Comparison of cholesterol determinations by nurses using a Reflotron analyser in a general practice setting showed a good correlation with plasma cholesterol determinations by wet chemistry in a clinical biochemistry laboratory. A limited number of comparisons did, however, give a much lower result on the Reflotron. In an experimental situation, small sample volumes (which could result from poor technique) were shown to produce falsely low readings. A simple method which may immediately detect falsely low Reflotron readings is discussed.