Filament Line-Mass Function (FLMF)

• FLMF is the statistical distribution of filament mass per unit length in molecular clouds, essential for understanding fragmentation and star formation.
• Different methodologies (mean, local segments, or pixel-wise) affect FLMF measurements, influencing criticality and the derived power-law slopes.
• Empirical studies show a Salpeter-like power-law tail in FLMF, connecting filament properties with the prestellar core mass function and stellar IMF.

Searching arXiv for the cited FLMF/filament line-mass literature to ground the article in current papers. arxiv.search query: "Filament line-mass function FLMF filament linear density function molecular clouds IMF" The Filament Line-Mass Function (FLMF) is the statistical distribution of filament mass per unit length in molecular clouds. In recent observational work, especially “A Salpeter-like filament linear density function across nearby molecular clouds,” the same quantity is denoted by the filament linear density $\Lambda$, and the corresponding distribution is called the filament linear density function (FLDF); no substantive distinction is introduced between FLMF and FLDF, and both refer to the distribution of filament line masses (Zhang et al., 15 Apr 2026). In this literature, the FLMF has become a central descriptor of filament populations because line mass is the variable most directly tied to filament criticality, fragmentation, and the proposed link between filamentary gas structure, the prestellar core mass function (CMF), and the stellar initial mass function (IMF) (André et al., 2019).

1. Terminology and mathematical forms

The underlying physical quantity is the filament mass per unit length. Different papers write it as $M_{\rm line}$, $m_\lambda$, $\lambda$, or $\Lambda$, but the meaning is the same: line mass or linear density. In the seven-cloud study of nearby molecular clouds, the preferred notation is $\Lambda$, and the FLDF is written as

$$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$

with cumulative form $N_{\rm C}(>\Lambda)$ also analyzed (Zhang et al., 15 Apr 2026). André et al. define the FLMF as the differential distribution of filament masses per unit length, $$g(M_{\rm line}) \equiv {\rm d}N/{\rm d}\log M_{\rm line}$$, emphasizing that it is distinct from the filament mass function in total filament mass because local fragmentation physics depends primarily on $M_{\rm line}$, not on $M_{\rm line}$0 (André et al., 2019).

The literature does not employ a single operational realization of the FLMF. Some studies use one mean line mass per filament, typically $M_{\rm line}$1, and then histogram the resulting filament population (André et al., 2019). Other studies define a local FLMF from short independent segments along filament crests. In the California GMC, the emphasized FLMF is built from local line masses measured in $M_{\rm line}$2 pc chunks, because “it is the local rather than the average value of the mass per unit length of a filament that determines its ability to fragment” (Zhang et al., 2024). Still other work adopts pixel-wise local line masses derived from column density and a fixed fiducial width, so that the “FLMF” becomes a histogram of filamentary pixels rather than of individual filament objects, as in hub-filament systems such as W3(OH), W3 Main, and S 106 (Kumar et al., 8 Aug 2025).

Theoretical papers also differ in formal convention. In the filament-fragmentation model of Hennebelle–Chabrier type extended to filaments, the filament mass-per-length function is written in linear form,

$M_{\rm line}$3

that is, $M_{\rm line}$4 rather than $M_{\rm line}$5 (Lee et al., 2017). This difference is not merely notational: logarithmic and linear differentials shift the power-law exponent by unity, so quoted slopes must always be interpreted together with the paper’s precise definition.

2. Measurement strategies and sample construction

Observational FLMF measurements depend strongly on how filaments are extracted, how backgrounds are removed, and whether the statistic is object-based or local. In the seven-cloud FLDF study, filaments are extracted with the multiscale method $M_{\rm line}$6, and the differential FLDF is fit in logarithmic bins of width $M_{\rm line}$7 dex, retaining only bins with $M_{\rm line}$8, while cumulative fits are restricted to $M_{\rm line}$9 (Zhang et al., 15 Apr 2026). That same work explicitly shows that the line-mass statistic becomes Salpeter-like only after integrating over the full hierarchy of spatial scales, making the treatment of multiscale structure integral to the definition of the population function (Zhang et al., 15 Apr 2026).

In the California GMC, $m_\lambda$0 is again used, but the measurement philosophy is more explicitly local. Filaments are separated from compact sources and large-scale background on the filamentary component map, then skeletons are divided into $m_\lambda$1 pc chunks, with an alternative $m_\lambda$2 pc segmentation used as a robustness test. Appendix simulations show that the observed sample is $m_\lambda$3 complete above $m_\lambda$4 in measured $m_\lambda$5, and that $m_\lambda$6 typically underestimates the true line mass by about $m_\lambda$7 for $m_\lambda$8 (Zhang et al., 2024). This is a direct reminder that FLMF slopes and turnover masses are conditional on extraction depth, width recovery, and background subtraction.

Synthetic-observation studies sharpen these methodological issues. In “From filaments to clumps,” line masses are computed from background-subtracted synthetic Herschel column-density maps after FILFINDER extraction of 8,832 filaments and 110,193 branches, and the resulting FLMF-like distributions are fit with MLE using the powerlaw package (Ma et al., 7 Jun 2026). In contrast, the galaxy-scale MHD analysis of filament hierarchies greater than $m_\lambda$9 pc emphasizes line mass as the key stability variable but does not publish a standalone $\lambda$0 or $\lambda$1 for its 325-filament sample, showing that many filament surveys remain interpretive rather than directly statistical from the FLMF standpoint (Pillsworth et al., 1 Apr 2025).

A further variation appears in hub-filament systems. There the FLMF is constructed from filamentary pixels selected by a Hessian minimum-curvature method, with local line mass assigned as

$\lambda$2

using a fixed $\lambda$3 pc. The resulting histograms are then compared separately inside and outside a hub ellipse defined from a broken-power-law radial column-density profile (Kumar et al., 8 Aug 2025). This is not equivalent to a catalog-level object FLMF, but it is explicitly used as an evolutionary diagnostic of dense-gas redistribution in HFSs (Kumar et al., 8 Aug 2025).

3. Empirical forms and observed slopes

The main observational result of the recent nearby-cloud analysis is that the combined line-mass distribution across seven nearby molecular clouds follows a power law only when the full spatial hierarchy is included, with

$\lambda$4

a slope described as mirroring the Salpeter IMF slope of $\lambda$5 (Zhang et al., 15 Apr 2026). The same study also reports that the median linear densities of filaments increase approximately linearly with spatial scale, $\lambda$6, and that the fraction of supercritical filaments varies strongly from cloud to cloud, from a few per cent to over $\lambda$7 (Zhang et al., 15 Apr 2026). The implication is that a Salpeter-like FLMF is not simply the line-mass distribution at one arbitrarily chosen scale, but the result of hierarchical integration.

Earlier nearby-cloud work by André et al. found that the filament mass function and FLMF have very similar shapes and are both consistent with a Salpeter-like law in the thermally supercritical regime,

$\lambda$8

in clear contrast to the much shallower mass functions of clouds and clumps (André et al., 2019). The California GMC study reinforced this result with a local-segment FLMF,

$\lambda$9

and also reported a K-S-compatible cumulative slope of $\Lambda$0 when the fit is restricted to the thermally supercritical regime $\Lambda$1 (Zhang et al., 2024).

Synthetic Herschel observations likewise recover a high-line-mass power-law tail. For full synthetic filaments the measured distribution is

$\Lambda$2

while the Hi-GAL comparison sample yields

$\Lambda$3

Branches form a steeper population with slope $\Lambda$4, demonstrating that internal decomposition changes the measured FLMF appreciably (Ma et al., 7 Jun 2026).

The FLMF can also vary by environment within a single system. In W3(OH), the high-line-mass end from roughly $\Lambda$5 to $\Lambda$6 is described as a relatively smooth Salpeter-like power law, whereas W3 Main and S 106 show separated hub and filament components, with hub slopes around $\Lambda$7 and $\Lambda$8, and filament-region slopes around $\Lambda$9 and $\Lambda$0, respectively (Kumar et al., 8 Aug 2025). This does not negate the Salpeter-like measurements in nearby-cloud samples; rather, it shows that the FLMF can split into distinct structural components when measured locally within HFS subregions.

4. Criticality, support, and the high-line-mass end

The physical significance of the FLMF derives from the fact that line mass is the principal stability variable for an approximately cylindrical filament. Several papers adopt the classical thermal critical line mass,

$\Lambda$1

for $\Lambda$2 K gas, as the reference threshold separating thermally subcritical and supercritical regimes (Zhang et al., 2024). In the California GMC, the concentration of prestellar cores near locally supercritical filamentary structures is one of the central empirical arguments that the FLMF is directly relevant to star formation (Zhang et al., 2024).

The criticality picture becomes more complex once turbulence and magnetic support are included. In galaxy-scale MHD simulations, the thermal criterion alone makes almost all large filaments appear supercritical, whereas a turbulent+magnetic critical-line-mass framework produces a roughly $\Lambda$3 split between average subcritical and supercritical systems and shows that local criticality along a filament becomes more informative than global $\Lambda$4 on $\Lambda$5 pc scales (Pillsworth et al., 1 Apr 2025). A plausible implication is that the observed high-line-mass end of the FLMF should not be interpreted against a single universal threshold.

Analytical MHS calculations make this dependence explicit. For isothermal filaments threaded by a lateral magnetic field, the maximum supported line mass is

$\Lambda$6

with magnetic support becoming significant when

$\Lambda$7

(Tomisaka, 2014). In negative-index polytropic filaments threaded by a lateral magnetic field, the maximum line mass is likewise not universal but depends on both thermal structure and magnetic flux, through an empirical formula involving $\Lambda$8 and $\Lambda$9 (Kashiwagi et al., 2021). These results show that any FLMF high-end truncation or broadening can reflect variations in $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$0, $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$1, external pressure, and mass loading as much as simple counting statistics.

Formation environment also matters. In MHD shock simulations, local line-mass histograms built from filament skeleton pixels show that strong shocks can generate filaments with $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$2 within about $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$3 Myr and can produce a Salpeter-like high-line-mass tail $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$4 in the plotted quantity, whereas weaker-shock environments yield slower buildup and greater dependence on turbulence and self-gravity (Abe et al., 2020). This suggests that the FLMF is shaped not only by equilibrium support limits but also by the dominant filament-formation channel.

5. Relation to the CMF, IMF, and hub formation

The strongest scientific interest in the FLMF comes from its repeated proximity to Salpeter-like slopes and its use as an intermediary between cloud structure and stellar masses. André et al. argued that because the FLMF of thermally supercritical filaments is Salpeter-like, and because most prestellar cores form in transcritical or supercritical filaments, the prestellar CMF and hence the IMF may be at least partly inherited from the FLMF through gravitational fragmentation of individual filaments (André et al., 2019). The California GMC study sharpened this argument by replacing whole-filament averages with local $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$5 pc segment line masses and proposing that the global CMF is the weighted integral of the CMFs produced by individual filament segments, with a weight proportional to $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$6 (Zhang et al., 2024).

A different but complementary route comes from the statistics of longitudinal line-mass fluctuations. The Herschel power-spectrum analysis of nearby filaments did not construct a population FLMF, but it measured a beam-corrected fluctuation spectrum $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$7, close to the slope assumed by Inutsuka’s model for generating a Salpeter-like CMF tail (Roy et al., 2015). This does not replace the FLMF; it identifies an internal stochastic structure of $$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$8 that can seed the fragment mass spectrum inside a filament population.

Analytical CMF models built directly on a filament line-mass distribution make the weighting role of the FLMF explicit. In the magnetized fragmentation theory of individual filaments, the cloud-wide CMF is

$$\frac{{\rm d}N}{{\rm d}\log\Lambda} \propto \Lambda^{-\alpha},$$9

where $N_{\rm C}(>\Lambda)$0 is the filament mass-per-length function, taken in fiducial form as $N_{\rm C}(>\Lambda)$1 (Lee et al., 2017). In that framework, the observed CMF is not the CMF of one representative filament but a population average over filaments with different line masses.

The FLMF also enters hub-formation models. In the analytical treatment of filament-filament collisions, the global hub mass function inherits the power-law exponent of the FLMF, so that if

$N_{\rm C}(>\Lambda)$2

then the global hub mass function has the same exponent (Tomisaka et al., 13 Mar 2026). The paper’s explicit interpretation is that a massive hub is formed by the collision of two massive filaments (Tomisaka et al., 13 Mar 2026). This suggests that the FLMF may regulate not only fragmentation along filaments but also the mass spectrum of higher-order hub structures in HFSs.

6. Scope, ambiguities, and methodological caveats

The term “FLMF” hides substantial heterogeneity. Some papers measure a distribution of one mean $N_{\rm C}(>\Lambda)$3 per filament, others a distribution of segment line masses, and others a pixel-wise histogram of local line masses. Still others focus on quantities adjacent to, but not identical with, the FLMF. The galaxy-scale filament-hierarchy study emphasizes average and local line mass and its comparison to thermal, turbulent, and magnetic critical thresholds, yet does not publish an explicit FLMF slope (Pillsworth et al., 1 Apr 2025). The Herschel fluctuation-spectrum study characterizes $N_{\rm C}(>\Lambda)$4 statistically through PDFs and 1D power spectra, but not through a population distribution $N_{\rm C}(>\Lambda)$5 (Roy et al., 2015). These differences matter whenever slopes are compared across papers.

Projection, extraction, and resolution effects can shift the inferred line-mass population appreciably. Synthetic Herschel analyses show that FILFINDER can join structures into extended networks, altering lengths and therefore $N_{\rm C}(>\Lambda)$6, even though the high-$N_{\rm C}(>\Lambda)$7 tail appears comparatively robust in the synthetic-versus-Hi-GAL comparison (Ma et al., 7 Jun 2026). Cloud Factory synthetic observations show that lower resolution shifts the observed line-mass PDF upward, that projection effects can be severe in crowded regions, and that most tracked filaments evolve in the observational $N_{\rm C}(>\Lambda)$8-$N_{\rm C}(>\Lambda)$9 plane primarily by accretion, with segmentation and dispersal introducing additional changes in $$g(M_{\rm line}) \equiv {\rm d}N/{\rm d}\log M_{\rm line}$$0 (Feng et al., 2024). A plausible implication is that any empirical FLMF must be read as an observation-conditioned population statistic rather than as a direct census of intrinsic 3D filament line masses.

A further caveat is that the line-mass threshold used to define “supercritical” is itself model-dependent. Thermal thresholds are numerically simple and widely used, but MHS and MHD work shows that magnetic flux, polytropic structure, turbulence, and scale can all shift the relevant critical line mass (Tomisaka, 2014). Consequently, the supercritical fraction inferred from an FLMF is not an invariant quantity across methodologies. The recent nearby-cloud FLDF result, in which the supercritical fraction ranges from a few per cent to over $$g(M_{\rm line}) \equiv {\rm d}N/{\rm d}\log M_{\rm line}$$1 across seven clouds, is therefore best understood as a cloud- and method-specific population measure rather than a universal constant (Zhang et al., 15 Apr 2026).

Taken together, the current literature supports a precise but nontrivial picture. The FLMF is the statistical distribution of filament line masses, but its measured form depends on whether “filament” means an extracted object, a local segment, or a filamentary pixel; on whether masses are raw, background-subtracted, or source-subtracted; and on whether the relevant physics is interpreted through thermal, virial, or magnetic criticality. Within those constraints, multiple independent studies nevertheless converge on a high-line-mass power-law regime close to Salpeter, making the FLMF one of the clearest observational interfaces between filamentary ISM structure and the mass spectrum of star formation (Zhang et al., 15 Apr 2026).