Equivalent Lagrangians

Given a Lagrangian L for a specific system we can obtain a new lagrangian L that leads to the same Euler-Lagrange equations that L if

This is an exercise from Hand and Finch Analytical Mechanics, page 28. The third item is proven by the definition of action directly. In problem 7 of the same page it is shown that it has to do with the equivalence of inertial frames.

This is related with @olver86 page 249: two Lagrangians L and L~ have the same Euler-Lagrange expressions if and only if the differ by a divergence

L=L~+Div(P).

Here Div is the total divergence.

Particular case

If two Lagrangians, L=L(x,u,u1) and L~=L~(u,u1), give rise to the same Euler-Lagrange equations, then they must differ by a total derivative with respect to x. Specifically, there exists a function f(x,u) such that:

L(x,u,u1)=L~(u,u1)+ddxf(x,u).

This is a consequence of the fact that the Euler-Lagrange equations remain unchanged when the Lagrangian is modified by a total derivative term. Therefore, the two Lagrangians are related by what is known as a "gauge transformation" in this context.