THE 

 LONDON, EDINBURGH and DUBLIN 



PHILOSOPHICAL MAGAZINE 



AND 



JOURNAL OF SCIENCE. 



[FOURTH SERIES.] 



OCTOBER 1858. 



XXVI. On the Distribution of Heat over Islands, and especially 

 over the British Isles. Bij Henry Hennessy, F.R.S., 

 M.R.I. A., Professor of Natural Philosophy in the Catholic 

 University of Ir eland''''. 



NO element among the conditions of terrestrial climate is so 

 important, none has engaged so much attention, nor has 

 any other been systematically observed for so long a period, as 

 temperature : yet the time is comparatively recent when philo- 

 sophers commenced to consider the laws of its distribution over 

 the earth's surface in a truly scientific spirit. In 1 779 appeared 

 the mathematical inquiries of Lambert, in which an attempt was 

 made to estimate the difference between the heat received by 

 the earth from the sun and the heat which it loses by radiation. 

 Mayer had about the same time deduced his well-known law 

 from theoretical grounds, and by considering solar radiation alone 

 as the source of terrestrial heat. Towards the close of the last 

 century oiu- countryman, Richard Kirwan, attempted for the 

 first time to compare Mayer's law with existing observations, 

 and thus to arrive at general views regarding the climate of our 

 planet. Humboldt followed up this step by one of still more 

 importance when he published his essay on Isothermal Lines. 

 Laying aside speculative considerations, he presented the results 

 of actual observation in a way at once novel and luminous. 

 Having found the mean annual temperatures of a great number 

 of stations, he compared them together and marked on a map 

 the places which had equal temperatures. The points of equal 

 temperature having been joined by curves, the forms of the iso- 

 thermal lines thus produced present a graphical picture of the 

 * From the Atlantis for July, 1858 : revised and communicated by the 

 Author. „ 



Phil. May. S. 4. Vol. 16. No. 107. Oct. 1858. R 



