Optics

The focus the eye loses a little every year

Somewhere in their forties nearly everyone starts holding books further away, and the change feels sudden: one year print is fine, a year or two later the arms are not long enough. Nothing sudden has happened to the eye. The range over which its lens can change its power has been shrinking by about a third of a dioptre a year since childhood, along a nearly straight line. What arrives at forty-five is the moment that line crosses the two and a half dioptres a book at forty centimetres asks for, twice over. The near point, being one over that range, races away just as the range runs out.

Assumes: What a lens is doing, and why three rays are enough · The focus that is a slab, not a plane

Ask someone in their late forties when they started needing reading glasses and they will usually name a year. Before it, print was fine; after it, menus were held at arm’s length, then further, and then glasses appeared. The experience is of a threshold. The physics is of a straight line, which has been sloping down since before they could read.

The eye focuses on near things by changing the shape of its lens. What a lens is doing described a lens by its power, the reciprocal of its focal length, measured in dioptres: a lens of one dioptre focuses parallel light at a metre, one of four at twenty-five centimetres. A relaxed, normal eye is focused on the far distance. To bring an object at distance dd into focus, the eye must add a power of 1/d1/d dioptres to its lens, two and a half dioptres for a book at forty centimetres, four for one at twenty-five. The amount it can add is called its amplitude of accommodation, and it is the quantity that decides how close anything can be seen.

How a lens changes its power

The lens does its focusing with a small index step and a large change of curvature. It sits in fluid, the aqueous humour in front and the vitreous behind, with refractive indices of about 1.336, while the lens itself averages about 1.42. The eye that loses its lens in water found that most of the eye’s power comes from the cornea’s front surface, where light enters from air; the lens’s surfaces, between tissue and fluid, have an index step only a fifth as large, and so contribute far less power for the same curvature. A surface’s power is the index step over its radius of curvature. For the lens’s front surface, with a step of 0.084, a radius of ten millimetres gives 8.4 dioptres; rounding it to six millimetres gives 14. The back surface steepens too, less, and the total change in a young adult is around ten dioptres.

That rounding is passive. The lens is held in a ring of fine fibres attached to the ciliary muscle; when the muscle relaxes, the ring is wide and the fibres pull the lens flat, and when it contracts, the ring narrows, the fibres go slack, and the lens’s own elastic capsule pulls it rounder. So the eye focuses near by letting go. The muscle keeps working in old age; it is the lens that stops responding.

A line that starts falling in childhood

The amplitude has been measured in thousands of people since the 1920s, most influentially by Alexander Duane, and in 1950 Henry Hofstetter fitted three straight lines to Duane’s data that have been used ever since: one for the largest amplitude found at each age, one for the average, one for the smallest.

The eye's focusing range shrinks steadily from childhood. How many dioptres of extra focusing power the eye can add for near work, against age, by Hofstetter's three lines through Duane's measurements. The mean falls 0.3 D a year from about 16 D at eight, without a break. Print at 40 cm needs 2.5 D, and holding it comfortably for long needs about twice that; the mean line crosses 5 D at 45, which is when reading first gets hard, and 2.5 D at 53, when it cannot be done at all. Nothing happens at forty-five except that a falling line crosses a fixed one. Past sixty the measured range levels off near half a dioptre, the eye's depth of focus.
Fig. 1 The eye’s amplitude of accommodation against age, by Hofstetter’s maximum, mean and minimum lines through Duane’s measurements. The mean falls 0.3 D a year from about 16 D at eight. Print at 40 cm needs 2.5 D and is comfortable with twice that in hand; the mean crosses 5 D at 45 and 2.5 D at 53. Past sixty the measured range levels off near half a dioptre.

The mean line is 18.5−0.3×age18.5 - 0.3 \times \text{age} dioptres. At eight it is sixteen dioptres, enough to focus at six centimetres, which is why a child can hold a toy against their nose and see it sharply. At thirty it is nine and a half. At forty-five it is five. The decline does not accelerate and does not start in middle age. It is as steady as anything in biology: the lens grows throughout life, adding layers of fibres, and becomes stiffer, so the muscles that relax its suspension and let it round up get less change of shape for the same effort.

Two lines across the figure mark what reading asks of it. A book at forty centimetres needs two and a half dioptres. But an eye working at the very limit of its range tires quickly and blurs, and people read comfortably for long only when they use about half their amplitude, keeping the rest in reserve. So the practical threshold is five dioptres. The mean line crosses it at forty-five. It crosses two and a half, the point at which the book cannot be focused at all however hard the eye tries, at fifty-three. The minimum and maximum lines put the first crossing between forty and fifty.

The threshold depends on the distance, and the distance has been changing. A phone is typically held at thirty centimetres or closer, which asks for 3.3 dioptres, and comfortable use for long periods asks for twice that, 6.7. The mean line reaches 6.7 dioptres at about thirty-nine. A generation that reads mostly from phones meets the same straight line six years earlier than one that read books at arm’s length, without any change in its eyes. The age at which “presbyopia begins” is a property of the reading habits of the time as much as of the lens.

That is the whole of the onset. Nothing happens at forty-five except that a line that has been falling for thirty-seven years crosses a line that does not move. The suddenness lies in the threshold, in the distance chosen for books, not in the eye.

Why the near point runs away

What a person notices is not the amplitude but the near point, the closest distance at which they can see sharply, and that is the reciprocal of the amplitude.

The book that has to move away. The nearest distance at which the average eye can focus, the reciprocal of its focusing range, against age, on a logarithmic scale. It recedes slowly for decades and then fast: 6 cm at ten, 11 cm at thirty, 20 cm at forty-five, 50 cm at fifty-five. Because the distance is one over a range that falls in a straight line, the same yearly loss moves the near point a few millimetres at twenty and tens of centimetres at fifty-five, and the arm runs out.
Fig. 2 The average eye’s near point, one over its amplitude, against age, on a logarithmic scale: 6 cm at ten, 11 cm at thirty, 20 cm at forty-five, 50 cm at fifty-five, past the 40 cm of a book and towards the end of an arm.

A reciprocal turns a straight line into a curve that hugs the axis and then shoots up. At twenty the amplitude is falling by 0.3 dioptres a year from about twelve, and the near point moves from 8.0 centimetres to 8.2: a fifth of a millimetre a month, imperceptible. At fifty the same 0.3 dioptres a year is taken from three and a half, and the near point moves from 29 centimetres to 31 in one year. At fifty-five it moves from 50 to nearly 60. The yearly loss in the eye is the same; its visible consequence grows as the square of the near point. By the time anyone notices, the near point is moving centimetres a year and passes the book, and then the length of an arm, within a few years.

The same reciprocal is behind every remark about presbyopia that sounds like a sudden change: “my arms got too short”, “the print got smaller”. The eye’s optics are a sum of powers, and the world is laid out in distances, and one is the reciprocal of the other. Every lens calculation has this shape: the condition a lens must meet and the thin-lens equation itself add reciprocals of distances, which is why a small change in power near infinity moves a focus a long way and the same change at close range barely moves it.

The slab of focus

A lens focused at one distance images that distance sharply and every other distance with some blur, but a small blur is invisible. The focus that is a slab, not a plane worked out how thick the slab of acceptable sharpness is for a camera, and the eye has one too, which is what lets an eye with almost no accommodation still read in good light.

A small pupil buys focus the lens no longer has. How far an eye can be out of focus, in dioptres either way, before its image visibly blurs, against the diameter of its pupil. The geometric blur grows with the pupil, so a tolerance of 2 arcminutes allows ±0.23 D at 2.5 mm and ±0.10 D at 6 mm; diffraction adds tolerance only at the smallest pupils, by the quarter-wave rule. A pupil narrowed to 2.5 mm in bright light tolerates ±0.29 D against ±0.10 D at 6 mm, which is part of why reading is easier by a window. The two tolerances are added in quadrature by convention.
Fig. 3 The defocus, in dioptres either way, that an eye tolerates before its image blurs past 2 arcminutes, against the diameter of its pupil. Geometric blur allows more defocus the smaller the pupil; diffraction adds tolerance only at the smallest pupils. A 2.5 mm pupil tolerates about ±0.29 D, a 6 mm one ±0.10 D.

The blur of a point out of focus by ΔD\Delta D dioptres, seen through a pupil of diameter pp, is a disc subtending an angle of about p ΔDp\,\Delta D radians. If the eye cannot tell a blur of two arcminutes from a point, about twice the finest detail normal vision resolves — a limit the eye a fly cannot afford traced to the eye’s own balance between the size of its pupil and the diffraction it causes — it tolerates a defocus of ΔD=β/p\Delta D = \beta/p, which is ±0.23 dioptres through a 2.5 millimetre pupil and ±0.10 through a 6 millimetre one. At small pupils diffraction makes the image fuzzy anyway, and adding a little defocus to a diffraction-limited image makes little further difference; by Rayleigh’s quarter-wave rule that adds tolerance as the inverse square of the pupil. The figure combines the two in quadrature, a convention rather than a law.

The pupil narrows in bright light, to two or three millimetres in sunshine, and widens to six or seven in the dark. Older eyes also have smaller pupils at every light level, by about a millimetre between twenty and sixty. Both help: a presbyopic eye reads better in bright light, partly because the print is brighter and partly because the narrower pupil deepens the slab of focus. Hofstetter’s lines level off near half a dioptre past sixty for the same reason. That residual half dioptre is not accommodation at all; the lens has stopped changing shape, and what is measured is the depth of focus of a small pupil.

A pinhole at sixty

Push the narrowing further, past what any pupil does, and an eye with no focusing range left can still read.

Reading through a pinhole at sixty. The blur in the image of print at 40 cm for an eye 1.5 D short of focusing on it — a typical sixty-year-old's — against the diameter of the opening the light comes through. Defocus blur shrinks with the opening and diffraction grows, and the sum is smallest at 0.94 mm, where it is 6.9 arcminutes: about half the height of a small letter, soft but legible. A pinhole in a card lets an eye with no focusing range left read, dimly; a pupil narrowed to 2 mm by bright light does part of the same.
Fig. 4 The blur in the image of print at 40 cm for an eye 1.5 D short of focusing on it, against the diameter of the opening the light comes through. Defocus blur shrinks with the opening and diffraction blur grows; the total is least at 0.94 mm, 6.9 arcminutes, about half the height of a small letter.

A sixty-year-old with one dioptre of range left is one and a half dioptres short of focusing on a book at forty centimetres. Through a normal pupil the blur is ten or fifteen arcminutes, as large as the letters. Look through a pinhole in a card and the defocus blur shrinks in proportion to the hole, while diffraction, which grows as the hole shrinks, sets a floor. The hole that makes the sharpest picture found exactly this balance in a pinhole camera, where the best hole is the one at which the geometric shadow of the hole and its diffraction pattern are the same size. Here the balance falls at a hole of

p∗=2.44 λΔD,p^\ast = \sqrt{\frac{2.44\,\lambda}{\Delta D}},

about 0.94 millimetres for green light and 1.5 dioptres of defocus. The blur there is seven arcminutes, about half the height of a lower-case letter at reading distance: soft, but legible, and very dim, since a millimetre hole admits a twenty-fifth of the light of a five-millimetre pupil. Pinhole spectacles, sheets of opaque plastic perforated with small holes, are sold on exactly this principle, and they work, at the cost of brightness and of a field of view broken into dots.

The surprising connection is that the camera and the ageing eye meet at the same optimum for the same reason. A pinhole camera has no lens and so infinite defocus on everything; an old eye has a lens that has stopped adjusting and a fixed defocus on near things. Both are rescued by the same trade between the shadow of a small aperture and the diffraction it produces, and both give an image limited to a few arcminutes.

Glasses that close a window

Reading glasses add a fixed power, usually between one and three dioptres, in front of the eye. They do not restore the lost range; they shift it.

Reading glasses give a window that closes with age. The distances an average eye sees sharply through +2.5 D reading glasses, against age, beside its unaided near point. The glasses move the far limit in to 40 cm, and the near limit is set by whatever focusing range remains: at forty-five they show everything from 13 to 40 cm; at sixty-five, with the range nearly gone, only 33 to 40 cm. A monitor at 60 cm is outside the window at any age, which is why a single pair of glasses stops being enough and lenses with graded power exist.
Fig. 5 The distances seen sharply through +2.5 D reading glasses, against age, beside the unaided near point. Through the glasses the far limit is fixed at 40 cm and the near limit is set by the range that remains: 13 to 40 cm at forty-five, 33 to 40 cm at sixty-five.

Through a lens of +2.5+2.5 dioptres, a relaxed eye focuses at forty centimetres instead of infinity, and the remaining amplitude brings its near point in to 1/(A+2.5)1/(A + 2.5). At forty-five, with five dioptres left, the glasses show everything from thirteen to forty centimetres, a generous window that comfortably holds a book. At sixty-five, with half a dioptre left, the window is from thirty-three to forty centimetres, seven centimetres deep. A book must be held at a particular distance; a computer screen at sixty centimetres, a dashboard, a face across a table, are all out of focus through the glasses and need the glasses removed.

That is why reading glasses that served at fifty seem not to serve at sixty-five, though the eye’s lens has hardly changed in between; it had little range left at fifty. The remedy is more lenses rather than stronger ones: bifocals with two powers, or progressive lenses whose power varies continuously from the top of the lens to the bottom, so that tilting the head chooses a distance. Every one of them is a way of doing by moving the head what the eye’s lens used to do by changing shape. The flat scene that comes back curved noted that no single lens images a whole flat scene sharply, and a progressive lens adds a deliberate version of the same compromise, with zones at its sides where the power changes sideways and vision is blurred.

The eyes that never need reading glasses

One group of people is spared the experience, partly. An eye that is short-sighted by two and a half dioptres is focused, when relaxed, not at infinity but at forty centimetres, and it reads a book at that distance with no accommodation at all. Myopes in their sixties take off their distance glasses to read, and find print perfectly sharp at the distance their eyes have always been focused at. They have lost exactly as much range as anyone else; their window has simply always been placed near rather than far.

Some people with healthy eyes are deliberately given the same arrangement in one eye only, a technique called monovision, in which one eye is corrected for distance and the other left or made a dioptre or two short-sighted. The brain attends to whichever image is sharper. And the most common surgery in the world, the replacement of a clouded lens with a plastic one in cataract surgery, turns everyone into a perfect presbyope: the implant has a fixed power, chosen for distance or for near, and an accommodation of zero, which is the end point the straight line was heading for all along.

Where the straight lines stop being right

Hofstetter’s lines are fits to population averages, and individuals scatter around them by two or three dioptres at any age; a particular forty-year-old may already struggle with print or may read comfortably at fifty. The measurements behind them were made subjectively, by moving a target towards the eye until the subject reported blur, which overestimates the true accommodation by the depth of focus; objective measurements of the lens’s actual change in power give lower amplitudes, especially in older eyes, and level off at zero rather than at half a dioptre.

The figures treat the eye as a thin lens and the pupil as its only aperture, ignoring the eye’s own aberrations, which are larger at large pupils and partly explain why vision in dim light is soft even for young eyes. They treat reading distance as fixed at forty centimetres, which is a convention; people with shorter arms or smaller phones read nearer and cross the threshold earlier. The “half in reserve” rule is a clinical rule of thumb rather than a measured law; the real comfortable fraction varies between people and with how long they read. And the eye’s own chromatic aberration, which the colour fringe no aperture can close found to be about two dioptres across the visible spectrum, means that an eye focused for green is out of focus for blue and red by more than the tolerances drawn here; vision copes by attending mostly to the middle of the spectrum, where the eye’s sensitivity peaks.

What survives the approximations is the shape of the argument: a range that falls in a straight line, a threshold set by a reading distance, and a near point that is one over the range and so runs away at the end.

Still open: why the lens stiffens

That the lens loses its ability to change shape is clear; why it does so is still argued. The lens keeps growing throughout life, so an older lens is larger and its suspension geometry changes; its proteins cross-link and its interior stiffens, by a factor of a thousand or more between youth and old age by some measurements; and the capsule that encloses it, whose elasticity rounds the lens when the suspension relaxes, changes too. How much of the loss is due to each, and whether any could be slowed or reversed, are questions on which treatments for presbyopia — eye drops, softening lasers, replacement lenses that accommodate — are currently being tried. None has yet restored a young eye’s range.

The habit worth carrying away is to look for the straight line under a sudden change. The eye’s range of focus falls by about 0.3 D a year from childhood, crossing the 5 D that comfortable reading at 40 cm needs at about forty-five and the 2.5 D it needs at all at about fifty-three; the near point, one over the range, moves a fifth of a millimetre a month at twenty and ten centimetres a year at fifty-five, and a pinhole about a millimetre across, balancing defocus against diffraction exactly as a pinhole camera does, lets an eye with no range left read again. Presbyopia is not an event. It is a reciprocal, seen at the end.

Part 16 of 16

This essay is one argument about Imaging. The others:

The objects named here

The third axis, after the field and the reading path: the things themselves, and every essay that touches each one.

AccommodationDepth of fieldDiffraction limitDioptreNear pointPinhole camera