A Bitter Sweetness: Behavioural Resistance and the Rewiring of Taste in Blattella germanica

1. Introduction: a different kind of resistance

Almost everything written about insecticide resistance describes the same basic event. A toxicant enters an insect and fails to kill it, because the insect metabolises it, excludes it, or carries a target site the molecule no longer binds well. The bed bug literature reviewed elsewhere in this journal is a detailed account of exactly that, layered four deep.

Glucose aversion in the German cockroach is a categorically different phenomenon, and the difference matters more than it might first appear. Glucose-averse cockroaches reject foods containing glucose even at relatively low concentrations, which protects them from ingesting lethal amounts of toxic bait.¹ The toxicant was never the obstacle. The insect simply stopped eating.

Put another way: a physiologically resistant insect wins the fight. A behaviourally resistant insect declines it.

1.1 What this paper argues

Our thesis is that glucose aversion is not a curiosity at the edge of resistance management but a template for a class of problem the industry is poorly equipped to detect. We develop that in four moves: the mechanism, the cascading consequences, the fitness cost that makes the trait unusual, and the diagnostic blind spot that lets it spread unobserved.

2. Why the German cockroach was primed for this

Before the mechanism, the population biology. Several features of this species favour rapid resistance evolution generally, and they are worth stating because they also explain why the trait spreads as fast as it does once it appears.

Populations of B. germanica have evolved multiple resistance mechanisms to a wide array of insecticides with multiple modes of action.² The conditions favouring this are structural. Populations are limited to indoor environments and are therefore relatively isolated, with limited gene flow between them. The species is flightless, which further limits gene flow while favouring inbreeding. As a result, when resistance mechanisms appear in a population there is little opportunity for dilution by susceptible alleles arriving from nearby.²

Add the general risk factors: frequent use of a single active ingredient or mode of action, combined with short generation time and high fecundity, heightens the risk of resistance evolution.² The German cockroach has all of these.

2.1 The obligate indoor association

One detail deserves emphasis because it connects directly to the housing analysis published elsewhere in this journal. B. germanica has a global distribution but is obligately associated with human-built structures.³ There is no wild reservoir. There is no external population of susceptible individuals drifting in to dilute an allele that selection has favoured indoors.

The consequence is that each building, or each closely connected group of units, functions as a semi-isolated evolutionary arena. Selection applied inside it is not buffered from outside. This is the same structural fact that makes attached housing difficult for bed bug management, viewed from the population-genetics side rather than the dispersal side.

3. Discovery and genetics

3.1 Field origin

Glucose aversion was first described by Silverman and Bieman in 1993, working with a field-derived strain that rejected glucose-containing bait.⁴ The observation came out of practice rather than theory. Baits were failing in ways that resistance bioassays did not explain.

3.2 The inheritance pattern

The genetics turned out to be simple and, from a management perspective, unfavourable. Glucose aversion is a partially dominant autosomal trait.¹ Once it appears in a population it readily spreads under continued selection pressure, with glucose-averse cockroaches displacing wild-type individuals.¹

Partial dominance matters. A recessive trait hides in heterozygotes and spreads slowly, giving a control programme time to notice something is wrong. A partially dominant trait expresses in heterozygotes, which means selection acts on it immediately and the frequency climbs from the first generation of bait exposure.

3.3 Speed

The rate of population replacement under selection is startling. Work by Wada-Katsumata and colleagues documented replacement of wild-type by glucose-averse cockroaches after exposure to glucose-containing toxic bait over a period of five days.⁵

Five days is shorter than the interval between a first bait application and the follow-up visit in most commercial programmes. A technician applying bait on Monday and returning the following week may be returning to a population whose composition has already shifted.

4. The neural mechanism

The reason this case is so well characterised is that the mechanism was traced to individual sensory neurons, which is unusual in applied entomology.

4.1 The sensory apparatus

Taste in this species is distributed across four sensory appendages: antennae, maxillary palps, labial palps and paraglossae, each bearing sensilla that contain gustatory receptor neurons (GRNs).⁶ Glucose aversion is a taste polymorphism based on changes in gustatory neurons within these sensilla, especially the paraglossae.¹⁷

4.2 The inversion

Here is the finding that makes this case remarkable. In wild-type cockroaches, glucose stimulates appetitive, sweet-responding GRNs within these sensilla, driving appetitive behaviour.¹⁸ In glucose-averse cockroaches, glucose also stimulates gustatory receptor neurons that normally respond to bitter compounds, driving aversive behaviour.¹⁸

Note the word also. The sweet response was not abolished and replaced. A deterrent channel was recruited alongside it, and the aversive signal dominates the behavioural output. The insect is not indifferent to glucose. It is repelled by it.

4.1b What a sensillum is, briefly

A hair-like or peg-like projection of the cuticle housing the dendrites of gustatory neurons, open at the tip to the surrounding fluid. An insect tasting a surface samples it through pores in its own exoskeleton, which is why contact chemoreception can be measured by recording from a single sensillum while a test solution bathes it.

That description is ours, supplied because the finding below is a recording from exactly such a preparation and the method is not obvious to a reader outside the field. It also explains why the location of the change matters: detection occurs at first contact, by touch, before anything is swallowed.

4.2b The primary measurement, stated precisely

The foundational experiment recorded from the neurons directly, and its result contains a control that is easy to read past. In both wild-type and glucose-averse (GA) cockroaches, D-fructose and D-glucose stimulated sugar-gustatory receptor neurons (GRNs), whereas the deterrent caffeine stimulated bitter-GRNs. In contrast, in GA cockroaches, D-glucose also stimulated bitter-GRNs and suppressed the responses of sugar-GRNs.¹²

Fructose behaves normally in averse animals and caffeine behaves normally in both. The change is specific to one molecule, which is what makes this a taste polymorphism rather than a general sensory defect, and the distinction matters because a degraded sensory system would carry costs across the whole diet rather than at one sugar.¹²

4.2c Two changes, not one

The same sentence reports an aversive signal added and an appetitive signal removed: glucose also stimulated bitter neurons, and suppressed the responses of the sugar neurons.¹² Either alone would reduce feeding; together they produce the reversal in modal quality described above. The authors state the result as a dual processing, with glucose functioning as both a phagostimulant and deterrent in GA cockroaches.¹²

4.2d A gain of function

Which should surprise a reader of this journal's other resistance articles. The senior author, on publication: most times, genetic changes, or mutations, cause the loss of function. In this case, the mutation resulted in the gain of a new function, triggering bitter receptors when glucose is introduced.¹³

An anticoagulant resistance allele degrades an enzyme. A sodium channel substitution degrades a binding site. A receptor substitution degrades a pocket. Each is a protein working less well at something it used to do. This one adds a response that was not there, and gain-of-function changes are rarer precisely because there are more ways of breaking a protein than of giving it a new specificity. That comparison is ours, and it is part of why the trait has proved as durable as it has.

4.3 Why the peripheral location matters

The change sits in the peripheral chemosensory system rather than in central processing. Reviews of this literature place it in a broader pattern in which changes in peripheral chemosensory systems drive adaptive shifts in food preference across insects.⁶

For control purposes the peripheral location is bad news. A centrally mediated aversion might be overcome by learning or by physiological state such as starvation. A receptor-level change is closer to hardwired. The signal arriving at the brain has already been recoded.

5. The first cascade: disaccharides and saliva

If the trait affected only glucose, formulators could simply switch sugars. The reason they cannot is a second finding that is, in its own way, more elegant than the first.

Glucose-aversion behaviour extends to more complex di- and trisaccharides that contain glucose monomers, because saliva hydrolyses these sugars and releases glucose, which then stimulates the aversion.¹ The behavioural phenotype extends to all oligosaccharides that contain glucose.³

5.1 The consequence for formulation

Sucrose is glucose bonded to fructose. Maltose is two glucose units. Both taste acceptable for a moment and then, as salivary enzymes act, become bitter in the mouth of an averse cockroach. The insect starts to feed and stops.

This closes off the obvious reformulation strategy. A bait manufacturer cannot simply substitute a different common sugar, because most of the palatable options liberate glucose on contact with saliva. Fructose remains acceptable, which is why fructose-based formulations appear in the literature as a comparator.¹

5.2 An observation about experimental design

This finding also carries a warning for anyone evaluating bait palatability. A short-duration choice test may record initial acceptance and miss the rejection that follows once hydrolysis has occurred. The temporal dimension of the response is part of the phenotype.

6. The second cascade: coprophagy and secondary mortality

Bait programmes do not rely solely on direct ingestion. A substantial part of their effect comes from horizontal transfer, in which a cockroach that has fed on bait returns to harbourage and poisons nestmates through faeces, regurgitate and, after death, its own carcass. Horizontal transfer and the resulting secondary mortality have been documented in German cockroaches, including in insecticide-resistant strains.¹

Nymphs are particularly dependent on coprophagy, which is a normal part of their nutrition. This is what makes bait efficient against a population where most individuals never encounter the bait station.

6.1 The hypothesis and the finding

McPherson and colleagues asked the obvious next question. If glucose-averse cockroaches reject glucose, and if bait-fed adults excrete glucose in their faeces, would averse nymphs also reject the faeces?¹

They hypothesised that ingestion of baits containing glucose or glucose-containing disaccharides would result in behaviourally relevant glucose levels in faeces, deterring coprophagy by averse nymphs. They fed adult females hydramethylnon baits rich in glucose, fructose, sucrose or maltose, and compared secondary mortality of averse and wild-type nymphs through coprophagy.¹

The paper's title states the result: glucose- and disaccharide-containing baits impede secondary mortality in glucose-averse German cockroaches.¹

6.2 Why this compounds rather than adds

The two effects multiply. Direct bait rejection removes the primary kill. Faecal rejection removes the secondary kill. A bait programme against an averse population loses both of its mechanisms of action at once, from a single underlying trait.

This is the part practitioners most often miss. The visible failure is that the bait is not being eaten. The invisible failure is that the transfer network the programme was counting on has also shut down.

7. The third cascade: courtship and assortative mating

The most surprising consequence has nothing to do with pest control at all, and it is the finding that elevates this from an applied result to a genuinely important piece of evolutionary biology.

7.1 The nuptial gift

During courtship, the male German cockroach offers the female a nuptial secretion, and this gift contains maltose, which expedites copulation.⁹ The female mounts the male to feed on it, and that positioning is what allows mating to proceed.

7.2 The conflict

The female's saliva rapidly hydrolyses maltose into glucose.⁹ In a glucose-averse female, this converts a courtship gift into a deterrent partway through consumption. She disengages, and copulation fails.

Researchers describe this directly as the adaptation causing females to reject courting males, because their salivary enzymes degrade the sugars in the male's nuptial gift into glucose.¹⁰ The trait is characterised as a rapid evolution of an adaptive taste polymorphism that disrupts courtship behaviour.¹⁰

7.3 Sensory conflict and its resolution

The framing offered by Wada-Katsumata, Hatano and Schal is worth quoting in substance. Human imposed selection can lead to adaptive changes in sensory traits, but rapid evolution of the sensory system can interfere with other behaviours, and animals must overcome such sensory conflicts.⁹ Their work describes a gustatory polymorphism mediating a new adaptive courtship strategy.⁹

In other words the species did not simply absorb the cost. Behavioural adjustment followed, including changes in male courtship tactics. Selection imposed by a bait formulation propagated into the sexual communication system of the species and then produced a secondary adaptation there.

7.4 The management-relevant part

Beyond its intrinsic interest, this produces assortative mating. Glucose-averse females are more likely to mate successfully with males whose courtship the female tolerates, which biases pairings toward like genotypes. Assortative mating accelerates the fixation of the trait relative to random mating, because it reduces the production of heterozygotes that would otherwise dilute it.

A control programme is therefore not merely selecting for a trait. It is selecting for a trait that has acquired its own mechanism for spreading faster.

8. The fitness cost and what it implies

Everything above describes a trait that looks unstoppable. The literature contains one important countervailing finding.

8.1 The evidence

Work on persistence of the sugar-rejecting genotype under various dietary regimes found that after six months, populations restricted to a 20% glucose food source contained a significantly higher frequency of wild-type individuals than expected. After twelve months, all dietary treatments contained significantly more wild-type individuals than expected.¹¹

8.2 Interpretation

Glucose aversion carries a cost. Refusing a major and widely available carbohydrate is not free in nutritional terms, and the courtship disruption of §7 imposes a reproductive cost as well. In the absence of bait selection, wild-type individuals recover ground.

This distinguishes behavioural resistance of this kind from target-site mutations such as kdrwhich frequently carry small enough costs that they persist for long periods after the selecting compound is withdrawn. Glucose aversion appears to be genuinely reversible on a timescale of months to a year.

8.2b And a general point the case makes about fitness costs

A trait imposing a nutritional restriction and an interference with mating still reached high frequency across the world. The reason is that the selection differential is enormous: a wild-type animal in a baited building does not reproduce at a reduced rate, it dies, and three-quarters of a normal reproductive output exceeds nothing by an unbounded margin.

Which makes almost any cost payable while selection is present. This journal's article on reversion reported resistance alleles persisting for decades after selection stopped; this case supplies the other half of the same picture. Fitness cost is a good predictor of what happens after withdrawal and a poor predictor of what happens during use, and a cost argument made without specifying the intensity of selection it is set against does not carry much. That formulation is ours.

8.3 The caveat

Laboratory populations under controlled diet are not buildings. The dietary environment of a real infestation is heterogeneous and largely outside anyone's control, and the recovery observed under experimental conditions should not be assumed to proceed at the same rate in a restaurant kitchen. The direction of the effect is supported. The magnitude in the field is not established.

9. Why this is invisible to standard diagnostics

This section states what we consider the most practically important argument in the paper.

9.1 The diagnostic mismatch

Resistance monitoring in structural pest control rests overwhelmingly on bioassays that expose insects to a known dose and measure mortality. That design assumes exposure. It measures what happens after the insect receives the toxicant.

Behavioural resistance acts entirely upstream of that. An averse cockroach placed in a forced-contact bioassay with hydramethylnon will die on schedule, because the assay removes the behaviour that constitutes the resistance. The population will be scored as fully susceptible while bait stations in the field remain untouched.

9.2 What the failure looks like in practice

The field signature is distinctive once you know to look for it: bait consumption is low or zero, populations persist, the product tests fine in the laboratory, and switching to a different bait active ingredient does not help provided the new bait uses the same sugar base. Operators frequently interpret this as a bad batch, a placement error, or competing food sources.

Competing food sources are a genuine and common cause of poor bait uptake, which is what makes the misdiagnosis so easy. The distinguishing test is simple in principle: offer a fructose-based formulation or a non-sugar matrix alongside the glucose-based one and compare uptake.

9.2b And the record will look like resistance developing

A site treated with product A that fails, then product B, then product C, reads on paper as a population that has acquired resistance to three unrelated actives.

If all three were gel baits sharing a phagostimulant base, the record is equally consistent with a single behavioural trait that was present before the first product was applied. Distinguishing the two requires knowing whether the bait was being eaten, which is an observation almost no programme records. That reading is ours.

9.3 The monitoring gap

There is, to our knowledge, no routine commercial diagnostic for glucose aversion available to a structural operator in Canada. The trait is assayed in research settings by two-choice feeding tests and by electrophysiology, neither of which is practical in the field. This is a genuine gap, and it is unusual: a well-characterised, rapidly spreading resistance mechanism with no deployable field test.

10. What behavioural resistance means for bait strategy

Several operational conclusions follow directly from the evidence above.

Rotate the matrix, not only the active ingredient. Conventional resistance management rotates modes of action. Against glucose aversion this accomplishes nothing if the carbohydrate base stays the same, because the trait acts on the base rather than the toxicant.

Treat low uptake as a diagnostic signal. Uneaten bait is data. In a population with adequate sanitation pressure and correct placement, persistently untouched bait warrants a palatability hypothesis before a placement hypothesis.

Expect the loss of secondary mortality. Because faecal transfer is impeded in averse populations,¹ the reach of a bait programme contracts to individuals that feed directly. Station density assumptions calibrated on horizontal transfer no longer hold.

Consider deliberate withdrawal. Given the fitness cost,¹¹ a planned period away from glucose-containing baits may restore susceptibility. This is speculative as field practice and we flag it as such, but the underlying biology supports the attempt.

Do not assume the trait is rare. It has been identified in multiple cockroach populations under strong and persistent selection pressure from toxic baits,⁶ and field-collected averse strains feature in the experimental literature.⁶

10.1 Which means rotating the matrix, not the active

Mode-of-action rotation guidance instructs the operator to alternate between chemical classes. For a behavioural failure that instruction is inert, because every bait built on the same phagostimulant is refused identically regardless of what is dissolved in it.

The rotation that would work is a rotation of formulation. This journal's article on fipronil argued that class membership does less predictive work than resistance-management guidance assumes; this is the same lesson approached from the formulation side, and the connection is ours.

10.1b And it removes a standard management lever

High-dose strategies in resistance management rely on the resistance allele being recessive and on untreated refuges supplying susceptible mates, so that resistant survivors mostly breed with susceptible partners and produce heterozygous offspring the dose still kills. Neither condition holds for a partially dominant trait in a closed indoor population with no refuge outside the treated space.

10.2 With a practical caution

Product labelling does not generally make the phagostimulant identifiable, and two products from different manufacturers with different actives may share a base. So the workable version of the advice is to trial a small number of physically different bait types at a failing site and observe consumption, rather than to select on any stated property of the formulation. We could not find a public source listing phagostimulant composition by product.

11. The Winnipeg case

Combining the population biology of §2 with the housing analysis published elsewhere in this journal produces a specific local picture.

11.1 Why multi-unit housing is the likely reservoir

The species is obligately indoor,³ flightless, and has limited gene flow between populations.² A Winnipeg apartment block with a persistent cockroach problem is therefore an isolated population under repeated selection, with no inflow of susceptible alleles.

Bait is the dominant professional approach against German cockroaches in residential settings, and for good reasons: it avoids the dispersal problems of repellent sprays and it reaches harbourage that no application can. It is also, by construction, a sustained and uniform selection pressure of exactly the kind that produced glucose aversion elsewhere.

11.2 The compounding with consumer products

Consumer bait products are widely available and are used heavily in the same buildings. As with the pyrethroid problem in bed bug management, the professional programme does not control the total selection pressure on the population it is managing.

11.3 What we do not know locally

We are not aware of any published survey of glucose aversion frequency in Canadian German cockroach populations, and certainly none for Manitoba. Given that the trait can be detected with a two-choice feeding assay that requires no specialised equipment beyond basic laboratory discipline, this is a tractable question. It is also, we would argue, a more useful piece of local research than another round of conventional bioassays.

11.4 And one measurement problem this raises for our own reporting

Two kinds of evidence are routinely offered for whether a bait programme is working. A susceptibility bioassay establishes whether the compound still kills. A bait uptake measurement establishes whether the material is being consumed.

This journal has criticised each of them in isolation, in different articles and for different reasons. Susceptibility testing without uptake data misses behavioural resistance entirely, for the reasons set out in section 9. Uptake data without susceptibility testing misses physiological resistance, because material can be eaten by animals it no longer kills. Neither measure is sufficient alone, and our own articles should have said so together rather than separately.

11.5 Which is a cheap fix rather than an expensive one

Placement weight before and after, or simple visual assessment of whether a bait deposit has been touched, distinguishes eaten from untouched at close to no cost. Our monitoring articles describe this as data most programmes already generate in the course of servicing and then discard. Here it is the single observation that separates two failure modes requiring opposite responses.

12. A wider lesson about anthropogenic selection

It is worth stepping back from control for a moment, because this case says something general.

Reviewers describe glucose aversion as a remarkable example of how anthropogenic selection drove the evolution of an altered gustatory trait that reshapes both foraging ecology and sexual communication.³ The chain runs from a commercial formulation decision, to a change in peripheral neuron response, to a change in what the species eats, to a change in how it mates.

Pest control is usually framed as an attempt to reduce a population. This case makes clear that it is also, unavoidably, an evolutionary intervention. We are not only killing insects. We are choosing which insects exist, and occasionally rewriting their sensory world in the process.

That framing is not an argument against control. It is an argument for humility about what control does, and for taking the evolutionary consequences of a formulation choice as seriously as its immediate efficacy.

13. Limitations and open questions

No Canadian prevalence data. Everything about frequency and distribution in this paper is inferred from populations studied elsewhere. We have found no survey of glucose aversion in Canadian or Manitoba populations.

Field reversal is unquantified. The fitness cost is demonstrated in laboratory populations under controlled diets.¹¹ Whether withdrawal of glucose baits restores susceptibility at a useful rate in an occupied building is untested, and our suggestion in §10 should be read as a hypothesis rather than a protocol.

Interaction with physiological resistance is not well characterised. Populations carrying both behavioural and metabolic or target-site resistance presumably exist, and the combined phenotype has not, to our knowledge, been systematically described.

The assortative mating argument is partly ours. The courtship disruption is documented.⁹¹⁰ The inference in §7.4 that this accelerates fixation is a reasonable population-genetic consequence but we have not located a study that models or measures it directly.

Chart values are directional. The genotype-frequency figure presents the direction and rough magnitude of the reported effect for readability. Exact frequencies vary by dietary treatment and should be taken from the source.¹¹

13.5 On the sources added to this article

The foundational 2013 paper reaches us through its published abstract and the author copy of the abstract and supplementary material, not through the full text, so the recording methods, the number of populations tested and the statistical treatment are not things we have inspected.¹² One source is a science news service, used for author quotation and framing rather than for any finding, and press coverage of research is not evidence of the research.¹³

13.6 And what remains missing across all of them

No source we consulted gives a frequency for the trait in any named field population, or a proportion of bait failures attributable to it. Worldwide, readily spreading and displacing wild-types are the terms available, and they are qualitative. That gap determines how often the practical advice in section 10 actually applies, and it is the most important thing we could not establish.

14. Conclusion

Glucose aversion is the most completely documented case of behavioural resistance in urban entomology, and it has been characterised at a level of detail that applied entomology rarely achieves: from a field bait failure in the early 1990s, to a partially dominant autosomal trait,¹⁴ to the specific finding that glucose recruits bitter-responding gustatory neurons in averse individuals.⁸

Its consequences did not stop at the bait station. Salivary hydrolysis extended the aversion to disaccharides, closing the obvious reformulation route.¹ Faecal glucose impeded the coprophagic transfer that delivers secondary mortality.¹ And maltose in the male nuptial gift carried the conflict into courtship, where it disrupted mating and prompted a further behavioural adaptation.⁹¹⁰ One change in one class of sensory neuron propagated through feeding, social transfer and reproduction.

For practitioners the operational content is compact. Bait that is not eaten is a finding, not an inconvenience. Rotating active ingredients while holding the sugar base constant does not address this failure mode. Secondary mortality cannot be assumed in an averse population. And because the trait appears to carry a real fitness cost,¹¹ the population may be recoverable in a way that target-site resistance generally is not, which is an opportunity the industry has not seriously tried to use.

The broader point is harder to act on but worth holding. A decision about which sugar to put in a bait matrix, made for reasons of cost and shelf stability, reshaped the taste perception and the courtship behaviour of a species distributed across the entire inhabited world. That is what selection pressure at industrial scale can do, and it is a reason to think carefully about the evolutionary bill attached to any control method that works well enough to be used everywhere.

References

1. McPherson, S., Wada-Katsumata, A., Silverman, J. & Schal, C. (2023). Glucose- and disaccharide-containing baits impede secondary mortality in glucose-averse German cockroaches. Journal of Economic Entomology116(2), 546–553. doi:10.1093/jee/toad030. Source for bait rejection at low glucose concentrations, partial dominance and population displacement, paraglossal sensilla, salivary hydrolysis of disaccharides, and the coprophagy experiment. https://academic.oup.com/jee/article/116/2/546/7072740
2. McPherson et al. (2023), introduction. Source for multiple resistance mechanisms across modes of action in B. germanicaand for the population features favouring resistance evolution: indoor restriction, limited gene flow, flightlessness, inbreeding, short generation time and high fecundity. https://schal-lab.cals.ncsu.edu/wp-content/uploads/sites/80/2023/08/2023-McPherson-Glucose-and-disaccharide-containing-baits-impede-secondary-mortality.pdf
3. Glucose aversion: a behavioral resistance mechanism in the German cockroach. Current Opinion in Insect Science. Source for the obligate association with human-built structures, extension of the phenotype to all glucose-containing oligosaccharides, olfactory learning-based bait avoidance, assortative mating, and the characterisation of anthropogenic selection reshaping foraging ecology and sexual communication. https://www.sciencedirect.com/science/article/abs/pii/S2214574524000245
4. Silverman, J. & Bieman, D.N. (1993). Glucose aversion in the German cockroach, Blattella germanica. Journal of Insect Physiology39, 925–933. Original description of the trait and its inheritance. See also Ross, M.H. & Silverman, J. (1995), Journal of Insect Behavior8, 825–834. https://www.nature.com/articles/srep46361
5. Wada-Katsumata et al. (2014), reproduced in Frontiers in Cellular Neuroscience: population replacement from wild-type to glucose-averse cockroaches after exposure to glucose-containing toxic bait for five days. https://frontiersin.org/articles/10.3389/fncel.2018.00281/full
6. Wada-Katsumata, A. & Schal, C. (2018). Changes in the Peripheral Chemosensory System Drive Adaptive Shifts in Food Preferences in Insects. Frontiers in Cellular Neuroscience12, 281. Source for the four sensory appendages and sensillar GRN structure, the review of peripheral gustatory mechanisms, and the observation that strong persistent bait selection has induced behavioural resistance in multiple cockroach populations including field-collected strains. https://pubmed.ncbi.nlm.nih.gov/30210303/
7. Wada-Katsumata, A., Silverman, J. & Schal, C. (2011). Differential inputs from chemosensory appendages mediate feeding responses to glucose in wild-type and glucose-averse German cockroaches, Blattella germanica. Chemical Senses36, 589–600. https://www.nature.com/articles/s42003-022-03415-8
8. Wada-Katsumata, A., Silverman, J. & Schal, C. (2013). Changes in taste neurons support the emergence of an adaptive behavior in cockroaches. Science340(6135), 972–975. doi:10.1126/science.1234854. The primary electrophysiological demonstration that glucose stimulates bitter-responding GRNs in glucose-averse cockroaches. https://blog.plantwise.org/2013/05/24/glucose-leaves-a-bitter-taste-as-cockroaches-learn-to-avoid-insecticide-baits/
9. Wada-Katsumata, A., Hatano, E. & Schal, C. (2023). Gustatory polymorphism mediates a new adaptive courtship strategy. Proceedings of the Royal Society B. Department of Entomology and Plant Pathology, North Carolina State University. Source for the maltose nuptial gift, salivary hydrolysis during courtship, and the sensory conflict framing. https://schal-lab.cals.ncsu.edu/wp-content/uploads/sites/80/2023/04/2023-Ayako-Proc-B.pdf
10. Rapid evolution of an adaptive taste polymorphism disrupts courtship behavior. Communications Biology (2022), 5, s42003-022-03415-8. Source for glucose aversion supporting greater survivorship under bait selection and for female rejection of courting males through salivary degradation of the nuptial gift. https://www.nature.com/articles/s42003-022-03415-8
11. Changes in taste neurons support the emergence of an adaptive behavior in cockroaches. The foundational 2013 paper, read as its published abstract together with the author copy of the abstract and supplementary material hosted by the corresponding author's university laboratory. Source for the statement that in response to the anthropogenic assault of toxic baits populations have rapidly evolved an adaptive behavioural aversion to glucose, described as a phagostimulant component of baits; for the electrophysiological result that in both wild-type and glucose-averse cockroaches fructose and glucose stimulated sugar gustatory receptor neurons while the deterrent caffeine stimulated bitter ones, and that in averse animals glucose also stimulated bitter neurons and suppressed the responses of sugar neurons; and for the conclusion that glucose is therefore processed as both phagostimulant and deterrent, and that this newly acquired peripheral taste sensitivity underlies the aversion in multiple averse populations. https://schal-lab.cals.ncsu.edu/wp-content/uploads/sites/80/2019/01/2013WadaKatsumata-Science-SM.pdf
12. Science news service report of the 2013 study. Press coverage of research, cited as attributed material and used for author quotation rather than for findings. Source for the senior author's observation that most genetic changes cause loss of function whereas this one resulted in the gain of a new function, triggering bitter receptors when glucose is introduced, and for his characterisation of the result as giving the cockroach a new and highly adaptive behaviour. https://www.sciencedaily.com/releases/2013/05/130523142959.htm
13. Persistence of a sugar-rejecting cockroach genotype under various dietary regimes. Scientific Reports (2017), 7, 46361. Source for the recovery of wild-type frequency after six months on a 20% glucose diet and across all dietary treatments after twelve months. See also Silverman, J. & Selbach, H. (1998), Journal of Insect Behavior11, 93–102. https://www.nature.com/articles/srep46361