The samples are the units that provide the experimental observations, such as tablets sampled for potency, patients sampled for plasma cholesterol levels, or tablets inspected for defects. The sampling procedure is an essential ingredient of a good experiment. An otherwise excellent experiment or investigation can be invalidated if proper attention is not given to choosing samples in a manner consistent with the experimental design or objectives. Statistical treatment of data and the inference based on experimental results depend on the sampling procedure. The way in which samples should be selected is not always obvious, and requires careful thought.