A note on geometrically reductive groups