Direction Fields (Slope Fields)

19 min read

Quick Reference

The page numbers above are confirmed against Logan, 3rd ed., subsection 1.1.3.

Try This First

Take the equation $x' = t + x$ and three points: $(0, 1)$, $(1, 0)$, and $(-1, -1)$. At each point, work out the single number $t + x$.

Before reading on, write down what you think each of those three numbers tells you about a solution that happens to pass through that point.

The Idea Before the Formula

Most first-order differential equations cannot be solved with a clean formula. That sounds discouraging, and yet a first-order equation hands you something almost as useful for free. At every point of a window in the plane, the equation tells you the slope a solution must have there.

Picture a current map of a river, drawn as thousands of tiny arrows that show which way the water flows at each spot. The differential equation is the rule that sets the current: give it a location and it returns the direction of flow there. A solution is the path of a leaf dropped onto the water. The leaf does not choose its own heading; it follows the local current at every instant. Drop the leaf at different starting points and you trace different paths, but every one of them obeys the same current map.

This geometric view is useful precisely because it needs no solution formula. You can see whether solutions rise or fall, where they level off, and what they approach in the long run, all by reading the slope the equation assigns across the plane.

In Logan’s notation the equation is written $x' = f(t, x)$, where $t$ is the independent variable (often time) and $x$ is the unknown function. The Stewart and OpenStax bridge writes the same idea as $y' = F(x, y)$. Both say the same thing: feed the right side a point, read back a slope.

Prerequisite Hub

Builds on

Unlocks

Self-check before you begin

Official Definitions

The following definitions are stated as in Logan, 3rd ed., subsection 1.1.3 “Geometric Approach”.

Direction field (slope field). What does a differential equation $x' = f(t, x)$ tell us geometrically? At each point $(t, x)$ of a window of the $tx$ plane we draw a short line segment (a dash) with slope $f(t, x)$. The collection of all these line segments, or mini-tangents, forms the direction field, or slope field, for the equation. A solution curve fits into this field so that at each point its tangent line has the slope given by the field.

Source: Logan, 3rd ed., §1.1.3 Geometric Approach, pp. 11-12.

Nullclines. The nullclines of the differential equation $x' = f(t, x)$ are the set of points $(t, x)$ for which $f(t, x) = 0$. Along a nullcline the slope field is zero (horizontal mini-tangents), and these curves separate regions where the slope field is positive from regions where it is negative.

Source: Logan, 3rd ed., §1.1.3, p. 12.

Isoclines. An isocline is a curve along which the slope field has a constant value, that is $f(t, x) = k$ for some fixed $k$. Every solution curve crosses an isocline with the same slope $k$.

Source: Logan, 3rd ed., §1.1.3, p. 12 (and Example 1.8, p. 15).

A nullcline is the special isocline with $k = 0$. Isoclines give a fast hand-sketching method: pick a few values of $k$, draw each curve $f(t, x) = k$, and draw segments of slope $k$ all along it.

Quick Reference

Key Concepts

1. A First-Order Equation Is a Slope Machine

Write a first-order equation as $x' = f(t, x)$. This says the slope of a solution at the point $(t, x)$ equals $f(t, x)$. The equation does not yet give a formula for the solution, but it gives the solution’s slope everywhere.

Example 1. For $x' = t + x$, find the slope a solution would have at $(0, 1)$, $(1, 0)$, and $(-1, -1)$.

Predict first: two of these points sit where $t + x$ is positive and one where it is negative, so expect two upward dashes and one downward dash.

Substitute each point into $f(t, x) = t + x$: \[ f(0, 1) = 0 + 1 = 1, \qquad f(1, 0) = 1 + 0 = 1, \qquad f(-1, -1) = -1 + (-1) = -2. \] A solution through $(0, 1)$ has slope $1$ there; through $(1, 0)$ slope $1$; through $(-1, -1)$ slope $-2$ (a moderately steep downward dash). The prediction holds: two upward, one downward. Reading off slopes point by point is the entire method.

2. Nullclines: Where the Field Is Flat

The nullcline is the set of points where $f(t, x) = 0$. Along it every dash is horizontal, and the sign of $f$ flips as you cross it. Finding the nullcline is usually the first move in sketching a field, because it splits the plane into a region of upward dashes and a region of downward dashes.

Example 2. Find the nullcline of $x' = t + x$ and describe the field on either side.

Predict first: the right side is a sum, so the nullcline should be a straight line.

Set $f(t, x) = t + x = 0$, which gives the line $x = -t$. That line is the nullcline; every dash on it is horizontal. Above the line, $t + x > 0$, so dashes tilt upward; below it, $t + x < 0$, so dashes tilt downward. The field pushes solutions upward above $x = -t$ and downward below it, with a flat seam along the line. The prediction (a straight nullcline) is confirmed.

Important: a dash shows slope, not the location of a solution. A point’s dash tells you how steeply a solution climbs or falls if it passes through that point. It does not say a solution is actually there. Many different solution curves pass through different points; the field is the shared slope rule they all obey.

3. Isoclines: Curves of Constant Slope

An isocline is a curve $f(t, x) = k$ along which every dash has the same slope $k$. The nullcline is the isocline with $k = 0$. Choosing several values of $k$ and drawing the matching curves is a fast, accurate way to sketch a field by hand: along each curve you simply repeat one slope.

Example 3. For $x' = t + x$, find the isoclines of slope $k = 1$ and $k = -1$, and use them to sketch the field.

Predict first: each isocline of $t + x$ should be a straight line parallel to the nullcline $x = -t$.

Set $t + x = 1$, giving the line $x = 1 - t$; along it every dash has slope $1$. Set $t + x = -1$, giving the line $x = -1 - t$; along it every dash has slope $-1$. These two lines are parallel to the nullcline $x = -t$, as predicted. Drawing slope-$1$ dashes along $x = 1 - t$, horizontal dashes along $x = -t$, and slope-$-1$ dashes along $x = -1 - t$ already outlines the whole field: the dashes steepen upward as you move up and to the right.

4. Solution Curves Flow Along the Field

A solution curve is a curve tangent to the field dash at every point it touches. Picture dropping a leaf into a current: the leaf traces a path that always points along the local flow. That path is a solution curve.

Because a starting point selects one curve, there is one solution curve through each point, and they never cross (where they appear to meet, the field would have two slopes at once, which it cannot).

Example 4. A solution of $x' = t + x$ passes through $(0, 1)$. Describe its initial direction and how it bends.

At $(0, 1)$ the slope is $f(0, 1) = 1$, so the curve leaves the point rising at slope $1$. The point sits above the nullcline $x = -t$ (since $0 + 1 = 1 > 0$), so the field tilts upward there and grows steeper as the curve climbs into the region where $t + x$ is larger. The solution curls upward, steepening as it rises.

5. Autonomous Equations: Slopes Depend Only on $x$

When the right side depends on $x$ alone, $x' = f(x)$, the equation is called autonomous. The defining feature in the field is that along any horizontal line (constant $x$) every dash has the same slope, because $f$ does not see $t$. The field looks like horizontal stripes of identical slope.

Example 5. Describe the direction field of the autonomous equation $x' = x$.

Since $f(t, x) = x$ ignores $t$, the slope depends only on the height $x$. Along $x = 1$ every dash has slope $1$; along $x = 2$ every dash has slope $2$; along $x = -1$ every dash has slope $-1$. The dashes steepen upward as you move up and steepen downward as you move down, and they are flat along $x = 0$. Solutions are the exponential curves $x = Ce^{t}$, which the field reveals as growing away from the flat line $x = 0$.

6. Equilibria: Constant Solutions and Long-Run Behavior

For an autonomous equation $x' = f(x)$, a value $c$ with $f(c) = 0$ gives a horizontal nullcline of zero-slope dashes, so the constant function $x = c$ is a solution. It is called an equilibrium solution. Reading the sign of $f$ on either side tells you whether nearby solutions move toward the equilibrium (stable) or away from it (unstable).

Example 6. Find the equilibria of the logistic equation $x' = x(1 - x)$ and classify them.

Predict first: the right side factors, so expect two equilibria, one at each factor’s root.

Set $f(x) = x(1 - x) = 0$: the equilibria are $x = 0$ and $x = 1$, matching the prediction. Check the sign of $f$ in each region:

• For $0 < x < 1$: both factors positive, so $f(x) > 0$, solutions rise toward $x = 1$.
• For $x > 1$: the factor $1 - x < 0$, so $f(x) < 0$, solutions fall toward $x = 1$.
• For $x < 0$: the factor $x < 0$ and $1 - x > 0$, so $f(x) < 0$, solutions fall away from $x = 0$.

Solutions on both sides of $x = 1$ move toward it, so $x = 1$ is stable. Solutions on both sides of $x = 0$ move away from it, so $x = 0$ is unstable. This is the logistic field, and it shows carrying-capacity behavior directly: any positive population is drawn toward $x = 1$.

7. Matching Equations to Their Fields

A common skill is to look at a field or a candidate solution graph and decide which equation produced it. The reliable test is to check a few easy features: where the slopes are zero (the nullcline), where they are steep, and whether the slope depends on $t$, on $x$, or on both.

Example 7. Which of $x' = 1 + t^2 + x^2$ and $x' = t$ has a direction field whose slopes are never negative?

Predict first: a sum of $1$ and two squares can never be negative.

Examine the sign of each right side. For $x' = 1 + t^2 + x^2$, the right side is at least $1$ everywhere (a sum of $1$ and two squares), so the slope is always positive; no dash ever points downward, and the equation has no nullcline at all. For $x' = t$, the slope is negative whenever $t < 0$, so that field has downward dashes on the left half-plane and the nullcline is the vertical line $t = 0$. The equation whose slopes are never negative is $x' = 1 + t^2 + x^2$, confirming the prediction.

Common Misconceptions

Leveled Practice

Level 1: Direct Application

Problem 1. For $x' = 2t - x$, find the slope of a solution at $(0, 0)$, $(1, 1)$, and $(2, -1)$.

Problem 2. Find the nullcline of $x' = t - x$.

Level 2: Reading the Field

Problem 3. Is the equation $x' = 3x - 2$ autonomous? What is its nullcline?

Problem 4. For $x' = t + x$, find the isocline of slope $k = 2$ and describe the dashes along it.

Problem 5. Find and classify the equilibrium of the autonomous equation $x' = 3x - 2$.

Level 3: Reading the Field, Harder

Problem 6. For the autonomous equation $x' = (x - 2)(x + 1)$, find the equilibria and classify each.

Problem 7. A direction field has identical upward dashes along every horizontal line, getting steeper as $x$ increases, and flat dashes along $x = 0$. Which equation fits: $x' = t$ or $x' = x$?

Level 4: Matching and Modeling

Problem 8. A cooling object obeys $\dfrac{dT}{dt} = k(T - 20)$ with $k < 0$. Using a nullcline-and-sign argument, explain what happens to $T$ in the long run, for any starting temperature.

Problem 9. Without drawing, decide which equation has a direction field unchanged when you replace $t$ with $-t$ (a left-right flip of the plane): $x' = t + x$ or $x' = t^2 + x$.

Level 5: Reasoning and Justification

Problem 10. Explain why two distinct solution curves of a first-order equation $x' = f(t, x)$ cannot cross, in terms of the direction field. How would you convince a classmate?

Mastery Checklist

You have mastered this skill when you can do all of the following without referring to notes:

Novice (Level 1-2):

• I can interpret $x' = f(t, x)$ as assigning a slope to every point of the plane.
• I can compute the slope a solution would have at a given point.
• I can find the nullcline of an equation and state what it separates.

Competent (Level 3-4):

• I can find isoclines for chosen slopes and use them to sketch a field by hand.
• I can identify an autonomous equation by its $t$-independent field.
• I can find equilibrium solutions where $f(x) = 0$ and classify them as stable or unstable.
• I can read long-run behavior of solutions from the field.

Proficient (Level 5):

• I can match an equation to its field by checking sign, nullcline, and variable dependence.
• I can explain why distinct solution curves cannot cross, in field terms.

Mental Model

A direction field is a current map of a river, drawn as thousands of tiny dashes showing which way the water flows at each spot.

The differential equation is the rule that sets the current: hand it a location and it returns the direction of flow there. A solution curve is the path of a leaf dropped onto the water; the leaf cannot choose its own direction, it follows the local current at every instant. Drop the leaf at different starting points and you trace different solution curves, but all of them obey the same map.

This picture makes the qualitative questions easy. Where the current is flat (the nullcline, zero slope), the flow neither rises nor falls; for an autonomous equation those flat lines are equilibria. Where solutions on both sides drift toward a flat line, it is a stable equilibrium, a place the flow settles into. Where they drift away, it is unstable. You can answer “what happens in the long run” by reading the map, long before you can solve for the leaf’s exact path.

Holding this view also explains why the next two methods exist. Euler’s method walks the leaf forward one short step at a time along the current (a numerical answer), and separation of variables finds a formula for the leaf’s path when the current is simple enough (a symbolic answer). The field is the shared geometric foundation underneath both.

Resources

Connections

Within MATH347 (Differential Equations)

• Introduction to differential equations: Reading a solution’s slope directly from the equation, introduced there, becomes the systematic direction-field method here.
• Euler’s method: The numerical companion to the direction field. It steps along the field dash by dash to approximate a solution curve when no formula is available.
• Separable equations: When the field comes from a separable equation, the curves it suggests can be found exactly by separating variables. The field is the geometric preview; separation is the symbolic confirmation.
• The logistic equation: The logistic field $x' = x(1 - x)$ shows its stable carrying-capacity equilibrium and unstable zero equilibrium at a glance.

Built From

• Tangent slope via limits (MATH161): A field dash is a tangent slope, the calculus idea developed through limits.
• Increasing and decreasing test (MATH161): Reading where solutions rise or fall from the sign of $f$ reuses this test.
• Evaluating functions (MATH161): Building a field requires evaluating $f(t, x)$ at many points.

Audience Notes

For students who find math intimidating: You do not need to solve anything to draw a direction field. Pick a point, plug it into the right side, get a number, and draw a little dash with that slope. Repeat. The picture tells the story of the solutions even when the algebra is hard. A wrong turn here is common and shows something worth understanding; there is no need to rush.

For students interested in proof: That solution curves do not cross rests on the existence-and-uniqueness theorem: under mild conditions on $f$, exactly one solution passes through each point. Where those conditions fail, curves can in fact meet, which is itself an instructive edge case.

For students interested in careers: Direction fields and their higher-dimensional cousins, phase portraits, are everyday tools in control engineering, ecology, and dynamical-systems analysis. Engineers read stability of an equilibrium straight off the picture before committing to a design.

For gifted and curious students: For systems of two autonomous equations, the direction field becomes a phase plane, and equilibria can be spirals, saddles, or centers rather than simple stable or unstable points. The predator-prey (Lotka-Volterra) model appears in this richer geometry, and the single-variable stability analysis here is the first step toward it.

Back to MATH347 Skills | Previous: Modeling with Differential Equations | Next: Euler’s Method