Reproducing the colors of living things faithfully in a scientific image is far harder than pointing a camera and pressing a button. Between the moment light bounces off a bird feather, a stained tissue slide, or a fluorescing coral, and the moment a reader looks at a published figure, dozens of translation steps can distort color: the sensor’s spectral sensitivity, the monitor’s gamut, the color map chosen for data visualization, and even the reader’s own visual biology. Achieving biological accuracy with color in science means controlling those translation steps so the final image or figure honestly represents the specimen, the data, or both.
Human Vision Sets the First Limit
Before worrying about cameras and printers, it helps to acknowledge that human eyes are not the gold standard for color. Our trichromatic vision, built on three classes of cone photoreceptors sensitive to roughly red, green, and blue wavelengths, covers a spectral window of about 400 to 700 nanometers. That sounds generous until you compare it with the rest of the animal kingdom. Birds, reptiles, and many freshwater fish are tetrachromats with four receptor types, extending their sensitivity into the ultraviolet range and perceiving finer hue distinctions than we can. Insects and even mice use ultraviolet light routinely. Dogs, horses, and some primates are dichromats, roughly equivalent to a red-green colorblind human.1Current Biology. Unconventional colour vision The upshot for science is that any color palette built on human perception is inherently subjective. A “biologically accurate” rendering of a flower pollinated by bees misses the ultraviolet nectar guides bees actually see, while a rendering of reef fish coloration may wash out signals visible to tetrachromatic predators.
Visual ecologists have developed perceptual color spaces that model how a given animal’s photoreceptors would interpret a stimulus, allowing researchers to move beyond the human-centric assumption and ask what a color signal looks like to its intended receiver.2PubMed. Colour spaces in ecology and evolutionary biology These models range from rough, generalist frameworks useful for broad comparisons to species-specific systems that incorporate the neural processing behind color perception. For any biologist studying animal signaling, “accurate color” is not a single target. It shifts depending on whose eyes you are trying to represent.
Physical Color Standards and Why They Still Matter
Long before digital imaging, taxonomists needed a way to describe color consistently. You cannot mail a living fungus to a colleague in another country, so nineteenth- and early-twentieth-century biologists developed printed color atlases as references. Robert Ridgway’s 1912 atlas, “Color Standards and Color Nomenclature,” became one of the most widely used guides in natural history, containing over a thousand named color swatches. A careful comparison of five surviving copies of Ridgway’s atlas found remarkable consistency: variation among copies was less than one Munsell hue unit in 83 percent of the 1,140 comparisons tested.3Mycological Research. Colour standards: a review and evaluation in relation to Penicillium taxonomy That is impressively stable for a printed product from 1912. Other standards, like the Methuen Handbook of Colour, have shown edition-to-edition discrepancies in hue, a reminder that physical color chips are only as reliable as their printing process.
The acceptable printed references for fine color distinctions remain the Munsell Book of Color and Ridgway’s atlas.3Mycological Research. Colour standards: a review and evaluation in relation to Penicillium taxonomy Even in a digital age, Munsell notation persists in soil science, mycology, and paleontology because it provides a device-independent language for color. A Munsell code printed in a 1920 journal paper can be matched to a chip today without any assumptions about monitor calibration or software rendering.
From Eyeballing to Spectrophotometry
For much of the history of biology, color was recorded the way most of us still describe it: someone looked at a specimen and wrote down a name. That approach is fast, cheap, and reproducible enough for broad categories, but it stumbles on subtle distinctions. Two researchers may disagree on whether a feather is “dark brown” or “rufous,” especially under different lighting. Spectrophotometry bypasses this problem by measuring the percentage of light reflected at each wavelength, producing an objective spectral curve rather than a subjective label.
A recent study assessed whether the traditional visual color categories used in paleontology for melanin-based feather colors actually hold up when tested against reflectance data. By measuring black, brown, grey, and rufous feathers with a spectrophotometer and analyzing the resulting spectra statistically, the researchers could quantify how well human-assigned categories correspond to measurable spectral differences.4Royal Society Open Science. Quantitative characterization of melanin-based feather colour using reflectance spectrophotometry This kind of ground-truthing is essential when the goal is biological accuracy rather than qualitative convenience, and it matters especially in fields like fossil reconstruction, where the original organism cannot be re-examined.
Structural Colors Pose a Special Challenge
Not all biological color comes from pigment. Many of the most vivid hues in nature, the iridescent blue of a morpho butterfly wing, the shimmering green of a beetle’s carapace, are structural colors produced by nanoscale physical structures that interfere with light. Structural colors shift with viewing angle, polarization, and illumination in ways that pigmentary colors do not, making them notoriously difficult to capture accurately in a photograph or reproduce in print.
Investigating these systems requires a broader toolkit than simple spectrophotometry. Because biological photonic structures are often more elaborate than synthetic ones, researchers need electron microscopy, optical modeling, and angle-resolved reflectance measurements to fully characterize how the color is produced and how it would appear under natural conditions.5PubMed Central. Physical methods for investigating structural colours in biological systems A flat photograph of a structurally colored feather taken from a single angle and under a single light source may look accurate on screen but misrepresents the feather’s appearance in the wild, where changing angles produce shifting color. Any published image of such a specimen should ideally note the viewing geometry and lighting conditions, though in practice few do.
Color Maps That Distort Your Data
Scientific color palettes are not only used to represent what a specimen looks like; they are also used to encode numerical data, as in heat maps, elevation models, and medical scans. The choice of color map here is not cosmetic. A perceptually non-uniform map like the classic rainbow (jet) palette assigns visual weight unevenly: the eye perceives sharp boundaries between certain color bands and smooth gradients between others, creating phantom features in the data that do not exist. The distortion introduced by such a color map can diverge from an objective representation by more than seven percent of the displayed data variation.6Nature Communications. The misuse of colour in science communication
A perceptually uniform color map, by contrast, spaces its colors so that equal steps in the data produce equal steps in perceived lightness. Maps like viridis, inferno, and cividis were designed with this property and also retain legibility in greyscale, which matters both for accessibility and for readers who print papers on black-and-white printers. Switching from rainbow to a perceptually uniform palette is one of the simplest and highest-impact changes a researcher can make to improve the fidelity of their figures. The data does not change, but the reader’s ability to interpret it honestly improves.
Colorblind Accessibility in Science Figures
Roughly eight percent of men and half a percent of women of European descent have some form of color vision deficiency, with deuteranopia (reduced sensitivity to green) being the most common type. That translates to a substantial fraction of any scientific readership who may struggle with red-green coded figures. A survey of published cell biology papers found that 45 percent contained at least some fluorescence microscopy images that were inaccessible to readers with deuteranopia, and 17 to 34 percent of papers contained color annotations invisible to someone with the condition.7PLOS Biology. Creating clear and informative image-based figures for scientific publications
The fixes are well established. Green-and-magenta or cyan-and-orange pairings replace red-and-green without losing information for colorblind viewers. Simulators like Color Oracle let you preview a figure through a deuteranopic viewer’s eyes before submitting. Best-practice protocols now recommend checking every figure with such a simulator during preparation.8PubMed. Choosing Suitable Color Palettes for Accessible and Accurate Science Figures Despite this, the adoption rate in published literature remains uneven. Journals increasingly require accessible color schemes in their submission guidelines, but enforcement varies.
RGB, CMYK, and the Printing Gap
A figure that looks perfect on a monitor may shift color dramatically when printed. Monitors display color in the RGB color space, which mixes red, green, and blue light additively. Printing presses use the CMYK space, which subtracts light using cyan, magenta, yellow, and black inks. The RGB gamut is generally larger than the CMYK gamut, meaning bright, saturated colors visible on screen, particularly vivid greens and electric blues, cannot be reproduced in print. The result is an unexpected shift in appearance when an RGB image is converted to CMYK for publication.9PubMed Central. Color Figures in BJ: RGB versus CMYK
For a fluorescence micrograph in which a researcher has carefully tuned green and red channels to represent two different proteins, this color shift can be more than an aesthetic annoyance. It can change what a reader perceives as signal versus background. The practical advice is to soft-proof CMYK output before submitting to a print journal, and when possible to use color choices that fall within the overlap region of both gamuts. Digital-only journals sidestep CMYK entirely, but even there, differences in monitor calibration among readers mean that the precise hue you see on your screen is not guaranteed to be the hue someone else sees on theirs.
Fluorescence Microscopy and Spectral Unmixing
Modern fluorescence microscopy routinely labels different cellular structures with distinct fluorophores, each emitting light at a slightly different wavelength. The color palette in the resulting image is therefore assigned by the researcher: Channel 1 might be mapped to blue, Channel 2 to green, Channel 3 to red. The biological accuracy of these images depends on how cleanly the signals from overlapping fluorophore spectra can be separated. When emission spectra overlap, one channel bleeds into another, creating false co-localization that looks like two proteins are in the same place when they are not.
Spectral unmixing algorithms address this problem by mathematically decomposing mixed signals into their individual components. A machine-learning-based method called LUMoS demonstrated the ability to cleanly separate up to six fluorophores in biological samples imaged with only four detectors, using unsupervised learning to identify each fluorophore’s spectral signature without prior calibration.10PubMed Central. Robust blind spectral unmixing for fluorescence microscopy using unsupervised learning This kind of approach matters because the more fluorophores you can image simultaneously, the richer the biological story you can capture in a single experiment, but only if the downstream color separation is trustworthy.
Histological Stains and Their Surprising Complexity
The most widely used stain in medical histology, the hematoxylin-and-eosin (H&E) stain, produces images in which nuclei appear blue-purple (basophilic) and most cytoplasm and extracellular structures appear pink (eosinophilic). It sounds simple, but the H&E stain has stood as the standard for histologic examination of human tissues precisely because this two-dye combination highlights fine cellular structures with a richness that more complex staining protocols sometimes fail to match.11PubMed. The wonderful colors of the hematoxylin-eosin stain in diagnostic surgical pathology
The challenge for color accuracy here is batch-to-batch variability. Different laboratories use slightly different fixation protocols, staining durations, and reagent concentrations, which produce visually distinct shades of blue and pink from the same tissue type. A pathologist trained in one lab develops an internal color reference that may not match slides from another institution. This variability becomes a serious problem in digital pathology, where algorithms trained to detect cancer on one lab’s slides may fail on another’s. Generative adversarial networks (GANs) have been applied to the color normalization task, treating it as an image-to-image translation problem that maps every pixel from the source lab’s color profile to a standardized target while preserving tissue architecture.12Expert Systems with Applications. Generative models for color normalization in digital pathology and dermatology: Advancing the learning paradigm Without this kind of correction, a deep-learning diagnostic tool could confuse staining artifacts with genuine pathology.
When Specimens Lose Their Color
Preserved biological specimens do not stay the color they were in life. Formalin fixation, the standard for preserving soft tissue, strips the natural reds and pinks from organs by oxidizing hemoglobin and myoglobin. A heart that was deep red on the operating table becomes a muted grey-brown in a jar within weeks. For teaching collections and anatomical museums, this color loss erases information a student or clinician would otherwise use to distinguish healthy tissue from pathological tissue.
A technique originally developed in Hungary by Romhányi uses pyridine, nicotine, and sodium dithionite to form red-colored complexes with ferro-hemochromes in the tissue, restoring a natural-looking redness. A modified version of this approach, incorporating citrate-phosphate-dextrose buffer and glycerin to control pH and osmolality, produced specimens that maintained stable, natural red coloration with no histological deterioration for more than 18 years.13PubMed. Re-colorization of museum specimens: a modification of Romhanyi’s technique based on pyridine/nicotine hemochromogen reactions More recently, 3D photometric scanning has been used to digitally archive formalin-preserved heart specimens as high-resolution, color-accurate 3D models that capture complex internal structures, offering a way to preserve color information digitally even as the physical specimen continues to degrade.14PubMed. Digital curation of formalin-preserved heart specimens via 3D photometric scanning: A report on recent archiving techniques and optimizations
Field Photography and the Calibration Problem
Field biologists frequently use digital photography to record animal or plant coloration in the wild. The appeal is obvious: a camera is lighter than a spectrophotometer, and photographs capture spatial patterns as well as color. But converting pixel values into biologically meaningful color data requires careful calibration. Light conditions in the field change constantly, white balance settings vary between camera bodies, and sensor response curves differ across manufacturers.
One common calibration approach uses a standardized color chart, such as a 24-patch ColorChecker, photographed alongside the specimen. Software then builds a statistical filter to map the camera’s response to known color values. However, a comparative study found that statistical filters trained on either a 24- or 45-color chart varied wildly in accuracy, with R-squared values ranging from 0.11 to 0.94, leading the authors to recommend that such filters should not be trusted and their use should be discontinued.15Methods in Ecology and Evolution. A comparison of photographic and spectrometric methods to quantify the colours seen by animal eyes For researchers who need quantitative color data rather than qualitative impressions, spectrophotometry remains the more reliable tool. Photography is useful for spatial context and pattern documentation, but the color numbers it produces should be treated with caution unless validated against spectral measurements.
False Color in Microscopy and Medical Imaging
Not all color in science is meant to replicate what something “really” looks like. Electron microscopy, for example, produces inherently grayscale images because it uses electron beams rather than visible light. Colorizing SEM images has become common practice to make surface details easier for the human eye to parse, distinguishing cell types in a tissue or highlighting features of a microorganism’s surface.16Oxford Academic (Microscopy Today). Color (and 3D) for Scanning Electron Microscopy The color in such images is entirely fabricated, applied after capture, and should be understood as a visualization aid rather than a record of the specimen’s actual appearance.
In medical imaging, false-color maps serve a different purpose: they encode functional data. A dynamic contrast-enhanced MRI scan, for instance, tracks how quickly contrast agent washes into and out of tissue, a signal that can distinguish malignant from benign tumors. Displaying this data in color rather than greyscale has been shown to produce significantly higher diagnostic quality ratings and better correlation between signal amplitude and vascular scoring.17PubMed Central. Comparison of diagnostic quality and accuracy in color-coded versus gray-scale DCE-MR imaging display The palette choice matters here not for biological realism but for diagnostic performance: the “right” colors are whichever ones help clinicians detect disease most reliably.
Multispectral Imaging Beyond the Visible
Plants reflect light in patterns that carry information invisible to the naked eye. Healthy leaves absorb most visible red light for photosynthesis but strongly reflect near-infrared wavelengths. Stressed plants shift this balance before any visible symptoms, like wilting or yellowing, appear. Remote sensing exploits this by computing vegetation indices from multispectral images, essentially comparing red and near-infrared reflectance to produce a false-color map of plant health.18PubMed Central. Utilization of single-image normalized difference vegetation index (SI-NDVI) for early plant stress detection The resulting images use color palettes where green means “healthy” and red means “stressed,” but these colors are assigned by convention, not by how the plant actually looks.
Plant reflectance is governed by leaf surface properties, internal structure, and the concentration and distribution of biochemical components like chlorophyll and water.19Trends in Plant Science. Science Color Palette: Achieving Biological Accuracy A similar logic applies to bioluminescence research, where the goal is to measure the spectral output of organisms that produce their own light. Specialized instruments can capture emission spectra from single photophores using fiber-optic-coupled photodiode arrays, recording data across a spectral window that far exceeds what any photograph could capture.20PubMed. Marine bioluminescence spectra measured with an optical multichannel detection system In both cases, the “color palette” is a tool for visualizing data that the human eye cannot see directly, and accuracy is measured against instrumental standards, not against subjective visual impression.
3D Models and the Promise of Photogrammetry
One of the more promising developments in biological color documentation is photogrammetry, which reconstructs three-dimensional models from sets of two-dimensional photographs. A protocol developed for studying flower morphology used a digital SLR camera, a turntable, and a portable lightbox to generate 3D models that captured both the geometry and the color information of the original specimens.21PubMed. Studying flowers in 3D using photogrammetry Unlike a single photograph, a 3D model can be rotated, measured, and examined from any angle, making it possible to study how color varies across a petal’s surface or how structural coloration changes with viewing geometry.
Photogrammetry has also been applied to the archival problem for preserved specimens, such as the formalin-fixed heart models described above. The technique produces models that preserve color at the time of scanning, creating a permanent digital record even if the specimen later degrades. For herbarium sheets, insect collections, and coral fragments, 3D photogrammetric models may eventually replace photography as the standard documentation method, giving future researchers access to both shape and color information that a flat image cannot provide.