Continuity, said carefully
The previous guide built the complex limit and showed how demanding it is — every road toward z_0 must agree on one value L. Continuity is the calm sequel. A function f is continuous at z_0 when three things all hold: f is defined there, the limit of f(z) as z approaches z_0 exists, and that limit equals the actual value f(z_0). Strip away the jargon and it just says: the value f wants as you sneak up on z_0 is the value it really has there. No hole, no jump, no surprise.
Geometrically this is gentle and intuitive in the two-plane picture you have been carrying. Nudge the input point z a little inside the z-plane, and the output point f(z) moves only a little inside the w-plane. A continuous map never tears the plane apart and never welds two distant pieces together at the seam. The translations, rotations, squaring and inversion maps you sketched earlier are all continuous wherever they are defined — that smoothness is exactly why their images of nice regions came out as nice, connected shapes rather than shredded confetti.
Two surfaces in disguise
Here is the idea that gives this guide its title and most of its mileage. Write the function as f = u + i v, splitting it into the real-valued field u(x, y) and the real-valued field v(x, y). Then f is continuous at z_0 = x_0 + i y_0 exactly when both u and v are continuous at the point (x_0, y_0) in the ordinary two-variable sense. This is the continuity of u and v, and it is an equivalence — not a one-way street. One complex continuity is precisely two real continuities standing back to back.
f(z) = u(x, y) + i v(x, y)
f continuous at z_0 <=> u continuous at (x_0, y_0)
AND
v continuous at (x_0, y_0)
(the link: |f(z) - f(z_0)|^2 = (u - u_0)^2 + (v - v_0)^2)Why does the equivalence hold? Because distance in the w-plane is built from u and v by Pythagoras: |f(z) - f(z_0)|^2 = (u - u_0)^2 + (v - v_0)^2. If the left side is small, neither term on the right can be large, so both u and v are pinned close to their target values; conversely, if both right-hand terms are small, the whole sum is small. Smallness leaks both ways. Nearness of the complex output is the same fact as nearness of each real coordinate.
Borrowing the whole toolbox
The payoff of the u-and-v bridge is that you never have to re-prove continuity from scratch. Every continuity rule from multivariable calculus crosses over intact: a sum, a difference, a product of continuous functions is continuous; a quotient is continuous wherever the denominator is non-zero; and a composition of continuous maps is continuous. Constants and the identity map z are obviously continuous, so by building up, every polynomial in z is continuous on the whole plane.
Watch it work on the maps you already know. The squaring map gives u = x^2 - y^2 and v = 2 x y; both are polynomials in x and y, hence continuous everywhere, so z^2 is continuous on the whole plane. The inversion 1/z is a quotient with denominator z, so it is continuous everywhere except z = 0, where it is not even defined — exactly the one point it cannot reach. Even the conjugation map z-bar, the troublemaker that failed to have a complex derivative, is perfectly continuous: its parts are u = x and v = -y, both as tame as a function can be. Continuity is forgiving in a way that differentiability, as you will soon see, is emphatically not.
When continuity earns you more: compact sets and uniformity
Continuity by itself is a local promise — about what happens near each single point. But upgrade the region and continuity starts paying dividends across a whole set at once. On a compact set in the plane — concretely, one that is closed and bounded, like a filled disk or a solid rectangle including its edges — a continuous function is automatically bounded and actually attains its largest and smallest modulus somewhere inside. This is the boundedness on a compact set, the planar twin of the extreme value theorem you trusted on a closed interval.
On that same compact set, continuity also gets stronger in a second way: it becomes uniform. Ordinary continuity lets the delta you need shrink as you move from point to point; uniform continuity promises a single delta that works everywhere on the set at once, no matter where z_0 sits. You do not need this refinement to talk about maps, but it is the quiet workhorse later: it is what lets you approximate integrals over curves, justify swapping limits, and trust that a function does not wobble arbitrarily fast on a bounded patch.
What continuous maps do to curves
The objects a map acts on most vividly are not lone points but whole roads. Describe a curve by letting a real parameter t run over an interval and writing z(t) = x(t) + i y(t); as t sweeps along, the point z(t) traces a path through the z-plane. Push that path through a continuous f and you get a new path w(t) = f(z(t)) in the w-plane. Continuity is precisely the guarantee that this image road is itself unbroken — a connected curve, never snapped into separated pieces. That is what lets you draw a map at all: feed it a grid of lines and watch the image grid bend smoothly rather than fracture.
- Pick a test shape in the z-plane — say the horizontal line y = 1, parametrized as z(t) = t + i for t real.
- Apply the map and simplify to w = u + i v: for f(z) = z^2 this gives w = (t^2 - 1) + i (2 t).
- Read the image as a curve: here u = t^2 - 1 and v = 2 t, and eliminating t gives u = (v/2)^2 - 1 — a sideways parabola.
- Because f is continuous, that parabola is a single unbroken arc — continuity is exactly what licenses you to trace it without lifting your pen.
Keep continuity in proportion, though. It is a genuinely mild condition — the SAME strength you knew from real calculus, merely asked of a two-dimensional input. It does not yet force anything magical. The famous rigidity of complex analysis — where one derivative compels infinite differentiability and a power series — lives one rung up, in complex differentiability, where we demand a limit not of f but of the difference quotient, and insist it agree along every direction. Continuity sets the stage; the next rung is where the drama actually begins.