416 RECORD OF SCIENCE FOR 1887 AND 1888. 



land, Virginia, and North Carolina in the early morning following a 

 clear evening. 



(e) A report of rain or snow-fall during any interval indicates that 

 the process of cloud formation has gone onto its extreme limit; the 

 quantity of rain is an index to the relative amounts of latent heat tem- 

 jiorarily left behind in the cloud stratum to be slowly lost by radiation ; 

 hence it shows the disturbance over different parts of the country of 

 the normal distribution of density. Where however only little rain 

 has fallen while the cloud stratum is extensive and dense, the total dis- 

 turbance due to the general warmth of air and cloud over the whole 

 region may be equally great, but not especially intense at any one point. 



(VII) Prediction of the weather. — The preceding considerations, which 

 are those that guided me in my earliest predictions in Cincinnati and 

 Washington, will enable any one to enter upon any series of charts of 

 existing conditions a system of numbers showing the places over which 

 the atmospheric disturbance in the cloud region is — or is likely to be — 

 appreciable, and to judge of its relative intensity as depending on the 

 excess or deficiency in the density of the air. From this we can reason 

 as to the first tendency of the movements of the air or the initial wind. 

 The consideration of the orography of the surrounding country will then 

 give us the basis for a conclusion as to whether in any region the gen- 

 eral movement of the atmosphere will tend to cool or warm any mass of 

 air (especially by carrying it over rising or resisting ground and forcing 

 it up to the higher levels or by pushing cold air under warm air), and 

 thereby initiate a still further condensation and exaggerate the disturb- 

 ances already existing, or possibly producing local rain and wholly new 

 centers of disturbance. By thus considering as fully as possible the 

 effects of solar heat, atmospheric moisture, orography, etc., we are able 

 to enter upon the weather map a series of figures showing our estimate 

 of all the disturbances likely to occur during the immediate future. In 

 general, we find that convective disturbances of equilibrium tncrease 

 from the morning map to the afternoon map, and diminish from the 11 

 p. M. to the 7 A. M. map; during the interval from 3 p. M. to 11 P. M. 

 the larger disturbances, such as thunder storms, tornadoes, and hurri- 

 canes, generally attain a maximum and begin *to diminish. The phe- 

 nomena specially due to terrestrial radiation have their maximum effect 

 between the 11 P. m. and 7 A. m. reports; between these hours we find, 

 for instance, that many extensive storms begin over regions for which at 

 11 P. M. the reports show clear weather. As the prevailing and probable 

 future direction and force of the wind are so important, it becomes very 

 necessary, in dealing with a limited part of the earth's surface, to ex- 

 amine the reports from outlying border stations in order to ascertain 

 the possible existence at a distance of unseen centers of disturbance ; a 

 slight abnormal fall in the barometer, or rise of temperature, or shift of 

 winds, or increase in cloud and rain may be the indicator of a distant 

 storm ; equally frequent is it that a general rise of pressure or increase 



