@kbsparky: The fixtures are wired using an electricians loop, so a single bad fixture shouldn't be causing the problem.....
Since you later disclosed that the entire thing was in a metal conduit system, this makes perfect sense. The sparks were coming from the fittings of the conduit, not out of the fixtures themselves.
This condition makes one ponder about the effectiveness of using metal conduit as your only source for an equipment grounding conductor. While recognized by the Code as compliant, I prefer to use a supplemental EGC if the metal conduit system itself is isolated.
I keep forgetting it so I have to borrow one from the HO. They're less bulky [the dryer, not the home owner] than a toaster oven.
I do have a 100w 100Ω rheostat some guy gave me for free at a Hamfest. Unlike an incand. bulb, its resistance does not depend on the current through it. It might stand 5A for a few seconds but I don't want to push my luck.
I'd like to get an air-cooled water heater element for a heavy 240v load but it's just too inconvenient.
Pulling a lot of current from wiring is a stress test and weaknesses [bad neutral, bad wirenuts] show up as too much voltage drop.
Or just buy Ideal's #65-165 tester. When it pulls heavy current for testing it's only for a short pulse so they don't need air cooling or a big heat sink.
If not, you go low-tech and carry a calculator around, like me. A meter that can display 123.45v, 4-1/2 digits, 20,000 counts, also comes in handy.