The balloon did not know it had been shot

At 3:25 a.m. on August 24, 1998, researchers launched a helium balloon from Vanscoy, Saskatchewan. MANTRA—Middle Atmosphere Nitrogen Trend Assessment—carried instruments for measuring gases involved in stratospheric ozone chemistry.

The measurements mostly happened. The landing did not.

A valve intended to release gas was obstructed, and both the termination system and its backup failed. The balloon drifted east. Off Newfoundland, two CF-18s fired more than 1,000 cannon rounds at it. The researchers had asked the pilots to aim for the suspension point, hoping the instrument package would detach and descend by parachute.

It did not. Neither did the balloon.

This became a dependable story about military embarrassment, but the embarrassment depends on treating every balloon like the rubber kind at a birthday party. MANTRA was about 150 metres high at altitude; laid flat, its envelope would cover several football fields. It was not stretched tight around highly pressurized gas. Small holes were small holes, not an invitation to pop.

The balloon continued over the Atlantic, attracting attention from British and American aircraft, entering Russian airspace and eventually coming down in Finland on September 2. The equipment was returned and reused, with bullet holes in the package and parachute.

A later MANTRA mission document gives the less cinematic accounting. The 1998 flight collected a useful scientific dataset, though some instruments malfunctioned and not every objective could be completed. Subsequent flights built on it.

The recovery system failed spectacularly. The experiment failed selectively. The balloon was excellent at remaining a balloon.