Dehumidifier for springtails and centipedes

Springtails (Collembola) are one to two millimetres long, usually grey, white or faintly purple, and they spring away when disturbed — that is the behaviour that names them. They eat mould, mildew, algae and decaying plant material. They do not eat wood, fabric, food or people, and they do not bite. What makes them useful to you is their skin: a springtail has almost no waterproofing on its cuticle and loses water straight through the body wall. In genuinely dry air a springtail desiccates within hours. It cannot survive away from a substrate that is at or very near saturation.

So a springtail on your bathroom floor is not a pest that wandered in. It is an animal that can only be there because something within a few feet of it is holding water and growing fungus.

House centipedes (Scutigera coleoptrata) are the fast, long-legged ones that cross a wall at speed and make people leave the room. They are predators. They eat springtails, silverfish, small cockroaches, spiders, fly larvae and anything else soft enough. They damage nothing — no wood, no fabric, no stored food. They too lose moisture readily and stay where the humidity is high.

Which gives you the whole diagnosis in one line: springtails mark the wet, and centipedes mark the springtails. Finding centipedes is second-hand evidence — it tells you the building is feeding a predator, which means the prey population is already established somewhere you have not looked yet.

The mechanism is identical for the two animals, which is why they belong on one page rather than two. But it reaches them by different routes, and the difference matters for what you should expect to see.

On springtails, the effect is direct. Lower the humidity and two things happen at once. The animals themselves start losing water faster than they can replace it, and — more slowly but more permanently — the fungal growth they feed on stops. Mould needs sustained high moisture at the surface it is growing on. Take that away and the food supply ends, whatever happens to the individuals.

On centipedes, the effect is indirect and delayed. A centipede responds to prey, not to a hygrometer. When the springtails and silverfish thin out, the centipedes have nothing to hunt and disperse. That takes considerably longer, and there is a counter-intuitive stage in the middle: you may briefly see more centipede activity, not less. A hungry centipede ranges further and crosses open floor more often, so it becomes more visible at exactly the point the treatment is working. If you did not know that was coming, it reads as failure.

The band worth aiming for in an occupied space is 30–50% relative humidity, with 45% a reasonable target. Above roughly 60%, sustained, you are in the range that supports mould growth — which is the springtails’ food supply, and therefore the whole problem. Below about 30% the air gets uncomfortable for people and starts opening joints in timber and flooring, so drying harder is not better.

Here is the part most guidance skips. Springtails do not live in the air. They live on surfaces, and a surface can be soaking wet in a room that reads a perfectly ordinary humidity — because what wets a surface is not the room’s relative humidity but whether that surface is colder than the room’s dew point.

Dew point is the temperature at which the air currently in the space has to give up its water. Anything colder than that number is wet, permanently, no matter what the hygrometer on the wall says.

Read the middle rows. A crawl space at 70°F and 70% RH has a dew point of 59.8°F. Foundation walls in contact with soil, cold-water supply lines and uninsulated ductwork routinely sit in the low-to-mid 50s through the summer. Every one of those surfaces is therefore carrying a film of liquid water, and the local humidity right at that film is 100% regardless of what the room average is. That film is the springtail habitat.

It is also why pulling the room down by ten or fifteen points has such a disproportionate effect. Going from 70% to 50% moves the dew point from 59.8°F to 50.5°F — below the temperature of almost anything in the space. The wet surfaces do not get slightly drier. They stop being wet.

Every location on the following list is somewhere the dew point argument above plays out locally, whatever the rest of the house is doing.

This is the most important paragraph on the page, and it is the one that argues against buying anything today.

A living space sitting at an ordinary 55% humidity does not produce hundreds of springtails. General background humidity produces the occasional one. Numbers — dozens on a bathroom floor, a rim of them around a basement wall, a fresh crop every morning — mean a specific wet thing, close by, that is keeping a small area permanently saturated. Until you find it, a dehumidifier treats the room while the source keeps producing.

Work through these before you spend money:

Run a dehumidifier without doing this and the numbers will genuinely fall — and they will come back within a week of switching it off, because the source never stopped. The dehumidifier is the second thing you do, not the first.

One free step worth taking first: most US states run a Cooperative Extension service that will identify an insect from a clear photograph at no cost. It is worth confirming you have springtails rather than something that needs a different answer entirely — we cannot identify anything for you, and neither can a product page.

Once you have found the source and want a machine to hold the space dry, one thing will trip you up before any product choice does.

In 2019 the US Department of Energy moved the dehumidifier test condition from 80°F to 65°F. A dehumidifier removes less water from colder air, so the same physical machine now earns a lower number — roughly 71% of its old figure. A modern “50 pint” unit removes about what a pre-2019 “70 pint” did. Every sizing chart written before 2019, and most of the ones circulating now because they were copied from those, is on the old scale. Follow one with a modern machine and you undersize by about a third. We publish both numbers and always say which is which.

Note what we did not size on. Three units in our own product data all rated 155 pints/day advertise coverage of 7,500, 6,000 and 6,000 square feet, and a 160-pint unit advertises 2,500. There is no standard behind the square-footage claim, so it cannot be compared between brands. Size on pints per day, which is actually tested.

One more detail specific to crawl spaces: there is usually nowhere to gravity-drain to, so the condensate has to be lifted out. At 22 pints/day the duty is about 0.115 gallons per hour — trivially small. Every condensate pump on the market is enormously oversized for that flow. You are buying the pump for lift, not for volume, and a pump’s rated capacity falls as the lift rises, so check the head your run actually needs.

A crawl-space unit is built to sit in a hostile space and run unattended for years, with a pump and a duct kit. In a finished basement, laundry room or bathroom you do not need any of that, and the sizing requirement is much smaller because you are not carrying the crawl-space uplift.

For a room the constraint is different: you want continuous drainage so the machine does not stop when a tank fills, and you want it small enough to live with.

Nothing on this page kills anything on contact, and it is worth being blunt about what the first fortnight looks like, because most people switch the machine off too early.

Hours to a day. Air humidity falls. This is the only fast part, and it is the part that shows on your hygrometer, which is misleading — the number moves long before the problem does.

Two to six weeks. Wet materials give up their water. Subfloor timber, joists, block walls, stored cardboard and carpet underlay hold far more water than the air does and release it slowly. This is the stage that actually matters, because it is the wood’s moisture — not the air’s — that feeds the mould that feeds the springtails. Timber that has been sitting above 20% moisture content for a season does not dry in a weekend.

Two to four weeks for the springtails. You should see a clear decline inside a month. Individuals die of desiccation quickly once their local surface dries, but you are waiting on the substrate, not the air.

Four to eight weeks for the centipedes, lagging the springtails, and possibly with a visible uptick first as hungry animals range further. If you evaluate the result at week two you will conclude it failed.

If you want them gone this afternoon, a vacuum cleaner is the correct tool. It removes the animals in front of you and does absolutely nothing about why they are there — which is fine, as long as you know that is the trade you are making.

We have never run any of these units in a crawl space, and we have never run a pest trial of any kind. What we do instead is publish the arithmetic, on both rating scales, so you can check it. The sizing figures come from the AHAM dampness classes converted onto the modern post-2019 scale; the dew points are the Magnus formula; the pint ratings are the manufacturers’ own DOE figures.

The biology on this page is standard entomology rather than anything we measured, and your state Cooperative Extension service will confirm or correct any of it for free. The timescales are practitioner-level generalisations — the shape is reliable, the exact weeks are not a measurement.

And the thing worth repeating: if the source is still producing water, no machine on this page fixes anything permanently. Find the leak first.