Recent advances in metabolic scaling theory have highlighted the importance of vascular network geometry in understanding the integration and scaling of whole-plant form and function. Additional work has demonstrated general scaling relationships for many leaf traits. However, it is unclear if a common theoretical framework can reveal the general rules underlying much of the variation observed in scaling relationships at the whole-plant and leaf level. Here we present an extension of the general model introduced by West, Brown and Enquist, that predicts scaling relationships in whole plants and leaves. Specifically, the model shows how the exponents describing how whole plant and leaf dimensions (surface area, length and diameter) should change with increasing mass (or with one another) as plants and leaves meet the demands of network flow in varying environments. As a result, our expanded network model predicts a highly constrained continuum of biological exponents that covary according to specific quantitative functions. Compilations of allometric data for a wide variety of plant taxa strongly support the extended models predictions. Together, our results (i) underscore the importance of network geometry in generating the diversity of allometric scaling relationships in biology, (ii) support the hypothesis that while selection has primarily acted to minimize the scaling of hydrodynamic resistance and, (iii) additional selection pressures for alternative branching geometries apparently govern much of the observed co-variation in intraspecific and interspecific scaling exponents. Further, our model provides a general theory to understand the origin and variation of many allometric trait correlations within complex integrated phenotypes.