Wow, that's a really nice meter.
Internally, they "do the math" by means of magnetic fields. It's a "moving coil" meter like a DC current/voltage meter, but using an electromagnet to provide the field instead of the permanent magnet,
The load current is routed through a large gauge wire that is wound around the stator, providing the field. The voltage is fed to the coil moving the pointer. Pointer movement is proportional to coil voltage multiplied by field current, smoothed in the time domain due to the mechanical mass. These meters work "wideband" from DC to some hundred hertz, the upper limit is given by the inductive characteristics.
The binding posts are labelled "line" and "load", so I'd suggest to connect the source to "line" and the load to "load", routing the load current through the meter. It says "compensated for potential current loss", so I'd guess the voltage sense (the moving coil) is connected across the "line" bolts, while the current sense (the field) is connected from one "line" bolt to one "load" bold, while the others are connected directly. So the meter doesn't display its own consumption. It would show positive power flowing from "line" to "load".
Edit: As this meter is designed to be wired directly "in-line", it would be rather difficult to modify its voltage and current ranges. A shunt wouldn't work at all, rather a current transformer (limiting its usage to AC). Depending on the internal construction, it might also be difficult to modify the voltage range without opening and modifying the meter itself.
Edit: "TSP #140 - How Do Purely Passive Watt-Meters Work?" shows excactly this kind of meter