Using data on 100 years of 19th century criminal trials at London's Old Bailey, this paper offers clear evidence of disparate treatment of Irish-named defendants and victims by English juries. We measure surname Irishness and Englishness using place of birth in the 1881 census. Irish-named defendants are 11% less likely to plea, 3% more likely to be convicted by the jury, and 16% less likely to receive a jury recommendation for mercy. These disparities are: (i) largest for violent crimes and for defendants with more distinctive Irish surnames; (ii) robust to case characteristic controls and proxies for signals associated with Irish surnames (social class, Irish county of origin, criminality); (iii) particularly visible for Irish defendants in cases with English victims; and (iv) spill-over onto English-named defendants with Irish codefendants. Disparate treatment is first visible in the 1830s, after which it grows, then persists through to the end of the century. In particular, the gap in jury conviction rates became larger during the twenty years after the Irish Potato Famine-induced migration to London. We do not find evidence, however, that the first bombing campaign of the Irish Republican Brotherhood (in 1867 and the 1880s) further exacerbated these disparities.