A selector comprises every part of a rule set up to—but not including—the left curly brace {. A selector is a pattern, and the declarations within the block that follows the selector are applied to all the elements that match this pattern. In the following example rule set, the selector is h2:

h2 {
  color: #666;
  font-weight: bold;
}

This selector—which is comprised of a single simple selector—will match all elements of type h2 in an HTML document. A simple selector can either be an element type selector or the universal selector, (*), optionally followed by attribute selectors, ID selectors, or pseudo-classes.¹ A selector can comprise a number of simple selectors separated by combinators, but it can contain only one pseudo-element, which must be appended to the last simple selector in the chain.

Here’s a more complex selector:

h2+p.warning:first-line {
  color: #666;
  font-weight: bold;
}

This selector consists of two simple selectors separated by an adjacent sibling combinator (the + character), and a pseudo-element (:first-line). The first simple selector (h2) is a type selector. The second simple selector contains a type selector (p) and an attribute selector—in this case, a special form of attribute selector called a class selector, which will match HTML class attributes containing the word “warning.”

As such, the selector above would match the first line of text within any p element that has a class attribute value of "warning" and is an adjacent sibling to an h2 element.

Finally, two or more selectors can be grouped, separated by commas (,); the declaration block that follows applies to both selectors. Consider these two rules:

#main ol {
  margin-left: 2em;
}
#main ul {
  margin-left: 2em;
}

They can be grouped like this:

#main ol, #main ul {
  margin-left: 2em;
}

You can read about selectors in detail in the selector reference section.