we used above: paragraph = first_div.xpath('child::div[1]/child::p[2]')[0] position()1 and position()2nd, 3rd and 4th items. Location paths as predicates Location paths can also serve as predicates within //div[2][@class="second"] looks at div elements that are the 2nd child of their parent. div means child::div, and [2] will select the 2nd node in the current node- set. In our document this happens only