Since 1951, when the percutaneous renal biopsy was introduced as an adjunctive method for study of patients with renal disease, reports of some 4,000 kidney biopsies have appeared in the literature. Only about 250 of these, however, have been performed in children. A biopsy specimen containing 5 to 10 glomeruli has been reported to be adequate for interpretation and to be representative of the total renal parenchyma in 84% of the cases with diffuse renal disease. Using a biopsy technique similar to that described by Kark, we have obtained an adequate specimen in 92% of 205 kidney biopsies performed in 168 children with diffuse renal diseases. Seven deaths have been previously reported in the literature. The circumstances surrounding the death of these seven patients and of the one death that occurred in our series are described. Perirenal hematoma has had a reported incidence of 0.4%. It has been our experience, as well as that of the other investigators, that if blood boss is replaced, the patient has an otherwise uneventful course and the mass subsequently disappears. Gross hematuria has had a reported incidence of 5.2%. Microscopic hematuria, lasting for 6 to 12 hours after biopsy, has been found to be the rule rather than the exception. The complications which have occurred have been associated with bleeding, and therefore a careful history concerning bleeding tendency and a study of the clotting mechanism is essential if the risk of needle renal biopsy is to be minimized. In addition to a bleeding tendency or defect in clotting mechanism, most investigators are agreed that the presence of only one kidney or an uncooperative patient are absolute contraindications to renal biopsy. The renal biopsy is primarily, at present, an additional and most useful investigative tool in the elucidation of the pathogenesis, natural history (by serial studies) and effectiveness of specific therapy upon the various renal diseases. It is of practical clinical importance in the selection of those patients with the nephrotic syndrome in whom glucocorticoid therapy is likely to be beneficial or the patient with anuria whose renal lesion is probably reversible with time; and, as a guide to the effectiveness of therapy in patients with pyelonephritis or lupus nephritis. It is not a technique that can be recommended for general or casual use. A classification of the pathohistobogic findings of diffuse glomerulonephritis, patterned after Ellis, is presented and discussed. This classification will be used in the description and discussion of various renal diseases and systemic diseases with associated nephritis in the three subsequent papers.