Newton’s cosh iteration
Newton's rule is N(z) = z − f(z) / f′(z). Here f(z) = cosh z − 1 and f′(z) = sinh z. Each step finds the correction (cosh z − 1) / sinh z, then takes it away from z.
The equation cosh z = 1 has infinitely many solutions: z = 2πik for every integer k. Each is a simple root: cosh z grows linearly close by, so Newton's method converges quadratically there. Far from the real axis, cosh and sinh grow exponentially. Their quotient may be tiny or enormous, and that is where the image's sensitive seams take shape.
Try z₀ = 0.2i. Since i·sinh(0.2) ≈ 0.201i and cosh(0.2) ≈ 1.020, the first correction is small; the orbit quickly reaches the root at z = 0. Now move to z₀ = (0.5 + 3.0i), farther from every root. The correction now contains large exponential terms. The orbit can cross several basin boundaries before settling, or wander chaotically between roots for many iterations before one captures it.
The function repeats with period 2πi in the imaginary direction, and the Newton map does too: N(z + 2πi) = N(z) + 2πi. That repeated rule is the source of the horizontal banding.

