WAVEGUIDE AS A COMMUNICATION MEDIUM 1245 



conversion comjx^u'nts were ])r('s(>nt ; similar data were ohscrvcd on llic 

 sliorter h^ijiths of lino. 



Azinmthal (listril)uti()ns of pi'oUc rcs))()iisc showed coiu-lusixrly lliat 

 v(My little conversion was ])i-cseiit to the \'A '\'\'],,„ and T.M,,,,, modes 

 having an index "/?" of four or moic 



A much more precise though more eiahoiale met hod olC\ alual ing the 

 eon\'ersion coefficients invoh'es use of coupled waxc transducers. ' Such 

 devices resj-jond to only one mod(^ and ha\-e known s(Misiti\-ities, and 

 therefore permit truly (iuantitati\-e m(>asur(Miients. (.\t t!i(> "pi'es(Mit time 

 this ap]n-oach requires a separate transducei' foi' each mode 1o he e\ahi- 

 ated, a somewhat cumbersome pi'ocedure, hut in principal the mode 

 transducers can l)e made "tunable" for a series of modes.) A. ('. Heck 

 and ]\[. Aronoff applied this more accurate method of conx-ersion coeffi- 

 cient determination to several modes of the oOO-foot hue, and 'liable 1 

 shows the vakies averaged over the frecpiency band. 



Table I — Average Ratio of Conversion to TEm Heat Loss 

 WITH TEoi Excitation 



Total 0.47 



The estimated absokite accuracy is between 10 per cent and 20 per 

 cent for these ratios. The variation of the conversion coefficients as a 

 function of frequency is shown in Fig. 24 for two of the important modes 

 and for the total of the modes given in Table I. Th(> total of the modes 

 measured is consistent with the probe indications, though the acou'acy 

 of the latter is low enough that this should not be interpreted as proxing 

 there are no other important conversion contributors. 



The above direct evaUiation of mode conversion in the .")()0-fooi line 

 yielded magnitudes that are sufficient to explain the conx'ersion-recon- 

 version process as already outlined. We wished to extend our experience 

 with this particular fine to higher-fre(|ueiicy lines which might be built 

 and to absolute tolerances that might l)e ])laced on the construction of a 

 new hue. Toward this end, theoretical relations wei'c derived by S. P. 

 Morgan, Jr., for the mode conversions to be exp(>cted due to waveguide 

 ellipticity, and due to the tilt and offset which may be exi)ected to occur 

 at the junction of two sections. Experimental work was done by M. 



