Twelve Centimetres Between Drowning and Drought

Stared at three pieces of dragon stone for forty minutes before realizing the problem wasn’t the rocks.

The aquascaping tank sits on my left, finally clear after six weeks of bacterial cycling. The terrarium jar sits on my right, sealed since May, its miniature weather system condensing and raining every eight hours. Between them, an empty 30-litre glass cube. I’d been thinking about this for days: build something that’s both. Water and land. Aquatic and terrestrial. A paludarium.

The word comes from Latin palus—swamp, marsh, land near water. Not aquarium-with-decoration. Not terrarium-with-a-puddle. Something that lives at the boundary.

Which is harder than it sounds, because boundaries don’t want to stay put.

The shop owner—same one who sold me the refractometer last week—listened to my plan and said: “You’re going to over-engineer the water and under-engineer the land. Everyone from reef tanks does.”

He was right. I’d been mentally speccing a protein skimmer for a 15-centimetre-deep freshwater pool that would hold maybe four ember tetras. Meanwhile my substrate plan was “pile dirt behind some rocks.”

The actual engineering problem: build a dam that holds saturated soil above a waterline without collapsing. Physics wants the soil to slump into the water. Capillary action wants to wick moisture up from the pool and flood the terrestrial section. And you want it to look like a riverbank, not a retaining wall.

Two approaches exist. First: hardscape barrier method—stack rocks, glue them with cyanoacrylate or epoxy, create a physical dam. Second: floating shelf method—attach an acrylic platform to the tank walls just above the waterline, maximize water volume, sacrifice aesthetics. The shelf keeps parameters stable (larger water volume = slower chemistry changes, everything I learned calibrating that reef refractometer applies). But it looks like furniture floating in a swamp.

I picked hardscape. Bought two more kilograms of dragon stone, the same grey-green limestone texture from the aquascaping build. Spent those forty minutes arranging them outside the tank—dry-fitting, testing stability, looking for pieces that interlock under their own weight. Guides warn about this constantly: saturated soil weighs more than you expect. A 30L tank can hit 70 kilograms total. Drop a rock carelessly during construction and you crack the glass.

Filter cloth goes down first as a cushion. Then rocks, stacked to create a barrier that rises from the front-left corner and curves toward the back. Superglue on contact points—thick viscosity, sets in thirty seconds. The glue joint is the failure point. If it gives, the soil collapses, and I’m draining the tank to rebuild.

Substrate behind the barrier: drainage layer of lava rock gravel (2 cm), then a mix of peat moss, perlite, and coco coir (5 cm). Not garden soil. Not aquasoil. Sterile substrate that won’t introduce pathogens into the humid closed environment. The terrestrial plants will root here and stay dry while water sits 10 cm away on the other side of the stone dam.

Aquatic substrate in front: fine sand. No nutrients needed—I’m planting Anubias and Java moss, both epiphytes that attach to rock and don’t root in substrate.

Then the critical decision: emersed plants.

Three categories exist. Submersed: fully underwater. Terrestrial: fully dry. Emersed: rooted below the waterline, growing into air. Emersed plants are the gradient species—evolved for riverbanks that flood seasonally. Anubias barteri will grow submerged, but given the chance it’ll push leaves above water and eventually flower. That’s the inversion: in an aquarium you’re holding it underwater, hoping it tolerates the conditions. In a paludarium you’re letting it do what it actually prefers.

Planted three Anubias rhizomes at the waterline, tied to rocks with fishing line. Positioned Java moss on the barrier stones—it’ll grow both underwater and terrestrially if kept moist. Terrestrial section gets Fittonia (nerve plant) and a small fern, same species from the terrarium build.

Waterfall next. Routed the canister filter outlet hose up through the back rocks, positioned the spout so water cascades down a groove in the dragon stone before returning to the pool. Simple. Functional. Except now evaporation accelerates—splashing increases surface area, humidity spikes, I’ll be topping off water three times weekly instead of weekly. The waterline becomes a moving target. As water evaporates, the “land” boundary creeps upward. Too much evaporation and the filter intake goes dry. Too little top-off and terrestrial plants drown.

This is the part that inverts reef experience. Reef chemistry is about precision: ±0.001 specific gravity per 5°C temperature change, calcium dosing to maintain 400-450 ppm, tight parameter windows. Paludariums want 90%+ humidity but you can’t seal the tank or the glass fogs and you lose visibility. Humidity management is qualitative—watch the moss, mist when it starts browning. No number to chase. No refractometer reading. Just observation.

Filled the water section slowly, watching the dam. Rocks held. Substrate stayed put. Filter started. Water cascaded down the stone face, splashing into the pool, cycling through the canister, returning. The Anubias leaves broke the surface at a 30-degree angle, half-submerged, exactly where they’re adapted to be.

Glass fogged within five minutes. Condensation running down the sides. The terrarium’s weather system, but open-topped, unstable, requiring adjustment. I’m not sure yet if that’s a feature or a failure.

The reef tank is still cycling—week two, ammonia down to 0.5 ppm, nitrite starting to climb. The aquascaping tank finally grew its carpet after eight weeks of patience. This one’s somewhere between: faster to establish than a reef, slower to stabilize than a terrarium. A gradient system with gradient timescales.

Waterline sits at 12 cm. Dam holds. Tomorrow I’ll check if the terrestrial substrate stayed dry or if capillary action won.