210 PEACTICAL PHOTO-MICEOGEAPHY 



vulcanite or even hard wood is quite efficient between which 

 are clamped by means of three screws two slips of glass separated 

 by an ordinary india-rubber ring (Fig. 64 b). By releasing the 

 screws the glass can be removed and replaced quickly and easily. 



The simplest of ready-made filters are pieces of ordinary 

 coloured glass. These may be obtained in a fairly large variety 

 of tints, but their absorption-spectrum is often widely different 

 from what might be expected, having regard only to their visual 

 colour. For ordinary purposes a rather deep yellow, green, 

 blue-green, and blue glass should be chosen. These should if 

 possible be what is known as pot-glass that is, glass in which 

 the colouring is carried right through and is not merely flashed 

 on to the surface. Such may be obtained from Messrs. Hetley 

 of Soho Square, London. While they may answer their pur- 

 pose when the very best obtainable results are not necessary, 

 they are open to considerable objection for work of a critical 

 nature. 



The illustrations (Fig. 67) are photographs of a few typical 

 absorption-spectra, taken with a spectroscope with aMiffraction- 

 grating mounted on a prism, so that the projected spectrum is 

 practically at right angles to the optical axis of the spectro- 

 scope. Such absorption-spectra may therefore be obtained by 

 combining an ordinary long-extension camera with a direct- 

 vision diffraction spectroscope. 



The arc-spectrum of iron has been chosen as the standard, 

 for the reason that it has a large number of lines well distributed 

 throughout the whole of the visible spectrum (Fig. 67 a). 



Fig. 67 b is the same spectrum taken with a piece of green 

 glass intervening. It will be seen that very great general 

 absorption of light has taken place, although the exposure given 

 was about three times that given to the bare iron arc-spectrum, 

 and in addition the light transmitted is not confined to the 

 green portion of the spectrum, but extends without any very 

 well-defined boundary in each direction. 



Fig. 67 c is taken with the glass known as signal-green. This 

 absorbs nearly the whole of the red end of the spectrum, but it 

 has no particular selective absorption effect on the remainder, 

 and results in considerable loss of light. It is therefore clear 

 that both from the point of view of general loss of light, 

 and from the difficulty of getting glass with a well-defined 



