Appears the force of the impact spewed the fuel out of the fillers.
I'm coming to this thread late.
There is a three-page spread in Dave Freidman's book 'Shelby GT40' which covers this test in detail.
The person who conducted the test, Bill Holbrook, shares his memories, and besides the above photo (in color) there are several photos of the car after the impact test.
He says that they used Stoddard Solvent in the tanks because it replicates fuel almost perfectly. They wanted to crash the car into the barrier at 60 mph but they were hampered by all the cars and cables; it actually hit at 51 mph. The wheels crushed into the front of the rocker panels, and the pressure causes the fuel to go up the filler neck and blow the fuel lids off, spraying the whole car with fuel.
After the test, they fitted thicker, stronger rubber fuel bladders and installed stronger roll cages.
The various replicas out there typically don't have any sort of protection at all; the fuel just sloshes around in bare metal tanks, like a 1950s F-1 car/time bomb. Rubber bladders are probably a good idea, but they have longevity issues, and cost a fortune to boot.
I guess the ultimate solution is to just not crash?
Try as I might, I can't positively identify which chassis was used for this test. Although Holbrook says the car was a write-off, the various registries make no mention of it that I can see. By the tail end of the '66 season, the GT40s that were being raced by Shelby and Holman-Moody had been retrofitted with rather hefty roll cages (barely visible in the photo above, but very clear in the still photos). Still, the roof is visibly tweaked and the rake of the windshield is much steeper after the crash, which would lead me to speculate that this tub was history.
The car in the photo has one very unique distinguishing feature--the cutouts behind the front wheel. So far as I can tell, that only appeared on P/1016. It is present throughout the 1966 season, but then when the car re-appears in '67 they're gone. There's no mention of P/1016 being used for crash-testing in any of the books I have--but then again, there's no mention of ANY specific Mk II being used for crash testing and yet obviously one was.
So, barring any contrary evidence, I'm speculating that it's P/1016 that got mauled by the barrier. I'm no expert though, I'm just a dork who looks at a lot of books and tries to piece evidence together. :laugh: