A good paper starts with a good experiment, which in turn starts with a clearly stated hypothesis. The introduction must end with this hypothesis, and the final paragraph of your paper must say what you conclude about that hypothesis. Write a draft introduction before you start your experiment. You will update this before submission, but your hypothesis should not change. As you do your research, start writing sentences for your methods. Make sure they describe exactly what you are doing. This means you have read the original methodology, even if citing it through another paper. If you are also keeping track of the references you will cite, that makes 3 big parts of your paper that you have drafted before finishing your study. I prefer combining results and discussion text in one section and put most of my results in either a table or figure. Some reviewers want discussion of every result, but I submit papers that contain “minor” results in some table rows that I do not discuss. If a reviewer asks, add text to the discussion about these results. The less you say in a discussion, the more you focus on the core objective. I write for my reviewers, so I would rather have to add distracting discussion text at their request after they have read my focused discussion. Certainly, anything else interesting that comes up is worth mentioning but don’t elaborate or speculate on ideas beyond the scope of your experiment. Do know how your results compare with the existing literature to make sure something is not way off—you do have to discuss anything that seems unusual. A good paper comprises unambiguous, direct, short, minimally complicated sentences that each convey a single idea. Integrate related ideas using paragraphs of simple sentences. Have others read your paper, especially people outside your immediate research group. Writing is an iterative process, so expect to modify your precious writing in response to reviewers. Respect the input of all your reviewers, and if you failed to make them understand something, rewrite your paper so that anyone reading the final version is not similarly misled. Know the journal guidelines.