Messier's catalogue contained many ‘nebulæ’, but spectroscopic work carried out during the 1860s by William Huggins, a brilliant English amateur, showed that there were two different types. Some, such as M 42 in Orion, were gaseous, while others, such as M 31, were apparently made up of stars. As long ago as 1755, Immanuel Kant had suggested that the starry nebulæ might be ‘island universes’ at vast distances, but there was no obvious way of deciding, and it was more generally believed that all nebulæ were members of the Milky Way.

There was a major development in 1845, when the third Earl of Rosse, using his remarkable home-made 72-inch reflector at Birr Castle in Ireland, found that M 51, in the constellation of Canes Venatici, was spiral in form; we now call it the Whirlpool. Other spirals were soon found, although for some years only the Birr telescope was capable of showing them as such. Other starry nebulæ were spherical, looking very like globular clusters, while others were elliptical or irregular. All were too far away to show measurable parallax, but in 1898 Agnes Clerke, a leading astronomical historian, referred to the island-universe theory as ‘a half-forgotten speculation’.

In 1920 there was a famous debate between two leading American astronomers, Harlow Shapley and Heber D. Curtis. By studying short-period variables in globular clusters, Shapley had given the first reasonably accurate value for the size of the Milky Way system, but he regarded the spirals as minor features of our Galaxy, while Curtis believed them to be external.