With the recent Nature publication showing that M dwarfs did not form in prior epochs as frequently as we had thought, what implications does this have on galaxy mass estimates and, by extension, the amount of dark matter necessary for galaxies to have the dynamics we observe?

My understanding is that because of how faint red dwarfs are, we have had to rely pretty heavily on assumptions about the ubiquitous concentration of red dwarfs with some variability, but this article seems to suggest a pretty big discrepancy between what we previously believed the concentrations to be and what they actually are.

  • 10
    $\begingroup$ For those (like me) that kept wondering what IMF stands for after reading the answers, I quote the Nature article abstract: "the stellar initial mass function (IMF), namely the distribution of the stellar mass formed in each episode of star formation". Wikipedia also has an article on it: en.wikipedia.org/wiki/Initial_mass_function $\endgroup$
    – Pere
    Feb 7 at 18:17

3 Answers 3


Probably nothing changes

Several reasons for this:

  • The study actually asserts (last paragraph of main text) that:

    The IMF variation also calls for an extensive revision of star formation rates and stellar mass in the Milky Way and external galaxies, especially those with extreme physical conditions and in the early Universe.

    (IMF here stands for Initial Mass Function, which is a function that describes the initial masses of stars in the galaxy.)

    This implies that the result only affects star formation rates & the total mass of stars, which makes intuitive sense given that the study deals with the IMF. The mass of the galaxy itself is unaffected. (It seems there is a missing word in the abstract where it says "mass estimation of galaxies" instead of "stellar mass estimation of galaxies".)

  • The mass of galaxies is generally estimated by methods other than summing up all the visible matter (because there's dark matter, and because there is ordinary matter that is not luminous). For example, the evidence for dark matter in galaxy dynamics comes from rotation curves & velocity dispersion, both of which are insensitive to the amount of stars in the galaxy. More stars could mean less dark matter, and less stars could mean more dark matter, but the mass of the galaxy itself is insensitive to this result.

  • There is also cosmological evidence for the abundance of dark matter, which is also insensitive to this result. In fact, one could argue that the cosmological angle is more important, because when we say "dark matter" we usually refer to non-baryonic dark matter. Non-baryonic dark matter cannot form stars that we can see, so this result has no bearing on it. Furthermore, cosmologists have already been investigating MACHOs (MAssive Compact Halo Objects, which faint red dwarves would fall under) as dark matter candidates; this has also been largely ruled out.

  • Finally, cosmological constraints also place a fairly firm bound on the baryonic density at about 5% of the critical density. Baryons in our universe can take the form of gas, stars, planets, and so on. Changing the amount of stars mean there is less gas/planets etc., but it does not change the 5% number.


Barely, because the estimates of dark matter are not sensitive to the IMF, they use (a) the dynamics of objects in galaxies to estimate the total mass of that galaxy, (b) observations of kinematics in galaxy clusters that depend on the total mass present, (c) measurements of the cosmic microwave background combined with measurements of primordial abundances of deuterium and helium that directly give the fraction of dark matter in the universe.

While the latter is really the gold standard and doesn't depend on the stellar IMF at all, the first two techniques do to some extent depend on some assumption about the stellar IMF through the stellar mass to light ratio. This is in order to separate the inferred total mass distribution into a baryonic (stellar) and dark component. However, although M dwarfs contribute almost no light (compared with higher mass stars) and are very common in the solar neighbourhood, they also contribute little mass compared with the overall total of all matter, so the results are insensitive to the numbers of M dwarfs unless they are wrong by orders of magnitude.

A further point to consider is that because all the low-mass stars ever born are still alive, the IMF that is has been previously estimated and used for low-mass stars would be the time-averaged IMF. But that is the correct thing to use when calculating the stellar mass of the average population in a galaxy. What might be affected is how much baryonic matter was made into stars at what point in the past - i.e. the star formation history.

I will read the paper with interest and come back to this. Estimating the ages of M dwarfs is extremely difficult.


I don't know how representative for the whole galaxy (never mind for other galaxies) one should consider a study that has sampled stars merely within a radius of about 300 pc of the Sun. This is just about 0.03% of the volume of the Milky Way!

Secondly, the masses of the stars have not been obtained as independent observational quantities in this study but are inferred from theoretical mass-luminosity relationships, which however are notoriously inaccurate for low mass stars. Masses can only be measured as such by observing binary stars, and the MASSIF study that was made about 20 years ago (based on Hubble Space Telescope data of binary stars) showed actually that existing models (at the time) underestimated the number of low mass stars in the solar neighbourhood by about a factor 5.

So I would certainly not be jumping to any conclusions here.

  • $\begingroup$ Neither of these points addresses whether we would infer more dark matter as a result of the claimed variation in the IMF. The absolute accuracy of the IMF, while relevant to estimating the amount of stellar mass, is not especially relevant in deciding whether it changes significantly. $\endgroup$
    – ProfRob
    Feb 9 at 23:05
  • $\begingroup$ @ProfRob My point is that the results of this study can not be seen as a statement about the global galactic mass of stars in the first place a) because of its local nature and b) because of the inherent uncertainties in the theoretical models used to relate the luminosities to stellar masses How this relates to the dark matter question is therefore really irrelevant. $\endgroup$
    – Thomas
    Feb 10 at 8:27

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .