It is widely believed that polyanionic plasma proteins, such as albumin, are normally prevented from penetrating the glomerular capillary wall by the presence of intrinsic fixed negative charges in the wall. There is also strong support for the corollary that proteinuria occurs in glomerular disease as a result of loss of such charges. Morphologic studies using cationic 'stains' have indicated that anionic groups can be detected in normal glomeruli and that, in proteinuric states, there is an apparent reduction in such staining. Because most of the latter claims are based largely upon light microscopic studies, critical reassessment at the electron microscopic level is necessary, particularly using the recently developed technique of in situ drip-fixation of superficial glomeruli during good blood flow. In normal rats, this technique results in heavy cationic colloidal iron staining of podocytic epithelial surfaces but little or no staining in the basement membrane or within endothelial fenestrae. In two experimental proteinuric models, rat nephrotoxic nephritis and aminonucleoside nephrosis, there was no loss of colloidal iron staining on podocytic epithelial surfaces at any stage. There was, however, a striking alteration in glomerular architecture in each model, particularly affecting podocytic epithelium. Thus, there was extensive replacement of foot processes by flattened expanses of epithelial cytoplasm, associated with scattered focal gaps in the epithelial covering of the basement membrane. It therefore appears that the reduction of glomerular polyanion seen by light microscopy in glomerular disease results simply from a decrease in visceral epithelial surface area rather than a loss of intrinsic polyanion from the filtering portion of the glomerular capillary wall.