1489 
were selected because of their resemblance to the theoretical 
ones rather than on the basis of known apriori values of Cans 
An objective mathematical method of analyzing the experimental 
curve to obtain the experimental E has not been successfully 
developed, and at this time it is only possible to guess at the 
angles by qualitative comparison of the shapes of the experimental 
and theoretical curves. 
Figure 3 illustrates the very close agreement between 
theory and experiment as to the progressively changing shape 
of the first bottom reflection with increasing angle of incidence, 
including the complete phase inversion predicted at glancing 
angles. 
Figure 2 compares theoretical and experimental curves 
for an entire record of several milliseconds duration. Here 
account is taken in the theoretical curves of the superposition 
of the various reflections which follow closely upon each other 
and of the cumulative phase shifts taking place in higher order 
reflections undergoing more than one incidence at the bottom. 
The theoretical curves are calculated using the assumed 
values of 2.7 for © /é » 1.15 for C,/C,, and the values of 
QO, for the various reflections to establish the values of eG OK 
The values of 6, for each reflection are obtained from the 
geometry of the particular experiment. (The shupes of the curves 
are not critically sensitive to the value adopted for Co/Cy so 
that a comparison of experimental results with theoretical shapes 
= Lie 
