In 1950, Dutch astronomer Jan Oort hypothesized that comets came from a vast shell of icy bodies about 50,000 times farther from the Sun than Earth is. A year later astronomer Gerard Kuiper suggested that some comet-like debris from the formation of the solar system should also be just beyond Neptune. In fact, he argued, it would be unusual not to find such a continuum of particles since this would imply the primordial solar system has a discrete "edge."
This notion was reinforced by the realization that there is a separate population of comets, called the Jupiter family, that behave strikingly different than those coming from the far reaches of the Oort cloud. Besides orbiting the Sun in less than 20 years (as opposed to 200 million years for an Oort member), the comets are unique because their orbits lie near the plane of the Earth's orbit around the Sun. In addition, all these comets go around the Sun in the same direction as the planets.
Kuiper's hypothesis was reinforced in the early 1980s when computer simulations of the solar system's formation predicted that a disk of debris should naturally form around the edge of the solar system. According to this scenario, planets would have agglomerated quickly in the inner region of the Sun's primordial circumstellar disk, and gravitationally swept up residual debris. However, beyond Neptune, the last of the gas giants, there should be a debris-field of icy objects that never coalesced to form planets.
The Kuiper belt remained theory until the 1992 detection of a 150-mile wide body, called 1992QB1 at the distance of the suspected belt. Several similar-sized objects were discovered quickly confirming the Kuiper belt was real. The planet Pluto, discovered in 1930, is considered the largest member of this Kuiper belt region. Also, Neptune's satellites, Triton and Nereid, and Saturn's satellite, Phoebe are in unusual orbits and may be captured Kuiper belt objects.
To isolate and subtract the effects of cosmic ray strikes on the WFPC 2's electronic detectors, which could mimic the faint signature of a comet, thirty-four images were taken of the same piece of sky. The cosmic ray hits change from picture to picture, but real objects remain constant. However, pinpointing comets was even trickier because they drift slowly along their orbit about the Sun. Although the orbital periods of these objects are 200 years or longer, the HST has sufficient spatial resolution to see them move in just a few minutes. This means the comets change position from picture to picture, just as cosmic ray strikes would. However, cosmic ray strikes are randomly placed events while the motions of the comets are well defined.
To distinguish between the comets and cosmic ray effects, the 34 images were then digitally shifted and stacked to the predicted offset to account for the expected drift rate of comets. It's like having a fixed camera on a tripod take a rapid series of snapshots of someone walking in front of the lens. The resulting snapshots could be stacked so that the person appeared stationary.
The researchers tested the reliability of this approach by shifting the stacked pictures in the opposite direction of the expected comets' motion. Ideally, no comets should have appeared, but random alignments added up to 24 anomalous detections.
When the team stack-shifted the pictures in the direction of the predicted comet motion, they came up with 53 objects. Assuming that 24 of these are, statistically, anomalous too, leaves a remainder of 29 objects considered "real."
The shift-stack technique was further tested by dividing the images into two groups and running an automated search algorithm to look for objects that showed up in the same position on sets of exposures.