The Duke Physics prelim policy can be found here: http://www.phy.duke.edu/graduate/preliminary.ptml
The purpose of this site is to give you some informal advice on how to prepare for the preliminary exam.

1. As you prepare your abstract, prelim report, and slides for your prelim presentation, be sure that you can explain every sentence, equation, figure, or concept in detail.  For example, if you show an equation, be ready to explain why it scales the way it does with relevant parameters.  Go back to the original derivation so that you understand what assumptions went into the derivation, under what circumstances the equation is valid, and what the equation physically means.  If you graph a quantity, be ready to explain why it has the shape it has.  If you mention an alternative experimental technique, be ready to describe the pros and cons of that technique as compared to the one you did use.

2. Practice your presentation in front of an audience several times.  Have your audience ask many questions and obtain feedback from them both on your presentation and on your response to questions. Do this not only so that you develop a talk that is as clear as possible, but so that you have many opportunities to field questions from both experts and non-experts.  Your own peers are a good source of audience members.

3. Since you and your advisor chose the members of your committee from those you thought would give interesting insight into your eventual thesis project, consider the individual interests of your committee members. What is the overlap of that committee member with your field of presentation?  What might your committee member want to hear more about?  Often this is something related to his/her own field of research.  So attempt to view your presentation through the eyes of each committee member.  This will allow you to clarify possible points of confusion and to anticipate questions the committee member may ask during the exam.

4. Utilize the expertise of students, postdocs, and faculty members as you prepare.  Ask questions.  You are surrounded by experts whose job it is to help you learn to be an excellent scientist.

5. While your first task in preparing for the exam is to master the physics directly related to your presentation, your second is to begin to make connections outside of that presentation.   For example, you are expected to have mastered the concepts in all of your core graduate classes.  Expect a question based upon a concept learned in a core course which has some relevance to your project.  For instance, if your project involves scattering into a detector, you may be asked questions about scattering theory at the level of your core graduate courses.  As another example, you are expected to attend all of our general physics colloquia.  Expect a question about how your project might relate to and/or be affected by the results presented in a colloquium on a related field.  If you have taken a general survey course, specifically in your sub-area of physics, it is wise to review what you have learned in that course.

6. Ultimately your prelim committee will want to know how much you know, starting with your project and working outwards.  They will ask you difficult questions. They do not expect that you will be able to answer all of them.  Do not panic when committee members ask questions beyond your level of knowledge. They do this to help probe the limits of your understanding and to see how you "think on your feet" on topics with which you are not completely familiar.  In some cases, when you are having trouble with a question, the questioner might back up and ask you a more basic question to help guide you to the answer to the more difficult question.  Thus, there is some back-and-forth during the exam as the committee works with you to help define the current boundaries of your knowledge.  With the information obtained during your exam, the committee can suggest areas that you need to focus on as you enter the thesis phase of your graduate career.  In cases where your knowledge has serious holes in many areas, you will not pass the exam.

7. Ask your advisor about his/her expectations at the preliminary exam. Obtain a list of topics you should understand.  If uncertain of the level of expertise required, ask for examples.  Make sure that you understand from your advisor what, in his/her opinion, is required in order to pass the exam.

8. It is likely better to pick one result and explain it well than to try to present everything you know about a subject.  You may wish to prepare slides on additional topics to refer to if the opportunity arises during the question period, but the prepared talk should tell a single, logical story.  The committee wants to see that you can focus on essential points and present a satisfying explanation of a phenomenon -- and one good example is enough.  However, it is best to speak to your advisor about this point since what is appropriate may vary from project to project.

9. At your prelim exam, you are the speaker and should be afforded the respect due to a speaker.  That means that all questions from committee members should be addressed to you, not to your advisor or other committee members.  At the same time, the prelim is an exam and the "audience" consists of your examiners.  They are expected to ask you questions to determine if you have reached the level of expertise and ability required for Ph.D. candidacy. If you cannot answer a question, you may ask clarifying questions of the committee.  Committee members may give you hints as to how to think about the question.  If you still cannot answer the question, you can explain what you do know and/or what you would do in order to figure out the answer to the question.  The committee is just as interested in your thought processes when faced with a problem with which you are unfamiliar as they are in determining that you have the necessary physics facts at your fingertips.