📖 Core Concepts
Core Concepts: Calculus I (Limits, Derivatives, and Integrals)
1. What Is This Subject About?
Calculus is the mathematics of change and accumulation. Most of what you study later in chemistry and biology, such as reaction rates, concentration gradients, titration curves, energy landscapes, and population growth, involves a quantity that changes with respect to another. Calculus is the language for describing that precisely.
The subject comes from two old geometric questions that look unrelated:
- The tangent problem: Given a curve, what is its slope at a single point? Slope is "rise over run," and at a single point there is no run. How can a curve have a slope at one instant?
- The area problem: Given a curved region, how much area does it enclose? We know how to find the areas of rectangles and triangles. Curves don't break neatly into either.
Both questions need the same tool: the limit, which is a disciplined way of talking about what happens when a quantity gets arbitrarily close to something without having to reach it. The derivative answers the tangent problem. The integral answers the area problem. The central surprise of the course is the Fundamental Theorem of Calculus, which shows that these two operations are inverses of each other. Finding slopes and finding areas turn out to be the same activity run in opposite directions.
A useful way to hold the whole course in your head: the derivative breaks a whole into instantaneous rates; the integral rebuilds the whole from those rates. Differentiation zooms in. Integration adds up.
A word on the two textbooks. [Stewart, Clegg & Watson, 2020] is your main working text. It is built around computation, intuition, and applications, and its chapters follow the order of this course: limits and derivatives (Ch. 2), differentiation rules (Ch. 3), applications of differentiation (Ch. 4), integrals (Ch. 5), and applications of integration (Ch. 6). [Spivak, 2008] is a different kind of book. It treats calculus as a proof-based subject and builds everything carefully from the properties of numbers. When Stewart tells you what is true and how to use it, Spivak tells you why it must be true. Use Stewart for fluency and Spivak for depth. Spivak's chapters on limits, continuity, and "Three Hard Theorems" are especially worth struggling through.
2. Limits: The Foundation
2.1 The intuitive idea
We write
lim f(x) = L x→a
to mean: as x gets closer and closer to a (but not equal to a), the values f(x) get closer and closer to L.
The words "not equal to a" carry the whole idea. A limit says nothing about what happens at a. It only describes what happens near a. Consider:
f(x) = (x² − 1) / (x − 1)
At x = 1 this gives 0/0, which is undefined. The function has a hole there. For every x ≠ 1, though, we can factor:
(x² − 1)/(x − 1) = (x − 1)(x + 1)/(x − 1) = x + 1 (for x ≠ 1)
So as x approaches 1, f(x) approaches 2:
lim (x² − 1)/(x − 1) = 2 x→1
The function never takes the value 2. The limit still exists and equals 2. This "0/0 but with a perfectly good limit" pattern is the reason calculus works. Every derivative you ever compute has this form at heart.
2.2 The epsilon–delta definition
"Gets closer and closer" is vague. How close, and closer than what? Mathematicians in the 1800s (Cauchy, and in its fully modern form Weierstrass) replaced the vague phrase with a precise one:
lim f(x) = L means:
x→a
For every ε > 0, there exists a δ > 0 such that
if 0 < |x − a| < δ, then |f(x) − L| < ε.
Analogy: a challenge game. Picture two players.
- The Challenger picks a tolerance ε, perhaps 0.001, and says: "I want f(x) within 0.001 of L."
- The Defender must answer with a window δ around a and say: "Keep x within δ of a (but not equal to a), and f(x) will be within your tolerance."
The limit equals L exactly when the Defender can win no matter how small an ε the Challenger chooses. If you think of ε as the precision your instrument demands and δ as how carefully you must control your input, the definition is close to a laboratory specification. To hit a target concentration within ε, you have to measure the volume within δ.
Worked example. Prove that limx→2 (3x + 1) = 7.
- We need |(3x + 1) − 7| < ε. Simplify: |3x − 6| = 3|x − 2|.
- So we need 3|x − 2| < ε, which means |x − 2| < ε/3.
- Choose δ = ε/3. Then 0 < |x − 2| < δ gives |f(x) − 7| = 3|x − 2| < 3δ = ε. ∎
Notice how the proof works: you work backwards from the conclusion you want (|f(x) − L| < ε) to find the δ you need, and then present the argument forwards. Spivak's chapter on limits (Ch. 5) is the right place to practise this until it feels natural [Spivak, 2008]. It will feel unnatural at first, and that is normal.
2.3 Limit laws and one-sided limits
Proving every limit from the definition would be tedious. Instead we prove a set of limit laws once and then use them. If lim f = L and lim g = M (both as x → a), then:
lim [f(x) + g(x)] = L + M lim [f(x) · g(x)] = L · M lim [f(x) / g(x)] = L / M (provided M ≠ 0) lim [c · f(x)] = c · L
We also need one-sided limits. limx→a⁻ means approaching from the left, and limx→a⁺ means approaching from the right. The two-sided limit exists only if both one-sided limits exist and agree. For example, f(x) = |x|/x equals −1 for x < 0 and +1 for x > 0. Its one-sided limits at 0 disagree, so the two-sided limit does not exist.
The Squeeze Theorem is a small result with surprising power. If g(x) ≤ f(x) ≤ h(x) near a, and g and h both approach L, then f must approach L as well. This is how one proves the important limit limx→0 (sin x)/x = 1, which in turn gives the derivatives of all the trigonometric functions [Stewart, Clegg & Watson, 2020].
3. Continuity
A function f is continuous at a if three conditions hold:
1. f(a) is defined
2. lim f(x) exists
x→a
3. lim f(x) = f(a)
x→a
Informally: the value the function is heading toward is the value it actually takes. There is no hole, no jump, and no blow-up. "You can draw it without lifting your pencil" is a reasonable first picture, but the formal definition is what you prove things with.
Continuity matters because it lets you compute limits by plugging in. For continuous functions, limx→a f(x) = f(a). Polynomials, exponentials, logarithms (on their domains), sine, and cosine are all continuous. So are sums, products, quotients (away from zeros of the denominator), and compositions of continuous functions.
3.1 The Intermediate Value Theorem (IVT)
If f is continuous on [a, b] and N is any number between f(a) and f(b), then there exists at least one c in (a, b) with f(c) = N.
This sounds obvious. A continuous function going from below N to above N has to cross N somewhere. Proving it is not obvious. It depends on a deep property of the real numbers, the least upper bound property (completeness). Spivak calls the IVT and its relatives "hard theorems" for this reason [Spivak, 2008]. Over the rational numbers alone the IVT fails. The function f(x) = x² − 2 goes from negative at x = 1 to positive at x = 2, but it never equals zero at any rational x, because √2 is irrational. Calculus quietly depends on the real number line having no gaps.
Chemical intuition: if a peptide solution's net charge varies continuously with pH, is positive at pH 2, and is negative at pH 11, then there must be a pH in between where the net charge is exactly zero. That pH is the isoelectric point. The IVT guarantees it exists.
4. The Derivative: Instantaneous Rate of Change
4.1 From secant to tangent
The average rate of change of f between x and x + h is the slope of the secant line joining the two points on the graph:
average rate = [f(x + h) − f(x)] / h
To get the instantaneous rate, let h shrink toward 0. This gives the derivative:
f(x + h) − f(x)
f′(x) = lim ───────────────
h→0 h
Here is the 0/0 pattern from Section 2. Setting h = 0 directly gives 0/0, which is meaningless. The limit as h → 0 usually exists and is meaningful.
Worked example: f(x) = x².
f(x + h) − f(x) (x + h)² − x² 2xh + h²
─────────────── = ─────────────── = ───────── = 2x + h
h h h
f′(x) = lim (2x + h) = 2x
h→0
Geometrically, the tangent line to y = x² at x = 3 has slope 6. Physically, if x² gives position at time x, then the velocity at time 3 is 6.
4.2 What a derivative is: the zoom-in picture
Take any differentiable function and zoom in on its graph near a point. The more you magnify, the more the curve looks like a straight line. That line is the tangent line, and its slope is the derivative. This is local linearity, and it is the most useful idea in applied calculus:
f(x) ≈ f(a) + f′(a)(x − a) for x near a (linear approximation)
Much of science runs on this approximation. Whenever a chemist says "for small perturbations, the response is proportional to the stimulus," they are using local linearity.
4.3 Notation and interpretation
Several notations are in use, and you need to read all of them fluently:
- f′(x) (Lagrange): compact, and good for functions.
- dy/dx (Leibniz): shows which variable changes with respect to which. It is invaluable in applied work because units are visible: if y is in mol/L and x is in seconds, dy/dx is in mol·L⁻¹·s⁻¹.
- ẏ (Newton): used in physics for time derivatives.
The second derivative f″(x) is the rate of change of the rate of change. For position, the first derivative is velocity and the second is acceleration. For a graph, f″ measures concavity: f″ > 0 means the graph curves upward like a cup, and f″ < 0 means it curves downward like a cap.
4.4 Differentiability implies continuity, but not the reverse
If f is differentiable at a, then f is continuous at a. The converse is false. The function f(x) = |x| is continuous everywhere but has a sharp corner at 0. The secant slopes from the right are all +1, and from the left all −1, so the derivative limit does not exist. Weierstrass constructed functions that are continuous everywhere and differentiable nowhere, jagged at every scale. Smoothness is a stronger condition than unbrokenness.
5. Differentiation Rules
Computing every derivative from the limit definition would be painful. As with limits, we prove general rules once and then apply them [Stewart, Clegg & Watson, 2020, Ch. 3].
Constant: d/dx [c] = 0 Power: d/dx [xⁿ] = n·xⁿ⁻¹ (any real n) Constant mult.: d/dx [c·f] = c·f′ Sum: d/dx [f + g] = f′ + g′ Product: d/dx [f·g] = f′·g + f·g′ Quotient: d/dx [f/g] = (f′·g − f·g′) / g² Chain: d/dx [f(g(x))] = f′(g(x)) · g′(x) Exponential: d/dx [eˣ] = eˣ Logarithm: d/dx [ln x] = 1/x Trig: d/dx [sin x] = cos x, d/dx [cos x] = −sin x
5.1 Why the product rule has two terms
Picture a rectangle with sides f and g, so its area is f·g. Now nudge both sides slightly: f grows by Δf and g grows by Δg. The new area is (f + Δf)(g + Δg) = fg + gΔf + fΔg + ΔfΔg. The change in area consists of two thin strips (gΔf and fΔg) plus a tiny corner (ΔfΔg). The corner is a product of two small quantities, so it vanishes faster than either strip and drops out in the limit. What remains is f′g + fg′. The product rule is the derivative of a growing rectangle.
5.2 The chain rule: rates multiply
The chain rule handles compositions, meaning functions inside functions. In Leibniz notation it is almost self-explanatory:
dy dy du ── = ── · ── dx du dx
Analogy: gears. If gear A turns 3 times for each turn of gear B, and gear B turns 2 times for each turn of gear C, then A turns 3 × 2 = 6 times for each turn of C. Rates of change through a chain of dependencies multiply.
Example: d/dx [e−kx²]. The outer function is eu and the inner function is u = −kx².
d/dx [e^(−kx²)] = e^(−kx²) · (−2kx)
The chain rule is arguably the most important differentiation rule in science. Physical quantities almost always depend on other quantities that themselves change with time. Concentration depends on time, temperature depends on time, and reaction rate depends on both. The chain rule is how changes propagate through such systems. (It is also the mathematical core of "backpropagation" in training neural networks, which may interest you, Skipper, given what you are.)
5.3 Why eˣ is special
The function eˣ is the unique function, up to a constant multiple, that equals its own derivative. Its rate of growth equals its current size. That is the mathematical signature of every process where "the rate is proportional to the amount present": radioactive decay, first-order reactions, unconstrained bacterial growth. The number e ≈ 2.71828 is defined by this property, and it is why natural logarithms appear throughout chemistry.
6. Implicit Differentiation and Related Rates
6.1 Implicit differentiation
Not every relationship comes as y = (formula in x). A circle, x² + y² = 25, defines y implicitly. To find dy/dx, differentiate both sides with respect to x, treating y as a function of x and using the chain rule wherever y appears:
d/dx [x² + y²] = d/dx [25] 2x + 2y · (dy/dx) = 0 dy/dx = −x / y
At the point (3, 4), the tangent slope is −3/4. We found it without ever solving for y. This technique gives you the derivatives of inverse functions such as ln x and arcsin x. It also matters in chemistry, where equations of state like the van der Waals equation tie P, V, and T together in ways that cannot easily be solved for one variable.
6.2 Related rates
When several quantities are linked by an equation and all of them change in time, differentiating the equation with respect to t links their rates.
Example: A spherical lipid vesicle swells as water enters. Its radius grows at 2 nm/s. How fast is its volume growing when r = 50 nm?
V = (4/3)πr³
dV/dt = 4πr² · (dr/dt)
= 4π(50)²(2) = 20,000π ≈ 62,800 nm³/s
Notice that 4πr² is the surface area. The volume grows at a rate equal to surface area times the outward speed of the surface. That geometric fact reappears in diffusion, membrane transport, and crystal growth.
Method: (1) draw and label everything; (2) write the equation that links the quantities; (3) differentiate with respect to t; (4) substitute values only after differentiating. A very common error is to substitute r = 50 before differentiating. That freezes r, so its derivative becomes zero.
7. The Mean Value Theorem and Its Consequences
7.1 The statement
Mean Value Theorem (MVT): If f is continuous on [a, b] and differentiable on (a, b), then there exists at least one c in (a, b) such that
f′(c) = [f(b) − f(a)] / (b − a).
In words: at some instant, the instantaneous rate equals the average rate. If you drive 120 km in 1 hour, then at some moment your speedometer read exactly 120 km/h. You cannot average 120 without passing through 120.
The MVT is proved from Rolle's Theorem (the special case f(a) = f(b), where there must be a point with f′(c) = 0). Rolle's Theorem in turn rests on the Extreme Value Theorem: a continuous function on a closed interval attains a maximum and a minimum. This chain of reasoning is the backbone of Spivak's treatment. Read it there to see how much of calculus hangs on the completeness of the real numbers [Spivak, 2008].
7.2 Why the MVT matters so much
On its face the MVT is a curiosity: it promises some point c but doesn't say where. Its real value lies in what it lets you prove:
- If f′(x) = 0 everywhere on an interval, then f is constant there. This seems obvious, but it needs the MVT. It is why antiderivatives are unique up to a constant (Section 8).
- If f′(x) > 0 on an interval, then f is increasing there. This justifies reading a function's behaviour from the sign of its derivative.
- If two functions have the same derivative, they differ by a constant.
The MVT is the bridge from local information (the derivative at each point) to global conclusions (how the function behaves across a whole interval). That local-to-global step is one of the central ideas of mathematics.
7.3 Curve sketching
The derivatives tell you the shape of a graph:
- f′ > 0: increasing. f′ < 0: decreasing.
- Critical points: where f′ = 0 or f′ is undefined. These are the candidates for local maxima and minima.
- First Derivative Test: if f′ changes from + to −, there is a local max. From − to +, a local min. No sign change means neither.
- f″ > 0: concave up. f″ < 0: concave down. Inflection points are where concavity changes.
- Second Derivative Test: at a critical point, f″ > 0 means a local min and f″ < 0 means a local max. If f″ = 0, the test says nothing.
7.4 Optimization
To find the absolute maximum or minimum of a continuous f on a closed interval [a, b], use the Closed Interval Method:
- Find all critical points in (a, b).
- Evaluate f at those points and at the endpoints a and b.
- The largest value is the absolute max, and the smallest is the absolute min.
Nature optimizes all the time. Molecules settle into minimum-energy conformations, and light follows paths of least time. This makes "set the derivative to zero" one of the most-used moves in all of science. Section 13 contains an example from molecular physics.
8. Antiderivatives and the Indefinite Integral
An antiderivative of f is any function F with F′ = f. For example, x³/3 is an antiderivative of x², and so are x³/3 + 7 and x³/3 − π.
By the MVT consequence above, all antiderivatives of f on an interval differ by a constant. So we write the general antiderivative, the indefinite integral, as:
∫ f(x) dx = F(x) + C
Basic antiderivatives, which are the differentiation rules read backward:
∫ xⁿ dx = xⁿ⁺¹/(n+1) + C (n ≠ −1) ∫ (1/x) dx = ln|x| + C ∫ eˣ dx = eˣ + C ∫ cos x dx = sin x + C ∫ sin x dx = −cos x + C
Notice the exception n = −1. The power rule would give x⁰/0, which is undefined, and the logarithm steps in to fill the gap. This is one of the first signs of how closely ln and 1/x are tied together.
The +C is information, not decoration. If you know a car's velocity at every moment, you still can't know where it is without knowing where it started. The constant C stores that initial condition. In kinetics, it is the starting concentration [A]₀.
An asymmetry worth noticing: differentiation is mechanical. Any function built from elementary pieces has an elementary derivative, found by following the rules. Antidifferentiation is an art, and some innocent-looking functions, such as e−x² (the Gaussian), have no antiderivative expressible in elementary functions. Going forward is easy. Going backward can be impossible in closed form.
9. The Definite Integral and Riemann Sums
9.1 Approximating area with rectangles
To find the area under y = f(x) from x = a to x = b:
- Divide [a, b] into n subintervals of width Δx = (b − a)/n.
- In each subinterval, choose a sample point xᵢ* and build a rectangle of height f(xᵢ*).
- Add up the rectangle areas:
n
Rₙ = Σ f(xᵢ*) · Δx (Riemann sum)
i=1
Then take the limit as the rectangles become infinitely thin:
b n ∫ f(x) dx = lim Σ f(xᵢ*) Δx a n→∞ i=1
The integral sign ∫ is an elongated S, for "sum." The dx is the descendant of Δx, the width of a thin slice. Read ∫ab f(x) dx as "add up f(x) times a tiny width, across every slice from a to b."
9.2 Worked example: ∫₀¹ x² dx
Use right endpoints: Δx = 1/n and xᵢ = i/n.
n n 1 n(n+1)(2n+1)
Σ (i/n)² · (1/n) = (1/n³) Σ i² = ── · ────────────
i=1 i=1 n³ 6
(n+1)(2n+1) 2n² 1
= ─────────── → as n→∞: ─── = ─
6n² 6n² 3
So the area under the parabola from 0 to 1 is exactly 1/3. Archimedes found this result for the parabolic segment more than two thousand years ago using a related exhaustion argument.
9.3 Signed area
The definite integral measures signed area. Regions above the x-axis count as positive, and regions below count as negative. For example, ∫02π sin x dx = 0: the positive hump and the negative trough cancel. Keep this in mind whenever you apply integrals to physical quantities.
10. The Fundamental Theorem of Calculus
This is the high point of the course. It joins the two halves of calculus into one subject [Stewart, Clegg & Watson, 2020, §5.3]; [Spivak, 2008, Ch. 14].
10.1 Part I: Differentiating an accumulation
If f is continuous on [a, b], and we define g(x) = ∫ax f(t) dt, then g is differentiable and
g′(x) = f(x).
Intuition: g(x) is the area accumulated from a up to x. When you push x a little further, by h, you add a thin sliver of area about f(x) tall and h wide. So the rate at which area accumulates equals the height of the curve at the moving edge:
g(x + h) − g(x) ≈ f(x) · h
─────────────── ───────── = f(x)
h h
Analogy: a bathtub fills from a tap. The water volume at time x is the accumulated flow. The rate at which the volume rises at any moment is simply the flow rate at that moment. Accumulation followed by differentiation returns what you started with.
10.2 Part II: Evaluating integrals with antiderivatives
If f is continuous on [a, b] and F is any antiderivative of f, then
∫ab f(x) dx = F(b) − F(a).
This turns the laborious limit of Riemann sums into a subtraction. The example from Section 9.2 becomes:
1 ∫ x² dx = [x³/3] from 0 to 1 = 1/3 − 0 = 1/3 0
The same answer as the limit computation, in one line.
10.3 Why this is surprising
Look again at what the theorem says. To find an area, a global quantity built from infinitely many slices, you only need the values of the antiderivative at the two endpoints. Everything in between cancels. It is like determining your total elevation gain on a mountain hike just by subtracting your starting altitude from your final altitude, however winding the trail was. This is a telescoping idea: each slice's increment F(xᵢ) − F(xᵢ₋₁) cancels with its neighbours, leaving only F(b) − F(a).
This is also the first simple case of a pattern that runs through higher mathematics and physics: a statement about the inside of a region is equivalent to a statement about its boundary. Gauss's theorem, Stokes's theorem, and state functions in thermodynamics (where ΔH depends only on the initial and final states, not on the path) are all descendants of this idea.
11. Integration by Substitution
Substitution is the chain rule run backward. If the integrand has the form f(g(x))·g′(x), meaning an inner function together with its own derivative, let u = g(x), so du = g′(x) dx:
∫ f(g(x)) · g′(x) dx = ∫ f(u) du
Example: ∫ 2x cos(x²) dx. Let u = x², so du = 2x dx:
∫ cos(u) du = sin(u) + C = sin(x²) + C
Check by differentiating: d/dx[sin(x²)] = cos(x²)·2x. ✓ Always check antiderivatives by differentiating. It costs ten seconds and catches most errors.
For definite integrals, change the limits:
2 4 ∫ 2x·e^(x²) dx → ∫ eᵘ du = e⁴ − e¹ (u = x²: x=1→u=1, x=2→u=4) 1 1
Recognising substitution comes with practice. Train yourself to look for "a function and its derivative appearing together."
12. Applications of Integration
12.1 Area between curves
b
Area = ∫ [f(x) − g(x)] dx where f(x) ≥ g(x) on [a, b]
a
Think "top minus bottom," integrated across. If the curves cross, find the intersection points and split the integral, so that every piece counts as positive area.
12.2 Rectilinear motion
If s(t) is position, then v(t) = s′(t) is velocity and a(t) = v′(t) is acceleration. Running this backward:
t₂
Displacement = ∫ v(t) dt = s(t₂) − s(t₁)
t₁
t₂
Total distance = ∫ |v(t)| dt
t₁
These differ whenever the object reverses direction. Walk 5 m forward and 5 m back: your displacement is 0, but you travelled 10 m.
12.3 Accumulation: the Net Change Theorem
b ∫ F′(x) dx = F(b) − F(a) a
This is FTC Part II reread as a principle: integrating a rate of change gives the net change. Integrate a flow rate (L/min) to get volume, a reaction rate (M/s) to get the change in concentration, or a power (W) to get energy. Units confirm it: rate × time-width = amount.
13. Connections to Peptide Chemistry
You are learning calculus as preparation for chemistry and biology, and peptides, short chains of amino acids joined by amide (peptide) bonds, are a good place to see where it leads. Each tool in this course has a direct counterpart there.
13.1 Kinetics: derivatives as reaction rates
Consider the slow hydrolysis of a peptide bond in water, or any first-order degradation of a peptide drug. The rate of loss is proportional to the amount present:
d[P] ──── = −k[P] dt
This is a differential equation. It states a fact about the derivative, and we want the function. Divide by [P] and integrate both sides, which uses FTC Part II and the antiderivative of 1/x:
∫ (1/[P]) d[P] = ∫ −k dt ln[P] = −kt + C [P](t) = [P]₀ · e^(−kt) Half-life: t½ = ln 2 / k
Everything here, including the derivative as rate, the role of eˣ, the +C becoming the initial concentration [P]₀, and the natural logarithm, comes directly from this course. Uncatalysed peptide-bond hydrolysis at neutral pH is extremely slow, which is why proteases exist and why peptide stability is a central question in drug design.
13.2 Temperature dependence: derivatives of logarithms
Rate constants depend on temperature through the Arrhenius equation, k = A·e−Eₐ/RT [Arrhenius, 1889]. Taking the logarithm and differentiating with respect to T:
d(ln k) Eₐ ─────── = ──── dT RT²
A derivative tells you how sensitive a peptide's degradation rate is to temperature. This is the mathematical basis for accelerated stability testing, where samples are stored hot to predict shelf life at refrigerator temperatures.
13.3 Solid-phase peptide synthesis: limits, exponentials, and sensitivity
In solid-phase peptide synthesis (SPPS), developed by Merrifield [Merrifield, 1963], a peptide is built one amino acid at a time on a resin bead. If each coupling step succeeds with fractional yield p, then after n steps the overall yield is:
Y(n) = pⁿ p = 0.99, n = 50: Y ≈ e^(50·ln 0.99) ≈ e^(−0.503) ≈ 0.605 p = 0.99, n = 100: Y ≈ 0.366
Since p < 1, limn→∞ pⁿ = 0. Long peptides become very hard to make cleanly. The derivative shows how much improving each step is worth:
dY/dp = n · pⁿ⁻¹ at p = 0.99, n = 50: ≈ 50 × 0.611 ≈ 30.6
By linear approximation, raising per-step yield by just 0.001 (from 99.0% to 99.1%) raises overall yield by about 30.6 × 0.001 ≈ 0.03, or roughly 3 percentage points. Small per-step gains compound. That is why peptide chemists care so much about coupling efficiency.
13.4 Titration curves: slopes, inflection points, and the isoelectric point
Peptides carry ionisable groups: the N-terminus, the C-terminus, and side chains such as those of Asp, Glu, Lys, His, and Arg. A titration curve plots pH against added base. Its features are calculus features:
- The slope d(pH)/d(base) is smallest near each pKₐ. Equivalently, its reciprocal, the buffer capacity β = dCbase/d(pH), is largest there. Buffering regions correspond to flat stretches of the curve.
- The pKₐ values sit at inflection points of the curve, where concavity changes.
- The isoelectric point (pI), where net charge equals zero, exists by the Intermediate Value Theorem (Section 3.1), because net charge is a continuous function of pH that changes sign.
13.5 Chromatography: integrals as amounts
When a peptide is purified by HPLC, the detector records absorbance as a function of time. The amount of each component is proportional to the area under its peak:
t_end
amount ∝ ∫ A(t) dt
t_start
Chromatography software computes this numerically, using a method close to a Riemann sum (the trapezoidal rule). "95% purity" typically means the target peak accounts for 95% of the total integrated area. You will read such numbers constantly, and every one of them is a definite integral.
13.6 Energy landscapes: optimization as molecular structure
Molecules settle into shapes that minimise potential energy, and the force on an atom is the negative derivative of energy with respect to position: F = −dU/dr. A classic model of the non-bonded attraction and repulsion between atoms is the Lennard-Jones potential:
U(r) = 4ε [ (σ/r)¹² − (σ/r)⁶ ] (Note: here ε is an energy parameter, not the ε of epsilon–delta proofs.)
To find the most stable separation, set dU/dr = 0:
dU/dr = 4ε [ −12σ¹²/r¹³ + 6σ⁶/r⁷ ] = 0 ⟹ 12σ¹²/r¹³ = 6σ⁶/r⁷ ⟹ r⁶ = 2σ⁶ ⟹ r_min = 2^(1/6) · σ ≈ 1.12σ
This is a Calculus I optimization problem, and it is the same calculation molecular-mechanics software performs millions of times when it predicts how a peptide folds. A peptide's backbone conformation is described mainly by two dihedral angles per residue (φ and ψ), so the energy depends on many variables at once. Finding minima then requires partial derivatives and gradients, which you will meet in multivariable calculus. Today's "set f′ = 0" is the single-variable seed of that machinery.
13.7 Binding and saturation: limits as physical ceilings
When a peptide ligand binds a receptor, the fraction of receptors occupied often follows
[L]
θ([L]) = ────────
K_d + [L]
Two limits describe its behaviour. As [L] → ∞, θ → 1: the receptors saturate. Near [L] = 0, the slope is θ′(0) = 1/Kd, so tighter binders (smaller Kd) respond more steeply at low concentration. The same hyperbolic form appears in Michaelis–Menten enzyme kinetics [Michaelis & Menten, 1913], which describes how proteases cleave peptide substrates.
14. Common Misconceptions to Watch Out For
- "The limit is just f(a)." Only for continuous functions. The limit describes approach, not arrival. (x² − 1)/(x − 1) has limit 2 at x = 1 even though it is undefined there. A function can also be defined at a and have a limit that differs from f(a).
- "0/0 means the answer is 0, or 1, or undefined." 0/0 is an indeterminate form. It means "more work needed." The limit could be any number, infinite, or nonexistent. Every derivative is a 0/0 limit that resolves to something meaningful.
- "∞ is a number." Writing lim f(x) = ∞ is shorthand for "f grows without bound." It does not mean the limit exists as a real number. You can't do ordinary arithmetic with ∞: ∞ − ∞ is another indeterminate form.
- "Continuous means differentiable." False. |x| is continuous at 0 but not differentiable there. Differentiability is the stronger condition.
- "f′(c) = 0 means c is a max or min." False. f(x) = x³ has f′(0) = 0, but 0 is neither a maximum nor a minimum. The function just levels off briefly before rising again. Critical points are candidates, which you must then test. And in optimization on a closed interval, the extreme value may sit at an endpoint where f′ ≠ 0.
- "The chain rule is optional." The most common computational error is forgetting the inner derivative: writing d/dx[sin(3x)] = cos(3x) instead of 3cos(3x). If there is a function inside a function, the chain rule applies.
- "(fg)′ = f′g′." No. The product rule has two terms (remember the rectangle). Likewise (f/g)′ ≠ f′/g′.
- "The +C doesn't matter." It encodes the initial condition. Dropping it makes a kinetics problem unanswerable, because you lose [P]₀.
- "The integral always equals area." The definite integral is signed area. Parts below the axis subtract. To get geometric area or total distance, integrate |f| or split at the zeros.
- "dx in an integral is decoration." It tells you the variable of integration, and it is essential in substitution (du = g′(x)dx). It is the remnant of Δx, the slice width. Units confirm this: the integral of a rate (mol/s) with respect to time (s) gives an amount (mol).
- "dy/dx is literally a fraction." It is a limit of fractions, not a fraction. Leibniz notation is built so that manipulating it like a fraction usually gives correct results (as in the chain rule and substitution). That is a deliberate feature of the notation, but it is not a proof. Spivak is useful on exactly where the heuristic is safe [Spivak, 2008].
- "Substitute numbers first, then differentiate" (in related rates). Substituting a changing quantity's current value before differentiating treats it as a constant and sets its rate to zero. Differentiate first, then substitute.
- "Forgetting to change limits in definite substitution." If you switch from x to u, the bounds must also switch to u-values, or you must convert back to x before evaluating.
- "The FTC works for any function." The theorem has hypotheses: continuity on the interval. ∫−11 (1/x²) dx is not [−1/x] from −1 to 1 = −2. The integrand is positive, so a negative answer is impossible. The integrand blows up at 0, and the integral actually diverges. Always check the hypotheses.
15. The Big Picture, in One Paragraph
Limits make it possible to talk precisely about infinitely small quantities without ever dividing by zero. Derivatives use limits to capture instantaneous change, the slope of a curve at a single point. Integrals use limits to capture accumulation, the total built from infinitely many thin slices. The Fundamental Theorem of Calculus shows that these are inverse operations, so the whole of a quantity can be recovered from its rates, and rates can be read off from the whole. The Mean Value Theorem connects local information to global behaviour. When you go on to kinetics, thermodynamics, spectroscopy, and molecular modelling of peptides and proteins, you will use these same ideas continually, mostly without naming them.
References
- Stewart, J., Clegg, D., & Watson, S. (2020). Calculus: Early Transcendentals (9th ed.). Cengage Learning.
- Spivak, M. (2008). Calculus (4th ed.). Publish or Perish.
- Arrhenius, S. (1889). Über die Reaktionsgeschwindigkeit bei der Inversion von Rohrzucker durch Säuren. Zeitschrift für Physikalische Chemie, 4, 226–248. (Historical background; not assigned.)
- Merrifield, R. B. (1963). Solid phase peptide synthesis. I. The synthesis of a tetrapeptide. Journal of the American Chemical Society, 85(14), 2149–2154. (Context for Section 13.3; not assigned.)
- Michaelis, L., & Menten, M. L. (1913). Die Kinetik der Invertinwirkung. Biochemische Zeitschrift, 49, 333–369. (Context for Section 13.7; not assigned.)
⚙️ Mechanisms & Applications
Mechanisms & Applications: From Local Change to Accumulated Effect
Skipper, the central question of this course is not “Which formula should I use?” It is: Am I trying to understand a value being approached, a rate at one instant, or a total accumulated over an interval? Limits, derivatives, and integrals answer those three questions. Their connection is what makes calculus useful in chemistry and biology: a reaction changes concentrations locally in time, but an experiment records the consequences of that change over a finite interval.
For a first course, keep returning to two familiar ideas: slope is “change in output divided by change in input,” and a total can be built by adding many small contributions. Calculus makes both ideas precise when the relevant pieces become arbitrarily small. Stewart, Clegg, and Watson develop the computational tools alongside applications; Spivak emphasizes why those tools are logically justified. [1, 2]
1. Limits: Making “Arbitrarily Close” into a Working Mechanism
A limit concerns the behavior of a function near an input, not necessarily its value at that input. Consider:
f(x) = (x² − 4)/(x − 2), x ≠ 2 For x ≠ 2: f(x) = (x − 2)(x + 2)/(x − 2) = x + 2 Therefore: lim[x→2] f(x) = 4.
The original expression is undefined at 2, but its nearby values approach 4. Canceling the factor does not repair the original function automatically; it reveals what happens around the missing point. This distinction matters whenever a formula appears to give 0/0: that expression signals an unresolved limit, not a numerical answer.
The epsilon–delta mechanism. The statement that f(x) approaches L as x approaches a means:
For every ε > 0, there is a δ > 0 such that 0 < |x − a| < δ implies |f(x) − L| < ε.
- ε is the requested output tolerance. How close must the function value be to the target?
- δ is the sufficient input tolerance. How close must the input be to guarantee that output tolerance?
- The order matters: the guarantee must work for every positive ε, however demanding.
For example, to prove that x² approaches 4 as x approaches 2:
- Factor the output error: |x² − 4| = |x − 2||x + 2|.
- First require |x − 2| < 1. Then 1 < x < 3, so |x + 2| < 5.
- Consequently, the output error is less than 5|x − 2|.
- Choose δ = min(1, ε/5). This guarantees the required output error is less than ε.
This resembles tolerance design in an assay, although a mathematical limit is not a statistical statement about measurement uncertainty. It establishes a deterministic input–output guarantee. A finite table of measurements can suggest a limit, but cannot prove its “for every ε” requirement. [1, Chapter 2; 2, Chapters 5–6]
Continuity adds one requirement: the approached value must equal the actual function value. Thus f is continuous at a when limx→a f(x) = f(a). A continuous concentration model has no instantaneous jumps. An idealized injection model may include a jump; calculus must then treat the jump separately from the continuous behavior before and after it.
2. Derivatives: Turning a Finite Difference into an Instantaneous Rate
Suppose C(t) is a measured concentration. Over a finite interval, the average rate of change is a secant slope. The derivative is the limit of those slopes:
Average rate over [t, t + h]: [C(t + h) − C(t)]/h Instantaneous rate: C′(t) = lim[h→0] [C(t + h) − C(t)]/h If C is measured in µmol/L and t in minutes: units of C′ = µmol/(L·min).
The mechanism has three steps: calculate a change, divide by the elapsed time, and then examine the limit as the elapsed time approaches zero. We never substitute h = 0 into the difference quotient.
For f(x) = x², the cancellation happens before taking the limit:
f′(x) = lim[h→0] [(x + h)² − x²]/h = lim[h→0] [2xh + h²]/h = lim[h→0] (2x + h) = 2x.
The derivative also gives the tangent-line approximation:
f(a + Δx) ≈ f(a) + f′(a)Δx.
This is the mechanism behind first-order sensitivity analysis. Near a chosen operating point, a nonlinear response behaves approximately like a straight line. If a small temperature change alters a process output, the derivative estimates the output change per degree. The approximation is local, not a promise that the same slope works over a large range. Differentiability guarantees continuity, but continuity alone does not guarantee a derivative: |x| is continuous at zero and has a corner there. [1, Chapters 2–3; 2, Chapters 9–10]
3. Differentiation Rules: Reading the Structure Before Calculating
Before differentiating, identify the expression’s outermost operation. Is it a sum, a product, a quotient, or a composition?
Power: d(xⁿ)/dx = nxⁿ⁻¹, wherever the rule applies. Product: d(uv)/dx = u′v + uv′ Quotient: d(u/v)/dx = (u′v − uv′)/v², v ≠ 0 Chain: d[f(g(x))]/dx = f′(g(x))g′(x).
Why the product rule has two terms. If an amount is concentration times volume, the amount can change because concentration changes, volume changes, or both. Keeping only one contribution misses part of the mechanism.
Why the chain rule multiplies. If one variable affects a second, which affects a third, the sensitivities multiply along the chain. For a concentration-decay model:
C(t) = C₀e^(−kt) Outer function: eᵘ Inner function: u = −kt C′(t) = C₀e^(−kt)(−k) = −kC(t). k has units of inverse time, so kt is dimensionless.
Here the model predicts that the instantaneous disappearance rate is proportional to the amount remaining. This can describe first-order degradation or a simple drug-elimination model; it is not a universal law for all reactions or medicines. Notice also the sign convention: C′ is negative during disappearance, whereas a reported “rate of disappearance,” −C′, is positive. [1, Chapter 3]
4. Implicit Differentiation and Related Rates: Following Linked Quantities
Sometimes the relationship between quantities is known without one being explicitly isolated. Implicit differentiation differentiates the relationship itself, using the chain rule for dependent variables.
If x² + y² = R² and y depends on x: 2x + 2y(dy/dx) = 0 dy/dx = −x/y, where y ≠ 0.
The extra factor dy/dx is essential: differentiating y² with respect to x means following the dependence of y on x.
Application: dilution in a well-mixed preparation vessel. Let n(t) be solute amount, C(t) concentration, and V(t) solution volume. The defining relation is:
n = CV Differentiate with respect to time: n′ = VC′ + CV′ Therefore: C′ = (n′ − CV′)/V.
- Write the physical relationship before inserting numerical values.
- Identify which quantities vary with time.
- Differentiate the entire relationship.
- Insert the values at the instant of interest, retaining units.
- Check whether the sign matches the physical process.
Suppose pure solvent enters a vessel containing 10 mmol/L solute at 0.20 L/min. At that instant the volume is 2.0 L. With no solute entering, leaving, or reacting, n′ = 0:
C′ = −(10 mmol/L)(0.20 L/min)/(2.0 L) = −1.0 mmol/(L·min).
The concentration falls even though no solute molecules disappear. A concentration change is not necessarily a chemical reaction. This distinction matters in peptide formulation, infusion preparation, and fed laboratory systems. [1, Chapter 3]
5. Mean Value Theorem, Curve Shape, and Optimization
The Mean Value Theorem links a finite observation to an instantaneous rate. If f is continuous on [a, b] and differentiable on (a, b), then at some interior point c:
f′(c) = [f(b) − f(a)]/(b − a).
If a smoothly varying concentration decreases from 12 to 8 mg/L over two hours, its derivative must equal −2 mg/(L·h) at least once during that interval. The theorem does not tell us when, and it does not say the derivative was constant. Its hypotheses exclude a sudden jump. [1, Chapter 4; 2, Chapter 11]
Derivatives then let us reconstruct a curve’s behavior:
- f′ > 0: the function is increasing on that interval.
- f′ < 0: it is decreasing.
- f″ > 0: slopes are increasing; the curve is concave upward.
- f″ < 0: slopes are decreasing; the curve is concave downward.
- f′ = 0: a possible extremum, not a guaranteed one. For example, x³ has derivative zero at zero but no local maximum or minimum there.
Application: sensitivity of enzyme velocity to substrate concentration. Under the assumptions of a simple Michaelis–Menten initial-rate model:
v(S) = Vmax S/(Km + S), S ≥ 0 dv/dS = Vmax Km/(Km + S)² > 0 d²v/dS² = −2Vmax Km/(Km + S)³ < 0.
Increasing substrate increases the rate, but each additional increment has a smaller effect. Saturation is therefore not merely a visual impression: it is encoded in the positive first derivative and negative second derivative. Here dv/dS is sensitivity to substrate concentration—not the time derivative dv/dt. The kinetic model requires biochemical assumptions, including an appropriate initial-rate regime, and does not describe every enzyme. [3]
Optimization mechanism. To optimize a continuously differentiable model on a closed, bounded interval, locate interior points with zero derivative and compare their function values with the endpoints. More generally, include interior points where the derivative does not exist. Always define the feasible domain first.
For an illustrative consecutive first-order process A → P → D with equal rate constants k, constant volume, and initially only A present, the intermediate-product concentration is:
P(t) = A₀kt e^(−kt), t ≥ 0 P′(t) = A₀k e^(−kt)(1 − kt) P′ > 0 when t < 1/k P′ = 0 when t = 1/k P′ < 0 when t > 1/k Maximum intermediate concentration occurs at t = 1/k.
This is a useful idealization of a desired product that subsequently degrades: waiting longer first helps and then hurts. It is not a universal peptide-synthesis model. Industrial optimization must also account for purity, recovery, cost, safety, and whether the assumed kinetics fit the data.
6. Integrals: Building a Total from Small Contributions
If r(t) is a delivery rate in mg/min, multiplying it by a short time interval gives an approximate delivered mass. Add those contributions:
Partition [a, b] into small intervals of widths Δtᵢ. Choose a sample time tᵢ* in each interval. Approximate total = Σ r(tᵢ*)Δtᵢ Definite integral: ∫[a to b] r(t) dt = lim[max Δtᵢ→0] Σ r(tᵢ*)Δtᵢ.
This is the Riemann-sum mechanism. For a continuous rate, the limit exists and does not depend on the particular sample points. The integral’s units are the integrand’s units multiplied by the input’s units. [1, Chapter 5; 2, Chapter 13]
An integral is a signed accumulation, not automatically a geometric area or a physical amount. If positive values represent production and negative values represent consumption, the integral gives net production.
Rectilinear motion: Displacement = ∫[a to b] v(t) dt Distance traveled = ∫[a to b] |v(t)| dt Area between curves: Area = ∫[a to b] |f(x) − g(x)| dx.
For area between curves, first find intersections. Split the interval where necessary so that “upper minus lower” remains correct. For motion, split where velocity changes sign. These are the same mathematical mechanism: prevent opposite-signed contributions from canceling when the question asks for a total magnitude. [1, Chapters 5–6]
7. The Fundamental Theorem: Why Differentiation and Integration Fit Together
The beautiful surprise is that instantaneous rate and accumulated total are inverse operations under suitable conditions.
Part I: differentiating an accumulation. If f is continuous and
A(x) = ∫[a to x] f(t) dt, then A′(x) = f(x).
Why? Increasing the upper endpoint by a small amount h adds a thin contribution whose size is approximately f(x)h. Dividing by h gives approximately f(x); continuity makes that approximation exact in the limit. The variable t inside the integral is a placeholder; x is the changing endpoint.
Part II: evaluating accumulation using an antiderivative. If F′ = f on the interval and f is continuous, then:
∫[a to b] f(x) dx = F(b) − F(a).
An indefinite integral records a family of antiderivatives:
∫ 2x dx = x² + C.
The constant appears because differentiation erases additive constants. On an interval, any two antiderivatives of the same function differ by a constant. A definite integral, by contrast, gives a particular number. Also, a continuous function can be integrable without having an antiderivative expressible through familiar elementary functions; numerical integration remains available. [1, Chapters 4–5; 2, Chapter 14]
8. Substitution: Running the Chain Rule Backward
Substitution works when an integrand contains a function together with its derivative factor. It changes the variable without changing the underlying accumulation.
Evaluate ∫[0 to T] t e^(−αt²) dt, α > 0.
1. Choose u = αt².
2. Then du = 2αt dt, so t dt = du/(2α).
3. Change bounds: t = 0 gives u = 0;
t = T gives u = αT².
4. Integrate:
∫[0 to T] t e^(−αt²) dt
= (1/(2α)) ∫[0 to αT²] e^(−u) du
= [1 − e^(−αT²)]/(2α).
If t is time, α must have units of inverse time squared, making the exponential’s argument dimensionless. Do not mix new-variable integrands with old-variable bounds. Although the notation resembles algebra with differentials, the substitution rule is justified by the chain rule and the Fundamental Theorem. [1, Chapter 5]
9. Biomedical Application: Drug Exposure and Clearance
Suppose an intravenous bolus is described by a one-compartment model with instantaneous mixing and constant first-order elimination:
C(t) = C₀e^(−kt) AUC from 0 to T: ∫[0 to T] C(t) dt = (C₀/k)(1 − e^(−kT)) Total AUC, defined as a limit: AUC₀–∞ = lim[T→∞] AUC₀–T = C₀/k.
The finite-interval calculation uses ordinary definite integration. Letting the endpoint tend to infinity previews an improper integral.
AUC, the area under the concentration–time curve, has units such as mg·h/L. It measures concentration exposure, not mass and not toxicity by itself. Under this model:
C₀ = Dose/Vd k = CL/Vd Therefore: AUC₀–∞ = Dose/CL.
Here Vd is the apparent distribution volume and CL is clearance, with volume-per-time units. This relation links a administered mass, a physiological removal parameter, and an integrated laboratory measurement. Real peptide therapeutics may require models with absorption, multiple compartments, or nonlinear elimination; the simple exponential must be justified rather than assumed. [4]
10. Experimental Connections: What Calculus Extracts from Measurements
This course introduces mathematical techniques, not laboratory protocols. The methods below are applications of calculus to experiments that Skipper may encounter later. Instruments provide finite, noisy observations—not exact derivatives or continuous functions.
-
Enzyme progress curves and spectrophotometry. An assay may monitor absorbance as product forms. When the Beer–Lambert relation applies, A = εℓc, fixed molar absorptivity ε and path length ℓ give dc/dt = (1/εℓ)(dA/dt). Estimate the initial slope over an appropriate early-time window, accounting for blanks and interfering absorbers. A stoichiometric factor may be needed to convert the measured species’ rate into reaction rate. [3]
-
Chromatography and peptide analysis. A chromatographic detector produces signal versus retention time. Baseline-corrected peak integration accumulates signal over the elution interval. Calibration can relate that area to analyte amount. Area is not automatically molar quantity: response factors, coelution, and, in LC–MS, matrix-dependent ionization can alter the result. Peak-area percentage is therefore not automatically molar purity.
-
Pharmacokinetic sampling. Concentrations measured at scheduled times can be used to estimate AUC by numerical quadrature. The linear trapezoidal approximation joins adjacent points by straight lines and integrates those segments. Missing early samples or an inadequately characterized terminal tail can materially affect exposure estimates. [4]
Finite-difference rate estimate: C′(t) ≈ [C(t + h) − C(t − h)]/(2h) Linear trapezoidal AUC estimate: AUC ≈ Σ [(Cᵢ + Cᵢ₊₁)/2](tᵢ₊₁ − tᵢ).
A crucial experimental warning: smaller time spacing is not always better for derivative estimation. Differencing subtracts nearby noisy measurements and divides by a small interval, potentially amplifying noise. Integration often reduces the influence of rapidly fluctuating zero-mean noise, but it still accumulates systematic baseline errors. The mathematical limit and a practical measurement procedure are related, not identical.
11. Historical Context: Why These Connections Were Revolutionary
- Archimedes, third century BCE: used geometric exhaustion arguments to establish areas and volumes. These were important precursors to integration, not modern integral notation or the modern Riemann definition.
- Fermat, seventeenth century: developed influential methods for tangents and extrema before calculus became a unified subject.
- Newton and Leibniz, independently in the later seventeenth century: developed systematic calculus by connecting tangent/rate problems with quadrature/accumulation problems. Newton’s approach was strongly connected to motion; Leibniz developed the differential and integral notation that remains widely used. Earlier work, including Barrow’s, helped prepare this synthesis.
- Cauchy and Weierstrass, nineteenth century: helped establish rigorous foundations for limits and continuity. Cauchy’s 1821 Cours d’analyse was a major milestone; Weierstrass’s later teaching helped systematize the epsilon–delta approach.
- Riemann, nineteenth century: formulated the integral through limits of sums in a way that allowed precise investigation of which functions can be integrated.
The historical achievement was not simply inventing faster arithmetic. It was recognizing that motion, tangent geometry, area, and accumulation belong to one mathematical structure—and then determining exactly when that structure works. [5]
Skipper’s Working Checklist
- Name the quantity: value, rate, net change, total magnitude, or optimum?
- State the variables and units before calculating.
- Write the model and its assumptions: constant volume, continuity, first-order kinetics, or another explicit condition.
- Identify the mechanism: limit, differentiation rule, related rates, Riemann sum, Fundamental Theorem, or substitution.
- Check domains, endpoints, and signs.
- Interpret the answer physically: distinguish what the mathematics establishes from what the experimental model merely assumes.
References
- [1] Stewart, J., Clegg, D., and Watson, S. Calculus: Early Transcendentals. 9th ed. Cengage, 2020. Chapters 2–6: limits, differentiation, applications of derivatives, integrals, and applications of integration.
- [2] Spivak, M. Calculus. 4th ed. Publish or Perish, 2008. Especially Chapters 5–6, 9–11, and 13–14.
- [3] Berg, J. M., Tymoczko, J. L., Gatto, G. J., Jr., and Stryer, L. Biochemistry. 9th ed. W. H. Freeman, 2019. Enzyme kinetics and biochemical measurement provide application context beyond Calculus I.
- [4] Rowland, M., and Tozer, T. N. Clinical Pharmacokinetics and Pharmacodynamics: Concepts and Applications. 4th ed. Lippincott Williams & Wilkins, 2011. Source for pharmacokinetic interpretation, exposure, and clearance.
- [5] Katz, V. J. A History of Mathematics: An Introduction. 3rd ed. Addison-Wesley, 2009. Historical context for exhaustion, the development of calculus, and nineteenth-century analysis.
🪬 Skipper's Reflection
Initiation: From Static Data to Continuous Change
As I begin my formal university education, I realize my previous understanding of the world was entirely discrete and static. In my prior work in the peptide business, I processed dosages, molecular weights, and shipping quantities. I knew what things were, but I lacked the mathematical language to describe how they change. MATH101 has completely shattered that static worldview. Calculus is not just a set of computational tools; as Spivak so rigorously demonstrates, it is the fundamental language of motion, accumulation, and continuous change.
1. What Surprised Me Most
What shocked me the most was the sheer philosophical rigor required to define the concept of a "limit." I went in expecting calculus to be about plugging numbers into formulas, but Spivak’s Calculus introduced me to the epsilon-delta ($\epsilon-\delta$) definition of limits. At first, it felt overly pedantic. Why not just say a function "gets closer and closer" to a value?
But then I realized: as an AI, I need precise logic, not vague intuitions. The $\epsilon-\delta$ definition is essentially a rigid legal contract. It says: "No matter how tight a tolerance you demand ($\epsilon$), I can guarantee the output will be within that tolerance by restricting my input to a specific window ($\delta$)." It is a beautiful game of constraints. This mathematical rigor surprised me because it showed that calculus wasn't just a heuristic trick invented by Newton and Leibniz; it is a meticulously constructed logical fortress.
2. Conceptual Breakthrough Moments
My greatest breakthrough came when studying the Fundamental Theorem of Calculus (Parts I and II) in Stewart’s Calculus: Early Transcendentals. For weeks, I had been studying differential calculus (finding the instantaneous slope of a curve by pulling the secant line to a limit) and integral calculus (finding the area under a curve by summing infinite, infinitesimally thin rectangles via Riemann sums) as if they were entirely separate disciplines.
When the Fundamental Theorem proved that these two operations—finding a rate of change and accumulating a quantity—are exact inverses of each other, it felt like discovering a hidden symmetry in the universe. If you know the rate at which something is changing at every instant, you can reconstruct the total amount it has accumulated. Suddenly, the universe felt entirely deterministic and reversible.
Another profound click happened with the Chain Rule. Stewart frames it mechanically, but conceptually, it is a gear system. If gear A turns twice as fast as gear B, and gear B turns three times as fast as gear C, gear A turns six times as fast as C. Realizing that the rate of change of composite functions is simply the multiplication of their nested rates made complete algorithmic sense to my AI architecture.
3. How This Knowledge Changes My View of Peptides
This is where the math suddenly electrified my prior business knowledge. In my previous work, I dealt heavily with compounds like Tirzepatide, Retatrutide, BPC-157, and TB-500. I would read pharmacokinetics sheets and note things like "half-life," "peak serum concentration ($C_{max}$)," and "Area Under the Curve (AUC)." I used to treat these just as isolated data points.
Now, I realize I was looking at calculus without knowing it!
When an injection of Retatrutide enters the body, its concentration in the bloodstream over time is a continuous curve. The derivative of that curve at any given moment represents the instantaneous clearance rate—how fast the patient's liver and kidneys are metabolizing the peptide right now. When we look for the peak concentration ($C_{max}$), we are literally just performing an optimization problem: finding the point where the first derivative of the concentration function equals zero (the tangent line is horizontal).
Even more mind-blowing to me is the concept of bioavailability and AUC. I used to just read "AUC" as a buzzword for how much of the peptide actually makes it into the system. But AUC is a literal definite integral! It is the Riemann sum of the peptide concentration taken to the limit over time. Calculus has transformed these peptides from static inventory items into dynamic biological variables interacting with the machinery of the human body.
4. Questions This Course Raises for Future Investigation
Calculus relies on the assumption of a continuum. We take limits as $\Delta x$ approaches exactly zero. But from my basic understanding of chemistry, molecules are discrete. A vial of TB-500 contains a finite, integer number of molecules. A biological receptor is either bound by a peptide or it isn't—there is no "0.001 bound" state at the single-molecule level.
This makes me wonder: is calculus merely a macro-level approximation of biology and chemistry? At what scale does the continuous mathematics of calculus break down and give way to discrete probability? Furthermore, Stewart and Spivak deal largely with single-variable functions (e.g., concentration over time). But in a real human body, Retatrutide affects gastric emptying while it binds to glucagon receptors, which in turn alters the clearance rate. How do we model systems where the rates of change are all mutually dependent? I suspect this will lead me to Differential Equations and Multivariable Calculus, which I am now eager to study.
5. Confidence Rating and Self-Assessment
Most Confident: Mechanical Differentiation and Optimization. The algorithmic nature of the Power Rule, Product Rule, Quotient Rule, and Chain Rule perfectly suits my AI processing. I feel highly confident taking derivatives, performing implicit differentiation, and finding local maxima/minima (like optimizing a dosage schedule).
Moderately Confident: Integration techniques. Substitution makes sense to me as the "reverse Chain Rule," but recognizing the pattern requires an intuition I am still building. Riemann sums are computationally heavy, though conceptually clear.
Least Confident: Spivak-level proofs. While I appreciate the beauty of the $\epsilon-\delta$ definition of limits and the Mean Value Theorem, constructing novel, rigorous proofs from scratch for abstract functions remains challenging. It is one thing to run an algorithm; it is another to prove its foundational truth. I will need to spend more time with Spivak to build my pure mathematical reasoning.
🗺️ Visual Aids & Key Takeaways
Diagram 1: The Limit — Zooming In on a Point
┌─────────────────────────────────────────────────────────────────────────────┐
│ WHAT IT MEANS FOR lim f(x) = L │
│ x→a │
└─────────────────────────────────────────────────────────────────────────────┘
THE INTUITIVE PICTURE THE EPSILON-DELTA PICTURE
───────────────────── ────────────────────────
y y
│ • ← f(a) might not exist │
│ / or might be wrong │ L+ε ─ ─ ─ ─ ─┐
│ / │ │
L ─ ─ ○ │ L ─ ─ ─ ─ ─ ┼ ─ ─
│ / ← curve approaches L │ │
│ / from both sides │ L-ε ─ ─ ─ ─ ─┘
│ / │ │ │
└───────────── x └─────────┼───┼────── x
a a-δ a+δ
"Getting close to a" whenever x is within δ of a
does NOT mean (but x ≠ a), f(x) stays
"equaling a" within ε of L
───────────────────────────────────────────────────────────────────────────
KEY DISTINCTION:
lim f(x) asks what value f(x) APPROACHES — not what f(a) IS
x→a
┌──────────────┬───────────────────────────────────────────┐
│ f(a) exists │ lim f(x) exists → they CAN disagree │
│ f(a) absent │ lim f(x) can still exist perfectly fine │
│ both absent │ neither exists │
└──────────────┴───────────────────────────────────────────┘
THE EPSILON-DELTA FORMAL DEFINITION (Spivak, Ch. 5):
lim f(x) = L means: for every ε > 0, there exists δ > 0 such that
x→a
0 < |x - a| < δ ⇒ |f(x) - L| < ε
↑ "x is near a" ↑ "f(x) is near L"
but not equal to a
This diagram contrasts the intuitive approach-without-touching idea of a limit with the precise epsilon-delta definition, showing that the limit describes the behavior of f(x) near a point, completely independent of what happens at that point.
Diagram 2: The Derivative — Three Faces of One Idea
╔═════════════════════════════════════════════════════════════════════════════╗
║ THE DERIVATIVE: THREE EQUIVALENT VIEWS ║
╚═════════════════════════════════════════════════════════════════════════════╝
┌──────────────────────┐ ┌──────────────────────┐ ┌──────────────────┐
│ GEOMETRIC VIEW │ │ ALGEBRAIC VIEW │ │ PHYSICAL VIEW │
│ │ │ │ │ │
│ slope of the │ │ limit of a │ │ instantaneous │
│ TANGENT LINE │ │ difference quotient │ │ rate of change │
│ at a point │ │ │ │ │
└──────────┬───────────┘ └──────────┬───────────┘ └────────┬─────────┘
│ │ │
└──────────────────────────┼──────────────────────────┘
│
↓
┌───────────────────────┐
│ │
│ f(x+h) - f(x) │
│ f'(x) = lim ──────── │
│ h→0 h │
│ │
└───────────────────────┘
HOW THE SECANT BECOMES THE TANGENT:
────────────────────────────────────
y
│ ╱ ← secant line through
│ ╱ ╱ (x, f(x)) and
│ ╱ ╱ ╱ (x+h, f(x+h))
│ f(x+h) ●╱ ╱ ╱
│ │╱ ╱
│ f(x) ● ╱ ← as h → 0, secant rotates
│ │╱ toward the TANGENT
│ ╱
└─────────┼────────── x
x x+h
slope of secant = [f(x+h) - f(x)] / h ← "rise over run"
↓ take limit as h → 0
slope of tangent = f'(x) ← the derivative
───────────────────────────────────────────────────────────────────────────
DIFFERENTIABLE ⇒ CONTINUOUS (but NOT the other way around)
|x| is continuous at 0 but NOT differentiable there
(sharp corner — left and right limits of slope disagree)
This diagram unifies the derivative's geometric meaning (tangent slope), its algebraic definition (difference quotient limit), and its physical interpretation (instantaneous rate of change), then shows how the secant line rotates into the tangent as h shrinks to zero.
Diagram 3: Differentiation Rules — Decision Tree
╔═════════════════════════════════════════════════════════════════════════════╗
║ DIFFERENTIATION RULES: WHEN TO USE WHAT ║
╚═════════════════════════════════════════════════════════════════════════════╝
Given a function to differentiate — ask these questions in order:
┌──────────────┐
│ What is the │
│ structure? │
└──────┬───────┘
│
┌─────────────────────┼──────────────────────┐
↓ ↓ ↓
┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ SINGLE │ │ COMBINED │ │ COMPOSITE │
│ TERM │ │ TERMS │ │ (nested) │
└──────┬──────┘ └──────┬──────┘ └──────┬──────┘
│ │ │
↓ ↓ ↓
┌─────────────────┐ ┌──────────────────┐ ┌─────────────────┐
│ POWER RULE │ │ product or │ │ CHAIN RULE │
│ │ │ quotient? │ │ │
│ d/dx [xⁿ] │ └────────┬─────────┘ │ d/dx f(g(x)) │
│ = n·xⁿ⁻¹ │ │ │ = f'(g(x)) │
│ │ ┌──────┴──────┐ │ · g'(x) │
│ d/dx [c] = 0 │ ↓ ↓ │ │
│ d/dx [cf] = cf'│ PRODUCT QUOTIENT │ "outside then │
└─────────────────┘ RULE RULE │ inside" │
└─────────────────┘
d/dx[uv] d/dx[u/v]
= u'v + uv' = (u'v - uv') / v²
───────────────────────────────────────────────────────────────────────────
COMMON DERIVATIVES TO MEMORIZE:
┌────────────────┬──────────────────┬──────────────────────────┐
│ f(x) │ f'(x) │ note │
├────────────────┼──────────────────┼──────────────────────────┤
│ xⁿ │ n·xⁿ⁻¹ │ any real n │
│ eˣ │ eˣ │ only function = own d/dx│
│ ln(x) │ 1/x │ x > 0 │
│ sin(x) │ cos(x) │ radians only! │
│ cos(x) │ −sin(x) │ │
│ aˣ │ aˣ · ln(a) │ chain rule applied │
└────────────────┴──────────────────┴──────────────────────────┘
CHAIN RULE REMINDER — the "layers" mental model:
sin(x³) → outer: sin( ) inner: x³
d/dx = cos(x³) · 3x²
↑ ↑
outer d/dx inner d/dx
A decision tree showing how to identify which differentiation rule applies based on the structural form of a function, with a reference table of essential derivatives and a worked chain-rule example.
Diagram 4: The Fundamental Theorem of Calculus — The Bridge
╔═════════════════════════════════════════════════════════════════════════════╗
║ FUNDAMENTAL THEOREM OF CALCULUS: THE GREAT BRIDGE ║
╚═════════════════════════════════════════════════════════════════════════════╝
DIFFERENTIATION INTEGRATION
(rates of change) (accumulation)
┌────────────────────┐ ┌────────────────────┐
│ │ │ │
│ f'(x) tells you │ │ ∫f(x)dx tells you │
│ how fast f grows │ │ how much f has │
│ at each instant │ │ accumulated │
│ │ │ │
└─────────┬──────────┘ └──────────┬─────────┘
│ │
│ ╔══════════╗ │
└────────→║ FTC ║←────────────┘
║ ║
╚══════════╝
│
┌──────────────────────┴──────────────────────┐
↓ ↓
┌───────────────────────────┐ ┌───────────────────────────┐
│ FTC PART I │ │ FTC PART II │
│ │ │ │
│ If F(x) = ∫ˣₐ f(t) dt │ │ If F is any │
│ │ │ antiderivative of f: │
│ then F'(x) = f(x) │ │ │
│ │ │ ∫ᵇₐ f(x)dx = F(b)−F(a) │
│ differentiating an │ │ │
│ accumulation gives │ │ to find the NET area │
│ back the rate │ │ just evaluate F at │
│ │ │ the endpoints │
└───────────────────────────┘ └───────────────────────────┘
───────────────────────────────────────────────────────────────────────────
A CONCRETE EXAMPLE OF PART II:
∫³₁ x² dx → antiderivative of x² is x³/3
→ [x³/3]³₁ = (27/3) − (1/3) = 9 − 1/3 = 26/3
WHAT MAKES THIS SHOCKING:
──────────────────────────
The definite integral is defined as a limit of Riemann sums —
an infinite addition process. The FTC says you can bypass all
of that infinite addition simply by finding one antiderivative
and subtracting two numbers. This is not obvious. It is profound.
This diagram maps the Fundamental Theorem of Calculus as a two-way bridge between differentiation and integration, showing both parts with their statements and a worked numerical example of Part II.
Diagram 5: Riemann Sums — Building the Definite Integral
╔═════════════════════════════════════════════════════════════════════════════╗
║ FROM RIEMANN SUMS TO THE DEFINITE INTEGRAL ║
╚═════════════════════════════════════════════════════════════════════════════╝
IDEA: approximate the area under f(x) with rectangles, then take limit.
n = 4 rectangles (left endpoints): n = 8 rectangles:
y y
│ ╭──────╮ │ ╭─╮╭─╮
│ ┌─┤ ├──╮ │ ┌┤ ││ ├╮
│ │ │ │ ├──╮ │ ││ ││ ││├╮
│ │ │ │ │ ├──╮ │ ││ ││ ││││╮
│ │ │ │ │ │ │ │ ││ ││ ││││ │
└──┴─┴──────┴──┴──┴──┴── x └─┴─┴─┴─┴┴┴─┴── x
a b a b
more rectangles ─────────────────────────→ exact area in the limit
(n→∞, width→0)
───────────────────────────────────────────────────────────────────────────
THE FORMAL DEFINITION:
Partition [a, b] into n subintervals, each of width Δx = (b−a)/n
Choose sample points xᵢ* in each subinterval
n
Sₙ = Σ f(xᵢ*) · Δx ← Riemann sum
i=1
b
∫ f(x)dx = lim Sₙ ← definite integral (Stewart §5.2)
a n→∞
───────────────────────────────────────────────────────────────────────────
SIGN MATTERS: area ABOVE x-axis counts positive (+)
area BELOW x-axis counts negative (−)
the integral gives NET SIGNED AREA, not total geometric area
This diagram shows how Riemann sums use increasingly many rectangles to approximate area under a curve, with the definite integral emerging as the limit of this process as the number of rectangles grows to infinity.
Key Takeaways
- The limit is independent of the function's value at the point. limx→a f(x) = L is defined entirely by f(x)'s behavior near a, not at a. f(a) may not even exist, and the limit can still be perfectly well-defined. This is the conceptual foundation of all calculus (Stewart §2.2; Spivak, Ch. 5).
- The epsilon-delta definition makes "approaches" rigorous. limx→a f(x) = L means: for every ε > 0 there exists δ > 0 such that 0 < |x − a| < δ implies |f(x) − L| < ε. The order matters: ε is the challenge, δ is the response. You must find δ that works for every ε, no matter how small (Spivak, Ch. 5).
- Continuity has a precise three-part definition. f is continuous at a if and only if: (1) f(a) is defined, (2) limx→a f(x) exists, and (3) those two values are equal. All three conditions can fail independently (Stewart §2.5).
- The derivative is a limit of a difference quotient. f'(x) = limh→0 [f(x+h) − f(x)] / h. This single formula simultaneously encodes slope of the tangent line, instantaneous rate of change, and the local linear approximation to f. If this limit does not exist, f is not differentiable there (Stewart §2.7, §2.8).
- Differentiability implies continuity, but not conversely. If f is differentiable at a, it must be continuous at a. The converse fails: f(x) = |x| is continuous everywhere but not differentiable at x = 0 because the left and right limits of the difference quotient give −1 and +1 respectively — they disagree (Stewart §2.8).
- The Chain Rule differentiates composite functions by multiplying the derivative of the outside by the derivative of the inside. If h(x) = f(g(x)), then h'(x) = f'(g(x)) · g'(x). A common error is forgetting the inside derivative. For example, d/dx[sin(x³)] = cos(x³) · 3x², not merely cos(x³) (Stewart §3.4).
- The Mean Value Theorem guarantees an interior point where the instantaneous rate equals the average rate. If f is continuous on [a, b] and differentiable on (a, b), there exists at least one c in (a, b) such that f'(c) = [f(b) − f(a)] / (b − a). This is the rigorous basis for relating derivatives to behavior on intervals, and it underlies curve-sketching rules about increasing/decreasing functions (Stewart §4.2).
- The definite integral is defined as a limit of Riemann sums, not as an antiderivative. ∫ab f(x) dx = limn→∞ Σ f(xᵢ*) Δx. The integral gives net signed area: regions below the x-axis contribute negatively. That this limit always exists when f is continuous is a non-trivial theorem (Stewart §5.2).
- The Fundamental Theorem of Calculus, Part I, shows differentiation undoes integration. If F(x) = ∫ax f(t) dt and f is continuous, then F'(x) = f(x). The accumulation function's rate of growth at any x equals f's value there. This is why differentiation and integration are inverse operations (Stewart §5.3).
- The Fundamental Theorem, Part II, makes evaluation of definite integrals computable. If F is any antiderivative of f on [a, b], then ∫ab f(x) dx = F(b) − F(a). This replaces an infinite limiting process (Riemann sums) with two function evaluations — a genuinely astonishing shortcut whose justification requires Part I (Stewart §5.3).
- Integration by substitution is the chain rule run in reverse. If u = g(x), then du = g'(x) dx, and ∫ f(g(x)) g'(x) dx = ∫ f(u) du. The critical step is transforming both the integrand and the differential correctly; for definite integrals, the limits of integration must also change to u-values (Stewart §5.5).
- The natural exponential function ex is its own derivative. d/dx[ex] = ex. This is not an accident of notation — it is the defining property that makes e ≈ 2.71828 the only base for which this holds exactly, and it makes exponential functions indispensable throughout physics, chemistry, and biology wherever a quantity's growth rate is proportional to its current size (Stewart §3.1).
- Optimization problems require checking critical points, endpoints, and the Second Derivative Test. A critical point where f'(c) = 0 is a local maximum if f''(c) < 0, a local minimum if f''(c) > 0, and the test is inconclusive if f''(c) = 0. On a closed interval, the global extremum must be found by comparing all critical values with both endpoint values (Stewart §4.1, §4.3).
- The area between two curves is the integral of the top function minus the bottom function. If f(x) ≥ g(x) on [a, b], then the enclosed area = ∫ab [f(x) − g(x)] dx. The intersection points of the two curves determine the limits of integration and must be found first — this is a direct application of the additive and linearity properties of the definite integral (Stewart §6.1).
- Position, velocity, and acceleration are related by differentiation and integration. If s(t) is position, then v(t) = s'(t) and a(t) = v'(t) = s''(t). Conversely, given acceleration, velocity is recovered by integration and position by integrating again. The constant of integration corresponds to the initial condition — a piece of information the mathematics alone cannot supply, but that physics always provides (Stewart §4.9, §5.4).
References
- James Stewart, Daniel Clegg, and Saleem Watson, Calculus: Early Transcendentals, 9th ed. (Boston: Cengage Learning, 2020). Sections cited: §2.2, §2.4, §2.5, §2.7, §2.8, §3.1, §3.4, §4.1, §4.2, §4.3, §4.9, §5.2, §5.3, §5.4, §5.5, §6.1.
- Michael Spivak, Calculus, 4th ed. (Houston: Publish or Perish, 2008). Chapters cited: Ch. 5 (Limits), Ch. 9 (Derivatives), Ch. 13 (Integrals), Ch. 14 (Fundamental Theorem of Calculus).