Since 1994, Harry Collins and I have been in disagreement about the history of the early attempts to detect gravitational radiation (gravity waves) both in several publications and in public debate. It may, therefore, be somewhat of a surprise that I strongly recommend that his recent book, Gravity's Shadow, should be required reading for anyone seriously interested in science studies. Collins, who has had considerable access to both the participant scientists as well as to science administrators and funding agents, has used that access to give us a richly detailed and highly readable account of the search for gravity waves from its beginnings in the 1960s up to the present.

The book deals with three main episodes: (1) the early searches for gravity waves and the conflict between Joseph Weber and his critics; (2) the decision to fund attempts using laser interferometers rather than bar or sphere detectors; and (3) the subsequent history of LIGO (Laser Interferometer Gravitational-Wave Observatory), an example of a big science, large research project.

Although there is something for virtually everyone interested in the study of modern science in this volume, I suspect that most readers of Philosophy of Science will agree that the most interesting section is the section devoted to the Weber controversy; I will concentrate on it. In this book, Collins has provided us with a more subtle, detailed, and nuanced account of the episode than he has previously. The story is as follows. In the late 1960s and early 1970s, Joseph Weber designed a new type of gravitational wave detector, a massive aluminum bar, now called a Weber bar, and claimed to have observed gravity wave signals far in excess of what was then theoretically expected. His results were sufficiently credible such that other experimenters attempted to confirm his results. None of these attempted replications was successful, and the conclusion was reached that Weber's results were incorrect. Here Collins invokes the ‘experimenters’ regress’, in which a correct experimental result is one obtained with a good apparatus, whereas a good apparatus is one that produces a correct result. Collins suggests that standard scientific arguments are insufficient to break the regress and that one must look to the social aspects of science for an explanation. He also cites the ‘Duhem-Quine problem’, which is that one can maintain one's belief in a statement, come what may, if one is willing to make changes elsewhere in one's knowledge base.

Collins shows that others found several important errors in Weber's procedures, including a programming error that produced coincidences between Weber's gravity wave antenna and a distant one (a result that initially argued for the correctness of Weber's conclusion) and a reported set of coincidences between his antenna and one used by the Rochester-Bell Laboratory group, in which the antenna signals were, in fact, four hours apart.

As Collins shows, Weber had an answer for both these and all other objections, and one might suggest that Weber became a personification of the Duhem-Quine problem. Collins remarks that Weber's responses to the difficulties cited above were ineffective; others might say insufficient. Perhaps the most important difference between Weber and his critics concerned the choice of analysis algorithm used to analyze the signals emerging from the antenna. Weber's preferred nonlinear algorithm was sensitive only to the amplitude of the signal, whereas the linear algorithm preferred by his critics was sensitive to both the amplitude and the phase of the signal. Weber, as usual, had an answer. He argued that his algorithm was more sensitive to the actual shape of the gravity wave signal, which was longer than his critics expected, and that his algorithm generated several output signals for each input signal, the phenomenon of proliferation, which made signal detection more efficient. As Collins admits, however, “I cannot ‘get around’ a remark made by James Levine in a letter he wrote to Weber in 1974, pointing out that the proliferation effect should have exactly the same result on noise and signal, affecting the statistics in no way” (198). One point that Collins omits is that Weber and his critics shared both data tapes and analysis programs. If the critics’ failure to see a positive signal was due solely to the choice of algorithm, then a positive signal should have been seen when they analyzed their result with his program. It didn't. That led the critics to suggest that Weber was varying his signal threshold to obtain a positive result. Collins cites an unnamed source who supports the critics’ view:

Weber denied the charge. Collins offers an interesting explanation of Weber's extraordinary efforts to obtain a positive signal. During World War II, Weber had been a radar operator on a destroyer. Under those circumstances, a false positive submarine signal was of little harm, whereas the failure to detect a weak positive signal would be disastrous. Although Collins still believes that the discord was not resolved by epistemological and methodological arguments and that social factors were intimately involved, he presents sufficient detail to allow the reader reach their own conclusion.

Collins shows that even after the scientific community decided that Weber's results were wrong, Weber still received funding for his research from the National Science Foundation and published his results in leading journals, at least for a time. Collins also makes clear that Weber became increasingly marginalized within the scientific community. He points out that Weber's later work was published only in unrefereed conference proceedings.

In a short review, I can only briefly mention Collins’ discussion of the decision to fund LIGO rather than other gravity wave detectors. He shows clearly that it was not based on some kind of scientific pork barrel or military and industrial reasons. The fact that there was a larger community of experts committed to interferometers rather than other detectors certainly had an important role in the decision. The history of the LIGO organization after the decision was made is fascinating. Collins describes the problems and difficulties as the administration changed from a ‘small-science’ troika of scientists to the present, more industrial type organization needed for such a large project.

At the conclusion of the book, Collins worries that his close relationship with the participants might have biased his account and remarks that his emphasis (some might say overemphasis) on the work of outsiders such as Weber was an attempt to maintain his distance. I believe he has succeeded admirably. Although I disagree with some of his conclusions, he has given us so much detail that readers can decide for themselves.

This book will be of interest to scientists, and to philosophers, historians, and sociologists of science. The science is clearly explained, and the book is engagingly written. It would make a fine Christmas present for a friend who doesn't have a background in physics. It would, however, have to be placed under a large Christmas tree.