GNU Octave: fplot exponent problem

I use GNU Octave 4.4.1 (x86_64-pc-linux-gnu on Arch Linux), and I have problem with fplot function. I would like to know why fplot(@(x) 1/(1+25*x.^2),[-1,1]) Have a different result of fplot(@(x) (1+25*x.^2).^-1,[-1,1]) (the latter returns the correct graphic) Had the **exact same problem** with my students in a windows version of Octave 4.4.1. I would like to know if it is an actual error or there's a fancy explanation for why the latter command is correct and the first one is wrong (so I can try to convince my students that they should use GNU Octave. OK, I'll try it anyway).