[turretslug]

Cg1-a will be low (the valve data may claim hundredths of a pF) but not always negligible in a high-gain tuned amplifier, and compromised by layout. Cg3-a will be somewhat higher- hence the mere single turn on the IFT, as mentioned it would be interesting to know if the phasing is degenerative or regenerative in action. Even the actual value of g2's decoupling capacitor can sometimes enter into the situation of IF strip stabilisation- some circuits used a specified value as part of a neutralising bridge, though seemingly not here.