Intuition The one core idea
A Bernoulli equation is almost an easy ("linear") equation, spoiled by a single term where y is raised to a weird power. The whole method is: rename a chunk of y as a new letter so that annoying power disappears , leaving a friendly equation you already know how to solve.
This page assumes you have seen none of the notation. We build every symbol from the ground up, in the order they need each other, then hand you back to the parent Bernoulli topic .
Before any equation, agree on the actors.
Definition The unknown function
y and the input x
x is a plain number we are free to pick — think of it as "how far along the horizontal axis" we are.
y is not a fixed number. It is a rule that gives a height for each x . We write y ( x ) to mean "the height when the input is x ."
Picture: a curve drawn on paper. Slide your finger to a value of x on the bottom axis; the curve tells you y , the height there.
Why the topic needs this: a differential equation is a statement about a curve we don't know yet . Everything below is a hunt for that curve.
d x d y = slope of the curve
Zoom in on the curve at one point until it looks like a straight line. Its steepness — how much height y you gain per one step of x — is called the derivative, written d x d y or y ′ .
Going uphill → d x d y > 0 .
Flat → d x d y = 0 .
Going downhill → d x d y < 0 .
Why two names? d x d y reminds you what is changing (the ratio of a small change in y to a small change in x ); y ′ is just shorthand when the meaning is clear. They are the same object .
Why the topic needs it: a Bernoulli equation contains d x d y . It tells us the slope at every point in terms of where we are — a recipe the curve must obey.
Definition Ordinary Differential Equation (ODE)
An equation that links a curve's height y to its slope d x d y (and the input x ). "Ordinary" = only one input variable x . "First-order" = the highest slope-taking we do is once (no y ′′ ).
Intuition Read an ODE as a "slope field"
An ODE like d x d y = (something involving x , y ) hands you a little arrow at every point of the plane telling you which way a curve through that point must lean. Solving = threading a curve that always follows the arrows.
Everything special about Bernoulli lives in a power of y . So nail powers.
Why the topic needs y 1 − n specifically: the substitution names v = y 1 − n . Where does 1 − n come from? Watch the arithmetic: to cancel the power y n from the equation we divide by it, turning y into y 1 − n = y 1 / y n (using the division rule above). That single subtraction 1 − n is the whole trick's fingerprint.
1 − n
n = 2 ⇒ 1 − n = − 1 ⇒ v = y − 1 = y 1 .
n = 2 1 ⇒ 1 − n = 2 1 ⇒ v = y .
n = 4 ⇒ 1 − n = − 3 ⇒ v = y − 3 .
P ( x ) and Q ( x ) — coefficients that may wobble with x
These are just numbers that depend on where you are , not on y . In d x d y + P ( x ) y = Q ( x ) y n :
P ( x ) scales the plain y term (e.g. P = x 1 ).
Q ( x ) scales the troublesome y n term (e.g. Q = x ).
Picture: dials whose settings change as you slide x , but which never peek at y .
Why the topic needs this distinction: the reason we can rescue the equation is that P and Q never contain y . All the y -difficulty is packed into that lone y n . Isolate it, kill it, done.
Definition Linear first-order ODE
An ODE is linear in y when y and d x d y appear only to the first power , never multiplied together, never inside roots or squares:
d x d y + P ( x ) y = Q ( x ) .
No y 2 , no y , no y ⋅ y ′ .
Intuition Why "linear" is the promised land
Linear equations have a guaranteed recipe (the integrating factor — see Linear First-Order ODEs — Integrating Factor ) that always cracks them. The y n in Bernoulli is the one thing that breaks linearity. Remove it and you are home.
Common mistake "Linear" does
not mean "straight line"
Why it feels right: the word linear evokes lines. The trap: here it describes how y enters the equation (first power only), not the shape of the solution curve — which can be wildly curvy.
This is the single most important tool, so we build it carefully.
If v is built out of y , and y in turn changes with x , then v 's slope with respect to x is
d x d v = how v reacts to y d y d v × how y reacts to x d x d y .
Read it as a relay: a nudge in x nudges y , which nudges v . Multiply the two "nudge rates."
the choice
Look at that result: it contains exactly y − n d x d y — the very lump the parent topic creates by dividing the equation through by y n . So v ′ swallows the troublesome term whole , and what remains has no bad power left. The exponent 1 − n was reverse-engineered so that d y d v spits out y − n . That is why not some other substitution.
You'll meet this fully in Linear First-Order ODEs — Integrating Factor ; here just decode the symbols so nothing is a black box.
μ ( x )
∫ ⋯ d x = the antiderivative : undo a derivative, i.e. find the function whose slope is inside. Geometrically, accumulated area under a curve.
e ( ) = the exponential function, the curve that grows at a rate equal to its own height . It appears because it is the one function that "undoes" a product-of-slopes into a clean derivative.
So μ ( x ) = e ∫ ( 1 − n ) P d x is a cleverly chosen multiplier that turns the left side of the linear ODE into the derivative of a single product ( μv ) ′ , ready to integrate.
Why the topic needs it: once Bernoulli is flattened into linear form v ′ + ( 1 − n ) P v = ( 1 − n ) Q , this μ is the crank you turn to finish.
C — the family label
Integration always leaves an unknown constant C , because many curves share the same slope field, differing only by a vertical shift or scaling. C is the name tag that picks one curve out of the infinite family. A single extra fact (an initial condition like y ( 1 ) = 2 ) pins C to a number.
derivative dy/dx as slope
ODE links height and slope
divide by y^n to expose y^1-n
P of x and Q of x coefficients
Test yourself — each line hides the answer.
What does d x d y measure, in one word? Slope (steepness) of the curve at a point.
Rewrite y − 2 and y 1/2 without exponents. What is y 0 , and why does that break the n = 1 case? y 0 = 1 ; with n = 1 , v = y 1 − 1 = y 0 = 1 is constant — useless.
Do P ( x ) and Q ( x ) ever depend on y ? No — they depend on x only.
State the chain rule for d x d v when v depends on y and y on x . d x d v = d y d v ⋅ d x d y .
Differentiate v = y 1 − n with respect to x . d x d v = ( 1 − n ) y − n d x d y .
What makes an ODE "linear in y "? y and y ′ appear only to the first power; no y n , no products of y with y ′ .
What does ∫ ⋯ d x do? Finds the antiderivative — the function whose slope is the integrand (area accumulation).
Why does an arbitrary constant C appear after integrating? Many curves share a slope field; C picks which one.