2 TlIOKP, Grating Filuis and Colo7it- Photography. 



and that the brightness increased as the prism was slightly 

 tilted, with a consequent lengthening of the spectrum 

 also. This efYcct was not produced from a replica of 

 another grating bright in the first order. The form of the 

 groove had evidently, therefore, given rise to the increase 

 in brightness, and since the original paper was read I 

 have succeeded in making a grating of small size, the 

 film replica of which causes nearly all the transmitted 

 light to be concentrated in the first order on one side. 

 Further experiments on the production of a more perfect 

 grating of this kind are in progress. 



For solar prominence observation a prism-grating 

 forms a very efficient battery of itself, and by tilting it to 

 an angle somewhat greater than is necessary to give 

 minimum deviation, its dispersion is quite equal to six 

 flint glass prisms of 60^ in the red portion of the spectrum, 

 whilst still retaining all its sharpness of definition. 



As might be expected there is a certain shrinkage of 

 the film upon drying, and this fact is made u.se of to 

 determine both the number of lines to the inch and the 

 quality of the replica. On placing the mounted film face 

 downward upon the original grating with the lines parallel 

 to each other, interference bands arc seen. When these 

 appear parrallel and equidistant it is confidently assumed 

 that the replica is suitable for work of a high class and 

 the number of bands per unit distance added to that of 

 the original grating, gives the actual number of lines on 

 the film. With a good grating prism as described several 

 lines are very distinctly seen between the two D lines in 

 the solar spectrum, with a low sun. The B group is par- 

 ticularly well seen and the individual lines distinguishable. 

 It has been found possible to silver the grooved side 

 of the mounted films by the ordinary glass silvering pro- 

 cess, but occasionally much difficulty in doing so is 



