Are natural pest controls safe for bees?

Are natural pest controls safe for bees?

Not automatically, and the assumption that they are is the most common mistake in this area. Some of the most acutely bee-toxic insecticides in use are botanical. Natural pyrethrins, extracted from chrysanthemum, are toxic enough to honey bees that EU and US regulators flagged acute contact risk in their reviews. Saponins - the active class in our own Slug Control · were among the most toxic compounds in a screen of 63 plant allelochemicals fed to honey bees, causing mortality at dietary concentrations between 0.003% and 0.6%.

Origin is not a safety category. What matters is the specific compound, the dose, the formulation, and above all when you spray.

Two actives in this range that bees meet directly

Geraniol is not just tolerated by honey bees - they make it. Geraniol was identified in 1962 as the active attractant component of the Nasonov gland secretion, the pheromone worker bees release to orient nestmates and mark food sources and nest entrances. Commercial swarm lures use it: a slow-release blend of citral, geraniol and nerolic and geranic acids drew 19 of 23 reproductive swarms to artificial nests, against 4 for pheromone-free controls.

That is genuinely reassuring about recognition and attraction, and it is not the same as a safety clearance. Bees detect geraniol as a signal - an electroantennogram study confirms a distinct antennal response - but the same paper notes that detailed toxicity data for geraniol are still lacking. No acute LD50 for geraniol in Apis mellifera has been published that we could locate. Anyone quoting one should be asked for the source. What can be said is that geraniol is a contact material with short persistence, so exposure is governed almost entirely by timing.

Thymol is the opposite case: bees are deliberately dosed with it. Thymol is the active in registered in-hive Varroa treatments such as Apiguard and Api Life Var, applied inside colonies at concentrations far above anything a foliar spray would deliver. That sounds like a clean bill of health. The literature says otherwise, and this is worth knowing if you use thymol products near hives.

In-hive thymol produces measurable sublethal effects. Colonies treated with a 74% thymol miticide showed reduced phototactic behaviour weeks and months after treatment. A residue study measured a median body burden of 4.3 µg per bee at 24 hours, including 11 ng in the brain, and found thymol persisting in wax at around 10 mg/kg a full year after treatment. In vitro larval rearing found both lethal and sublethal developmental effects at in-hive-relevant doses. And in a direct selectivity comparison, thymol and carvacrol raised bee mortality significantly above solvent control while γ-terpinene did not - selectivity ratios of 48 to 91-fold in favour of the mite, which is a workable margin but not an absence of effect.

So thymol is well characterised in bees, deliberately used in bees, and still carries documented sublethal costs. That is the honest position.

Why LD50 is the wrong question

Acute lethal dose measures whether a bee dies in 48 hours. It does not measure whether a forager can still navigate home.

A widely cited study tracked free-flying foragers by RFID and found that a sublethal, non-lethal dose caused enough homing failure to compromise colony dynamics - bees that survived exposure simply did not return. The canonical review of sublethal effects in beneficial arthropods makes the general argument: physiological and behavioural endpoints are required for a complete risk picture, and acute mortality alone systematically understates impact.

This applies to natural actives too. Chronic oral exposure to carvacrol and thymol produced measurable sublethal effects at doses well below acute lethality.

Timing is the strongest lever you control

For short-persistence contact materials - which is most of what this range contains - when you spray matters more than what you spray.

The regulatory framework for this is RT25, the residual time to 25% bee mortality. The US EPA maintains a database of field-weathered residue toxicity tests. Products with an RT25 under roughly 8 hours pose minimal hazard if applied in late evening or at night, because residues decay below harmful levels before foragers return. Above 8 hours, extended residual toxicity requires additional mitigation.

Field evidence supports the principle. A seven-week replicated field study of truck-mounted ULV applications made when bees were not foraging found no significant difference in bee mortality, colony health or detoxification enzyme activity against untreated controls.

One honest caveat: a 2024 systematic review recalculating RT25 across 50 studies found that nearly a third of 155 EPA-registered labels stated residual toxicity inconsistently with the calculated value. Timing guidance is only as good as the residue data behind it.

Practical rules:

• Do not apply to open bloom.
• Apply at dusk, after dark, or in early morning before foraging begins.
• Avoid application when temperatures will keep bees flying late.
• Check for hives within foraging range - typically up to about 3 km - and tell the beekeeper before you spray.
• Do not let spray or rinsate reach flowering weeds beneath the crop.

The adjuvant nobody reads

The active is not the only thing in the tank. Ingestion of 20 µg of an organosilicone surfactant impaired olfactory learning in honey bees in a proboscis extension reflex assay - organosilicones were more active than nonionic adjuvants, and crop oil concentrates showed no effect. “Inert” is a regulatory classification, not a biological one.

For balance: a later study found a tested organosilicone adjuvant did not increase the lethal effects of a pyrethroid or a carbamate, so the picture is about sublethal impairment rather than acute synergy.

If you tank-mix, the adjuvant is part of your pollinator risk assessment.

Biologicals are not automatically exempt

Bacillus thuringiensis is widely assumed bee-safe. A semi-field and laboratory study found Bt spores detectable in all hive matrices for the full test duration, with chronic exposure reducing adult and larval survival in a concentration-dependent way. Not a reason to avoid Bt - a reason not to treat it as a free pass.

Entomopathogenic nematodes were long considered non-susceptible: a 1995 study sprayed four species onto combs and found under 10% adult mortality with no evidence of infection on dissection. That finding should no longer be presented as settled. More recent work reports complete worker mortality in bioassays with some Steinernema species, so the honest statement today is that entomopathogenic-nematode bee safety is less certain than it was.

Trichoderma-based products showed no worker mortality and no adverse reproductive effects in a laboratory study on Bombus terrestris.

If you think bees have been harmed

Stop applications. Contact the beekeeper if hives are nearby, and report the incident to your national pesticide incident scheme - in the UK the Wildlife Incident Investigation Scheme, in the US your State Lead Agency for pesticide regulation. Retain the product label, batch number and a record of the application, because an investigation is only as useful as the record behind it.

This page is stewardship guidance, not a regulatory determination. Product registrations, permitted uses and bee-protection label statements vary by jurisdiction, and the label on the product you hold is the governing document.

Ratings for the actives in this range

Ratings reflect what is known, and “insufficient data” is stated as such rather than defaulted to safe.

Which compounds are rated, and how

Ratings come from the compound records, so the same vocabulary appears here and on each compound page.