A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 



23 



for water nuclei. I have not, however, with the data now in hand been able to 

 complete this to my satisfaction, and have for this reason confined the present 

 chapter to experimental work, to the exclusion of theoretical considerations of 

 the kind given tentatively elsewhere.' 



7. Apparatus. — If in the receiver, or condensation chamber, .4, the 

 metalhc pipe, e, joining at b, leads directly to the tubular condenser, C (radii 

 1.05 and .34 cm., length, 50 cm.), the apparatus, figures i, 2, takes the fonn 

 adapted for measuring the initial ionization of the nuclei. If the cocks k and d 

 (filter) are closed and the fine radial jets are put in action by opening the water 

 faucet h, the charged air is gradually expelled through b as the w^ater level in A 

 rises. When an efficient jet is used the rate is usually about 2 Hters/ minute. 

 This velocity may be increased or diminished by aid of the flask, M, figure i, 

 attached at k. 



Since the jets, y, impinge on the walls of the vessel, this is kept uniformly 

 moist or better coated with water, and therefore continually put to earth by 

 the hydrant connection. Similarly the pipe, e, leading to the outer coating of 

 the condenser is with this continually put to earth. The core of the condenser, 

 insulated by long hard laibber supports, retains charge well even at high poten- 

 tials, and in spite of the damp gases, because of the remoteness of the supports. 



8. Results. Initial charges. — The following table gives the data obtained, 

 when the nuclei generated by the spray are at once passed into the tubular con- 

 denser, whose inner surface is charged as stated, the outer being put to earth. 



TABLE 5.— IONIZATION OF WATER NUCLEI. dV/dt=2 LIT./MIN. TOTAL 

 CAPACITY OF ELECTROMETER AND CONDENSER, 72 cm. 



The method consisted in testing the insulation of the condenser, immedi- 

 ately before and after the introduction of water nuclei. The table gives the de- 

 flection in centimeters, after interv^als of i niin., \ min., and i min., in each of the 

 cases, respectively. The conduction, a = S ([ogs)/St, is computed by assuming 

 Ohm's law; but in case of the medium of water nuclei, it is seen at once that 

 Ohm's law does not apply, and that the conduction, a, increases enomiously as 

 ' American Journal of Science (4), xv, 1903, p. 105; ibid., p. 217. 



