All batteries have a nominal voltage slightly higher than the voltage required of the equipment but drops to the specified voltage once a load is placed on it. The mercury battery had a nominal voltage of 1.4v just like the zinc-air battery. As has already been stated the only difference is in the lifespan of the zinc-air as opposed to the mercury. That is the reason I prefer using a diode in line and using silver oxide batteries. I get the same voltage curve (as the mercury cell) and the long life.. This also translates to a more accurate reading when the battery is near it final stage. The zinc-air tends to be a bit unstable at that point whereas the silver oxide is rock solid stable right up to the point of battery death.
Something to remember is that battery death does not mean the battery is depleted of any voltage. Battery death is when the needed voltage drops below what is needed to drive the meter...
I do agree tests are in fact done at calibration extremes.. Old batteries will show their weakness at the extreme load end.
Most camera meters have a current load of approx. 50µA-200µA so there isn't a lot of difference. Hand held meters usually higher.. In case you are wondering µA is micro-amps..
Measured my kit to hand
OM1n dull light /1.8 lens 15.0 micro amps
V625PX (Hg off load) 1.5v - it is an old cell
The meter current would be below nudging my meter with a lens cap on but a lot higher in bright sun the CdS cells vary their resistance radically note I did the current test with a fresh alkaline to avoid any internal resistance problems.
I don't actually use camera meters so don't have zinc batteries to hand.
I did do a calibration test for the meter with zinc cells and from dull to bright the OM1n was better than 1/2 a stop with an OM4, the CdS resistors are old and do drift.
Old batteries can develop internal resistance which alters the meters effective calibration especially in bright light.
Selecting a diode is dependent on current taken and you need to solder it in or make an adapter.