This page drills Heron's formula until no case can surprise you. We deliberately hunt the edges: right triangles, obtuse ones, acute scalene, integer-perfect "Heronian" cases, degenerate (flat) triangles, forbidden side-lengths, scaling, and a real word problem.
Intuition Read this first
Heron takes three lengths a , b , c , builds s = 2 a + b + c , and returns
Area = s ( s − a ) ( s − b ) ( s − c ) .
The only thing that can go wrong is what happens under the square root . If all four factors are positive → a real triangle with real area. If one factor hits zero → the triangle has been squashed flat (area 0 ). If one goes negative → the three sticks can't close into a triangle at all. Every "scenario" below is really a story about the sign of those four factors.
Cell
What it tests
Under-root behaviour
Example
A. Right triangle
cross-check vs 2 1 × base × height
all > 0 , perfect square
Ex 1 (3-4-5)
B. Obtuse triangle
cos C < 0 , does Heron still work?
all > 0
Ex 2 (2-3-4)
C. Acute scalene
generic, all angles < 9 0 ∘ , irrational area
all > 0 , non-square inside
Ex 3 (4-5-6)
D. Symmetric / equilateral
general side a , algebra not numbers
all > 0
Ex 4
E. Heronian (integer area)
messy sides, clean answer
perfect square inside
Ex 5 (5-12-13, 9-10-17)
F. Degenerate (collinear)
a + b = c exactly
one factor = 0 → area 0
Ex 6
G. Forbidden sides
a + b < c
one factor < 0 → no triangle
Ex 7
H. Scaling / limiting
multiply all sides by k
area scales by k 2
Ex 8
I. Word problem (real world)
pick sides from context, units
all > 0
Ex 9 (triangular field)
J. Exam twist (reverse)
given area, find missing side
solve under root
Ex 10
We now hit every cell.
How to read this figure (three panels, left → right). Each panel shows the same base drawn in gray along the horizontal base axis, with the two slanted sides in orange and the apex (top vertex) marked as a coloured dot; the vertical axis is the apex's height above the base. The two orange sides are the sticks whose lengths feed the factors ( s − a ) and ( s − b ) .
Left panel (green apex, high up): a genuine triangle. All four under-root factors are positive, so the area (the blue shaded region) is a real positive number.
Middle panel (orange apex, on the base line): the apex has slid all the way down until it sits on the base — the shape is flat. Exactly one factor, s − c , has reached 0 , so the area is 0 .
Right panel (red apex, red dashed gap): the two orange sticks are now too short to meet — there is a visible gap where the apex should be. One factor has gone negative , and no triangle exists.
So the single left-to-right motion of one dot — high, then on the line, then failing to meet — is the entire scenario matrix in one animation frame set.
Worked example Example 1 — the 3-4-5 triangle
Sides a = 3 , b = 4 , c = 5 . Find the area with Heron, then confirm with base×height.
Forecast: it's a right triangle (legs 3 and 4), so before doing any Heron algebra, guess the area from 2 1 × 3 × 4 . Hold that number.
s = 2 3 + 4 + 5 = 6 .
Why this step? Heron needs the semi-perimeter first — every later factor ( s − a ) , ( s − b ) , ( s − c ) is measured from this single reference length, so we cannot start anywhere else.
s − a = 3 , s − b = 2 , s − c = 1 .
Why this step? These are the "slack" each side leaves against the semi-perimeter. All three positive is the algebraic signature of a genuine triangle (Cell A confirmed).
Area = 6 ⋅ 3 ⋅ 2 ⋅ 1 = 36 = 6 .
Why this step? Multiply the four factors, then undo the squaring that the derivation built in by taking the root.
Verify: the 5 is the hypotenuse (3 2 + 4 2 = 25 = 5 2 ), so the legs 3 , 4 are perpendicular and Area = 2 1 ( 3 ) ( 4 ) = 6 . ✅ Matches. See Law of Cosines : here cos C = 24 9 + 16 − 25 = 0 , so C = 9 0 ∘ — Heron doesn't need to know that, but it's consistent.
Worked example Example 2 — the 2-3-4 triangle (obtuse)
Sides a = 2 , b = 3 , c = 4 . Does Heron care that the biggest angle is obtuse?
Forecast: c = 4 is longer than what a right triangle would allow (2 2 + 3 2 = 13 ≈ 3.6 ), so the angle opposite c is obtuse and cos C < 0 . Guess: does the formula flinch? (It won't.)
s = 2 2 + 3 + 4 = 2 9 = 4.5 .
Why this step? Build s as always; note it need not be an integer — Heron works over the reals.
s − a = 2.5 , s − b = 1.5 , s − c = 0.5 .
Why this step? The smallest slack s − c = 0.5 is tiny because c = 4 is the "greedy" long side — but it is still positive, so the triangle is real. Obtuseness lives in the angle , not in the sign of any factor .
Area = 4.5 ⋅ 2.5 ⋅ 1.5 ⋅ 0.5 = 8.4375 ≈ 2.9047 .
Why this step? Plug in and root; the result is irrational, which is perfectly normal.
Verify: with Law of Cosines , cos C = 2 ⋅ 2 ⋅ 3 4 + 9 − 16 = 12 − 3 = − 0.25 < 0 — confirmed obtuse. Then sin C = 1 − 0.0625 = 0.9375 and 2 1 ab sin C = 2 1 ( 2 ) ( 3 ) 0.9375 = 3 0.9375 ≈ 2.9047 . ✅ Identical. Heron never asked about the sign of cos C ; the derivation already swallowed it.
Worked example Example 3 — the 4-5-6 triangle (all angles acute, all sides different)
Sides a = 4 , b = 5 , c = 6 . Find the area, then confirm every angle is genuinely acute.
Forecast: no side is "special," no right angle, no symmetry — this is the everyday triangle. Predict whether the area comes out as a clean number or an irrational one.
s = 2 4 + 5 + 6 = 2 15 = 7.5 .
Why this step? Build s ; here it is a half-integer, reminding us Heron isn't limited to integer inputs.
s − a = 3.5 , s − b = 2.5 , s − c = 1.5 .
Why this step? All three slacks are comfortably positive, and none is tiny — the hallmark of a "fat," well-proportioned triangle rather than a thin one.
Area = 7.5 ⋅ 3.5 ⋅ 2.5 ⋅ 1.5 = 98.4375 = 4 15 7 ≈ 9.9216 .
Why this step? The product under the root is 2 15 ⋅ 2 7 ⋅ 2 5 ⋅ 2 3 = 16 1575 , and 1575 = 15 7 , so the exact area is 4 15 7 — irrational, exactly as a generic triangle should be.
Verify (all angles acute): by the Law of Cosines , an angle is acute exactly when the square of its opposite side is less than the sum of the other two squares. Largest side is c = 6 : 6 2 = 36 vs 4 2 + 5 2 = 41 , and 36 < 41 , so even the biggest angle is acute — hence all angles are acute. Cross-check the area with 2 1 ab sin C : cos C = 2 ⋅ 4 ⋅ 5 16 + 25 − 36 = 40 5 = 0.125 , sin C = 1 − 0.12 5 2 = 0.984375 , and 2 1 ( 4 ) ( 5 ) 0.984375 = 10 0.984375 ≈ 9.9216 . ✅ Matches Heron.
Worked example Example 4 — equilateral of side
a , and isosceles a , a , b
Derive the equilateral area, then generalise to any isosceles triangle with equal sides a and base b .
Forecast: you probably remember 4 3 a 2 for equilateral. Can Heron reproduce it and give the isosceles version in one shot?
Equilateral: sides a , a , a , so s = 2 3 a and s − a = 2 a (three times).
Why this step? Symmetry means all three ( s − ⋅ ) factors are equal, so the four-factor product collapses to s ⋅ ( s − a ) 3 — a huge simplification we get for free from the symmetry.
Area = 2 3 a ( 2 a ) 3 = 16 3 a 4 = 4 3 a 2 .
Why this step? Collect powers of a under the root; a 4 = a 2 pulls the length-squared out, leaving the pure constant 4 3 .
Isosceles a , a , b : s = 2 2 a + b = a + 2 b . Then s − a = 2 b (twice) and s − b = a − 2 b .
Why this step? Two equal sides force two equal factors, so again the product partly collapses — this is why symmetric cases are algebraically easy.
Area = ( a + 2 b ) ( 2 b ) 2 ( a − 2 b ) = 2 b a 2 − 4 b 2 = 4 b 4 a 2 − b 2 .
Why this step? Pull the repeated 2 b out of the root; the remaining two factors are a difference of squares ( a + 2 b ) ( a − 2 b ) = a 2 − 4 b 2 — the same Difference of Squares Factoring move that powered the original derivation.
Verify: set b = a in step 4: Area = 4 a 4 a 2 − a 2 = 4 a 3 a = 4 3 a 2 . ✅ The isosceles formula collapses to the equilateral one. Also the classic height check: isosceles height h = a 2 − ( b /2 ) 2 , area = 2 1 bh = 2 b a 2 − b 2 /4 — same thing.
Worked example Example 5 — 5-12-13 and 9-10-17
Show both give whole-number areas (that's what makes them Heronian Triangles ).
Forecast: 5 , 12 , 13 is another Pythagorean triple — bet on a clean area. 9 , 10 , 17 is not right-angled; will it still be clean?
Triangle 5-12-13:
s = 2 5 + 12 + 13 = 15 . Why? Build s ; it lands on an integer because the perimeter 30 is even — a good omen for a clean area.
s − a = 10 , s − b = 3 , s − c = 2 . Why? All positive → a real triangle, and all integers, so the under-root product will be an integer.
Area = 15 ⋅ 10 ⋅ 3 ⋅ 2 = 900 = 30 . Why? 900 = 3 0 2 is a perfect square, which is the precise condition for an integer area.
Triangle 9-10-17:
s = 2 9 + 10 + 17 = 18 . Why? Build s ; perimeter 36 is even so s is an integer.
s − a = 9 , s − b = 8 , s − c = 1 . Why? All positive, so valid — but s − c = 1 is small because c = 17 is nearly 9 + 10 = 19 , i.e. a "thin" triangle sitting close to degenerate.
Area = 18 ⋅ 9 ⋅ 8 ⋅ 1 = 1296 = 36 . Why? 1296 = 3 6 2 → integer area, even without a right angle.
Verify: 5-12-13: 5 2 + 1 2 2 = 25 + 144 = 169 = 1 3 2 , so it's right-angled, 2 1 ( 5 ) ( 12 ) = 30 . ✅ 9-10-17: not a triple, but 18 ⋅ 9 = 162 , 8 ⋅ 1 = 8 , 162 ⋅ 8 = 1296 = 3 6 2 . ✅ Both integer — genuinely Heronian.
Worked example Example 6 — sides 3, 4, 7 (a squashed triangle)
Compute the "area" of a triangle with a = 3 , b = 4 , c = 7 .
Forecast: notice 3 + 4 = 7 = c exactly. The two short sticks laid end-to-end reach the far endpoint — the triangle is flat as paper. Guess the area before computing.
s = 2 3 + 4 + 7 = 7 . Why? Build s ; note s equals c here, which is the tell-tale of the boundary case.
s − a = 4 , s − b = 3 , s − c = 0 . Why this step? The factor s − c = 0 is the alarm bell — it happens precisely when c equals the sum of the other two, the equality edge of the triangle inequality.
Area = 7 ⋅ 4 ⋅ 3 ⋅ 0 = 0 = 0 .
Why this step? Anything times 0 is 0 ; the root of 0 is 0 , so Heron reports the flat shape's true area cleanly.
Verify: geometrically the three points are collinear, enclosing no space — area must be 0 , and Heron returns exactly 0 . This is the boundary case set by the Triangle Inequality (a + b ≥ c with equality). ✅ Look at the middle panel of the figure: the apex has landed on the base.
Worked example Example 7 — sides 1, 1, 5
Try Heron on a = 1 , b = 1 , c = 5 .
Forecast: two 1-length sticks can never span a gap of 5. Predict what the formula does when handed impossible data.
s = 2 1 + 1 + 5 = 2 7 = 3.5 . Why? Build s mechanically — the formula doesn't yet "know" anything is wrong.
s − c = 3.5 − 5 = − 1.5 < 0 . Why this step? One factor is negative , which is the exact algebraic fingerprint of the Triangle Inequality a + b > c being violated (2 < 5 ).
Area = 3.5 ⋅ 2.5 ⋅ 2.5 ⋅ ( − 1.5 ) = − 32.8125 → not a real number .
Why this step? A negative under the square root signals "no real triangle exists," not "small triangle." The formula is honestly refusing .
Verify: the product s ( s − a ) ( s − b ) ( s − c ) = 3.5 ⋅ 2.5 ⋅ 2.5 ⋅ ( − 1.5 ) = − 32.8125 < 0 . ✅ Negative ⇒ Heron correctly reports impossibility. Always test a + b > c before trusting a numerical output.
Worked example Example 8 — double every side of the 3-4-5
Take the 3-4-5 triangle and scale all sides by k = 2 to get 6-8-10. Predict the new area.
Forecast: lengths double, but area is a two-dimensional quantity — guess whether it doubles or quadruples.
New sides 6 , 8 , 10 , so s = 2 6 + 8 + 10 = 12 .
Why this step? Because s is a sum of lengths, scaling every side by k scales s by k too — here s went from 6 to 12 , exactly × 2 . This linearity is the seed of the whole scaling law.
s − a = 6 , s − b = 4 , s − c = 2 .
Why this step? Each slack is also exactly k = 2 times the original (3 , 2 , 1 ), because subtraction of two scaled lengths scales by k as well — so all four under-root factors scale by k together.
Area = 12 ⋅ 6 ⋅ 4 ⋅ 2 = 576 = 24 .
Why this step? Four factors each × k multiply the product by k 4 , so the root grows by k 2 ; taking 576 = 24 confirms it.
Verify: original area was 6 ; since the root scales by k 2 = 4 , we expect 6 × 4 = 24 . ✅ Area scales as the square of the length factor — a universal fact, visible directly in Heron's four-factor product.
Worked example Example 9 — a triangular field
A farmer's field has straight edges of 13 m , 14 m , and 15 m . Seed costs $2 per square metre. What's the seeding bill?
Forecast: the sides 13 , 14 , 15 are a famous Heronian set — you may recall the area is 84 . Predict the cost.
s = 2 13 + 14 + 15 = 21 m . Why? Semi-perimeter, in metres; keeping units on s means the final area lands automatically in m 2 .
s − a = 8 , s − b = 7 , s − c = 6 (all in metres). Why? All positive → a valid, buildable field.
Area = 21 ⋅ 8 ⋅ 7 ⋅ 6 = 7056 = 84 m 2 . Why? 21 ⋅ 8 = 168 , 7 ⋅ 6 = 42 , 168 ⋅ 42 = 7056 = 8 4 2 — a perfect square, so a clean integer area.
Cost =84\text{ m}^2\times\ 2/\text{m}^2=$168. ∗ W h y t hi ss t e p ? ∗ M u l t i pl y a r e ab y u ni tp r i ce ; t h e \text{m}^2in t h e a r e a c an ce l s t h e \text{m}^2$ in the denominator of the price, leaving pure dollars.
Verify: units — \text{m}^2\cdot(\ /\text{m}^2)=$.✅ N u m er i c — 84^2=7056. ✅ The bill is **\ 168**.
Worked example Example 10 — given the area, find the third side
An isosceles triangle has two sides of 10 and area 48 . Find the base b (there are two answers!).
Forecast: from Example 4, isosceles area = 4 b 4 a 2 − b 2 . Setting this to 48 with a = 10 is a hidden quadratic in b 2 — expect possibly two valid bases.
Area = 4 b 4 ( 10 ) 2 − b 2 = 4 b 400 − b 2 = 48 .
Why this step? Use the isosceles Heron result so only one unknown b remains — turning a triangle problem into an algebra problem.
Square both sides: 16 b 2 ( 400 − b 2 ) = 4 8 2 = 2304 , i.e. b 2 ( 400 − b 2 ) = 36864 .
Why this step? Squaring removes the root; introduce u = b 2 so the equation becomes a plain quadratic.
u ( 400 − u ) = 36864 ⇒ u 2 − 400 u + 36864 = 0 . Solve with the quadratic formula: u = 2 400 ± 40 0 2 − 4 ⋅ 36864 = 2 400 ± 160000 − 147456 = 2 400 ± 12544 .
Why this step? Standard quadratic; the whole difficulty is now hidden inside one square root.
Find 12544 by factoring, not guessing: 12544 = 2 7 ⋅ 98 = … more cleanly, 12544 = 64 ⋅ 196 , and both are perfect squares, so 12544 = 64 ⋅ 196 = 8 ⋅ 14 = 112 .
Why this step? Splitting a big radicand into a product of known squares (64 = 8 2 , 196 = 1 4 2 ) is the reliable way to extract a perfect square by hand.
u = 2 400 ± 112 = 256 or 144 , so b = 256 = 16 or b = 144 = 12 .
Why this step? b = u ; both are positive and satisfy the Triangle Inequality (b < 2 a = 20 ), so both are legitimate.
Verify: for b = 12 : 4 12 400 − 144 = 3 256 = 3 ⋅ 16 = 48 . ✅ For b = 16 : 4 16 400 − 256 = 4 144 = 4 ⋅ 12 = 48 . ✅ Two triangles (a "tall thin" one and a "short wide" one) share the same two legs and the same area — exactly the twist an exam loves.
Recall Which single quantity decides every scenario?
The sign of the product s ( s − a ) ( s − b ) ( s − c ) .
Positive ::: real triangle, real area.
Zero ::: degenerate (collinear, area 0 ).
Negative ::: no triangle exists (triangle inequality violated).
Recall How does area respond to scaling all sides by
k ?
It multiplies by k 2 (area is 2-dimensional; the four under-root factors each scale by k , so the root scales by k 2 ).
Recall Why can one area value give two different triangles (Ex 10)?
Because Heron squares to a quadratic in b 2 ; two positive roots can both satisfy the triangle inequality — a "tall" and a "wide" solution.
Recall How do you know a triangle is acute from its sides alone (Ex 3)?
Square the longest side and compare to the sum of the other two squares. Less than ::: all angles acute. Equal ::: right angle. Greater than ::: obtuse.
Mnemonic The four-factor traffic light
Green (all + ) → go, real area. Amber (one = 0 ) → flat, area 0 . Red (one − ) → stop, no triangle.
Heron's formula (derivation using trig) — the parent formula being drilled.
Law of Cosines — checks acute (Ex 3), obtuse (Ex 2) and right-angle (Ex 1) cases.
Area of a triangle = ½ab sin C — the sanity cross-check in Ex 2 and Ex 3.
Triangle Inequality — governs Cells F and G (zero / forbidden).
Semi-perimeter and Incircle (r = Area/s) — every example computes s first.
Difference of Squares Factoring — the isosceles collapse in Ex 4.
Heronian Triangles — Ex 5 and Ex 9 land on integer areas.