Labov's 1963 study of /r/ in New York City department stores had three principal findings: 1) social stratification: use of consonantal /r/ in coda position (r-1) was correlated with the status of the store, i.e. more (r-1) in Saks than Macys 2) an age distribution suggesting ongoing 'change from above' towards increased (r-1) use 3) linguistic conditioning: more (r-1) in word-final position and emphatic repetitions. These observations have subsequently been reinvestigated, in 1986 by Fowler, and in 2009 by Mather, effectively providing a real-time trend series by replicating the original methodology. In this paper we replicate Labov’s methodology. The results indicate continuing progression in the direction Labov predicted. This is a unique case in variationist studies of change in progress; no other change has been so frequently sampled with controlled methodology across so long a time frame. The present study extends the duration of this series to 54 years. The original study investigated three department stores that were socially stratified by price level and target clientele: Saks, Macys and S. Klein. Klein, the low-end store, closed in 1976, but the other two still operate in their original locations in Manhattan and were investigated for the present study using Labov's original methodology. 160 speakers were observed in each store. The results show continued advance of (r-1) in real time. The percentage of speakers using all (r-1) has increased by a factor of 2.8 in both stores since Labov’s study, but the rate of change has accelerated considerably since Mather’s 2009 data collection. Social stratification of the variable is still apparent: Saks, the high-end store, continues to show higher rates of (r-1) than Macys. In Labov’s apparent time results, (r-1) use increased among younger speakers at Saks, but was higher among older speakers at Macys, suggesting that the change originated among higher status speakers, and spread to lower status speakers by diffusion in adult life. In the current study, this pattern has shifted. The age distribution at Saks is flattening out at a high level (speaker groups under age 50 all produce over 80% (r-1) tokens), while Macys now shows a conventional apparent time pattern, with (r-1) advancing among younger speakers. The linguistic conditioning on the process is moderating as the change approaches completion: Macys speakers showed moderate increases in (r-1) of about 7% between internal ('fourth') and final position ('floor'), and between non-emphatic and emphatic productions, but Saks employees are essentially uniform across all conditions.