What Is a Keratometer? K Readings, K1/K2 and How to Choose One

What is a keratometer? It is an instrument — historically called an ophthalmometer — that measures the curvature of the front surface of the cornea and reports the axis and extent of corneal astigmatism. That is the whole job. It projects an illuminated target onto the cornea, measures the reflected image, and converts what it sees into two numbers per eye plus an axis. Everything else an eye care practice does with keratometry — fitting contact lenses, calculating intraocular lens power, checking whether refractive cylinder and corneal cylinder agree — runs downstream of those numbers.

Keratometry is the measurement itself — a clinical reference describes it as "a method to measure the two principal meridional radii of curvature of the central cornea by measuring the size of mire reflections from the corneal surface." The instrument is the keratometer; keratometry is what it does; the K reading is what comes out.

This article covers how to read a K reading, what unit keratometry is expressed in, the optics behind the measurement, why two keratometers can disagree about the same eye, and how to decide between a standalone keratometer, a combined autorefractor/keratometer and a corneal topographer. If you have already made the category decision and want a model shortlist, our autorefractor buying guide is the commercial companion to this page.

The short version of the purchase, before the physics: for a general practice equipping a pretest lane, the instrument to shop for is usually not a standalone keratometer but a combined autorefractor/keratometer, which captures refraction and keratometry in the same patient seating. Our write-up of the Ezer ERK-BH autorefractor keratometer works through one unit in that category in detail. The cases where a standalone K is still the right buy are real but narrow, and they are listed further down.

What a keratometer measures — and what it does not

The National Library of Medicine's StatPearls review defines the keratometer as "an instrument used to measure the curvature of the anterior corneal surface and assess the extent and axis of astigmatism." Two things in that sentence do a lot of work: anterior, and curvature.

Anterior only. A keratometer reads the front surface of the cornea by reflection. It has no information about the back surface, so posterior corneal astigmatism is not in the reading at all — it is estimated by convention, not measured.

A very small zone. Keratometry samples a fixed chord of roughly 2–3 mm, described elsewhere as a 3–4 mm diameter ring, and it takes that reading from just four locations — two points on each of the two principal meridians. An adult cornea is about 11.7 mm across horizontally. The keratometer is therefore characterising the whole optical surface from a zone barely a quarter of its width, at four points.

A geometric assumption. The instrument assumes the cornea is a sphere or a spherocylinder. Real corneas are aspheric and radially asymmetric — they flatten toward the periphery, and after ocular surgery they may be frankly irregular. When the assumption holds, keratometry is fast, repeatable and entirely adequate. When it does not, the reading is still produced; it is just less meaningful. Knowing which situation you are in is the practical skill.

How to read a K reading: K1, K2, the axis and the units

This is the part the instrument manual assumes you already know, and it is what most people arriving at this page are actually asking. A keratometer produces two numbers per eye plus an axis. Here is what each one is.

K1 and K2 — the flat meridian and the steep one

The American Academy of Ophthalmology states the convention plainly: "K1 is the flattest meridian of the cornea (in dioptric power) and K2 is the steepest meridian of the cornea (in dioptric power)." So the lower number is K1, the higher number is K2, and both are powers in diopters. StatPearls separately describes what the pair is for: keratometry assesses "the extent and axis of astigmatism." This page does not publish an arithmetic rule linking the two K values to the astigmatism figure, because an automated instrument publishes corneal astigmatism and its axis as measurements of their own, with their own ranges and increments on the spec sheet — see the printout section below.

Diopters or millimetres — what unit is keratometry expressed in?

Both, and the distinction matters more than it looks. StatPearls notes that "the parameters of the keratometers are always labeled for the radius of curvature and dioptric power." The radius is what the instrument physically measures, in millimetres. The dioptric power is calculated from it, using the relationship StatPearls gives as D = (n−1)/r — with the conventional index substituted in, D = 0.3375/r, and inverted, r = 0.3375/D (with r in metres).

Two practical consequences follow. First, a steeper cornea has a smaller radius and a larger power, so the two columns run in opposite directions — the flat meridian K1 is the larger millimetre figure. Second, because the diopter column is computed through an assumed index and the millimetre column is not, the radius is the honest quantity to compare between two instruments. That is the subject of the 1.3375 section below, and it is the most common source of "the new unit reads differently" calls.

The rest of the printout

An automated AR/K reports more than the two meridian powers. Reading the published specification for the Ezer ERK-BH as an example, the instrument publishes a keratometry radius of curvature range of 5.0–13.0 mm in 0.01 mm increments, a corneal power range of 25.96–67.50 D in selectable 0.05/0.12/0.25 D increments, a corneal astigmatism range of 0.00 to −15.00 D in the same increments, and an axis range of 1–180° in 1° steps. The astigmatism value and its axis are published as their own measurement ranges on the spec sheet rather than being left for you to work out, and the selectable increment is worth noticing on a quote: an instrument set to 0.25 D steps and one set to 0.05 D steps need not print the same K value for the same eye.

How a keratometer works: mires, doubling and the reflected image

Keratometry models the tear-film-covered cornea as a convex mirror. Project an object of known size at a known distance, and the size of the reflected image varies with the mirror's curvature. StatPearls gives the working relationship as r = 2ul/o, where r is the radius of curvature, o the object size, l the image size and u the distance. Measure the image, and you have the radius.

What the mires actually are. The "mire" is the illuminated target the instrument projects, and what is measured is the size of its reflection. StatPearls describes the classical target as a circular mire carrying two plus and two minus signs — the alignment marks the operator brings together. A clinical reference puts the whole method in one sentence: keratometry measures "the two principal meridional radii of curvature of the central cornea by measuring the size of mire reflections from the corneal surface." Everything else — the doubling optics, the drums, the software — exists to make that one measurement survive a moving eye.

The complication is that the eye never holds still. Micro-movements shift the reflected image continuously, so measuring it directly is hopeless. The fix, which every keratometer since Helmholtz uses, is doubling: the optics deliberately produce two images of the mire and the operator aligns them until they just touch. Because both images move together when the eye moves, the alignment stays valid through the motion. It is an elegant piece of nineteenth-century engineering that survives unchanged inside modern automated instruments.

Helmholtz versus Javal-Schiotz — one position or two

The two classical designs differ in which variable they hold fixed.

• Helmholtz-type (variable doubling). The object size stays fixed and the amount of doubling is varied until alignment is achieved. The Bausch & Lomb instrument is the familiar implementation: four-aperture diaphragms with base-up and base-out doubling prisms produce two image pairs offset by 90°, so both principal meridians can be read from a single instrument position. This is the "one-position" keratometer.
• Javal-Schiotz-type (variable object size). The doubling is fixed — classically by Wollaston prisms exploiting birefringence — and the object size or its angle is varied instead. Reading both meridians requires rotating the instrument, which is why this design is called "two-position."

For a buyer in 2026 this is mostly context rather than a decision point: nearly every instrument you will be quoted is either a one-position optical keratometer or a fully automated unit. It matters when you are training staff on an older instrument you already own, or when a used unit's documentation calls a design by name and you need to know what technique it demands.

The 1.3375 problem: why two keratometers disagree about the same eye

This is the single most useful thing on this page, and it is the part spec sheets never explain.

A keratometer measures a radius, in millimetres. It reports a power, in diopters. Converting one to the other requires a refractive index, and the index used is not the cornea's real one. The true refractive index of corneal tissue is 1.376. Keratometers instead use a lower "keratometric" or effective index — conventionally 1.3375 in the United States, giving the familiar D = 0.3375/r — which fudges the anterior and posterior surfaces into a single number. The convention traces back to Helmholtz's original calibration.

The catch: the convention is not universal. 1.3320 is used in Europe, and instruments in the field publish other values too. Among the units we carry, the Ezer ERK-BH states its corneal power range "when cornea equivalent refractive index is 1.3375," while the Ezer ERK-770 states 1.337. Those are small differences that produce real, systematic offsets in the diopter column.

What to do about it: when you replace or add a keratometer, find the assumed index in the spec table before you compare readings between the old and new instrument, and before you compare a K value to one from an optical biometer or a topographer. If the numbers shift by a few tenths of a diopter on day one, the index is the first thing to check — not the instrument. Comparing the radius in millimetres sidesteps the issue entirely, because that is the quantity actually measured.

Measurement range and why older instruments need auxiliary lenses

A classical keratometer's native range runs roughly 36–52 D (6.5–9.38 mm), extendable to about 30–61 D with a −1.00 D or +1.25 D auxiliary lens over the objective. That limit is a genuine constraint on very flat or very steep corneas, and the auxiliary-lens workaround adds a step and a conversion table to the workflow.

Modern instruments have largely engineered the problem away. The Luxvision KR-800C publishes 35–68 D natively; the Ezer ERK-BH publishes 25.96–67.50 D. If your practice sees post-surgical or unusually steep corneas, native range is worth checking on the quote — but it is rarely the deciding specification any more.

What K readings do in a working lane

StatPearls lists keratometry's applications as intraocular lens power calculation, contact lens fitting, assessment and monitoring of astigmatism, and estimating refractive error where the media are hazy. Operationally, that translates into a handful of recurring uses:

• Contact lens fitting. Base curve selection starts from K. This is also where the small sampled zone bites hardest — a soft lens covers the entire cornea and beyond, while the keratometer has characterised the central few millimetres.
• Cross-checking the refraction. Corneal cylinder that disagrees sharply with refractive cylinder is a signal worth investigating rather than dispensing around.
• Pre-operative measurement. K values are a required input to IOL power formulas.
• A starting point when subjective refraction is difficult. Keratometry is objective and takes seconds.

For the clinical reasoning behind objective measurement in the pretest sequence, see our companion article on the importance of autorefractors. This page stays on the instrument and the purchase.

Four ways to get a K reading

Manual versus automated, concretely. The optical principle does not change; the detector does. A clinical reference describes autokeratometry this way: the devices "also use reflective techniques, but the images of the reflected infrared mires are collected with infrared detectors and analyzed to calculate radii of curvature." The reasonable reading of that sentence is that the alignment judgement a manual instrument asks of the operator is made by the detector and the software instead. StatPearls summarises the trade the same way — automated keratometers are "compact, less time-consuming, and require less skill to operate." That last clause is the staffing argument, and it is the one worth pricing: the less operator skill a capture demands, the wider the set of staff who can be trained to take it.

One clarification that catches buyers out: "handheld autorefractor" and "handheld keratometer" are not the same instrument. The Ezer ERH-770 handheld autorefractor, for example, is a Shack-Hartmann aberrometer that publishes sphere, cylinder and axis — and no keratometry at all. If you need portable K specifically, confirm the keratometry specification is in the published table rather than assuming a portable refraction device includes it.

Keratometer versus corneal topographer

The difference is coverage, and it decides the instrument.

Keratometry gives you the two principal meridians from four points inside a 3–4 mm zone. A topographer characterises a much wider area, with Placido-based systems sampling along the mires at intervals as fine as one degree, and can derive a simulated K for continuity with your existing records. For a regular, roughly spherocylindrical cornea the two will broadly agree, and the keratometer is faster and far cheaper to put in a lane.

Where they part company is irregularity. As the clinical literature puts it plainly, K values used in IOL calculations "should not be measured with keratometry in eyes that have undergone keratorefractive surgery or are irregular for any other reason." A cornea reshaped by LASIK or PRK breaks the fixed-index assumption described above, and four sampled points cannot describe a surface that is no longer symmetric.

The practical conclusion for most practices is not "topographer instead of keratometer." It is that the AR/K handles the daily volume and a topographer such as the Ezer Onyx corneal topographer earns its place when your case mix includes specialty lens fitting, post-surgical eyes or premium IOL planning. They answer different questions, and they sit at very different levels of complexity and investment.

The buying decision: standalone keratometer or an AR/K?

Most practices buying today choose a combined autorefractor/keratometer, and for most that is correct: one seat in the pretest lane, one instrument to train on, one printout, refraction and K captured in the same sequence. The autorefractor/keratometer category is where the volume is.

The honest way to frame the decision is as a question about the lane, not the instrument: what else has to happen while the patient is in that chair? If the answer includes sphere, cylinder and axis — and in a general practice it almost always does — then a standalone keratometer buys you one of the two measurements and still costs you the seat, the training and the workflow step. That is the whole case for the AR/K, and it is why a category page and not a keratometer page is usually the right next click. Our write-up of the Ezer ERK-BH autorefractor keratometer works through one unit in that category in detail, and the autorefractor buying guide compares the field.

A standalone keratometer still makes sense in specific situations:

• You already have refraction covered. If a digital refractor or an existing autorefractor handles sphere and cylinder, you may only need K.
• Teaching and technique. A manual keratometer shows a trainee the mires, the tear film and the doubling principle directly. Automated instruments hide all of it.
• A backup that cannot go down. A purely optical instrument has no display, no board and no firmware. It works during a power cut and it does not become unsupported.
• Portability. A handheld keratometer goes where a 20 kg table-top cannot.

Against that: a standalone occupies bench space and a step in the workflow to produce one of the two numbers an AR/K gives you together, and a manual unit's readings depend on operator technique in a way an automated capture does not.

Specifications that actually matter when you compare

• Keratometry radius range (in mm) — the quantity actually measured, and the only one you can compare across instruments without worrying about the index.
• Corneal power range and the assumed index. Compare like for like. A range quoted at 1.3375 is not directly comparable to one quoted at 1.337.
• Corneal astigmatism range and increment. Increments of 0.05/0.12/0.25 D are typical on current automated units.
• Corneal diameter measurement — useful for lens sizing; ranges differ meaningfully between models.
• Contact lens base curve (CLBC) mode. If you fit lenses, this lets the instrument read the base curve of a lens directly. It is present on some units and absent on others, and it is easy to miss on a spec table.
• Peripheral keratometry. Some AR/K units add a peripheral K mode, which extends what the instrument can say about the cornea outside the central zone.
• Minimum pupil diameter (for the refraction side of an AR/K) — it determines how often you get a reading on a small or difficult pupil.
• Output and connectivity. A built-in printer, and whether the unit offers USB, RS-232C, Bluetooth or Wi-Fi. Note that a connector on a spec sheet is not the same as a working link to your records system — confirm what the connection actually delivers with your specific software.
• Footprint and weight. Depth is usually the binding constraint on an instrument table, not width.
• Service, spare parts and warranty. An instrument that measures in hundredths of a millimetre needs a support path when it drifts or a part fails.

Keratometry across the US Ophthalmic range

Every figure below is taken from the current product page for each model. Where a specification is not published, it is left blank rather than estimated.

A few things worth drawing out of that table. The ERK-BH is the only unit here that publishes both a contact lens base curve mode and a peripheral keratometry mode, and it has the widest radius range at 5.0–13.0 mm — relevant if you fit lenses or see flatter corneas. The ERK-9000 A trades some of that range for a larger 8″ display and a motorized chinrest. The ERK-770 is the connectivity option. And the KR-800C is the one instrument on the list that will still work when the power is out.

Ezer is US Ophthalmic's own brand, and every unit above is supported by our in-house technical and spare-parts department with US warranty and service — which is the part of a keratometry purchase that only shows up two years later, when something needs adjusting.

What to confirm before you sign

• Which keratometric index the instrument uses, if you will be comparing readings with an existing unit or a biometer.
• Whether CLBC mode is included, if contact lens fitting is part of your practice.
• The printer consumable — roll width and how you reorder it.
• What the data interface will actually deliver with your specific records software, in writing.
• Your instrument table's usable depth against the unit's depth, not its width.
• The warranty term and exactly which assemblies — chinrest motor, printer, touchscreen — are covered.
• Current availability and lead time on the specific model, which moves.

Frequently asked questions

What is a keratometer used for?

It measures the curvature of the front surface of the cornea and reports the axis and extent of corneal astigmatism. Those readings are used as inputs to contact lens base curve selection, intraocular lens power calculation, monitoring of astigmatism over time, and estimating refractive error when the ocular media are hazy.

What do K1 and K2 mean on a keratometry reading?

K1 is the flattest meridian of the cornea and K2 is the steepest, both expressed in dioptric power, per the American Academy of Ophthalmology. K1 is therefore the lower of the two diopter figures and, because power and radius run in opposite directions in the formula D = 0.3375/r, the larger radius in the millimetre column. An automated unit publishes corneal astigmatism and its axis as separate measured values with their own ranges, rather than leaving them to be derived from K1 and K2.

What unit is keratometry expressed in?

Both millimetres and diopters. The radius of curvature, in millimetres, is what the instrument measures. The dioptric power is calculated from it — StatPearls gives the relationship as D = (n−1)/r, which with the conventional keratometric index becomes D = 0.3375/r and inverts to r = 0.3375/D, with r in metres. Because the diopter figure passes through an assumed index and the millimetre figure does not, the radius is the safer quantity to compare between two different instruments.

What is the difference between a keratometer and an autorefractor?

An autorefractor measures the refractive error of the whole eye and reports sphere, cylinder and axis. A keratometer measures only the curvature of the anterior cornea. A combined autorefractor/keratometer (AR/K) does both in one capture sequence, which is why most practices buy the combined instrument rather than two separate ones.

Should I buy a standalone keratometer or an autorefractor keratometer?

For a general practice equipping a pretest lane, the combined autorefractor/keratometer is the default: it produces the K reading and the refraction in one patient seating, on one seat of bench space, with one instrument for staff to learn. A standalone keratometer earns its place when refraction is already covered by another instrument, when the purpose is teaching the technique, when a purely optical backup that cannot lose firmware support is wanted, or when the reading has to be taken away from the table. Compare the combined units in the autorefractor/keratometer range and request a quote for current configurations.

How accurate is a keratometer?

Current automated instruments resolve the corneal radius to 0.01 mm, and the Ezer EZ-KR-1800 handheld publishes a precision of ±0.05 mm with axis deviation within ±2°. The larger source of variation is not the instrument's resolution but its assumptions: keratometry samples four points inside a 3–4 mm zone and models the cornea as a sphere or spherocylinder. On a regular cornea that is a good model. On an irregular or post-surgical cornea it is not, and topography is the appropriate instrument.

Why do two keratometers give different diopter readings for the same eye?

Because the diopter value is calculated from the measured radius using an assumed refractive index, and that index is a convention rather than the cornea's true value of 1.376. Instruments in the United States conventionally use 1.3375; 1.3320 is used in Europe, and other values appear in the field — the Ezer ERK-BH states 1.3375 and the ERK-770 states 1.337. Comparing the radius in millimetres avoids the discrepancy entirely.

Do I need a keratometer if I already have a corneal topographer?

Most topographers can output a simulated K, so the reading is available. Practices still keep a keratometer or an AR/K in the pretest lane because it is faster, cheaper to place, and captures K alongside refraction in one patient seating. The topographer earns its place on the cases keratometry cannot describe: irregular corneas, post-refractive-surgery eyes and specialty lens fitting.

Is there a handheld keratometer?

Yes. The Ezer EZ-KR-1800 is a handheld corneal curvature instrument measuring 240×90×60 mm with a single measurement time of 0.03 s, a CLBC mode and a wireless infrared thermal printer. Be aware that a handheld autorefractor is a different instrument — the Ezer ERH-770, for example, is a Shack-Hartmann aberrometer and does not publish keratometry.

What is a normal K reading?

Keratometers are built around the range human corneas actually occupy: a classical instrument's native range is about 36–52 D (6.5–9.38 mm), extendable with auxiliary lenses, and current automated units publish wider ranges — 25.96–67.50 D on the Ezer ERK-BH and 35–68 D on the Luxvision KR-800C. Interpretation of an individual reading belongs with the examining clinician.

What is the difference between a one-position and a two-position keratometer?

A one-position instrument, following the Helmholtz principle with fixed object size and variable doubling, reads both principal meridians without being rotated — the Bausch & Lomb design is the classic example. A two-position instrument, following the Javal-Schiotz principle with fixed doubling and variable object size, must be rotated to read the second meridian. One-position is faster and is what most manual instruments in current use are.

Sources

Optical principles, the mire target, the radius/diopter relationship, sampled zone, keratometric index, measurement range and limitations: Keratometer, StatPearls, National Library of Medicine. Definition of keratometry, the mire-reflection method and autokeratometry's infrared detection: Keratometry and Topography (clinical reference). K1 as the flattest meridian and K2 as the steepest, in dioptric power: How to Read Corneal Topography, American Academy of Ophthalmology. Corneal dimensions and true refractive index: Anatomy of the Cornea. All model specifications are taken from the current US Ophthalmic product pages linked above and were verified on the product pages on 7 August 2026 and the ERK-BH keratometry specifications were re-verified on 18 September 2026. Request a quote for current configurations, availability and lead times.