Step out of bright noon sun and walk into a dim hallway. For about thirty seconds the world is a black wall — you put a hand out to find a doorframe. After a minute the shapes come back. After ten minutes you could read large print. After twenty you are, broadly, as sensitive as you are going to get. Try this on a moonless walk outside instead, and at the end of that twenty minutes you can see stars you didn't know were there.
The recovery curve hidden in that minute-to-twenty interval is dark adaptation, and it is one of the prettiest things in human vision. Plot threshold light intensity against time and it doesn't fall as a single smooth slope. It falls fast for a few minutes, plateaus, then sets off again on a second, slower descent — two segments joined by a sharp little kink. That kink, the cone–rod break, is the moment your visual system hands the job from one set of photoreceptors to another. By the time it's done, sensitivity has improved by roughly a million-fold over your daylight-bleached starting point — about five to six log units.
This post is a tour of the curve: the two-stage biology, what slows it down, and what dark adaptation has to do with the parts of vision an at-home contrast sensitivity test (like ours) is and isn't measuring.
The two-stage curve
Hecht, Haig and Chase ran the classical experiment in 1937: bleach the photopigment with a bright field, then have the observer in the dark report the faintest test flash they can see, minute by minute, for half an hour. The plot they got is the one vision-science textbooks still reprint (Hecht, Haig & Chase, 1937).
The cone segment is the first descent. It drops fast — most of it in the first two or three minutes — and finishes by around five to eight minutes in, plateauing at a moderate sensitivity. The cones are the daytime workforce: high-acuity, colour-capable, concentrated in the macula, fast to recover. But they have a sensitivity ceiling. After about seven minutes they have done what they can do, and the system idles on them while something else keeps regenerating in the background.
The cone–rod break is the kink. At about five to ten minutes in (the timing shifts with how bright the prior light adaptation was), the curve takes a corner and starts descending again. That corner is the moment rod sensitivity has caught up with and overtaken cone sensitivity, and the rods take over the threshold detection job. The break is sharp because the two photoreceptor systems are largely independent: when the slower one finally beats the faster one, threshold flips abruptly from one curve to the other.
The rod segment is the second, slower descent. It continues for another fifteen to twenty minutes, plateauing somewhere between twenty and thirty minutes in — gaining another two to three log units beyond where the cones left off. Daylight-to-dark, the total stacks to something like a factor of a million.
What is taking so long? Pigment regeneration. The light-detection molecule in rods is rhodopsin, a complex of opsin (a protein) and 11-cis-retinal (a small molecule derived from vitamin A). When a photon hits rhodopsin, the 11-cis-retinal isomerises to all-trans-retinal and the molecule releases its grip on the opsin — the rod is now "bleached." To reset, the cell has to ship the all-trans-retinal out, re-isomerise it back to 11-cis-retinal in the retinal pigment epithelium next door (the visual cycle, or retinoid cycle), and ship it back to opsin to recombine into fresh rhodopsin. Lamb and Pugh's 2004 review makes the molecular case that after a large bleach, dark-adaptation recovery is rate-limited by the delivery of 11-cis-retinal to opsin in the bleached rod outer segments (Lamb & Pugh, 2004). The kinetics of the cycle set the kinetics of the curve. Cones recycle their pigment faster via additional retinoid routes — which is why the curve has two segments instead of one.
What slows it down
A healthy young adult dark-adapts on a textbook schedule. Move along the demographics and the conditions, though, and that schedule starts to bend in characteristic ways.
Age. Dark adaptation slows with age, even in healthy older eyes, and the slowing falls disproportionately on the rod segment. Owsley and colleagues at the University of Alabama at Birmingham documented that healthy older adults have a measurably delayed rod intercept time — the time after a bleach for rod-mediated threshold to fall to a fixed criterion. The cone phase is largely preserved; the rods are the slow-down (Owsley et al., 2016). Some of this is RPE aging, some reduced retinoid delivery across an aging Bruch's membrane, some reduced photoreceptor numbers.
Early age-related macular degeneration. This is the dramatic one. Delayed rod-mediated dark adaptation is one of the earliest known functional signals of AMD — measurable in some patients before drusen are visible on a standard exam, and associated with incident AMD over a few years' follow-up (Owsley et al., 2016). The mechanism is plausibly the same one driving the contrast-sensitivity signal we cover in the AMD post: the RPE–Bruch's-membrane complex the rods depend on for retinoid recycling is the same complex AMD damages first. The clinical instrument here is the AdaptDx dark adaptometer (and its faster sibling, AdaptDx Pro), which a growing number of optometry practices have on hand. It runs a controlled bleach and reports a single rod-intercept-time number in around five to twenty minutes.
Vitamin A deficiency. The classical cause of "night blindness." Without 11-cis-retinal, the rods can't reset their rhodopsin, and dark adaptation fails. Frank deficiency is rare in well-fed populations but real in malabsorption syndromes (post-bariatric, severe Crohn's, advanced liver disease) and in highly restrictive diets. Importantly, it is treatable: replace the vitamin, the rods come back.
Retinitis pigmentosa and other rod dystrophies. RP is an umbrella for hereditary diseases that damage rod function first, then encroach on cones. One of the earliest symptoms is impaired dark adaptation, often beginning in childhood or adolescence. Stargardt disease, a juvenile macular dystrophy, similarly disrupts the visual cycle (via the ABCA4 retinoid transporter).
Some medications. Isotretinoin, amiodarone, and a handful of drugs with photoreceptor effects (sildenafil and its siblings, transiently). If your dark adaptation has changed since starting a new medication, raise it with the prescriber.
The shared theme: dark adaptation is a sensitive readout of the photoreceptor–RPE complex and the retinoid supply chain feeding it. When any link weakens, the curve slows. Like contrast sensitivity, it is a functional measure tuned to a part of vision the eye chart does not probe.
Dark adaptation and contrast sensitivity
These two measures are siblings, not twins. Both reflect parts of vision the Snellen chart misses; they don't measure the same thing.
A contrast sensitivity test is typically run in photopic (well-lit) conditions on a stationary stimulus. It probes the cone system at steady state and asks how faint a pattern your visual system can detect across a range of spatial frequencies. The classic clinical tools (Pelli-Robson, FACT, CSV-1000) and at-home tests like ours all live in this regime.
Dark adaptation is a temporal measure. It asks how quickly the rod system recovers after a bleach, and how sensitive it eventually gets. It is fundamentally about the dim and the transitional — walking into a dark room, driving at dusk, the first ten minutes of stargazing.
People who experience the world in low light a lot — dusk drivers, restaurant-goers, night-shift workers — depend on both systems. A reduced photopic CSF and a slow dark-adaptation curve are correlated but distinct problems. Either can be present without the other; either can produce the lived experience of "my eyes never quite catch up at dusk." Our low-light contrast post covers the steady-state mesopic regime once the system has finished adapting; this post is about the transition into it. In the AMD literature in particular, both signals have been documented as early — sometimes before standard acuity falls — and the clinical pattern of "both reduced" is more informative than either alone.
What you can do at home
There is no reliable DIY dark-adaptation test. The measurement requires a controlled bleach, a calibrated low-intensity test stimulus, a dark room, and a way to track threshold over many minutes — the AdaptDx and its lab cousins exist because the at-home environment can't deliver any of those constraints. We are not going to suggest a number you can put on your curve from your phone. Anyone who does is selling something.
What you can do — usefully, for free — is notice the symptoms that suggest the curve has changed:
- How long does it take you to see comfortably walking from a sunny street into a dim shop? Compared to a year ago, five years ago?
- How does dusk driving feel — specifically the first few minutes after sunset, when neither daylight nor headlights are running the scene?
- Do you trip in dim corridors more than you used to? Does your foot find the second-to-last stair when you expected the last one?
- Have you stopped doing things in the dark — dusk walks, stargazing, the night-time bathroom trip without the light — because they feel harder?
None of these is diagnostic of anything specific. All are reasonable inputs to bring to an eye exam, particularly if they have changed over months. The sentence to use in the clinic is: "I have noticed my eyes are slower than they used to be at adjusting from bright to dark. Would dark-adaptation testing be appropriate, given my age and risk factors?" That gives the clinician a reason to consider the AdaptDx — if the practice has one — or to refer you to a retinal specialist who does.
What our test can and can't tell you
Note. Our test measures photopic contrast sensitivity — the daytime curve — not dark adaptation. A normal contrast sensitivity result does not mean your dark adaptation is normal, and a reduced result does not specifically implicate the dark-adaptation pathway. They are different parts of vision measured by different instruments. If your dusk experience has changed but your contrast sensitivity result looks fine, that is not contradictory — dark-adaptation testing is a separate question worth pursuing through your eye doctor.
What our test is good at is giving you a baseline number for the steady-state, well-lit part of vision the eye chart underweights. A repeated result that drifts down over months is a real signal worth bringing in. The dark-adaptation story sits alongside that, not inside it.
Take the test (and ask the right next question)
Take our test now to set a photopic contrast sensitivity baseline. Three minutes, a normally-lit room, a saved result. If your dusk experience is the part that has changed — slow recovery walking into a dim room, harder night driving, lost stargazing comfort — that result is one piece of the picture and dark-adaptation testing is the other. Mention the dusk symptoms specifically at your next eye appointment.
The twenty-minute curve is one of the most beautifully physiological measurements in vision science: two photoreceptor systems, a molecular supply chain, a regeneration timescale you can almost feel happening as you sit in a dim room waiting to see. Most people never get it measured. A growing number of optometry practices can, and it is one of the more useful new tools in early-AMD screening. Worth knowing about; worth asking about, if dusk has gotten harder.
References
- Hecht, S., Haig, C., & Chase, A. M. (1937). The influence of light adaptation on subsequent dark adaptation of the eye. Journal of General Physiology, 20(6), 831–850. The classical experiment establishing the two-segment dark adaptation curve in humans: cone recovery in ~5–8 minutes, followed after the cone–rod break by a slower rod segment continuing to ~25 minutes. The figures from this paper are still the canonical reprint in vision-science textbooks.
- Lamb, T. D., & Pugh, E. N., Jr. (2004). Dark adaptation and the retinoid cycle of vision. Progress in Retinal and Eye Research, 23(3), 307–380. Molecular and biophysical review of dark adaptation, with a mathematical model showing that after a large bleach, recovery time is set by the rate at which 11-cis-retinal is delivered to opsin in the bleached rod outer segments. The standard reference for the kinetics-of-the-curve story.
- Owsley, C., McGwin, G., Jr., Clark, M. E., Jackson, G. R., Callahan, M. A., Kline, L. B., Witherspoon, C. D., & Curcio, C. A. (2016). Delayed rod-mediated dark adaptation is a functional biomarker for incident early age-related macular degeneration. Ophthalmology, 123(2), 344–351. Older adults with healthy maculas at baseline but delayed rod-mediated dark adaptation were several times more likely to develop early AMD within three years — establishing dark adaptation as a clinically useful early-AMD functional biomarker.
- Pelli, D. G., Robson, J. G., & Wilkins, A. J. (1988). The design of a new letter chart for measuring contrast sensitivity. Clinical Vision Sciences, 2, 187–199. Methodology anchor for clinical contrast sensitivity measurement and the argument that contrast sensitivity is a clinically meaningful complement to visual acuity.