Crystals (or resonators) in AT26 packages have vanishingly small capacitances, so I conjured a little fixture for my AADE L/C Meter IIB (*) that holds them securely under little fingers snipped from an EMI shield:
The finger on the right sits atop a snippet of rectangular brass tube so it need not bend so far.
The base is a snippet of double-sided PCB with copper tape soldered around the edges. I drilled the holes slightly oversize and soldered copper tape there, giving the top foil a direct connection to the terminals. The raggedy slot looks like it came from a hacksaw; no false advertising there.
The meter reports 6.5 pF of stray capacitance and nulls it to zero as usual. Without the fixture, it shows 2.5 pF.
With the crystal in that position, the meter measures Cpar, the parasitic capacitance from both terminals to the can, which should be (roughly) twice the capacitance from either terminal to the can.
Two more clips measure C0, the plate-to-plate capacitance:
The meter drive is about 200 mV at 700 kHz, far away from resonance. Assuming the resonator’s effective series resistance is 25 kΩ (tuning forks aren’t crystals!), it’s dissipating 1.5 µW (and less as the ESR goes up). That may be slightly hot for some resonators, but it’s surely survivable.
Some preliminary data on five 32.768 kHz crystals shows Cpar = 0.4 pF and C0 = 0.9 pF. I don’t trust those numbers very much, but they’re reproducible within 0.1-ish pF.
(*) Almost All Digital Electronics and its website vanished after the owner died; the meter continues to work fine. The cheap knockoffs flooding eBay and Amazon may get you close to the goal.