I didn't look closely at this circuit, but the Pierce seems to have
the
crystal from plate to grid, or drain to gate. This seems as though it is
easy to slam the crystal pretty hard (depends on the drain resistor).
I'd
worry about crystal dissipation as this causes drift and ageing issues.
In
VCOs, the more tightly you couple the tuned circuit to the device, the
poorer the phase noise. We always had to get the things to just barely
oscillate under worst case conditions to get the lowest phase noise.
There
are other ways, but memory fails...

I like the good old standby common collector colpitts used in just about
all
Motorola channel elements & oscillators. Hard to make one which won't
oscillate.
--
Steve N,

Agreed, and the statement about starting with ~2 volts indicates way too
much feedback at 12 volts. Certainly none of the above leads one to think
it's a low noise circuit, just going to build one to measure. Suspect that
the "no free lunch" still applies and that to get really low noise operation
you still need current limiting, AGC 'd or otherwise and tricks like feeding
the output thru the crystal to clean it up. Also believe there's probably a
contribution by the draftsman again and the 22 meg gate resistor is really a
second 2.2 Megohm .

W4ZCB