The Trojan Horse: Horizontal Transfer, Tertiary Mortality, and Why Slow Is the Point

1. Introduction: a mechanism used daily and rarely explained

Gel bait is the standard approach to German cockroaches in occupied buildings, and for good reasons set out elsewhere in this journal: it avoids the dispersal that repellent chemistry causes, and it reaches harbourage no application can.

The second of those claims deserves closer examination, because a gram of gel on a cabinet hinge does not reach anything. What reaches the harbourage is a cockroach.

Horizontal transfer occurs when insects contact or ingest an insecticide, return to an aggregation or nest, and transfer the insecticide to other conspecific insects through contact.² The phenomenon occurs when the most active members of a colony, often foraging adults, become exposed to an insecticide residue which is subsequently transferred to unexposed members upon returning to the nest.²

2. The original demonstration

The finding is now foundational but it had to be proven, and the method was elegant.

Secondary mortality was first reported in the German cockroach, Blattella germanica, using radiolabelled hydramethylnon.¹

2.1 Why radiolabelling mattered

Observing that untreated cockroaches die near treated ones establishes correlation. Tagging the active ingredient with a radioisotope and then finding the label inside insects that never visited the bait establishes the physical movement of the molecule.

That distinction is what turned an operational impression into a mechanism, and it is why the 1991 work anchors everything that follows.

3. The four routes

Four mechanisms have been shown to facilitate horizontal transfer of insecticides in the German cockroach: contact, coprophagy, emetophagy, and necrophagy and cannibalism.¹

Related literature describes the same set, noting that mortality can occur in cockroaches not only through direct interaction with the bait but also through necrophagy, coprophagy, emetophagy and physical distribution of bait in the area.⁵ Mutual grooming and trophallaxis appear across the wider horizontal transfer literature.²

3.1 Taking each in turn

Coprophagy is ingestion of faeces from insects that consumed bait.³ Undigested active ingredient passes through the gut and remains biologically available in the frass.

Emetophagy is ingestion of regurgitated, bait-laden material.¹

Necrophagy and cannibalism involve feeding on poisoned or dead nestmates. This was shown to play an important role when cockroaches that consumed fipronil-containing baits were subsequently wholly or partially consumed by third instars, inducing substantial mortality in them.²

Contact transfer occurs during grooming and ordinary physical interaction within the aggregation.¹

4. Why first instars are the target

The reason coprophagy matters most is a piece of cockroach natural history that determines the whole strategy.

Coprophagy is especially important in targeting first instars, which represent a significant proportion of the total population under normal conditions yet may be the most difficult to reach with baits, because they do not forage independently, spend the majority of their time within the shelter, and rely on adult faeces for nutrition.³

Adult cockroaches that feed on bait directly can therefore be used to deliver it to more sedentary stages such as young nymphs.³ The paper describing this calls the foraging adults Trojan horses,¹ which is exactly right.

4.1 Why this defeats surface treatment

A residual application treats surfaces that foraging adults cross. First instars do not cross them, because they do not forage. A treatment that reaches only foragers leaves the sedentary majority untouched, which is a substantial part of why spraying underperforms against this species even when resistance is not involved.

5. Tertiary mortality

The most striking single experiment in this literature followed the toxicant through three levels of a colony.

Only one study has reported horizontal transfer beyond secondary mortality. Researchers offered bait containing 0.6 per cent indoxacarb to an adult male, quantified secondary mortality of first instars exposed to that male, and then offered the dead nymphs to starved adult males. Eighty one per cent of those males died, largely from eating the dead nymphs.¹

5.1 What three levels means

One bait-feeding event killed an adult, that adult killed nymphs that never touched the bait, and those nymphs killed further adults that never touched the bait or the original donor. The active ingredient remained lethal after passing through two insects.

This is why baiting can suppress an aggregation that a treatment physically reached only at its margins.

5.2 The appropriate caution

It is a single laboratory study under controlled conditions with starved recipients, and the literature itself notes that only one study has reported transfer at this level.¹ Starvation makes necrophagy more likely than it might be in a well-provisioned kitchen. The result demonstrates that tertiary transfer is possible, not that it is routine.

6. Why delay is the mechanism

We can now state the design principle that the rest of the paper depends on.

Every transfer route in §3 requires the exposed insect to do something after exposure. Coprophagy requires it to return to harbourage and defecate there. Emetophagy requires it to regurgitate among nestmates. Necrophagy requires the carcass to be where others will find it. Contact requires interaction.

An insect killed at the bait station does none of these. It dies in the open, away from the aggregation, having delivered a dose to exactly one individual.

6.1 The commercial tension

This is awkward for an industry that sells visible results. A product producing rapid knockdown looks effective and demonstrates less. A product producing no visible effect for a day looks like nothing is happening and is doing the work.

Client expectation is calibrated by consumer aerosols that kill on contact, and the professional approach is deliberately the opposite. Managing that expectation is part of deploying the method, in the same way it is for thermal treatment as discussed elsewhere in this journal.

6.2 The qualification

Delay is not unlimited. Coprophagy is described as most effective with slow-acting insecticides,³ but a compound too slow to kill before the population reproduces is not controlling anything. The optimum is long enough to get home, short enough to matter, and §7 shows the industry has products at different points on that curve.

7. The active ingredient determines the route

Here is the finding we think is least known among practitioners and most useful.

The transfer pathway is not a fixed property of cockroaches. It depends on the speed of the active.

In studies of fipronil, a fast-acting neurotoxic insecticide, fipronil excreted from bait-fed B. germanica caused secondary mortality in untreated cockroaches in the same cage. The presence of dying foragers, and contact between them and untreated cockroaches, facilitated distribution and caused high mortality. Their faeces, on the other hand, appeared to play a minor role in fipronil transfer, in contrast to hydramethylnon, a delayed-action insecticide that is transferred by coprophagy.⁶

7.1 Why the difference arises

A delayed-action compound gives the insect time to return, feed normally and defecate repeatedly in the harbourage, loading the frass over hours or days. A fast-acting compound kills before much defecation occurs, so the dominant route becomes the dying insect itself, through contact and subsequent necrophagy.

7.2 The practical consequence

Because first instars are reached principally through faeces,³ a product whose transfer runs mainly through contact with dying adults may reach the sedentary nymphs less efficiently than one transferred by coprophagy.

That is a real distinction between products that share a label claim. We are not in a position to rank commercial baits on it, and we would not do so in a journal published by a company that buys them, but a practitioner choosing between actives should know the variable exists.

8. The oral secretion finding

A fifth route emerged from work combining radiotracers with video observation.

Fipronil was the first major insecticide effective both by ingestion and by contact, which prompted evaluation of topical, residual and oral introduction pathways and of transfer from treated donor to untreated recipient cockroaches by coprophagy, trophallaxis and necrophagy.³

That work identified a novel mechanism of translocation involving oral secretions. Through a combination of radiotracer techniques and time-lapse video analysis, small nymphs were shown to be fatally attracted to the oral secretions of cockroaches that had ingested fipronil.³

8.1 Why this is an unsettling result

The nymphs were not incidentally exposed. They were attracted. Something about the secretions of a dosed cockroach draws the life stage least able to survive the dose.

This fit with earlier observations that first instars behave differently from other life stages.³

Whether the attraction is to the secretion itself or to something the intoxicated insect produces is not something we can resolve from the available material, and we flag it as a question rather than asserting an answer.

9. Comparative bait performance

Recent work provides comparative figures across commercial products, which is unusual and useful.

A study examined horizontal transfer potential of a 0.1 per cent isocycloseram gel on a susceptible German cockroach strain, compared against commercial baits containing 0.6 per cent indoxacarb, 0.05 per cent fipronil, 0.5 per cent dinotefuran, or 0.5 per cent clothianidin with 0.5 per cent pyriproxyfen.⁴

9.1 Coprophagy and emetophagy

Adult males were given access to bait in a choice test against dog food as an alternative food source, which killed 85 to 100 per cent of cockroaches within 24 hours. Bait and carcasses were then removed and new groups of adult males were placed in arenas containing only the excretions of the previous males.⁴

That secondary exposure killed 93 to 100 per cent of cockroaches within seven days, with mean survival time of 1.6 days for 1 per cent isocycloseram and 1.0 to 3.5 days for the other baits.⁴

9.2 Necrophagy

To examine necrophagy, adult males were exposed to bait for 24 hours, which killed 92.5 to 97.5 per cent of cockroaches across all baits.⁴

9.3 What the numbers establish

Cockroaches exposed to nothing but the excretions left behind by bait-fed insects died at rates of 93 to 100 per cent.⁴ The excretions alone were nearly as lethal as the bait.

The caveat is important and we state it plainly: this was a susceptible strain.⁴ As set out elsewhere in this journal, field populations frequently are not susceptible, and where glucose aversion is present the primary feeding event that starts the whole cascade may not occur at all.

10. Ants use a different pathway

A brief comparison, because the same logic applies with different biology.

Similar toxicant distribution mechanisms occur in ants, where workers translocate toxicants mainly through trophallaxis and necrophoresis.⁵

Trophallaxis is mouth-to-mouth food sharing and is a routine, high-frequency social behaviour in ants rather than an opportunistic one. Necrophoresis is the removal of corpses from the nest, which is a sanitation behaviour that brings workers into contact with dead nestmates.

10.1 Why this matters for pharaoh ants

As set out in the healthcare article published elsewhere in this journal, pharaoh ant colonies carry hundreds of queens in inaccessible nests, and trophallaxis is the mechanism by which bait reaches queens no application could contact.

The global distributions of pharaoh ants and German cockroaches overlap, with infestation hot spots reported in the eastern United States, Europe and the tropics.⁵ These are the two species for which the Trojan horse approach is most necessary, and in both cases for the same underlying reason: the individuals that matter never leave the nest.

11. What defeats the cascade

Understanding the mechanism makes its failure modes legible.

Primary feeding never occurs. Glucose aversion, discussed at length elsewhere in this journal, prevents the initiating event. No donor, no cascade.

Faeces are rejected. Also documented in averse populations, which removes the coprophagic route specifically and therefore the route to first instars.

The insect dies too fast. Whether through an overly fast active or through contamination of bait with a repellent or fast-acting spray, a donor that dies at the station delivers nothing.

Repellent contamination. Aerosol applied near a bait placement can render it aversive, and cleaning products can do the same.

Bait is removed. Placements wiped away during cleaning end availability before the cascade completes.

Resistance to the active. Secondary doses are by definition smaller than the primary dose, so a population with partial resistance may survive transfer even where the primary feeding kills.

12. Practical implications

Placement should target foragers, not nymphs. The foragers are the delivery system. Placing bait where adults travel is correct even though the nymphs are the larger problem.³

Undisturbed harbourage is a feature. The cascade happens inside the aggregation. Aggressive disturbance of harbourage during a baiting programme works against the mechanism.

Do not mix repellent chemistry into a baiting programme. This is the same conclusion the resistance and healthcare articles reach by different routes.

Judge progress on population trend, not on speed. Adults and subadults die first, so early visible mortality is expected while the nymph suppression that determines the outcome happens out of sight over subsequent weeks.

Leave bait available. The cascade in §9 operated over seven days,⁴ and the colony-level effect takes longer.

13. Limitations and open questions

Most of this is laboratory work. The transfer studies are arena experiments with defined groups.¹⁴ Field aggregations are larger, better provisioned and more heterogeneous.

The tertiary result is a single study. The literature itself says only one study has reported transfer beyond secondary mortality.¹

Susceptible strains. The comparative bait work used a susceptible strain,⁴ and generalisation to resistant field populations is not warranted.

We have not ranked products. The comparative study reports ranges across several commercial baits.⁴ We have reported the ranges and declined to make recommendations, since we purchase these products and are not an appropriate source for that comparison.

Some sourcing is secondary. Several primary papers, including Silverman et al. 1991, Kopanic and Schal, Buczkowski et al. and le Patourel, are cited here through reviews rather than read in full.¹²

The oral secretion mechanism is reported through an interview account. The description of the radiotracer and video work comes from a trade publication quoting the supervising researcher.³

14. Conclusion

Bait works because cockroaches carry it home. Secondary mortality was demonstrated with radiolabelled hydramethylnon,¹ and four routes have since been documented: coprophagy, emetophagy, necrophagy and cannibalism, and contact,¹ with a fifth involving oral secretions to which small nymphs are fatally attracted.³ Indoxacarb has been carried through three levels of a colony, killing 81 per cent of adults that ate nymphs killed by an adult that ate the bait.¹ Cockroaches exposed only to the excretions of bait-fed insects died at 93 to 100 per cent within a week.⁴

The route depends on the active. Delayed-action compounds move principally through faeces, while fast-acting fipronil moves through dying foragers and contact with faeces playing only a minor role.⁶ Since first instars are reached mainly through faeces,³ that difference determines which part of the population a given product actually reaches.

All of which rests on a single requirement: the insect has to survive long enough to get home. Delayed action is not something bait chemistry tolerates. It is the property that turns a feeding event by one insect into a population event, and it is why the most effective thing a bait can do in its first hour is nothing at all.

References

1. Cockroaches as Trojan Horses for Control of Cockroach Aggregations With Baits. Journal of Economic Entomology, 116(2), 529. Source for the first report of secondary mortality in Blattella germanica using radiolabelled hydramethylnon (Silverman et al. 1991), the four documented transfer mechanisms of contact, coprophagy, emetophagy, and necrophagy and cannibalism with their primary citations, and the tertiary mortality experiment in which 0.6 per cent indoxacarb bait fed to an adult male produced secondary mortality in first instars whose carcasses then killed 81 per cent of starved adult males (Buczkowski et al. 2008). https://academic.oup.com/jee/article/116/2/529/7025293
2. Transfer of Ingested Insecticides Among Cockroaches: Effects of Active Ingredient, Bait Formulation, and Assay Procedures. Source for the general definition of horizontal transfer, the description of foraging adults as the exposed members transferring insecticide on return to the nest, the listing of mutual grooming, trophallaxis, necrophagy and coprophagy as mechanisms, and le Patourel's observation that cockroaches consuming fipronil baits were consumed by third instars with substantial resulting mortality. https://www.researchgate.net/publication/11670690_Transfer_of_Ingested_Insecticides_Among_Cockroaches_Effects_of_Active_Ingredient_Bait_Formulation_and_Assay_Procedures
3. Tertiary Mortality in the German Cockroach. Pest Control Technology. Trade publication reporting work from Coby Schal's laboratory. Source for the account of coprophagy, necrophagy and emetophagy as transfer mechanisms, the explanation of why first instars are difficult to reach with bait yet rely on adult faeces for nutrition, the statement that coprophagy is most effective with slow-acting insecticides, fipronil as the first major insecticide effective by both ingestion and contact, and the identification of oral secretion translocation with small nymphs fatally attracted to secretions of fipronil-dosed cockroaches using radiotracer and time-lapse video methods. https://www.pctonline.com/article/-cockroach-control--tertiary-mortality-in-the-german-cockroach/
4. Horizontal Transfer of a Novel Isocycloseram Bait on German Cockroaches (Blattodea: Ectobiidae). International Conference on Urban Pests, 2025. Source for the comparison of 0.1 per cent isocycloseram gel against commercial baits containing 0.6 per cent indoxacarb, 0.05 per cent fipronil, 0.5 per cent dinotefuran and 0.5 per cent clothianidin with 0.5 per cent pyriproxyfen; the choice test killing 85 to 100 per cent within 24 hours; secondary exposure to excretions killing 93 to 100 per cent within seven days with mean survival times of 1.0 to 3.5 days; and the necrophagy exposure killing 92.5 to 97.5 per cent across all baits, all on a susceptible strain. https://www.icup.org.uk/conferences/2025/papers/horizontal-transfer-of-a-novel-isocycloseram-bait-on-german-cockroaches-blattodea-ectobiidae/
5. Toxicant translocation and colony impact in the Pharaoh ant. Source for the summary of cockroach transfer through necrophagy, coprophagy, emetophagy and physical distribution of bait, for the description of ant toxicant translocation principally through trophallaxis and necrophoresis citing Choe and Rust 2008, and for the overlapping global distributions of pharaoh ants and German cockroaches with hot spots in the eastern United States, Europe and the tropics. http://www.chowyang.com/uploads/2/4/3/5/24359966/309.pdf
6. Buczkowski, G. & Schal, C. Horizontal transmission of fipronil in the German cockroach. Purdue University. Source for the finding that fipronil excreted from bait-fed B. germanica caused secondary mortality in untreated cockroaches, that the presence of dying foragers and contact with untreated cockroaches facilitated distribution, and that faeces played only a minor role in fipronil transfer in contrast to hydramethylnon which is transferred by coprophagy. https://www.entm.purdue.edu/ants/pubs/1.pdf