538 Prof. Arthur Schuster on 



white light does not consist of alternate bright and dark 

 bands, hut is represented by the curve C (fig. 7), and has only 

 one well-defined minimum. There would be no interference 

 at all, in the sense the word is used here, if we could have a 

 source of light giving out all radiations with equal intensities. 

 If we can draw a curve such as that given, it is only because 

 we know that the energy of our available sources of white 

 light is distributed in a way which cannot differ much from 

 that represented by the curve D (fig. 7). There would be no 

 difficulty in extending our investigation to the case of thin 

 films and other cases in which the source of light is multi- 

 plied more than twice. The general result, however, would 

 not be altered ; whatever interference there is, must depend 

 on the prevalence of some particular wave-length in the 

 original spectrum. 



20. Imagine, now, a photographic film to be substituted 

 for the screen in the last section. The curve G (fig. 7) would 

 evidently not represent the effect on the film, for the bands 

 commonly associated with interference -effects would appear ; 

 this indicates an interference phenomenon which is not in- 

 cluded in the cases we have hitherto discussed. To explain 

 how the photographic film can produce interference where 

 it previously did not exist, we may think of a pendulum 

 which is set in motion by a blow and moves without fric- 

 tion. Suppose that after a certain lapse of time the blow 

 is repeated. The second blow may increase or diminish 

 the momentum due to the first blow according to the rela- 

 tion between the interval of time which intervenes between 

 the blow and the time of vibration. It will be quite con- 

 sistent with our use of the word if we say that the second 

 blow has "interfered" with the first. But if friction be 

 introduced so that, before the second blow is delivered, the 

 motion due to the first has practically died out, the second 

 blow will produce its own effect whatever the interval of time 

 between the blows. There is now no interference, and the 

 interference in the previous case is seen to depend on the fact 

 that the effect of the first blow is a periodic motion continued 

 for a sufficient number of periods. From the case of a single 

 blow we may proceed to the discussion of impulses delivered 

 at regular intervals. 



Consider two exactly equal pendulums vibrating in a 

 period t. Let blows be delivered to the first at regular 

 intervals t, and to the second at intervals t-\-r. If n blows 

 have been delivered, the first being given simultaneously in 

 both cases, the blows given to the second pendulum will be 

 behind those given to the first by an interval nr. If that 



