Scientists Discover Moth Wings Can Smell, And They’re Laser-Focused on Just Two Scents

A tobacco hawkmoth hindwing, snipped off at the base and trimmed to fit, sits between two recording electrodes in a smear of conductive gel. Filtered air runs across its upper surface through an aluminum tube, and every so often a 200 ms puff of something scented joins the stream. The wing has been separated from the moth, from its antennae, from every neuron in its head. It answers anyway, with a small electrical deflection that shows up on the trace.

Antennae are the insect nose. That has been the working assumption for the best part of a century, but wings have turned out to be busier organs than anyone reckoned, feeling airflow, registering vibration and, in vinegar flies at least, tasting the cuticle of a passing mate.

What separates a nose from a finger, in insect terms, comes down to holes. A mechanosensory bristle sits in a raised socket and is otherwise sealed; a taste bristle carries a single pore near its tip, so a molecule has to make contact to register; an olfactory sensillum is perforated along its wall, letting airborne molecules diffuse through the cuticle to the neuron waiting inside. Ahmed Reda Ismaieel and colleagues at the Max Planck Institute for Chemical Ecology in Jena wiped the scales off Manduca sexta wings, cut away the margins, sputtered them with gold and put them under a scanning electron microscope. Two kinds of hair alternate along the edge: slender ridged scales roughly 120 micrometers long, and shorter, stouter bristles averaging 80, somewhere between 21 and 32 of them per margin (counts from five animals of each sex). The bristles had the subapical pore you would expect on a taste organ. They also had pores scattered up the wall, which nobody had described on a butterfly or moth wing before.

Morphology is suggestive. It’s not, on its own, a response.

Thirteen odors, two hits

So Ismaieel wired the wings up and ran them, using a wing version of the electroantennography that olfaction labs have leaned on for decades, against 13 synthetic odorants and the collected headspace of three plants. Pyrrole, pyridine, nicotine, trimethylamine, putrescine, linalool, a fruity ester, two acids, a Datura flower sealed in a bag: nothing. Two compounds, and only two, pulled a response out of the tissue, both of them saturated cyclic amines that smell to us like something going off at the back of the fridge, and both of them concentration dependent down to a threshold at a dilution of 1:1000.

The odorants were pyrrolidine and piperidine. Sonja Bisch-Knaden, who supervised the work with Bill Hansson, had a reason to be looking at receptors in the first place: “The tobacco hawkmoth was recently found to have receptor proteins in its wings that could detect different smells and tastes”, she says, pointing back at a 2022 survey of chemosensory gene expression across the moth’s body.

The expression work in this study says something slightly unusual about how any of that would function. Using a NanoString probe set against the moth’s chemosensory receptor families, the team detected 15 receptor genes expressed in the wing margins, about half the number the earlier survey picked up in whole wings. Eleven of the 15 were odorant receptors, the family that handles volatiles in most insects. But ORCo, the co-receptor without which those receptors do not assemble into a working channel in higher insects, was absent. By NanoString and again by the more sensitive RT-PCR. What was present were IR76b and IR25a, the ionotropic co-receptors that in vinegar flies partner with amine-sensing receptors, which is a tidy match for a wing that answers to amines and nothing else.

To narrow down which receptor, Ismaieel pushed the wing-expressed proteins through AlphaFold2, kept the ten models scoring at least 80 on the confidence metric, and docked both amines into their predicted binding pockets. Most of the fits came back weak and inconsistent, the computational equivalent of a shrug. Two did not: MsexIR7d.2 and MsexIR7d.3, members of a moth and butterfly specific receptor clade nobody has functionally characterized, each holding both molecules with a hydrogen bond to the same acidic residue (Asp211 and Asp218 respectively) and three further shared contacts apiece.

The edges were not the point

Then a complication. Ismaieel cut the margins off the hindwings, taking every one of those wall-pore bristles with them, recorded again, and found the response to both amines unchanged. Which means the sensors are not confined to the edge. So the team went hunting for them across the wing surface and came up empty, a null result the study’s press release leaves out, while carrying Bisch-Knaden’s inference that the moth has sensory hairs across its wings. Their best explanation is a size problem: hawkmoth wing scales run roughly twice the length of a margin bristle and could be hiding smaller sensilla underneath.

The docking, meanwhile, is a hypothesis generator rather than a result, and the authors say so. Pinning either candidate to pyrrolidine would mean expressing the genes in something like Xenopus oocytes and testing them directly, which has not been done. Nor is there behavioral evidence that a wing smelling an amine changes what the moth does (a gap the paper states and the release rather glides past) closing instead on the line that the wings definitely smell and could steer them toward egg-laying plants. There’s a counting difference too: the release credits the team with identifying 33 taste and scent receptor mRNAs in the wings, a figure that appears to combine this study’s 15 margin genes with the whole-wing set published back in 2022.

Why amines, though? Pyrrolidine and piperidine are the chemical scaffolding of the alkaloids that nightshades build, and nightshades are what Manduca sexta caterpillars eat, so a wing tuned to those two molecules and deaf to everything else is at least pointed in a sensible direction. Undamaged leaves do not release them, and the wings ignored the headspace of intact plants, which leaves the researchers reaching for a scenario in which a hovering female scrapes the leaf surface with the spines on her tarsi and pulls the smell out of it. An animal that tastes with its wings, smells with a grooming organ on its foreleg and now, it seems, smells with its wings too starts to look less like an insect with a nose than an insect that is more or less all nose, sampling the world from every surface it happens to own.

Frequently Asked Questions