Eyebrows for a Square Face: the Jaw Corner, the Brow Bend and the Band Between

What is the best eyebrow shape for a square face?

A range, and the range is the answer rather than an evasion. Nine sources with retrievable publication dates give this face a brow between November 2013 and March 2026. Six say soften it with a curve. Two say build it up strong and angular. One says produce elongation with a well arched brow and does not say whether an arch is a curve or an angle, which is the problem in miniature. Of the six that say soften, three immediately take back both ends of soft, and the fourth — the most recent of them, from March 2026 — asks for a brow that is soft but defined, so that it complements the bone under it without looking overly sculpted. Not one of the nine prints a measurement of a brow against this face.

The word soften is doing all the work and none of it. A 2014 waxing guide says a strong and thick brow helps soften a heavy jawline, and that a defined sharp peak at the top of the brow makes the brow appear stronger, which helps create balance. Read that twice. The mechanism offered for softening is sharpening: put a hard corner above a hard corner and the face balances. Six years later a clinic guide says you can either balance a strong square jaw with a strong brow or soften it with curved eyebrows, which is at least honest about there being two moves, and then recommends a thick, angular and well defined shape as the best eyebrow shape for a square face, adding that a softer curve also works if what you want is softening. Both of those sources are using soften and balance as though the words settle something. They settle nothing, and every source that follows inherits the ambiguity without noticing it.

Count what is countable, because the count is the finding. Across nine positions and twelve years and four months, the numbers printed are: a peak landmark given twice and differently — above the outer tip of the eye or above its middle, from a 2021 academy guide; above the outer edge of the iris, from a 2024 product brand — and those are not the same point, since the outer corner of the eye sits well outside the outer edge of the iris. A method for finding the face shape, given twice: trace the outline in a mirror with a lipstick or a liner and compare the widths; or pull the hair back and look at the overall outline. A start and an end for the brow, anchored to the nose twice, by two sources four years apart. A maintenance interval, given once, as touch-ups every few weeks, with nothing said about what the few weeks count from. And a product count: ten shaped stencils in one kit. That last is the only integer any of the nine prints that is not a landmark.

So the direct answer to the heading is that there is no single best eyebrow for a square face. The best eyebrows for a square face are not a shape but a bend with a floor and a ceiling, and neither the floor nor the ceiling has ever been published for this face. What has been published is the floor for brows in general — 135 degrees, with a spread of 9 either side, from people positioning their own ideal rather than being told one — and the corner this face is carrying underneath, at 127.9 degrees on this site's rule set. Put those two together and the band has numbers on it for the first time. That is section four's job, and the arithmetic in section three is what makes it more than a slogan.

One caveat that belongs up here rather than at the bottom, because it changes how much weight the 127.9 can carry. This site publishes, on its about page, that on the calibration photographs behind its rule set the jaw angle did not separate square faces from round ones cleanly: the two groups measured about the same. The ordering that puts this outline's corner at 127.9 and the round face's at 142 comes from the literature, carried into a measured range rather than read off a measurement of you. Two of the five readings this site takes — the length ratio and the chin taper — are digit for digit the values the round face carries. The corner is the reading that decides, and it is also the harder of the two to measure on yourself. That is why the result page reports a margin against all seven outlines and a confidence rather than a single name, and it is why the number below should be treated as the literature's corner and checked against your own.

Soften or echo: the fork that gets named once in twelve years

There are two opposite things you can do to a strong jaw with a brow, and only one of them is usually offered. You can soften it, by putting a curve above a corner so the eye reads a rounded line where the bone turns. Or you can echo it, by putting a hard line above a hard line so the two read as a matched set rather than as a fault. The nine dated positions overwhelmingly offer the first. Exactly one of them, in May 2020, says the second exists: you can either balance a strong square jaw with a strong brow or soften it with curved eyebrows. Having named the fork, it picks the angular arm and moves on. Nobody after it acknowledges that a choice was made.

The 2014 source is the interesting case, because it runs both arms through the same sentence without appearing to. Strong and thick eyebrows can help soften the heavy jawline. A defined sharp peak will make the eyebrows appear stronger, which helps to create a balance of the face. Softening is the stated goal, sharpening is the stated method, and balance is the stated result. Whichever arm you think you are being handed, that paragraph will read as agreement with you. This is the mechanism by which a fork goes unnoticed for twelve years: the vocabulary is shared and the direction is not.

The clearest statement of the echo arm anywhere in the pool is not about eyebrows at all. In a 2025 exchange, a person described their face as blocky — specifically, that the back corner of the lower jaw and the cheekbones sit on the same line as each other, which is this outline's width reading stated without a number — and twenty-one people answered with something to do. One of the twenty-one went to jewellery. The necklace in the photograph was thin and delicate; the advice was to wear something more rigid and chunky, to match the jawline, because matching the jewellery to the lower third of the face makes everything read as less blocky since it all matches. That is the echo argument, made explicitly, with a reason attached, in the same years that nine brow guides were telling this face to soften and never mentioning that the alternative existed.

It is worth being precise about what the echo argument claims, because it is not the opposite of softening in the way it first looks. It says the blockiness reads as a fault when nothing else on the face agrees with it, and as a feature when something does. Under that reading the brow could go either way and the deciding variable is the rest of the outfit, not the jaw. None of the nine considers this. The nearest any of them comes is a 2023 guide's aside that you might look great with sharper brows if your hairstyle is doing something softening to your features — a conditional on the hair, which is the same structure of argument pointed at a different feature, and the only time in twelve years that a source admits the brow is not being chosen on its own.

The fork also explains a disagreement that otherwise looks arbitrary. Two of the nine name public figures as examples of this outline. One 2020 guide lists three; a 2021 guide lists three; the two lists share exactly one name. And that name appears elsewhere in the pool as the illustration of a fault — an arch that starts too far in, so that the section descending after it takes up nearly half the brow instead of the outer third, which is what turns a bend into an angry triangle. The same face is this outline's example and the cautionary tale about its brow, in material written two years apart by people who had plainly not read each other. Neither is wrong about what they are pointing at. They are pointing at different things, and nothing in the record says so.

The brow has a corner, and so does the jaw

Both of them are angles, which is the reason this face's advice can be given a number at all. The jaw's corner — the one below the ear, where the line turns instead of curving — is centred at 127.9 degrees on this outline in this site's rule set, the sharpest of the seven it carries; the guide to this face's glasses works out what that corner does and why a frame is worn across it, and nothing here needs re-deriving. The brow's corner is the one at the peak, between the inner section and the tail. That angle has been measured exactly once in the published record. A July 2017 study asked 60 women to position their own ideal brow with a finger, five photographs each, and defined the apex angle as the angle between the medial portion of the brow — usually, it says, the medial two-thirds — and the tail, the lateral one-third. The mean was 135 degrees with a standard deviation of 9. The paper prints the unit as millimetres, which is a typographical error; its own second table gives 134.66 ± 8.12 and 135.44 ± 9.22 for the two age groups, and those are degrees.

Two things about that number have to be said before it is used. It is not a square-face number: the words face shape do not appear anywhere in the study, so the 60 were never sorted by outline and the 135 is what people asked for, not what any face shape needs. And it is a preference, not an optimum — the method was a finger on a photograph. What makes it usable anyway is that it is the only brow angle in the record with a spread attached, and a spread is what turns a single number into a range. No source in the nine gives this face a bend. This one gives everyone a bend, with an error bar.

The subtraction is this page's own, and no source anywhere in the record makes it. 135 minus 127.9 is 7.1 degrees. Divided by the study's own standard deviation of 9, that is 0.79 of a standard deviation — the published average preference already sits further from the jaw's corner than four fifths of a spread. Expressed as a share of the ±1 SD band, the jaw's corner sits 1.9 degrees inside a band 18 degrees wide, which is 10.6 per cent of it counted from the sharp end. In other words, roughly a tenth of the acceptable band is sharper than your own jaw, and the remaining nine tenths are more open than it. Soften, translated, means: land anywhere in the nine tenths.

Degrees do not tell you what to do with a pair of tweezers, so here is the same thing in millimetres. Take the brow as two straight runs, an inner one and a tail one, meeting at the peak, with the two ends level — which is what every source in the nine assumes without saying, since the tail is forbidden from dropping below the inner end and the level case is the boundary. Take the inner run as two thirds of the brow's length and the tail as one third, which is the split the 2017 study's own definition of the apex angle uses. Call the total length L, measured along the brow from its inner start to the tip of its tail, and the angle at the peak A. Then the peak's height above the line joining the two ends works out to a fixed fraction of L for each A, because the geometry scales: with S standing for 180 − A, the inner run's slope satisfies tan α = sin S ÷ (2 + cos S), and the rise is two thirds of L times sin α.

On a 60 mm brow — 60 divided by 25.4 is 2.3622, so 2.36 in, and the two runs are 40 mm and 20 mm, which are 1.57 in and 0.79 in — the four angles that matter give: 144 degrees, a rise of 8.19 mm, which is 8.19 ÷ 25.4 = 0.3225, so 0.32 in. 135 degrees, 10.11 mm, and 10.11 ÷ 25.4 = 0.3980, so 0.40 in. 127.9 degrees, the jaw's own corner, 11.56 mm and 0.46 in. 126 degrees, 11.94 mm and 0.47 in. The whole ±1 SD band, from 126 to 144, is 11.94 − 8.19 = 3.74 mm, and 3.74 ÷ 25.4 = 0.1473, so 0.15 in. The gap between the measured mean and the jaw's corner is 11.56 − 10.11 = 1.45 mm, or 0.06 in. Every one of those is exact division by 25.4, not an approximation: the international inch has been defined as exactly 25.4 mm since 1959.

The table below runs the same four angles across six brow lengths, because your brow is probably not 60 mm and the fraction is what transfers. To use it on a length that is not in the table, multiply: the rise as a share of brow length is 0.136542 at 144 degrees, 0.168483 at 135, 0.192640 at 127.9 and 0.198925 at 126. On a 52 mm brow at the mean, 0.168483 × 52 = 8.7611, so 8.76 mm, and 8.76 ÷ 25.4 = 0.3449, so 0.34 in. Near the mean the sensitivity is about 0.21 mm per degree on a 60 mm brow — nine degrees either side of 135 buys 1.83 mm going sharper and 1.91 mm going flatter — which is the practical size of the whole argument. It is also why one degree is not worth chasing: no pencil, no stencil and no mirror in the record resolves one degree, and 0.21 mm is under the width of a hair.

Read the fourth column against the second and the whole instruction is 1.45 mm on a 60 mm brow. Read the fifth column against the second and the width of everything anyone has ever called acceptable for this face is 3.74 mm, or 0.15 in. The two angle columns at the edges are ±1 SD around the only brow angle ever measured, and that measurement recorded no face shape, so the edges are this page's arithmetic rather than anyone's finding. Conversions are exact division by 25.4. Lengths are measured along the brow, not straight across, which is why a 60 mm brow spans rather less than 60 mm of face.

The brow shape for a square face is a range, and here are its two edges

Four sources state the edges of the band without ever putting a number on one. A 2021 academy guide bans two shapes in a single breath: thin strings are not allowed, this face will look ridiculous in them, and the same applies to wide horizontal eyebrows, which expand the square and make it rough. A 2025 chain guide bans two more, in the sentence after prescribing the first: a soft curved brow will provide roundness and a lift, and a brow shape that is too rounded or flat should be avoided as it can accentuate a square shaped face even more. A 2023 brand guide bans a third thing — a full semi-circle, which is liable to backfire and clash with your features — and offers four named styles instead, one of which has no arch at all. And a 2025 compilation, which lists its own sources as search results, other people's writing, what the author has seen and heard, and personal experience, states the band in one clause: medium arches prevent the brow from appearing flat or overly sharp.

So the edges are flat on one side and overly sharp on the other, and both are named. What is missing is where they sit, which is the reason four different guides can forbid four different things and none of them can be checked against any of the others. Too rounded, too flat, overly sharp, a full semi-circle, thin strings, wide horizontals: six prohibitions, no thresholds. A reader who has been told to avoid a brow that is too flat and a reader who has been told to avoid a brow that is overly sharp are being given the same instruction from opposite ends, and neither can tell whether the brow in their mirror is inside or outside it.

The band can be numbered from the material in section three, and this is the step no source takes. Take 135 degrees as the centre, because it is the only brow angle anyone has measured. Take ±9 degrees as its width, because that is the measured spread and there is no other candidate. The edges are then 144 and 126. At 144 the bend is very nearly gone — the two runs meet at an angle only 36 degrees short of straight, and on a 60 mm brow the peak stands 8.19 mm above the ends, which is 13.7 per cent of the brow's length. That is the flat edge, the one two sources ban. At 126 the peak stands 11.94 mm above the ends, 19.9 per cent of the length. That is the sharp edge, the one two other sources ban. Between them: 3.74 mm, 0.15 in, eighteen degrees, and every shape name in the nine.

The jaw's corner sits inside that band, and its position is the useful part. At 127.9 degrees it is 1.9 degrees inside the sharp edge and 16.1 degrees inside the flat one — 10.6 per cent of the band counted from the sharp end, 89.4 per cent counted from the flat one. Which means the two prohibitions that look like a pair of separate warnings are one warning with two wordings. Do not echo the jaw and do not go too angular are the same instruction, and between them they rule out a tenth of the band. Everything the record tells this face to avoid on the sharp side is avoided by a single move: open the bend until it is more obtuse than the corner underneath it. On a 60 mm brow that move is worth at most 0.38 mm of peak height, since the sharpest permitted bend is 11.94 mm and the jaw's own is 11.56 mm. It is the smallest instruction on this page and the only one that is load-bearing.

The flat edge is where the record actually disagrees with itself, and the disagreement is resolvable. A 2023 guide offers a flat but upward tilted brow as one of its four styles for this face, with the condition that the overall brow angle still tilts upward, and notes it can work well on a fairly low forehead. The largest body of brow advice in the pool says the same thing from the other direction: even a straight brow should have an upward direction rather than being a horizontal line, because a horizontal line reads as sad on a resting face and looks unnatural, and keeping the tail higher prevents that. Two pools, no contact, one instruction. Flat is available to this face; horizontal is not. What separates them is not the bend at the peak but whether the tail is above, level with, or below the inner end — and the same advice that permits the flat brow forbids the tail dropping, as does a 2021 guide, which warns against letting the tail dip too low because it brings the eye shape down.

That leaves thickness, which the band does not describe and the nine do not settle. Strong and thick, from 2014. Thick, angular and well defined, from 2020. Not too thick but not sparse, medium or slightly larger in width, from 2021. A fuller brow without straight lines, from 2023. Keep the brow looking full, from 2026. And in the pool, one person setting out the actual trade rather than a recommendation: when the brows are thinner the eyes stand out, when they are thicker the two balance each other, so the question is whether you want your eyes to stand out or you want the face to look harmonious. Every one of those is compatible with the band, because the band measures the bend and not the mass in it. Thickness is the setting this face's advice talks about most and numbers least.

Put it on paper before anything comes out. The sheet below carries the two angles side by side, the rise table at six brow lengths with the multiply-it-yourself fractions beside it, the band with both edges and the jaw's position in it marked, the corner-rounding method for the front of the brow, the four no-removal ways to lower an arch, and a stencil log to fill in before anyone touches a hair: Download the square-face brow chart (printable) ↓

Which corner to round, and why the front is the one that reads square

A brow has two corners, and this face's advice never separates them. There is the corner at the peak, which is the one the band in section four measures, and there is the corner at the front, where the brow turns up out of the nose. The second is where the squareness that gets complained about actually lives. A 2023 guide warns against a blocky shape at the front of the brows, especially when filling them in, saying it can look harsh on anyone but especially on this face. A person asking for help in 2016 put the same thing in their own words: they did not like how their brows were more squared off at the beginning, and had stopped filling them in because the result looked too harsh.

The only answer in the entire pool that separates the two corners came from that exchange, and it is one sentence. Keep the bottom inner corner square while rounding the top corner, which softens the expression. Not round the brow. Not remove the corner. Round one of its two edges and leave the other alone, so the front keeps the vertical line that stops the brow looking drawn and loses the right angle at the top of it. The same reply adds a second move with the same structure: train the hairs to stay up and do not trim them, so the wispy ends extend past wherever you fill in, which softens the outline by blurring it rather than by removing anything.

This is worth separating from the phrase people actually type, because the two mean opposite things. Square shaped eyebrows describes a fault on this face, not a goal — a front that turns a corner instead of curving, which is what the 2023 guide and the 2016 asker are both describing. A square arch eyebrows is a different thing again: an arch whose bend is a corner rather than a curve, which is the fault the pool's largest body of advice illustrates with a brow that gets suddenly bigger around the arch instead of holding an even thickness, so that the arch ends up thicker than the inner corner and the whole shape reads as a semicircle. Front corner and peak corner. Two faults, two fixes, and the record talks about them as though they were one.

The semicircle is the clearest case of the two pools contradicting each other on a single shape. A 2021 academy guide prescribes one: highlighted cheekbones and forehead should be softened with a semicircular, soft arch. A 2023 brand guide forbids it: do not go for a full semi-circle, it is liable to backfire and clash with your features, and a fuller brow without straight lines with just a hint of rounding at the arch works best. And the largest body of advice in the pool uses it as the name of a mistake — the brow that thickens at the arch until it looks like a semicircle. Three positions, one shape, and the difference between them is thickness rather than curvature. A semicircular arch drawn at an even thickness is what the 2023 guide means by a hint of rounding; the same outline drawn with the mass piled at the peak is what the other two are describing. None of the three says so.

There is a usable test for whether your own bend is a corner or a curve, and it comes from the pool rather than the nine. A person relayed what a technician told them at a counter: the arch counts as arched if it is still visible when the brows are resting, unbrushed and unset. That is a check you can run in a mirror in two seconds, before and after anything, and it is the only criterion in the record for the thing the band measures. Brushed up and set with a gel, almost any brow reads as arched; at rest, only the ones whose bend is in the hair rather than in the styling do. If your arch survives being left alone, the peak corner is already a curve and the front corner is the one to work on. If it does not, the reverse is true, and the reversible methods in the next two sections are the ones that apply.

One more instruction that only makes sense once the two corners are separated. A 2021 guide says the tip can be made a little longer than usual and that short lines should be avoided; a 2025 compilation says to taper the tails gradually instead of cutting them abruptly. Both are about the outer end, and both are about how the bend finishes rather than where it peaks. On a brow whose front corner has been rounded and whose peak has been opened, the tail is the only remaining place a hard line can appear, and a tail that stops abruptly puts one back. Length here is not the length of the brow along the face — it is the length of the taper at the end of it.

The top edge is the lever on this face, not the forbidden zone

Elsewhere in this subject the top of the brow is where you are told not to go. On this face one of the nine tells you to go there first. A 2023 guide describes a tail-arched style, in which the arch starts closer to the tail than to the centre, and the method it gives for reaching it is filling in from the top and tweezing minimally at the bottom of the arch. The shape is built downward from above with a pencil and almost nothing is removed from underneath. It is the only published method in the nine that says which edge to work on, and it says the top one.

The pool arrives at the same place by removal, from three directions independently. Take hair off the top of the arch and off the tail to flatten the bend. Have the very top of the arch taken off by a technician to shave down a harsh curve, on the reasoning that taking height off the top reduces the severe look. Take hair off the top arched part and fill in below the brow at the same time — the two moves together, which lowers the peak and keeps the mass. A fourth reply describes the same result reached by shaving rather than plucking, and reports that the arch ended up looking at least half as high.

And there is a version that removes nothing at all, which is the one to try first and the one with the most detail in it. Set the brow with a wax, then comb the front section upward rather than to the side — that alone reduces the arch angle. Then when adding colour, put more emphasis on the top front portion of the brow and less on the bottom, going as far as adding colour to the skin above the brow, which makes the arch read gentler. Three moves: direction of the set, placement of the colour, and a line drawn above the existing outline. All three wash off. The reply that describes them also says it could not attach a picture, which is the reason the description is unusually precise — it had to be.

The ceiling matters more here than the methods, because opening a bend has a limit that closing one does not. One reply in the 2022 exchange about a high arch states it flatly: there is nothing to make them straighter other than to draw them in at the arch, and the alternative would be a cosmetic procedure along the lines of implanting hair into a beard or a hairline, which is a bit extreme. That is the whole of the non-surgical range. You can lower the appearance of an arch by setting, by colour, by removing from the top, or by filling below — and once those are exhausted, the arch is bone and hair and it stays where it is. Anyone promising to flatten a strong natural arch into a straight brow by removal alone is promising something the record says cannot be done.

The cost side is stated twice, by two people in two different years, and the two agree. Shaving is safer than plucking, because shaved hairs grow back and plucked ones can stop; the person who wrote that had plucked their own ends, was growing them out, and reported one brow looking a bit sad as a result. A commenter in the largest exchange in the pool gives the same rule for the same reason, and adds the motive: styles and faces change over time. Nobody in the nine mentions reversibility once. The only maintenance figure any of them prints is touch-ups every few weeks, which tells you how often to go back and not what happens if the shape you went back for was wrong.

Which is the reason the top edge is a lever rather than a zone on this face. The instruction from section four is to open the bend, and opening a bend means either taking height off the peak or adding height below it — the top edge and the bottom edge, and only one of those two moves is undoable in an evening. Work from the top with colour first, since colour is the reversible half of the same move. If the result is right, the removal version of it is a small, bounded amount of hair in a place you have already confirmed. If it is wrong, nothing has come out and the next attempt costs the same as the last. The order is not caution. It is the only sequence in which the third attempt is as informative as the first.

Why the reversible route is the only one that tells you anything here

There is a named failure of judgement in this subject and two people in the pool use the term for it independently. Eyebrow blindness. One of them lists it among the reasons a decade of attempting a winged liner had not worked, alongside a hooked nose, high cheekbones with a low jawline, bronzer that goes muddy, concealer that makes them look half dead, and a full face that ends up worse than no face. The other uses it as an explanation and a remedy: it is real, almost everyone who tries a bolder brow panics at first and thinks it looks bad, so get input from someone else before deciding. The word describes a specific deficit — you cannot see your own brows accurately, because you have seen them every day for years — and it is fatal to any method whose only feedback is a mirror.

It is especially fatal on a band this narrow. The whole width of acceptable bend for this face is 3.74 mm of peak rise on a 60 mm brow, and the move that matters most, from the jaw's corner to the measured mean, is 1.45 mm. A judgement instrument that cannot resolve a change of that size, operated by someone known to misread their own result, is not going to produce a reliable answer in one sitting. Which is why the sequence in the previous section is not a preference.

The pool contains three ways to see the shape before committing to it, and the nine contain one. An overlay: there are filters that lay a guide over your face, and one person reports that it almost always looks too far, too thick or too arched on screen until you actually do it, at which point it looks right — so the overlay is calibrated for the wrong thing but is still the only way to see four shapes in a minute. A stencil: a reply to someone unsure of their own face shape suggests getting stencils and trying a few shapes to see what suits, which is the same method with plastic in it; a 2024 product guide recommends a kit of ten shaped stencils, and that integer is the only countable number in the nine. And a trace: two of the nine give a method for establishing the outline of the face itself — pull the hair back, look at the outline in a mirror, compare forehead, cheekbone and jaw widths; or take a lipstick or a liner and trace the face with an imaginary line to measure the width between the widest points. Neither trace is for the brow. Both are for the thing under it.

The strongest argument for the reversible route in the whole pool is not about brows. The person who asked the blocky-face question in 2025 said, in passing, that they expected their own features to change over the following months and would revisit the jaw afterwards. That is the situation the shape-name advice never accounts for: a decision made against a face that is a moving target. And the nine make it worse without meaning to, because one of them conditions the brow on the hair — you might look great with sharper brows if your hairstyle is doing the softening — which means a haircut can invalidate the brow you chose on the strength of a guide.

There is a further reason, and it is the one the record keeps stumbling over. The largest body of advice in the pool declines to organise itself by face shape at all, on the grounds that the shape is not the variable that decides; it organises itself by distance from the eye, by thickness, by where the brow starts and ends, and by the tail. A 2026 magazine guide reaches a similar place by a different route, quoting one artist's rule that brows should follow bone structure and natural growth patterns, not trends, and another's that the brow bone acts as a natural blueprint for where the brow should live. Two pools, opposite methods, same conclusion: the thing to read is your own hair and the bone under it. That is a reversible-route argument, because bone and growth pattern are the two inputs you can inspect without removing anything.

So the honest version of the instruction is procedural rather than shaped. Overlay the four candidate shapes and pick two. Set the hairs upward and see whether the arch is still there at rest. Draw the front corner's top edge round with colour, leave its bottom edge square, and look at it for a week. Only then take hair out, only from the top, and only as much as the drawing told you. On a band 0.15 in wide, judged by an eye that is documented to misjudge its own brows, on a face that may be changing, that sequence is not the cautious way to do it. It is the only way that produces an answer.

Where the advice went instead: hair, colour, jewellery and a clinic

Twenty-one people answered a face described as blocky, in June 2025, with something to do. The breakdown by category, counting mentions rather than people, since several replies span three: hair, six — curtain bangs, face-framing layers that start closer to the jaw instead of at the cheekbone, a curve put into both sides, volume at the roots, a chin-length bob, a side part. Where to put colour, nine — contour the jaw and the sides of the forehead to cast shadows that make the face read rounder; blush on the apples blended up toward the outer corners of the eyes, because that direction elongates the face; blush taken up the temples toward the side of the brow; highlight the middle of the nose; curl the lashes; wear lipstick. Injectables, five — four separate replies naming the masseter muscle, one naming chin and mid-face filler. Someone else's face to copy, three. Jewellery, one. Eyebrows, four mentions across four different replies, a line each.

The four brow mentions are worth reading in full, because they are the whole of what this exchange had to say about the feature this page is about. Choose an eyebrow pencil in a similar colour to the brows and fill them in slightly, to draw the eye up and pull focus. Use a brow pencil to style them with more of an arch. An eyebrow fixer, alongside a hairstyle with volume at the roots. And, from the reply that also suggested the masseter muscle and the blush placement, that layers and defined brows genuinely enhance this face shape — accompanied by a retouched photograph showing what the adviser meant. The asker looked at it and said they were not sure they liked that look, but would think about it. Three of the four are about colour and one is about shape. None is about the bend, the band, or the corner.

That distribution is the finding, and it is not a criticism of the people answering. A brow is one line among twenty-one because it is the smallest lever in the set: hair covers the outline, colour moves the widest point, an injection moves the bone's appearance, and a brow moves 3.74 mm of peak height. The same exchange makes the ordering explicit without meaning to — the reply about the necklace argues that the lower third of the face should be matched rather than softened, and the reply about the blush argues that the direction of a stroke changes the face's apparent length. Those are levers with visible travel. The brow's travel is under a sixth of an inch.

What the pool has that the nine do not is any acknowledgement of this. Not one of the published sources compares the brow to the other things that could be done instead, and not one says how much of the result the brow is responsible for. The nearest is the 2023 guide's conditional on the hairstyle, which at least admits that another feature can be doing the work. The nearest in the pool is the asker's own summary of their routine — lash extensions, filling in the brows, lipstick — and their report that every time they attempted a full face they ended up looking worse than before they started. Twenty-one answers to that, and the four about the brow are all additions.

Two of the nine do route the reader elsewhere, and both routes are worth naming because they are not brow advice. A 2025 chain guide's suggestion for this face is a service: have the brows done and tinted. A 2023 guide's suggestion is a method for finding the face shape, then a referral to eyebrow mapping for the start and end points — in the same document that says measuring is less helpful here, since a square face is all about the angles. That is the measurement contradiction in one page: the outline is not to be measured, the brow within it is. Whether a technician should then choose the shape is argued in the two pages this one links to at the foot, and the pool is split on it in the same years — one compilation in 2025 recommends getting the shape done professionally and then maintaining it, while the largest exchange in the pool, two years earlier, advises against leaving the choice to a professional on the grounds that they do the same thing to everyone or take the easiest option given what is already there.

Square and oblong are the same jaw at a different length, and the brows are opposite

The nine contain a paired instruction, stated twice, and it is the sharpest thing in the record about this outline. A 2026 magazine guide describes the rectangle — also called the oblong — as similar to a square face but noticeably longer than it is wide, with the forehead, cheeks and jawline typically the same angular width. For the square it recommends soft angles, a gentle arch to counterbalance the angular jawline, shaped lifted with the tail tapering naturally and the brow kept full. For the oblong, in the same article, it recommends a flatter, more horizontal brow to shorten the face visually and balance the width. A 2025 chain guide compresses the pair into one sentence in its questions section: flat brows suit longer faces, while curved or angled brows balance heart- and square-shaped faces.

This site's own numbers say what separates the two, and it is one reading. The length ratio — face length against cheekbone width — is centred at 1.103 for this outline and 1.261 for the oblong, which on a spread of 0.048 is 3.29 standard deviations apart. Every other reading is close or identical. The forehead ratio is 0.951 for both, exactly. The jaw ratio is 0.809 against 0.803, which is 0.23 of a standard deviation and nothing. The jaw angle is 127.9 against 133.8, or 1.31 standard deviations — real, but a third of the length gap. So the two outlines that are handed opposite brows differ decisively on one measurement, and on the width of the jaw they are the same face.

A 2021 academy guide makes the pair and then declines to resolve it. Its rectangular row asks for a wide, straight brow, positioned not too close together, with the hairs near the nose taken out and the tail moderately thin, and says high arcs and sharp bends are completely inappropriate. Its square row asks for a semicircular soft arch with the bend raised high. Then, one paragraph later, it says that owners of a rectangular face, as well as a square one, should visually soften the geometry and should not choose sharp corners in the brow arch. The two faces share an exclusion and are given opposite shapes, and the guide does not say which of the two instructions wins if your measurements fall between them.

One source documents a person who falls between them, and it is the most useful thing in the nine. A salon guide from 2013 — the earliest row in the table, re-dated a decade later — asks out loud whether a particular public figure has a square face or a long one, decides that a prominent chin adds elements of both, settles on square as the dominant type, and then notes that the photographs show her wearing her brows two ways: shaped for a square face with a prominent arch at some times, and for a long face with the arch less prominent at others. The guide states its own rule in the same paragraph — the eyebrow shape for a square face must produce some sort of elongation — which is precisely what the oblong rows above forbid. One face, two brows, chosen by which of its two readings the stylist was working to that day, in a document that never notices the collision.

For anyone near that boundary, which the numbers above suggest is a wide region rather than a thin one, the practical consequence is that the brow cannot be chosen from the shape name. The name is decided by a ratio, and the ratio is decided by two tape measurements: face length from the middle of the hairline to the tip of the chin, and cheekbone width across the highest points. At 1.103 you are on this page's instruction. At 1.261 you are on the opposite one. In between, the honest answer is the one that salon guide gave by accident — that both are defensible, that they produce different brows, and that the reversible route from section seven is what lets you find out which you prefer without having decided correctly first.

There is a second pairing in the record and it points the other way, which is why it belongs here. The same 2026 guide that separates this outline from the oblong also quotes an artist saying that square faces often have strong bone structure, which can support a beautiful brow, and that she keeps the brow soft but defined so it complements the natural structure without looking overly sculpted. Support and soften, in one recommendation: the bone is an asset and the brow should not compete with it. That is neither arm of the fork in section two. It is a third position — echo the structure with definition, soften the corner with the bend — and it is the most recent thing in the record.

Brows for a square face, in the order the advice was given

Nine rows, one per dated position, oldest first, spanning 10 November 2013 to 23 March 2026. The first column gives the date the source itself prints where it prints one, with the machine-readable date alongside when the two differ; the ordering follows the printed date, which is why the first row carries a re-dating ten years after its original. The second column describes what kind of source it is, without naming it, which is this site's convention for comparison tables. The third is the instruction this face is given, in the source's own terms. The fourth is the one to read down: whether the source printed a number.

Read the fourth column and the shape of the record appears. Nine positions across twelve years and four months, and not one of them contains a measurement of a brow against this face. Two print a landmark for the peak, and the two landmarks are not the same point. Two print a method for finding the outline of the face, one by tracing in a mirror with a liner and one by pulling the hair back and comparing widths. Two anchor the brow's start and end to the nose, four years apart, in the face of an argument — made in between them, in the largest body of advice in the pool — that the nose is the wrong reference because a wide nose sets the brows too far apart and a narrow one too close together. One prints an interval, every few weeks, without saying what it counts from. The only integer in the column that is not a landmark is the count of stencils in a kit.

Read the third column and what moves is the adjective, not the reasoning. Strong and thick with a defined sharp peak, in 2013 and 2014. Thick, angular and well defined, in 2020, from the one source that names the alternative. A higher, more curved arch, in 2021. A semicircular soft arch with no sharp corners, in the same year. A soft medium arch, with a higher one permitted if the forehead can take it, in 2023, alongside three other styles including one with no arch. Three shapes at once — softly arched, rounded, angled — in 2024, with the caveat that the angled one should not be too harsh. A soft curved brow, but not too rounded and not flat, in 2025. Soft angles, soft but defined, not overly sculpted, in 2026. The word swings from sharp to curved to angular to soft to all three at once and back to soft, and at no point does anyone measure anything.

Nine rows, 10 November 2013 to 23 March 2026 — 4,516 days, or twelve years, four months and thirteen days. Where a source prints a date that differs from its machine-readable one, both are given and the printed one decides the order; that is the first row, whose own text supplies an original date a decade before the one in its markup. Four sources were retrieved and excluded rather than used to pad the count: one carries no date of any kind, one returned an access-error page instead of its article, one returned a 403, and one is an image collection with no publication date. Sources are described here by what they are and never named or linked, and the fourth column records only what each source printed itself — never this page's estimate.

The best brow shape for a square face, and what the record never settled

Open the bend past 127.9 degrees and you are inside the band; the measured mean sits at 135 and the band runs from 126 to 144. On a 60 mm brow that is a peak rise between 8.19 mm and 11.94 mm — 0.32 in and 0.47 in — with the jaw's own corner at 11.56 mm, a tenth of the way in from the sharp end. The entire width of everything anyone has called acceptable for this face is 3.74 mm, or 0.15 in. The move that matters is 1.45 mm, and it is a floor rather than a target, since nine tenths of the band lies on the open side of it.

That is the answer the nine will not give, because none of them measured. Ask any of them for the best brows for a square face and what comes back is a direction — soften — with four prohibitions around it, and the direction is right while the prohibitions are unenforceable, since not one of them has a threshold attached. The reason the band can be numbered at all is that somebody measured a brow angle once, on 60 women who were never sorted by face shape, and reported a spread with it. Every number on this page comes from a source that was not looking at this outline. The subtraction between the two angles, the position of the jaw inside the band, and the rise table are this page's arithmetic, and they are the whole of what the record could not be bothered to do.

Six things remain unsettled, and naming them is more useful than resolving them, because none of them is resolvable from inside the record. Which arm of the fork you are on: named once in twelve years and four months, by a source that then chose the arm the other six forbid. Which corner is at fault: the front one, on the only evidence in the pool, and no published source separates the two. Whether the top edge is available: one of the nine builds the shape from above, three people in the pool lower an arch by removing from above, a fourth does it without removing anything, and nobody states the cost of being wrong. Where the peak goes: two landmarks, three positions, no agreement. How thick: five adjectives and one trade-off, and the trade-off is the only honest statement among them. And which way you want the face to read, since two entries in the pool describe a higher, thinner-tailed brow as feminine and a flatter, straighter, medium-thickness one as masculine, with one of them noting that covering the brows and repainting them higher is a technique for exactly that reason — an axis no published source in the nine has any vocabulary for.

Four cases are not addressed anywhere, in the nine or in the pool. Brows that are faint rather than wrongly shaped, where the problem is colour depth and the only colour-first answer in the record came in 2016 to someone who never named a face shape: lighten them with a brow gel a shade toward soft or ash brown. Sparse brows, which none of the four named styles in the 2023 row assumes and none of the prohibitions accounts for — a ban on thin strings is no use to someone whose brows are already thin. Structural asymmetry at the peak, which the two conflicting landmarks make worse rather than better, since on an asymmetric face the outer tip of the eye and the outer edge of the iris give two different answers on the two sides. And the view the shape is being judged in: one person in the pool reported that their arch looked fine from the side and wrong from the front, and not one of the nine says which view it is prescribing for.

What the record does settle, narrowly: flat is permitted and horizontal is not, from two pools with no contact between them, and the difference is whether the tail is above or level with the inner end rather than anything about the peak. That is a real consensus, it is checkable in a mirror, and it is the only one in twelve years and four months. Everything else is a direction, four prohibitions and a tenth of a band wide enough to argue about.

• Find your own corner before comparing it to 127.9. The angle below the ear, where the jaw line turns rather than curves, is the reading that decides this outline, and it is also the one this site says does not separate square from round cleanly on the photographs behind its rule set. Take a photograph straight on with the hair back, mark the corner, and measure it — or use the detector, which reports a margin against all seven outlines rather than a name, because the corner is genuinely hard to read.
• Measure the brow along itself, not across the face, and put the length into the table. Inner start to the tip of the tail, following the curve. On a 60 mm brow the band is 3.74 mm of peak rise and the jaw's own corner is 11.56 mm; on a 52 mm brow multiply the fractions — 0.136542, 0.168483, 0.192640, 0.198925 — by 52, and divide by 25.4 for inches. If your brow is not in the table, the fractions are, and they scale exactly.
• Round the top of the front corner and leave the bottom of it square. That is the only instruction in the record that separates the brow's two corners, it is one sentence long, and it needs no hair removed: it is a filling decision. Check it against the other fault first — a brow that thickens at the arch instead of holding an even thickness reads as a semicircle, and a semicircle is prescribed by one source, forbidden by another and used as the name of a mistake by a third.
• Test the bend at rest. Leave the brows unbrushed and unset and look: if the arch is still visible, it is in the hair and the front corner is the one to work on. If it is only visible once the hairs are brushed up, the bend is styling, and the reversible methods apply to the whole shape rather than to one corner.
• Lower an arch in this order, and stop as soon as it looks right: set the hairs and comb the front upward rather than to the side; put colour on the top front and less on the bottom, including on the skin above the outline; then, only if you still want less, take hair from the top of the arch. There is nothing beyond that short of drawing the arch in, and the record says so in terms. Shaving is the reversible version of the third step; plucking is not, and one person in the pool reports a brow that never came back even.
• Decide which way you want the face to read before deciding the bend, because the record will not decide it for you. Sharpen and the jaw reads as a matched set. Open the bend and the corner reads as softened. Thinner and the eyes stand out; thicker and the two balance. Higher and thinner-tailed reads one way, flatter and medium reads another, and no source among the nine admits the second axis exists.