Comparison of the values for an elastic property of a particular metal or alloy quoted in the various data books and published in original papers indicates a remarkably wide spread. For any given material, values differing by 20% are quite common. Therefore any calculation using elastic properties as input data should explicitly declare the values that have been taken.

The sources of error in measurement of clastic properties arise partly from difficulties in measurement of clastic strain. The quantity to be measured is at most a few parts per thousand changes in length and to be measured with two to three figure accuracy. Other sources of error are more fundamental. For example, in the preparation of specimens for measurement of elastic properties, a certain degree of preferred orientation of the individual grains is introduced, thus affecting the properties in the ways described in Section 522. It should also be realized that data for isotropic material may not best represent the properties of components that might have developed preferred orientation of the grains during fabrication. Finally, both at room temperature and to a greater extent at elevated temperatures and at high stress levels stress-strain response of a specimen is time dependent and structure sensitive.

Where high accuracy is required as for example, in attempting to understand the performance of a particular fabricated heat exchanger, the best procedure would be to measure the elastic properties on samples of the actual materials at the appropriate temperature and strain rate.

Table 1 and Table 2 show single crystal clastic coefficients for hexagonal and cubic metals, respectively (Tang et al., 1969 and Smithells, 1967). They are included here to indicate the extent to which elastic anisotropy can be expected in wrought polycrystalline materials with preferred orientation of the grains.