The FOOT-STRUT split, which has its origins in 17th century English, is notable for never having occurred in the speech of Northerners in England; thus, stood~stud are still homophones. The present study analyses 122 speakers from Manchester in the North-West of England. Although the vast majority of speakers have no distinction between the two vowels in minimal-pair production and judgement tests, vowel height is correlated with socio-economic status: the higher the social class, the higher the F1 of STRUT. Surprisingly, in statistical models the predictor of vowel class remains significant. This means that, for a speech community without the split, there remains an effect in the expected direction: STRUT vowels are lower than FOOT vowels in the vowel space (i.e. they have a higher F1). We suggest that co-articulatory effects of surrounding consonants explain this acoustic difference, as they have significant lowering/heightening effects on F1 but are not fully captured by our statistical model. We argue that the perplexing nature of the historical split can be partially accounted for in this data, as the frequency of co-occurring phonetic environments is notably different for FOOT than in STRUT, resulting in cumulative effects of co-articulation. We also present evidence of age-grading which suggests that middle class speakers may develop a phonetic distinction as they age.