I had always heard the "pear shaped tone" description used in reference to early efforts at analysis of the spectral content of speech and singing.

I'll leave it to others to determine whether the terminology predates the development of instrumentation for such direct measurement, but will note that oscillographs and crude oscilloscopes date to about the time when the Edison recording and telephone instruments appeared - early 1900s.

On an oscilloscope or oscillograph, the "pear shaped tone" looks exactly (more or less) pear shaped - in a display of amplitude versus frequency. Regularly spaced harmonics vary smoothly in amplitude, with most of the content in the lower harmonics, tapering smoothly to lower amplitudes. The "ideal" tone contains "rich" harmonics though the upper second octave - the body of the "pear." Into about the third octave, there are "rich" harmonics, but with lesser amplitude - the neck of the "pear," with little content above about the third or possibly into the lower fourth octave.

A nasal or "thready" tone (or a free reed like a harmonica) looks more like a rather "scraggly" christmas three.

If the terminology existed before then, it is certainly nice to see that "modern" instruments confirmed its appropriatness.

John