What a root is
A root runs a power backwards. Squaring 9 gives 81, so the square root of 81 is 9. Cubing 4 gives 64, so the cube root of 64 is 4. The general version carries an order: the nth root of a number a is the value that, multiplied by itself n times, lands back on a.
ⁿ√a = b means b × b × … × b = a
(n copies of b)
The number under the sign has a name worth knowing, because textbooks use it without explanation: the radicand. The small number sitting in the crook of the sign is the index or order. A square root leaves the index off, since two is assumed; every other root writes it in.
One habit causes more confusion in this corner of algebra than anything else. Eighty-one has two square roots, 9 and −9, because a negative times a negative is positive. The radical sign, though, is defined to return only the non-negative one, called the principal root. That is why solving x² = 81 gives x = ±9 with the sign written in by hand, while √81 on its own is simply 9. This page returns principal roots, the same convention every scientific calculator follows.
Using the three tabs
The three tabs do the same arithmetic with different amounts of typing. Square root and cube root each take one number and fix the order at 2 or 3. Any root takes both, so you can ask for a seventh root or a root of order 1.5 without hunting for a key.
Whichever tab you are on, the answer appears as you type — there is no need to press Calculate, though the button is there if you prefer it. The panel on the right carries four things:
- The decimal, to fourteen significant digits, which is roughly the limit of what double-precision arithmetic can promise.
- The check line, showing the answer raised back to the power you asked for. If that line does not reproduce your original number, something is wrong, and you can see it rather than take it on faith.
- The simplified radical form, when one exists.
- A bar chart of nearby roots of the same number, which makes the shape of the function visible: roots crowd together quickly as the order climbs.
Simplified radical form
Homework often wants the exact answer rather than a decimal, and "exact" usually means simplified radical form. The rule is to pull out of the radical every factor that is itself a perfect power of the right order.
Take √72. Seventy-two factors into 2 × 2 × 2 × 3 × 3. A square root pairs factors up: the two 3s make a pair, and two of the three 2s make another. Each pair escapes the sign as a single copy, so 3 and 2 come out, multiply to 6, and one lonely 2 stays behind:
√72 = √(36 × 2) = 6√2 ≈ 8.4852813742386
Cube roots want triples instead of pairs. ∛54 factors into 2 × 3 × 3 × 3; the three 3s form one triple, 3 steps outside, and the 2 remains:
∛54 = ∛(27 × 2) = 3∛2 ≈ 3.7797631496846
The calculator does this factoring for you and prints the simplified form beneath the decimal whenever the radicand is a whole number and the order is a whole number from 2 to 12. When no factor can be pulled out — √2, √15, ∛7 — the radical is already as simple as it gets, and nothing extra appears.
A small warning about the phrase. "Simplified" is a convention, not a mathematical improvement: 6√2 and √72 are the same number. What the simplified form buys you is comparability. Two expressions that look different often turn out to be identical once both are reduced, which is exactly what a marker is checking for.
Why 10,000,000 sometimes prints as 9,999,999.99999999
Ask most calculators for the cube root of 1021 and you get something like 9999999.999999998. The true answer is ten million exactly. Nothing is broken; the machine is doing what it was built to do.
Computers store ordinary numbers in binary with a fixed budget of bits, and a root is normally computed as a raised to the power 1/n — a logarithm, a multiplication and an exponential, each rounded. One third cannot be written exactly in binary any more than it can in decimal, so the exponent is already a hair off before the arithmetic starts, and the result lands a hair below the whole number it should have hit.
This page handles it differently. Before showing a result it asks whether the radicand is a whole number and the order a whole number, and if so it searches for an exact integer root using arbitrary-precision arithmetic — the kind that does not round. It uses Newton's method on integers, then verifies the candidate by raising it back to the power and comparing against your input digit for digit. Only an exact match is reported as exact.
The upshot is that perfect powers come back clean. The cube root of 1021 reads 10,000,000. The seventh root of 78,125 reads 5. And when the answer genuinely is irrational, the decimal is shown without a false claim of exactness attached to it.
Working a root out by hand
It is worth knowing how to get a root without a calculator, if only because the method explains what the calculator is doing quickly.
For a square root, guess, divide, average, repeat. To find √40, start with 6, since 6² = 36 is close. Divide 40 by 6 to get 6.6667. The true root lies between those two, so average them: 6.3333. Divide again, 40 ÷ 6.3333 = 6.3158, and average once more to reach 6.3246. The actual value is 6.3245553203368, so three rounds of simple arithmetic have already bought five correct digits. Each further round roughly doubles the number of correct digits, which is why the method feels almost unreasonably fast.
For higher orders the same idea works with a small adjustment. To find the fourth root of 500, guess 4 and note that 44 = 256, which is low. Try 5: 54 = 625, too high. The answer sits between the two, and nearer the top since 500 is much closer to 625 than to 256 — test 4.7 and you get 487.97, a little low, so nudge up to 4.73 and you get 500.55, a little high. Two more adjustments land on 4.7287, against a true value of 4.7287080450159. Bracket, test, narrow: it is slower than the averaging trick but needs nothing beyond multiplication.
Negative numbers under the sign
Whether a negative radicand has an answer depends entirely on whether the order is odd or even, and the reason is short. An odd number of negative factors stays negative; an even number turns positive.
So the cube root of −27 is −3, because (−3) × (−3) × (−3) = −27, and this page returns it without complaint. The square root of −4 has no real value, because nothing real squares to a negative. The same goes for the fourth root, the sixth, and every other even order.
Mathematics does not stop there — √−4 is 2i in the complex numbers, where i is defined as the square root of −1. Complex roots matter enormously in electrical engineering and signal processing. This calculator works in real numbers only and will tell you so rather than silently returning nothing.
Fractional and negative orders
Nothing requires the order to be a whole number, and writing roots as exponents shows why. A root of order n is the same thing as raising to the power 1/n:
ⁿ√a = a1/n ∛a² = a2/3
Once a root is an exponent, every exponent rule applies to it unchanged, which is the real reason the notation is worth learning. A root of a root multiplies the orders. A root of a product splits across the factors. An order of 0.5 is a square root; an order of 1.5 is a cube root of a square.
Negative orders flip the result into a reciprocal: a root of order −2 is one divided by the square root. Accept both here and both behave as the algebra says they should. The one order with no meaning at all is zero, since a1/0 asks for division by zero, and the calculator says so instead of producing a number.
Fractional orders do come with a caveat worth stating plainly: a negative radicand under a fractional order has no real value even when the fraction happens to reduce to an odd whole number, because the intermediate step goes through a power that is undefined for negatives. Write such a case with whole numbers if you want an answer.
Where roots turn up
Roots are far more common outside the classroom than their reputation suggests, and usually they are undoing a growth or an area.
Finance. Turning total growth into an annual rate is a root. An investment that grows 60% over seven years has an annual rate of the seventh root of 1.60, minus one — about 6.94% a year, which is a far smaller number than most people guess from 60%.
Geometry. Area gives side length through a square root; volume gives edge length through a cube root. Scaling a shape so its area doubles means multiplying every length by √2, roughly 1.414, which is why paper sizes from A4 to A3 use that exact ratio.
Statistics. Standard deviation is the square root of variance, taken specifically so the spread is reported in the same units as the data rather than in units squared. The standard error of a mean shrinks with the square root of the sample size, which is why quadrupling a survey only halves its margin of error.
Physics and engineering. A pendulum's period follows the square root of its length. The speed of a wave on a string follows the square root of tension over density. Impact speed from a fall follows the square root of height.
If a quantity is defined through squaring or cubing something, the root is how you get back — which is most of what roots are for.
Frequently asked questions
What is the difference between a root and a radical?
The radical is the symbol and the expression written with it; the root is the value that expression evaluates to. The square root of 49 is 7, and √49 is the radical expressing it. In ordinary use the two words are swapped freely without anyone minding.
Does every number have a square root?
Every non-negative real number does. Negative numbers have no real square root, because any real number squared is positive or zero. They do have square roots among the complex numbers, where √−4 is 2i, but those are not real values and this calculator works in real numbers.
Why does √81 give 9 and not −9?
Both 9 and −9 square to 81, but the radical sign is defined to return only the non-negative one, called the principal root. When you solve x² = 81 as an equation you write ±9, adding the second solution by hand, because the equation has two answers while the symbol returns one.
How do I put an answer in simplified radical form?
Factor the radicand, pull out every group of n identical factors as a single copy outside the sign, and leave the rest inside. For √72 the factors are 2·2·2·3·3, the pair of 3s and a pair of 2s come out as 6, and 6√2 remains. This page prints that form automatically under the decimal.
Why do some calculators show 9999999.999999998 instead of 10000000?
Because a root is normally computed through logarithms and exponentials in binary floating point, and 1/3 cannot be represented exactly, so the result lands a fraction below the whole number. This page checks every whole-number case with exact integer arithmetic before printing, so perfect powers come back clean.
Can the root order be a fraction or a negative number?
Yes to both. An order of n means raising to the power 1/n, so an order of 1.5 is a cube root of a square, and a negative order gives the reciprocal of the ordinary root. Only an order of zero is undefined, since it asks for division by zero in the exponent.
Results are shown to fourteen significant digits, the practical limit of double-precision arithmetic. Whole-number cases are verified against exact integer arithmetic before being reported as exact. For work where the last digit carries real consequences, confirm with a computer algebra system that keeps symbolic values rather than decimals.