Again, due to the breadboard affects on crystals, my- circuit was slightly different than Alan's by the lucky trial and error of my incorrect cap value components removal - including a floating caramic cap at the base of TR1 - at full PSU 30V before this thing kick started at a stable 15.9982MHz.
I removed the top 10nF cap and one 470pF (mine was a different value anyway) from ground., leaving one floating...(aerial!?). So touchy! Just connecting the probe ground can stop it..
17V gives a 1Vpp sine wave - if you don't move anything!:
Arduino, Amps, Kits etc.