The Outer Solar System Origins Survey (OSSOS), a wide-field imaging program in 2013-2017 with the CanadaFrance-Hawaii Telescope, surveyed 155 deg 2 of sky to depths of m r = 24.1-25.2. We present 838 outer solar system discoveries that are entirely free of ephemeris bias. This increases the inventory of trans-Neptunian objects (TNOs) with accurately known orbits by nearly 50%. Each minor planet has 20-60 Gaia/Pan-STARRS-calibrated astrometric measurements made over 2-5 oppositions, which allows accurate classification of their orbits within the trans-Neptunian dynamical populations. The populations orbiting in mean-motion resonance with Neptune are key to understanding Neptune's early migration. Our 313 resonant TNOs, including 132 plutinos, triple the available characterized sample and include new occupancy of distant resonances out to semimajor axis a ∼ 130 au. OSSOS doubles the known population of the nonresonant Kuiper Belt, providing 436 TNOs in this region, all with exceptionally high-quality orbits of a uncertainty σ a 0.1%; they show that the belt exists from a  37 au, with a lower perihelion bound of 35au. We confirm the presence of a concentrated low-inclination a ; 44 au "kernel" population and a dynamically cold population extending beyond the 2:1 resonance. We finely quantify the survey's observational biases. Our survey simulator provides a straightforward way to impose these biases on models of the trans-Neptunian orbit distributions, allowing statistical comparison to the discoveries. The OSSOS TNOs, unprecedented in their orbital precision for the size of the sample, are ideal for testing concepts of the history of giant planet migration in the solar system.

How black holes accrete surrounding matter is a fundamental yet unsolved question in astrophysics. It is generally believed that matter is absorbed into black holes via accretion disks, the state of which depends primarily on the mass-accretion rate. When this rate approaches the critical rate (the Eddington limit), thermal instability is supposed to occur in the inner disk, causing repetitive patterns of large-amplitude X-ray variability (oscillations) on timescales of minutes to hours. In fact, such oscillations have been observed only in sources with a high mass-accretion rate, such as GRS 1915+105 (refs 2, 3). These large-amplitude, relatively slow timescale, phenomena are thought to have physical origins distinct from those of X-ray or optical variations with small amplitudes and fast timescales (less than about 10 seconds) often observed in other black-hole binaries-for example, XTE J1118+480 (ref. 4) and GX 339-4 (ref. 5). Here we report an extensive multi-colour optical photometric data set of V404 Cygni, an X-ray transient source containing a black hole of nine solar masses (and a companion star) at a distance of 2.4 kiloparsecs (ref. 8). Our data show that optical oscillations on timescales of 100 seconds to 2.5 hours can occur at mass-accretion rates more than ten times lower than previously thought. This suggests that the accretion rate is not the critical parameter for inducing inner-disk instabilities. Instead, we propose that a long orbital period is a key condition for these large-amplitude oscillations, because the outer part of the large disk in binaries with long orbital periods will have surface densities too low to maintain sustained mass accretion to the inner part of the disk. The lack of sustained accretion--not the actual rate--would then be the critical factor causing large-amplitude oscillations in long-period systems.

Both physical and dynamical properties must be considered to constrain the origins of the dynamically excited distant Solar System populations. We present high-precision (g-r ) colors for 25 small (H r >5) dynamically excited Trans-Neptunian Objects (TNOs) and centaurs acquired as part of the Colours of the Outer Solar System Origins Survey (Col-OSSOS). We combine our dataset with previously published measurements and consider a set of 229 colors of outer Solar System objects on dynamically excited orbits. The overall color distribution is bimodal and can be decomposed into two distinct classes, termed 'gray' and 'red', that each has a normal color distribution. The two color classes have different inclination distributions: red objects have lower inclinations than the gray ones. This trend holds for all dynamically excited TNO populations. Even in the worst-case scenario, biases in the discovery surveys cannot account for this trend: it is intrinsic to the TNO population. Considering that TNOs are the precursors of centaurs, and that their inclinations are roughly preserved as they become centaurs, our finding solves the conundrum of centaurs being the only outer Solar System population identified so far to exhibit this property (Tegler et al. 2016). The different orbital distributions of the gray and red dynamically excited TNOs provide strong evidence that their colors are due to different formation locations in a disk of planetesimals with a compositional gradient.

Observations show that 100 km class Kuiper Belt objects (KBOs) can be divided into (at least) two color groups, red (R; g−i<1.2) and very red (VR; g−i>1.2), reflecting a difference in their surface composition. This is thought to imply that KBOs formed over a relatively wide range of radial distance, r. The cold classicals at 42au <r<47 au are predominantly VR, and known Neptune Trojans at r ; 30 au are mostly R. Intriguingly, however, the dynamically hot KBOs show a mix of R and VR colors and no correlation of color with r. Here we perform migration/instability simulations where the Kuiper Belt is populated from an extended planetesimal disk. We find that the color observations can be best understood if R objects formed at r<r * and VR objects at r>r * , with 30au <r * <40 au. The proposed transition at 30au <r * <40 au would explain why the VR objects in the dynamically hot population have smaller orbital inclinations than the R objects, because the orbital excitation from Neptune weakens for orbits starting beyond 30 au. Possible causes of the R-VR color bimodality are discussed.

We discuss the detection in the Outer Solar System Origins Survey (OSSOS) of two objects in Neptune's distant 9:1 mean motion resonance at semimajor axis a ≈ 130 au. Both objects are securely resonant on 10 Myr timescales, with one securely in the 9:1 resonance's leading asymmetric libration island and the other in either the symmetric or trailing asymmetric island. These objects are the largest semimajor axis objects with secure resonant classifications, and their detection in a carefully characterized survey allows for the first robust resonance population estimate beyond 100 au. The detection of these objects implies a 9:1 resonance population of 1.1 × 10 4 objects with H r < 8.66 (D 100 km) on similar orbits (95% confidence range of ∼ 0.4 − 3 × 10 4 ). Integrations over 4 Gyr of an ensemble of clones spanning these objects' orbit fit uncertainties reveal that they both have median resonance occupation timescales of ∼ 1 Gyr. These timescales are consistent with the hypothesis that these objects originate in the scattering population but became transiently stuck to Neptune's 9:1 resonance within the last ∼ 1 Gyr of solar system evolution. Based on simulations of a model of the current scattering population, we estimate the expected resonance sticking population in the 9:1 resonance to be 1000-4500 objects with H r < 8.66; this is marginally consistent with the OSSOS 9:1 population estimate. We conclude that resonance sticking is a plausible explanation for the observed 9:1 population, but we also discuss the possibility of a primordial 9:1 population, which would have interesting implications for the Kuiper belt's dynamical history.