Why the Order of the Bounds Matters
A definite integral records signed accumulation across an interval. The bounds specify not only which interval is involved, but also its orientation: \(\int_a^b f(x)\,dx\) proceeds from \(a\) to \(b\). If the bounds are reversed, the direction of accumulation is reversed too.
This orientation is reflected directly in the Fundamental Theorem of Calculus. When the integrand has an antiderivative, evaluate the antiderivative at the upper bound and subtract its value at the lower bound, even when the upper bound is numerically smaller. Reversing the bounds reverses the sign of the integral, but the resulting sign also depends on whether the integrand is positive or negative.
The Endpoint Formula with Reversed Bounds
The Fundamental Theorem of Calculus, Part II, states that if \(f\) is continuous on an interval containing \(a\) and \(b\), and \(F\) is an antiderivative of \(f\) there, then
The formula applies whether \(a<b\), \(a=b\), or \(a>b\). In the reversed case, the numerical upper bound is smaller than the lower bound, but it remains the first value in the endpoint difference: evaluate at the stated upper bound, then subtract the value at the stated lower bound.
For comparison, the integral with the bounds in increasing order is \(\int_v^u f(x)\,dx=F(u)-F(v)\). Therefore the two integrals are negatives of one another:
Why Reversing the Bounds Negates the Integral
The sign change is part of the meaning of an oriented integral. For \(v<u\), the integral from \(u\) to \(v\) is defined as the negative of the integral from \(v\) to \(u\). This convention lets the endpoint formula work consistently in either direction.
By the Fundamental Theorem, \(\int_v^u f(x)\,dx=F(u)-F(v)\), provided \(f\) is continuous and \(F'=f\) on the interval.
By the definition of orientation, \(\int_u^v f(x)\,dx=-\int_v^u f(x)\,dx\).
Thus \(\int_u^v f(x)\,dx=-[F(u)-F(v)]=F(v)-F(u)\), which is exactly the endpoint formula with the bounds as stated.
This derivation explains why it would be incorrect to evaluate \(F(u)-F(v)\) after reversing the bounds: that difference belongs to the integral from \(v\) to \(u\), not the integral from \(u\) to \(v\).
Worked Example: A Positive Integrand with Reversed Bounds
Worked Example: A Positive Integrand with Reversed Bounds
Evaluate \(\int_4^1 (x+2)\,dx\), then compare its value with the geometric area between the graph and the \(x\)-axis.
The integrand \(x+2\) is positive throughout \([1,4]\). An antiderivative is \(F(x)=\frac{x^2}{2}+2x\), since \(F'(x)=x+2\).
Apply the endpoint formula in the order of the given bounds:
The geometric area over the same interval is positive. With bounds in increasing order, its value is \(F(4)-F(1)=16-\frac52=\frac{27}{2}\). Thus the reversed integral is \(-\frac{27}{2}\), while the area is \(\frac{27}{2}\).
Check: Differentiating \(F(x)=\frac{x^2}{2}+2x\) gives \(F'(x)=x+2\), the integrand. Also, reversing the bounds gives \(\int_4^1(x+2)\,dx=-\int_1^4(x+2)\,dx\), consistent with the computed values.
Signed Integral and Geometric Area
When \(f(x)\ge 0\) on an interval, the integral in increasing order gives the area between the graph and the \(x\)-axis. Reversing the bounds makes that integral negative, even though the geometric region has not become a negative area. When \(f(x)\le 0\), the integral in increasing order is nonpositive, while the geometric area is nonnegative.
More generally, a definite integral is signed accumulation: contributions where the function is above the \(x\)-axis count positively, and contributions where it is below count negatively. Unsigned geometric area counts both kinds of regions positively. If the function changes sign, the unsigned area is not in general the absolute value of the signed integral, because positive and negative contributions can cancel.
Orientation and the sign of the integrand are separate issues. Reversing the bounds always negates the signed integral. It does not mean the resulting answer must be negative: if the integral in increasing order is negative, reversing its bounds makes the answer positive.
Worked Example: A Negative Integrand with Reversed Bounds
Worked Example: A Negative Integrand with Reversed Bounds
Evaluate \(\int_3^1 -2x\,dx\). This example shows why reversed bounds do not always give a negative answer.
An antiderivative is \(F(x)=-x^2\), because \(F'(x)=-2x\). On \([1,3]\), the integrand \(-2x\) is negative.
The answer is positive. Indeed, the integral in increasing order is \(F(3)-F(1)=-9-(-1)=-8\), so reversing the bounds changes it to \(8\). The positive answer does not mean the graph lies above the \(x\)-axis: the function is negative on the interval, and the direction of integration is reversed.
Check: \(F'(x)=\frac{d}{dx}(-x^2)=-2x\). The endpoint difference \(F(1)-F(3)=-1-(-9)=8\) confirms the evaluated integral.
Worked Example: Reversing Bounds for a Trigonometric Integrand
Worked Example: Reversing Bounds for a Trigonometric Integrand
Evaluate \(\int_{\pi/2}^0 \cos x\,dx\).
An antiderivative of \(\cos x\) is \(F(x)=\sin x\), since \(F'(x)=\cos x\). The integrand is nonnegative on \([0,\pi/2]\).
In increasing order, the integral is \(\int_0^{\pi/2}\cos x\,dx=1\), which also equals the geometric area here because \(\cos x\ge0\) on this interval. Reversing the bounds changes the signed integral to \(-1\), not the area of the region.
Check: Differentiating \(\sin x\) gives \(\cos x\), and the endpoint evaluation \(0-1=-1\) agrees with the rule that reversing the bounds negates the integral.
When the Integrand Changes Sign
If the function changes sign within the interval, the signed integral combines contributions with opposite signs. Reversing the bounds negates that combined result, but does not remove any cancellation that occurred. Consequently, the absolute value of the integral can be smaller than the unsigned geometric area.
For example, consider \(f(x)=x-1\) on \([-1,2]\). The function is negative for \(x<1\) and positive for \(x>1\). Its signed integral is computed from the antiderivative \(F(x)=\frac{x^2}{2}-x\):
Therefore \(\int_2^{-1}(x-1)\,dx=\frac32\). The reversed integral is the negative of the forward signed integral. However, because the function has both negative and positive values on the interval, the magnitude \(\frac32\) is not the total unsigned geometric area: the signed contributions partially cancel in the forward integral.
A Common Sign Error
A frequent mistake is to reverse the bounds mentally but keep the endpoint subtraction in the original order.
Worked Example: Correcting an Endpoint-Order Error
Consider \(\int_5^2 3x^2\,dx\). An antiderivative is \(F(x)=x^3\), since \(F'(x)=3x^2\).
Incorrect calculation: Using \(F(5)-F(2)\) gives \(125-8=117\). This evaluates the integral from \(2\) to \(5\), not from \(5\) to \(2\).
Correction: Preserve the stated bound order and subtract the lower-bound value from the upper-bound value:
Check: \(F'(x)=3x^2\), as required. Also, \(\int_2^5 3x^2\,dx=117\), so the reversed integral is \(-117\).
A Reliable Evaluation Routine
When evaluating a definite integral with reversed bounds, use the following checks:
- Confirm that the integrand is continuous on the interval and identify an antiderivative.
- Write the endpoint formula with the bounds exactly as given: upper-bound value minus lower-bound value.
- Substitute each endpoint carefully, including parentheses when an endpoint is negative.
- Interpret the sign as signed accumulation, not as a statement that geometric area is negative.
- If useful, compare with the integral whose bounds are reversed; the two values must be negatives of each other.
The sign check is especially useful: if changing only the order of the bounds does not change the sign of your result, revisit the endpoint subtraction.
Check Your Understanding
Use the endpoint formula and distinguish signed integral from geometric area where appropriate.
- Evaluate \(\int_3^0 (x+1)\,dx\). Is the result the same as the geometric area?
- If \(\int_1^4 f(x)\,dx=-6\), what is \(\int_4^1 f(x)\,dx\)?
- Can an integral with reversed bounds be positive? Explain how the sign of the integrand affects the answer.
- For \(f(x)=x-2\) on \([0,4]\), is the absolute value of the signed integral necessarily the unsigned geometric area? Explain.
- State the endpoint formula for \(\int_b^a f(x)\,dx\) when \(a<b\) and \(F'=f\).
