Where Fibonacci hides in Pascal's triangle

2026-05-08

Pascal's triangle is built on a single rule. Each row begins and ends with 1; every other entry is the sum of the two directly above it. Run that rule eleven times and you get a tidy little pyramid of integers.

Read horizontally and you find familiar things. The 1s along each edge. The natural numbers in the next diagonal. The triangular numbers in the diagonal after that. None of these arrive by surprise; they were baked in by the rule.

Now tilt your eye. Cut a band that goes shallow - more sideways than vertical, two rows up for every three cells to the right. Sum the entries the band touches.

Press play. Each band lights up, the cells inside it carry their values into a sum at the bottom, and the running sequence grows: 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89. Those are the Fibonacci numbers.

Pascal's triangle was not constructed with Fibonacci in mind. The only ingredients are 1s and integer addition. And yet, sliced shallow, every Fibonacci number sits there waiting.

Why it has to be Fibonacci

Number rows and columns from zero. The cells the band touches are exactly the entries where row plus column equals some constant d. Call the sum across that band S_d. We want to show S_d = S_d-1 + S_d-2, with S₀ = S₁ = 1. That alone forces the sequence to be Fibonacci.

The triangle's rule, written algebraically, is Pascal's identity:

C(n, k)  =  C(n-1, k)  +  C(n-1, k-1)

Apply that to every cell on diagonal d. The sum across the band splits into two:

S_d  =  Σ C(d-k, k)
 =  Σ C(d-1-k, k)  +  Σ C(d-1-k, k-1)

The first sum collects the entries on the band one step shallower - row plus column equals d-1. That is S_d-1. The second sum, after a shift of index, collects the band two steps shallower - row plus column equals d-2. That is S_d-2.

S_d  =  S_d-1  +  S_d-2

The single rule that builds the triangle - "each cell is the sum of two earlier cells" - is the same rule, projected into one dimension, that builds Fibonacci. Slicing shallow re-runs the rule along a different axis. The structure comes through unchanged.

Same rule, different dress

The Fibonacci numbers keep showing up. They drive the ratios that chase φ, they steal a cell out of the chessboard, and they nest inside every pentagon. In each of those, something in the construction is plainly asking for them.

Pascal's triangle isn't. It is the pure combinatorial object - the count of paths, the binomial coefficients, the row-sums that are powers of two. Nobody summons Fibonacci to its construction. He turns up anyway, the moment you cut through the triangle sideways.

One rule. A shallow tilt. Fibonacci.