This paper illustrates how the concept of the malignant cell per se as the prime and only target in cancer therapy may be erroneous. The micro-vasculature evoked to satisfy nutritional requirements of solid tumors, and the inadequacy of this nutrition for all tumor cells, provide novel targeting concepts. The vascular architecture and the microenvironmental gradients (VAMP) will differ from one tumor to another and may determine whether current therapies succeed or fail. Many agents have a different toxicity or mode of action at the pathophysiological oxygen tensions that prevail in solid tumors. This warrants more attention. The hypoxic cell or the immature proliferating endothelial cell may provide tumor specificity that is more general than, and greater than, that conferred by the process of malignant transformation. The poor vasculature of solid tumors is often regarded as a problem by the oncologist. It limits the access of cytotoxic drugs, monoclonal antibodies, cytokines, etc. It also leads to hypoxic radioresistance because of diffusion limited chronic hypoxia and perfusion limited intermittent hypoxia, resulting from transient vessel closure. However, it can also be seen as a potential target, since prolonged vessel occlusion can lead to an avalanche of cell death. Strategies to prevent further expansion of the vascular network (anti-angiogenesis) should stabilize tumors and prevent further growth. Vascular targeting, aiming to damage the microvascular function and cause occlusion, can lead to extensive cell death. The target may relate to the excessive proliferation of endothelial cells in tumors or to abnormal functional aspects, such as altered cell shape (influencing permeability) adhesiveness to leukocytes or steps in the coagulation cascade. These microvascular features and microenvironmental gradients, and the phenotypic consequences of them, have been relatively neglected. The altered milieu and inadequate neovasculature is a common feature of all types of solid tumor, whereas the genetic changes that can give rise to a malignancy are very variable, from tumor site to site and even within a site from individual to individual. It seems, therefore, that therapies that could be of widespread general applicability might more easily be found from the micro-environmental or anti-vascular approaches than from gene therapy targeted at specific oncogenes. This approach will require cross fertilisation between scientists from quite disparate backgrounds, whose paths seldom cross, and who may not read, or even scan, each other's literature. If the endothelium or the low oxygen tension in subsets of tumor cells are the key to successful cancer treatment in mice, there are considerable implications for screening methods in vitro and for predictive and prognostic tests made on homogenized tumor samples.