I invite comment on the following hypothesis and possible lines of evidence on the Late Heavy Bombardment.
Hypothesis: The sun was, until about 3.8 billion years ago, a member of a galactic star cluster. The Late Heavy Bombardment was the concluding phase of the consequences of the membership of the Sun, with its solar system, in that cluster. As a member of that cluster, the sun passed through a stellar background that was considerably more congested than the setting it has been in at present, or has been for the last 3.8 billion years.
The result was that the comets in the Oort cloud and the bodies in as far as the Kuiper Belt, were stirred into unstable orbits much more frequently than after the Sun escaped the gravitational influence of that cluster.
The sun's orbit within that cluster took it at times near the "core" of the cluster and at other longer periods in the nether reaches of that cluster (though even then in a considerably more crowded stellar background than that outside such clusters.
Thus, after the first 100 million or so years, when the sun and the solar system collected the debris of the refractory material and the material around the gas giant planets of the primordial nebula, there was a prolonged period in which the main influence on the planetesimals of the Oort cloud and the other more remote bodies of the outer solar system was the passage of the Sun's sister stars. This probably happened in pulses, as the Sun passed through the cluster core or near other stars in the outer parts of the cluster rather than at an even pace.
The overall rate gradually declined as the cluster gradually diffused. Finally, the Sun itself escaped, or was ejected from that cluster. When that happened, what remained of the Oort Cloud stabilized and is now disturbed much less frequently. There would still be pulses as the Sun crosses the plane of the galaxy or as random stars approach the sun closely, but the 500-fold decline in the rate of cratering between 3.8 and 3.2 billion years ago reflects the Sun's transition from a member of a family of stars to its present status as a solitary star.
Argument and possible ways to test hypothesis:
Stars are typically formed in groups in nebulae. As the material in the nebula is used up by newly-formed stars or is dispersed by shock waves from the explosions of supernovae or from external gravitational influences, some stars escape but others remain gravitationally linked.
Galactic clusters gradually disperse, but their dynamical lifetimes can range into the hundreds of millions of years (examples--The Hyades and Praesepe (M44), thought to be in the range of 700 million years old; even the rather diffuse Coma Cluster (MEL 111)) or even the billions of years in exceptional sheltered settings (M67, NGC 188) if their trajectories take them sufficiently far from the galactic plane.
A hypothesis of this type would turn on the question of whether a galactic cluster of sufficient density to last 750 million years or more would be so dense that it would too rapidly and completely strip away an Oort Cloud, even if the Sun's orbit within the cluster was quite eccentric and it spent most of its time in the relatively sparse outer region of the cluster. But at least this is the operating premise.
The second line of evidence turns on the bombardment pattern of other planets relative to the Moon, particularly Mercury and Mars. Comparisons with Mars may be complicated by the unanswered question about the extent to which it had more water and a denser atmosphere than at present and whether such differences would influence the diameter and distribution of craters from impacts of a given size.
Venus is apparently resurfaced too frequently and other factors may complicate information about the outer Jovian moons Ganymede and Callisto and the influence of Jupiter on their bombardment patterns.
One line of reasoning was explained by Robert G. Storm in "The Geology of the Terrestrial Planets," (NASA SP-469) and Eugene Shoemaker in "The New Solar System" (Third Edition).
However, if the relative contribution of comets to asteroids was much greater during the Late Heavy Bombardment period, the bombardment rate on Mercury could have been considerably higher than it is now. In that case, Caloris would be more like the age of Imbrium or Orientale than that of Nectaris.
In addition, such a change in the rate of the two categories of impactors would allow for a later formation of the smooth plains on Mercury. Obviously, there is no way of knowing whether this is the case short of sending a landing craft to Mercury to test whether this is the case. Still, if the overall hypothesis described is plausible, such a test would be useful on a future launch to Mercury.
The third approach is through evidence for which I have no information. This requires an analysis of evidence regarding the distribution of craters, particularly on the Moon, as they relate to the angle and distribution of impact.
This approach starts with the assumption that the Moon became tidally locked to the Earth in its orbit less than 750 million years after the Solar System's formation.
If that was the case, and if the inclination of the orbit to the ecliptic of the Moon around the Earth has remained at about 5.5 degrees above or below the terrestrial orbit relative to the Solar equator, and if the inclination of the orbit of the Earth to the plane of the Solar equator has remained in the range of 2.5 degrees, the inclination of the Lunar orbit to the Sun should be no more than about eight degrees.
The second assumption is that most of the asteroids are in orbits that are relatively close to the plane of the solar equator. This would suggest that most of them, when disturbed, should collide with the Moon (and other terrestrial planets) in a fairly direct line with the equator. They may strike the limbs of the planetary bodies, but the shapes of the craters and the form of their rims should indicate collisions that tend to be most often distributed in roughly the plane of the equator.
On the other hand, most of the collisions with the polar regions will indicate an impact at an angle from or near the plane of the equator.
In short, there would be fewer right-angle impacts at the poles than at the equator.
On the other hand, the comets in the Oort cloud are theorized to be fairly well distributed in a sphere around the solar system. Thus, they are more or less equally likely to approach the Sun and planets at any angle. The distribution of right-angle to low-angle impacts of comets would be isotropic. This distribution would be expected to change if the primary source of impactors shifted from the Oort Cloud to asteroids.
Thus, most craters on the lunar poles would be low-angle during the Eratosthenian and Copernican periods but there would be a significantly higher proportion of right-angle impacts during the Pre-Nectarian and Nectarian and early Imbrian periods, with the ratio to low-angle impacts declining during the late Imbrian period.
This pattern might be less distinct on Mercury and Mars as the axes of these planets might have varied. Of course, this pattern also wouldn't hold if major impacts themselves caused major changes in the axis of the Moon.
The pattern might be fairly distinct on the outer Jovian satellites as they should have become tidally locked too Jupiter quite early. But the best data would probably be lunar.
This is a hypothesis I developed several years ago so the sources are old, but as I haven't seen any discussion, one way or the other on this hypothesis, I send it as an inquiry.
