SPECIAL ANTAKCTIC RECORDS. 



151 



In the Physical Observation tables there are 16 days for which Declination ranges are given at both 

 stations and 12 for which Horizontal-Force ranges are given (the curve not being beyond the limit of 

 registration). Deriving arithmetic means from these, wo find the following results for the daily ranges : 



The excess of the Antarctic figures is here very much less than for the hourly ranges. 



52. Plate XI gives a copy of a quick-run record obtained during a solar eclipse on September 21, 1903 

 (September 20, G.M.T.). According to Mr. BERNACCHI'S calculations the local times of the beginning, the 

 central phase, and the end of the eclipse were respectively 3.40 p.m., 4.38 p.m., and 5.37 p.m. The 

 quick run was started at about 3.39 p.m. and continued until 5.46 p.m. The curves shown extend 

 from 3.39 p.m. to about 5.39 p.m. They are copies taken with the Schmidt tracer, retaining the original 

 scales for the ordinates, but reducing the time scale about 25 per cent. 



The curves are shown in two portions, the upper panel representing the earlier half. The base lines 

 D Do and H Ho have been shifted from their positions in the originals so as to bring them within the 

 range of the plate. The Declination, or D, curve and the Horizontal-Force, or H, curve cross one another 

 in the second panel. The Declination scale is shown on the left, the Horizontal-Force scale on the right. 

 The Vertical-Force curve to the eye was practically a straight line and so has been omitted along with 

 the corresponding base line and the temperature curve to avoid confusing the trace. 



The primary reason for reproducing the curves is that it is a question of interest whether an eclipse of 

 the sun does exercise an appreciable influence on the variation of the magnetic elements. On various 

 occasions there have been schemes of special observations during eclipses, and several authorities have 

 been inclined to think that the results have shown a diminution in the changes as compared to those 

 ordinarily encountered at the same hour of the day. 



The curves in Plate XI show considerably more disturbance than those on June 15 and July 1, 1902, in 

 Plates IV and V. They answer, however, to a time of year when the changes, both regular and irregular, 

 were larger. It can be readily realised how useful open time-scale curves like these might prove for 

 comparing disturbances of considerable magnitude at different stations. 



53. Plate XII gives an example of a series of regular waves, or " pulsations " as they have been very 

 appropriately termed by Dr. VAN BEIIMELEX. The curves are exact copies of Antarctic quick-run curves 

 of February 2, 1903. The letters D, H, V, T have their usual significance, the base lines being distin- 

 guished by the suffix 0. The scale values appear at the left-hand side. 



The pulsations are most apparent in the Horizontal-Force curve between 9.10 and 9.25 p.m. But 

 they may be detected at all events in the original throughout most of the H curve between 9 

 and 10 p.m., and can be recognised distinctly in the D curve about 9.15 p.m. 



The special feature of pulsations is that they represent wave-like oscillations of force not varying much 

 in amplitude and of at least approximately constant period during a considerable number of wave periods. 

 In the present case there seem to be about 19 complete periods in 1 cm., and, as the time scale is about 

 240 mm. to the hour, the period is accordingly about 8 seconds. 



Very short-period pulsations cannot be seen except in open time-scale curves, but pulsations of 2 or 3 

 minutes' period are not infrequently seen in ordinary curves from Kew pattern magnetographs. The 

 Eschenhagen type of magnetograph is better adapted for showing pulsations, but the Antarctic 

 conditions were so disturbed that pulsations with periods of 2 or 3 minutes would be difficult to recognise 



