A problem with graduate textbooks, especially in the sciences, is their cost. It is quite common for a graduate student to shell out several hundred dollars to purchase required textbooks for graduate courses. In fact, because graduate courses are typically taken early on, when a student is fresh in graduate school and may not have yet acquired the necessary cash management skills, this expense can deal a serious blow to a student’s finances.
In some cases, one could conceivably accept the argument that the usefulness of graduate textbooks extends beyond the duration of the course, as these are books which a professional scientist often consults. Therefore, one might look at this as an investment. And even if their career takes them out of research or science, seldom will physics Master’s or PhD graduates get rid of their graduate textbooks.
In principle a student could also do without purchasing a graduate textbook, as the university library with typically carry one copy for consultation. Clearly, however, it is vastly more convenient to have one’s own copy, and borrowing one from a fellow graduate student may not always be a viable option (because they are so expensive, these books are seldom loaned).
One could ask the question: is a textbook really necessary for a graduate course ? In my experience, even though students at that level should acquire the habit of consulting different books, providing at least one “official” reference makes things easier, both for the instructor as well as for the students. A textbook not only provides an overview of the subject and a sense for the logical connection between the different topics, but also facilitates more mundane matters such as assigning homework.
Sometimes, however, things get really complicated. This term I have been assigned Quantum Mechanics, and for a number of reasons I have decided to base the course on Feynman’s Path Integral approach . I mostly teach the course off my notes, which are, of course, constantly evolving and summarize my own understanding of this subject. I would like to be able to provide a textbook too, but the obvious reference is out of print. It is possible to find copies, but price can be quite formidable (over one thousand dollars).
This has been the situation for this book for quite some time. Surprisingly, publishers such as Dover, famous for producing affordable paperback copies of classic physics textbooks, have shied away from reprinting this particular book.
What is an instructor supposed to do in these cases ? In principle I could make photocopies of my own personal copy of the book, and distribute them to all students who have registered for the course. In my case, things would be even easier, because I happen to have a PDF version of the entire book, made a few years back by an uncommonly industrious and talented graduate student.
Of course, I am hesitant to do that because there are copyright restrictions, and I am not sure whether distributing such a copy of an out-of-print book is legal, or the circumstances under which it would be considered acceptable. I do not want to get into a discussion of copyright restrictions, a subject I know nothing about, but I am wondering whether in the name of copyright a publisher can prevent someone from acquiring a book. Should this book be regarded as non-existent until someone officially reprints it ?
 My feelings about graduate courses are expounded here. I see no point in rehashing undergraduate material, and the path integral formulation is not commonly included in undergraduate Quantum Mechanics syllabi.