Left: ratio of the expected number of detections with primary mass in the MG (defined here as 2.5–5 M_⊙) compared to the expected number of NS detections (defined here as 1–2.5 M_⊙). Right: ratio of the expected number of detections with an NS primary mass (in the mass range 1–2.5 M_⊙) compared to a BH primary mass (in the mass range 5 M_⊙–100 M_⊙). Here GWTC-1 contains zero events with m₁ in the MG, one BNS, and 10 BBHs, leading us to measure an mass-gap–NS ratio of ∼zero and an NS–BH ratio of ∼0.1; the 90% highest posterior density credible intervals on these values are shown by the gray shaded regions (see the text for more detail). The simple POWER-LAW model predicts at least as many mass gap detections as NS detections. Meanwhile, when we allow for both a dip and a break in the power law, we lower the expected fraction of mass-gap detections and raise the expected fraction of NS detections relative to BH detections, allowing us to better fit the observed number of detections in each bin.