March 24, 1922] 



SCIENCE 



325 



the 1/ distribution tetween walls. The case for 

 R 90° is now a compound harmonic which 

 seems to dip into stationary wave trains at a 

 and b. The other curve is quite similar in 

 general character, only less pronounced. Hori- 

 zontal wave lengths exceeding 35 cm. may 

 again be detected at the double inflections. 



5. Survey in the vertical direction (z). This 

 was carried ovit by allowing the resonator to 

 rest horizontally on the table, in a direction 

 normal to the f" pipe, the latter being raised 

 successively in steps of 10 em., keeping it in 

 the same azimuth of 0°. 



The graphs (figure 5) for two positions of 

 the resonator are essentially identical, indi- 

 cating stationary waves produced by reflection 

 from the table. The z distance between crests 

 and troughs, however, now varies between 24 

 and 25 cm. and thus corresponds very closely 

 to the semiwave length, 24 cm., of the f" pipe. 

 In all cases the pipe must be raised some dis- 

 tance (40 cm.) before the periodic distributions 

 begin. 



The behaviour here in evidence is very much 

 like Melde's experiment, though it is now made 

 with a string of air (as it were) between the 

 actuating organ pipe as one attachment and 

 the table as the other. The only adjustment 

 possible is thus the length of the string. Since 

 the resonator lies on the table, certainly to be 

 regarded as a nodal surface, we would be 

 inclined to look for the maximum of wave 

 production, when the direct and return wave 

 train coincide in phase at the mouth of the 

 pipe. This will take place at intervals of 

 X/2, or A,? := 24 cm. apart, conformably with 

 the graphs. It would seem, however, that the 

 maxima (in view of the loss of 1/2 at the 

 table) should lie at 5r = 5X/4, 7X/4, etc., where- 

 as in the graphs they lie at 2)v/2, 3X/2, etc. 

 The latter demand a node at the mouth of the 

 f" pipe. 



As the table is certainly a nodal surface, we 

 here encounter the result of special sensitive- 

 ness to nodes on the part of the mouth of the 

 pin hole resonator. The case is tested in figure 

 6, where the pipe, P, in azimuth 0° is «; = 90 

 cm. above the table and the resonator vertically 

 below the pipe is raised from the table at 

 2 = 0. The evidence given bv the curve is 



very satisfactory, the wave lengths of the 

 graph being 24 to 25 cm., or semi wave lengths 

 of the f" pipe. There is complete absence of 

 deflection at 10 to 12 cm. above the table; i. e., 

 at X/4 for the pipe, so that the ventral seg- 

 ment is inactive. As the pipe at ^; = 90 cm. is 

 approached by the resonator, the deflections 

 naturally increase, but they do so very slowly. 

 Obviously the present disposition with a raised 

 pipe and with the resonator between pipe and 

 table is conclusive; but because of the im- 

 portant evidence obtained I rejjeated it for an 

 f" pipe at ^; = 70 cm. (nearly 3X/2). The re- 

 sults in figure 7 are of the same kind as to 

 wave length, inactivity for a resonator 12 em. 

 (X/4) above the table, the marked effectiveness 

 (maximum) of the distant node at the table 

 (z — 0) and a maximum near the pipe (^; = 68 

 em.). Troughs and crests lie at positions 

 which are multiples of « = 12 em. 



The above results for normal reflection may 

 be summarized as follows: Both the organ 

 pipe and the pin hole resonator are stimulated 

 in proportion as their mouths lie in a nodal 

 region or surface; they remain relatively un- 

 influenced by a ventral segment. Consequently 

 an even number of half wave lengths lie be- 

 tween pipe and resonator when the response 

 is a maximum. Although the mouths of the 

 respective pipes are necessarily ventral seg- 

 ments, the anomalous features of these results 

 disappear when it is remembered that the 

 nodes are alternately dense and rare. 



6. Beflectdon from plates. Using a plane 

 about ls.5 square meters in area, displaced 

 along X and normal to it, in steps of 10 cm. 

 from the origin successively, the effect of reflec- 

 tion (as I shall show elsewhere) came out 

 beautifulh'. It was possilile, by compounding 

 the direct and reflected rays in each case, to 

 interpret the harmonics and compute the wave 

 length of the pipe accurately. At the distance 

 of the wall (174 cm.), however, the reflection 

 effect had dwindled to 10 to 15 scale parts. 

 Diminution of the reflection effect also occurred 

 when the plane was placed oblique to x, but 

 not as abruptly as the law of reflection would 

 predict. Furthermore the distribution of 

 height in the successive maxima in the different 

 reflection curves was quite as remote from 



