On the Hawking effect