Vol. 8, 1922 
PHYSICS: C. BARUS 
99 
There is no symmetry of graph at the middle of the table (y = 33) and 
the two edges T, T' show no correspondence. The graphs may actually 
be traced into free air beyond the table, as indicated in the prolongation 
of figure 12. Finally if the intensity of the supposedly chief maxima be 
laid off in terms of the pipe position an harmonic figure 12 of relatively 
constant amplitude appears in reasonable contrast with the curve of di- 
minishing amplitude in figure 2' , for the raised pipe. In both cases, the 
steps of 20 cm. between pipe positions are too large for sharp discrimina- 
tions; but it is noticeable that the distance between crests in figures 2' , 12 
and the earlier reflection figures is about = 40 cm. The prevailing 
A;V in the single graphs 2, 6, etc., is about 30 cm. In figure 12 the points 
beyond 7^ = 80 cm. were obtained in similar investigations, here omitted. 
4. Summary. — The above work as a whole has shown that when the 
reflecting plate is relatively near, to the pipe and resonator, the position 
of maxima and minima may be predicted as a case of ordinary interfer- 
ence and the wave-length computed satisfactorily. The distribution of 
intensity among the crests and troughs remains harmonic and has not 
been foreseen. 
If the reflecting surfaces are relatively remote the positions of crests 
and troughs cannot, as a rule, be found by the same method satisfactorily. 
In certain instances, there seemed to be an approximate fit; but as a whole 
the attempt was unsuccessful and the distribution of nodal intensities 
equally puzzling. Acoustic topography is not symmetrical to the pipe. 
It appeared, however, that, when the organ pipe, definitely placed, is 
sounded, the occurrence of nodal surfaces of a fixed position in free air is 
demonstrable everywhere. Distributions between walls are consistent 
and differ from distributions between wall and door. It is not improbable, 
moreover, that such surfaces are regularly grouped and may for reasonable 
distances be approximately parallel. If then they are intersected obliquely, 
for instance by the plane of the table, and if a is the mean angle be- 
tween the latter and the nodal surface, the distance between crests on 
the table should be A3; = X/(2 sin a). Hence this intercept, Ay, may 
have any value depending upon the general shape of the room. If the 
prevailing value A3; = 30 cm. be taken, sin a = 24/30 or a = 53^ 
roughly. This angle between surfaces is also the angle of incidence of the 
rays, and has been encountered more or less closely in other instances 
(fig. 5). The distribution of intensities would then also depend on the 
fixed pipe position in relation to the given room, regarded integrally as an 
enveloping reflector. Such an explanation is plausible and flexible, no 
doubt; but it has the great disadvantage that none of the results can be 
reproduced by computation. The conviction that some other explana- 
tion may be found is not removed and the answer probably lurks in such 
curves as figures 2' and 12. This is particularly so, because the pin- 
