More or less parallel with this line are the lines on which annually nearly the same number of Auroræ are seen. The line for one Aurora annually went from Bordeaux, through Switzerland, past Krakau, south of Moscow and Tobolsk, to the northern end of Lake Baikal, on to the Sea of Ochotsk and to the Southern Aleutes, thence through Northern California to the mouth of the Mississippi and to Bordeaux. The line for five Auroræ annually went from Brest through Belgium, Stettin, Wologda, between Tobolsk and Beresow, parallel to the previous line to Ochotsk, and on to Brest, &c. Almost exactly with the line of greatest frequency coincides the line forming the boundary of the direction of visibility of the Northern Light towards the Pole or towards the Equator; while northwards of this line the Polar Light is seen in the direction towards the Equator; and from all stations the Northern Lights are seen in directions which are pretty much normal to that curve and the entire system of isochasms. [Sidenote: Assumed connexion between Aurora and ice-formation.] Professor Fritz has remarked that the curves of greater frequency tend towards the region of atmospheric pressure, and also that they bear some relation to the limit of perpetual ice—tending most southward where, as in North America, the ice limit comes further south. He also endeavours to show a connexion between the periods of maximum of Auroræ and those of ice-formation, and considers ice to be an important local cause influencing their distribution. These being most frequently seen over open water in the Arctic regions, has been referred to as noticed by Franklin and others. _Extent and principal Zone of the Aurora._ [Sidenote: Extent and principal zone of Aurora. M. Moberg’s Finland observations (1846-55) compared by Prof. Fritz with those in other regions.] The Finland observations, published by M. Moberg in his ‘Polarlichter Katalogue’ of Northern Lights in the years 1846-55, numbering 1100, have been compared by Prof. Fritz, in his paper in the ‘Wochenschrift für Astronomie,’ with the auroral phenomena of the same period in all other regions. The Table shows that of 2035 days of the months August to April on which Northern Lights were seen, 1107 days were those of Northern Lights for Finland. On 794 they were visible simultaneously in America, and mostly also in Europe; on 101 days in Europe only, and on 212 days in Finland only. On 958 days Northern Lights were visible in Europe and America which were not visible in Finland. All these numbers refer only to the months August to April, as in the remaining months the brightness of the night in Finland makes such observations impossible. The conclusion is arrived at that a large portion of Auroræ have no very great extension, or that the causes producing the phenomena must often be of a very local character; while in another portion of the phenomena the extent, or the regions of simultaneous appearance are very considerable.