The Harvard College Observatory’s glass plates, which record a century of changes in the sky, have now been converted into digital form.

Orion in glass plate form
Cut-out is from a glass plate of Orion
DASCH / Harvard

Deep within the photographic plate stacks at the Harvard College Observatory, curator of astronomical photographs Thom Burns slips a delicate glass plate out of its brown paper sheath. He places it on a large, specialized scanner — custom-built for this purpose — preparing it for digitization, the astronomical markings encoded on its surface soon to be forever preserved in an online database. While there isn’t anything unique about this particular glass negative — its pale, filmy surface speckled with dark stars is a common sight in the stacks — there is nonetheless cause for tension.

This moment is the culmination of a project that has spanned decades: Digital Access to a Sky Century @ Harvard, or DASCH. Within the labyrinthian walls and filing cabinets of the Harvard College Observatory’s Plate Stacks reside some 600,000 glass historical plates, their collection spanning 1880 to 1990. Members of DASCH have endeavored to digitize all plates from which accurate brightness data can be obtained — which is no small feat, considering there are about 430,000 of them. By digitizing such an expansive collection, the project has granted astronomers unprecedented access to a century of sky. With more than 400 terabytes of imaging data now at their fingertips, astronomers unlock the potential for exciting discoveries.

A glass plate is placed on the custom-built scanner.
Kayleigh MacDonald

As Burns sets the plate in place and flicks off the lights, the room falls silent. Soon, the machine begins its work — red lights flickering hypnotically as the scanner measures the position and brightness of each tiny region, then shifts the plate over with microscopic precision. The entire process usually takes about 85 seconds, according to the project’s operation lead Peter Williams, which means that when the project was underway, the team would complete on average around 200 to 400 plates a day.

This painstaking “labor of love,” as Williams puts it, finally comes to an end as the final plate is scanned. As the lights flick on to enthusiastic applause, the plate’s data is sent off to an online database, where it will eventually be made available on the DASCH website.

After the plate-scanning, the project’s founder, Jonathan “Josh” Grindlay of Harvard University, offers the entire team praise for the devotion they’ve shown to the project since 2005, when the first plates were scanned.

Rho Ophiuchus in glass plate form
The first glass plate to be scanned, in 2005, showed the Rho Ophiuchus star-forming complex.
DASCH / Harvard

A Century of Fleeting Phenomena

DASCH provides astronomers with the benefit of hindsight, says Harvard astronomy professor Daniel Eisenstein: “We’re at a time now where astronomers are actively studying events that are arising now with very accurate new telescopes,” he explains. “But lacking that overall 100-year baseline can be a problem. So here, we can really unlock that and deliver it not just to scientists here, but around the world.”

Glass plate with annotations
Glass plates often contain annotations, sometimes hundreds of them.
Kayleigh MacDonald

In particular, astronomers are excited about the dataset’s ability to unearth new insights in areas of time-domain astrophysics, which includes the study of objects such as X-ray binaries and quasars that can evolve on human timescales.

However, the collection is more than just a repository of research data — the beginnings of photography and the progress of astronomical data collection can be traced through its images. “The collection represents the birth of photography itself,” Burns explains, from daguerreotypes to analog astrophotography. 

The collection boasts the first daguerreotype of the Moon, the discovery plate of the Horsehead Nebula, and the oldest photograph of a comet, among numerous other images of the Sun, Moon, eclipses, and stellar spectra.

Glass plates in envelopes
The envelopes housing the glass plates have become brittle with age. They were photographed to retain the information recorded before rehousing the glass plates in archival envelopes.
Lindsay Smith Zrull et al. / DASCH / Harvard

The archive also traces the work of astronomers at the Harvard College Observatory. DASCH preserves the meticulous records of 146 women, dubbed “computers,” who created, studied, and annotated the plates, including the likes of Annie Jump Cannon, Henrietta Swan Leavitt, Williamina Fleming, and Cecilia Payne-Gaposchkin.

Annotations include notes on spectral typing, on which Annie Jump Cannon built her spectral classification system, as well as plates marking the Cepheid variable stars that Henrietta Swan Leavitt studied to discover her famous period-luminosity relation. By digitizing the plates, “there’s so many more stories that we get to tell,” Burns says.

The Future: Project StarGlass

One project Burns still hopes to incorporate into DASCH is Project PHaEDRA, a “crowdsource transcription project” that records more than 2,500 logbooks from individual women computers. Burns hopes to link these logbooks to the corresponding digitized plates via their next big project: StarGlass.  “It's not only making the DASCH data more accessible to a wider group of people, but it is creating a system that we can continue on with digitizing,” Burns said.

Ben Sabath, a software engineer, has been developing the software to house StarGlass. “We’re still in the beta stage,” he explains as he clicks through the website, demonstrating how one can search for an object and sort by the different dates it was observed and by the astronomers who cataloged it.

The final plate in DASCH was scanned on Thursday, March 28th, 2024. Access the full DASCH archive here and see what StarGlass is all about here!

A team member notes the last day of scanning in the logbook.
Kayleigh MacDonald


Image of Andrew James

Andrew James

April 2, 2024 at 7:08 pm

Brilliant! Thanks so much for this article.

The availability of the data and images is extraordinarily useful. For me, it will add some of the research that I'm doing on historical basis. I've already looked at two or three images, I can now understand some of the comments that were made on southern open clusters.

Now reading Harlow Shapley's book on 'Star Clusters' (1930) means we can re-investigate the early classification methods they determined from these plates. Out early knowledge about the Milky Way and its structure come from these plates.

I wonder how Shapley would feel about this release? After all, he was the director of the Harvard college Observatory between 1921 and 1952

You must be logged in to post a comment.

Image of Rob


April 8, 2024 at 7:37 am

I found it rather curious that for a project of this nature was using a handwritten log book to document the scanning of each plate. Surely they should have used a digital notebook?

You must be logged in to post a comment.

You must be logged in to post a comment.