January 30, 1903.] 



SCIENCE. 



173 



led to believe that the indications at 3 m 

 were real. The gi-eat variation in the in- 

 tensity of these lines may be due, in part, 

 to the fact that the slit subtends different 

 angles in the two regions, and that the 

 suspected line at 3 /^ is isolated, while the 

 others occur in a group in which the inten- 

 sity of each one is influenced by those ad- 

 joining it. 



The presence of the bands of larger 

 wave-lengths than 4 ^ indicates that the 

 true radiant efficiency of the arc is lower 

 than the values found in a previous in- 

 vestigation. 



Experiments concerning Very Brief Elec- 

 trical Contacts: Herschel C. Parker, 

 Columbia University. 

 A series of electrical contacts giving a 

 fairly accurate range of adjustment from 

 0.1 second to 0.00001 second would fur- 

 nish a valuable means of investigation. A 

 gravity contact key devised by Dr. Charles 

 Forbes gives promise of fulfilling the above 

 conditions. The writer has made many 

 determinations of the times of contact 

 given by the various devices employed on 

 this key, and has also investigated the 

 times of contact of several forms of pen- 

 dulum. 



The method employed was as follows: 

 a condenser of knowTi capacity (F, farads) 

 was charged during the time of contact 

 (T) and the deflection on discharging 

 noted. This deflection (if a good mica 

 condenser is used which has no absorp- 

 tion) is proportional to the electromotive 

 force (E) and the capacity (F). The 

 condenser is again charged through a re- 

 sistance (R) and the deflection (Q) ob- 

 served. Then : 



and. 



Q = EFX{i—e — TIRF) 

 T= — BFX\oge{l — QIEF). 



The 'gravity key' consists essentially of 

 a rectangular weight falling on metal 



guides, the key being furnished with a 

 scale divided in fractions of a second, ac- 

 cording to the law of falling bodies, and 

 the weight actuating the various forms 

 of switches employed. If two switches 

 are used, one to make the contact and the 

 other to break the contact, by placing them 

 at dift'erent distances apart on the scale, 

 times of contact varying from 0.4 second 

 to 0.001 second may be obtained. For 

 shorter times a single switch that makes 

 and breaks the contact is made use of, and 

 the time made faster or slower by placing 

 in different positions on the scale so that 

 the falling weight strikes it with varying 

 velocities. 



In one form, the weight moves the short 

 arm of a lever, the long arm passing over 

 a contact strip. Another form is one in 

 which the fulcrum of the lever changes, 

 first giving contact and then breaking the 

 circuit immediately afterwards. In still 

 another type the falling weight strikes a 

 lever arm and releases a spring, which 

 makes the contact, and a further motion 

 of the lever breaks the contact, thus giving 

 a differential effect between the velocity 

 of the weight and the rapidity of the 

 spring. With this key it is possible to 

 obtain a contact of only 0.000017 second 

 and with careful adjustment it seems pos- 

 sible to reach 0.00001 second. 



Experiments made with pendulums con- 

 sisting of a steel ball suspended by a wire, 

 and striking against a steel anvil, gave very 

 positive and satisfactory contacts. Using 

 a pendulum with the suspension wire about 

 four meters long and the steel ball two 

 inches in diameter, an arc of |° gave 

 0.00039 second, while a pendulum with a 

 short suspension wire using one-half-inch 

 steel ball, through an arc of 90° gave 

 0.000079 second. 



It is interesting to note that in worldng 

 with condensers the best mica condenser 

 gives no appreciable variation in capacity 



