A lamp can satisfy every dimension in part 1 and still fail to light in a holder that also satisfies part 2, because the two tolerance bands can combine in the worst possible way. The contact-making test exists to catch exactly that combination, and it is the only routine check that measures function rather than geometry.
The gauge is a holder made to the least favourable limits for contact: the socket is as deep as the standard allows and the contact is as far back as the standard allows. The lamp is inserted, and the circuit must close. A cap whose contact sits even slightly high will fail, which is the point, because that same cap in a real holder at the other end of the tolerance band will also fail, just less predictably.
The circuit must close on first insertion, without rocking, and without forcing the lamp beyond its normal seating. A lamp that lights only after being pressed hard or rotated slightly has failed the test even though current flowed, because the fitting will not apply that force.
The contact-making gauge has a spring or a fixed contact that wears. Its critical dimension is the contact depth, and it should be checked at every calibration with a depth measurement, not merely a functional pass with a known good lamp. A reference lamp that has itself drifted is not a calibration standard.
