264 



Mr. A. M. Worthington on the 



[June 1,5, 



tage of diffusing light through its mass. The mercury, on the contrary, 

 allows no light to penetrate its interior, and for this reason the form of 

 the drop is less easy to ascertain ; for the contrast between the brightly 

 illuminated raised or convex parts and the parts that are hollowed or in 

 shadow is so great that it is often uncertain whether the dark portion of 

 the figure is black plate or unilluminated mercury. This was at first a 

 serious cause of error, as in some earlier arrangements which I tried, the 

 light of the spark was reflected laterally from some distance by means of 

 a concave mirror, when the deception was very great, as will be seen by 

 reference to fig. Ila, Set 6. By means of the cardboard box, how- 

 ever, from the surface of which light was diffused from all directions 

 on the plate, and by having the spark close to the plate, I became quite 

 certain of the figures. Owing to the appreciable amount of time required 

 for the demagnetization of the electromagnet, I was not able to obtain 

 figures of the first portion of the spreading out of the drop. 



Owing also to accidental causes difficult of control, such as the varia- 

 tions of the contact at E, the variations of the pressure with which the 

 iron bar of the relay was initially pressed down by hand on the ends of 

 the electromagnet, and to the oxidation and irregularities of the surface 

 of the mercury at E-, the time of appearance of the spark after the first 

 contact of the drop varied slightly, so that the stages seen with a given 

 depth of immersion of the wire P and a given tension of the spring were 

 not always the same, but varied between narrow limits. But a little 

 judgment enables the observer to tell whether the stage seen is before or 

 after the mean stage most frequently seen with that arrangement of 

 level of mercury and velocity of withdrawal of the platinum wire. 



Sets 1, 2, and 4 are figures of milk ; 



Sets 5, 6, 7, 8, and 9 of mercury. 



Explanatory notes are attached where necessary. 



The figures suggest a few general observations. 



The existence of the radial arms is a deviation from perfect symmetry 

 of figure round the vertical axis ; and some slight initial disturbance of 

 symmetry must be required to determine the formation of arms. 



Such a disturbing cause is always at hand in the irregularities of the 

 surface of the glass or of the layer of smoke, which allow the drop to 

 spread with less frictional resistance in one direction than in another, 

 and also in the oscillations of the drop about its mean figure while it 

 falls through the air. The occurrence of such a figure as III, Set 6, 

 confirms this view, as it is an approximation to perfect symmetry when 

 the disturbing cause has been probably very small. 



The fact, too, that a very slight dirtiness of the plate was sufficient to 

 cause great irregularities of figure is confirmative of this, as is also the 

 fact that the tendency to form radial arms increases with the height 

 of fall. 



In seeking the explanation of the fact that the arms which are first 



