Microsoft recently patented search that uses 'visual position' of the text in page to rank pages. I don't think SEs use external CSS files, they may have started doing in last couple of years.
Spiders will parse through CSS that is embedded directly on the page but will ignore an external CSS files. In order to decrease your code-to-text ration you want to move as much CSS off the page and in to an external CSS file as possible.
They can read them, but they don't incorporate them into search rankings. Instead, they use it as a quality control measure to identify those that try to manipulate the search rankings by using display: none; and same/similar foreground/background colors to stuff keywords and do other things I won't get into here.