They didn't use a collector because they were using the slag purely for fire assay analysis.
The first smelt breaks down the alloy complex, releasing a portion of the metals (esp. Pd) and rendering the slag suitable for FA Analysis.
The second smelt releasing a large portion of the remaining metals locked in the alloy complex (esp. gold). It looks like the Pd is locking up the gold.
This is looking easier and easier and pretty bloody obvious.
- Crush rock --> Primary smelt with collector --> Electrolysis recovers most of the PGMs, some gold;
- Resmelt slag --> Electrolysis recovers most of the gold, remaining PGMs
should recover massive amounts of metal. Cheap and simple.