SPECIAL TYPE OF DISTURBANCE. 



189 



104. Table LXX summarises the results from the whole 82 occurrences as to the hour at which the 

 first phase ended. An occurrence at an exact hour was assigned to the hour then commencing. It will 

 be seen that more than half the occurrences happened between 7 and 9 p.m. There is no marked seasonal 

 variation in the hour of occurrence. A larger proportion of occurrences took place after 8 p.m. in April, 

 May and June than in July, August and September, but that might be accidental. The great concentra- 

 tion of occurrences near 8 p.m. is obviously a vital point when considering the real significance of the 

 sequence presented in the disturbances of June 28 and 29, 1903. 



TABLE LXX. Hour of Occurrence of End of First Phase. 



105. Table LXXI gives particulars of the average duration of the two phases, again from the whole 

 82 occurrences. There seems a distinct tendency for the second phase to last the longer, though instances 

 in which the reverse is true are not rare. There seems no marked seasonal variation in the duration of 

 the phases. There is an apparent difference between 1902 and 1903 which may be real. 1902 was a year 

 of very small sun-spot frequency (WOUBR'S relative value 5'0), while 1903 had a considerably larger 

 frequency (WOLFER'S value 24 - 4) and showed larger magnetic ranges both for the regular and irregular 

 variations. 



TABLE LXXI. Duration of Phases (Minutes). 



106. An attempt was made to arrive at a more exact idea of the type of disturbance by forming means 

 based on a number of days. Only those occurrences were used for which the trace was complete. 

 Incompleteness was naturally most common in the larger disturbances, so that the mean results obtained 

 somewhat underestimate the average amplitude. Occurrences in which the phases in the different 

 elements did not synchronise, or which were ill defined, were also omitted. The 51 occurrences which 

 remained were grouped according to the season of the year as shown in Table LXXII, which gives the 

 mean results obtained. The Declination results are expressed not in minutes of arc, but in units of force, 

 replacing 1' by T92y as the force acting perpendicular to the magnetic Meridian necessary to alter the 

 direction of the Declination needle by one minute of arc. 



