The Disappearing Constant in Definite Integral Antiderivatives

The Fundamental Theorem of Calculus says this: if $f$ is continuous on $[a, b]$ and $F$ is any antiderivative of $f$ on that interval, then

$$\int_a^b f(x)\,dx = F(b) - F(a).$$

The word carrying the weight is any. If $F$ is an antiderivative of $f$, so is $F + C$ for every constant $C$, because differentiating destroys the constant. Feed $F + C$ into the theorem and watch:

$$\big[F(b) + C\big] - \big[F(a) + C\big] = F(b) - F(a).$$

The constant is added once and subtracted once, so it is gone before you reach the answer. This is not a convention someone agreed on. It is what the subtraction does.

Carry the $+C$ through if it steadies you:

$$\int_1^2 (3x^2 + 2)\,dx = \Big[x^3 + 2x + C\Big]_1^2 = (8 + 4 + C) - (1 + 2 + C) = 12 + C - 3 - C = 9.$$

Writing $9 + C$ on the final line is wrong, and not on a technicality. A definite integral is one number. The expression $9 + C$ stands for a family of numbers, one for each value of $C$, so it answers a different question. A complete solution evaluates an antiderivative at both limits, subtracts, and finishes on a number.

In one study, 150 calculus students at two American colleges were asked open questions about what a definite integral means^[1]. The two answers that came up most often were the area under a curve and the antiderivative; far fewer described it as adding up many small pieces^[1].

That second answer is the one to sit with. If a definite integral simply is the antiderivative in your head, the antiderivative's constant travels along with it. Dropping it then feels like discarding part of the answer. That study measured what students say an integral means, not what they do with $C$, so that step is my inference rather than the paper's^[1].

There is a second way to describe this. Researchers have put forward an account in which a new mathematical procedure starts life as a ritual^[2]. The steps get carried out because those are the steps. Only later are they carried out for the result they are meant to produce. This is a framework those authors proposed, not something the paper measured^[2]. "Always add $+C$" is a rule triggered by the sight of an integral sign, not a question you have asked.

The theorem underneath is genuinely hard, well past this stage. In a small teaching study, 19 senior and graduate mathematics students struggled with the Fundamental Theorem^[3]. The researcher traced that to shaky ideas about rate of change, and about how two changing quantities move together^[3]. It is a small, unusual group, and the study says nothing about constants of integration^[3].

The constant is not decoration; some problems fall apart without it. Take a particle with velocity $v(t) = 3t^2 + 2$, whose position at time $1$ is $s(1) = 5$. Antidifferentiating gives

$$s(t) = t^3 + 2t + C.$$

The given condition pins $C$ down. Substituting $t = 1$ gives $1 + 2 + C = 5$, so $C = 2$ and $s(t) = t^3 + 2t + 2$. Check it: differentiating returns $3t^2 + 2$, and $s(1) = 1 + 2 + 2 = 5$. Here $C$ has one specific value, and omitting it gives the wrong position.

The two problems now fit together. The position at time $2$ is $s(2) = 8 + 4 + 2 = 14$, which needed $C$. The change in position from time $1$ to time $2$ is $s(2) - s(1) = 14 - 5 = 9$, the definite integral from the top of this post, which never needed $C$. The constant tells you where you are. The definite integral tells you how far you moved. Subtracting throws away the first question and keeps the second.

Sometimes the constant arrives under a different letter. Evaluate

$$\int_3^5 4x\,dx = \Big[2x^2 + K\Big]_3^5 = (50 + K) - (18 + K) = 50 + K - 18 - K = 32.$$

The $K$ cancels exactly as $C$ did. The letter was never the point. Symbolic limits behave no differently: $\int_a^b f(x)\,dx = F(b) - F(a)$, an expression in $a$ and $b$ with no constant attached. No definite integral keeps a constant of integration. An indefinite one always does: with nothing to subtract, $\int (3x^2 + 2)\,dx = x^3 + 2x + C$, and dropping that $C$ is a real error, not a tidy-up.

Here is a habit worth trying, borrowed from research on other topics and younger learners^[4][5]: do not study only correct work.

In one experiment, 74 fourth- and fifth-graders learning about decimals were split at random into two groups^[4]. The group that compared a wrong worked answer against a right one learned the idea better, and held fewer wrong beliefs afterwards, than the group comparing two correct answers^[4]. That was young children in a brief one-to-one session^[4].

Closer to this reader's age, two experiments ran inside real first-year algebra classes, the first of them with students solving linear equations^[5]. Those who explained worked examples as part of their practice understood the ideas better than those who only practised^[5]. The same work found signs that wrong examples may be especially useful for that understanding, though the researchers called this a suggestion from their results, not a settled finding^[5].

Applied here: write both final lines in your notes, $9 + C$ and $9$, with a sentence underneath saying which answers the question and why the other does not. A review that gathered up the studies on explaining work to yourself found a small to moderate improvement on tests taken right after the lesson^[6]. Prompting learners to say why a common mistake is wrong is something that review recommends on the strength of that work, not a result it measured^[6].

One last problem. Evaluate $\int_0^1 (6x + 3)\,dx$. Carrying the constant through gives $\Big[3x^2 + 3x + C\Big]_0^1 = (3 + 3 + C) - (0 + 0 + C) = 6 + C - C = 6$. Now write one line in your own words: why is $6 + C$ not an answer to that question?

None of these studies measures how often students actually write $9 + C$. The study of 150 calculus students asked what a definite integral means, not which errors show up on homework^[1]. My sense that the slip is common comes from watching people work, not from a finding.

Nobody has tested wrong-example practice on this particular mistake either. The decimal experiment involved fourth- and fifth-graders, and the algebra work sat inside first-year algebra classes^[4][5]. A calculus student facing a constant of integration is a different case, and borrowed advice can fail.

The review of explaining your work is honest about its own edges. Its improvement showed up on tests taken right after the lesson, while the evidence that the habit helps in an ordinary classroom, or still helps after a delay, is much thinner^[6]. Writing one sentence about $6 + C$ costs a minute, which is why I would still do it. That is a judgment about cost, not a promise from the research^[6].

The account of a procedure performed as a ritual is a proposed framework, and its authors did not measure calculus students^[2]. It gives you a name for the habit, not evidence that the habit is yours.