Multivariable calculus
Limit of x^2y/(x^2+y^4) at the Origin
The limit of x^2 y / (x^2 + y^4) as (x, y) approaches the origin is 0. Write the function as y times x^2/(x^2 + y^4). The second factor always lies between 0 and 1, so the size of the function is at most |y|, and any approach to the origin forces that to 0.
The limit exists
| Path | Limit along it |
|---|---|
| along y = 0 | 0 |
| along y = x | 0 |
| along the parabola x = y^2 | 0 |
Paths, including the curve that usually breaks these
Start with the axis and the diagonal, then substitute the parabola .
The third path is the parabola , chosen deliberately because it makes the two denominator terms the same size. On problems of this shape that is the approach most likely to expose a failure, and here it still returns . That raises confidence a lot, and still proves nothing.
The bounded factor argument
Split the function so that one piece obviously tends to zero and the other is obviously bounded.
The second factor is a nonnegative number divided by something at least as large, so it lands in at every point other than the origin.
Any approach to the origin drives , so the squeeze gives a limit of .
Polar coordinates are a poor tool here. They give , where the two denominator terms carry different powers of , so no single factor of comes out cleanly. When the denominator is not homogeneous, reach for a bound rather than for polar.
The mistake: confusing this with the standard counterexample
Move the fourth power to the other variable and the answer flips completely.
For , along the parabola the value is , while along it is . Two paths, two values, so has no limit at the origin. The two functions look nearly identical and behave in opposite ways.
The structural difference is which terms match. In the denominator contains , exactly the factor sitting in the numerator, so that pair divides out of the size estimate and only the lone is left over. In the denominator's terms are and , neither of which matches , and along the numerator and denominator are both of size , which lets a nonzero value survive all the way in.
Frequently asked questions
How do I know which curved path to try?
Match the degrees. Setting the two denominator terms equal in size, , suggests . That is the approach along which the denominator is as small as it can be relative to its parts, so if any path is going to produce a nonzero value, it is that one.
If every path I try gives 0, is the limit 0?
Only if you check every path, which cannot be done one at a time. That is why the bound matters: it holds at every point near the origin simultaneously, so it settles all approaches in a single line.