Straight Eyebrows: Two Measurements Before a Face-Shape Rule

Two readings behind the word straight

The brow has two ends and, usually, one highest point. The inner end is where the hair starts beside the nose. The tail tip is the outer end. The peak is the highest point of the upper edge between them. Put a horizontal line through the upper edge at the inner end and two vertical distances fall out: how far the peak rises above that line, and how far the tail tip sits above or below it. The first is the arch rise. The second is the tail delta; a negative value means the tail ends lower than the inner end.

Two more readings belong beside them. Brow length runs along the brow from the inner end to the tail tip, following the hair rather than a straight chord. Lift is the vertical gap from the upper lash line to the underside of the brow at its narrowest point. Lift is not arch rise. A brow can be almost flat and sit high above the eye, or carry a clear peak and sit close to it.

None of these readings has a published cutoff that turns it into the word straight. The clinical study from 1994 measured eyebrow height but used a different eyelid reference. The 2019 image study measured curvature in normalized pixels. The two 2026 guides read for this page publish no millimetres or degrees at all. So the table below is a method, not a standard: it says where to put the ruler, not what number makes a brow straight.

A worked example makes the difference between the readings concrete. On a 58 mm (2.28 in) brow, a peak 4 mm (0.16 in) above the inner end is an arch rise of 4 mm; a tail 2 mm (0.08 in) below the inner end is a tail delta of minus 2 mm; a 12 mm (0.47 in) gap from the lash line is a lift of 12 mm. The four numbers describe four different things, and changing one does not necessarily change the others. The next section turns the first two into ratios so that brows of different lengths can be compared.

The same three millimetres on three different brows

A raw millimetre reading does not travel between faces. Three millimetres of arch rise on a 52 mm brow is a different proportion of the brow than three millimetres on a 64 mm brow. Dividing each vertical reading by brow length gives a number that can be compared across faces, which is why the method uses ratios rather than a single millimetre target.

The arithmetic is simple enough to do on a phone. Divide the arch rise by the brow length and multiply by 100 for a percentage. Do the same with the tail delta, keeping the sign. The table prints the result for three brow lengths and three arch rises. It is calculated for this page from those lengths; it is not a published norm.

The same four millimetres can read as a gentle rise on one brow and a pronounced one on another. Four millimetres is 7.7 percent of a 52 mm brow, 6.9 percent of a 58 mm brow and 6.3 percent of a 64 mm brow. That is a difference of about a fifth of the reading, which is larger than the precision of a hand-held ruler. The ratio does not make the measurement exact; it makes two measurements comparable.

What this means for you: write the brow length beside the arch and tail figures. A 4 mm rise on a 58 mm brow and a 4 mm rise on a 64 mm brow are not the same reading, even though the ruler shows the same number.

The same division works for a tail that drops. A 2 mm drop is 3.8 percent of a 52 mm brow, 3.4 percent of a 58 mm brow and 3.1 percent of a 64 mm brow. A one-millimetre error in either vertical reading moves the ratio by 1.9, 1.7 or 1.6 points respectively, so the percentages should be rounded to the nearest whole number rather than reported to a decimal place. The lift can be divided by brow length the same way, but the result is descriptive: no study read for this page gives a target range for that ratio.

Straight, flat, soft-angled: what the labels share

The labels are not a taxonomy. The March 2026 guide calls the brow it recommends for a rectangle face a flat brow and describes it as flatter and more horizontal. The September 2026 mapping guide calls the same idea a horizontal line for a long face. Neither uses the word straight as a measured category, and neither prints a threshold. A reader can be forgiven for treating the three words as synonyms; in practice they overlap without matching.

A working set of distinctions helps more than a dictionary. These are descriptions of what the words usually point at, not standards shared by the sources. The label that matters at a mirror is not the name but the reading you are changing. Moving the peak outward changes the position of the arch without changing its height much. Drawing the tail higher changes the tail delta without touching the arch. Raising the whole brow changes the lift while leaving both other readings alone.

What the words share is a line; what they hide is the lift. That is why the same brow can be called straight by one guide and flat or soft-angled by another without either guide being wrong about the hair. The four readings in section one are the way out of the vocabulary: they describe the line, the tail and the gap above the eye separately, so a change in one does not get reported as a change in another.

How to take the readings at home

Work in even light with a mirror at eye level. Take the reading before brow makeup if you want the line the hair actually grows; take it after makeup if you want the line you draw. Write down which one you took. A filled brow reads thicker and can change the apparent arch, so the two readings are not interchangeable.

Mark three points on each brow: the inner end, the peak and the tail tip. For brow length, run a flexible tape or a strip of paper along the hair from the inner end to the tail tip, mark the end, and lay the strip against a ruler. For the two vertical readings, hold a ruler or the edge of a book level through the upper edge at the inner end, then measure straight up to the peak and straight up or down to the tail tip. For lift, measure from the upper lash line to the underside of the brow at the narrowest point. Take each reading twice, on two days if possible, and keep the average.

The two common errors are head tilt and pressure. Tilting the head back raises the apparent tail on both sides; tilting it forward lowers it. Pressing the ruler into the brow flattens the hair and changes the upper edge. A photograph adds a third error: a camera above eye level makes the tail look lower, and a camera below eye level makes it look higher. Keep the camera at eye level and the head in the same position for both sides.

A worked example shows how much precision is available. On a 58 mm brow, an arch rise of 4 mm is 6.9 percent, a tail delta of minus 2 mm is minus 3.4 percent, and a lift of 12 mm is 20.7 percent of brow length. A one-millimetre error in the arch or tail moves the ratio by 1.7 points. Round the results to whole percentages and keep the millimetre readings beside them. The method does not need a laboratory; it needs the same method on the second day.

What this means for you: the useful record is four numbers and a note about makeup, not a shape name. A name can be applied to two different brows; the four numbers cannot.

Who a straight brow suits, read from the brow bone and the eye

The face-shape rule is the most common answer, and it has more evidence than the youth claim. In a 2007 study, 78 raters looked at five model faces morphed into round, square, oval and long shapes. The brows were removed digitally and drawn back in twice: once by a makeup artist working to the face shape, and once to the classic ideal. The modified brow won for the square face in 62.7 percent of the raters and for the long face in 58.7 percent, both statistically significant. The oval and round faces showed no significant difference. The paper suggests that a flatter brow may give a long face an illusion of fullness.

That is a preference result on morphed photographs, not a measurement of what a real face can hold. The March 2026 guide still assigns a flat brow to a rectangle face and a soft high arch to a round one, which keeps the rule alive; the same article quotes a brow artist saying that brows should follow bone structure and natural growth patterns, not trends. Both statements appear on the same page because they answer different questions: one is about the outline, the other about the material.

The brow bone is the material. A flat brow bone gives little natural arch to work with; a prominent one casts a shadow that can look like a peak even when the hair is flat. The bone does not change when a face-shape label changes, and a label cannot tell you where the bone is. That is why the five face-specific pages on this site each begin with the hair and the bone before they discuss the outline.

Lift is the second input, and it has a measured perceptual effect. In the 2015 experiments, straight and arched brows were shown at five inclinations. The eyes were perceived to tilt in the same direction as the brows. Lower brows made the eyes appear larger; higher brows made them appear smaller; the curvature of the brow had no effect on the perceived size of the eye. The study measured an illusion, not a verdict on beauty, but the direction is useful: raising a brow changes the eye area even when the line of the brow stays straight.

Eye spacing is the third input, and it is where the standard map is weakest. The three-point map keys the inner start to the side of the nose, the peak to the pupil or iris, and the tail to the outer corner of the eye. Nose width varies more than eye spacing does, so a wide nose pushes both starts outward and a narrow nose pulls them in. The map is a convention, not a measurement of the bone. Use the inner corner of the eye as a cross-check before moving the start.

The face outline is the fourth input, and it is the only one a face-shape label can describe. It changes the contrast between the brow line and the rest of the face; it does not move the brow bone, the lift or the eye spacing. A straight brow can suit a long face because the horizontal line answers the vertical outline, and it can suit a square face for a different reason: the flat line leaves the jaw as the dominant angle. The same brow on a round face may read as a second horizontal line where the outline already supplies softness. None of those outcomes is a rule that holds for every face.

The first three inputs are measured on the face; the fourth is a description of the outline. A face-shape rule can summarize the fourth but not the first three. That is the limit the published preference studies leave open: they changed the drawn brow and asked which version the raters preferred, not which version the face could grow.

What the published record actually measured

The table collects seven dated sources: one clinical measurement, two preference experiments, one perception experiment, one image-database study and two consumer guides. The units are not the same. One row reports millimetres, two report shares of raters, one reports normalized pixels, and the two 2026 rows report words rather than numbers. Read the third column as what each source published, not as a common scale.

The span runs from February 1994 to September 2026, thirty-two years and seven months. Sorting by the date each source was published makes the stable parts and the moving parts easier to see than sorting by the shape each one recommends.

The 1994 eyebrow height is not the arch rise defined above: it uses a different eyelid reference and was taken on 143 subjects in a clinical setting. The 2007 percentages are shares of raters, not shares of a population. The 2019 bending degree is in normalized pixels and comes from a model-generated preference score, not a fresh human panel. The two 2026 guides are consumer publications; neither reports a sample, a standard deviation or a repeat-measurement error. Because the units differ, the table does not rank the rows or average them.

What changed, what did not

Three things changed across the thirty-two years. The instrument did: a 1994 clinical series measured brow height with calipers against eyelid landmarks, a 2019 image study measured bending degree in normalized pixels, and the 2026 guides describe a three-point map of start, peak and end with no number attached. The evidence base changed too: the 2007 experiments asked panels of raters to choose between drawn variants, while the 2019 study let a trained model score 300 photographs, so its score depends on the model's preferences rather than a room of people. The vocabulary multiplied as well: flat, horizontal, straight and soft-angled now circulate as separate recommendations.

What did not change is more useful. None of the sources read for this page publishes a cutoff in millimetres or degrees that separates a straight brow from an arched one, and none reports a repeat-measurement error for a brow reading, so the size of a difference that counts is unknown. The face-shape assignment that the 2007 study tested, and that both 2026 guides still print, has not acquired a larger sample or a second outcome measure in the nineteen years since; it remains one experiment with 78 raters and five morphed faces.

The stable part is the geometry, not the advice. A brow has an inner end, a peak and a tail; the vertical distance from the inner end to each of the other two can be read on any photograph that includes a scale, and both are divided by brow length. The 2026 guides describe the same three points and the studies measure subsets of them; none of the sources contradicts that geometry. What moves is which reading a guide decides to reward, and for which outline. Treat the geometry as the measurement and the assignment as a convention, and the record reads as a sequence of preferences rather than a sequence of discoveries.

What a straight brow cannot do

It cannot be guaranteed to make a face look younger. No study read for this page measured the perceived age of a wearer by brow shape. The 2007 study that is often cited for a youthful brow measured what younger and older raters preferred; the age in that result belongs to the person doing the looking, not to the face being looked at. Treating the two as the same is a category error, and it is the most common one in the advice this page read.

It cannot be guaranteed to suit a face shape. The 2007 preference study found a significant effect for square and long faces and none for oval or round; the sample was 78 raters and five morphed faces. A 2026 guide still assigns flat to rectangle and another still assigns horizontal to long, but neither publishes a threshold or a sample. A rule that changes with the guide is a convention, not a measurement.

It cannot be reduced to one number. The arch rise and the tail delta are independent. A brow can have a small rise and a tail that drops, or a larger rise and a tail that finishes level; both can be called straight. The lift is a third number and it sits outside the brow entirely. Anyone who gives a single figure for a brow shape has collapsed at least one of these readings into another.

What the method can do is give the same brow a repeatable description and make two changes comparable. If you draw the brow, the line can be removed at the end of the day; if you remove hair, the change stays until the hair grows back. The readings do not decide which of those is right for you. They only keep the description honest while you decide.

Questions this page answers