Cosmic spherules, asteroidal collisions and the possibility of detecting changes in the solar constant

Summary. The Poynting–Robertson effect, applied to ‘iron’spherules in dated deep‐sea cores, should provide a test of whether or not cosmic spherules originate in asteroidal collisions. If D is the overall diameter of the spherule and δ its bulk density, the effect predicts that a number of straight sloping lines should appear in a plot of D δ against arrival time for spherules down a core. From a Pacific core, 223 iron spherules ( D ≥ 45 μm), in a variety of states, were examined to estimate δ. But, a decision was made to use only 100 spherules in a final D δ plot as δ had dubious accuracy for spherules with rusted or missing metal globules and for those with enclosed or burst bubbies. In this plot, two parallel sloping lines appear with some clarity. They are widely spaced in time and could be due to infrequent collisions amongst the few asteroids at ∼ 3.4 au . Possibly, there are more frequent collisions in the main belts, but lines showing these are not clearly revealed. During the examination (i.e. by cracking the spherules), 16 iron spherules of a new type were found. Instead of a metal globule, a black bead of glass is eccentrically enclosed in the oxide shell. The bead is a true non‐magnetic glass, ruby‐red in thin section. It is rich in oxides of Fe, Ni and Si. The metal globules from 59 spherules were analysed for Ni. This allowed the Ni percentage to be estimated for the original metal, prior to its partial oxidation. This Ni distribution is similar to the distribution in meteoritic iron. If an asteroidal origin is correct, it follows, from mass influx estimations, that an appreciable fraction of spherules must exist in the zodiacal cloud. Catastrophic asteroidal collision is suggested in order to produce a copious supply of melt‐ejecta by frictional heating. It is also presumed that meteoritic chondrules were large droplets formed in ancient catastrophic asteroidal collisions. To detect changes in the Sun's luminosity, very accurate plotting is required. It is shown that the errors of the method are too great to detect any possible changes during the glacial cycles. However, long‐period oscillations or a steady trend over ∼ 1 Myr are within range of the method for changes ∼ 3 per cent of the mean Sun.