I am not in a position to-night to imitate my predecessors, and give at present the results of any research into archives or correspondence. My contribution is purely theoretical in its character, though I trust it may have a practical bearing. I simply wish to call attention to some of the more obvious dangers now besetting historical writing, and to attempt to estimate how far certain new methods of work may avert those dangers. My paper is merely a recommendation of ‘co-operation in the production of historical works.’