Squaring can remove a square root from an equation, but it can also erase sign information. Solving a radical equation therefore has two distinct stages: produce possible solutions by algebra, then check which possibilities satisfy the original equation.
The square-root symbol chooses the nonnegative root
The expression equals 3, not The equation has two solutions, but the square-root function itself returns the nonnegative one. This distinction explains why a radical equation may reject a negative candidate.
For over the real numbers, the radicand must satisfy and the other side must satisfy Squaring then gives The squared equation alone no longer remembers that B was required to be nonnegative.
Example: isolate before squaring
Solve First isolate the radical: Square both sides to obtain giving Checking:
If you square before isolating, you must square the whole left side, including its cross-term. In general, is not Isolating the radical avoids unnecessary expansion and reduces mistakes.
Example: reject an extraneous candidate
Solve The original equation requires Squaring gives or The algebraic candidates are 2 and
At 2, so the equation holds. At which is not The value is extraneous: it solves the squared equation but not the original one. The only solution is 2.
Squaring both 1 and gives 1, which is exactly the information loss that created the extra candidate. Checking is logically necessary, not merely a way to catch arithmetic slips.
Example: a radical with a linear expression
Solve The right side must be nonnegative, so Squaring yields or The quadratic formula gives
Since it violates the original sign requirement. The remaining candidate is To check exactly, its right side is and the square of that value is equal to Thus the positive square root is indeed
Keeping exact radicals avoids accidentally accepting or rejecting a candidate because of a rounded decimal.
Two radicals may need two squaring steps
If an equation contains two square roots, isolate one radical first. After the first squaring, a cross-term may still contain a radical. Isolate that remaining radical and square again. Record restrictions and check all final candidates in the original equation, since each squaring can introduce possibilities that were not previously allowed.
Odd roots behave differently. Cubing is one-to-one on real numbers, so cubing both sides of an equation does not lose a sign in the same way. For example, gives hence with no requirement that the cube-root output be nonnegative.
Try it yourself
Solve Find all candidates after squaring, then identify the actual solutions.
Answer: Squaring gives so The candidates are 3 and Since the square root is nonnegative, is impossible; checking gives The only solution is 3.
What to remember
Isolate, state real-domain and sign requirements, square carefully, and check in the original equation. A candidate is not a solution until it survives that final check.