The sample of reviews on this site is small, but it has an unusual distribution - five out of six reviews (including this one) give 4 stars. Thus it qualifies as a book which almost all readers found excellent without giving it their ultimate approval. And this is exactly the right conclusion. It is very well done, and loaded to the gunwales with interesting material - both historical and mathematical. But there is a missing ingredient.
Derbyshire's brilliant earlier book on Riemann (actually the only popular book on the subject that deserves to stay in the literature) was focussed by the clarity, and the difficulty, of its goal: to get the reader to understand the Riemann Hypothesis and the state of recent research. In slightly disappointing contrast, Unknown Quantity flows nicely along on a current of Whiggish historical progress, chugging downstream from Ahmes and Diophantus to the broad modern concept of algebra. Inevitably, this is less exciting.
The feature that should nevertheless make this book a continuing success is that Derbyshire provides unusually direct historical explorations of the way the mathematics was created. So, for instance, his exposition of Abel's proof of the unsolvability of the quintic is not a discussion of some later simplified proof of the same result. Moving on to Galois, we get a real sense of how the same result appears in a more general setting. All of this done using essentially elementary tools (with extensions in the useful "primers" that are dotted about the book).
The UK paperback edition has a curiously meaningless cover illustration (a sort of sprocket replacing the iconic "x" that adorns the US version). And there is a tantalising list of illustration credits; tantalising because only one of the illustrations has made it into this edition.