TUK STKUCTUKK OK THE NUCLEUS. 65 



placed in these rates of subsidence, complicated as they are with so many vague 

 secondary phenomena. 



A comparison of tables 5 and shows that the Jinal velocities in the experi- 

 mental curves (.017-.050 cni./sec.) correspond very closely to the velocities com- 

 puted in table 5 (.021 cm./sec, etc.) for the particles produced in the earlier orders 

 of exhaustion (z = 0, etc.). The two tables, therefore, mutually corroborate each 

 other, proving that the order of values investigated for the size of the cloud 

 particles is reasonably cori-ect ; pi'oving also that the initial apparent subsidence ia 

 spurious, as anticipated. In fact, if water is pi-ecipitated by sudden cooling from 

 atmospheric air in an otherwise dry vessel {Of. Chapter VI, § 10), the apparent sub- 

 sidence is astonishingly I'apid, being of the order of more than meter/second. 

 Clearly, from the excessively small particles precipitated, this can be nothing but 

 quick evaporation undei' conditions of minimum moisture. As the temperature 

 rises after the exhaustion, the air at once ceases to be saturated. 



A few rough estimates may be made for the globe, 30 cm. in diameter. 

 The descent of the body of fog per minute will be about 1.5 cm. during the first 

 exhaustions (s = 1), and about 11 cm. during the final exhaustions {z = 26). The 

 loss of particles may be taken as the ratio (^f the volume submerged to the total 

 volume of fog, and the i-esults show that dui-ing the first exhaustions a loss some- 

 what exceeding 5 %, during the last exhaustions a loss approaching 50 %, is to be 

 apprehended. Hence, the time during which subsidence occurs (presence of fog 

 within the receiver) must be reduced to a minimum, only maintained just long 

 enough to make the observation. But even in this case the subsidence error 

 toward the end of the series becomes increasingly menacing, and to it are to be 

 ascribed the jagged outlines of the data for s/fN in tables 3 and 4. Very 

 fortunately, the subsidence error is in large measure counterbalanced by the evapo- 

 ration of particles, and for this reason the removal of particles by subsidence has 

 been ignored in this chapter. Any correction applied would be arbitrary; but 

 there is no question that the results could be raateiially improved if a new form of 

 apparatus were contrived adapted to insure the utmost dispatch between the 

 exhaustions and observations. 



12. Normal and other coronas. — With particles as near together as they are 

 originally (.02 to .03 cm.), it is not remarkable that other optical phenomena may 

 make their appearance due to the mutual action of the edges of the particles, and 

 it is not improbable that this mutual effect contributes to the colored central patch 

 of the coronas, occui-ring as a superposition of a new phenomenon on the old. The 

 normal coronas do not begin until the number of particles is 530 or 600 per cub. 

 cm., or about 8 per linear cm., putting them somewhat more than 1 millimeter apart. 

 Normal coronas are seen until the pai'ticles are about 2 millimeters apart, after 

 which they shrink beyond observation. The particles themselves are but .0014 to 

 .0025 cm. in diameter, at the beginning and the end of the series of normal coronas 

 respectively, and are thus very small as compared with their distance apart. Again, 

 initially— ^. e., immediately after nucleation, whei'e the particles are but .02 cm. 

 apart— their dimensions are .0003. In all cases, therefore, the interstices are large. 



