We present a highly reliable flux-limited census of 18,949 point sources in the Galactic midplane that have intrinsically red mid-infrared colors. These sources were selected from the Spitzer Space Telescope Galactic Legacy Infrared Midplane Survey Extraordinaire (GLIMPSE) I and II surveys of 274 deg 2 of the Galactic midplane, and consist mostly of high-and intermediate-mass young stellar objects (YSOs) and asymptotic giant branch (AGB) stars. The selection criteria were carefully chosen to minimize the effects of positiondependent sensitivity, saturation, and confusion. The distribution of sources on the sky and their location in the Infrared Array Camera and the Multiband Image Photometer for Spitzer 24 μm color-magnitude and colorcolor space are presented. Using this large sample, we find that YSOs and AGB stars can be mostly separated by simple color-magnitude selection criteria into approximately 50%-70% of YSOs and 30%-50% of AGB stars. Planetary nebulae and background galaxies together represent at most 2%-3% of all the red sources. 1004 red sources in the GLIMPSE II region, mostly AGB stars with high mass-loss rates, show significant (0.3 mag) variability at 4.5 and/or 8.0 μm. With over 11,000 likely YSOs and over 7000 likely AGB stars, this is to date the largest uniform census of AGB stars and high-and intermediate-mass YSOs in the Milky Way Galaxy.

Throughout the Milky Way, molecular clouds typically appear filamentary, and mounting evidence indicates that this morphology plays an important role in star formation. What is not known is to what extent the dense filaments most closely associated with star formation are connected to the surrounding diffuse clouds up to arbitrarily large scales. How are these cradles of star formation linked to the Milky Way's spiral structure? Using archival Galactic plane survey data, we have used multiple datasets in search of large-scale, velocity-coherent filaments in the Galactic plane. In this paper, we present our methods employed to identify coherent filamentary structures first in extinction and confirmed using Galactic Ring Survey data. We present a sample of seven giant molecular filaments (GMFs) that have lengths on the order of ∼100 pc, total masses of 10 4 -10 5 M , and exhibit velocity coherence over their full length. The GMFs we study appear to be inter-arm clouds and may be the Milky Way analogs to spurs observed in nearby spiral galaxies. We find that between 2 and 12% of the total mass (above ∼10 20 cm −2 ) is "dense" (above 10 22 cm −2 ), where filaments near spiral arms in the Galactic midplane tend to have higher dense gas mass fractions than those further from the arms.

We present Atacama Large Millimeter/submillimeter Array (ALMA) line and continuum observations at 1.2 mm with ∼0.3 ′′ resolution that uncover a Keplerian-like disk around the forming O-type star AFGL 4176. The continuum emission from the disk at 1.21 mm (source mm1) has a deconvolved size of 870±110 AU × 330±300 AU and arises from a structure ∼8 M ⊙ in mass, calculated assuming a dust temperature of 190 K. The first-moment maps, pixel-to-pixel line modeling, assuming local thermodynamic equilibrium (LTE), and position-velocity diagrams of the CH 3 CN J=13-12 K-line emission all show a velocity gradient along the major axis of the source, coupled with an increase in velocity at small radii, consistent with Keplerian-like rotation. The LTE line modeling shows that where CH 3 CN J=13-12 is excited, the temperatures in the disk range from ∼70 to at least 300 K and that the H 2 column density peaks at 2.8×10 24 cm −2 . In addition, we present Atacama Pathfinder Experiment (APEX) 12 CO observations which show a large-scale outflow from AFGL 4176 perpendicular to the major axis of mm1, supporting the disk interpretation. Finally, we present a radiative transfer model of a Keplerian disk surrounding an O7 star, with a disk mass and radius of 12 M ⊙ and 2000 AU, that reproduces the line and continuum data, further supporting our conclusion that our observations have uncovered a Keplerian disk around an O-type star.

Context. The past decade has witnessed a large number of Galactic plane surveys at angular resolutions below 20 . However, no comparable high-resolution survey exists at long radio wavelengths around 21 cm in line and continuum emission. Aims. We remedy this situation by studying the northern Galactic plane at ∼20 resolution in emission of atomic, molecular, and ionized gas. Methods. Employing the Karl G. Jansky Very Large Array (VLA) in the C-array configuration and a large program, we observe the HI 21 cm line, four OH lines, nineteen Hnα radio recombination lines as well as the continuum emission from 1 to 2 GHz in full polarization over a large part of the first Galactic quadrant. Results. Covering Galactic longitudes from 14.5 to 67.4 deg and latitudes between ±1.25 deg, we image all of these lines and the continuum at ∼20 resolution. These data allow us to study the various components of the interstellar medium (ISM): from the atomic phase, traced by the HI line, to the molecular phase, observed by the OH transitions, to the ionized medium, revealed by the cm continuum and the Hnα radio recombination lines. Furthermore, the polarized continuum emission enables magnetic field studies. In this overview paper, we discuss the survey outline and present the first data release as well as early results from the different datasets. We now release the first half of the survey; the second half will follow later after the ongoing data processing has been completed. The data in fits format (continuum images and line data cubes) can be accessed through the project web-page. Conclusions. The HI/OH/Recombination line survey of the Milky Way (THOR) opens a new window to the different parts of the ISM. It enables detailed studies of molecular cloud formation, conversion of atomic to molecular gas, and feedback from Hii regions as well as the magnetic field in the Milky Way. It is highly complementary to other surveys of our Galaxy, and comparing the different datasets will allow us to address many open questions.

The surroundings of massive protostars constitute an accretion disc which has numerically been shown to be subject to fragmentation and responsible for luminous accretion-driven outbursts. Moreover, it is suspected to produce close binary companions which will later strongly influence the star's future evolution in the Hertzsprung-Russel diagram. We present threedimensional gravitation-radiation-hydrodynamic numerical simulations of 100 M pre-stellar cores. We find that accretion discs of young massive stars violently fragment without preventing the (highly variable) accretion of gaseous clumps onto the protostars. While acquiring the characteristics of a nascent low-mass companion, some disc fragments migrate onto the central massive protostar with dynamical properties showing that its final Keplerian orbit is close enough to constitute a close massive proto-binary system, having a young high-mass and a low-mass component. We conclude on the viability of the disc fragmentation channel for the formation of such short-period binaries, and that both processes -close massive binary formation and accretion bursts-may happen at the same time. FU-Orionis-type bursts, such as observed in the young high-mass star S255IR-NIRS3, may not only indicate ongoing disc fragmentation, but also be considered as a tracer for the formation of close massive binaries -progenitors of the subsequent massive spectroscopic binaries -once the high-mass component of the system will enter the main-sequence phase of its evolution. Finally, we investigate the ALMA-observability of the disc fragments.

We study the fragmentation of the nearest high line-mass filament, the integral shaped filament (ISF, line-mass ∼ 400 M ⊙ pc −1 ) in the Orion A molecular cloud. We have observed a 1.6 pc long section of the ISF with the Atacama Large Millimetre/submillimeter Array (ALMA) at 3 mm continuum emission, at a resolution of ∼3 ′′ (1 200 AU). We identify from the region 43 dense cores with masses about a solar mass. 60% of the ALMA cores are protostellar and 40% are starless. The nearest neighbour separations of the cores do not show a preferred fragmentation scale; the frequency of short separations increases down to 1 200 AU. We apply a twopoint correlation analysis on the dense core separations and show that the ALMA cores are significantly grouped at separations below ∼17 000 AU and strongly grouped below ∼6 000 AU. The protostellar and starless cores are grouped differently: only the starless cores group strongly below ∼6 000 AU. In addition, the spatial distribution of the cores indicates periodic grouping of the cores into groups of ∼30 000 AU in size, separated by ∼50 000 AU. The groups coincide with dust column density peaks detected by Herschel. These results show hierarchical, two-mode fragmentation in which the maternal filament periodically fragments into groups of dense cores. Critically, our results indicate that the fragmentation models for lower line-mass filaments (∼ 16 M ⊙ pc −1 ) fail to capture the observed properties of the ISF. We also find that the protostars identified with Spitzer and Herschel in the ISF are grouped at separations below ∼17 000 AU. In contrast, young stars with disks do not show significant grouping. This suggests that the grouping of dense cores is partially retained over the protostar lifetime, but not over the lifetime of stars with disks. This is in agreement with a scenario where protostars are ejected from the maternal filament by the slingshot mechanism, a model recently proposed for the ISF by Stutz & Gould. The separation distributions of the dense cores and protostars may also provide an evolutionary tracer of filament fragmentation.

We carried out a large program with the Karl G. Jansky Very Large Array (VLA): "THOR: The H , OH, Recombination line survey of the Milky Way". We observed a significant portion (∼100 deg 2 ) of the Galactic plane in the first quadrant of the Milky Way in the 21 cm H  line, 4 OH transitions, 19 radio recombination lines, and continuum from 1 to 2 GHz. In this paper we present a catalog of the continuum sources in the first half of the survey (l = 14.0−37.9• and l = 47.1−51.2 • , |b| ≤ 1.1 • ) at a spatial resolution of 10−25 , depending on the frequency and sky position with a spatially varying noise level of ∼0.3−1 mJy beam −1 . The catalog contains ∼4400 sources. Around 1200 of these are spatially resolved, and ∼1000 are possible artifacts, given their low signal-to-noise ratios. Since the spatial distribution of the unresolved objects is evenly distributed and not confined to the Galactic plane, most of them are extragalactic. Thanks to the broad bandwidth of the observations from 1 to 2 GHz, we are able to determine a reliable spectral index for ∼1800 sources. The spectral index distribution reveals a double-peaked profile with maxima at spectral indices of α ≈ −1 and α ≈ 0, corresponding to steep declining and flat spectra, respectively. This allows us to distinguish between thermal and non-thermal emission, which can be used to determine the nature of each source. We examine the spectral index of ∼300 known H  regions, for which we find thermal emission with spectral indices around α ≈ 0. In contrast, supernova remnants (SNR) show non-thermal emission with α ≈ −0.5 and extragalactic objects generally have a steeper spectral index of α ≈ −1. Using the spectral index information of the THOR survey, we investigate potential SNR candidates. We classify the radiation of four SNR candidates as non-thermal, and for the first time, we provide strong evidence for the SNR origin of these candidates.

Context. Filamentary structures in the interstellar medium are crucial ingredients of the star formation process. They fragment to form individual star-forming cores, and at the same time they may also funnel gas toward the central gas cores, providing an additional gas reservoir. Aims. We want to resolve the length scales for filament formation and fragmentation (resolution ≤0.1 pc), in particular the Jeans length and cylinder fragmentation scale. Methods. We have observed the prototypical high-mass star-forming filament IRDC 18223 with the Plateau de Bure Interferometer (PdBI) in the 3.2 mm continuum and N 2 H + (1-0) line emission in a ten-field mosaic at a spatial resolution of ∼4 (∼14 000 au). Results. The dust continuum emission resolves the filament into a chain of at least 12 relatively regularly spaced cores. The mean separation between cores is ∼0.40(±0.18) pc. While this is approximately consistent with the fragmentation of an infinite, isothermal, and gravitationally bound gas cylinder, a high mass-to-length ratio of M/l ≈ 1000 M pc −1 requires additional turbulent and/or magnetic support against radial collapse of the filament. The N 2 H + (1−0) data reveal a velocity gradient perpendicular to the main filament. Although rotation of the filament cannot be excluded, the data are also consistent with the main filament being comprised of several velocity-coherent subfilaments. Furthermore, this velocity gradient perpendicular to the filament resembles results toward Serpens south that are interpreted as signatures of filament formation within magnetized and turbulent sheet-like structures. Lower-density gas tracers ([CI] and C 18 O) reveal a similar red-and blueshifted velocity structure on scales around 60 east and west of the filament. This may tentatively be interpreted as a signature of the large-scale cloud and the smaller scale filament being kinematically coupled. We do not identify a velocity gradient along the axis of the filament. This may be due to no significant gas flows along the filamentary axis, but it may also be partly caused by a low inclination angle of the filament with respect to the plane of the sky minimizing such a signature. Conclusions. The IRDC 18223 3.2 mm continuum data are consistent with thermal fragmentation of a gravitationally bound and compressible gas cylinder. However, the high mass-to-length ratio requires additional support -most likely turbulence and/or magnetic fields -against collapse. The N 2 H + spectral line data indicate a kinematic origin of the filament, but we cannot conclusively differentiate whether it has formed out of (pre-existing) velocity-coherent subfilaments, whether magnetized converging gas flows, a larger-scale collapsing cloud, or even whether rotation played a significant role during filament formation.