Originally posted by David113i think i've seen this before and it was a relatively cute problem... don't remember the path towards solution though. maybe using the definition of the "choose" numbers, and seeing that some factor not cancelled in n!/k!(n-k)! must also necessarily be present in n!/(k+i)!(n-[k+i])! for any number i<(n-k)? something to that effect?
If two numbers greater than 1 appear in the same row in Pascal's triangle, then they cannot be coprime.