METHODS OF STUDYING VISION IN ANIMALS 71 



mended as the standard source in all general high intensity 

 work. This lamp, supplied by the Arthur H. Thomas Company, 

 Philadelphia, complete for $102, is simple in its mechanism. 

 The feed is purely magnetic. As the two carbons burn away 

 they offer greater and greater resistance to the flow of current; 

 finally a portion of the current is shunted through the feeding 

 mechanism. The carbons are then fed toward each other until 

 the resistance becomes so small that no further current passes 

 through the feeding mechanism. The forward motion of the 

 carbons, therefore, ceases before they are brought into actual 

 contact. The lamp is always practically noiseless, with the 

 exception of the " cluck " or series of " clucks " made by the 

 feeding mechanism. A drop of from i to 3 amperes is suffi- 

 cient to start the feednig device. 



It was long ago shown by Abney that even a large decrease 

 in amperage does not alter the intensit}^ of the crater of the 

 positive carbon. It merely decreases the size of the crater. 

 Since the height of the slit used is less than the diameter of the 

 image, it follows that the variation in the energy of the 

 monochromatic bands is hardly detectable. Of course there 

 are slight imperfections in the carbon, which occasionally pro- 

 duce a sudden flicker and slight shiftings of the positions of the 

 carbons, but such disturbances are momentary and negligible. 

 It is essential to have a large crater. This can be obtained only 

 by the use of a high amperage. Twenty-five amperes is a neces- 

 sity. A positive cored carbon (Elektra) 9-16" in diameter and 

 a cored negative carbon 7-16" in diameter are very satisfactory. 

 If other sizes of carbons are used, or if the amperage is low, 

 there will be a shifting of the crater and a tendency for the 

 two carbons to burn unevenly. Suitable carbons may be pur- 

 chased with the lamp. 



These lamps are very durable. One of them at the Johns 

 Hopkins University has now been in operation two years and 

 gives entire satisfaction. It burns for i^ hours without change 

 of carbons. It has burned daily for months without attention 

 and during that time has never jammed the carbons nor in any 

 other way interrupted an experiment. 



If energy measurements are to be made, it is essential to 

 have both an ammeter and a voltmeter in the circuit with the 

 arc. The voltmeter should read 0-130 by i-volt divisions. 



