Glass thickness and room heating: where the humidity actually goes

The room reads 32 percent, the tank reads 82, and the glass is wet on the inside from the waterline up. The keeper assumes the tank is leaking humidity through 6 mm glass and starts shopping for a thicker build. Meanwhile the actual loss is going out of a 40 by 15 cm mesh panel on top, and no amount of glass will touch it.

Water leaves a vivarium through a small number of routes, and glass is the least important of them. Knowing which route dominates in your setup turns a vague humidity problem into a specific fix, usually for under twenty pounds.

Where the water actually goes

Four routes carry almost all of it. The lid or mesh top is first by a wide margin, because it is a large area with free air movement above it. Gaps around doors and cables are second. Condensation running down the glass and into the drainage layer is third, and it is not a loss at all, just a relocation. Direct evaporation from an open water section is fourth.

Glass itself is effectively a vapour barrier. A 4 mm pane and a 10 mm pane pass so little water vapour that the difference is unmeasurable in a terrarium context. What glass does do is conduct heat, and that matters because a cold pane creates a condensation surface. Water that condenses on the glass is water that has left the air, and in a tank with a drainage layer it ends up at the bottom rather than in the substrate.

What the cold pane is doing to your numbers

A single-glazed tank in a room at 19 C with an internal temperature of 26 C has an internal glass surface around 22 to 23 C. That is below the dew point of air at 85 percent, so water condenses continuously. The relative humidity inside the tank then reads lower than the water content suggests, because the air has been dried by the glass.

This is why tanks in cold rooms seem to drink water. They are not losing it to the room through the glass. They are losing it to the glass, where it runs down and sits in the drainage layer, and the substrate above slowly dries as it tries to rebalance. A tank with a good drainage layer and a substrate that does not touch the water table will dry from the top while the bottom stays wet, which is a confusing thing to watch.

Radiators, walls and the cold corner

Placement changes the numbers more than construction does. A tank against an external wall in an unheated room has one cold face and will condense heavily on that side, and the substrate near that face will run wetter than the rest. A tank above a radiator has a warm base and a cold top, which inverts the normal gradient and dries the surface faster than the floor.

The worst arrangement we see is a rack against a single-glazed window with a radiator underneath. The top tubs run at 40 percent, the bottom tubs at 90, and the keeper spends a fortune on gel trying to average the two. Move the rack a metre and the spread halves.

Three fixes ranked by cost

• Cover 60 to 70 percent of the mesh top with glass or acrylic, leaving a strip for air exchange. Under ten pounds, biggest single effect
• Insulate the back and one side with 5 mm foam board or cork. Under twenty pounds, cuts condensation and stabilises temperature
• Seal cable gaps and door edges with foam strip. Under five pounds, small but reliable effect

Do them in that order and measure after each one. A tank that drops from 82 to 65 percent in two days will usually hold at 80 for four days after the lid is covered, which is the difference between a twice-weekly routine and a daily one.

What you should not do is add gel to compensate for an open lid. The gel will hold more water, the lid will still remove it at the same rate, and you end up with a heavier, wetter floor that fails faster when the reservoir runs down. Fix the exit before you increase the tank. If you want the reference numbers for how much a given gel dose holds, the SAP powder pages are the place to look, but the lid is still the first thing to change.

Take a photo of your tank from the front with the light on, at the end of a normal day. The condensation pattern in that photo shows you exactly where the heat is going. It is the cheapest diagnostic in the hobby and almost nobody does it.