All galaxies formed at about the same time approximately 13 billion years ago. The classical model on the origin of the galaxies says that they formed from huge gas clouds larger than the resulting galaxy. The clouds began collapsing because their internal gravity was strong enough to overcome the pressure in the cloud. If the gas cloud was slowly rotating, then the collapsing gas cloud formed most of its stars before the cloud could flatten into a disk. The result was an elliptical galaxy. If the gas cloud was rotating faster, then the collapsing gas cloud formed a disk before most of the stars were made. The result was a spiral galaxy. The rate of star formation may be the determining factor in what type of galaxy will form. But, perhaps the situation is reversed: the type of galaxy determines the rate of star formation. Which is the ``cause'' and which is the ``effect''? As of now, it is not known.
A more recent variation of the classical model says that there were extremely large gas clouds that fragmented into smaller clouds. Each of the smaller clouds then formed a galaxy. This explains why galaxies are grouped in clusters and even clusters of galaxy clusters (superclusters). However, the model predicts a very long time for the collapse of the super-large clouds and fragmentation into individual galaxy clouds. There should still be galaxies forming today and that is not observed.
The other model builds galaxies from the merging of smaller clumps about the size of a million solar masses (the sizes of the globular clusters). These clumps would have been able to start collapsing when the universe was still very young. Then galaxies would be drawn into clusters and clusters into superclusters by their mutual gravity. This model predicts that there should be many more small galaxies than large galaxies---that is observed to be true. The dwarf irregular galaxies may be from cloud fragments that did not get incorporated into larger galaxies. Also, the galaxy clusters and superclusters should still be in the process of forming---observations suggest this to be true, as well.
Collisions take place over very long timescales compared to the length of our lifetime---several tens of millions of years. In order to study the collisions, astronomers use powerful computers to simulate the gravitational interactions between galaxies. The computer can run through a simulation in several hours to a few days depending on the computer hardware and the number of interacting points. The results are checked with observations of galaxies in different stages of interaction. Note that this is the same process used to study the evolution of stars. The physics of stellar interiors are input into the computer model and the entire star's life cycle is simulated in a short time. Then the results are checked with observations of stars in different stages of their life.
In the past, computer simulations used several million points to represent a galaxy to save on computer processing time. A simulation of several million points could take many weeks to process. However, galaxies are made of billions to trillions of stars, so each point in the simulation actually represented large clusters of stars. The simulations were said to be of ``low resolution'' because many individual stars were smeared together to make one mass point in the simulation. The resolution of a computer simulation does affect the result, but it is not known how much of the result is influenced by the resolution of the simulation and how much the ignorance of the physics plays a role. Computer hardware speeds and the programming techniques have greatly improved, so astronomers are now getting to the point where they can run simulations with several billion mass points in a few weeks time.
When two galaxies collide the stars will pass right on by each other without colliding. The distances between stars is so large compared to the sizes of the stars that star-star collisions are very rare when the galaxies collide. The orbits of the stars can be radically changed, though. Gravity is a long-range force and is the primary agent of the radical changes in a galaxy's structure when another galaxy comes close to it. Computer simulations show that a small galaxy passing close to a disk galaxy can trigger the formation of spiral arms in the disk galaxy. Alas! The simulations show that the arms do not last long. Part of the reason may be in the low resolution of the simulations.
The stars may be flung out from the colliding galaxies to form long arcs. Several examples of very distorted galaxies are seen with long antenna-like arcs. In some collisions a small galaxy will collide head-on with a large galaxy and punch a hole in the large galaxy. The stars are not destroyed. The star orbits in the large galaxy are shifted to produce a ring around a compact core.
Select the ``Antennae Galaxies formation movie'' link below to show a movie of a computer simulation from Joshua Barnes showing the formation of the Antennae Galaxies. It is a Quicktime movie, so you will need a Quicktime viewer. The red particles are the dark matter particles and the white and green are stars and gas, respectively. Other collision movies are available on Barnes' Galaxy Transformations web site and on Chris Mihos' Galaxy Collisions and Mergers website.
Here are some photographs of examples of these collisions. The first is the Antennae Galaxies (NGC 4038 & NGC 4039) as viewed from the ground (left) and from the Hubble Space Telescope (right). Note the large number of H II regions produced from the collision. The second is the Cartwheel Galaxy as seen by the Hubble Space Telescope. A large spiral was hit face-on by one of the two galaxies to the right of the ring. The insets on the left show details of the clumpy ring structure and the core of the Cartwheel. Selecting the images will bring up an enlarged version in another window. See Mihos' galaxy modelling website for simulations of the creation of the Cartwheel Galaxy.
The gas clouds in galaxies are much larger than the stars, so they will very likely hit the clouds in another galaxy when the galaxies collide. When the clouds hit each other, they compress and collapse to form a lot of stars in a short time. Galaxies undergoing such a burst of star formation are called starburst galaxies and they can be the among the most luminous of galaxies.
Though typical galaxy collisions take place over what to us seems a long timescale, they are short compared to the lifetimes of galaxies. Some collisions are gentler and longer-lasting. In such collisions the galaxies can merge. Computer simulations show us that the big elliptical galaxies can form from the collisions of galaxies, including spiral galaxies. Elliptical galaxies formed in this way have faint shells of stars or dense clumps of stars that are probably debris left from the merging process. Mergers of galaxies to form ellipticals is probably why ellipticals are common in the central parts of rich clusters. The spirals in the outer regions of the clusters have not undergone any major interactions yet and so retain their original shape. Large spirals can merge with small galaxies and retain a spiral structure.
Some satellite galaxies of the Milky Way are in the process of merging with our galaxy. A recently discovered dwarf galaxy in the direction of the Milky Way's center is stretched and distorted from the tidal effects of the Milky Way's strong gravity. A narrow band of neutral hydrogen from other satellite galaxies, the Magellanic Clouds, appears to be trailing behind those galaxies as they orbit the Milky Way. The band of hydrogen gas, called the ``Magellanic Stream'', extends almost 90° across the sky away from the Magellanic Clouds and may be the result of an encounter they experienced with the Milky Way about 200 million years ago.
The giant ellipticals (called ``cD galaxies'') found close to the centers of galaxies were formed from the collision and merging of galaxies. When the giant elliptical gets large enough, it can gobble up nearby galaxies whole. This is called galactic cannibalism. The cD galaxies will have several bright concentrations in them instead of just one at the center. The other bright points are the cores of other galaxies that have been gobbled up.
If collisions and mergers do happen, then more interactions should be seen when looking at regions of space at very great distances. When you look out to great distances, you see the universe as it was long ago because the light from those places takes such a long time to reach us over the billions of light years of intervening space. Edwin Hubble's discovery of the expansion of the universe means that the galaxies were once much closer together, so collisions should have been more common. Pictures from the Hubble Space Telescope of very distant galaxies show more distorted shapes, bent spiral arms, and irregular fragments than in nearby galaxies (seen in a more recent stage of their evolution).
The formation of galaxies is one major field of current research in astronomy. Astronomers are close to solving the engineering problem of computer hardware speeds and simulation techniques so that they can focus on the physical principles of galaxy formation. One major roadblock in their progress is the lack of understanding of the role that dark matter plays in the formation and interaction of galaxies. Since the dark matter's composition and how far out it extends in the galaxies and galaxy clusters is unknown, it is not known how to best incorporate it into the computer simulations. Faced with such ignorance of the nature of dark matter, astronomers try inputting different models of the dark matter into the simulations and see if the results match the observations. Some simulations suggest that a spiral galaxy without a massive dark matter halo may form a bar across the middle of it---a barred spiral galaxy is the result.
There will be many new fundamental discoveries made in the coming years, so this section of the web site will surely undergo major revisions of the content. Although the content of our knowledge will be changed and expanded, the process of figuring out how things work will be the same. Theories and models will be created from the past observations and the fundamental physical laws and principles. Predictions will be made and then tested against new observations. Nature will veto our ideas or say that we are on the right track. Theories will be dropped, modified, or broadened. Having to reject a favorite theory can be frustrating but the excitement of meeting the challenge of the mystery and occasionally making a breakthrough in our understanding motivates astronomers and other scientists to keep exploring.
Go back to previous section --
Go to next section
last updated: June 3, 2004