634 BIRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, 19021903. 



It is impossible, however, to discover from the above Table any distinct increase in the length of 

 the tail when the comet is in the vicinity of the plane of the sun's equator. The greatest length of tail 

 is always found after the passage of the perihelion, and this indicates that a prominent part is played by 

 evaporation of the constituents of the nucleus, brought about by the radiant heat of the sun. 



It would appear, however, from the graphic representation (fig. 229) of the variation with time in 

 the length of the tail, that it is not the passage of the perihelion alone that is decisive. The passage 

 of the perihelion is marked P, the time when /? is marked Z?j or B z , answering respectively to the 

 first and second intersections with the plane of the equator. Finally we have the point of time, A, at 

 which the angle has its minimum, in those cases in which this point falls within the period of time 

 under consideration. 



It may be remarked as a general characteristic, that the curves about the maximum of length of 

 tail have very steeply ascending and descending branches. Further, this maximum sometimes occurs a 

 comparatively short time after the passage of the perihelion - e. g. the comet 1862 III and sometimes a 

 comparatively long time after e. g. the comet 1861 II. On the whole, the length of this interval varies 

 considerably, and there does not appear to be any simple connection; the impression is rather, that the 

 great development in the length of the tail about the maximum takes place at the time when the comet 

 is passing certain especially favorable strata or zones. This is especially marked, for instance, in the 

 comet 1862 III. 



There are two other circumstances in particular to be considered here, namely, whether the light 

 of the moon can obliterate the faint light of the comet's tail, and whether, during the period under con- 

 sideration, the tail of the comet has moved much farther from, or much nearer to, the earth. 



In only the first of the cases considered is it noted that the light of the moon has interfered, and 

 this is shown in the curve. 



With regard to the second of the above-mentioned circumstances, it is easy to estimate from the 

 angles a and r whether the distance from the tail of the comet varies so greatly as to have any signi- 

 ficance in judging of the light. In no case does it appear to exert any real influence during the period 

 about the various maxima. 



On closer inspection it appears that the great development of the tail occurs most frequently at a 

 certain distance from the sun's equator, answering to values of ft of between 15 and 30. 



In this connection, one recalls how the sun-spots also occur most frequently in about 20 helio- 

 centric latitude. 



The comet 1618 III exhibits a peculiar circumstance, the curve for the length of its tail having a 

 distinct intermediate minimum. This might be due to the comet's having passed through two layi ; 

 pencils of rays from the sun, one immediately after the other; but it is perhaps just as likely that the 

 peculiar condition might be due to internal causes in the comet, or to the disturbing influence of moon- 

 light, or to unfavorable atmospheric conditions. 



It would be natural, therefore, to compare the above-mentioned layers that were favorable to the 

 development of comet's tails with the pencils of the strongest and magnetically stiffest corpuscle-rays 

 which we imagine to emanate from the region surrounding the sun-spots, and which, when they sweep 

 past our earth, produce powerful magnetic disturbances. It may be that it is these very rays, with their 

 abundance of energy, that can charge the comet mass to a high negative tension, and thus occa 

 sion the secondary electric discharge from the comet into space. 



One circumstance that speaks strongly in favour of a hypothesis such as this, is the greater de- 

 velopment thought to have been found 'in years of sun-spot maxima than in years of sun-spot minima. 

 This has been demonstrated, for instance, in Encke's comet, by BERBERICH and BOSLER, the latter having 

 given an exceedingly interesting graphic representation of this condition, which is reproduced here. 



