324 Notices of Memoirs. — Colliery Explosions. 



II.— Colliery Explosions and Weather. Being an abstract of a 

 paper read before the Eoyal Society on the 18th May, drawn up 

 by RoBT. H. Scott, F.E.S., Director of the Meteorological Office, 

 and W. Galloway, Esq. 



AFTER a preliminary reference to previous papers on the subject, 

 and especially to tiie diagrams published by Mr. Joseph Dickin- 

 son and by Mr. Bunning, of Newcastle-on-Tyne, the authors of the 

 paper referred specially to Mr. Dobson's paper, published in the 

 reports of the British Association. They showed that the periodicity 

 alleged by him to exist in these explosions had no real foundation 

 in fact, for ori plotting the dates of the explosions for the last 20 

 years in two 10-year periods very slight resemblance was seen be- 

 tween the two curves. The number of accidents, all fatal ones, on 

 which this statement was based, was 1369. 



In the progress of this inquiry it had come out that the number 

 of serious accidents, involving the loss of 10 lives or more, had 

 materially increased during the last 5 years, the numbers being — 

 1851-5 XIII. 1856-60 ZV. 



1861-5 XII. 1866^70 XXI. 



These numbers ajopear to be well WiOrfhy of remark. 



For the special purpose of the paper the continuous records from 

 Stonyhurst, one of the observatories in connexion with the Meteoro- 

 logical Oiffice, were taken, and the curves for the barometer and 

 themometer were plotted for the 3 years 1868-70. The records of 

 fatal explosions were obtained from the published reports of the 

 inspectors, while the dates of the non-fatal accid^ents were obtained 

 from the inspectors themselves, who, almost without exception, 

 replied to the communications addressed to them, and furnished the 

 desired information. Mr. Dobson, in his paper, having spoken of 

 the explosions occurring principally at the commencement of a 

 storm, the authors showed that it was not in some cases until two 

 or three days after the barometer had reached its lowest point that 

 the accident happened. They showed also why, during a period of 

 continued violent oscillation of the barometer, the passage of each 

 successive barometrical minimum is not characterized by an equal 

 number of explosions, the largest groups of accidents being reported 

 where a serious break occurred after a period of calm weather. 

 The effect of a high temperature of the air in interfering with 

 ventilation, and especially with natural ventilation, was also ex- 

 plained, and it was shown how the first hot days in spring were 

 marked by explosions. 



The actual dates of the explosions for the three years in question 

 were then compared with the meteorological records, and it was 

 shown that out of 550 explosions — 



266 or 48 per cent, might be attributed to the state of the barometer; 

 123 or 22 per cent, to the state of the tbermoraeter ; 

 161 or 30 per cent, remained unaccounted for on meteorological 



grounds. 



The next point touched upon in the paper was the action of a 

 more or less impure ventilating current in increasing the explosive 



