294 



BELL SYSTEM TECHNICAL JOURNAL 



dielectric as above but surrounded by metal. It will be noted that 

 the limits of this curve are essentially the same as for the unshielded 

 guide but that the two curves follow rather different courses. If, 

 in this case, the dielectric constant were protjressively reduced the 

 curve shown would gradually shrink into the miniature replica shown 

 dotted in the lower left corner. This is of course the practical case 

 of a hollow conductor to be discussed shortly. The above discussion 

 leads naturally to the view that a wave guide is a propagating medium 

 bounded by a dielectric discontinuity. In one case the discontinuity 

 is the interface between the dielectric of the guide and the surrounding 

 medium. In the other it is the interface between the dielectric and 

 a surrounding conductor. 



It will be noted from Fig. 5 that at the highest frequencies the phase 

 velocity in shielded guides, like that for unshielded guides, is the 

 same as the velocity along ordinary conductors in that medium but 

 at frequencies near the cut-off this velocity approaches infinity. The 

 solid curve is calculated on the assumption that the medium had a 

 dielectric constant of 81. The indicated points are the results of 

 experiments made with water as a dielectric. For this experiment 

 the water was supported in three-foot cylinders of copper, six inches and 

 ten inches in diameter respectively. The same range of frequencies 

 was used as above. A somewhat closer agreement between calcula- 

 tion and experiment would have resulted if a value of dielectric con- 

 stant of 78.9 had been assumed in the computations. 



TABLE II 

 Velocity Ratios for Hi Waves in Hollow Conductors 



1 Probable error 0.4 per cent. 



^ Probable error 0.4 per cent. 



' Probable errors (arising from error in X/rf) range from 0.2 per cent for X/rf = 0.98 

 to 1.8 per cent for \ld = 1.55. Note that agreement in most cases is within probable 

 error. However the fact that in all but (jiie observation differences have same sign 

 suggests some systematic relation. 



I 



I 



