The downstream development of drag reduction following wall injection of concentrated polyethyleneoxide solutions into a turbulent water flow was studied by simultaneous pressure drop and flow visualization measurements at Reynolds number 55000 in a transparent acrylic test pipe of 15.7 mm ID and L/D = 330. The experiments partially modeled drag reduction in the Trans-Alaska Pipeline System, by roughly matching (i) the pipe friction velocity u¦" ~ 0.15 m/s, (ii) the backbone chain links Nbb ~ 0.5x106 and downstream concentrations Cline ~ 5 wppm of the polymer additive and (iii) the wall injection of concentrated polymer through equi-spaced circumferential ports at friction-normalized injection velocities Vinj+ ~ 1 << mean velocity V+ ~ 20. Polymer dissolution into the flow was assessed two ways, (i) from the initial increase in fractional drag reduction DR towards the eventual downstream DRline, which provided a characteristic development distance (L/D)dr and (ii) from visual observation and flash photography to reveal a characteristic distance (L/D)pv when red-dyed strands of the injected solution disappeared. Specific results were: (i) Friction factors for solvent, distilled water, adhered to the Prandtl-Karman law for fully-developed turbulent flow in smooth pipes, and injection of red-dyed water "blanks" gave (L/D)pv = 45±5, with corresponding eddy diffusivity to kinematic viscosity ratio (¦Å/¦Í) = 140, of the magnitude expected at our Re°Ìf = 4000. (ii) Concentrated polymer solutions with Cinj from 500 to 5000 wppm provided downstream polymer concentrations Cline from 1 to 50 wppm that exhibited DRline from 0.49 to 0.74, the latter close to the asymptotic maximum drag reduction MDR = 0.78 possible at the present Re = 55000. (iii) Drag reduction development distances (L/D)dr = 79±7 were essentially constant over the 10- and 50-fold ranges of Cinj and Cline respectively traversed. (iv) Dyed polymer strand disappearance distances were (L/D)pv = [140±30, 250±50, ~330, >330] at Cline = [1.0, 2.0, ~5, °Ý10] wppm, respectively, all of the latter attained with each of two Cinj = 500 and 1000 wppm. The insensitivity of (L/D)dr to Cline possibly reflects the rapid initial disruption and dissolution of the original polymer "threads" injected into a Newtonian turbulence, whereas the exquisite sensitivity of (L/D)pv upon Cline might mirror slow dissolution of the final polymer "strands" in a drag-reduced turbulence.