A note on saddle point behavior for ordinary and functional differential equations