Horizontal And Vertical Tangents From Partial Derivatives

Why Partial Derivatives Identify Tangent Direction

An implicit equation \(F(x,y)=0\) describes a curve without necessarily expressing \(y\) explicitly as a function of \(x\). At a point on the curve, the partial derivatives \(F_x\) and \(F_y\) reveal whether the local tangent is horizontal or vertical, provided the point is regular.

The key is to distinguish a candidate from a verified tangent. A point must first satisfy the curve equation. Then the relevant partial derivative must vanish, while the other must be nonzero. That nonzero condition ensures that the curve has a well-behaved local branch through the point.

The Classification Rule

Assume \(F\) has continuous first partial derivatives near \((a,b)\), and \(F(a,b)=0\). If \(F_y(a,b)\ne 0\), the implicit function theorem gives a local branch \(y=y(x)\). Differentiating the identity \(F(x,y(x))=0\) by the chain rule gives

Consequently, at the point \((a,b)\), the slope of this branch is

Thus, when \(F_y(a,b)\ne 0\), the tangent is horizontal exactly when \(F_x(a,b)=0\). A zero slope gives the horizontal tangent line \(y=b\).

For a vertical tangent, use the other local graph description. If \(F_x(a,b)\ne 0\), the implicit function theorem gives a local branch \(x=x(y)\), and differentiating \(F(x(y),y)=0\) gives

If \(F_y(a,b)=0\), then \(dx/dy=0\): the curve is locally expressible as \(x=x(y)\) with zero slope in that description. Its tangent line is vertical, \(x=a\).

The condition that the other partial derivative is nonzero is essential. If both partial derivatives vanish, these tests do not classify the point. Such cases require separate analysis; vanishing one partial derivative alone is not enough to verify a tangent.

Why the Rule Works

The argument follows from differentiating the curve constraint along a local branch. The two possible graph descriptions give the two slope formulas. The following steps make explicit how the hypotheses connect to the conclusion.

This also explains why the regularity check matters: the nonzero denominator is what permits the corresponding local graph and slope formula. The rule is not merely a pattern for guessing tangent directions; it is a consequence of the chain rule together with the local graph guaranteed by the implicit function theorem.

Worked Examples

A Common Mistake: Reversing the Partial Derivatives

A frequent error is to associate \(F_x=0\) with a vertical tangent and \(F_y=0\) with a horizontal one. The partial derivatives are components of a normal direction to the curve, so it can be tempting to read their zeros as tangent directions directly. The slope formula shows why that assignment is reversed.

For example, take the circle \(x^2+y^2-4=0\) at \((0,2)\). Here \(F_x=2x=0\) and \(F_y=2y=4\). A mistaken classification might call this vertical because \(F_x=0\). But the valid local slope formula gives \(y'=-F_x/F_y=-0/4=0\), so the tangent is horizontal. The point lies on the circle since \(0^2+2^2-4=0\), and \(F_y=4\ne0\) verifies that the local \(y(x)\) branch exists. Its tangent line is \(y=2\).

A Reliable Procedure

When a problem asks for horizontal or vertical tangents of an implicit curve, organize the work so that candidate generation and verification remain distinct.

1. Compute \(F_x\) and \(F_y\).
2. For horizontal candidates, solve \(F(x,y)=0\) together with \(F_x(x,y)=0\).
3. For every resulting point, evaluate \(F_y\). Keep it as a verified horizontal tangent only if \(F_y\ne0\).
4. For vertical candidates, solve \(F(x,y)=0\) together with \(F_y(x,y)=0\).
5. For every resulting point, evaluate \(F_x\). Keep it as a verified vertical tangent only if \(F_x\ne0\).
6. State the tangent line: \(y=b\) for a horizontal tangent at \((a,b)\), or \(x=a\) for a vertical tangent.

Solving only \(F_x=0\) or \(F_y=0\) is not sufficient: the point must also lie on \(F=0\). Likewise, finding a point on the curve where one partial derivative is zero does not complete the argument until the other partial derivative is checked. That final check is what establishes a regular local branch and makes the tangent classification valid.

Why This Method Is Useful

Implicit equations often describe curves that are awkward or impossible to solve globally for \(y\) in terms of \(x\). Partial derivatives allow tangent directions to be found directly from the defining equation, without first rearranging it into an explicit formula. They also make the logical conditions visible: the curve equation locates the point, one vanishing partial identifies a candidate direction, and the other nonzero partial verifies the appropriate local graph.

The method therefore gives both an efficient calculation and a safeguard against overclaiming. A zero derivative condition produces candidates, not automatically valid tangents. Checking the point and regularity keeps the conclusion tied to the local geometry the implicit equation actually guarantees.