If I had to decide what makes a good science teacher, it would be the ability to demonstrate how experiments fit into the proverbial "scheme of things". Nothing kills interest in science faster than 1. not being able to accurately relay the structure of the big picture and 2. just tossing a bunch of apparently random experiments at the students and expecting them to figure out how the pieces fit together. You wouldn't attempt to put a jigsaw puzzle together in the dark, would you?
A second skill that can make a science teacher go from great to outstanding is the ability to motivate students into figuring out the answers themselves; good research skills, the desire to seek answers, and the ability to effectively communicate the results are traits that must be cultivated.