days was in the same direction, but the amplitude 

 of the changes on such occasions was usually below the mean, especially in H. The 

 number of " fourth " days was so small that the results are not of much significance. 

 "Third" days in the case of H and V resemble "second" days, but the non-cyclic 

 changes appear smaller. 



The isolated days seem intermediate in character between " first " and " second " 

 days, resembling the former most closely in the case of H, but the latter in the case 

 of D and V. The isolated days represented on the whole the shortest storms, and if 

 the phases vary in length with the length of the storm, one day in a short storm may 

 suffice for the development of phenomena that require two days in a longer storm. It 

 must, however, be remembered that it often depends on the hour of commencement 

 whether highly disturbed conditions extend over one day or over two, also that the 

 largest movements occur sometimes near the beginning, sometimes near the end, of a 

 storm, and in the latter case the day on which the storm commences will very 

 likely not be classed as highly disturbed. Another important point is that on 

 individual days much turns on the " accident " of whether midnight comes at 

 the crest or the hollow of the oscillatory movement in progress at the time. In 

 a station at some distance east or west of Greenwich, e.g., Falmouth, the sign even 

 of the non-cyclic change would sometimes be altered if local time were substituted 

 for G.M.T. 



The relationship between the phenomena on the first and second days of a long 

 storm can hardly be a simple case of action and reaction. The rise in V on the 

 average ." second " day distinctly exceeds the fall on the average "first" day. At 

 the end of the disturbed period H on the average is still depressed, but D and V are 

 slightly enhanced. The depression in H sometimes continues for a number of days 

 or even weeks. 



Another consideration is that the fall in H during ." first " days of disturbance is 

 often not the first change of the kind associated with the storm. When the storm 

 has a sudden commencement, the sequence of changes in H at Kew is usually as 

 follows : In a minority of cases there is a distinct but minute fall lasting only one or 

 two minutes, followed immediately by a considerably larger rise. In other cases no 

 preliminary fall can be made out. In either event, in practically all cases, at the 

 end of 5 minutes from the commencement H is above its initial value. There is a 

 rapid rise usually of the order of 20y, but sometimes 50y, 70y, or even lOOy. The 

 rise may occupy only 2 or 3 minutes, but usually longer, and the time during 

 which the force remains enhanced may be only a few minutes or over an hour. 



