Simple Summary
A termite trap today is like a doorbell that rings “someone’s here” but cannot tell you how many. That headcount is exactly what decides whether to treat and how much bait to spend. We sealed known crowds of termites, about a hundred to five hundred workers, into boxes with a wood block, then let cheap sensors catch the carbon dioxide they breathed out over 24 h. The gas climbed almost hand-in-hand with the number of termites, so a single day’s reading could estimate the colony’s size inside. A cheaper methane sensor mirrored the same ranking, and termites showed up within about an hour of sealing. The one snag: readings drifted between batches run days apart, so a field unit would need re-tuning to its own site and season before you would trust the number. Still, the study shows a trap can report “how many,” not just “who’s there.”
Abstract
Termite monitoring traps signal only presence or absence, leaving treatment decisions blind to infestation size. Termites respire CO2, so the gas accumulating inside a trap could in principle report how many workers it holds; existing work, however, relates CO2 to colony biomass rather than to termite number. Here we present a low-cost networked sensing system that records CO2, a metal-oxide (MOx) methane channel, temperature, and humidity inside sealed termite traps, and we use it to establish, to our knowledge, the first in-trap dose–response indexed to termite number rather than biomass. Across three temporally disjoint batches, each with its own blank control, eight single-chamber loads of 105–500 Coptotermes formosanus workers were tested under controlled laboratory conditions; net 24 h CO2 accumulation rose near-linearly with worker number (6.75 ± 0.95 ppm termite−1 24 h−1, R2 = 0.893) and retained leave-one-out skill (Q2 = 0.80), while a second, metal-oxide CH4 channel ordered the chambers consistently (ρ = 0.970) as a device response rather than a methane flux. We report this as a descriptive laboratory dose–response for worker number rather than as a validated abundance estimator: with one chamber per load, termite number is confounded with chamber, node and colony differences, and the calibration is not yet transferable between batches (11.9–21.4% overprediction), whereas the sensor noise floor corresponded to fewer than ten termites. In the eight laboratory chambers tested, the 59.8 ppm presence threshold was crossed within 0.16–1.45 h (10–87 min) of sealing; four-hour triage and grading abundance at the moment of detection remain potential applications that require site- and season-specific calibration and replicated field validation.