To make it simple, consider two resistors in series in series across a voltage source.
The voltage across each resistor will be in the inverse ratio of the resistances.
Now consider two transformers in series across a voltage source. One transformer is fully loaded and the other transformer has no load.
The impedance of the loaded transformer will be much less than the impedance of the unloaded transformer.
The voltages across the transformers will tend to be in the inverse ratio of their impedances up to a point.
When the voltage across the unloaded transformer approaches the saturation point, the impedance of the unloaded transformer will drop and the voltage across the unloaded transformer will stabilize near the saturation voltage. The balance of the applied voltage will be seen across the loaded transformer.
This is a simplification.
In the real world, the effective X:R ratio of the loaded transformer (X:R ratio of the transformer plus the reflected X:R ratio of the load) will be different than the X:R ratio of the unloaded transformer. The division of the reactive current will probably be different than the division of the real current.
Of course when you apply these basics to a three phase transformer bank, things get more complicated.
And last but not least, if the transformer has a three legged core, the phantom delta effect tends to stabilize the voltages.
Further to GG's comment on distribution transformers, many but not all transformers intended for line to neutral distribution service are constructed with only one high voltage bushing.
Bill
--------------------
"Why not the best?"
Jimmy Carter