Velilind's Laws of Experimentation: (1) If reproducibility may be a problem, conduct the test only once. (2) If a straight line fit is required, obtain only two data points.