As regards the ultra-violet, which may be considered to begin at about X3,9oo, prisms of light flint glass will transmit to the neighbourhood of X3,3oo, while prisms of the so-called uviol glass will extend this to about X 2,900. Beyond this, if prisms are used, it is necessary to construct them from natural crystals, usually Iceland spar or quartz, and principally the latter. With an instrument built up from prisms and lenses of quartz, it is possible to photograph as far as about 2,IOOA with ordi nary plates and to about 1,84oA with specially prepared plates. It is in fact possible to see the whole of this part of the spectrum by allowing it to fall on a fluorescent screen such as is used in X-ray observations. The need for special plates for the most refrangible part of this region is due to the fact that gelatine ab sorbs the short waves before they can reach the sensitive salts which are embedded in it. An extremely useful method is that introduced by Duclaux and Jeantet, in which an ordinary plate is coated with an oil, such as paraffin, or nujol, which becomes fluorescent under the stimulus of ultra-violet light so that the spectrum is actually photographed in light of longer wave-length; when this method is applied the oil must be removed, as by wash ing the plate with benzene, before development. Plates prepared by Schumann's method, in which the gelatine is reduced to the smallest quantity that will allow adherence of the silver salts to the glass, however, are the most suitable for the extreme ultra violet. Such plates are now prepared commercially by the firm of Adam Hilger, Ltd., London.
As quartz has the property of double refraction, it is necessary to cut prisms of this material so that the optic axis of the crystal is at right angles to the refracting edge and parallel to the base.
Even then, however, there would be a slight doubling of the spectrum lines arising from the property of quartz by which it rotates the plane of polarization. This is overcome by Cornu's device (fig. 1o) in which one
half of the prism is made from right and the other from left-handed quartz. The two lenses in a quartz spectograph are also of opposite rotations, and in order that a long range of spectrum may be in good focus, the photo graphic plate must be inclined to the axis of the camera, and bent to an appropriate curve in the plate-holder.
For wave-lengths still shorter than 1,84oA, fluorite is the most useful substance at present known. With prisms and lenses of this material it is possible to photo graph as far as 1,2ooA, providing that Schumann plates are used and that the absorption of air is eliminated by exhausting the spectrograph and allowing no air to intervene between the source and the slit. Victor Schumann was the first to investigate by this method and the region from 1,85oA to 1,200A is conveniently called the Schumann region. Even one millimetre of air at atmos pheric pressure will absorb all light of wave-lengths less than about 1,7ooA. Among the sources which are available for investi gations in this region, the slit being covered with a thin plate of fluorite, are the "vacuum arc" (an electric arc in an exhausted vessel), the spark in an atmosphere of hydrogen (this gas being transparent for short waves), vacuum tubes, and the "vacuum spark" of Millikan and Bowen in which short sparks are pro duced by heavy discharges between terminals in a high vacuum. In each case, the vessel containing the source is sealed to the slit plate of the instrument.
The region of the spectrum lying beyond the limit set by fluorite can at present only be investigated by the use of concave diffraction gratings, in which no lenses are used, and the path of the rays from the source to the photographic plate is entirely in vacuo. With such instruments the spectrum has been photographed as far as 136A by Millikan and Bowen, and the region beyond this—the soft X-ray region— has been photographed by Jean Thibaud, using a grating near grazing incidence.