How to keep a lab journal
Keeping a lab journal is, above all, an act of discipline. You need consistency and a set of rules on which you have agreed yourself. Getting started is easy, but making it grow is hard. I have seen a lot of lab journals that are simply useless. They are a mere collection of dates and uncomprehensible notations about experiments that someone performed.
Remember, a lab journal is a bet for the long term, which means it must be written in a clear and consistent way. There are many reasons to answer why keeping a lab journal, and the only way of being successful in all of them is to know how to approach it.
Start with a date
A lab journal is chronologically ordered. Every time you sit down to write something on the journal you must be sure you are writing under the current date. This is important not only for potential legal claims, it is fundamental to find data back. When perform experiments, data will most likely be saved according to the date. This gives the possibility both of understanding what you were doing if you find the data or the other way around: find where the data of a given experiment is located.
It is normal that some dates are missing from the journal. You won't work every day in the lab. Whether each day should start on a new page is a matter of taste. I like writing in blocks so all the relevant information is on the same page. If I see a block won't fit in the remaining of the page when starting a new day, I simply start on a new one. If you are strict, a lab journal should not have empty spaces, to prevent altering it retroactively.
Write what you do
When you are about to perform a measurement, make sure you have all the relevant parameters written down. It is impossible to write down absolutely everything, but there are some things that are more important than others. For example, if you have different samples, you must write down which one you are using. Sometimes you don't know what parameters are important until you decide to change something.
These are the other notes that link to this one.