Calculus tool

Multivariable Limit Calculator

Q: How many approach paths does a multivariable limit actually need to prove existence?

In principle, infinitely many, since there are infinitely many directions and curves through a single point in the plane. In practice, mathematicians prove existence with the formal epsilon-delta definition or a polar-coordinate substitution, not by testing a finite list of paths. Testing a handful of representative paths (as this calculator does) is a fast, genuinely useful sanity check, and it is a complete proof of non-existence the moment two paths disagree, but it cannot fully prove existence on its own.

Q: Why does a straight-line-only check sometimes miss that a limit does not exist?

Because some functions are constructed so that every straight line through the target point gives the same value, while a curved path (very often a parabola shaped to match the function's own denominator) gives a different one. The classic textbook example is x²y / (x⁴ + y²) at the origin: every straight line y = mx gives 0, but the parabolic path y = x² gives 1/2. That is exactly why this calculator always includes a parabolic path alongside the five straight lines.

Q: What does it mean if the calculator says a path's value is "no clear value"?

It means the sampled values along that specific path did not settle down to a stable number as the sample points shrank toward the target (they kept growing, oscillating, or hitting an undefined point, for example division by zero along that particular path). That path is inconclusive on its own; the overall consistency check across all six paths still applies.

Q: Can this tool handle a limit at a point other than the origin?

Yes. Enter any target values for a (the x-coordinate) and b (the y-coordinate); they don't need to be 0. Continuous functions like x² + y² at a non-singular point are a good sanity check: every path should agree with simple direct substitution.

Enter a function of two variables, f(x, y), and a target point (a, b). This calculator samples the function along six standard approach paths (horizontal, vertical, and three diagonal lines of different slopes, plus a parabolic path) and reports whether every path agrees on the same value or whether two paths disagree, which proves the limit does not exist. A single substitution can never prove a two-variable limit exists, because there are infinitely many directions to approach a point in the plane; this tool is upfront about testing several representative paths rather than pretending to check them all.

Function f(x, y)

x approaches (a)

y approaches (b)

Try:

lim(x,y)→(0,0) f(x,y) = –

Was this tool helpful?

Give feedback

Why a two-variable limit is fundamentally harder than a single-variable one

For a single-variable limit lim_x→a f(x), there are only two directions to approach a from: the left and the right. If both agree, the limit exists. For a two-variable limit lim_{(x,y)→(a,b)} f(x,y), there are infinitely many directions and infinitely many curved paths to approach (a, b) from in the plane, and the limit only exists if f(x, y) approaches the same value along every single one of them. That asymmetry is the whole difficulty of multivariable limits: a handful of paths agreeing is strong evidence the limit exists, but it is never a complete proof, while just two paths disagreeing is a complete proof that the limit does not exist.

Worked example: a limit that does not exist

Consider lim_{(x,y)→(0,0)} xy / (x² + y²). Along the path y = 0 (approaching the origin along the x-axis), f(x, 0) = 0 for every x ≠ 0, so the value along this path is 0. Along the path y = x (approaching along the diagonal), f(x, x) = x² / (2x²) = 1/2 for every x ≠ 0, a constant value regardless of how close x gets to 0. Since one path gives 0 and another gives 1/2, the two paths disagree, which is a complete proof that the two-variable limit does not exist, even though the function is perfectly well-defined and bounded near the origin along any single path.

Path y = 0 (horizontal): f(x,0) = (x·0)/(x²+0²) = 0 for all x ≠ 0.
Path y = x (diagonal, slope 1): f(x,x) = (x·x)/(x²+x²) = x²/2x² = 1/2 for all x ≠ 0.
0 ≠ 1/2: the paths disagree.
Conclusion: the limit does not exist, because a limit must be the same along every approach path.

Worked example: a limit that does exist

Consider lim_{(x,y)→(0,0)} x²y / (x² + y²). This one looks similar to the example above but behaves very differently. Every straight-line path through the origin, y = mx for any slope m, gives f(x, mx) = mx³ / (x² + m²x²) = mx / (1 + m²), which shrinks to 0 as x → 0 for every value of m, since the denominator stays bounded away from 0 while the numerator vanishes. A useful general bound confirms it further: since x² ≤ x² + y², the whole expression satisfies |f(x,y)| ≤ |y|, which forces f(x,y) → 0 as (x,y) → (0,0) regardless of the direction of approach. This calculator's six tested paths (including the parabolic one, which a purely linear check would miss) all agree on 0, consistent with that bound.

The six paths this calculator tests, and why a parabola is included

Testing only straight lines through the target point (horizontal, vertical, and a couple of diagonals) is a classic trap: some functions genuinely disagree with every straight line but reveal path-dependence only along a curved path, most famously a parabola matched to the function's own shape. That's why this calculator always includes a parabolic path (y = b + (x - a)²) alongside the five straight-line paths, so a limit that only breaks down along a curve is not silently reported as "exists" just because every straight line happened to agree.

Horizontal: y = b (approach along a line parallel to the x-axis).
Vertical: x = a (approach along a line parallel to the y-axis).
Diagonal, slope 1: y = b + (x - a).
Diagonal, slope -1: y = b - (x - a).
Diagonal, slope 2: y = b + 2(x - a).
Parabola: y = b + (x - a)² (catches path-dependence that only shows up along a curved approach, which straight lines alone can miss).

What "consistent across every tested path" does and does not prove

When all six tested paths agree on the same value, this calculator reports that as strong evidence the limit exists, not a formal proof. A rigorous proof that a two-variable limit exists uses the formal epsilon-delta definition (showing that for every ε > 0 there is a δ > 0 such that |f(x,y) - L| < ε whenever the distance from (x,y) to (a,b) is less than δ, for every direction at once) or polar-coordinate substitution, not a finite list of paths. Six agreeing paths rule out every common counterexample pattern taught in a standard calculus course, but a pathological function could in principle disagree along a path none of the six happen to trace. When a path-dependence is actually found, on the other hand, that conclusion is airtight: two disagreeing numbers are a complete proof on their own, no further checking needed.

Related tools

For an ordinary single-variable limit, x → a, see the general-purpose limit calculator, which covers direct substitution, factoring, and infinity limits. For an indeterminate 0/0 or ∞/∞ ratio of single-variable functions, see the L'Hopital's rule calculator, which differentiates numerator and denominator symbolically instead of testing approach paths.

Frequently asked questions

How many approach paths does a multivariable limit actually need to prove existence?

In principle, infinitely many, since there are infinitely many directions and curves through a single point in the plane. In practice, mathematicians prove existence with the formal epsilon-delta definition or a polar-coordinate substitution, not by testing a finite list of paths. Testing a handful of representative paths (as this calculator does) is a fast, genuinely useful sanity check, and it is a complete proof of non-existence the moment two paths disagree, but it cannot fully prove existence on its own.

Why does a straight-line-only check sometimes miss that a limit does not exist?

Because some functions are constructed so that every straight line through the target point gives the same value, while a curved path (very often a parabola shaped to match the function's own denominator) gives a different one. The classic textbook example is x²y / (x⁴ + y²) at the origin: every straight line y = mx gives 0, but the parabolic path y = x² gives 1/2. That is exactly why this calculator always includes a parabolic path alongside the five straight lines.

What does it mean if the calculator says a path's value is "no clear value"?

It means the sampled values along that specific path did not settle down to a stable number as the sample points shrank toward the target (they kept growing, oscillating, or hitting an undefined point, for example division by zero along that particular path). That path is inconclusive on its own; the overall consistency check across all six paths still applies.

Can this tool handle a limit at a point other than the origin?

Yes. Enter any target values for a (the x-coordinate) and b (the y-coordinate); they don't need to be 0. Continuous functions like x² + y² at a non-singular point are a good sanity check: every path should agree with simple direct substitution.

Working with a single variable instead?

Use the general limit calculator for x approaching a finite value or infinity.

Go to the limit calculator

Free. No sign-up.