Can a largemouth bass see red? What the cone data says

The short answer

Probably yes, with limits. A 2019 laboratory study of largemouth bass (Micropterus salmoides) found a long-wavelength cone that peaks at 614 nm, which is orange-red, and bass trained to red could tell it apart from the other colours they were tested on, including ones of similar brightness. The folklore that bass “can’t see red” does not match that result.

What the study does not show is how much red matters to a bass in a real lake, at depth, in water that is not clear. That depends on how much red light is left by the time it reaches the fish, and the study is silent on that.

What the study measured

Mitchem and colleagues set out to document the photoreceptors of largemouth bass, use them to model colour perception, and test the model in a behavioural assay. According to the abstract, they found:

Cell type	Peak sensitivity
Single cones	535 nm
Twin cones	614 nm
Rods	528 nm

Two cone types make the fish dichromatic. The authors conclude that bass have dichromatic vision and use chromatic cues, not only brightness, when making visual decisions.

A simple model built from those receptors made three predictions. Bass should not tell chartreuse yellow from white. They should not tell green from blue. They should tell red from every grey-scale stimulus.

What the behavioural trials showed

The behavioural test was a training task. Bass learned that a target colour meant a food reward, then were tested on whether they could separate that colour from one of similar brightness.

The abstract reports that bass trained to red and to green easily told their training colour apart from all other colours, including the ones matched in brightness (white and black, respectively). The authors say the model’s predictions were “partially upheld”. The abstract does not say which predictions failed, so this guide does not either.

The abstract also gives no sample size, tank conditions or light levels, so those are not discussed here.

What the study cannot say

• Wild fish. The trials were a food-reward training task. They show the fish can discriminate colours. They do not show which colours wild bass prefer to attack.
• Other species. The result is for largemouth bass. It says nothing about smallmouth, pike, walleye or trout.
• Other light. Discrimination at close range under test lighting is not discrimination at 6 m in a green lake.
• Strike behaviour. Seeing a colour and choosing to eat something that colour are separate steps.

Why the predictions matter

The model is useful because it can be wrong in a specific way. If bass could not separate red from grey, the red-trained fish would have failed the test. They did not fail it. The abstract says the predictions were only partly upheld, without saying which ones, so treat all three as a model that held up in part.

For a reader choosing between colours, the practical point is narrow: the hardware for seeing long wavelengths is there, and some colour pairs that look different to a person may look alike to a bass. Both statements come from a model of the eye plus a small number of trained-choice tests, and neither is a field result.

What water does to red

Colour vision only matters if the colour reaches the eye. Water absorbs and scatters light, and it does so unevenly across wavelengths.

The USGS Water Science School page in this guide’s sources covers turbidity: the amount of light scattered by material in the water, including clay, silt, algae, dissolved coloured organic compounds and plankton. Higher scattering means higher turbidity, and turbid water looks cloudy or opaque. That page does not give attenuation by depth or by colour, so it cannot tell you how fast red fades. Numbers for that come from the simulation’s model, not from the sources.

The model uses a standard exponential decay, I = I₀·e^(−Kd·d), with a separate Kd for each of eight wavelength bands. Taking the site’s coefficients at face value, here is what the model leaves at 5 m of path in two water types:

Water type	540 nm (green) remaining	620 nm (red) remaining
Clear	about 69%	about 29%
Green / algal	about 55%	about 22%

These are model assumptions, flagged as uncalibrated against field data on How it works. They show the direction of the effect. Red falls faster than green in both, and the gap widens with depth. In the model a deep, red lure is dimmer to a bass, but not invisible.

What the simulation does with it

The Fish Vision Camera takes the bass cone peaks (535 nm and 614 nm, matching the study) and runs the chain from sunlight to eye: attenuation going down, reflection off the lure, attenuation across to the fish, a backscatter veil, then integration by the cones. Light is blended toward rod vision as it falls, with a low-light gain of 1.6 for bass.

Per-water walkthroughs are on the clear water and green water colour pages, and the bass profile is summarised at largemouth bass colours.

The simulation simplifies things in ways that matter for red:

• Cone curves are Gaussians, while real ones are skewed.
• There is no UV channel and no polarisation.
• Water types are four fixed classes, not your lake.
• The Kd values and the low-light gain are model inputs, not measurements from the study.

How it works lists these limits, and says the results are comparative hypotheses, not predictions that a fish will strike.

What to check yourself

Open the Fish Vision Camera, pick largemouth bass and a red lure, then step through clear and green water at increasing depth. Watch what happens to the red against a green background. The study supports the idea that the fish has the receptors to see the difference. It leaves open how often that difference decides a strike.