A、During the first half of their careers. B、During the first third of their careers. C、During the second half of their careers.

(20)Students seeking success should look for mentors among rising young stars instead of established big names in their field, according to the first large-scale study of mentorship. Such findings might also extend beyond academia to business, the military and the arts, if future studies of mentorship come up with similar results. An exhaustive family tree of mathematicians dating back to Isaac Newton allowed researchers to study a sample of 7 259 mathematicians who graduated between 1900 and 1960.(21)The networking data showed that successful academics did a much better job in mentoring students during the first third of their careers , rather than the last third of their careers. "What we observe is that when proteges(门徒)select mentors who are relatively young prospective superstars, they in turn become hotshots themselves," said Dean Malmgren, a chemical and biological engineer at Northwestern University in Evanston. By contrast, students who hoped to ride on coattails of a big shot mentor late in that person’s career fared less well. Students trained by mathematicians who were in the first third of his or her career went on to train 29% more students than expected, while students trained by mathematicians in the last third of his or her career went on to train 31% fewer students than expected. The pattern proved itself as a consistently strong signal across the decades and not just a flash in the pan, Malmgren said.(22)He suspected that the amount of face time a mentor can spend with students has much to do with the different mentorship results, even if the study did not assess the exact causes behind the patterns.
20. What kind of mentors should students choose to achieve success?
21. During which period of career did mathematicians do better in mentoring students?
22. According to Malmgren, what might be related to the different mentorship results?