Triton has a very thin atmosphere, composed essentially of nitrogen and methane in lesser proportions, but enough to allow for winds to blow always in the same direction This nitrogen is what condenses on the surface as ice to make it so shiny.

Triton was only observed closely by the Voyager 2 probe in 1989. No other future project is dedicated to it.

Finally, the origin of this satellite remains a mystery to this day. It's orbit leads to believe that it was not formed with the planet but in the Kuiper belt before being captured by Neptune. Triton would therefore be similar to Pluto, the dwarf planet, that is now considered to be part of the Kuiper belt.

The colour of the surface in white with a slight pinkish tint in places. The surface of Triton can be split into two distinct parts. The Southern hemisphere is the shiniest and the Northern hemisphere is darker and so more pinky-red. The white tint is due to the nitrogen ice and the pink colouring is due to the methane ice which takes this colour when exposed to sunlight.

When Voyager 2 passed near Triton, it was spring in the Southern hemisphere. This was marked by the presence of vast nitrogen icecaps (up to the equator) that were only just starting to sublimate. The equator is actually covered in an even shinier layer if ice that the Southern hemisphere, made of the nitrogen that condensed last. This equatorial zone also marks a quite definite boundary between the two types of surface (see photo 1).

The Northern hemisphere and the equatorial zone show signs of a younger surface (less than 2 billion years old): The craters are very small and numerous and there are also graben stripes on the surface showing that there is internal activity. On the whole, the topography of the Northern hemisphere is irregular when seen as high resolution.

But the physical aspect of some areas in the Northern hemisphere and equator are very strange and unique in the solar system. These and circular depressions 30 to 50 km wide separated by rough crests. They could be impact craters due to the regularity and size of the depressions and their distances from each other, their exact origins remain unknown but are most likely signs of internal activity. These zones are called Cantaloupe-skin terrain (see photo 2).

Next to this strange region, we find depressions that could be the faces of very large old impact craters that were modified by the flow, melting and crumbling of the materials on the surface (see photo 3). Many episodes of successive emptying ad refining seem to have taken place. In the centre of the basin in the centre of the photo we can see a rough area that could be due to a recent irruption of material.

The faults/grabens cross the Cantaloupe-skin terrain and are 35 km wide. The crest in the centre of the grabens could be ice rising from deep inside the grabens (see photo 4). These faults are proof of the existence of tectonic activity.

The Southern hemisphere is globally smoother but dotted with smaller volcanoes that also give the impression That there must be internal activity. The darker material (Nitrogen with a little methane and carbon) are from the depths near the surface and are ejected by small craters up to heights of 8 km. They are then blown all in the same direction by the dominant winds (see photo 5). On this photo, we can see 50 of these little fumaroles. One must not assume however that these 50 volcanoes are all active at the same time. These are actually traces on the surface left by each volcano but only 4 geysers are active in the images from Voyager 2. These traces on the ground are several km wide and several hundred km long. The lifetime of these are estimated to be between 1 to 5 earth years.

Triton's cryovolcanism is surprising due to its distance from the Sun and low surface temperature. To explain the current internal activity and the characteristics of Triton's surface, it is necessary to envisage an very particular composition and differentiated internal structure.

The particular importance of the seasonal cycle for Triton are pointed out by the spring activity of the geysers exclusively in the Southern hemisphere. An internal heating due to the radioactive disintegration of the core's silicates is sometimes considered to be necessary as well as the solar energy, to reach fusion temperatures for the nitrogen/methane/carbon mix on the surface. The heat would then be transmitted from the rocky core to an ice coat, then to the iced surface made up of the volatiles materials, maintaining the appearance of geysers.

In some models, the core would have a radius of about 1000 km and would account for 70 % of the total mass. The existence of an internal ocean of water potentially containing some ammoniac, is sometimes put forward. This ocean would be at a depth of 20 km into the satellite.

If we compare the size of Triton with the dimensions of Uranus' satellites, we notice that this satellite is particularly large for a gaseous planet of average size. Further more Triton's orbit in irregular, it has a high inclination and orbits in the retrograde direction to the rotation of Neptune (in the opposite direction to the rotation of Neptune). The satellites that are made from the disc of material in orbit around the planet have an inclination close to 0 and a pro-grade orbit. (in the same direction are the planet's rotation which is in the direction of the disc of rotating material...).

It is therefore practically certain that Triton was not formed near Neptune but in a distinct heliocentric orbit and, due to a series of unfortunate events, was captured by Neptune's gravitational field. The events that lead to the capture of a satellites by a giant planet are in general as follows:

• The atmosphere of the gaseous planets are sometimes far vaster in the beginning of their formation that at the present time, thanks to the heat liberated in the accretion process. This atmosphere could have therefore induced a slowing down of objects in orbits close by by cooling.
• A sudden increase in mass of the planet in the accretion process can upset the trajectory of other objects near to it.
• The gravitational interactions and collision with already existing irregular system satellites or with other planetesimals on the heliocentric trajectories could lead to the capture of a future satellite.

But it is hard to explain Triton's situation from these hypotheses. Neptune formed slowly and never had a very large atmosphere which topples the first two hypotheses. Finally, the mass of Triton is such that is unlikely that other system's satellites and planetesimals could have affected it's orbit enough.

The fourth explanation was elaborated on measurements from Triton:

According to this idea, Triton was one of two elements of an couple similar to the one formed by Pluton and Charon. This couple came close to Neptune with a speed of centre of mass relatively small compared to the planet. But each component of the couple having its own speed relative to the centre of mass, one of the two had a smaller individual speed relative to Neptune than the escape velocity of the planet. This component, Triton, became a prisoner of the satellites where as the other continued on its course.

For this model to be valid, the accretion disc in which Neptune was formed must have contain lots of bodies of size comparable to that of Triton or Pluto, which is possible. A non negligible fraction of these bodies must also have been linked in couples. This is also possible as shown in observations of the Kuiper belt, what is left of the accretion disc of the solar system, which determined that it is 10 to 15 % binary objects (the Pluto-Charon is a prime example).