So apart from butchering a perfectly serviceable expression of metaphysical correspondence, and while I think it’s totally acceptable to have incongruous monsters in your dungeons, I think a dungeon-stocking principle might be to derive some underworld encounters from overworld encounters.

Maybe this is obvious to some. Others, like me, need help.

Encounter tables can (may? kind of?) form the core of player expectations for monsters both above- and below-ground. Another way to put it might be like, narrative concepts of foreshadowing. Sometimes literally as with a dragon’s shadow as it flies overhead. And your players soil themselves.

But an encounter table is also a good place for a GM to turn when they don’t know what to do — I think this happens most with WotC-era “story” GMs.

Using myself as an example.

It’s a good idea to stick with your encounter table rather than bringing in some new monster type. Sometimes you can have a rival group of the same type of monsters move in — these goblins don’t like those goblins.

But also, the monsters you’re less likely to encounter in the overworld — are probably be movers and shakers in the underworld. They might not be the top dogs, but they’re major players. (Or pawns to something else.)

One of the reasons they’re rarer “up top” is because they tend to be at home, underground, in their lairs. I mean, that’s one explanation.

SO, what’s this all about then?

Well, I made some more progress on the cornerstones, going so far as to cut together a few by using 3e and 5e for reference. I’ve noticed 5e has a more conservative estimate of monster capability.

To wit, 5e monsters have a CR about one-half that of their 3e equivalents.

That’s at a glance. Looking back on my time as a 3e GM, I can get behind this. Often those high-CR monsters were appealing, but the XP for their CR was too great a reward for their respective challenge.

At the time of this writing, I’ve hacked together about sixteen cornerstones. Most are composed of exactly four monsters of ascending CR. A couple have three, and a couple have five. I’m working to ensure they’re exactly four.

A couple of them are/were obvious, but most aren’t.

I know there was a time when monsters were invented to fill gaps in HD- such as between goblins, hobgoblins, and bugbears of the goblinoid family of monsters.

Beyond a few notable examples however, monsters don’t seem to “stand together” in the sense that you’d encounter them within a single biome or region. Which makes me glad for my biology research.

Also, the research I’ve done into various myths and legends.

Beyond simply grouping “like” monsters together, I’ve tried to make sure they are the sorts of monsters you might reasonably encounter together.

Thematically or otherwise.

Finally, I’ve tried to keep monsters from like mythologies together — which led to separate “Greek” monster cornerstones in some cases. Some critters are more “Arcadian,” and some are more “Olympian,” if that makes any sense.

In the end, if what you want is a Greek-themed campaign, you should be able to slot together several different Greek cornerstones together.