What do I think it takes to be a scientist?
I'm going to approach this question using a combination of my experience and reason. If other people have experience and reason they'd like to add, we can come up with a scientific understanding of what science is:
First, you need to be able to write well. If you can't write well, nobody is going to support your research projects. If you're fortunate enough to obtain a government grant, you're almost in the clear because you don't have to worry about your funds drying up except for changes in administration. Also, there is a constant battle for lab space with the other research scientists. If another scientist wants your lab space, you'd better make darned certain you have students working in there, 24/7, or he can make a case to the department that you aren't using the space efficiently. Every now and then, he may walk by the lab and take mental note as to how many students there are, nod a self-satisfied nod, and walk on. One thing that may help to maintain lab space is to offer another, non-threatening scientist access to your lab where he can put students; thus causing the appearance of efficient use of space. If you need resources (computers or other equipment, chemicals, etc.), you can either use some of your hard-earned grant money, or you can mooch off another scientist (while he isn't paying attention), or you can make a request to the department.
Sometimes you are required to teach a course. If this comes up, don't fret. Make a case to the department that your research is critically important, and request extra money for graduate students. Grad students are key in teaching courses. They gain valuable experience, help solve their own (precarious) monetary problems, and probably already know the particulars of what needs to be taught. Suffice to say, there's no need for you to "get your hands dirty" with said particulars.
When it finally comes time to publish (and those darned journals only publish a couple times a year), you need to stay up all hours of the night waiting on your incompetent students to email you a "final draft." "Final draft" is in quotes because, although it's a draft, it's almost never final. Grad students display all of the grammatical and spelling power of a second grader, and typically copy and paste the text they've written and formatted from MS Word to LaTeX. If you're really lucky, you have a paper that can be reworked to the point where it's ready to make the (extended) submission deadline.
Publishing books is another matter, altogether. If you've never published a book, you're in for quite a ride. I won't get into that, here. If you've already published, you can modify what you've written, update it with new examples, rework the outline, and republish it as a "second edition." This is particularly useful since every time you republish, your students have to rebuy. If you've managed to "get in good" with another research scientist who has already published, you can help him rework his book and be added to the list of coauthors on the second edition's cover. Publishing books is the most effective means you have of acquiring lab space.
"Publish or Die"