The evolution of the vertebrate kidney records three occasions, each separated by about 50 million years, when fish have abandoned glomeruli to produce urine by tubular mechanisms. The recurring dismissal of glomeruli suggests a mechanism of aglomerular urine formation intrinsic to renal tubules. Indeed, the transepithelial secretion of organic solutes and of inorganic solutes such as sulfate, phosphate, and magnesium can all drive secretory water flow in renal proximal tubules of fish. However, the secretion of NaCl via secondary active transport of Cl is the primary mover of secretory water flow in, surprisingly, proximal tubules of both glomerular and aglomerular fish. In filtering kidneys, the tubular secretion of solute and water is overshadowed by reabsorptive transport activities, but secretion progressively comes to light as glomerular filtration decreases. Thus the difference between glomerular and aglomerular urine formation is more a difference of degree than of kind. At low rates of glomerular filtration in seawater fish, NaCl-coupled water secretion serves to increase the renal excretory capacity by increasing the luminal volume into which waste, excess, and toxic solutes can be secreted. The reabsorption of NaCl and water in the distal nephron and urinary bladder concentrates unwanted solutes for excretion while minimizing renal water loss. In aglomerular fish, NaCl-coupled water secretion across proximal tubules replaces glomerular filtration to increase renal excretory capacity. A review of the literature suggests that tubular secretion of NaCl and water is an early function of the vertebrate proximal tubule that has been retained throughout evolution. Active transepithelial Cl secretion takes place in gall bladders studied as models of the mammalian proximal tubule and in proximal tubules of amphibians and apparently also of mammals. The tubular secretion of Cl is also observed in mammalian distal tubules. The evidence consistent with and for Cl secretion in, respectively, proximal and distal tubules of the mammalian kidney calls for a reexamination of basic assumptions in renal physiology that may lead to new opportunities for managing some forms of renal disease.