Getting into conjecturing at this point.
CO2 or a reducing environment would also help oxidation which at the surface or edges would give rise to the cooling scale or oxides and impurities.
In my mind, I imagine the cooling of metal in a melt can be similar to the forming of crystals from a saturated metal salt solution, the crystals can push out impurities towards the surface, or at least out of part of the crystal structures which form as salts from a solution or an alloy of metal, many factors can come into play composition of the solution or molten metals and gases, temperatures, cooling rate, and environment, which can greatly affect the crystal structure of the alloy as well as the surface and contaminants...
Silver loves to absorb oxygen in a melt, this oxygen and the base metals involved in the melt react to form oxides which on cooling the silver alloy will push out as the crystals cool.
These salts, impurities or metal oxides will also have different melting points than the alloy itself...
Impurities can act like seeds for crystals to begin to grow or places for oxides to form groups or larger areas of scale, Oxidation of base metals such as copper or iron in the alloy will try to get to the surface, climb the side of the mold edge to get to the surface picking up more oxides...
Not really related, I have very little experience with jewelry but when I have made rings with silver that Had PGM metals I noticed hard spots at the surface of casting (like a small impure lump) on the surface of the casting much harder than the surrounding alloy composition.
While just rambling I notice the red circles are on one end, would this be the cooler or more of the oxidizing range of your torch during the melt???