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.