ANTARCTIC TERM-HOUR RECORDS. 149 



then another quick-run portion covering about 5 minutes, then a gap representing the time occupied in 

 changing papers, and, finally, a slow-run portion answering to the period 11.18 a.m. to noon. It is interesting 

 to compare the quick- and slow-run portions of the D and H curves. At first sight one would hardly 

 realise that the curves were really as quiet towards the end of the hour as at the beginning. 



In all the curves the Declination has the same scale value, 1 mm. EE 1' -5. In the other elements there 

 is considerable variation. In H the value of 1 mm. varies from l'4y to 0'67y, in V it varies from 

 16 '5y to 3 - 2y. Full particulars will be found in Table I, or can be derived from the plates themselves. 

 The curves in Plates VI to X were obtained as follows : 



The original Antarctic curves were photographed at Kew by Mr. W. J. BOXALL, light being transmitted 

 through the sheet on to photographic paper immediately in contact with it. From the negatives thus 

 formed positives were made, and when unduly faint these were intensified with pen and ink. In some 

 cases the base lines shown represent exactly the originals, but in other cases the original base lines were 

 replaced by others occurring in greater juxtaposition to the curves to which they really belonged. This 

 was done to minimise the risk of confusion between the different elements. 



The latter part of the work was done by Mr. BROOKES at Teddington, under the general supervision of 

 Mr. F. J. SELBY. Every care was taken to avoid introducing departures from the original curves. 



In some cases, e.g. November 1 and 15, and December 15, 1902, there is, unfortunately, a good deal of 

 intercrossing of the D and PI curves. It is hoped, however, that by the aid of the letters attached it will 

 be found possible to distinguish the traces. 



The original V trace was lacking on October 1 and November 15, 1902, and on February 1, 1903, while 

 the H trace was off the sheet during the term hours of January 1 and February 15, 1903. 



The curves, being exact copies of the originals, are, of course, uncorrected for temperature, and, as will 

 be seen from the TT curves, when these exist, temperature was far from constant. As already explained, 

 the Vertical-Force magnet had a large temperature coefficient; thus a very appreciable temperature 

 correction would in reality be necessary to bring the values of V derived from measurements of the 

 V curves in some of the hours to a common basis. In other words, the base-line value in V was not really 

 constant, and the values actually assigned it refer to only one instant of the hour. 



The temperature variations seldom, however, showed rapid oscillations, and whilst the temperature 

 change is, in some cases, responsible for a slight general drift up or down the sheet in the V curve, it is 

 practically without effect on the range of the oscillations shown in the curves. 



51. Owing to the difference in the type of the Antarctic magnetographs from those in use at 

 Christchurch or Mauritius, the difficulty of identifying corresponding movements is materially increased. 

 The difficulty is further enhanced by the fact that we have, in the case of the Antarctic, high sensitiveness 

 in the instruments going along with large disturbance. But whatever the cause may be, I have not 

 succeeded in tracing any close parallelism between individual Antarctic term-hour movements and those 

 at eitjier Mauritius or Christchurch. There is, however, a certain amount of parallelism between the 

 Antarctic and Christchurch curves, as may be seen on reference to Table IjXI % which gives the ranges of 

 Declination and of Horizontal Force at the two stations during all the term hours for which information 

 existed. The Christchurch ranges are derived from tabulations, at 20-second intervals, made at Christ- 

 church. The parallelism referred to consists in this, that on the term hours for which the range in the 

 Antarctic is exceptionally low the range was also especially small at Christchurch. The cause, however, 

 so far as Christchurch is concerned, has little to do with disturbance. The hours for which there were 

 specially low ranges in the Antarctic occurred in Midwinter near midnight, and as midnight in the 

 Antarctic and midnight at Christchurch differed by only about 23 minutes, the hours in question were 

 naturally those for which the regular diurnal changes at Christchurch were least. 



