5.1 Engineered sampling plan 5.1.1 Divide the lot into at least two subdivisions(on a rational basis considering its form and condition) and take at least two subsamples from each subdivision. 5.1.2 For logs, a practical way to attain randomness is to subdivide the shipment into approximately equal quantities by carloads, truckloads, or cords, identify each subdivision by number, and select the subdivisions to be included in the sample by the use ofa random number table. From each equal subdivision so selectedan equal number of logs are to be taken at random during unloading or stacking. Thenumber of logs comprising the complete sample should be sufficient to produce at least twice the amount of sawdust or raspings required for the moisture content test and the chemical analyses desired. The amount ofmaterial obtained by sampling should be Teduced to the amount necessary for analysis by quartering(see TAPPIT605"Reducing aGross Sample of Granular or Aggregate Material to Testing Size") 5.1.3 subsamples of chips may be taken using a shovel or scoop after identifying subdivisions ofthe lot in the same way as for logs. 5.14 An engineered sampling plan requires accepting broad assumptions regarding the relative unimportance of, or the lack of differences in variation between and within subdivisions of the lot. It avoids questions regarding the limits of uncertainty in the decision made based on the sample, and it provides no real control over that uncertainty. It does not ensure a least-cost balance of sampling against the risk of economic loss due to a bad determination 5.2 Probability sampling plan 5.2.1 If the lot can be assumed to be homogeneous and a prior estimate of the standard deviation of the lot(derived from subsampling) is available, the number of subsamples(n) may be calculated by