ON THE STATISTICAL VIEW OF NATURE. 551 



a general and unbiassed glance at the world outside, 

 leaving our study, our observatory, our laboratory, our 

 dissecting- or our measuring-room, and ask ourselves 

 the simple question, By the work carried on in these 

 various secluded places, in the " sapientum templa 

 serena," how much of the world outside have we really 

 learnt to comprehend, or even only to describe and 

 picture to ourselves correctly and completely ? The 

 answer is hardly encouraging. The first thing we 

 notice in stepping .out of our door is a phenomenon 

 still as incalculable as it has ever been, and yet bound 

 up with the enjoyment of our lives and the success 

 of our work as much as ever the weather. What do 

 we know of it which is practically reliable and useful ? 

 The reply must be, " Next to nothing." Some general 

 astronomical and some more detailed physical and 

 chemical relations permit us to describe a few general 

 meteorological and a few recurring seasonable events, but 

 scarcely with more practical detail and certainty than the 

 unscientific ancients or the untaught children of nature 

 of to-day. We know in general the cause of storms, of 

 changes of temperature, of the seasons, of rain, hail, 

 drought, and cold, but we do not know much more of 

 the exact when and where of these various changes 

 than did our forefathers. The natural atmosphere and 

 climate which surround us are still elements of con- 

 jecture and uncertainty. 



Assume, however, that we go a step further, and 

 having accustomed ourselves to take the weather, good 

 or bad, as it is, enter into the artificial atmosphere and 

 surroundings of practical life, of industry, trade, and 



