294 



Messrs. Scott and Galloway on the Connexion [Apr. 1 8, 



subject of the present communication, but it does not call for special 

 notice. 



Mr. Dobson's is the most elaborate discussion of the subject which has 

 yet appeared, and we must devote a few words to it. 



One of the diagrams in his paper exhibits 491 explosions in five-day 

 periods, and from it the author concludes that the annual march of the 

 curve coincides with the march of temperature, with a relative maximum 

 occurring in the autumn, which he attributes to the frequency of serious 

 storms at that time of the year. 



To this we can only say that the maximum and minimum of annual 

 temperature do not occur at the precise epochs named by Mr. Dobson, 

 and that the last quarter of the year cannot be considered as preeminently 

 stormy, at least not more so than such a month as the January (1872) 

 which has just passed. 



In order to test this question of periodicity, we have collected from the 

 Reports of the Inspectors of Mines all the recorded explosions for the last 

 twenty years. These explosions are only the fatal ones, and include a few 

 accidents due to suffocation by "choke-damp," or carbonic-acid gas. We 

 consider that we may fairly count such accidents as due to causes closely 

 related to those which produce explosions of " fire-damp." 



We plotted the whole of these explosions in two decennial periods for 

 the seventy-three intervals of five days each, and found that the curves 

 hardly showed any agreement with each other, so that no confirmation 

 was obtained for Mr. Dobson's alleged periodicity. The curve for the 

 entire period of twenty years, including 1369 accidents, was also con- 

 structed, and all that is worth notice about it is that the number of acci- 

 dents is somewhat greater in October than in other months. The absolute 

 maximum falls at the end of January, and the absolute minimum in the 

 middle of September. 



From a comparison of the two decennial curves we learn that the number 

 of fatal accidents has decreased from 768 in the former (1851-1860) to 

 601 in the latter period (1861-1870), notwithstanding the great augmen- 

 tation of the output of coal during the time in question. 



We do not consider it worth while to reproduce these curves. 



At this juncture there is one matter which deserves attention, viz. that 

 the number of serious accidents, involving the loss of ten or more lives, 

 has very materially increased during the last five years. The number of 

 such explosions during the years 1851-1855 (inclusive) was 13; 1856- 

 1860, 15; 1861-1865, 12 ; and 1866-1870, 21. On this matter we do 

 not offer any opinion, but it seems to us desirable to place the figures on 

 record. 



. For the purposes of the present paper we have taken the continuous 

 records furnished by the Observatory of Stonyhurst, one of the seven 

 established by the Meteorological Committee. These records exist from 

 the 1st of January 1868, and the curves for the years 1869 and 1870 have 



