Boy nub vs girl nub: how to read a 12-week scan

Why the angle is everything

Everything begins with the genital tubercle angle, measured off a line through the baby's lower spine:

• Boy: nub tilts up and away, greater than 30°.
• Girl: nub stays flat and parallel, under 30°.

This isn't folklore — it's the rule from the original nub study (Efrat 1999), and later work confirmed the split: male tubercles averaged 35.9° versus 21.6° in females (Alfuraih 2024). The catch is that those ranges overlap near the middle, so a nub at 27–32° is genuinely a tough call. For the full method behind the measurement, see the angle of the dangle.

Why does the angle work at all? Because the male and female tubercle move in different directions as they develop. The male version lifts and rotates away from the spine; the female version stays tucked along the body line. By the time you reach 12–13 weeks, those two paths have separated enough to read — but only enough, not by a wide margin. That narrow gap is the whole reason nub guessing is exciting and unreliable in equal measure.

The genital tubercle: how it forms differently by sex

In the first trimester, every embryo begins with the same undifferentiated genital tubercle — the tiny bud you see on the scan. By week 9–10, chromosomal sex starts directing which path it takes. In males, androgens cause the tubercle to lift and elongate, rotating upward away from the body wall. In females, the structure stays lower, developing into the clitoris and labia without that upward rotation.

By 12 weeks the two trajectories are physically separate enough to see on ultrasound — but only just. The angular difference is real, it is measurable, and it is the basis of the nub method. But the degree of separation is modest at 12 weeks, which is why borderline angles are genuinely hard to call and why a week or two of extra development can make the read much clearer.

The >30° rule: what it means in practice

The 30° cutoff is not a rule of thumb — it is the threshold from Efrat et al. (1999), the study that first formalized the nub method in a peer-reviewed setting. In practice it means:

• If you can draw a line through the lower spine and the nub clearly tilts away from that line at 30° or more, that is the boy prediction.
• If the nub runs nearly parallel to the spine — loosely defined as under 30°, but in strong girl reads often under 15° — that is the girl prediction.
• If the angle is in the overlap zone (roughly 20–32°), it is borderline. A borderline nub is not a negative result — it is an inconclusive one. Do not force a read.

Later research found the average male angle is 35.9° and the average female angle is 21.6°, with the standard deviations overlapping between roughly 25° and 30°. That overlap is where mistakes cluster.

Boy vs girl: side-by-side marker table

The takeaway: trust the angle first. Direction and "stacking" are the kind of clues parents trade in forums, but they were never validated the way the angle was. Lead with the measured marker, treat the rest as supporting color.

A quick word on "stacking," since it comes up constantly. The idea is that layered, parallel white lines near the tubercle suggest a girl — sometimes called the "hamburger" look, mirroring the three-lines sign seen later on the anatomy scan. It can be a genuine clue in skilled hands, but it is easy to imagine into a grainy image, and it was never part of the peer-reviewed angle rule. Use it to support an angle read you already trust, not to overturn one. The same goes for "forward-facing": a tubercle that clearly points up and out is consistent with a boy, but only if you are truly looking at a side profile. In any other plane, "forward" doesn't mean what you think it does.

Mistakes people make when reading the angle

• Measuring on a non-sagittal image. Any rotation of the baby changes the apparent angle. If you are not looking at a perfect side profile with the spine as a clean line, there is no valid read — period.
• Confusing the umbilical cord for a nub. The cord runs between the legs and happily masquerades as a raised boy nub. Trace the cord before you commit to a read.
• Reading at 11 weeks. Efrat's data showed 56% of males were misread as female at 11 weeks. That is not a clue — that is a coin flip with extra steps.
• Forcing a read on a borderline angle. A nub in the high-20s to low-30s is inconclusive. The appropriate call is "too close to tell," not a reluctant guess.
• Confirmation bias. If you already want a girl, you will see a girl in an ambiguous photo. Ask someone neutral to measure the angle without knowing what you hope for.
• Over-trusting stacking. Layered lines can be shadows, cord loops, or tissue folds. The angle is measured; stacking is interpreted.

What good and bad scan images look like

A usable image has:

• A clear mid-sagittal (side-on) profile — the spine runs as a continuous line from the head to the base of the body.
• The nub clearly separated from the legs, with enough contrast to trace its direction.
• The gestation recorded — you need to know you're at 12 weeks or later.
• Enough resolution that you can see the angular difference without guessing at outlines.

An unusable image has:

• A tilted or front-on orientation — you are not seeing the angle you think you are.
• Legs crossed or tucked in a way that obscures the nub.
• Low resolution or heavy compression — if you have to squint to find the nub, the angle is a guess about a guess.
• No gestational age recorded — guessing the week from the image is another layer of error.

The rule of thumb: if you cannot draw a straight reference line through the spine and clearly see the nub angling away from or along that line, skip the read. A forced read on a bad image is worse than no read at all.

The borderline case: what if the nub is 20–29°?

This is the zone where the boy and girl angle distributions genuinely overlap. A nub at 25° is not a girl nub — it is a nub that could be either, sitting right in the middle of the overlap window.

What to do:

• Do not force a prediction. "Borderline" is a legitimate and honest result. The peer-reviewed studies did not pretend that every nub is classifiable, and neither should you.
• Weight other factors. Is the image near 13 weeks rather than 12? A borderline-flat nub at 13 weeks leans girl more than the same angle at 12 weeks, because most males have lifted noticeably by then.
• Wait for a better scan. A week of development can take a 27° nub and make it unmistakably 22° or 34°. If you have access to another scan at 13 weeks, that is often the cleanest resolution.
• Use it as one signal among several. The BabyPeek app never rests a prediction on the nub alone — it combines the angle with four other methods to reach a majority verdict, which handles borderline nubs far better than a single read.

How early is too early?

Twelve weeks is roughly the earliest the nub is readable at all. Before that, boys and girls look nearly the same, and early guesses flip like coins — Efrat's data showed 56% of males misread as female at 11 weeks, falling to near-zero by 13 weeks.

Under good 12–14-week conditions, accuracy is strong: male predictions hit 99–100%, and female accuracy rose from 91.5% in early 12 weeks to 100% by 13 weeks (Efrat 2006). Notice that boys read more reliably early — the female tubercle takes a little longer to settle into its flatter, unmistakable angle. For the full timing picture, see nub theory explained.

This asymmetry has a practical upshot for reading photos. If a 12-week scan shows a clearly raised, steep tubercle, a "boy" guess is on firm footing. If it looks flat and you want to call "girl," give it a little humility — at exactly 12 weeks a small share of those flat-looking nubs still turn out to be boys whose tubercles hadn't fully lifted yet. A confirming look closer to 13 weeks tightens the girl read considerably.

The short version: 10–11 weeks — too early. 12 weeks — readable under ideal conditions. 13–14 weeks — the best window. After 14 weeks the nub has typically become genitalia that read more clearly with a different kind of scan view.

Reading your own photo: a quick checklist

Before you call boy or girl from a photo, run through this:

• Is it a true side profile? You should see the spine running along one edge and the nub clearly separated from the legs. If you can't, stop — any angle you measure is fiction.
• Is the scan at least 12 weeks? Earlier than that, the markers haven't separated. A 10- or 11-week "nub" guess is mostly wishful.
• Where's the umbilical cord? It loves to photobomb between the legs and masquerade as a boy. Trace it before you commit.
• Is the angle clearly steep or clearly flat? A confident, well-separated angle is worth trusting. A borderline one in the high-20s to low-30s is not a verdict.
• Are you reading it, or hoping it? Confirmation bias is the silent killer of nub guesses. If you already want a girl, you'll see a girl in an ambiguous photo. Ask someone neutral.

None of this turns a guess into a diagnosis — but it keeps you from over-reading a fuzzy image. The structures themselves only become unmistakable later; for that full picture, see how to tell gender from ultrasound.

Honest limits: scan quality and position

Here's the part that humbles every confident nub guess.

• Scan quality. A grainy, low-resolution, or off-angle photo simply can't be measured. If you can't draw a clean spine line and see the nub clearly separated, the angle is a guess about a guess.
• Baby's position. You need a mid-sagittal side profile. If the legs are crossed, baby is curled up, or you're looking at the wrong plane, there's no valid read — full stop.
• Overlap near the cutoff. A borderline angle (high-20s to low-30s) is exactly where mistakes cluster.
• It's not a test. Even a textbook-perfect nub can be wrong. The reliable answer comes later — see how to tell gender from ultrasound and the 18–20-week anatomy scan.

Compare the nub method against all the other folk methods on the methods compared page to see where it fits in the broader picture.

FAQ

Sources

• Efrat, Akinfenwa & Nicolaides (1999) — male if tubercle angle >30°, female if parallel/under 30°.
• Efrat, Perri, Ramati, Tugendreich & Meizner (2006) — male accuracy 99–100%; female 91.5% early 12 wk to 100% by 13 wk.
• Alfuraih, Almajem & Alsolai (2024) — GTA mean 35.9° male vs 21.6° female; overlapping ranges near the cutoff.