I wish to take up your phrase of "can did not burst in the gun tube but split in flight or less likely upon hitting the ground". You wrote something similar in a prior post.
This was not the intended action as the artillery desired case bursting at the muzzle so that a cannister cone which spread from the muzzle of the piece. Griffith, on page 72 of his 1839 edition of The Artillerist's Manual states:
"CASE OR CANISTER SHOT The common case or canister shot consists of a number of balls packed in tin or iron canisters of a cylindrical form: the balls being of different weights according to the size of the gun. For field service the balls are counted into the case and laid in tiers, but for other purposes they are loosely thrown in till the case is filled. The top being soldered on, the case is secured to a wooden bottom by slips of tin soldered down. These shot should never be used at a greater distance than 300 yards, and even at that range (the case bursting immediately on leaving the Gun) they scatter so much as to be nearly unserviceable ; but at from 100 to 200 yards they are very destructive."
Now before you ask how one "knows" where the case "burst" or how the cannister balls spread without appropriate photography, I must these folks surely - after firing round after round - have seen the pattern to a sufficient degree.
I say this as when I shot skeet I could "see" the cloud of shot pellets (fired by me or others) at least for some distance. As a shooter, my eyes were focused elsewhere than on the pellets near the barrel, so I admit I did not see mine at that distance but later in their flight. Still, if one can see the cloud pattern of the pellets, the artillerist could see enough of a pattern of canister. It may have an aspect of a blur (as the pellets, but the canister balls were larger - the size of golf balls, which one can see individually, let alone a cloud of them), but it was not clear air.
The hits against walls helped the artillerists to record the pattern more scientifically. - R