The test fluids were three aqueous solutions of Natrosol 250 H hydroxyethyl cellulose. Their nominal concentrations were 0.3, 0.5, and 0.7%. Viscosity data for each solution wee filled with a generalized Newtonian viscosity model which properly describes the zero‐shear viscosity. The test geometries were two conical sections. Their vertex angles were approximately 14 and 21°. Laminar flow rate vs. pressure drop data were taken for each Natrosol solution in each conical section. Approximate expressions relating flow rate and pressure drop were derived for the limiting cases of very low and very high flow rates. The low flow rate (non‐Newtonian flow) expression was of form: pressuredrop=functionof (flow rate, geometry, viscosity model parameters). The high flow rate (inviscid flow) expression was of form: pressuredrop=functionof (flow rate, geometry, fluid density). These two expressions were in excellent agreement with data at low and high flow rates. The sum of these two expressions was in good agreement with data over the entire range of flow rates. This superposition expression in no way accounts for normal stresses, time‐dependent elastic effects, or the effect of the third invariant on viscosity. Its success in describing the data implies that these phenomena were not important. For engineering purposes generalized Newtonian viscosity models will probably be adequate for characterizing the flow of dilute polymer solutions in conical sections.