In the last few pages of the book, Diamond proposes that historians should treat their subject like a science. He says that many sciences, like geology and climate science, rely on "natural experiments" where scientists have to try to compare things that have already happened with one another and infer rules from the differences in what happened. Scientists have to do that because they cannot simply change some factor in the earth's atmosphere, for example, and see what happens. Diamond says historians should do the same with human history.
In the rest of the book, Diamond tries to do this. He does it most clearly in Chapter 2. There, he looks at Polynesian societies. He asks why various Polynesian societies turned out so differently. He comes to the conclusion that their differences were caused by the environments in which they arose. Another example of how Diamond does this is found in Chapter 15. There, Diamond looks at the differences between Aboriginal Australian societies and European societies in Australia. From the way they turned out, he concludes that Aborigines were not culturally inclined to reject farming and technology. Instead, he says that geography was behind the changes.
In ways like these, Diamond conducts "natural experiments" throughout his book.