Backtracking – How do explicit constraints depend on problem instance in backtracking?

I was reading the Chapter on Backtracking in Fundamentals of Computer Algorithms by Horowitz and Sahani, and I read the following line.

The explicit constraints depend on the particular instance I of the problem being solved.

I don’t understand why this statement is true? I have an understanding that the set S (to which all the elements of the n-tuple belong) would be the same for all instances of the problem, but this statement suggests that it depends on the problem instance.