Existence of periodic solutions for \(\ddot x+\omega^2\dot x- \mu F(x,\dot x,\ddot x)=0\) (Q1418399)

The authors consider the nonlinear third-order equation \[ \frac {d^3x}{dt^3}+ \omega^2{dx\over dt}- \mu F\Biggl(x,{dx\over dt}, {d^2x\over dt^2}\Biggr)= 0,\tag{1} \] searching for periodic solutions. However, the conditions imposed on this equation are extremely restrictive: It is assumed that the nonlinearity \(F\) is equal to the derivative with respect to \(t\) of a function \(f(x, {dx\over dt})\), which in fact reduces equation (1) to the second-order equation \[ {d^2x\over dt^2}+ \omega^2 x-\mu f\Biggl(x,{dx\over dt}\Biggr)= \text{const}.\tag{2} \] Thus, strictly speaking, the paper does not really treat genuine third-order equations. So, assuming Hopf bifurcation for equation (2) leads to the existence of infinitely many periodic solutions of equation (1). Similarly, the existence of infinitely many periodic solutions resp. homoclinic orbits of (1) is derived in the case \(f(x, -y)= f(x, y)\). Finally, these results are applied to equation (2) with a quadratic function \(f\).