What They Saw
They grew Azotobacter vinelandii UW (aka CA) or a tungsten-tolerant mutant with tungsten and Mo or V, then extracted the nitrogenase.
This mutant was derived by growing UW with W and ammonium over several passages to try to remove all the Mo the cells might be storing, and then plating the cells on plates without ammonium, so they would have to fix nitrogen in the presence of W and absence of Mo to survive. Stuff that grew was tungsten-tolerant: strain LM2.
Then to get nitrogenase, they grew LM2 with tungsten and UW with tungsten and ammonium (not sure why it would produce nitrogenase in that condition, though I guess it might when it ran out of fixed nitrogen).
Both UW and LM2 could grow on medium with Mo, though LM2 was sometimes slower. With W, LM2 grew faster or slower depending on amount of W, while UW did not grow.
They used something called rocket immunoelectrophoresis to measure how much MoFe nitrogenase components each strain produced with W. Each produced about the same amount of dinitrogenase reductase in all conditions, but UW with 1mM W produced about 56% the amount of dinitrogenase as it produced with Mo, and LM2 with 10mM W produced about 7% of the amount that UW produced with Mo. I wonder how accurate this technique is. But they also note that the specific acetylene reduction activity of crude extracts was about 6-8% for each strain grown in W compared to UW in Mo. Does this include alternative nitrogenase activity? They say no. But overall, it seems that both strains produce less Mo nitrogenase with W, LM2 less than UW, but LM2's is relatively more active.
I think what happened next was that they couldn't isolate nitrogenase well from LM2, so they studied it from UW. They had two kinds: typical Mo nitrogenase, and Mo/W nitrogenase that had one FeW-cofactor and one FeMo-cofactor. This latter showed less activity in every way: hydrogen under argon, nitrogen, or acetylene atmospheres, nitrogen reduction, and acetylene reduction. This fits with at least one previous study (045). There is still activity though, even with nitrogen reduction, so it's not clear why the cells can't grow in W; maybe its interaction with other important proteins?
What This Means
It's possible that instead of each molecule of Mo/W protein containing one W and one Mo, half the protein could have all Mo and the other half all W; this would give the same results, but seems a less likely explanation. Of course, both possibilities are pretty weird and confusing.
According to some chemistry stuff they did, it seems like the enzyme can't reduce FeW-cofactors, which could reduce the possible electron flux by half, I think. There are a lot of mysteries here.