The load-bearing assumption in stream-based dark matter work is that a tidal stream left undisturbed stays thin and smooth, so a gap or a kink implies something massive passed nearby [10]. The University of Washington run, published Aug. 27 in The Astrophysical Journal, inverts that experiment by removing the perturbers and looking at what remains [1][2]. Almost all of it emerged irregular anyway. Seventy streams out of about 15,000 came through five billion simulated years perfectly smooth [4], which is 0.47 percent [1]; the other 14,930, about 99.5 percent, acquired structure with no subhalo anywhere in the box [2].
The mechanism is unglamorous and, usefully, not random. The simulated disks had their stars spread unevenly, and streams crossing the denser patches were bent and torn by the lumpy potential [5]. Streams on orbits closer to the galactic core met those regions more often and came out more damaged [6]. The features the runs produced included wiggles, kinks, spurs, branches, gaps and clumps, and some streams did not survive as streams at all [7].
This work leaves open whether the gaps in real streams are baryonic in origin. There were no dark matter clumps in these simulations, so the work cannot weigh host perturbations against subhalo perturbations; it can only show that the null hypothesis behind that weighing was too clean. Lead author Arpit Arora says the host galaxies alone produced the same kinds of irregularities seen in real streams, and that predicting the host contribution is what makes isolating dark matter possible [8]. He plans a follow-up that puts the clumps back in to test whether their signatures are distinguishable [9]. That comparison will not be possible until the follow-up run is complete.
Two limits are worth holding onto. The sample is four host galaxies, not a statistical population. And "the same kinds of irregularities" is a morphological statement, not a matched distribution of gap depths, spacings or density power spectra, which is what a subhalo mass constraint actually consumes. The radius dependence looks like the more tractable handle: if host-induced damage rises toward the core [6], then feature frequency as a function of galactocentric radius becomes a population test rather than an argument about individual famous streams, and the Simonyi Survey Telescope at the NSF-DOE Vera C. Rubin Observatory is expected to supply that population, along with enough streams to build a taxonomy of their features [11]. Nora Shipp, a co-author, calls streams one of the sharpest tools inside the Milky Way laboratory [12]. Sharp tools still need their zero point measured, and this paper supplies a computed one where the field had only assumed it.
