What other respondents have described here is correct.
However, a much more accurate method is to spot meter distant scenes, very especially when there is mixed light (bright light, shadows, shifting light). Incident/reflective readings assume an overall average value of luminance across the scene, and this is the problem — such scenes I mentioned are not average in their distribution: there will be reflections, shadows, light areas, poorly defined areas...all have their own values.
I do, strongly, advocate photographers diversify their metering regime and not rely solely on incident readings from afar. This cannot be emphasised enough in saving people from potentially disastrous results using transparency film in lighting conditions that can vary from marginal to extreme.