This is very terse (and I don't have access to a list of theorems already proved), so we can't be sure of all the reasoning; but perhaps he concludes that t < x because x is being divided by 1+x, which is greater than 1. It is *not *derived from t < 1.