948 



SCIENCE. 



[N.S. Vol. XVI. No. 415. 



ogist, the specimen in my hands will be 

 turned over for examination to an expert. 

 John M. Clarke. 



SHORTER ARTICLES. 

 PRELIMINARY RESULTS ON THE CHANGES OF AT- 

 MOSPHERIC NUCLEATION.* 



1. Last May, Mr. Harvey N. Davis, at my 

 request, put up an apparatus in this labora- 

 tory for counting the number of nuclei in the 

 atmosphere by measuring the coronas pro- 

 ducible with such air under appropriate con- 

 ditions. The apparatus gave promise at once, 

 but Mr. Davis was unexpectedly called away 

 before the observations became fruitful and 

 the project was temporarily abandoned. Be- 

 lieving that an instantaneous method of at 

 least estimating the degree of atmospheric 

 nucleation is a desideratum,t and must throw 

 light eventually on the origin and character 

 of the nuclei in the atmosphere, I have recently 

 undertaken the work myself, and the results 

 obtained in October, after the indications of 

 the apparatus had become warrantable, are 

 given below. 



I may add that Mr. Davis, and later Mr. R. 

 Pierce, Jr., had been at work for some time on 

 the measurement of the daily variation of the 

 solar constant (a project recently set on foot 

 by the U. S. Weather Bureau) and that I 

 hoped from a coordination of the two classes 

 of data to reach conclusions of interest. 



2. Apparatus. — The original apparatus was 

 of an improvised kind, consisting of a large, 

 horizontally placed aspirator flask or receiver 

 (about 10 liters in capacity, 30 cm. long and 

 20 cm. in diameter) in which the coronas 

 were produced, an exhaustion reservoir and 

 appurtenances. Atmospheric air entered by a 

 quarter-inch lead pipe, and after passing 

 through a coil of pipe in a water-bath, kept 

 at room temperature as shown by a thermom- 

 eter, entered the receiver and was there satu- 

 rated with water. Some of the nuclei may be 

 absorbed in this necessarily long and thin in- 



* Read to the American Physical Society, 

 October 18, 1902. 



t The pioneering work of Aitken is well known. 

 His apparatus, however, would be inconvenient for 

 the purposes here in view. 



flux pipe, though the tests made did not bear 

 this out. At all events the absorption is pro- 

 portional to the nucleation and will not af- 

 fect the ratios of successive nucleations in the 

 lapse of time which are here chiefly in ques- 

 tion. In later experiments the receiver was 

 replaced by a cylinder 50 cm. long and IS 

 cm. in diameter, the walls of which, in the 

 absence of plate glass apparatus, ijroduced 

 less distortion. To measure the ai^ertures of 

 the coronas produced in the receiver, a hori- 

 zontal goniometer was placed about one metet" 

 in front of it and the small circular source 

 of light about two meters behind, all being 

 at the same level. 



3. Method. — It is necessary to take in the 

 air at some feet from' the laboratory; when- 

 ever the house is colder than the atmosphere, 

 there is a draft outward and one is apt to 

 catch the ventilation. The influx pipe must 

 be scrupulously without leaks for the same 

 reason. 



Since coronas actiially run as far as the 

 green centered types, considerable variation 

 is detected and a skilled eye may often dis- 

 pense with the goniometer. For this reason 

 distortion of the coronas due to the walls of 

 the flask is of little consequence at the outset. 



In the tables the date and hour, the aperture 

 of the corona, the character of the weather 

 and the temperature of the influx air were 

 taken. From these, the number of nuclei, n, 

 per cubic centimeter was computed. The 

 measurements showed (if s is the chord of the 

 aperture on the given goniometer), 



d — .002/.S ( 1 ) 



in centimeters, where d is the diameter of the 

 fog particle. At mean atmospheric tem- 

 perature (21° C.) one may put m = 79/10" 

 grams, as the moisture precipitated per cubic 

 centimeter of air, for the pressure difference 

 Sp = 17 centimeters of mercury; used 

 throughout. From this the number of nuclei 

 per cubic centimeter becomes by equation (1) 

 „,= 189Xs«. (2) 



4. Errors. — The occurrence of s' in (2) 

 makes it unreasonable to expect very sharp 

 data for n, since s is not obtainable from 

 sharp lines. Differentiating equation (2) 



