what the values of m.p+n.q, where m and n are natural integer numbers, could be?
In fact, m.p+n.q = (m.p' +n.q')*gcd(p,q), where gcd(p,q) indicates the greater common dividor of p and q, and p' and q' the respective ratios of p and q to gcd(p,q), such that gcd(p',q') =1. Therefore, could (m.p'+n.q') have any natural value? No. Indeed, its smaller non null values, if supposed that q' < p' for example, are q', 2q', 3q'..., [p'/q']q', p', ([p'/q']+1)q', p'+q', ([p'/q']+2)q', p'+2q', ([p'/q']+3)q', p'+3q',..., 2[p'/q']q', p'+[p'/q']q', 2p' and (2[p'/q']+1)q', p'+([p'/q']+1)q', 2p'+q' and (2[p'/q']+2)q'..., where [x] symbolizes the integer part of x. Morover, if m.p'+n.q' = y.p'+z.q', (m,n) and (y,z) being natural integers pairs, then (m-y)p' = (-n+z)q', and because gcd(p',q') =1, it exists an integer a such that z-n =a.p', and m-y =a.q', so z=n+a.p' and y = m-a.q'. Then, with z and y non negative, -[n/p'] =< a =< [m/q']. Thus for each (m.p'+n.q') there are ([m/q']+[n/p']) others pairs (m,n) which give the same value.