Well, maybe I’d like to ask a strange question.
Actually, “MyGraph->GetXaxis()->SetLimits(xmin, xmax);” is just a shortcut for “MyGraph->GetHistogram()->GetXaxis()->SetLimits(xmin, xmax);”.
However, there is a non-negligible difference between “MyGraph->SetM[in|ax]imum(ym[in|ax]);” and “MyGraph->GetHistogram()->SetM[in|ax]imum(ym[in|ax]);”.
And that difference exists not only for TGraph related classes but also for TF1, for example (well, maybe for more, but I didn’t trace it that much).
So, the question is … when should one use which method and why?