Readable books on advanced topics I realise that there are already a few questions looking for general book recommendations, but the motivation and type of book I'm looking for here is a little different, so I hope you can indulge me.
I enjoy reading quite a lot, and tend to prefer books that teach me something new, rather than straight up fiction. When I was younger I used to read a lot of popular science books, and at the time I thought these were teaching me something, but then I got older and went to university and studied theoretical physics, and that killed the genre for me.
Now, I know that I can learn about a topic simply by seeking out review papers on it, or finding an appropriate textbook. That's what I do when I need to do something for work. However, most of the time this material does not make for light reading, and requires a significant amount of effort to work through, and is not necessarily the kind of thing I'd want to use as a way to relax.
However, occasionally one comes across a very readable book on some aspect of physics. Nielsen and Chuang's Quantum Computation and Quantum Information strikes me as an example of this. There are probably better examples of this, but what are they?
This brings me to my question: 

Beyond the standard undergraduate topics, which areas of physics or mathematics have
  books which both provide fairly comprehensive introductions and which
  are actually enjoyable to read?

To be clear, by "enjoyable" I mean something more than that they simply be accessible or well written. I mean that it should be something that I could read for relaxation rather than work. To some people this will probably sound like an insane thing to want to do. However modern physics is huge, and it bothers me that I don't know very much about, say, string theory beyond bosonic string theory, and if it is something I can learn more about in my spare time, then that would be great.

UPDATE: There appears to be some confusion in the answers over the type of book I am looking for. I am not looking for popularisations. Rather I am looking for something at the level of a graduate text, but only those which are particularly readable.
 A: I really like The Princeton Guide to Mathematics 
(Amazon, Google)
Even though I believe it was meant more as a sort of Encyclopedia rather than an actual Textbook, I really like to read it as a "normal" book.
Basically, whenever I understand some new mathematical concept I look it up in the Guide and see how it branches out and often find new interesting topics I'd like to read about. 
Furthermore it gives an excellent overview of all the mathematical topics there are and is very well written and easy to understand.
A: A perfect example for what I think you are looking for is J.S. Bell's Speakable and unspeakable in quantum mechanics (Amazon, Google).  Although it is a collection of papers, this is a very readable account of some aspects of quantum foundations.
A: I find the book Geometry of quantum states: An Introduction to Quantum Entanglement by Ingemar Bengtsson and Karol Życzkowski both readable and useful. I already referred to it in three different occasions on this site. It focuses on the geometrical description of the spaces of quantum states and maps of finite level quantum systems.  This subject is rapidly developing and even the book was only recently published (2006), it may have become a little outdated. Nevertheless, I think that it is very useful as a research level introductory reference.
A: A brief list of miscellaneous "technical" (not popular science) books I read on the bus, and I enjoyed every page:
Everything from R.P. Feynman, from his "Lectures on physics" to "Statistical mechanics", the "Lectures on computation" or "Gravitation". "QED: the strange theory of light and matter" may be considered as (extremely good) popular science, but it contains more insight than most technical reviews.
General Relativity: Misner, Thorne and Wheeler, "Gravitation", but also Foster and Nightingale, "A short course on general relativity". Also Taylor and Wheeler, "Spacetime physics" on special relativity. J. Weeks, "The shape of space" on topology is a true jewel. 
About fractals and surface growth, A.L. Barabasi, "Scaling concepts in surface growth" is extremely readable. M. Schroeder's "Fractals, chaos and power laws" is both easy to read, lovely and accurate.
In abstract algebra, I. Stewart, "Galois theory" is very nice, but also the book from H.E. Edwards. They are both like novels.
Marsden and Weinstein, "Calculus unlimited". From M. Spivak, "Calculus" and "Calculus on manifolds". His books in differential geometry are also very nice. 
Jaynes, "Lectures on probability theory".
Ufff, this was just out of the box. The books I cite are a bit old, I know, but they are still highly recommendable.
A: I found Connes's Noncommutative Geometry a very good book.
A: Perhaps, the best mathematics book in this category is "What is Mathamatics?" by Richard Courant.  Courant wrote this book in the 40's to help his son.  It is still good.
A: Well, The road to reality by roger penrose is very good , it's not a popularization and it deals with advanced topics in Supersymmetry and covers even twistor theory , however you do not need to work through it , you can read it in your freetime . I highly recommend this book
A: The books which are good are without exception written by the people who did the research themselves and are describing work they either created or reproduced independently. The books which are bad are without exception written by those who are writing about someone else's research without having reproduced it internally from scratch. This means that your best bet is to only read those books by those people who are describing their own research, not that of others. That's a complete guide to the literature.
When dealing with historical material, this still applies, but it's difficult to read the Principia. But this works for anything past 1850 or so, and it's not a lot of literature until the goverment starts funding in a major way in the 1950s.
A: Personally, I found In Search of Schrödinger's Cat by John Gribbin very entertaining. It is quite old now and I'm not sure whether you'll find anything new in there after studying theoretical physics, though.
A: The New Physics by Paul Davies.
