How this calculation works
Differentiate the curve components symbolically, substitute the curve into the planar field and integrate the field/velocity dot product over the entered parameter bounds. Inspect every coarse node, a half-step comparison, sampled field arrows and the path with distinct start/end markers.
∫C F·dr = ∫[P(x(t),y(t))x′(t)+Q(x(t),y(t))y′(t)] dtA worked example
For F=(−y,x) and r(t)=(cos t,sin t), the velocity is (−sin t,cos t) and the dot product is one. The upper semicircle from t=0 to π has work π. A full counterclockwise circle gives 2π, and reversing its parameter bounds changes the sign.
Good to know
Planar vector line integrals only, not scalar ∫f ds. Bounded field expressions in x,y and curve expressions in t. Fixed Simpson integration requires even intervals and finite sampled path/velocity/field values. Finite sampling and half-step agreement do not certify smoothness, accuracy or absence of singularities. Numerical field arrows are not trajectories.
Reference: OpenStax Calculus: line integrals (opens in a new tab) ↗
Distinguish vector work from scalar arc-length accumulation
This page calculates the vector line integral ∫C F·dr for a planar field F=(P,Q). The dot product compares the field with the directed path velocity. It can accumulate positive and negative contributions, depending on whether the field points with or against the traversal.
A scalar line integral such as ∫C f ds is a different operation. It weights a scalar by arc length and uses a speed factor rather than the field/velocity dot product. Giving both calculations the same generic area label would hide a meaningful mathematical difference. When F represents a force, its vector integral can represent work if the force and coordinate units are consistent.
The Stokes theorem numerical verification calculator separately compares three-dimensional curl flux with a lifted rectangular boundary circulation. Numerical agreement alone does not establish its theorem hypotheses.
Reduce the path calculation to one parameter
Write the curve as r(t)=(x(t),y(t)) with stated start and end parameter values. Differentiate the two coordinate expressions to obtain x′(t) and y′(t). Then evaluate P and Q at the curve position, not at the parameter value itself. The resulting scalar integrand is P(x(t),y(t))x′(t)+Q(x(t),y(t))y′(t).
The table shows the parameter, position, symbolic-velocity values, field components and dot product at each coarse Simpson node. Keeping those quantities separate helps reveal errors such as substituting t for both coordinates or multiplying a component by the wrong velocity entry. The symbolic velocity expressions remain in the steps for an independent check.
Follow the rotational field around a semicircle
For F=(−y,x) and r(t)=(cos t,sin t), the field on the curve is (−sin t,cos t). The velocity is the same vector. Their dot product is sin²t+cos²t=1. Over t=0 to π, the curve follows the upper semicircle from (1,0) to (−1,0), giving integral π.
Extending the parameter end to 2π traces a full counterclockwise circle and gives 2π. Reversing the bounds gives the opposite orientation and changes the sign. The field and path picture retains separate start/end markers so an otherwise identical geometric curve is not mistaken for an identical directed traversal.
Read arrows as local field samples rather than motion predictions
The diagram draws sampled field arrows on a finite grid around the curve. One shared positive display factor preserves their relative sampled magnitudes and directions within that view. The red path is a separate object supplied by its parameterization. A field arrow does not prove that a moving particle would follow the entered path.
Undefined field values away from the path are omitted and counted. On the actual integration samples, the path, velocity and field must all be finite; an undefined required value stops the calculation rather than becoming a zero contribution. A finite grid can still miss singularities between nodes, so the picture does not establish that the field is smooth throughout the region.
Keep parameter speed and orientation separate
Under the usual regularity conditions, an orientation-preserving reparameterization traces the same directed geometric path and gives the same vector line integral. Changing speed does not by itself change the result because the velocity factor and parameter interval change together. Reversing orientation is different and changes the sign.
For example, r(t)=(cos(2t),sin(2t)) from zero to π/2 traces the same upper semicircle as the original example. Its dot product becomes two, but its parameter interval is half as long, so the total remains π. This reasoning does not authorize an arbitrary nonsmooth or discontinuous parameter change without checking the assumptions.
Use numerical comparisons without inventing a proof
Composite Simpson integration combines endpoint and interior samples with alternating weights. Its coarse interval count must be even. The tool also runs the same integral with twice as many intervals, reporting both values and their absolute difference. The table contains the coarse nodes used for the main answer.
Agreement between those runs can be useful evidence for a well-resolved smooth integrand, but it is not a certified error bound. Two grids can miss the same narrow feature or align with an oscillation. The tool does not prove smoothness, field conservativeness, path independence or that an unobserved singularity is absent.
The trapezoidal and Simpson integration calculator studies a scalar finite-interval quadrature directly. The Green’s theorem calculator uses a separate region-and-boundary relationship under its stated assumptions. Keep the specific path, orientation and model with a copied result or SVG rather than treating these related tasks as interchangeable.
A few questions, answered
Is this a scalar line integral ∫f ds?
No. It calculates the planar vector work/circulation integral ∫F·dr.
What happens if I reverse the parameter bounds?
The traversal orientation reverses and the signed integral changes sign when the model remains valid.
Does tracing the same path faster change the work?
An orientation-preserving regular reparameterization gives the same vector line integral under the usual assumptions.
Does zero net work mean the field was zero everywhere?
No. Contributions can cancel, or the field can be perpendicular to the path velocity.
Is the half-step difference an exact error estimate?
No. It compares numerical resolutions and is not a certified bound.