Presets
Playback
The hunter and the moths
Moth numbers and size, the bird's flight and handling time.
The night sky
Moon phase, height in the sky and cloud.
Camera
Quality
Layers
Model checks
The science
How far can a nightjar see a moth by moonlight?
A nightjar hunts by eye, watching moths as dark or pale shapes against the sky. Vision in dim light is limited by photon counting (the Rose-de Vries law): a target of angular size d/r and contrast C is just visible when (d/r) C √I / k reaches one, where I is the light level and k is the observer's threshold. Haze and vegetation dim the signal with distance, e−r/L. Setting the product to one gives r er/L = A with A = d C √I / k, solved by the Lambert W function: rd = L · W(A/L). The page also solves the same equation by bisection and compares the two (check 3).
The light level is starlight (about 2 mlux) plus the moon: 250 mlux × (lit fraction)³ × sin(altitude) × (1 − 0.75 cloud). The cube on the phase is a rough fit to the steep drop in moonlight between full and quarter moon; clouds also dim the stars. A 30 mm moth is visible out to about 9 m under a high full moon and barely 1 m by starlight alone, so the same meadow is a very different hunting ground from one night to the next.
Holling's type II functional response
The nightjar flies at speed v and sees every moth that comes inside a circle of radius rd. Moths that fly at speed u in random directions cross that circle at a rate a N with a = 2 rd 〈|v − u|〉 (the swept width times the mean relative speed). Each moth then takes a handling time h (a pursuit, a snap, a swallow) during which no new moth is noticed. A hunting cycle is a search of mean length 1/(aN) followed by h, so the intake is f = aN / (1 + aNh): it rises almost linearly while moths are rare and levels off at 1/h when they are plentiful. The picture of the bird is a time budget: the share of the night spent handling is f h.
Check 1 and check 2 run an independent agent simulation of the same hunt (hundreds of moths taken, several random seeds) and compare it with the formula at the current settings and at four times the moth density. Check 2 also verifies that search time plus handling time equals the flight time and that handling time equals moths taken times h.
Honest simplifications: the bird flies a gently wandering path at constant speed over a flat 80 m arena that wraps around at the edges, so the model has no depletion of the meadow (Holling's assumption: a taken moth is replaced in place by a new one a few seconds later, which keeps the moth density constant); moths fly at one speed in random directions, never dodge, and their number does not depend on the moon (in real life many moths are less active around a bright full moon, which would partly offset the better light); detection is a simple circle and only vision is included (no hearing, no wind); handling time is fixed; the visibility constants (a 30 mm moth reaching about 14 m before extinction at the full-moon zenith) are plausible values chosen for illustration, not measured nightjar data; there is no energy budget. The nightjar is drawn 3.2 times and the moths 22 times life size so they can be seen, and the screen brightness is compressed because the eye adapts to the dark. The aim is to show why light and prey density together set how well an insect-eating bird can feed, not to predict a real bird.
What to try
- Press New moon (starlight only): the detection cylinder shrinks to about a metre and the intake drops several times, even though the moth density has not changed.
- Use Moth swarm (150 per ha) and watch the share of time spent handling climb toward 100 percent: the intake bends toward the 1/h ceiling in the first graph.
- Drag the Moon age slider and watch the marker move along the lunar-month graph: only the few days around full moon give a long detection range, and the weeks around new moon are dark.
- Pick Moth level or Bird close-up to see which moths are lit up as visible inside the cylinder.