504 DanieJl on a new Hi/gromeler. 



<iay is very small, the indication of the weather must be taken 

 i^ore from the actual rise and fall of the point of condensation, 

 than from the difference between it, and the temperature of 

 the air. In showery weather the indications of the in- 

 strument vary rapidly three or four degrees, and a person 

 making observations at short intervals of time, may easily 

 predict the approach of a storm. On the 30th of August 

 I had occasion to make a great many experiments in the 

 course of the day, which was marked by many heavy 

 showers. The morning was fine, and the first observation 

 marked the point of precipitation 10° below the temperature of 

 the air. In half an hour after it was only 6°. At the interval 

 of another half hour, it had again returned to 10°. During 

 this period a very heavy cloud passed, and although there was 

 no rain at the place of observation, I afterwards found that a 

 smart shower had fallen within a thousand paces. Subse- 

 quent observations gave a difference of only 2°. Slight showers 

 also fall sometimes in summer, without affecting the hygro- 

 meter, but they are of very short continuance, and probably 

 come from very lofty regions. They are rather indications of 

 fine weather than the contrary. 



Fogs also and mists must be taken into consideration. They 

 produce the same effect upon the instrument as the greater pre- 

 cipitations of rain. Thus, on the evenings of the 15th and 

 18th of October, the hygrometer shewed the utmost point of 

 humidity during the fogs which prevailed, while the great fall 

 in the degree of condensation shewed that the^effect would not 

 be great or lasting. A change from fine weather to rain is 

 more quickly perceptible in low situations than one from wet 

 to fine, for the effect of a shower lasts rather longer than the 

 state of the atmosphere in higher regions would warrant, on 

 account of the damp exhalations from the moistened ground. 



Winds, moreover, sometimes produce an effect before their 

 influence is pointed out by other indications. Many instances 

 occur in the journal in which a change of wind has affected 

 the hygrometer before it has changed the direction of a 

 weathercock, or even the course of smoke. Thus, on the 3d 



