The 2.5k resistor in parallel with the inductor models the estimated loses due to the finite unloaded Q of the inductor. The switch is required to provide a noise kick at the start of the simulation to get the sim going correctly. Neither of these components should be added to the actual bread boarded circuit hardware. Start the sim transient analysis, then momentarily touch the switch to give the loop a noise jolt. The inductor was built by winding 17 turns of #22 AWG wire on a T37-2 powdered iron toriodal core. The osc was designed, built and tested several years ago using the schematic as shown (but without the 2.5k loss resistor and extra start up sim switch). The transient analysis takes a little time to settle to the final values, but the sim results match closely to the actual hardware measurements.
