730 Prof. E. H. Barton and Mr. W. B. Kilby on 



dust figures occurred vertically between two balls or a ball 

 and very blunt point. The spark-gap was usually o£ the 

 order 2 cm., but was varied slightly according to the degree 

 of confinement of: the disturbance by cover-plate or obstacles. 

 To ensure a sufficient quantity of electricity passing at each 

 spark, a ley den-jar was always used across the terminals in 

 shunt with the spark-gap. This jar was small, medium, or 

 large according to the requirements of the case. 



The number of sparks needed to produce a figure varied 

 with circumstances. In any case it was found that the 

 figure reached its best state at a certain stage. 



The sparking potential was obtained by an influence 

 machine or an induction-coil, the former being more suited 

 for certain patterns and the latter for others. 



A small spark-gap was arranged on the prong of a large 

 tuning-fork of frequency 128 per second. Sparks were then 

 passed while the fork was vibrating, the figures being 

 examined in a horizontal tube with one end near the spark. 

 But the fork's vibrations appeared to have no decided effect 

 on the stria?. 



An explosion of soap bubbles (blown with mixed hydrogen 

 and oxygen) w r as also tried near the end of a dust-tube, but 

 failed to produce any stria?. 



Shadowgraphs and Enlargements — Where the dust figures 

 were formed on a glass plate and a print the same size was 

 desired, the following method of photography without a 

 camera was found suitable. The photographic plate was 

 placed film-side down close over the dust figure, separation 

 being secured by several microscope-slips. An electric light 

 about six feet below then gave the desired photograph with 

 a few seconds' exposure. 



Where the dust figures were on metal, or an enlargement 

 was required, the photograph was taken with a camera, the 

 dust figure being placed in a truly horizontal position, and 

 the camera fixed above with its optical axis vertical. Some- 

 times, however, a camera enlargement was made from a 

 shadowgraph, as the latter has the advantage of showing the 

 stria? in black lines, whereas the camera photograph shows 

 light and shade on the opposite sides of the ridge. 



Results. 



Reflexion. — The natural case to test first of all was oblique 

 reflexion from a plane surface. But this failed to exhibit 

 striae corresponding to any clear distinction between the 

 incident and reflected waves. So, to separate the regions in 

 w T hich incident and reflected waves would occur, an obstacle 



