Differential Equations › Section A · Basic concepts › Topic A.2
Direction fields
A first-order equation tells you the slope of its solutions at every point, so you can draw them before you can solve them. A direction field turns the equation into a picture that shows how solutions behave and how that behaviour depends on where they start.
A picture before a formula
Many first-order equations cannot be solved with a formula, and even when they can, the formula may hide the behaviour you care about: does the solution level off, grow without bound, or depend sharply on its starting value? A direction field answers those questions directly from the equation. The Calculus topic "Slope fields and Euler's method" introduced the idea. This topic develops it as a working tool: drawing families of integral curves, using isoclines to sketch quickly, and reading long-term behaviour from the picture, including how it depends on the initial condition.
Building a direction field
Write the equation in the form \(y' = f(t, y)\). At each point of a grid in the \(ty\)-plane, evaluate \(f(t, y)\) and draw a short segment, or an arrow pointing right, with that slope. Positive values give rising segments, negative values falling ones, zero gives a horizontal segment, and large values give steep ones. The result is the direction field, also called the slope field. By hand you draw perhaps twenty or thirty segments to see the pattern; software draws hundreds in an instant. Either way, the picture contains no new information: it is the equation, drawn.
Integral curves
A solution curve of the equation is called an integral curve. At every point it passes through, it must be tangent to the segment drawn there, so integral curves are the curves that follow the field, like a leaf carried along a stream. Sketching a handful of them, starting from different heights, shows the whole family of solutions at once. For the equations in this topic, where \(f\) and its derivative with respect to \(y\) are continuous, two integral curves never cross or touch. The topic "Existence, uniqueness, and intervals of validity" explains why.
Worked example: a field without t
Sketch the field of \(y' = 6 - 2y\) and describe the solutions as \(t \to \infty\). 1. The slope depends only on \(y\), so every segment in a horizontal row is the same. 2. Tabulate rows: \(y = 3\) gives slope 0, \(y = 4\) gives \(-2\), \(y = 5\) gives \(-4\), \(y = 2\) gives 2 and \(y = 0\) gives 6. 3. Read the picture: above \(y = 3\) curves fall, below it they rise, and the horizontal line \(y = 3\) is itself a solution. 4. Conclusion: every solution approaches 3, whatever \(y(0)\) is. Check: the solution \(y = 3 + (y(0) - 3)e^{-2t}\) confirms it.
Isoclines
Drawing a field point by point is slow. A faster way is to find the isoclines, the curves along which the slope has a fixed value \(c\). They are the curves \(f(t, y) = c\). Along each isocline, every segment has the same slope, so you draw one curve and stamp identical segments along it. The isocline for \(c = 0\) is the most useful, because it marks where integral curves are horizontal: when \(f\) changes sign across it, solutions rise on one side and fall on the other. If an isocline happens to be a curve whose own slope equals \(c\), it is itself a solution.
Worked example: isoclines of a linear equation
Use isoclines to sketch \(y' = x - 2y\). 1. The isoclines are \(x - 2y = c\), that is \(y = \frac{x}{2} - \frac{c}{2}\): parallel lines of slope \(\frac{1}{2}\). 2. The \(c = 0\) isocline is \(y = \frac{x}{2}\); above it slopes are negative, below it positive. 3. The isocline with \(c = \frac{1}{2}\) has slope \(\frac{1}{2}\), equal to the field's slope along it, so the line \(y = \frac{x}{2} - \frac{1}{4}\) is a solution. 4. Check: its derivative is \(\frac{1}{2}\), and \(x - 2\left(\frac{x}{2} - \frac{1}{4}\right) = \frac{1}{2}\). Every other integral curve approaches this line as \(x\) grows.
Fields with and without t
When \(f\) depends on \(y\) alone, the equation is called autonomous and its field has a strong symmetry: slide the picture left or right and nothing changes. Shifting any integral curve sideways then gives another integral curve, and horizontal lines where \(f(y) = 0\) are constant solutions that no other curve can cross. When \(f\) involves the independent variable, as \(y' = x - 2y\) involves \(x\), the field changes from column to column and long-term behaviour may follow a moving curve, such as a line, rather than a constant. The topic "Equilibrium solutions and stability" studies the autonomous case in detail.
Worked example: a drug infusion
In an illustrative model, a drip delivers a drug at 10 mg per hour, and the body removes it at a rate of 0.5 times the amount present, so \(A' = 10 - 0.5A\), assuming a steady drip and clearance proportional to \(A\). 1. The slope is 0 when \(A = 20\), positive below and negative above. 2. So a patient starting with no drug climbs toward 20 mg and never exceeds it, while a patient starting above 20 mg drifts down to it. 3. The field predicts the level of 20 mg without solving. 4. Check: solving gives \(A = 20(1 - e^{-t/2})\), which is about 17.3 mg after 4 hours and approaches 20 mg.
What a field can and cannot tell you
A direction field shows qualitative behaviour reliably: where solutions rise and fall, which constant or straight-line solutions attract or repel neighbours, and how the outcome depends on the starting value. It also warns you of trouble, such as segments steepening without limit as curves race upward. What it cannot give is precise numbers; reading a value of \(y(2)\) from a hand sketch is a rough estimate at best. For numbers you need a formula from the solving methods that follow in this section, or a numerical method such as the one in "Euler's method and error".
Common mistakes to avoid
First, swapping the coordinates: at the point \((t, y) = (1, 3)\) evaluate \(f(1, 3)\), not \(f(3, 1)\). Second, drawing integral curves that cut across segments instead of running tangent to them, or that cross each other. Third, assuming an isocline is a solution; it is one only when its own slope equals the slope on it. Fourth, treating a field with \(t\) in it as if every row had one slope. Fifth, claiming exact values from a sketch, or a long-run limit without checking curves that start on both sides of a constant solution.
Key takeaways
For \(y' = f(t, y)\), a direction field draws a segment of slope \(f(t, y)\) at grid points, and integral curves follow those segments without crossing. Isoclines \(f(t, y) = c\) carry segments of equal slope; the \(c = 0\) isocline marks where solutions turn. An isocline whose own slope equals \(c\) is a solution. Autonomous fields repeat along each row, and their horizontal solutions organise the long-term behaviour. Fields give reliable qualitative answers, including how the outcome depends on \(y(0)\), but numbers need formulas or numerical methods.
Flashcards
Flashcard
Short segments of slope f(t, y) drawn at grid points
Test your understanding
Question 1 of 5
In the direction field of y′ = y − t², what is the slope of the segment at the point (t, y) = (2, 1)?
Question 2 of 5
For y′ = x + 2y, which line is the isocline along which every slope equals 4?
Question 3 of 5
For y′ = 8 − 2y, what happens to the solution with y(0) = 10 as t → ∞?
Question 4 of 5
In the direction field of an autonomous equation y′ = f(y), how do the segments along one horizontal line compare?
Question 5 of 5
For y′ = x − y, the isoclines are lines y = x − c. Which one is itself a solution?
