In short

No matter how perfectly you grind a lens or how smoothly you polish a mirror, the wave nature of light sets an ultimate limit on how close two point sources can be and still be seen as separate. Each point source, imaged through a circular aperture of diameter D, becomes not a point but an Airy disc of angular radius (first dark ring)

\theta_\text{Airy} \;=\; 1.22\,\frac{\lambda}{D}

Two sources at angular separation smaller than this merge into a single blob. They are "just resolved" — visible as two — when the central maximum of one Airy pattern falls on the first minimum of the other. This is the Rayleigh criterion:

\boxed{\;\theta_\text{min} \;=\; 1.22\,\frac{\lambda}{D}\;}

Resolving power of a telescope. The smallest angular separation the telescope can distinguish is \theta_\text{min} above; the resolving power is RP = 1/\theta_\text{min} = D/(1.22\lambda). Larger aperture, finer detail. (At a visible wavelength \lambda = 550 nm, a 10 cm telescope resolves \sim 1.4 arc-seconds; a 10 m telescope resolves \sim 0.014 arc-seconds — a hundred times better.)

Resolving power of a microscope. The smallest separation between two object points that can be resolved is

\boxed{\;d_\text{min} \;=\; \frac{0.61\,\lambda}{n \sin\alpha} \;=\; \frac{0.61\,\lambda}{\text{NA}}\;}

where \alpha is the half-angle of the cone of light the objective collects from the object, n is the refractive index of the immersion medium, and numerical aperture \text{NA} = n\sin\alpha. The resolving power of a microscope is 1/d_\text{min}.

The Abbe diffraction limit. Ernst Abbe's original form of the rule is d_\text{min} = \lambda/(2\,\text{NA}) — essentially the Rayleigh version with a coefficient just under 0.5 instead of 0.61, depending on which kind of criterion you apply. For visible light with the best oil-immersion objectives (\text{NA} \approx 1.4), the limit is about 200 nm — the reason optical microscopes cannot see individual proteins and why electron and X-ray microscopy were invented.

The resolving limit is not a flaw in the instrument. It is a fundamental fact about light — a diffraction pattern around every aperture is unavoidable, and Rayleigh's criterion is the first place that the fact becomes an engineering constraint.

Two stars lie 1 arc-second apart in the night sky over Kavalur, in the hills of Tamil Nadu where the Indian Institute of Astrophysics operates its 2.3-metre Vainu Bappu telescope. Observed with the naked eye from Chennai, they are a single point — unresolved. Observed with a cheap binocular on a rooftop, still a single point. Observed with a backyard 10-cm refractor, the point begins to look a bit oblong. Observed with the 2.3-metre Bappu, two clean, perfectly separated pin-pricks of light with a sliver of dark sky between them.

What changed? The glass got bigger. Why does bigger glass help? Not because it collects more light (it does, but that affects brightness, not sharpness). Not because it magnifies more (a cheap binocular at 50× does not resolve better than a premium one at 50×). The reason is diffraction. Light passing through a circular aperture — any aperture, from the pupil of your eye to the 25-metre dish of the Giant Metrewave Radio Telescope near Khodad in Maharashtra — diffracts. A point source does not image to a point; it images to a diffraction pattern. And when two nearby point sources each produce their own diffraction pattern, the two patterns can overlap so completely that the eye can no longer tell there are two sources at all.

The article on single slit diffraction derived the pattern a narrow rectangular slit makes. This article carries that same principle to the case that actually governs optical instruments — a circular aperture — and works out the precise limit it sets on resolution. The limit is called the Rayleigh criterion, and it explains why astronomers build bigger and bigger telescopes, why the best optical microscopes bottom out at about 200 nm, and why the cameras in the Chandrayaan orbiters were designed around an aperture carefully tuned to the resolution Moon geology demanded.

Why a point source does not image to a point

A telescope or a camera lens has an aperture — a circular opening, of some diameter D. Light from a distant point source reaches this aperture as a very nearly plane wave, because the source is so far away that the spherical wavefronts have flattened to planes. The aperture lets through a circular patch of wavefront. Behind it — at the focal plane, where the image forms — the parallel rays converge to a point, ideally.

But the aperture is itself a diffracting obstruction: a patch of wavefront passing through a circular hole, then converging, does not focus to a perfect point. It focuses to a bright central disc surrounded by faint concentric rings of progressively lower intensity. The bright central disc is called the Airy disc (after the first mathematician to work out the full formula). Its angular radius, measured from the centre of the disc to the first dark ring, is

\theta_\text{Airy} \;=\; 1.22\,\frac{\lambda}{D}

(The numerical coefficient 1.22 is the first zero of the Bessel function J_1 divided by \pi, and it comes from the exact integral for diffraction by a circular aperture. For a rectangular slit the coefficient would be 1.00; the circular geometry of a real aperture gives 1.22.)

About 84\% of the total light from the point source lands inside the Airy disc; the rest is scattered into the outer rings. For a well-focused instrument, this is the best "point image" you will ever get. No lens grinding can improve it, because the Airy disc is not an aberration — it is the Fourier-transform consequence of chopping off the infinite plane wave at the boundary of a finite aperture.

Airy pattern intensity profile for a circular aperture A horizontal axis showing angular position θ in units of λ over D. A bell-shaped central peak (Airy disc) rises to intensity 1 at θ = 0, falls to zero at ±1.22 λ/D (first dark ring), and has two faint secondary peaks at about ±1.64 λ/D and ±2.23 λ/D. intensity $-2.23$ $-1.64$ $-1.22$ $0$ $1.22$ $1.64$ $2.23$ $\theta/(\lambda/D)$ $1$ $0.5$ Airy disc first dark ring first dark ring
The Airy pattern — the image of a point source formed by a circular aperture of diameter $D$. Bright central disc of angular radius $\theta = 1.22\lambda/D$, surrounded by faint concentric rings. The rings are real — on a good night through a high-quality telescope, you can see the first ring around a bright star as a thin halo.

Two nearby sources and the Rayleigh criterion

Now put two point sources into the sky, separated by a small angle \theta_\text{sep}. Each produces its own Airy pattern in the focal plane; the two patterns overlap.

Rayleigh criterion. Two equally bright point sources are "just resolved" when the maximum of one Airy pattern lies exactly on the first minimum (the first dark ring) of the other.

At this separation, \theta_\text{sep} = 1.22\lambda/D — the angular radius of the Airy disc. The combined intensity profile shows two humps separated by a shallow dip; the intensity at the dip is about 73.6\% of the intensity at either peak. The eye, or a detector, can just tell there are two sources rather than one.

A separation much less than 1.22\lambda/D makes the dip disappear — the two merge. A separation larger than 1.22\lambda/D gives a clear two-peak profile. The Rayleigh criterion marks the boundary.

Exploring the criterion yourself

Drag the angular separation below, measured in units of \lambda/D. Below 1.22, the two Airy patterns merge; above 1.22, they separate.

Interactive: two-source resolution via the Rayleigh criterion A plot showing the combined intensity profile of two Gaussian-approximated Airy patterns separated by a draggable angle s, in units of lambda over D. A red point lets you select s; readouts show s and the dip-to-peak ratio that defines whether the two sources are resolved. separation $s$ in units of $\lambda/D$ dip / peak intensity 0 1.0 1.22 2.0 3.0 0 0.736 1.5 2 Rayleigh limit (dip ≈ 0.736) dip/peak$(s)$ drag the red point
The ratio of the intensity at the dip (midway between the two sources) to the intensity at each peak, plotted against angular separation. At small $s$, the dip is as bright as the peaks — there is no dip, the sources are unresolved. As $s$ grows, the dip deepens. At $s = 1.22\lambda/D$ the dip is at $73.6\%$ of the peak; this is the Rayleigh limit, the conventional boundary of "just resolved". For $s \gg 1.22\lambda/D$ the dip approaches zero and the two sources are comfortably separated.

Resolving power of a telescope — why bigger is better

Point a telescope at two stars separated by angle \theta_\text{sep}. The telescope's aperture has diameter D. The Rayleigh criterion says the two stars are resolved when

\theta_\text{sep} \;\gtrsim\; 1.22\,\frac{\lambda}{D}

The resolving power of the telescope is defined as the reciprocal:

\text{RP}_\text{telescope} \;=\; \frac{1}{\theta_\text{min}} \;=\; \frac{D}{1.22\,\lambda}

Three consequences fall out immediately.

  1. Larger D, finer resolution. This is the first-order reason telescopes get built bigger and bigger. The Vainu Bappu Observatory's 2.3-m at Kavalur resolves roughly 20× finer than the 10-cm refractor on a college rooftop.

  2. Shorter wavelength, finer resolution. A telescope tuned to blue light (shorter \lambda) resolves finer than the same aperture in red light. This is why ultraviolet satellites (like ASTROSAT, launched by ISRO) see structures in star-forming regions that optical telescopes blur over.

  3. Radio wavelengths need enormous apertures. The GMRT at Khodad works at wavelengths of tens of centimetres; to achieve sub-arc-second resolution at \lambda = 0.5 m, you need D of order kilometres. The GMRT achieves this by using thirty 45-m dishes spread over 25 km, operated as a single interferometric aperture. The effective D is the baseline between outlying dishes — about 25 km — and the resolution follows the Rayleigh formula applied to that effective D.

Numerical example: the Vainu Bappu Telescope

A star \lambda = 550 nm is observed through the Vainu Bappu Telescope's D = 2.34 m aperture. The resolving angle is

\theta_\text{min} = 1.22 \times \frac{550 \times 10^{-9}}{2.34} \approx 2.87 \times 10^{-7} \text{ rad}

Convert to arc-seconds: 1 \text{ rad} = 206265 arc-seconds, so

\theta_\text{min} \approx 2.87 \times 10^{-7} \times 206265 \approx 0.059 \text{ arc-sec}

In practice, the Earth's atmosphere limits resolution to about 0.5–1 arc-second even for the best telescopes on the best nights — the so-called seeing limit. This is why astronomers build adaptive-optics systems (and why space telescopes exist).

Resolving power of a microscope — the Abbe limit

For a microscope, the object is a tiny specimen centimetres from the objective, not a star at infinity. The geometry is different, and the Rayleigh criterion translates into a statement about the smallest linear separation d_\text{min} between two object points that can be resolved.

Assumption. The objective collects light from the object over a cone of half-angle \alpha. The space between the object and the objective has refractive index n (often air, n = 1; for high-power objectives, immersion oil with n \approx 1.5).

Step 1. Rewrite the Rayleigh criterion for the near-field geometry.

Two object points separated by d subtend an angle at the objective equal to d/f where f is the objective's focal length — but this is not quite the right quantity, because the objective does not image a point to a diffraction-limited point at infinity. Instead, the proper way to think about it: two nearby object points produce two wavefronts that enter the objective within the cone of half-angle \alpha. Abbe's argument (1873) says that for the image to preserve the two points as distinct, the path-difference between the two extreme rays of the accepting cone must differ for the two object points by at least one wavelength.

The geometric path difference between the two extreme rays for a single point object is 2d\sin\alpha. The object is illuminated through a medium of index n, and optical path differences are what matter for interference, so the optical path difference is 2n d \sin \alpha.

Setting this equal to one wavelength for the minimum resolvable d:

2 n d \sin\alpha \;\ge\; \lambda

giving, for the critical condition (self-luminous objects, Rayleigh-style),

d_\text{min} \;=\; \frac{0.61\,\lambda}{n \sin\alpha}

The coefficient 0.61 = 1.22/2 reflects the Rayleigh criterion — the "just resolved" convention — and the factor of 2 comes from the two-sided symmetry of the cone.

Why: the Rayleigh-criterion version of Abbe's argument says the critical path-difference is the path difference that produces the first zero of the Airy pattern, which in the two-source geometry reduces to the critical d above. Abbe's original argument, for coherently illuminated specimens, gives d_\text{min} = \lambda/(2\text{NA}) — the same result with a slightly different coefficient because the illumination condition is different. For CBSE/JEE purposes, the 0.61 form with Rayleigh is conventional.

Step 2. Introduce the numerical aperture.

The combination n \sin \alpha appears so often that it has its own name — the numerical aperture of the objective:

\text{NA} \;=\; n \sin\alpha
\boxed{\;d_\text{min} \;=\; \frac{0.61\,\lambda}{\text{NA}}\;}

Step 3. Compute the limit for typical microscope objectives.

A dry objective can approach \text{NA} = 0.95 (since \sin\alpha can approach 1 and n = 1). With \lambda = 550 nm:

d_\text{min} = \frac{0.61 \times 550}{0.95} \approx 353 \text{ nm}

An oil-immersion objective dips into a drop of oil (with n \approx 1.52) placed on the specimen. Now \text{NA} = n \sin\alpha can reach 1.4–1.5. With \lambda = 550 nm:

d_\text{min} = \frac{0.61 \times 550}{1.4} \approx 240 \text{ nm}

Use short-wavelength light (UV, \lambda \approx 200 nm) and oil immersion, and you push d_\text{min} down to roughly 100 nm. Below this, you need a completely different kind of imaging (electron microscopy, X-ray microscopy, or super-resolution fluorescence techniques like STED and PALM, the topic of recent Nobel Prizes).

This is the famous diffraction limit of optical microscopy. Individual virions (about 100 nm), individual ribosomes (\sim 25 nm), and certainly individual proteins (\sim 5 nm) are beyond the reach of a conventional optical microscope. An AIIMS pathology lab using an oil-immersion microscope can clearly see bacterial cells (1–10 μm) and large structures inside them, but not individual molecules.

Worked examples

Example 1: Resolving double stars with a 10-cm telescope

A backyard refractor has an objective diameter D = 10 cm. On a night of perfect seeing, two stars are separated by 1 arc-second in the sky. Observed at the centre of the visible spectrum (\lambda = 550 nm), will the telescope resolve them?

Two stars through a 10 cm aperture Two stars at angular separation 1 arc-second seen through a 10 cm aperture. Each produces an Airy disc at the focal plane. The computed Rayleigh limit is 1.39 arc-seconds, so the two Airy discs overlap and appear as a single elongated blob. star A star B $D = 10$ cm single blob Airy discs merge
Two stars $1''$ apart seen through $D = 10$ cm. The Rayleigh limit of the telescope is $\approx 1.39''$, larger than the separation, so the two Airy discs overlap into a single elongated blob.

Step 1. Write the Rayleigh formula.

\theta_\text{min} = 1.22\,\frac{\lambda}{D}

Step 2. Plug in numbers.

\theta_\text{min} = 1.22 \times \frac{550 \times 10^{-9} \text{ m}}{0.10 \text{ m}} = 1.22 \times 5.5 \times 10^{-6} = 6.71 \times 10^{-6} \text{ rad}

Step 3. Convert to arc-seconds.

\theta_\text{min} = 6.71 \times 10^{-6} \times 206265 \approx 1.38 \text{ arc-sec}

Why: The conversion factor 206265 arc-seconds per radian comes from 180 \times 3600 / \pi. Arc-seconds are the natural astronomical unit — 1 arc-second is 1/3600 of a degree, about the angle of a 1-cm coin seen from 2 km away.

Step 4. Compare to the 1 arc-second separation.

1.38'' > 1'', so the separation is smaller than the telescope's resolving limit. The stars are not resolved.

Result: A 10-cm refractor cannot separate two stars 1'' apart at visible wavelengths. You would need a telescope with D > 1.22 \times 550 \times 10^{-9} / (1/206265) = 1.22 \times 550 \times 10^{-9} \times 206265 \approx 13.8 cm to just resolve them.

What this shows: The Rayleigh formula sets a hard floor on telescope resolution that depends only on aperture and wavelength — not on magnification, not on eyepiece quality, not on the observer's eyesight. A 100× eyepiece on a 10-cm telescope reveals the same unresolved blob; you cannot magnify your way out of a diffraction limit.

Example 2: Resolving two points on a chromosome with an oil-immersion microscope

A pathologist at AIIMS uses an oil-immersion objective with \text{NA} = 1.4 to look at a stained chromosome. Violet light with \lambda = 400 nm is used. What is the smallest distance on the specimen that the microscope can resolve? How does this compare to the spacing of adjacent base pairs on a DNA strand (about 0.34 nm)?

Microscope cone of light with NA labelled A microscope objective as an inverted triangle accepting a cone of light from a specimen on a glass slide. The half-angle α of the cone and the numerical aperture NA = n sin α are labelled. An oil-immersion layer between objective and slide is shown. microscope slide specimen oil, $n = 1.52$ objective $\alpha$ NA $= n\sin\alpha = 1.4$
An oil-immersion objective accepts a wide cone of light from the specimen. The numerical aperture $\text{NA} = n\sin\alpha$ combines the refractive index of the immersion medium and the half-angle of the accepting cone.

Step 1. Write the Abbe/Rayleigh formula.

d_\text{min} = \frac{0.61\,\lambda}{\text{NA}}

Step 2. Plug in numbers.

d_\text{min} = \frac{0.61 \times 400 \text{ nm}}{1.4} = \frac{244}{1.4} \approx 174 \text{ nm}

Step 3. Compare to DNA spacing.

DNA base-pair spacing: 0.34 nm. 174 / 0.34 \approx 512. So d_\text{min} is about 500 base-pair-spacings.

Why: the factor of 500 says the optical microscope smears 500 consecutive base pairs into one "spot". You cannot resolve individual bases, individual nucleotides, or anything closer than \sim 100 nm apart, no matter how good the objective is — because diffraction is forbidding it.

Step 4. What about using blue-green or UV?

Push to \lambda = 250 nm (a UV objective working through quartz optics):

d_\text{min} = \frac{0.61 \times 250}{1.4} \approx 109 \text{ nm}

Still enormously larger than DNA base spacing. UV microscopy helps for bacteria and organelles, but it cannot reach molecular scales.

Result: d_\text{min} \approx 174 nm for this objective. The microscope can resolve individual chromosome bands (typical width \sim 1 μm) comfortably, and individual mitochondria (\sim 500 nm) well. It cannot resolve individual virions (\sim 100 nm) or individual proteins (\sim 5 nm).

What this shows: The Abbe diffraction limit is the reason biology switched to electron microscopy and, more recently, super-resolution fluorescence techniques (PALM, STORM, STED) — each of which is a clever way of sidestepping the Abbe limit rather than beating it within classical optics. Classical optics bottoms out around 100–200 nm, and the Rayleigh/Abbe formula says exactly why.

Example 3: GMRT resolution at 21 cm

The Giant Metrewave Radio Telescope at Khodad observes 21-cm wavelength radiation from neutral hydrogen clouds in our galaxy. As an interferometer with baseline 25 km, what angular resolution does it achieve? Compare to the Vainu Bappu optical telescope at Kavalur operating at \lambda = 550 nm with D = 2.34 m.

Step 1. GMRT Rayleigh limit.

\theta_\text{GMRT} = 1.22 \times \frac{0.21 \text{ m}}{25000 \text{ m}} = 1.22 \times 8.4 \times 10^{-6} \approx 1.02 \times 10^{-5} \text{ rad}

Convert to arc-seconds: 1.02 \times 10^{-5} \times 206265 \approx 2.1 arc-seconds.

Step 2. Vainu Bappu Rayleigh limit.

\theta_\text{VBT} = 1.22 \times \frac{550 \times 10^{-9}}{2.34} \approx 2.87 \times 10^{-7} \text{ rad} \approx 0.059 \text{ arc-sec}

Step 3. Compare.

Vainu Bappu wins on raw resolution by a factor of about 35 — 0.059'' vs 2.1''. Yet GMRT sees astronomical objects (cold hydrogen gas, pulsars, AGN cores) that are completely invisible to optical telescopes because those objects do not emit visible light. Each telescope's resolving power is competitive within the wavelength range it observes.

Why: the resolving formula 1.22\lambda/D is wavelength-divided-by-aperture. Radio waves are a million times longer than optical light, so a radio telescope needs a million-times-larger aperture to match optical resolution. GMRT's 25 km baseline is a million times the 2.3-m Bappu aperture, which is why the two are within a factor of \sim 35 of each other despite operating in wildly different bands.

Result: GMRT: \approx 2.1'' at 21 cm. Vainu Bappu: \approx 0.06'' at 550 nm. The Bappu sees finer detail in its band; the GMRT sees a completely different sky where cold hydrogen clouds are the stars.

What this shows: Bigger apertures don't just improve resolution — at long wavelengths they enable observations at all. The square-kilometre-array (SKA) being built partly in South Africa and partly in Australia will extend GMRT's idea to baselines of thousands of kilometres, pushing radio resolution below what even the best optical telescopes achieve.

Common confusions

If you can apply the Rayleigh criterion and the Abbe limit to problems, you are done. What follows is the exact Airy-pattern formula, the relationship between image and object spaces, and the super-resolution tricks that sidestep the diffraction limit.

The exact Airy pattern

For a circular aperture of radius a illuminated by a plane wave of wavelength \lambda, the diffraction pattern at the focal plane has intensity

I(\theta) = I_0 \left[\frac{2 J_1(x)}{x}\right]^2, \qquad x = \frac{2\pi a \sin\theta}{\lambda}

where J_1 is the Bessel function of the first kind of order 1. The zeros of this expression are at x = 3.832, 7.016, 10.173, \ldots (the zeros of J_1). The first zero gives \sin\theta = 3.832 \lambda/(2\pi a) = 0.61 \lambda/a = 1.22 \lambda/D, which is exactly the Rayleigh formula. The second zero (first inner dark ring on the outer side of the central disc) is at \theta = 2.23 \lambda/D, and so on.

The Rayleigh criterion is not a hand-waved threshold — it is the first zero of a specific analytic function, predicted by Maxwell's equations for a circular-aperture Fourier transform.

Connection to the Fourier transform

The amplitude pattern at the focal plane of an aperture A(u, v) (where u, v are transverse coordinates at the aperture) is the two-dimensional Fourier transform of A:

U(x, y) \;\propto\; \iint A(u, v) \, e^{-i 2\pi (ux + vy)/(\lambda f)} \, du\, dv

For a uniformly illuminated circular aperture, this integral evaluates to the Airy function. The Rayleigh limit is the first-zero spacing of the Fourier transform of a disc; the Fourier transform of a hair-thin slit (single-slit diffraction) gives a \sin x/x pattern with first zero at \lambda/a; these are the same principle applied to different aperture shapes.

Dawes' limit — and why real observers beat Rayleigh

Dawes (1867, empirical) found that experienced observers can just split equal-brightness double stars at \theta \approx 0.105/D_\text{inches} arc-seconds, which at visible wavelengths corresponds to roughly \theta \approx 0.98 \lambda/D — slightly tighter than the 1.22 \lambda/D Rayleigh criterion. Similarly, the Sparrow limit is even tighter (\sim 0.95 \lambda/D), at the point where the dip between the two peaks exactly vanishes. The three criteria differ because they define "just resolved" differently; Rayleigh is the most conservative and the one taught in introductory physics.

For unequal brightness or noisy data, practical resolution is worse than Rayleigh. For image-stacked or high-signal-to-noise data, it can be better. The Rayleigh formula is a benchmark, not a law.

Super-resolution fluorescence microscopy — beating the Abbe limit

For almost a hundred years after Abbe (1873), the diffraction limit was treated as the final word on optical resolution. Then, starting in the 1990s, a series of fluorescence tricks showed how to sidestep it:

  • Stimulated Emission Depletion (STED) microscopy uses a donut-shaped depletion beam to switch off fluorescence everywhere except in a central sub-diffraction-limited spot.
  • Single-Molecule Localisation Microscopy (PALM, STORM) detects individual fluorescent molecules one at a time and localises each to a few nanometres, then reconstructs the image from tens of thousands of localisations.
  • Structured Illumination Microscopy (SIM) illuminates the specimen with patterned light, aliasing sub-diffraction-limited structure into detectable frequencies.

These techniques achieve lateral resolution of \sim 20–50 nm with visible light — an order of magnitude below the Abbe limit. They earned Eric Betzig, Stefan Hell, and William Moerner the 2014 Nobel Prize in Chemistry. They do not violate the Rayleigh/Abbe formula; they violate the assumption that the object is a fixed intensity distribution. By making the object itself a set of switchable emitters, they turn the classical diffraction limit into a limit on each localisation event, which can then be beaten by statistics.

Radio interferometry and the effective aperture

GMRT's 30 individual 45-m dishes do not collect light as a single 25-km-wide mirror would — they collect amplitudes and phases separately at each dish and combine them afterwards in a correlator. The resolution they achieve is set by the baseline — the longest distance between pairs of dishes — not by the individual dish diameter. The effective aperture of the interferometer is the baseline; the effective collecting area is the sum of the individual dishes' areas.

This is why the Event Horizon Telescope (EHT) that imaged the supermassive black hole M87* achieved the angular resolution of a single Earth-sized radio dish: its baselines are continental, but its collecting area is just the sum of about a dozen millimetre-wave facilities around the world. The Rayleigh formula applies, with D being the baseline.

The seeing limit and adaptive optics

For ground-based optical telescopes, the atmosphere limits resolution to \sim 1'' on a good night — roughly the Rayleigh limit of a 15-cm aperture, regardless of how big the real telescope is. The fix is adaptive optics: a deformable mirror that corrects the wavefront in real time by sensing distortions against a bright reference star (or an artificial laser-guide star). Modern adaptive optics at the European Very Large Telescope in Chile can reach within a factor of a few of the diffraction limit of an 8-m telescope — \sim 0.02'' at near-infrared wavelengths. At Hanle, the Indian Astronomical Observatory's 2-m Himalayan Chandra Telescope is being equipped with similar adaptive-optics capability.

Where this leads next