AND CHLORINE UNDER THE INFLUENCE OF LIGHT. 
83 
§ 1. The Relation of the Initial Expansion to the Time of Insolation. 
The first point considered was the relation of the expansion to the time of 
illumination, the intensity of the illumination remaining constant. 
The influence of very small cpiantities of imjDurity was found to be merely to 
diminish the quantity of effect observed, not to alter the quality of the effects, so that 
impure gas and more intense light give the same effects as pure gas and less intense 
light. This is useful practically, as there is no need to take so much time for 
preparing the purest mixture obtainable in order to observe the qualitative effects in 
the initial expansion. 
In fig. 5 are plotted the results of four series of experiments. In each series the 
light was of constant intensity. For any single observation the illumination was for 
the time represented as abscissa and the expansion observed was that represented as 
ordinate. 
Fig. 5. Abscissae, time in seconds. Ordinates, expansion. 
Between each observation the bulb was left for a quarter of an hour in the dark, so 
that one experiment should not affect the next. 
These curves show that we have an increasing rate of expansion with the increase 
of the time of illumination. The relation between the expansion and the time of 
illumination was also determined, in the course of one expansion, by marking the 
position of the index at intervals of one second after the beginning of illumination. 
With rapid motion of the index this method is difficult, but it has the advantage 
over the other method of following the whole course of the expansion in one 
experiment. 
In the following table are given the results of experiments of this kind. The 
numbers in the last three columns represent the positions of the index at the time 
stated in the first column. 
M 2 
