Here’s kind of an unusual question: What is the density of integers that have an even number of prime factors with an exponent greater than 1?

To define the density, you take the proportion up to an integer *N* then take the limit as *N* goes to infinity.

It’s not obvious that the limit should exist. But if it does exist, you might guess that it’s 1/2; it’s a question involving whether something is even or odd, and so even and odd occurrences might balance each other out.

But the result, known as the Feller-Tonier constant, is bigger than 1/2. This seems more reasonable when you consider that zero is an even number and a lot of numbers may not have any prime factors with exponent bigger than 1.

Here’s a little Python code to compute the ratio for *N* = 10^{6}.

from sympy.ntheory import factorint def f(n): v = factorint(n).values() n = len([x for x in v if x > 1]) return n % 2 == 0 c = 0 N = 1000000 for n in range(N): if f(n): c += 1 print (c/N)

This gives a ratio of 0.661344. The limiting value is 0.661370494….