2. Standing on marble blocks in rock layer is the EASIEST way to set the marble-spawn flag true. In the best condition(when nothing goes wrong) with the marble-spawn flag true you can ensure 4/5 of the spawns are marble-styled.
I was finding a rate more like 1 in 5 marble mobs whether standing directly on 2 marble blocks, or with then 20 tiles away. Caveats to this finding holding true universally are: (A) RNG and a small sample size for me. (B) I was at the edge of the world, and do not know that this (literal) edge case might cause differences.
marble blocks under where the mobs would spawn
Seemed to more reliably boost marble mob rate. But again still lower than 4 in5, I think. Wonder if game progression (bosses downed) changes things possibly?
6. If the marble style is settled, then Medusa will spawn in 5/6 chance when there isn't a Medusa, and won't spawn when there is one.
"settled"? (You mean natural vs smooth?) Certainly I did seem to find that Medusa was posibly always first to spawn, each test, followed by a Hoplite, sometimes. And never got 2 at once.
do some biomes take precedence over there in spawns?
Well, generally it goes by biome of spawn surface block found during spawn event, so less a case of 'precedence' and more of randomness biased (or controlled) by terrain. Presumably a marble (or granite) block is equally likely to produce a marble (or granite) mob, respectively, regardless of the spawn surface block types...? That kind of thing would take a lot of effort to test to any degree of veracity.
anything special about the new underground desert?
I'm feeling that Antlion mobs are a little different too, since I had one turn up, in my above tests, when it's only possible source was half a dozen hardened sandstone back walls. Couldn't duplicate though. There's a lot of different sub-rules for spawning... Chaos Elementals require a lack of player placed wall behind you, for example (Diceman found in one of his farm builds). Can't remember if they need a halloed surface, or just a active biome...?