Almost all academic books have to have a bibliography, a list of all the books referenced in the text or which may have provided information not directly quoted. The humanities advance (or keep churning on anyway) largely by discussion of other books, and if you’re going to discuss books it’s rather important that your readers should be able to identify which books you’re talking about. And this doesn’t just mean the title and author. They need to know which edition you consulted because there could be subtle differences between the UK and the US editions, between this and that translation, or between the first and the second edition. You also need to display to readers the fact that you have indeed consulted a suitably wide range of sources, and have not willfully ignored any specific line of research.

The bibliography is not a place where your own good ideas should be tested. There are conventions which should be adhered to. Your idea might even be better, but this is pretty much how bibliographies are laid out, and testing a novel arrangement is only going to distract the reader. The whole purpose of the bibliography is to enable readers to check references, and the most efficient way to achieve this is to lay things out exactly as the reader is used to seeing things arranged.

The bibliography may be given the title References, in which case it will in theory only contain books footnoted in the text.

Bibliographies may be arranged by chapter and printed at the end of each chapter. This may lead to duplication and to my mind is less useful than a single bibliography. It’s just harder to find than a single source at the end of the book. I suspect that this chapter-wise arrangement will tend to happen the closer one comes to the scientific and pedagogical end of things, though it will always be a temptation in a collaborative multi-author volume. The convenience of the reader should be the deciding factor in how such matters are to be handled, rather than the ease of producing offprints.

There are two basic ways of dealing with the issue of referencing books in an academic text. Well three actually, although the third, spelling out full title, dates etc. in the text every time you allude to the book is so ludicrously cumbersome that nobody’d ever contemplate doing it except perhaps in the shortest of pamphlets where each book reference only occurred on a single occasion. (There is of course also a fourth method, which is just quietly to pass the assertion off as if it were your own, but this is not recommended.) The two real methods are the Harvard system in which you give author and year of publication only in the text (Alcock, 2002), putting everything else in the bibliography at the back of the book, ideally just before the index for ease of location. The other, the short-title system, will include a full bibliographical reference only at the first mention of a book, with only the author’s surname and a short title for subsequent references. It can be fairly annoying if books referred to in the short-title system are not also included in a bibliography at the back of the book, since when you come upon “Bertrand, Inscriptions p. 82″ you may fail to remember the full title, and spend hours flipping back through the book you are reading trying to find the full reference, date of publication, publisher, and so on.

The aim in academic communication should be to make it as easy as possible to check and build on assertions and conclusions. Omitting a bibliography doesn’t conduce to that end. In a brief, polemical work, it may be fine to omit a bibliography or to include only a select reading list. In more textbook-like writing a select annotated reading list will probably be more useful for students. But where academic debate is the point the lack of a bibliography is rather like an orator’s speech impediment.

This piece from Teleread is about annotated bibliographies, a staunchly scholarly category of book. I do have one annotated bibliography directed at the general reader, which I’ve always sworn by, The Modern Movement: 100 Key Books from England, France and America, 1880-1950 edited by Cyril Connolly. This book was jointly published in 1965 by André Deutsch Ltd. and Hamish Hamilton Ltd. neither of whom I guess wants to give up the honor of publishing Mr Connolly, the panjandrum of the day. That it was printed in Edinburgh by T. & A. Constable Ltd. further endears it to me. Connolly tells you just enough about each book to show you why it is you should get down to reading the book in question.

See also Professor of books