# The Skeleton of the Milky Way

Abstract

Recently, Goodman et al. (2014) argued that the very long, very thin infrared dark cloud “Nessie” lies directly in the Galactic midplane and runs along the Scutum-Centaurus arm in position-position-velocity (p-p-v) space as traced by lower density $$\textrm{CO}$$ and higher density $$\mathrm{NH_3}$$ gas. Nessie was presented as the first “bone” of the Milky Way, an extraordinarily long, thin, high-contrast filament that can be used to map our Galaxy’s ”skeleton.“ Here, we present evidence for additional bones in the Milky Way Galaxy, arguing that Nessie is not a curiosity but one of several filaments that could potentially trace Galactic structure. Our ten bone candidates are all long, filamentary, mid-infrared extinction features which lie parallel to, and no more than twenty parsecs from, the physical Galactic mid-plane. We use $$\textrm{CO}$$, $$\mathrm{N_2H+}$$, $$\textrm{HCO}^{+}$$ and $$\mathrm{NH_3}$$ radial velocity data to establish the three-dimensional location of the candidates in p-p-v space. Of the ten candidates, six also: have a projected aspect ratio of $$\ge 50\colon1$$; run along, or extremely close to, the Scutum-Centaurus arm in p-p-v space; and exhibit no abrupt shifts in velocity. Evidence suggests that these candidates are marking the locations of significant spiral features, with the bone called filament 5 (”BC_18.88-0.09") being a close analog to Nessie in the Northern Sky. As molecular spectral-line and extinction maps cover more of the sky at increasing resolution and sensitivity, we seek to find more bones in future studies, ultimately to create a global-fit to the Galaxy’s spiral arms by piecing together individual skeletal features.

# Submitted

This Authorea pre-print has now been submitted to The Astrophysical Journal, and the draft is now available on astro-ph. Please see http://arxiv.org/abs/1506.08807 for the most up-to-date version! Thanks!

# Introduction

Many surprisingly fundamental questions remain about the structure of the Milky Way. For instance, does the Milky Way have two (Jackson 2008, Francis 2009, Dobbs 2012) or four (Reid 2009, Bobylev 2013, Urquhart 2013) major spiral arms? What is the precise location of these arms in position-position-velocity (p-p-v) space? And what is the nature of inter-arm structure—is it made of well-defined spurs or more web-like structures? An understanding of the Milky Way’s true structure has eluded us, largely due to the fact that we are embedded in the galaxy we are attempting to delineate.

Much of our current understanding of the Milky Way’s three-dimensional structure stems from radial velocity measurements of gas. Making use of the Milky Way’s rotation curve (McClure-Griffiths 2007, Reid 2014), we can translate line-of-sight velocities into distances, thereby constructing a gross three-dimensional model of our Galaxy. Thanks to a wealth of spectroscopic surveys, velocity-resolved observations are readily available for much of the Galaxy’s molecular, atomic, and ionized gas. Extended tracers, like CO (Dame et al., 2001) or HI (Shane, 1972) provide the best constraints on the Galaxy’s overall anatomy. To probe finer structure, observations of high-mass star forming regions can also provide kinematic and distance information for high-density gas. For instance, measurements of trigonometric parallaxes and proper motions of masers from the BeSSeL survey produce accurate locations for several spiral arm segments, along with their associated pitch angles (Reid et al., 2014). Likewise, the Bolocam Galactic Plane Survey (BGPS, Schlingman et al., 2011; Shirley et al., 2013; Ellsworth-Bowers et al., 2013; Ellsworth-Bowers et al., 2015), the Millimetre Astronomy Legacy Team 90 GHz Survey (MALT90, Foster et al., 2011; Jackson et al., 2013), the $$\textrm{H}_2\textrm{O}$$ Southern Galactic Plane Survey (HOPS, Purcell et al., 2012; Walsh et al., 2011), and ATLASGAL follow-up spectral line surveys (Beuther et al., 2012; Wienen et al., 2012) have produced hundreds of high-spectral resolution velocity measurements of the dense gas in molecular clouds. Analyses of extinction data from surveys like Pan-STARRS1 complement this emission line data and can also be used to create three-dimensional models of the Galaxy’s structure (Green et al., 2014; Schlafly et al., 2014), with high resolution on the plane of the Sky, but much coarser resolution along the line of sight.

While the tools available for probing the Milky Way’s internal structure are diverse, none has especially high three-dimensional resolution over wide areas, nor are they capable of probing the densest gas on Galactic scales. To address these issues, Goodman et al. (2014) recently suggested that extraordinarily elongated filamentary infrared dark clouds, termed “bones” could be used to determine the structure of the Milky Way’s spiral arms, their location within the Galaxy, and their appearance as viewed by an outside observer. Goodman et al. (2014) presented Nessie as the first “bone” of the Milky Way. They found that Nessie was at least $$3^\circ$$ ($$\sim 140$$ pc), and possibly as long as $$8 ^\circ$$ ($$\sim 450$$ pc) in length, while being less than 0.1 $$^\circ$$ ($$\sim 0.6$$ pc) wide. They also conclude that Nessie lies very near the geometric mid-plane of the Milky Way Galaxy, at the 3.1 kpc distance to the Scutum-Centaurus arm. An analysis of the radial velocities of $${\rm NH}_3$$ emission and CO emission confirms that Nessie runs along the Scutum-Centaurus arm in p-p-v space, suggesting it forms a dense spine of that arm in physical space as well (Goodman et al., 2014)

Until very recently, no simulations had the spatial resolution to predict filaments as narrow as the bones, if they were to exist. In 2014, numerical simulations from Smith et al. (2014), using the AREPO moving mesh code, revealed dense filaments, with aspect ratios and column densities similar to Nessie, forming within and parallel to the mean plane of a simulated spiral galaxy. A detailed analysis of Nessie’s properties, along with these new simulation results, suggests Nessie may be the first in a class of objects that could trace our Galaxy’s densest spiral features (Goodman et al., 2014). It is reassuring to recognize that Nessie should be the easiest object of its kind to find. Nessie is located in the closest major spiral arm to the the Sun, perpendicular to our line of sight, slightly offset from the Galactic center. This placement makes Nessie clearly visible against the bright background of the Galactic center, and more elongated than objects more distant or inclined to our line-of-sight.

In this paper, we use large-scale mid-infrared imaging of the Galactic plane to search for bone candidates near locations where currently-claimed spiral arms should lie on the Sky. It is critical to appreciate, as explained in detail in Goodman et al. (2014), that the Sun’s 25 pc elevation above the Galactic mid-plane gives viewers on Earth a (very-foreshortened!) top-down perspective view of the Galaxy’s structure, so that arms lie at predictable offsets (e.g. $$b=-0.4 ^\circ$$ for Nessie) from the IAU zero of Galactic Latitude. We predict and exploit these offsets in our search for bone candidates. We investigate the best candidates’ relationship to the Milky Way’s spiral structure using molecular line data to establish likely distances. When used in conjunction with the kinematic, parallax, and extinction measurements outlined above, these new bones have the potential to pin down the Milky Way’s galactic structure on the Sky to pc-scale resolution in regions the vicinity of the nearest bones.

# Methodology

## Visual Search

To search for more bones, we looked for them around where they are expected to lie in $$l,b,v$$ space, according to our current understanding of the Milky Way’s structure. We began by calculating the expected $$l,b$$ paths of Galactic arms using a log-spiral approximation as described in recent literature (Dame 2011, Vallée 2008) and assuming a 25 pc height above the plane for the Sun (see Goodman et al., 2014, and references therein). The predicted positions of the Galactic arms (Scutum-Centaurus, Carina-Sagittarius, Norma-Cygnus, and Perseus) were then overlain on Spitzer GLIMPSE/MIPSGAL (Benjamin et al., 2003; Churchwell et al., 2009) images in World Wide Telescope (WWT)—a tool that facilitates easy visualization of several layers of data at scales from the full sky down to the highest-resolution details. We restricted our initial search to the MIPSGAL footprint ($$|l|<62^\circ, |b|<1^\circ$$), with particular attention given to the region $$-30^\circ<l<51^\circ$$, as the Scutum-Centaurus arm (the closest major spiral arm from our vantage point) has tangent points at these longitudes. Panning along the full Spitzer/MIPSGAL Survey in WWT, we searched for largely continuous, filamentary, extinction features near and roughly parallel to the Galactic mid-plane, where all of the overlain arm traces lie. This visual inspection yielded about fifteen initial bone candidates. A video showing how this search worked in WWT is available on YouTube, and the original WWT Tour, of which the video shows a capture, is available at the Bones of the Milky Way Dataverse.

## Probing Velocity Structure

For features that appear associated with spiral arms on the 2-D plane of the sky, radial velocity data is needed to establish whether 3-D association with a spiral feature is likely. Any good bone candidate must have similar line-of-sight velocities along its full length (i.e. no abrupt shifts in velocity of more than 3 km/s per 10 pc along the bone), and more importantly, the measured radial velocities should be very close to those predicted by the Milky Way’s rotation curve for arms at a known distance. To probe the velocity structure of the initial bone candidates identified in WWT, we employed radial velocity data from five separate radio surveys: HOPS (Purcell et al., 2012; Walsh et al., 2011), MALT90 (Foster et al., 2011; Jackson et al., 2013), BGPS spectral-line follow-up (Schlingman 2011, Shirley 2013, Ellsworth-Bowers 2013), GRS (Jackson et al., 2006) and ThrUMMS (Barnes 2011). The HOPS, MALT90, and BGPS surveys are all geared towards probing dense regions hosting the early stages of high mass star formation. We utilize $$\textrm{NH}_3$$ emission from HOPS, $$\mathrm{N_2H^{+}}$$ from MALT90, and $$\textrm{HCO}^{+}$$ from BGPS. All three of these lines trace dense molecular gas ($$\sim 10^{4}\textrm{ cm}^{-3}$$), and are often found in dense, cool clouds with temperatures less than 100 K (Purcell et al., 2012; Shirley et al., 2013). As infrared dark clouds tend to harbor cool, high density clumps of gas which fuel the formation of massive stars, all three of these data sets contain spectra for hundreds of regions within the longitude range of the potential Galactic bones. To complement these high density gas tracers, we probe the puffier envelopes ($$\sim 10^{2}\textrm{ cm}^{-3}$$) surrounding these bones using high resolution $$^{13}\rm{CO}$$ data from the GRS and ThrUMMS survey.

We investigate the velocity structure of our filaments in two ways: first, whenever possible, we establish the velocity contiguity of our candidates by performing a slice extraction along each filamentary extinction feature in Glue, a visualization tool that facilitates the linking of data sets. We link spectral p-p-v cubes from the GRS and ThrUMMS survey with GLIMPSE-Spitzer mid-infrared images and obtain velocity as a function of position along a path that traces the entire extinction feature; for a demonstration of how this was done, see appendix section Glue Demo I. The results of the slice extraction along the path of one of our strongest bone candidates is shown in figure \ref{fig:filament5_slice}. We are able to establish velocity coherence for all candidates lying within the coverage range of the GRS survey ($$18^\circ < l < 56^\circ$$) and the ThrUMMS survey ($$300^\circ < l < 358^\circ$$). We also confirm that this GRS or ThrUMMS-determined velocity agrees with the velocity of dense gas tracers to within 5 km/s.

Next, we ensure that the candidates are contiguous in velocity space as traced mainly by high-density emission from the HOPS, MALT90, and BGPS surveys. In cases where HOPS, MALT90, and BGPS catalog data are not available along the extinction feature, we also extract spectra from GRS and MALT90 p-p-v cubes using the spectrum extractor tool in Glue. We once again link spectral p-p-v cubes from the GRS survey with GLIMPSE-Spitzer mid-infrared images and use the spectrum-extractor tool to obtain velocities along different regions of the extinction feature; a demonstration of the procedure used to extract velocities in Glue is shown in the appendix (see section Glue Demo II). Since CO traces lower density gas ($$\sim 10^2 \textrm{ cm}^{-3}$$) and $$\mathrm{N_2H+}$$, $$\textrm{HCO}^+$$, and $$\textrm{NH}_3$$ trace high density gas ($$\sim 10^4 \textrm{ cm}^{-3}$$), the dense gas sources provide much stronger evidence for the velocity of cold, dense, filamentary IRDCs. However, where dense gas sources are not available, the complete and unbiased high resolution GRS survey, although less desirable, allows us to roughly gauge the velocity along entire lengths of filaments.

By overlaying the HOPS, MALT90, BGPS, and GRS determined velocities on a p-v diagram of CO emission, we establish whether these filaments are associated with an existing spiral arm trace. For this study, we first use the whole-galaxy Dame et al. (2001) CO survey to roughly locate each of the arms in p-p-v space. We then overplot spiral fits to CO and HI for various spiral arm models, to determine whether candidates are consistent with previously claimed spiral arm traces. Of the approximately fifteen candidates identified visually, ten of these candidates are within 10 km/s of the Scutum-Centaurus and Norma-4kpc/Norma-Cygnus arms. We plot these ten candidates in p-p-v space in figure \ref{fig:skeleton}. In addition to showing our Bone candidates, we show several different predictions of the positions of two spiral arms toward the inner Galaxy in longitude-velocity space, from Dame et al. (2011), Sanna et al. (2014), Shane (1972), and Vallée (2008). We also include Scutum-Centaurus and Norma fits from Reid et al. (2015, in prep), derived from trigonometric parallax measurements of high-mass star forming regions taken as part of the BeSSeL survey (Reid et al., 2014). Reid et al. (2015, in prep) produce fits with ($$l,b,v$$) loci that closely follow GMCs that trace the arms, producing a rough log-spiral approximation determined by trigonometric parallax rather than an assumed Galactic rotation curve.

## Establishing “Bone” Criteria

After narrowing down our list to ten filaments with kinematic structure consistent with existing spiral arm models, we develop a set of criteria for objects to be called “bones:”

1. Visually continuous mid-infrared extinction feature

2. Parallel to the Galactic plane, to within $$30^\circ$$

3. Within 20 pc of the physical Galactic mid-plane, assuming a flat galaxy

4. Within 10 km/s of the global-log spiral fit to any Milky Way arm

5. No abrupt shifts in velocity (of more than  3 km/s per 10 pc) within extinction feature

6. Projected aspect ratio $$\ge 50:1$$

The name and coordinates for these ten filaments, along with their average LSR velocities, the number of bone criteria they satisfy, and a “quality rating” are listed in figure \ref{fig:candidates}. In figure \ref{fig:mass_of_bones}, we summarize physical parameters for all ten bone candidates, including estimates of distance, volume, mass, and aspect ratio. We calculate mass by estimating an average H$$_2$$ column density of $$2 \times 10^{22}$$ cm$$^{-2}$$, consistent with the minimum IRDC peak column density to be included in the Peretto et al. (2009) catalog of 11,303 IRDCs. We calculate distances assuming all of our bone candidates (see Figure \ref{fig:skeleton}) are associated with the Dame et al. (2011) Scutum-Centaurus arm. When available, we also provide distance measurements from the Ellsworth-Bowers et al. (2013) catalog, which provides distances to 1710 molecular clouds from the BGPS survey, derived using a Bayesian distance probability density function.

Of the ten filaments with velocities consistent with galactic rotation, six of these meet all six bone criteria: filament 1 (“BC_026.94-0.30”), filament 2 (“BC_025.24-0.45”), filament 5 (“BC_018.88-0.09”), filament 7 (“BC_004.14-0.02”), filament 9 (“BC_335.31-0.29”), and filament 10 (“BC_332.21-0.04”), to varying degrees of excellence. We note that filament 10 has likely been disrupted by stellar feedback, making its aspect ratio and velocity structure more difficult to define. Since we predict that all galactic bones will likely be destroyed by stellar feedback and/or galactic shear, we include it here as part of a larger attempt to build a catalog of bones at all stages of their evolution. We also include filament 9, even though it has a $$p-v$$ orientation perpendicular to predicted fits of the Scutum-Centaurus arm (see Figure \ref{fig:skeleton}). As spurs and inter-arm structures are likely to lie close to the physical Galactic mid-plane, but with velocity gradients angled with respect to predicted arm fits, we do not require that a bone be parallel to arm $$p-v$$ traces, so as not to exclude potential spurs, feathers, or other inter-arm features.

Of the four remaining filaments that do not meet all six criteria—filament 8 (“BC_357.62-0.33”), filament 6 (“BC_011.13-0.12”), filament 3 (“BC_024.96-0.17”), and filament 4 (“BC_021.25-0.15”)—all of them fail criterion 6 (aspect ratio $$\ge 50:1$$). As our criterion 6 does not allow for projection effects in imposing an aspect ratio limit, we emphasize that those filaments lying more tangential to our line-of-sight will appear foreshortened, and could very well meet the 50:1 minimum limit if projection effects were removed. We plan to examine the the aspect ratios of all our candidates in a follow-up study, accounting for expected projection effects if they lie along the nearest spiral arm. The first of these candidates, filament 8, shows particular promise, lying within 2-3 pc of the physical Galactic midplane and tracing a prominent peak of CO emission in both p-p and p-v space (see appendix section on filament 8). The second filament, filament 6 (“the snake”), has already been well-studied from a star formation perspective, hosting over a dozen protostellar cores likely to produce regions of high-mass star formation (Wang et al., 2014; Henning et al., 2010). From a Galactic bone perspective, the snake strongly satisfies all criteria except number 6—it lies within 15 pc of the physical galactic midplane and 5 km/s from the Dame et al. (2011) Scutum-Centaurus global-log fit to CO, also tracing a prominent peak of CO emission in p-v space (see appendix section on filament 6). The remaining two filaments, filaments 3 and 4, are both awarded a quality rating of “C.” Filament 3 lies 10 km/s from the Shane (1972) fit to HI for the Scutum-Centaurus arm (at the upper limit of criterion 4) while filament 4 appears to be a potential interarm filament, lying between the Scutum-Centaurus and Norma-4kpc arms in p-v space. We also note a small break in the extinction feature of filament 4, though the filament has been confirmed to be contiguous in velocity space as traced by $$^{13}\rm{CO}$$ from the GRS Survey.

In summary, it is important to emphasize that some of the above criteria will likely be modified in the long run, as we learn more about the Skeleton of the Milky Way. Given our limited a priori knowledge of the Galaxy’s structure, it is presently easier to confirm Bones that are spine-like, lying along arms with velocities predicted by extant modeling (criteria 1, 5), and harder to find spurs off those arms or inter-arm features (e.g. filament 9, the velocities of which are hard to predict well. Similarly, since criterion 6 does not allow for projection effects in imposing an aspect ratio limit, bones which otherwise meet all criteria could fail if they lie close to the tangents of spiral arms. As we learn more about spiral structure from simulations and modeling, these criteria will also be adjusted to allow for bone-like features that represent spurs, inter-arm structures, and/or foreshortened structures lying close to our line of sight.

\label{fig:skeleton}: A position-velocity space summary of Bone candidates and spiral arm models. Blue background shows $$^{12}\rm{CO}$$ emission from Dame et al. (2001), integrated between $$-1^\circ < b < 1^\circ$$. Black dots show measurements of BGPS, HOPS, MALT90, and GRS-determined velocities, with particular candidate filaments identified by numbers (see table \ref{fig:candidates} for further identification), or, in the case of Nessie, by name. Lines of varying color show predicted p-v spiral arm traces from the literature (see text for references).

\label{fig:candidates}: (1) Central galactic coordinates for our filaments, prefixed with “BC” (bone candidate) and ordered by Galactic longitude. (2) Name by which each bone candidate is referred to throughout this paper. (3) Average VLSR of the bone candidate, computed by averaging the velocities of the sources for each filament seen in figure \ref{fig:skeleton}. (4) Number of bone criteria satisfied (see section 2.3). (5) We assign a quality rating to each bone candidate dictated by how strongly they satisfy the bone criteria; a score of “A” is given if the candidate strongly or moderately satisfies all bone criteria. A score of “B” is given if the candidate strongly or moderately satisfies five criteria, but weakly satisfies (or fails to satisfy) one criterion. A score of “C” is given if the candidate strongly or moderately satisfies four criteria, but weakly satisfies (or fails to satisfy) two or more criteria.

\label{fig:mass_of_bones}: A comparison of the physical properties of the Bone candidates, based on a similar table from Goodman et al. (2014). We estimate an average H$$_2$$ column density of $$2 \times 10^{22} \rm{cm}^{-2}$$, consistent with the minimum peak column density to be included in the Peretto et al. (2009) catalog of 11,303 IRDCs; this corresponds to an assumed equivalent Av. of 20 magnitudes and an assumed average density of $$3\times 10^{-19}$$ $$\rm{g}/\rm{cm}^{3}$$. We plan to perform extinction mapping on each candidate to physically measure this column densities in a follow-up study. (1) Central galactic coordinates for our filaments, prefixed with “BC” (bone candidate) and ordered by Galactic longitude. (2) Name by which each bone candidate is referred to throughout this paper. (3) We estimate distances by assuming all candidates are associated with the Scutum-Centaurus arm as fitted by Dame et al. (2011). (4) When available, we cite distances from Ellsworth-Bowers et al. (2013), which were derived by applying a Bayesian distance probability density function to 1710 molecular clouds from the BGPS survey. (5) Length in parsecs is calculated assuming the Scutum-Centaurus arm distance from column 3. (6) Radius in parsecs is calculated assuming the Scutum-Centaurus arm distance from column 3. (9) We assume the filaments are cylindrically shaped, and calculate the volume based on measured radius and length. (11) Aspect ratio does not account for projection effects.

\label{fig:filament5_slice} We show the results of performing a slice extraction along the filamentary extinction feature of our strongest bone candidate, filament 5. Using Glue, we link spectral p-p-v cubes from the GRS survey with GLIMPSE-Spitzer $$8\mu\rm{m}$$ mid-infrared images and extract velocities along a path tracing the entire extinction feature (red outline in top panel). We confirm that filament 5 is velocity contiguous as traced by lower density CO gas from the GRS survey (red boxed region in lower p-v panel). We also confirm that this GRS-determined velocity agrees with the velocity of dense gas tracers to within 5 km/s.

# Analysis of New Bones

Filament 5 is our strongest bone candidate, in that it is highly elongated ($$0.7^\circ$$ or 45 pc, with an aspect ratio of 140:1) and exactly along a previously-claimed spiral arm trace in p-p-v space, although its orientation makes it less somewhat elongated than Nessie on the sky. In figure \ref{fig:Candid5_pos_vel} we show a p-v diagram in the longitude range of filament 5 and overlay fits to the Scutum-Centaurus arm from Shane (1972), Vallée (2008), Dame et al. (2011), and Reid et al. (2015, in prep). We see that the HOPS, BGPS, and GRS-determined velocities associated with filament 5 are highly correlated with the Dame et al. (2011) and the Reid et al. (2015, in prep) global-log fits to CO and HI, suggesting that filament 5 is marking a “spine” of the Scutum-Centaurus arm in this longitude range. Moreover, filament 5 also lies along a CO peak in longitude-latitude space, as evident in figure \ref{fig:Candid5_pos_pos}. By overlaying a trace of the mid-IR extinction feature of filament 5 on a plane of the sky map (integrated in Scutum-Centaurus’s velocity range in the region around filament 5) we see that filament 5 lies in the center of the most intense CO emission. Finally, figure \ref{fig:Candid5_with_tilt} shows that filament 5 lies within $$\approx$$ 10 pc of the true physical mid-plane. All these figures taken together indicate that filament 5 is Nessie’s counterpart in the first quadrant, suggesting that Nessie is not a curiosity, but one of several bones that trace significant spiral features.

Our study is not the first follow up to the Nessie work in Goodman et al. (2014) to look for more long filaments associated with spiral structure. Ragan et al. (2014) and Wang et al. (2015) have undertaken similar studies. However, ours is the first study to specifically look for bones in regions we are most likely to find them, that is, elongated along the Galactic plane. Moreover, ours is the only study to create a quantitative set of criteria capable of defining this new class of objects (i.e. galactic “bones”).

Ragan et al. (2014) undertook a blind search (not restricted to latitudes where the mid-plane should lie) for long thin filaments (> $$1^\circ$$) in the first quadrant of the Milky Way, using near and mid-infrared images. In addition to confirming that Nessie lies along the Scutum-Centaurus arm, Ragan et al. (2014) find seven Giant Molecular Filaments (GMFs) of which only one, GMF 20.0-17.9, is said to be associated with Galactic structure (declared a spur of the Scutum-Centaurus arm). Our strongest bone candidate, filament 5, is a subsection of GMF 20.0-17.9, but, unlike Ragan et al. (2014), we argue that filament 5 runs right down the spine of the Scutum-Centaurus arm in p-v space. We believe the discrepancy arises due to a difference in methodology. Ragan et al. (2014) group neighboring IRDCs into a single filament, despite breaks in the extinction feature and kinks in velocity structure. Since grouping several IRDCs to make a longer structure violates our criteria 1 and 6, we only consider the continuous and kinematically coherent part of the filament, which is remarkably parallel to the Scutum-Centaurus arm in p-v space. Likewise, in figure 4 from Ragan et al. (2014) (analogous to our figure \ref{fig:skeleton}), they represent filaments as straight lines connecting velocities measured at the tips of the filaments while we represent filaments as sets of points whose velocities are determined by the BGPS, HOPS, MALT90, and GRS surveys. We compare our p-p and p-v analysis of filament 5 with the analysis from Ragan et al. (2014) in figure \ref{fig:ragan_comp}.

Ragan et al. (2014) find little or no association with their GMFs and Galactic structure, suggesting that we are perhaps not as sensitive to spiral arm filaments in the first quadrant, or that the frequency and orientation of spiral arm filaments in the first quadrant is different than the fourth. Our three bone candidates with an “A” quality rating (filaments 1,2, and 5) all lie in the first quadrant, so we speculate that Galactic bones are not subject to the same fourth quadrant bias that GMFs are potentially prone to. We also emphasize that the GMFs from Ragan et al. (2014) and our Galactic bones should be classified as fundamentally different objects. The lengths of the GMFs range from 60-230 pc, while our longest bone candidates is only 52 pc. By definition, GMFs are meant to be larger structures composed of several smaller, high-contrast elements, so no bone in itself will realistically be classified as a GMF. As is the case with filament 5, we expect that there will be significant overlap between the GMFs and Bone catalogs in the future, as our Galactic bones should be a subset of any spiral tracing GMFs yet to be discovered.

Like Ragan et al. (2014), Wang et al. (2015) search for large-scale filaments and establish their relationship to Galactic structure after the fact. Rather than searching for filaments elongated along the Galactic plane, Wang et al. (2015) search for the longest, coldest, and densest filaments (aspect ratio >>10) in Hi-GAL images, within the longitude range of $$15^\circ < l < 56^\circ$$. Filaments were initially identified using Hi-GAL 350 and 500 $$\mu\rm{m}$$ emission. Temperature and column density maps were created for each candidate, and those which exhibited systematically lower temperatures with respect to the background were selected. As in our study, Wang et al. (2015) confirmed velocity contiguity by extracting a p-v slice along the curvature of each filament.

Wang et al. (2015) highlight nine filaments as their most prominent, with one of the nine being Nessie. Only one of their filaments (Filament 6, or BC_011.13-0.12, the “snake”) overlaps with our sample, and is not classified as a bone due to its short aspect ratio ($$\approx 25:1$$). Seven other Wang et al. (2015) filaments fail one or more of our bone criteria. G24, G26, and G47 lie between 39-62 pc above the physical Galactic midplane (outside our $$\pm 20$$ pc criterion), while G28 and G65 have aspect ratios of around 19:1 and 38:1 (less than our 50:1 minimum aspect ratio criterion). Additionally, G29 and G49 are not largely continuous mid-infrared extinction features, violating our criterion 1. This is not surprising, as the Wang et al. (2015) study was designed to identify filaments emitting at longer Hi-GAL wavelengths, which are not necessarily seen continuously in absorption at mid-infrared wavelengths. In future studies, Wang et al. (2015) plan to extend their search to the entire Galactic plane. Despite differences in methodology, there should be some degree of overlap between the Wang et al. (2015) catalog and our bones catalog, and these, along with the Ragan et al. (2014) catalog, are expected to be highly complementary of each other.

An in-depth analysis of the other nine bone candidates can be found in the appendix.

\label{fig:Candid5_pos_vel}: Position-velocity diagram of $$\textrm{CO}$$, $$\textrm{NH}_3$$, and $$\textrm{HCO}^+$$ emission for filament 5. Blue background for the main panel shows $$^{12}$$CO (1-0) emission integrated between $$-1^\circ < \textrm{b} < 1 ^\circ$$ (Dame et al., 2001), while the purple sub panel shows higher resolution GRS $$^{13}$$CO (1-0) emission integrated over the same latitude range. Black dots show HOPS (NH$$_3$$ emission), BGPS (HCO$$^+$$ emission) and GRS (high resolution CO emission) sources associated with filament 5. The colored lines show fits to the Scutum-Centaurus arm (see text for references).

\label{fig:Candid5_pos_pos}: Plane of the sky map integrated between 35-55 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 5. A trace of filament 5, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map.The black dashed line indicates the location of the physical galactic midplane, while the solid black lines indicate ± 20 pc from the galactic midplane at the 3.7 kpc distance to filament 5, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The correlation between filament 5’s mid-IR extinction feature and Scutum-Centaurus’s most intense CO emission hints that it may be a spine of Scutum-Centaurus as traced by lower density CO gas.

\label{fig:Candid5_with_tilt}: Filament 5 lies within $$\approx$$ 10 pc of the physical galactic midplane. The background is a GLIMPSE Spitzer 8 $$\mu\textrm{m}$$ image. The dashed line is color-coded by Dame et al. (2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.7 kpc distance to filament 5, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares, triangles, and circles correspond to HOPS, BGPS, and GRS sources, respectively. The label “with tilt” indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 5 can be seen in the inset.

\label{fig:ragan_comp}: Top: Position-position analysis of filament 5 (this paper), as it compares to the (larger) feature known as GMF20.0-17.9 (Ragan et al., 2014). In green, we overlay the GRS $$^{13}\rm{CO}$$ integrated intensity contours that Ragan et al. (2014) uses to define GMF20.0-17.9; they group several neighboring IRDCs into a single filament. In the boxed yellow region, we show our filament 5, a subset of GMF20.0-17.9. In red, we show a path that connects the Ragan et al. (2014) IRDCs and traces BC_18.88-0.09. Bottom: Position-velocity analysis of filament 5, as it compares to GMF20.0-17.9. We show the results of performing a slice extraction along the red path in the upper panel. Using Glue, we link spectral p-p-v cubes from the GRS survey with GLIMPSE-Spitzer $$8\mu\rm{m}$$ images and extract velocity as a function of position along the path. As seen inside the yellow boxed region in the lower panel, the section of the path that corresponds to filament is remarkably kinematically contiguous, with velocities ranging between 45 and 49 km/s. In contrast, Ragan et al. (2014) group the 37 km/s emission at x=0 pc with the 50 km/s emission at x=115 pc and connect these two points with a straight line on a longitude-velocity diagram (green line in lower p-v panel).

# Discussion

Our initial search was intentionally limited: we specifically searched the Galaxy for the most prominent bones, near agreed-upon major spiral features, in a narrow longitude range (|$$l$$|<62$$^{\circ}$$). Through a more comprehensive search, there are potentially hundreds of bone-like filaments discoverable in the Milky Way. If we can find enough bones, it should be possible to piece them together to delineate the major structural features of our Galaxy, using other relevant information as well (from maser measurements, 3D extinction mapping, and kinemeatic constraints on both gas and stars). Astronomers have been trying to accurately model the location of the spiral arms in position-velocity space for decades. A bones-based approach should be able to resolve some of the discrepancies amongst the many arm models shown in figure \ref{fig:skeleton}. The level of disagreement on arm locations is large enough that finding even just a handful of Galactic bones marking sections of the spines of spiral arms will tie arm fits down with high fidelity at particular positions in p-p-v space. These “spinal” anchors will have especially large weights in statistical fits that seek to combine many measures of the Milky Way’s skeletal structure.

In the future, we plan to test and apply algorithms that “connect the dots” between markers of high density IRDC peaks, in a search for more skeletal features. Lenfestey, Fuller and Peretto (2015, in prep) have recently undertaken such a study, utilizing an IRDC catalog of $$\approx$$ 11,000 high density peaks from Peretto et al. (2009). Lenfestey et al. (2014) have grouped these catalog objects into long filamentary structures, using a Minimum Spanning Tree (MST) algorithm, identifying 100 structures in the region $$|l| < 65 ^{\circ}, |b|<1^{\circ}$$. Of these structures, 22 are linear features similar to the Nessie cloud. We plan to investigate the Lenfestey et al. (2014) filaments, as well as apply MST and related structure-finding procedures to additional surveys (e.g. ATLASGAL or Hi-GAL, Csengeri et al., 2014; Molinari et al., 2010), applying our initial bone criteria to all candidates found, thereby producing a larger population of bones capable of pinning down galactic structure.

Along with increasing our bone population, we plan to improve simulations in hopes of answering key questions about bones’ origin and evolution. For instance, synthetic observations of simulations should be able to tell us what fraction of highly-elongated dense clouds appear to be: a) aligned with arms; b) spur-like; c) inter-arm; or d) random long thin clouds unaligned with Galactic structure. And, the simulations should shed light on the likely origins of these types of objects, in part by predicting different velocity, density, or mass profiles for objects with different origins.

Not all long skinny filaments are expected to be associated with Galactic structure. Studies prior to Ragan et al. (2014) offer at least two examples of long molecular clouds that are not obviously Bone-like. The “Massive Molecular Filament” G32.02+0.06, studied by Battersby et al. (2014), does not appear to be tracing an arm structure. Likewise, the 500-pc long molecular “wisp” discussed by Li et al. (2013) also does not presently appear directly related to Galactic structure. Neither of these two clouds currently lies in any special position in p-p-v space. It is possible that these are bone remnants, disrupted by feedback or Galactic shear, but, without better Galaxy modeling, it is very hard to speculate on what fractions of long thin clouds were formerly bones, are currently bones, or were never bones.

While the Smith et al. (2014) Galaxy models are the first that provide high enough resolution to simulate our incredibly long and thin bones, they do not include stellar feedback or magnetic fields—either of which could cause disruptions in the appearance of the currently-simulated bone-like features. In the future, we hope to utilize more comprehensive, targeted high-resolution synthetic observations (e.g. of dust absorption and emission and of CO spectra), based on high-resolution simulations like the ones in Smith et al. (2014). Finally, we hope to use simulations to estimate the biases inherent in our selection criteria (how many spurious “Bones” should we expect to find randomly, by the chance alignment of discontinuous IRDC peaks?).

Though challenging, we plan to combine future high resolution synthetic observations with a wealth of existing data sets to build a skeletal model of the Milky Way. When used in conjunction with BeSSeL maser-based rotation curves (Reid et al., 2014), CO (Dame et al., 2001) and HI (Shane, 1972) p-v fitting, 3D extinction mapping (Schlafly, 2014b), HII region arm mapping (Anderson et al., 2012), and GAIA results, bones have the potential to not only redefine Galactic structure at unprecedented resolution, but also to resolve fundamental questions that have been plaguing Galactic astronomers for decades.

# Conclusion

We undertake a search for Galactic bones in the region $$|l|<62^\circ, |b|<1^\circ$$. We use large-scale GLIMPSE Spitzer images of the Galactic plane to search for skinny, largely continuous infrared dark clouds elongated along predicted spiral arms on a 2D plane of the sky map. We visually identify ten bone candidates that lie parallel to, and no more than twenty parsecs from, the physical Galactic midplane (assuming a flat Galaxy). We use radial velocity measurements derived from high and low density gas emission to establish velocity coherence and place the candidates in p-v space. Of the ten candidates, six are also contiguous in velocity space, lie within 10 km/s of the global log-fit to CO or HI for the Scutum-Centaurus and Norma-4kpc arms, and possess an aspect ratio of at least 50:1. The other four candidates only fail the minimum aspect ratio criterion, and could be reclassified as we refine this criterion to account for projection effects.

Our strongest candidate, filament 5 (“BC_18.88-0.09”) runs remarkably parallel to the physical Galactic midplane and lies just 10-15 pc above that plane. It also exhibits remarkable velocity contiguity and runs exactly along the Dame et al. (2011) fit to the Scutum-Centaurus arm in p-v space. Filament 5 also possesses an aspect ratio of at least 140:1, suggesting that it has formed as the result of a larger global spiral potential rather than the localized collapse of a giant molecular cloud. Cumulative evidence suggests that filament 5 and our other classified Galactic bones mark the location of significant spiral features and can be used to pin down the accuracy of spiral arm models to within one pc in regions near bones. In future we plan to follow-up on our ten bone candidates—performing extinction mapping and obtaining high resolution spectra with a suite of dense gas tracers. Ultimately, we plan to build a skeletal model of the Milky Way, accumulating, classifying, and synthesizing hundreds of galactic bones en route to mapping the spiral structure of our Galaxy in unparalleled detail.

# Acknowledments

This work is supported in part by the National Science Foundation REU and Department of Defense ASSURE programs under NSF Grant no. 1262851 and by the Smithsonian Institution

This publication makes use of molecular line data from the Boston University-FCRAO Galactic Ring Survey (GRS). The GRS is a joint project of Boston University and Five College Radio Astronomy Observatory, funded by the National Science Foundation under grants AST-9800334, AST-0098562, and AST-0100793. This publication also makes use of data from the H2O Southern Galactic Plane Survey and the The Millimetre Astronomy Legacy Team 90 GHz Survey.

This work is based [in part] on observations made with the Spitzer Space Telescope, which is operated by the Jet Propulsion Laboratory, California Institute of Technology under a contract with NASA.

The glue software is developed by the Seamless Astronomy group at the Center for Astrophysics, under contract to NASA’s James Webb Space Telescope program. WorldWide Telescope has been developed and supported by Microsoft Research, in part through gifts to the Seamless Astronomy program, and it will become an open-source resource available on GitHub in 2015.

# Glue Demo I

A demonstration of how a p-v slice was extracted along a path tracing the extinction feature, using the software visualization package Glue. First, using the “link data” function, we linked the Galactic latitude and longitude from the Spitzer image (upper left image) with the Galactic latitude and longitude from the GRS or ThrUMMS fits cubes (upper right image). Next, yellow circular regions were created along the extinction feature in the Spitzer image, marking the path along which the slice will be extracted. Since the data products are “linked”, Glue automatically overlays the same yellow regions on the GRS or ThrUMMS data cube (upper right image). Then, using the “slice extractor” tool, we trace a path through the centroids of the yellow regions in the CO fits cube, effectively creating a customized p-v slice along the extinction feature (bottom image).

# Glue Demo II

A demonstration of how radial velocities were extracted at specific Galactic coordinates along the filament, using the software visualization package Glue. First, using the “link data” function, we linked the Galactic latitude and longitude from the Spitzer image (upper left image) with the Galactic latitude and longitude from the GRS or MALT90 fits cubes (upper right image). Then, points were selected along the extinction feature (blues circles in upper left image) and which are then automatically overlaid on the $$^{13}\rm{CO}$$ (GRS) or $$\textrm{N}_2\textrm{H}^{+}$$ (MALT90) fits cube (blue points in upper right image). A spectrum (bottom image) was extracted from the the small red boxed region around the the blue point (top right image), using the “spectrum extractor” function in Glue. Then, a Gaussian was fitted around the highest peak in intensity.

# Filament 1 (“BC_26.94-0.30”)

Filament 1 is a confirmed bone, meeting all six criteria with a quality grade of “A.” With a length of 13 pc and a width of 0.12 parsecs, it is the shortest, thinnest, and least massive of our ten candidates, with an aspect ratio of approximately $$53:1$$. It runs exactly along the Reid et al. (2015, in prep) p-v fit to the Scutum-Centaurus arm and lies within 10 pc of the physical Galactic midplane.

We show the results of performing a slice extraction along the filamentary extinction feature of filament 1, which corresponds to the black line in the plane of the sky map below. Using Glue, we link spectral p-p-v cubes from the GRS survey with GLIMPSE-Spitzer mid-infrared images and extract velocities along a path tracing the entire extinction feature. We confirm that filament 1 is contiguous in velocity space as traced by lower density CO gas from the GRS survey.

Position-velocity diagram of CO, NH$$_{3}$$, and HCO$$^+$$ emission for filament 1. Blue background shows $$^{12}$$CO (1-0) emission integrated between $$-1^\circ < \textrm{b} < 1^\circ$$ (Dame et al., 2001). Black dots show GRS sources associated with filament 1. The black lines are fits to CO data for the Scutum-Centaurus arm (Sanna 2014, Dame 2011, Shane 1972)

Plane of the sky map integrated between 55-75 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 1. The black dashed line indicates the location of the physical galactic midplane, while the solid black lines indicate $$\pm$$ 20 pc from the galactic midplane at the 4.5 kpc distance to filament 1, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. A trace of filament 1, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map. The correlation between filament 1’s mid-IR extinction feature and the Scutum-Centaurus’s arm most intense CO emission hints that it may be a spine of Scutum-Centaurus as traced by lower density CO gas.

Filament 1 lies within $$\approx$$ 10 pc of the physical galactic midplane. The background is a GLIMPSE Spitzer 8 $$\mu\textrm{m}$$ image. The dashed line is color-coded by Dame et al. (2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 4.5 kpc distance to filament 1, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares, triangles, and circles correspond to HOPS, BGPS, and GRS sources, respectively. The label “with tilt” indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 1 can be seen in the inset.

# Filament 2 (“BC_025.24-0.45”)

Filament 2 is a confirmed bone, meeting all six criteria with a quality grade of “A.” Behind filament 5 (our strongest bone), it has the second longest aspect ratio of our ten candidates, at $$126:1$$. It lies about 2-3 km/s from the Shane (1972) p-v fit to the Scutum-Centaurus arm, as well as 20 pc from the physical Galactic midplane. When overlaid on a plane of the sky map (integrated in the velocity range of the Scutum-Centaurus in the region around filament 2), it traces a prominent peak in CO emission, forming a spine of Scutum-Centaurus as traced by lower density gas.

We show the results of performing a slice extraction along the filamentary extinction feature of filament 2, which corresponds to the black line in the plane of the sky map below. Using Glue, we link spectral p-p-v cubes from the GRS survey with GLIMPSE-Spitzer mid-infrared images and extract velocities along a path tracing the entire extinction feature. We confirm that filament 2 is contiguous in velocity space as traced by lower density CO gas from the GRS survey.

Position-velocity diagram of CO, NH$$_{3}$$, and HCO$$^+$$ emission for filament 2. Blue background shows $$^{12}$$CO (1-0) emission integrated between $$-1^\circ < \textrm{b} < 1^\circ$$ (Dame et al., 2001). Black dots show HOPS, BGPS, and GRS sources associated with filament 2. Colored lines show spiral fits from the literature for the Scutum-Centaurus arm (see text for references)

Plane of the sky map integrated between 55-75 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 2. The black dashed line indicates the location of the physical Galactic midplane, while the solid black lines indicate $$\pm$$ 20 pc from the galactic midplane at the 4.4 kpc distance to filament 2, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. A trace of filament 2, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map. The correlation between filament 2’s mid-IR extinction feature and the Scutum-Centaurus’s arm most intense CO emission hints that it may be a spine of Scutum-Centaurus as traced by lower density CO gas

Filament 2 lies within $$\approx$$ 20 pc of the physical galactic midplane. The background is a GLIMPSE Spitzer 8 $$\mu\textrm{m}$$ image. The dashed line is color-coded by Dame et al. (2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 4.4 kpc distance to filament 2, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares, triangles, and circles correspond to HOPS, BGPS, and GRS sources, respectively. The label “with tilt” indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 2 can be seen in the inset.

# Filament 3 (“BC_24.95-0.17”)

Filament 3 is not a confirmed bone, receiving a quality grade of “C.” With an aspect ratio of $$36:1$$, it fails our criterion 6 (minimum aspect ratio of $$50:1$$). It also weakly satisfies our criterion 4 (within 10 km/s of the p-v fit to any Milky Way arm), lying 10 km/s below the Shane (1972) fit to HI for the Scutum-Centaurus arm. Despite lying at the upper limit of criterion 4, it strongly satisfies criterion 3, falling about 3 pc from the physical Galactic midplane.

We show the results of performing a slice extraction along the filamentary extinction feature of filament 3, which corresponds to the black line in the plane of the sky map below. Using Glue, we link spectral p-p-v cubes from the GRS survey with GLIMPSE-Spitzer mid-infrared images and extract velocities along a path tracing the entire extinction feature. We confirm that filament 3 is contiguous in velocity space as traced by lower density CO gas from the GRS survey.

Top: Position-velocity diagram of CO, NH$$_{3}$$, and HCO$$^+$$ emission for filament 3. Blue background shows $$^{12}$$CO (1-0) emission integrated between $$-1^\circ < \textrm{b} < 1^\circ$$ (Dame et al., 2001). Black dots show HOPS, BGPS, and GRS sources associated with filament 3. Colored lines show fits for the Scutum-Centaurus arm (see text for references).

Plane of the sky map integrated between 55-75 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 3. The black dashed line indicates the location of the physical galactic midplane, while the solid black lines indicate $$\pm$$ 20 pc from the galactic midplane at the 4.3 kpc distance to filament 3, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. A trace of filament 3, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map.

Filament 3 lies within $$\approx$$ 3 pc of the physical galactic midplane. The background is a GLIMPSE Spitzer 8 $$\mu\textrm{m}$$ image. The dashed line is color-coded by Dame et al. (2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 4.3 kpc distance to filament 3, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares, triangles, and circles correspond to HOPS, BGPS, and GRS sources, respectively. The label “with tilt” indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 3 can be seen in the inset.

# Filament 4 (“BC_021.25-0.15”)

Filament 4 is not a confirmed bone, receiving a quality grade of “C.” It weakly satisfies our criterion 4 and appears to be an interarm filament that lies between the Scutum-Centaurus and Norma-4kpc arms, in p-v. With an aspect ratio of $$40:1$$, it also fails our criterion 6. Finally, it only weakly satisfies our criterion 1 (largely continuous mid-infrared extinction feature), as there is a small break in the feature around $$l=21.25^\circ$$.

We show the results of performing a slice extraction along the filamentary extinction feature of filament 4, which corresponds to the black line in the plane of the sky map below. Using Glue, we link spectral p-p-v cubes from the GRS survey with GLIMPSE-Spitzer mid-infrared images and extract velocities along a path tracing the entire extinction feature. We confirm that filament 4 is contiguous in velocity space as traced by lower density CO gas from the GRS survey.

Position-velocity diagram of $$^{13}\rm{CO}$$ and $$\rm{HCO}+$$ emission for filament 4. Blue background shows $$^{12}$$CO (1-0) emission integrated between $$-1^\circ < \textrm{b} < 1^\circ$$ (Dame et al., 2001). Black dots show GRS and BGPS sources associated with filament 4. Colored lines show fits for the Scutum-Centaurus and Norma-4kpc arms (see text for references).

Plane of the sky map integrated between 55-75 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 4. The black dashed line indicates the location of the physical galactic midplane, while the solid black lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.9 kpc distance to filament 4, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. A trace of filament 4, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map.

Filament 4 lies within $$\approx$$ 8 pc of the physical galactic midplane. The background is a GLIMPSE Spitzer 8 $$\mu\textrm{m}$$ image. The dashed line is color-coded by Dame et al. (2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.9 kpc distance to filament 4, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares, triangles, and circles correspond to HOPS, BGPS, and GRS sources, respectively. The label “with tilt” indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 4 can be seen in the inset.

# Filament 6 (“BC_011.13-0.12”)

Filament 6 is not a confirmed bone candidate, receiving a quality grade of “B.” At 25:1, it has the smallest aspect ratio of all ten candidates, failing to satisfy our criterion 6. Despite this, it lies within 5 km/s of both the Scutum-Centaurus and Norma-4kpc fits in p-v space, as well within 15 pc of the physical galactic midplane. Designated the “snake”, note that filament 6 has been well-studied for its star formation properties, hosting over a dozen pre-stellar cores likely to produce regions of high mass star formation (Wang et al., 2014; Henning et al., 2010).

Position-velocity diagram of CO, NH$$_{3}$$, N$$_2$$H$$+$$ and HCO$$^+$$ emission for filament 6. Blue background shows $$^{12}$$CO (1-0) emission integrated between $$-1^\circ < \textrm{b} < 1^\circ$$ (Dame et al., 2001). Black dots show HOPS, BGPS, and MALT90 sources associated with filament 6. Colored lines show fits for the Scutum-Centaurus and Norma-4kpc arms (see text for references).

Plane of the sky map integrated between 15-35 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 6. The black dashed line indicates the location of the physical galactic midplane, while the solid black lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.3 kpc distance to filament 6, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. A trace of filament 6, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map.

Filament 6 lies within $$\approx$$ 15 pc of the physical galactic midplane. The background is a GLIMPSE Spitzer 8 $$\mu\textrm{m}$$ image. The dashed line is color-coded by Dame et al. (2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.3 kpc distance to filament 6, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares, triangles, and diamonds correspond to HOPS, BGPS, and MALT90 sources, respectively. The label “with tilt” indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 6 can be seen in the inset.

# Filament 7 (“BC_4.14-0.02”)

Filament 7 is a confirmed bone candidate, with a quality grade of “B.” There is a slight break in the extinction feature around $$l=4^\circ$$, so it weakly satisfies criterion 1. It moderately satisfies the other five criteria, though we note that we were unable to confirm contiguity in velocity space using lower density gas tracers, as it was outside the coverage range of both GRS and ThrUMMS. It does, however, express contiguity as traced by higher density gas from the HOPS and MALT90 surveys. It lies on top of the Scutum-Centaurus fits to CO and HI , but its velocity gradient is slightly angled with respect to these fits, suggesting it could be a potential spur of this arm.

Position-velocity diagram of CO, $$\rm{NH}_3$$, and $$\rm{N}_2\rm{H}+$$ emission for filament 7. Blue background shows $$^{12}\rm{CO} (1-0)$$ emission integrated between $$−1^\circ<\rm{b}<1^\circ$$ (Dame et al., 2001). Black dots show HOPS and MALT90 sources associated with filament 7. Colored lines show fits for the Scutum-Centaurus arm (see text for references).

Plane of the sky map integrated between 0-20 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 7. The black dashed line indicates the location of the physical galactic midplane, while the solid black lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.1 kpc distance to filament 7, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. A trace of filament 7, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map.

BC_4.14-0.02 lies within $$\approx$$ 20 pc of the physical galactic midplane. The background is a GLIMPSE Spitzer 8 $$\mu\textrm{m}$$ image. The dashed line is color-coded by Dame et al. (2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.1 kpc distance to filament 7, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares and circles correspond to HOPS and MALT90 sources, respectively. The label “with tilt” indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 7 can be seen in the inset.

# Filament 8 (“BC_357.62-0.33”)

Filament 8 is not a confirmed bone candidate, failing to satisfy our $$50:1$$ minimum aspect ratio criterion. Otherwise, it moderately or strongly satisfies the other five criteria. It lies exactly on and parallel to the physical Galactic midplane. Though it lies about 6-8 km/s from the Scutum-Centaurus arm, its exhibits a similar velocity gradient and traces a prominent peak of CO emisson in p-v space.

We show the results of performing a slice extraction along the filamentary extinction feature of filament 8, which corresponds to the black line in the plane of the sky map below. Using Glue, we link spectral p-p-v cubes from the ThrUMMS survey with GLIMPSE-Spitzer mid-infrared images and extract velocities along a path tracing the entire extinction feature. We confirm that filament 8 is contiguous in velocity space as traced by lower density CO gas from the ThrUMMS survey.

Top: Position-velocity diagram of CO and $$\rm{NH}_3$$ emission for filament 8. Blue background shows $$^{12}\rm{CO} (1-0)$$ emission integrated between $$−1^\circ<\rm{b}<1^\circ$$ (Dame et al., 2011). Black dots show HOPS sources associated with filament 8. The colored lines are fits for the Scutum-Centaurus arm (see text for references).

Plane of the sky map integrated between -15-0 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 8. A trace of filament 8, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map. The black dashed line indicates the location of the physical galactic midplane, while the solid black lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.0 kpc distance to filament 8, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The correlation between filament 8’s mid-IR extinction feature and the Scutum-Centaurus’s arm more intense CO emission hints that it may be a spine of Scutum-Centaurus as traced by lower density CO gas.

Filament 8 lies right on the physical galactic midplane. The background is a GLIMPSE Spitzer 8 μm image. The dashed line is color-coded by (Dame et al., 2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.0 kpc distance to filament 8, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares correspond to HOPS sources. The label ”with tilt" indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 8 can be seen in the inset.

# Filament 9 (“BC_335.31-0.29”)

Filament 9 is a confirmed bone candidate, receiving a quality grade of “B.” It strongly satisfies our criterion 3 (lying exactly on the physical Galactic midplane) and moderately satisfies criterion 2 (lying at a slight $$15^\circ$$ angle with respect to the plane). Filament 9 is notable for lying perpendicular to the Dame et al. (2011) Scutum-Centaurus fit in p-v space, suggesting it could be a potential interarm filament.

We show the results of performing a slice extraction along the filamentary extinction feature of filament 9, which corresponds to the black line in the plane of the sky map below. Using Glue, we link spectral p-p-v cubes from the ThrUMMS survey with GLIMPSE-Spitzer mid-infrared images and extract velocities along a path tracing the entire extinction feature. We confirm that filament 9 is contiguous in velocity space as traced by lower density CO gas from the ThrUMMS survey.

Position-velocity diagram of CO, NH$$_{3}$$, and $$\rm{N}_{2}\rm{H}+$$ emission for filament 9. Blue background shows $$^{12}$$CO (1-0) emission integrated between $$-1^\circ < \textrm{b} < 1^\circ$$ (Dame et al., 2001). Black dots show HOPS and MALT90 sources associated with filament 9. Colored lines are fits to the Scutum-Centaurus arm (see text for references).

Plane of the sky map integrated between -50-30 km/s, the approximate velocity range of the Scutum-Centaurus arm in the region around filament 9. The black dashed line indicates the location of the physical galactic midplane, while the solid black lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.2 kpc distance to filament 9, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. A trace of filament 9, as it would appear as a mid-IR extinction feature, is superimposed on the CO emission map. The correlation between filament 9’s mid-IR extinction feature and the Scutum-Centaurus’s arm more intense CO emission hints that it may be a spine of Scutum-Centaurus as traced by lower density CO gas.

Filament 9 lies right on the physical galactic midplane. The background is a GLIMPSE Spitzer 8 $$\mu\textrm{m}$$ image. The dashed line is color-coded by Dame et al. (2011) LSR velocity and indicates the location of the physical galactic midplane. The solid colored lines indicate $$\pm$$ 20 pc from the galactic midplane at the 3.2 kpc distance to filament 9, assuming the candidate is associated with the Dame et al. (2011) Scutum-Centaurus model. The squares and diamonds correspond to HOPS and MALT90 sources, respectively. The label “with tilt” indicates that both the Sun’s 25 pc elevation above the IAU mid-plane and the small tilt of the plane caused by the offset of the Galactic Center from the IAU (0,0) have been accounted for in this view. A closer look at filament 9 can be seen in the inset.

# Filament 10 (“BC_332.21-0.04”)

Filament 10 is a confirmed bone candidate, receiving a quality grade of “B.” It weakly satisfies our criterion 1 (largely continuous mid-infrared extinction feature). We speculate that it is likely being broken apart by stellar feedback, making it more difficult to detect continuity in the extinction feature. Otherwise, it moderately or strongly satisfies the other five criterion, lying within 5 km/s of the Dame et al. (2011) Scutum-Centaurus fit in p-v space, and within 10-15 pc of the physical galactic midplane.

We show the results of performing a slice extraction along the filamentary extinction feature of filament 10, which corresponds to the black line in the plane of the sky map below. Using Glue, we link spectral p-p-v cubes from the ThrUMMS survey with GLIMPSE-Spitzer mid-infrared images and extract velocities along a path tracing the entire extinction feature. We confirm that filament 10 is contiguous in velocity space as traced by lower density CO gas from the ThrUMMS survey.