Focusing on the Harvard Observatory, she describes how two open-minded and pragmatic directors, Edward Pickering (from 1877 to 1919) and Harlow Shapley (1921-52), recruited scores of women at cut-rate wages to calculate, measure, classify and make sense of the glass universe produced by astrophotography. These archives of glass plates are the “glass universe” of Dava Sobel’s title the solution to the labour problem provides her subtitle. In Paris, Potsdam, Oxford, Algiers, Helsinki, Melbourne and Cambridge, Massachusetts tons of glass plates dotted with the imprint of millions of stars and their spectra piled up – and with them an urgent demand for highly skilled, conscientious and cheap labour to analyse them. Enterprising observatory directors all over the world launched grand astrophotographic initiatives to create new star maps, catalogues and spectroscopic classifications. In gleeful defiance of naysayers such as the French philosopher Auguste Comte, who only a few decades before had proclaimed the impossibility of astrophysics, astronomers set out to investigate what the stars were made of. Even more miraculously, the photographs could, with the use of diffraction gratings, yield stellar spectra striated by those characteristic dark lines that are the signatures of various elements.

Suddenly it became possible, with comparatively little effort, to capture millions more stars: whole swaths of the sky (including in the southern hemisphere, thanks to new observatories in colonial outposts such as South Africa and Australia) could be covered systematically stars too faint to be seen even with the best optical telescopes could be coaxed into visibility by increasing the exposure time and emulsion sensitivity of photographic plates. The invention of astrophotography in the mid-19th century intensified both the labour and archival challenges that had always confronted astronomers. (Compare the average physical lifetime of a CD, about 30 years.) Civilisations rise, civilisations fall, but astronomy endures. Nasa’s Five Millennium Canon of Lunar Eclipses (2000 BCE-3000 CE) includes observations originally recorded on cuneiform clay tablets by ancient Mesopotamian astronomers. As with spotting rarely visible phenomena, such as supernovae, it also requires techniques of record keeping and transmission that transcend the longevity of civilisations and the durability of most materials.

And so little time: what is a puny human lifetime compared with that of the universe, or even of a mere planet? It takes generations of diligent observers to detect the regularities that unfold over millennia – for example, the precession of the equinoxes: it takes nearly 26,000 years for the constellations to appear to rotate around the Earth. A stronomy has always had a labour shortage: so many stars, so few people to observe them.

The Glass Universe by Dava Sobel