The rule
- √(a · b) = √a · √b, for a ≥ 0 and b ≥ 0
The square root of a product is the product of the square roots. The root may be moved into each number separately.
- √(9 · 16) = √9 · √16 = 3 · 4 = 12, and sure enough √144 = 12
- √(4 · 25) = 2 · 5 = 10
- √(a²b) = a√b, if a ≥ 0
Why it holds
A square root is a power with exponent ½, and the exponent reaches everything in the bracket — this is the power of a product with exponent ½.
You can also check it backwards. Square the right-hand side:
(√a · √b)² = √a · √b · √a · √b = (√a)² · (√b)² = a · b
So the right-hand side is a number that squares to ab, and it is not negative. That is what √(ab) means, so the two are the same number.
The condition: no negatives
Both factors have to be zero or greater. With negatives the rule breaks: √(−4 · −9) is √36 = 6, but √(−4) · √(−9) is not a real number at all.
Worked examples
Simplify √72. Look for the largest square that divides 72: it is 36. Then √72 = √(36 · 2) = 6√2.
Simplify √50. 50 = 25 · 2, so 5√2.
Multiply two roots. √8 · √2 = √16 = 4. Read the other way, the rule is a way of merging.
With letters. √(16x⁴) = √16 · √(x⁴) = 4x², for every x, since x² is never negative.
Common mistakes
- Using it on a sum. √(9 + 16) is √25 = 5, not 3 + 4 = 7. The rule is about products, not sums.
- Not taking out the largest square. √72 = √(4 · 18) = 2√18 is true but not finished: √18 simplifies further.
- Using it on negative factors. See the condition above.
- Forgetting the factor outside. 3√2 · 2√2 is 6 · 2 = 12, not √2 · √2 = 2.
Common questions
Does it work for cube roots too?
Yes, and for every other index: ⁿ√(ab) = ⁿ√a · ⁿ√b. For odd indices the numbers need not even be
positive, since a cube root of a negative number exists. It is written here for the square root because
that is the one in daily use.
What does “simplifying” a root mean?
Taking every square out from under the root sign, so that what is left has no squares in it. √72
simplifies to 6√2, and there it stops.
Why write 6√2 rather than a decimal?
Because 6√2 is exact. The decimal 8.485… is rounded, and in a calculation with several steps the
rounding error grows. Keep the surd while you work and round only at the end.