First author

Stars form in dense clouds of cold, turbulent molecular gas, but astronomers have had difficulty dissecting these clouds' complex structures to determine the mechanisms that drive star formation. On page 63, Alyssa Goodman, an astrophysicist at the Harvard-Smithsonian Center for Astrophysics, and her colleagues adapt the dendrogram — a technique typically used to depict evolutionary relationships using a 'tree' diagram — to obtain a realistic representation of the internal structure of a molecular cloud. Goodman describes how her team created the first three-dimensional (3D) figure that can be manipulated by Nature readers.

Was it difficult to show your data in three dimensions?

We got lucky. At a conference in 2004, I spoke about how dissatisfied I was with the way astronomers typically deal with 3D data — by compressing them to 2D or showing a series of 'slices' through a cube. Afterwards, Michael Halle, a computer scientist at the Brigham and Women's Hospital in Boston, Massachusetts, said that his surgical-planning software, used to image organs ahead of and during surgery, could manipulate and analyse high-dimensional data and might solve my problems. Co-author Michelle Borkin, then an undergraduate student, tried it with our data, and a graphics company then helped us put the 3D pdfs together.

Why did you decide to use a dendrogram?

We knew we needed a way to describe the interconnectedness — how certain 3D objects are nested within other 3D objects — of different cloud regions. Drawing a tree is a natural way to explain that kind of hierarchical relationship, but is rarely used in our field.

Will your findings usher in a new era of research?

Yes. Our work shows that an object's own gravity is an important force working, over millions of years and on multiple length scales, to shape a cloud's internal structure. Theoreticians typically ignore gravity in star-formation simulations until the final stages, when fluctuations previously caused by turbulence need to collapse in order for a star to form. But now simulators will need to include gravity earlier on.

Will this help to explain how stars acquire their mass?

It should, because now we can isolate self-gravitating — presumably star-forming — gas from its more transient surroundings. At very high resolution, this procedure will soon allow us to directly compare the frequency distribution of self-gravitating gas mass with that of stellar mass.