Shared pesticide susceptibility

Why does every farmer spraying the correct dose leave the whole district with a chemical that no longer works?

Take a district where every grower does everything right. Nobody over-sprays, nobody under-doses, everyone follows the label and the local advisory to the letter. Eight seasons later the product has stopped working for all of them.

We usually reach for blame here — somebody must have cut the rate, or sprayed off-label. But suppose nobody did. Suppose the loss is the arithmetic consequence of everyone behaving correctly. What kind of thing would have to be at stake for correct individual behaviour to produce a collective loss?

Let us be precise about what is actually being consumed. When you spray, you buy dead pests on your own field. What you spend is something less obvious: the fraction of the local pest population that is still killable by that chemistry. Call that fraction susceptibility.

Now ask two questions about it, because the answers are what make this a hard problem rather than a merely sad one.

First: who owns it? Susceptibility is not a property of your field. It is a property of the pest population — specifically of the interbreeding population, which for a mobile moth or beetle spans a whole district, sometimes a region. Your spray does not deplete your own stock. It draws down everyone's.

Second: does using it destroy it? Here is the uncomfortable step. A spray does not merely remove insects; it removes them selectively. Resistance alleles are usually present at very low frequency before any chemical arrives — estimates commonly sit somewhere around one in a thousand to one in a million individuals. The spray kills the susceptible and spares the rare resistant, and the survivors are the ones that breed the next generation. Each application does not subtract from susceptibility so much as multiply the resistant share.

Do you see what follows? The correct dose is not a mitigation of this. The correct dose is precisely the thing that makes the selection sharp. A weak dose leaves survivors of every kind and selects weakly; a label-rate dose leaves only the genuinely resistant. The recommendation is optimised for the outcome the farmer is paying for — dead pests this week — and is silent about the outcome nobody is paying for.

So we have the shape now. The benefit of a spray is private, immediate and certain. The cost is spread across every grower who shares that pest population, delayed by years, and uncertain. If you unilaterally forgo a spray, you take a yield loss and the resistance is delayed by a negligible amount, because your neighbours are still selecting. Restraint by one is a gift with no recipient. That is a collective-action problem in its purest form, and it does not dissolve if everybody is virtuous, only if everybody is coordinated.

Which brings us to the interesting part: what would a coordinated solution even look like? Not simply spraying less — that only slows the clock. The clever answer, and the one that actually got mandated, works genetically rather than economically.

Suppose the resistant survivors, instead of mating with each other, almost always mate with a susceptible partner. Their offspring carry one resistant copy and one susceptible copy — and if the dose is high enough to kill those heterozygotes, resistance is effectively erased each generation. To make that true you need a large, reliable supply of untreated pests. So you deliberately grow some of the crop unprotected: a refuge.

That is the high-dose/refuge strategy that the US EPA required alongside Bt crops — structured refuges of unprotected acreage, in the range of roughly five to twenty per cent depending on the crop and whether the plant stacks more than one toxin. It is a strange thing to ask of a farmer. You are asking them to farm a slice of land badly, on purpose, so that a resource nobody can invoice for keeps existing.

And when compliance held, it worked. Pink bollworm was driven to eradication in the south-western United States by Bt cotton combined with refuges and sterile insect releases, declared complete in 2018. Where refuge planting was widely ignored, in Bt cotton in parts of India, the same pest evolved resistance to the same toxins within about a decade. Same technology, same pest, opposite outcome — the variable was collective compliance.

One honest caveat: susceptibility is not perfectly non-renewable. Resistance alleles often carry a fitness cost, so frequencies can fall when a chemical is withdrawn. But reversion is slow, and pests sometimes evolve modifier genes that cancel the cost, at which point the loss really is permanent.

The instinct to prescribe "spray less" is not wrong, but it misidentifies the resource. Rotating modes of action, using thresholds instead of calendars, and mixing in biological controls all extend susceptibility — but each is still a decision made on one farm about a stock held in common, so each carries the same free-rider problem. Something has to make restraint collective: a mandate, a licence condition, or an area-wide programme.

It is also worth noticing that the correct scale of governance is set by the pest, not by the map. For an insect that moves a few metres, the resource is nearly private and a single farm can manage it alone. For one that flies fifty kilometres, no farm, cooperative, or even county is the right unit. Getting the boundary wrong is how well-designed schemes fail.

• How resistance monitoring actually detects a rising allele frequency early enough to change practice.
• Why insecticide mode-of-action rotation schemes assume independence between classes, and what cross-resistance does to that assumption.
• The parallel case in human medicine, where antibiotic susceptibility is the same kind of shared stock with a very different governance structure.

Every term the collection defines is gathered in the glossary.