Writing reviews of scientific books is, in this sense, just another form of peer review, yet it is often more than that. A critique of a project proposal or a paper is usually technical and schematic; it never goes outside a small audience of grant officers or colleagues; its function is administrative. A book review, on the other hand, is a literary effort in its own right, and its readership extends beyond a small circle of specialists; its function is to inform a larger public.
Reviewing another's textbook or monograph in a professional journal offers the opportunity to give a personal overview of one's own ideas on the subject. Writing in a popular magazine gives one a forum to educate, provoke, even amuse. In addition, writing reviews is a painless and relatively direct way to break into print.
Yet how does one go about becoming a book reviewer? In a word: volunteer. There are thousands of science-related books published each year, and hundreds of journals and magazines that carry signed reviews. These include not just specialty journals, but also magazines aimed at educators, at scientists and educated laymen, and even the book review sections of some newspapers.
Review editors, who are in many cases not scientists themselves, refer books to researchers suggested to them by other contributors to the magazine, sometimes by a board of editors. Occasionally an editor will send a book to an established public figure, a major authority. But there are only a limited number of hours in a research scientist's day, and only a limited stock of Carl Sagans and B.F. Skinners. As a practical matter, editors, who usually have more books than reviewers, look not for scientists with high name recognition, but for thoughtful readers who can write well.
If that describes you (and you may not know it until you try), it is up to you to let book review editors know you exist. Sometimes this can be done through colleagues who are themselves contributors to the publication. But you can also do it yourself. Send a letter of inquiry, in your best literary style, stating your interests and areas of expertise. More often than not, you'll be given an opportunity to try your skills on whatever appropriate volume the review editor has accumulated on his groaning shelf. Often an editor will send along a copy of the table of contents or the jacket notes so that you can get some idea of whether the book interests you.
The first book offered you may not be exactly the one you have in mind, but unfamiliar ideas and novel approaches can provide seminal material for an essay, even if the book itself seems weak. Some magazines publish regular lists of books received. In these cases, it doesn't hurt to express interest in a particular volume or two (particularly if you've wanted to read it but can't afford to buy it yourself.), but don't count on getting your choice of assignments until you've established yourself as a reliable contributor. Avoid submitting a finished manuscript until you are asked to do so: most publications in my experience do not welcome unsolicited book reviews.
How you actually organize the review depends on where it's to be published. Review editors invariably send along a sheet of guidelines that will give you some idea of what they are looking for. Some journals expect a summary of the contents of the book, with only a few sentences on its high and low points. Some want a review to convey the flavor of the book without explicitly presenting its contents. Most want some sort of sober evaluation of what the author was trying to do, and how well he or she accomplished it.
The length of a review, to some extent, dictates the approach one can take. A short review, 300 words or so, can only touch on the high points of the book. If you are clever, you can evoke the quality of the work with a well-chosen quote or a few pithy sentences. In a longer review, you can devote more space to the author's text itself, and you can discourse on what the work teaches, how it compares with other books, where it fits into the current state of the art. In some cases you may be able to review several related works in a single comparative essay.
Naturally, the audience for the review is important, too. If you are writing a review of Cosmological Dimensions of Axion String Dynamics for the Meta-Astrophysical Journal, your comments can be highly technical and your criticisms as recondite as you choose. On the other hand, if you are reviewing The Secret Life of Charles Darwin for Popular Paleontology magazine, you will want to spend some time considering how to make its message understandable to a reader who may have had little prior knowledge of the field.
Writing a judicious review takes some restraint. If it's a good book, you have to resist becoming effusive. Few books merit masterpiece status; few will alter the course of Western intellectual history. Simply set out what you like about the book with plenty of enthusiasm, but use superlatives only when appropriate.
The best way to learn the trade, of course, is to practice it, submitting one's efforts to the red pencils of editors and learning from their comments. Some journals edit their submissions more heavily than others, but don't expect to have your text printed verbatim. Review editors are as sensitive to how their readers will respond as they are to the scientific reliability of the author. In popular science magazines in particular, style is important; working on book reviews in these publications is a good way to learn to write for a general audience.
Alas, a list of published reviews may never amount to much in the eyes of an employer, a dean or an academic personnel committee. Nor can it line one's pockets with gold—some popular magazines pay for reviews, but at the typical rate of $100 per review, you would have to write two or three each week to reach the U.S. minimum wage level. Writing reviews is an occupation for those who like to read and to write, and whose mark of success is a wall full of books, not a Bigelow on the floor.