148 Proceedings of the Royal Society of Edinburgh. [Sess. 
determined from the reading of the micrometer screw. All adjustments 
were made for the middle of the ultraviolet. 
A very steady arc was obtained with a positive electrode of 11 mm. 
diameter, and negative electrode of 8 mm. diameter on a voltage of 145 
volts with 15 ohms ballast resistance, an inductance consisting of one of 
the coils of a large electro-magnet being placed in series with the arc 
Kayser * recommends for the iron arc poles of 10 to 15 mm. diameter with 
a current of 10 to 15 amperes, while Pfund f recommends 3*5 amperes on 
a 2 20- volt circuit with a positive electrode 12 mm. and negative electrode 
6 mm. in diameter. It is interesting to note that Nutting J obtained a 
good ultraviolet spectrum for photographic purposes by using a spark 
source between terminals of an alloy of two parts of aluminium to one part 
of zinc; a large capacity was placed in parallel with the spark to make 
the spectrum continuous. 
While this arrangement gave a fairly strong beam of ultraviolet light, 
it was found necessary to purify it by resolving the incident radiation by 
means of an objective quartz prism ; only the ultraviolet was then allowed 
to enter the spectroscope. The intensity of illumination, however, when 
the beam was projected on an object of any considerable size was small, 
and very long exposures were required. 
The apparatus was therefore modified and the scope of the investigation 
limited to examining the appearance of different pigments and salts in 
ultraviolet light. This paper contains the results obtained for a number 
of pigments of various colours. Instead of illuminating these with mono- 
chromatic light, the whole iron arc spectrum as reflected diffusely from 
them was photographed and compared with the spectrum reflected from 
a selected piece of white cardboard. The usual method for producing a 
spectrum was employed ; a plan of the essentials is shown in the figure. 
The source S was the iron arc placed outside the window as before. 
A Cornu prism of 4’6 cms. side and 4‘0 cms. height was mounted on a 
levelling table B along with two quartz lenses of 2'7 cms. aperture and of 
22*9 cms. and 49*8 cms. focal length respectively, the prism being set for 
minimum deviation of wave-length 308 micromillimetres. The slit A in 
the window shutter was 2*0 cms. high and 0*5 mm. wide. Its spectrum 
lies practically in a single plane CD. 
Winsor and Newton’s oil colours were used. They were laid thickly 
on narrow strips of wood by means of a palette knife ; the oil was absorbed 
by the wood, leaving a hard even layer of pigment. The painted strip was 
* Handhuch der Spectroscopie, vol. i, p. 916. t Astroph. Jr., xxvii, p. 296 (1908). 
f Pliys. Rev., xiii, p. 193 (1901), 
