I do not have access to sour oranges. Since this is an experiment, I am going to try them with blood orange, because they are only in season for a few more weeks at best.
Take Oranges and core them on the side and lay them in water, then boil them in faire water till they be tender, shift them in the boiling to take away their bitterness, then take sugar and boil it to the height of syrup as much as will cover them, and so put your oranges into it, and that will make them take sugar. If you have 24 Oranges, beat 8 of them, till they come to paste, with a pound of fine sugar, then fill every one of the other oranges with the same, and so boil them again in your syrup: then there will be marmalade of Oranges within your oranges, and it will cut like an hard egg.[1]
The blood oranges I found were a bit on the sour side, so they would do for this exercise. I cored 6 oranges 2 of which would be beaten into a paste with sugar. I used 2x refined sugar for this experiments.
I would definitely do this again. I learned the bigger the ring, the easier they stayed up and resisted cracking when the sugar boiled over the tops of them. I also learned that oranges float, so you cannot have too much liquid (aka cover the oranges). But sugar will rise and cover the tops of the sugar. Filling the pan 2/3 up the orange is sufficient for complete coverage and total saturation. Also all oranges ended up the same color in the end. Variations on the skin all evened out to the same dark orange. Filling is a little loose. I dont know if the breed of orange makes a difference or if there was too much liquid in the pulpy filling. Total cooking time 3 hours. This resulted in 4 oranges, 2 of which I am happy with their appearance. I can see where this would be better in bulk.
[1] Platt, Hugh. “The Arte Of Preserving Conserving, Candying.” Delights for Ladies: To Adorne Their Persons, Tables, Closets, and Distillatories with Beauties, Banquets, Perfumes, and Waters. Reade, Practise, and Censure. London: Humfrey Lownes, 1609.