The objective of this study was to investigate by computer simulation the hypothesis that the focal generation of an atherosclerotic plaque is a consequence of self perpetuating macrophage (Mph) recruitment. Atherosclerosis always develops as focal plaques, in which monocytes adhere to the endothelium and enter the artery to give rise to intimal macrophages. These cells oxidise low density lipoprotein to form products, e.g. oxidised phospholipids, that can activate the endothelium to increase monocyte adhesion. Hence the focal form of the plaque may arise through self perpetuation of monocyte entry. As macrophages can release factors that promote smooth muscle accumulation, or die to produce extracellular cholesterol, the dynamics of these other important plaque components are also studied. A computer model was written in Microsoft Visual C++, in which a computer screen represents either a) a surface view of a flattened area of an arterial wall, or b) a hardware accelerated 3D rendered model of an intact tubular artery. In a) each pixel is a cell-sized area of the wall that can be occupied by multiple smooth muscle cells (SMCs), Mphs and extracellular lipid, simulating the components that can be present through the thickening intima in vivo. Cell numbers are governed by probability functions: 1) Mph recruitment from blood monocytes at each pixel depends on an initial very low value, which increases as a function of the number of macrophages present at that pixel and those adjacent. 2) SMC accumulation depends on Mph number. 3) Mphs have a probability of death, which can depend on their number, and lipid accumulation is the direct result. 4) SMCs can optionally inhibit macrophage recruitment. These functions are programmed by graphical interfaces. The program runs in multiple cycles, in which the cell contents at multiple randomly chosen pixels are adjusted according to the prevailing probabilities, using further randomisation to determine the result. Mphs and SMC are displayed on the screen, coded by colour for cell type and number. The part a) of the program also allows graphical display of the stacked cell components along any line on the arterial surface, simulating an intimal cross-section. The 3D image in b) is calculated from the same cellular data as in a), assuming thickening of the arterial wall proportionate to the number of cells present. If macrophage recruitment probability is allowed to be a very steeply rising function of macrophage numbers, the program readily generates realistic lesions. Small randomly distributed fatty streak-like foci of macrophages are produced initially, which may progress or regress. Larger dome-shaped Mph rich plaques are then generated, with central fibrous cap-like central regions of SMCs. The central location of the fibrous cap and the peripheral macrophage only zone are seen as a consequence of the centrifugal growth of the plaque. Lipid accumulates centrally in the lesions. Allowing SMCs to inhibit Mph recruitment permits depletion of the central region of the lesion of Mph, as is often observed. The 3D rendered image shows a remarkable resemblance to an atherosclerotic artery in vivo. The realistic modelling of atherosclerotic plaque morphology supports the view that plaques result from the centrifugal spread of a self-perpetuating macrophage-dependent process.