SOME SPECIAL PKOCESSES 277 



ance from a pictorial point of view to represent them in colour, 

 yet ordinary photography shows them as well, provided that 

 the detail and differentiation of structure is brought out. 

 Staining methods are themselves arbitrary, and are only 

 employed in order to differentiate structure ; therefore, when 

 the structures are photographed, their differences are usually 

 as well shown in monochrome as by any colour method. 

 Coloured lantern-slides of necessity involve a great deal more 

 work in their production ; the results to be obtained are not, from 

 a scientific point of view, usually so exact, and they are therefore 

 not to be recommended except in very special cases. Perhaps 

 in metallography their use is appreciated more than in any 

 other branch of work, as the surface iridescence, or difference 

 in colour to be observed in such specimens, is inherent in the 

 specimen and does not depend on any artificial coloration. In 

 that branch of work, therefore, some of the very best results 

 have been obtained. 



The chief methods available are four. The Sanger Shepherd 

 three-colour process (in which three separate negatives have to 

 be taken), the Lumiere autochrome process, and the Thames 

 and Dufay colour-plate processes. In the latter three the colour 

 effect is obtained directly on a single plate. It must be ad- 

 mitted that although the first-mentioned process is by far the 

 most laborious one, the results to be obtained by it are from 

 a photo-micrographic point of view superior to any other, as 

 the image is absolutely grainless and is very transparent. 



The whole of the necessary materials for carrying out the 

 first process are to be obtained from Messrs. Sanger, Shepherd 

 & Co., and full detailed instructions for the work are supplied 

 with the outfit. The method consists essentially in obtaining 

 on a panchromatic plate three separate negatives : one each 

 through a red, green, and violet filter. The exposures to be 

 given will be dependent on the source of light, but need not be 

 of any great length. The time of exposure will not be the same 

 for each screen, but will vary according to the colour-sensitive- 

 ness of the plate and the spectroscopic value of the source of light. 

 As an example, with a given illuminant the exposure might be 

 20 seconds through the red screen, 100 seconds through the 

 green screen, and 25 seconds through the violet screen. It 

 is important to ensure that the relative exposure is correct, 



