Four biological options get sold for fungus gnats. The soil predatory mite Stratiolaelaps scimitus, beneficial nematodes, the rove beetle Dalotia coriaria, and Bacillus thuringiensis subsp. israelensis — the last of which isn't even an organism you keep.
They're usually presented as alternatives — as though the question were which one is best.
It isn't, and there's a specific reason. No published trial has put all four against each other under a single protocol. What exists is a decent body of single-agent data, a handful of pairwise combination studies, and a set of results that answer genuinely different questions — which is not the same as answering yours.
Speed isn't the same question as persistence. Persistence isn't the same question as which life stages get reached. And at least one popular combination is documented to cancel itself out.
The short version
- Fastest
- Bti and nematodes — dead larvae inside one to two days
- Longest lasting
- S. scimitus, which can hold a season. Rove beetles colonize and persist
- Reaches pupae
- Rove beetles. Mites don't attack pupae, and Bti only works on feeding larvae
- Works deepest
- Nematodes — they swim the water film through the whole profile
- Best documented disappointment
- Bti: 92% larval mortality in the lab, 26% emergence reduction in a greenhouse
- Never combine
- Rove beetles with predatory mites. Documented predation in both directions
Four different mechanisms
Worth being precise about, because the mechanism predicts nearly everything else — how fast it acts, how long it lasts, what it can reach, and what it's compatible with.
| Agent | What it is | How it kills |
|---|---|---|
| Stratiolaelaps scimitus | A predatory soil mite, about 0.4mm, living in the top inch of substrate | Hunts and eats larvae directly. A handful a day, reproducing in place, and scavenging algae and debris between pest events |
| Steinernema feltiae | An entomopathogenic nematode, microscopic, moving through the water film on substrate particles | Infective juveniles enter the host and release symbiotic bacteria. Death by septicemia, usually inside one to two days |
| Dalotia coriaria | A 3–4mm rove beetle. Adults and larvae are both predatory, and the adults fly | Hunts actively through the surface layer, taking larvae and pupae. Highly mobile — it relocates itself to the hotspots |
| Bti | A bacterium applied as a drench. A biological insecticide rather than an organism you establish | Eaten by feeding larvae, then the toxin crystals rupture the gut lining. No reproduction, no persistence |
Two of these you're establishing — a standing population that lives in the substrate and keeps working. Two of them you're applying — an intervention that does its job and is gone.
Most disappointment with soil biologicals comes from one thing — expecting a category you apply to behave like a category you establish.
Which life stages each one reaches
A fungus gnat spends its life as an egg, four larval instars, a pupa and a short-lived adult. No single agent covers all of it — and the gaps are where reinfestation comes from.
| Eggs | Larvae | Pupae | Adults | |
|---|---|---|---|---|
| S. scimitus | Negligible | All instars, prefers 1st | Not attacked | No |
| Nematodes | No | Yes | Limited | No |
| D. coriaria | Some | Yes | Yes | No |
| Bti | No | Feeding larvae, best on 1st instars | No | No |
| Sticky cards | No | No | No | Yes — monitoring, and some mass trapping |
The pupal gap
The foundational laboratory work on S. scimitus found it consumed all larval instars of its sciarid prey, that egg predation was negligible, and that pupae were not attacked at all.
This is the single most useful mechanical fact in the comparison. It explains why a mite release into an established population produces two to three more weeks of adult emergence, and why the rove beetle — which does take pupae — behaves differently on the same timeline.
It also explains why nothing in the soil touches the adults, and why sticky cards aren't optional.
Speed versus persistence
These trade against each other almost perfectly — and the trade is the actual decision.
Bti — fast, gone in two days
UConn puts the toxic window at about two days, best against first instars, with two or three repeat applications at high rates often needed. Nothing establishes.
Nematodes — fast, active 10–14 days
Death follows infection inside one to two days, and extension guidance puts active persistence in the medium at roughly ten to fourteen. A treatment, not a resident.
S. scimitus — slow, potentially a season
Kansas State says one application can establish a population for an entire growing season. It survives three to four weeks with no prey and doesn't diapause.
D. coriaria — moderate, and it colonizes
A generation in about 17 days at 79°F, roughly 90 eggs per female, adults living around 48 days. It builds a standing population and moves itself around.
Which gives a clean rule. If the question is how do I stop this now, the answer is nematodes or Bti. If the question is how do I stop having this problem, the answer is a mite or a beetle established in the substrate before the problem starts.
Most people ask the first question — and buy the answer to the second.
What the trials actually measured
Since there's no head-to-head study, the honest thing is to give each agent its own best evidence and let the differences in test design show — because those differences are the story.
Predatory mites
Against thrips in greenhouse chrysanthemum, S. scimitus at roughly 1,000 mites per square meter brought counts from 53.7 per flower in untreated houses down to 13.5 — a 74.9% reduction that held through high summer temperatures.
Combined with an entomopathogenic fungus against thrips, the mite alone cut larvae by 72.6% and adults by 66.1%. The fungus alone managed 59.4% and 53.6%. Together they reached 86.0% and 82.2%.
That's the clearest published case of two soil agents beating either one alone — and it involves no mites eating each other, which is more than can be said for some combinations below.
Nematodes
Application rates of 1.25 and 2.5 × 10⁵ infective juveniles per square meter both significantly reduced fungus gnat populations in impatiens, and the higher rate bought nothing extra. Poinsettia needed more.
Nematode performance was diminished in nursery mix compared with other media, and application timing mattered in one crop and not the other. In separate work, doubling the commercial dose significantly reduced gnats at 21 days, and coir, pine sawdust and coir-plus-vermiculite all improved nematode mobility.
Rove beetles
This is the result that should be better known, and almost nobody quotes the second half of it.
In small laboratory containers, a single rove beetle cut adult gnat recovery from 5.4 in the untreated controls to 2.7 — and the growing medium alone moved the same measure from 6.0 in one commercial blend to 0.9 in another.
Then the same study ran the beetles in 2-liter, 15cm greenhouse pots. The rove beetle treatments did not differ significantly from the untreated control. The authors put it down to container volume and searching area.
A follow-up found that ten beetle adults per container regulated gnat larvae regardless of starting prey density — at which point you are, admittedly, running a beetle farm.
Read those together and the message isn't that rove beetles don't work. It's that the dose has to scale with the container, and a rate derived from a deli cup will disappoint you in a real pot.
Bti
The best-documented gap between laboratory and greenhouse performance in this whole comparison.
A 2023 study recorded 92% mortality of second and third instar larvae at 14 days after treatment in the lab — and only a 26% reduction in adult emergence in the greenhouse bioassay at the same concentration. The classic 1985 work established the dose–response and showed that continuous exposure from egg to pupa left only 8% surviving against 84% in water controls.
Both numbers are true. The difference between them is the difference between a petri dish and a pot with drainage — and it's the single most useful thing to know before you form an expectation.
The variable nobody accounts for
Two studies from the same laboratory found something inconvenient — the growing medium changed the outcome more consistently than the moisture level did.
Rove beetles suppressed gnat emergence in one commercial peat blend and not in another. Larval survival differed significantly between two other blends. Nematode efficacy was reduced in nursery mix relative to other media.
Nobody sells a mite or a nematode calibrated to your particular bag of substrate, and the published comparisons are between commercial blends rather than between materials — so peat versus coir isn't a question the literature answers cleanly.
What it does establish is that medium is a real variable of the same order as dose — worth remembering before you conclude that a product which worked for someone else has failed for you.
One physical result worth knowing
A 3.2cm layer of coarse recycled-glass granules over the substrate surface significantly reduced or delayed adult fungus gnat emergence, on its own, as a purely physical barrier.
In the same study, combining that barrier with rove beetles didn't suppress emergence any further.
A top dressing is a legitimate part of a fungus gnat program. It's also, like everything else here, not additive with everything else you might do.
What combines, and what cancels out
| Combination | Verdict | Evidence |
|---|---|---|
| Mites + nematodes | Recommended | Different depths, different speeds, different mechanisms. Extension guidance recommends exactly this pairing once populations are established |
| Mites + Bti | Recommended | Bti knocks the larval population down and the mites hold the surface afterwards. The standard extension recommendation for an active problem |
| Mites + entomopathogenic fungus | Good evidence | 86.0% larval and 82.2% adult thrips reduction combined, against 72.6% and 66.1% for the mite alone |
| Nematodes + rove beetles | Compatible | The beetle is a poor host for S. feltiae — highest mortality recorded was 25.4%, and the nematode can't reproduce in it. The authors concluded beetle populations shouldn't be significantly affected |
| Mites + rove beetles | Don't | Bidirectional intraguild predation. The commercial laelapid mites, S. scimitus included, attacked every larval stage of the beetle; beetle adults ate all twenty eggs of the sister species offered to them and almost all its young nymphs. Kansas State advises against simultaneous use |
| Mites + springtails | Don't | Laelapid soil mites prey on springtails. Direct evidence is for the sister species Gaeolaelaps aculeifer, which kills and eats Folsomia candida |
| Soil agent + foliar predator (thrips) | Depends on the crop | Better and faster than either alone in cyclamen. Added nothing over foliar predators alone across gerbera, chrysanthemum and rose |
The mite-plus-beetle case catches out the people trying hardest — both products are sold for the same pest, and buying both feels like diligence.
Worth restating plainly. They'll each reduce your fungus gnats, and they'll also reduce each other.
Choosing, in practice
Fresh substrate, no problem yet
S. scimitus at potting. This is the scenario where it's unambiguously the best of the four — establish it, keep the top inch damp, and one application may hold the season.
A few adults on the card
S. scimitus now, and start counting. You're ahead of the curve and the mite can stay ahead of it.
Weeks of visible adults
Nematodes or Bti for the knockdown, mites underneath for the hold. Don't expect the mite alone to reverse this — the eggs and pupae are out of its reach.
Large containers, deep profile
Nematodes reach through the profile and the mite works one inch down. Scale the mite dose to surface area and lean on nematodes for depth.
A greenhouse or a big grow space
Rove beetles, which fly and find hotspots on their own — but dose to the container volume, and then don't also run mites.
Reptile enclosure or bioactive setup
S. scimitus, and not springtails in the same substrate. Pick a crew.
The version that holds across all of them is simple enough. The soil is one layer of a program, and no single organism covers a whole life cycle.
Buy speed and persistence separately, put a sticky card in every pot you care about, and let the counts tell you which layer is doing the work.
The rest of the Stratiolaelaps series
- You Bought the Wrong Species. So Did Everyone Else.What the species actually is, what it eats, and which label claims survive the literature.
- Shake, Sprinkle, Wait. Mostly Wait.Where the rates come from, how they convert into pots, and what the first month looks like.
- Your Soil Mites Didn't Fail. Nine Things Did.Nine causes, each with the fingerprint that tells them apart.
- One Inch of Soil. That's the Whole Game.Pore space, water films, and why one inch of substrate decides everything.
All of Mite MattersEvery organism we sell, written up against the literature rather than the label.
Common questions
Which one should I start with?
It depends on which question you are actually asking, because these four answer different ones. If you need the larvae dead this week, Bti and nematodes kill inside one to two days. If you want something that keeps working, S. scimitus can hold for a season on one application and rove beetles colonize and persist. Speed and persistence trade against each other almost perfectly, and that trade is the real decision.
Can I use predatory mites and nematodes together?
Yes, and it is the standard pairing once a population is established. They work at different depths by different mechanisms — the mite hunts the top inch, the nematode swims the water film through the whole profile — so they cover between them a good deal of what neither reaches alone.
Can I use predatory mites and rove beetles together?
No. Intraguild predation is documented in both directions, and Kansas State’s guidance says simultaneous use should be avoided. Adding the second predator does not add coverage; it subtracts the first one.
Why does Bti look so much better in the lab than in my pots?
Because the lab is measuring a different thing. Bti produced 92% larval mortality in laboratory conditions and about a 26% reduction in adult emergence in a greenhouse. It only works on feeding larvae, its toxic window is roughly two days, nothing establishes, and two or three repeat applications at high rates are often needed. It is an intervention, not a resident.
Does anything in the soil kill the adult gnats?
No. Nothing you put in the substrate reaches the adults — not the mite, not the nematode, not the rove beetle, not Bti. That is why sticky cards are not optional: they are the only thing acting on the flying stage, and they are also how you read whether any of the rest is working.
References
References
- Wright, E.M. & Chambers, R.J. (1994). The biology of the predatory mite Hypoaspis miles (Acari: Laelapidae), a potential biological control agent of Bradysia paupera. Entomophaga, 39(2), 225–235. springer.com
- Herrick, N.J. & Cloyd, R.A. (2017). Effects of growing medium type and moisture level on predation by adult rove beetle, Dalotia coriaria, on fungus gnat, Bradysia sp. nr. coprophila, larvae under laboratory and greenhouse conditions. HortScience, 52(5), 736–741. ashs.org
- Herrick, N.J. & Cloyd, R.A. (2018). Effects of growing medium type and predator:prey ratio on rove beetle, Dalotia coriaria (Coleoptera: Staphylinidae), adult predation on fungus gnat, Bradysia sp. nr. coprophila (Diptera: Sciaridae), larvae. HortScience, 53(10), 1441–1446. doi.org
- Jagdale, G.B., Casey, M.L., Grewal, P.S. & Lindquist, R.K. (2004). Application rate and timing, potting medium, and host plant effects on the efficacy of Steinernema feltiae against the fungus gnat, Bradysia coprophila, in floriculture. Biological Control, 29(2), 296–305. doi.org
- Katumanyane, A., Ferreira, T. & Malan, A.P. (2018). Greenhouse application of Steinernema yirgalemense to control fungus gnats, Bradysia impatiens. BioControl, 63(5), 729–738. springer.com
- Osborne, L.S., Boucias, D.G. & Lindquist, R.K. (1985). Activity of Bacillus thuringiensis var. israelensis on Bradysia coprophila (Diptera: Sciaridae). Journal of Economic Entomology, 78(4), 922–925. doi.org
- Duarte, A.F. et al. (2023). Toxicity of Bacillus sp. on the fungus gnats, Bradysia aff. ocellaris larvae (Diptera: Sciaridae). Bioscience Journal, 39, e39089. doi.org
- Zhang, X., Wu, S., Reitz, S.R. & Gao, Y. (2021). Simultaneous application of entomopathogenic Beauveria bassiana granules and predatory mites Stratiolaelaps scimitus for control of western flower thrips. Journal of Pest Science, 94(1), 119–127. springer.com
- Jung, D.O., Hwang, H.S., Kim, S.Y. & Lee, K.Y. (2019). Biological control of thrips using a self-produced predatory mite Stratiolaelaps scimitus in the greenhouse chrysanthemum. Korean Journal of Applied Entomology, 58(3), 233–238. agris.fao.org
- Jandricic, S., Scott-Dupree, C.D., Broadbent, A.B., Harris, C.R. & Murphy, G. (2006). Compatibility of Atheta coriaria with other biological control agents and reduced-risk insecticides used in greenhouse floriculture IPM programs for fungus gnats. The Canadian Entomologist, 138(5), 712–722. doi.org
- Raudenbush, A.L., Cloyd, R.A. & Echegaray, E.R. (2014). Effect of a physical barrier on adult emergence and egg survival associated with the fungus gnat Bradysia sp. nr. coprophila under laboratory conditions. HortScience, 49(7), 905–910. doi.org
- Pozzebon, A., Boaria, A. & Duso, C. (2015). Single and combined releases of biological control agents against canopy- and soil-dwelling stages of Frankliniella occidentalis in cyclamen. BioControl, 60(3), 341–350. springer.com
- Manners, A.G., Dembowski, B.R. & Healey, M.A. (2013). Biological control of western flower thrips in gerberas, chrysanthemums and roses. Australian Journal of Entomology, 52, 246–258. doi.org
- Jensen, K., Toft, S., Sørensen, J.G., Sigsgaard, L., Kristensen, T.N., Overgaard, J. & Holmstrup, M. (2019). Prey-specific experience affects prey preference and time to kill in the soil predatory mite Gaeolaelaps aculeifer Canestrini. Biological Control, 139, 104076. doi.org
- Cloyd, R.A. (2023). Stratiolaelaps scimitus: Biological Control Agent of Fungus Gnats and the Western Flower Thrips. Kansas State University, MF3632. ksre.ksu.edu
- Pundt, L. Managing Fungus Gnats in the Greenhouse. UConn Extension Integrated Pest Management. ipm.cahnr.uconn.edu
- Pundt, L. Biological Control of Fungus Gnats. UConn Extension Integrated Pest Management — nematode persistence in the medium, the Bti toxic window, and Dalotia coriaria life history. ipm.cahnr.uconn.edu
Speed from one layer, persistence from another. They're different purchases.
