摘要: The utility of physical measurement depends on effective sampling. Every sample should represent a definite body or class of soil. Most commonly, the body or class will be a morphologically defined soil horizon or stratigraphic layer with clearly stated dimensions. But unless the morphological criteria used for defining the body or class correlate well with the physical measurements, the body or class forms an inappropriate basis for extrapolation. Published correlations between horizons defined using conventional morphology and soil physical properties vary considerably (Butler and Hubble 1977; McKeague et al. 1984; Bouma 1989; van Genuchten and Leij 1992; van Genuchten et al. 1999). In some instances horizons and soil profiles are effective carriers of soil physical information (Baker 1978; Wosten et al. 1985); in others correlations between recognizable horizons and physical properties are poor (eg Kooistra et al. 1985; McKenzie and MacLeod 1989; McKenzie et al. 1991). This Handbook draws attention to predictive relationships, orpedotransferfunctions, and Chapter 22 addresses the subject directly. Soil physical measurements will have long-term value if they have an associated site and profile description that conforms to standards defined in the Australian Soil and Land Survey Field Handbook (McDonald et al. 1990). At some experimental sites, a single description may suffice for all the plots. However, in many instances there is too much spatial variation in soil morphology, and in those situations several descriptions of the site and profile are needed. While there are problems in correlating physical measurements with …