EXPEUIMENTS AVITII tONIZKI) AIH. 75 



The results of figure 6 are surprisingly irregular, seeing tliat some pains was 

 taken with the work, but tliey map out a grapli of an exponential character to 

 which diflfereut interpretations may be given. Either tliere is a decided falling oif 

 of the ionizing potency of [)hosphorus, or the ions are subject to a potential or a 

 concentration gradient, or tliey destroy eucli other in the lapse of time. These 

 cases will be discussed at length below. It will be noted that the room was quite 

 warm, nearly 30° C. at the place of experiment. This may account for the I'emai-k- 

 ably large conductions (tf.s'/rfO obtained in these experiments, as compared with 

 the data below. The difference between positive and negative chai'ges was not 



tested. 



4. Modifed apparatus. — -Before continuing the work it seemed advisable to 

 make trials to secure moi'e ])erfect insulation. It was found to be nearly perfect 

 without the condenser (current /=:.0010 sc. p.), but uncertain when the condenser 

 was installed (/=.0010, .0005, .0002). This is often too near the data themselves 

 (for K 6, /=.00(JO in extreme cases of smallness); but, since the cui-rentsare derived 

 from differences, the effect of the leak is reduced. No means were found for 

 I'emovino- it j)erfectly (paraffine, rubber, etc., wei-e tried in succession with no ad- 

 vantage), and a dry, clean glass stem (7, figure 1, held in i)laee by a wide hard- 

 rubber sheath F, seemed the best remedy. 



Experiments were then tried to ascertain the difference Iwtween a condenser 

 closed as above and one left open for the stem to pass through, as shown in figure 

 2. Here the wire and ping attachment puts the outer surface to earth at E. The 

 closed condenser always introduced the larger leakage as shown for instance in the 

 "■raphs ' fio-ure 7 for the given condenser K 6. Care was also to be taken to avoid 

 friction in manipulating the stem 0. 



5. New observations. — Conformably with the experience gained tlie condenser 

 in the following experiments is left open above, as seen in figure 2, and is put to 

 earth at Ehy a brass plug wedged into the neck. The tube (J terminating in the 

 ball of yihosphorus is kept clean and dry, and extends to 30 centims. above the con- 

 denser. Unfortunately there is some escape of ions around the neck, and the access 

 of air here is a menace, but there seems to be no way out of the dilemma. 



To estimate the loss of ionizing activity of the pliosi)horus pellicle, one of the 

 condensers (K 6 and K 4, I'espectively) was treated as a standai'd and observations 

 were made upon this before and after each of the observations of the other con- 

 densers. Table 2 ami the corresponding curves, figure 8, are obtained in this way. 

 The scheme adopted limited all observations to an interval of 5 minutes. Readings 

 were taken per minute, and by dividing them into 2 parts the rates 6s/6t were 

 obtained linearly from data 3 minutes apai't. The table also contains other constants 

 to be discussed below. In figure 8 the numerals attached to the curves show the 

 number of the condenser and the order of measurement. 



Curiously K 6 shows a linear diminution of potential or constant current, while 

 in the other cases the currents usually fall oft". Again, the values of ds/St for K 6 



' Curves marked " closed " were obtained before and after the curve marked " open." This 

 eliminates the fluctuations of the ionizer. 



