342 ATMOSPHERE IN RELATION TO HUMAN LIFE AND HEALTH. 



Per cent. 



Motion of cirrus, fast 23 



Motion of cumulus, very slow 8 



Vertical height of cumulus compared with breadth, great 73 



A few waves or close ripples of well defined hard cirrus strata nearly overhead. 84 



Length of steam trail, moderate (estimated 90 yards) 52 



Color of clouds at dawn, pale yellow 58 



Regular or irregular distribution of clouds (?) 



Regularity or variability of temperature and humidity in adjacent strata, etc. ( ?) 



[Probability of rain in twenty-four hours.] 



Visibility, great - 70 



Audibility, great 61 



Humidity, difference of bulbs, 4 degrees 46 



Humidity (increasing or diminishing), diminishing 29 



High clouds, increasing 68 



Cirrus (straight or tangled), tangled 81 



Stars last night, much twinkling 71 



Smoke, tending downward 69 



Total 877 



Probability, rain. 



The number of items in the forecast might be much increased with 

 increasing knowledge, and the value of each sign would also increase 

 with continuous exact observation. Moreover, each sign should be 

 studied not as a single item, but as occurring with others, and when 

 considered in relation to others would gain much in value. Thus, visi- 

 bility is not infrequent in fine dry weather, and also occurs in moist 

 weather, before rain. If observed day after day in fine weather, its 

 value in forecasting is evidently much less than when occurring in 

 somewhat unsettled weather. In fact, each sign has properly a partic- 

 ular value in particular kinds of weather, and the special value has to 

 be ascertained. The length of time during which a certain type of 

 weather has continued is in some proportion to the probability of the 

 ensuing days being of a similar type. 



When the total of the various percentages exceeds a certain fixed 

 amount, the probability of bad weather rises to something approaching 

 certainty, and perhaps the probability of fine weather when the amount 

 is minus goes a little further still. When, in addition, the probability 

 announced by the central office from wide data is in the same direction, 

 it becomes justifiable to place reliance on the forecasts for agricultural 

 purposes and general district warnings. It will also eventually be of 

 great use to farmers to have telegraphic information forwarded to dis- 

 tricts toward which bad weather is moving, if there is reason to regard 

 the change as more than local when first noticed. 



On some Possible Modifications of Climate by Human 



Agency. 1 



There can be no doubt that some effect upon climate, shown more 

 by physiological influences upon mankind than by instrumental records, 



1 This section is derived from MS. written in 1891, but not in any way published. 



