View metadata,citationandsimilarpapersatcore.ac.uk arXiv:astro-ph/0701848v1 30 Jan 2007 xetd tesaclrto eas ti h quantity the is it because than acceleration bigger stress much I are velocities expected. be random characterized or to is rotation problem found the thus: statis- Sometimes is generic. least discrepancy and acceleration at real the checked, The and be plane. tically, can a on assumptions circles in par- geometric move a stars in galaxy that spiral ticular e.g. measured assumptions, Doppler geometric ac- with combining Those velocities by systems. inferred these are in celerations visible basis galax- actually the of matter on galaxies, the clusters made of spiral predictions large Newtonian of the exceed in well outskirts ies, swimming the galaxies in of clouds and of gas accelerations and the Yet stars well. physics extremely Newtonian motions So describe there. gravita- weak and are light’s, fields to tional compared low are Speeds galaxies. truth. the then to motions, approximation good rapid a grav- extremely is strong theory with extremely Newtonian or with that fields, situations belief itational the sidesteps is in one few comfort gravitation a if took until settings: up astronomers But ago new decades supposed. these Newton to than gravitation difficult crucial more of are aspects GR nonlinear The in galactic by etc. compact confirmed sources, nuclei, X-ray now galactic holes, of exotic black observations those myriad universe, of the existence of double the Hulse-Taylor denizens anticipated celebrated has of the and dynamics in pulsar, the system, GR in solar effects (GR). the subtle relativity approxi- predicted general an correctly 1915 iden- only has usually Einstein’s is theory, it with gravitation that tified relativistic known a been to long Neverthe- mation has centuries. three it over less, well for faithfully nology ∗ fJrslm ia a,Jrslm994 Israel; 91904, [email protected] Jerusalem Email: Ram, Givat Jerusalem, of uhrsaddress: Author’s u sta o osdrtestaini h el of realm the in situation the Consider so? that is But tech- and physics served has theory gravity Newtonian a ersosbefrteecc fMN,adcnld iha with conclude of and description full MOND, a of reach to efficacy needed the still are for developments cosmology responsible and be lensing T can confronting gravitational in on groups data o various observational by some work galaxies reviewing recent fo on of formulations After relativistic clusters of development for cosmology. the well recollect or less econom lensing doing observa elucidating while gravitational interpret in of scales, successful to paradigm all been way of had Milgrom’s alternative MOND already matter. an years dark provided many has For (MOND) matter”. “dark omr ain su xssi otmoayatohsc a astrophysics contemporary in exists issue salient more No aa nttt fPyis erwUniversity Hebrew Physics, of Institute Racah .Introduction 1. h oie etna dynamics—MOND Newtonian modified The n t mlctosfrnwphysics new for implications its and ao .Bekenstein D. Jacob should e V atr(M nssesrnigfo h eytenuous very the masses visible from dark with galaxies ranging much acceleration spheroidal systems dwarf of the in existence resolve (DM) the to matter postulating trend ha- by the invisible discrepancy started but Thus massive in los. embedded (including often galaxies disk are bar, significant that Since spirals) no hypothesis have the galaxies dispersion. justified spiral this velocity the little of third with about bar en- systems the are to theory, against Newtonian demic galaxies in which, disk instabilities stabilize forming would halos sive wrong? is what case, gravitational any the In of field. strength the measures directly which h ra lseso aaiswt bevdmse nthe in masses observed with 10 galaxies of clusters great the aeytt rvd needn vdneo Msexis- DM’s of evidence e.g. independent tence, provide to [4] yet experiments of have laboratory years thirty and But exploration Mu˜nozastronomical paradigm [3]). and DM Khalil by the provided to is introduction excessive easy the (an for account accelerations to provide pull to gravitational is missing role the DM’s The exists. discrepancy celeration yaisshm u owr yMlrm[,6 ]a an DM. as to 7] appeal 6, systematic [5, the Newtonian Milgrom to modified by alternative the forward on put scheme focuses dynamics review This DM? to itiuini prlglxe,ms ftervsbemass light visible the their from of radius most judge with galaxies, to velocity spiral in the because, distribution of vis- expected falloff the but naively a beyond center, Yet was well the to edge. of grows disk’s out radius ible moves the one as veloc- drops as (linear) hardly rises a first with which potential, center gravitational ity galaxy’s the each galax- of around spiral tracers of circle disks as it the serve [8], in coworkers which clouds and ies, gas Rubin that since as clear ever well been First, as has Bosma galaxies. facts of work disk empirical overarching the of two phenomenology the the of in stock take to well e tt oewt h atdfiiny comment I deficiency. last the with cope to it r twsraie yOtie n ebe 1 htmas- that [1] Peebles and Ostriker by realized was It ops ugeto Madatraie oi,i is it it, to alternatives and DM on judgment pass To 14 ,arltvsi moieto OD with MOND, of embodiment relativistic a S, M nisoiia omi ol o address not could it form original its in ; h ol novn ouosre matter. unobserved no involving world the hogotIakwa oto physics of sort what ask I Throughout . ⊙ ∗ alak[] nbifi n ytmweea ac- an where system any in brief in [2], ballpark dcsooyta hto h elusive the of that than cosmology nd h vdnei ao fMN,I MOND, of favor in evidence the f clytednmc fds galaxies disk of dynamics the ically γ in ihu pelt invisible to appeal without tions asfo t ea.I hr nalternative an there Is decay. its from rays .Sapnn h problem the Sharpening 2. pria fwa theoretical what of appraisal n oie etna dynamics Newtonian modified provided byCERNDocumentServer ∼ brought toyouby 10 7 M ⊙ to CORE 2 is rather centrally concentrated. Newtonian gravitation would thus predict that “rotation curves” (RC) should drop as r−1/2 outside the bright parts of these galaxies, but this is not seen in over a hundred extended RCs mea- sured with sufficient precision [9, 10, 11]. In fact in many spiral galaxies, particularly those possessing high surface brightness, the extended RC becomes flat away from the central parts (FIG. 1).
FIG. 2: The TF correlation for a sample of galaxies as a log- 10 log plot of the blue band luminosity LB in units of 10 L⊙ vs. the asymptotic rotational velocity Vrot in km/s. The straight line (mine) has slope exactly 4 and corresponds to MOND’s FIG. 1: Collage of RCs of nearby spiral galaxies obtained prediction, Eq. (2), for a reasonable mass (in units of ⊙) M by combining Doppler data from CO molecular lines for the to LB ratio ΥB = 3. Graph reproduced from Ref. 13 by central regions, optical lines for the disks, and HI 21 cm line permission of the American Astronomical Society. for the outer (gas) disks. The rotation velocity in units of km s−1 is plotted vs galactocentric radius R in kiloparsecs (kpc); 1 kpc 3000 light years. It is seen that the RCs are flat to well beyond≈ the edges of the optical disks ( 10 kpc). Graph from by fine tuning the halo’s parameters is it feasible to match Ref. 9, reprinted with permission from∼ the Annual Review of this kind of profile to the observed flat RCs as well as Astronomy and Astrophysics, Volume 39 (c)2001 by Annual would the isothermal sphere (the need to fine tune pa- Reviews www.annualreviews.org rameters of a halo model in order to fit the shape of the galaxy’s RC was already clear long ago [14]). And the Second, as originally pointed out by Tully and Fisher predicted 1/r “cusp” in the density profile is also obser- (TF) [12], the rotation velocity Vrot and the blue band vationally problematic [4]. luminosity LB of a galaxy are simply correlated (FIG. 2). The DM halo hypothesis is evidently an attempt to re- In a more informative form [10], the TF law states that solve the acceleration discrepancy within orthodox grav- for disk galaxies the luminosity in the near infrared band, itation theory. But in the coming to terms with this LK′ (a good tracer of stellar mass), is proportional to the discrepancy, suspicion fell on Newtonian gravity already fourth power of the rotation velocity in the flat part of early in the game. Zwicky, who had exposed the accel- the RC, with a universal proportionality constant. eration discrepancy in clusters of galaxies [15], opined The flat extended RC’s have provided the logic under- much later that the discrepancy may reflect a failure of pinning the DM paradigm as follows. In the flat part conventional physics [16]. Concrete proposals for depar- of a RC, the rotation velocity is r independent, so the tures from the Newtonian inverse square law in various centripetal acceleration goes as 1/r. Thus in the plane settings were put forward by a number of workers [17]. of the galaxy the gravitational field must be decreasing All of these still regarded gravitation as a linear inter- as 1/r. According to Poisson’s equation of Newtonian action with the strength of the field proportional to its theory, such a gravitational field, if assumed spherically source’s mass. symmetric, must be sourced by a mass distribution with As Milgrom realized [5], such a modification of the 1/r2 profile (isothermal sphere model). Since the visible gravity law, if relevant on the scales of galaxies, is in- mass density (in stars and gas) in the inner disk drops compatible with the TF law. For one it would imply much faster than this, it is consistent to assume that the that a mass m1 at r1 generates at r2 the acceleration total mass distribution in the outer parts is quasispher- field m1g(r2 r1), with g depending only on fundamen- ical. The conclusion is that each spiral galaxy must be tal constants− and on the relative positions of source and immersed in a roundish DM halo with mass density pro- field points in accordance with the principle of homogene- file tending, at large r, to 1/r2. ity of space. Then, obviously, for matter orbiting with But questions plague the halo hypothesis. The halo speed Vrot at radius R in the outskirts of a galaxy (with DM, though much searched for, has never been detected its source mass, M, concentrated at a fairly definite dis- 2 directly [4]. Cosmogonic simulations of DM halo forma- tance R), the acceleration, Vrot /R should be equal to tion within Newtonian physics predict that such a halo M g(R) . This would require not only g(R) R−1, | | 2 | | ∝ would have a density profile behaving like 1/r for small but also M Vrot , the last incompatible with the TF 3 ∝ r and gradually steepening to 1/r asymptotically. Only law. Then again g(r2 r1) with the correct dimen- − 3
−2 −1 sions, LT M , cannot be built by using only G, r2 r1 Introducing the ratio of mass to luminosity L, Υ, we have − −1 4 and the scale of length ℓ at which gravity departs from the added prediction L = (Ga0Υ) Vf . Although Υ in Newtonian form. And inclusion of mass in the construc- the blue band varies somewhat with galaxy color, this tion (presumably the source’s mass) would be against the last results explains the TF law, FIG. 2. Theoretically Υ spirit of linearity. The bottom line is that linear gravity, is much less variable in the near infrared (K’) band [18], even if non-Newtonian, is incompatible with the TF law. which accounts for the extra sharpness of the TF relation in that band (see Fig. 2 in Ref. 10).
3. The MOND paradigm
Milgrom [5, 6, 7] proposed a novel paradigm which can be interpreted as reflecting non-Newtonian as well as nonlinear character of gravity already at the nonrela- tivistic level. He called the scheme Modified Newtonian Dynamics (MOND). The essential part of it is the rela- tion
µ( a /a0) a = ∇Φ (1) | | − N between the acceleration a of a particle and the ambi- ent conventional Newtonian gravitational field ∇ΦN . If the function µ were unity, this would be usual− New- tonian dynamics. Milgrom assumes that the positive smooth monotonic function µ approximately equals its argument when this is small compared to unity (deep MOND limit), but tends to unity when that argument is large compared to unity (FIG. 3). The a0 is a natural constant, approximately equal to 10−10 ms−2. Itisafact that the centripetal accelerations of stars and gas clouds in the outskirts of spiral galaxies tend to be below a0.
ΜHxL 1 FIG. 4: The baryonic TF correlation for galaxies spanning 6 orders of magnitude in mass. This is a log-log plot of the 0.8 total baryonic (stars plus gas) mass b in units of the solar M 0.6 mass ⊙ vs. the observed asymptotic rotational velocity Vf in km/s.M For each of the 60 galaxies from Ref. 10 (deep blue 0.4 circles) the mass in stars comes from a fit of the shape of the RC with MOND, whereas for the eight dwarf spirals (light 0.2 blue circles) the mass in stars (relatively small) is inferred in x Ref. 19 directly from the luminosity. The green line, with 1 2 3 4 5 6 7 slope 4, is MOND’s prediction, Eq. (2). Graph reproduced from Ref. 19 by permission of the American Astronomical FIG. 3: Two popular choices for Milgrom’s µ function: the Society. “simple” function µ(x)= x/(1+x) (solid) and the “standard” function µ(x)= x/√1+ x2 (dotted). The impressive agreement of Eq. (2) with observations How does the MOND formula (1) help? Consider again is most clear from the baryonic TF law [19], FIG. 4. In stars or gas clouds orbiting in the disk of a spiral galaxy full agreement with MOND, the observed baryonic mass (mass M) with speed V (r) at radius r from its center. in spiral galaxies is accurately proportional to the fourth We must identify a with the centripetal acceleration power of the asymptotic rotation velocity. MOND’s sin- V (r)2/r. Sufficiently| | outside the main mass distribution gle formula thus unifies the two overarching facts of spiral we may estimate ∇Φ GM/r2. And at sufficiently galaxy phenomenology. | N | ≈ large r, a will drop below a0 and we shall be able to MOND is similarly successful in explaining the detailed approximate| | µ(x) x. Putting all this together gives shapes of RCs. FIG. 5 shows the measured RCs of fifteen 4 2 ≈2 V (r) /r a0GM/r from which we may conclude two galaxies together with the MOND fits based on surface things. First,≈ V (r) well outside the main mass distribu- photometry in the optical band and radio measurements. tion becomes independent of r, that is, the RC flattens The only parameter adjusted in these impressive fits is at some value Vf . Second, from the coefficients follows the disk’s stellar mass, or equivalently, the stellar mass- to-luminosity ratio Υ. The trend of the so determined −1 4 M = (Ga0) Vf . (2) values of Υ with galaxy color jibes with that predicted by 4
MOND dynamics [6]. (In galaxies like ours this occurs some way out in the disk, and that is why the RC may exhibit a brief drop before becoming flat asymptotically). On this basis Milgrom predicted that were disk galaxies with surface mass density everywhere below Σm to exist, they should show especially large acceleration discrep- ancies. A population of such galaxies became known in the late 1980’s [21], and all facets of Milgrom’s predic- tion were subsequently confirmed [23]. MOND obviously predicts that in these so called low surface brightness galaxies, the shapes of the RCs, which tend to be rising throughout the visible disk, should be independent of the precise way µ(x) switches from x to unity for x 1. Constraints of space forbid me from delving into∼ other MOND successes, e.g. for elliptical and dwarf spheroidal galaxies. But excellent reviews of these and other phe- nomenological aspects of MOND exist [10, 24, 25]. It is well, however, to stress one prominent empirical difficulty faced by the MOND formula (1). Clusters of galaxies exist which comprise hundreds of galaxies of various kinds moving in their joint gravita- 1 FIG. 5: MOND fits to measured RCs of a sample of spiral tional field with velocities of up to 103 kms− . The New- galaxies The galactocentric radius (horizontal axis) is in kpc tonian virial theorem (gravitational potential energy of a and the rotation velocity (vertical) in km/s. Each solid line is the MOND fit to the RC based on the distribution of light and quiescent system equals minus twice its kinetic energy) neutral hydrogen. The other curves are the Newtonian RCs can be used to estimate a cluster’s mass since the kinetic contributed by the gas (H and He) in the disk (short dashed), term is proportional to the total mass while the poten- by the stars (dotted) and by the central bulge (if present, long tial one goes like the square of mass. In this way it is dashed). The free parameter of the fit is the mass of the stars determined that the total gravitating mass in a typical in the disk (and the mass of the bulge if present). Figure cluster is 5-10 times the mass actually seen in galaxies as reproduced from Ref. 10 with permission, from the Annual well as in hot X-ray emitting gas, an oft constituent of Review of Astronomy and Astrophysics, (c)2002 by Annual clusters. A similar story is told by analyses of the hydro- Reviews, www.annualreviews.org. statics of the hot gas in light of its measured temperature. Finally, the gravitational lensing by several clusters has led to estimates of their masses commensurate with the stellar evolutionary models [18]; in this MOND does bet- mentioned diagnosis. ter than DM halo models [20]. To make similarly success- When the MOND formula, or even better, a MOND ful fits without just “putting DM where it is needed”, DM analog to the virial theorem [26], is used to estimate the theorists must adjust two extra parameters apart from Υ masses of clusters, one finds an improvement but not a in standardized halo models. As stressed in a recent crit- full resolution to the acceleration discrepancy. Clusters ical reappraisal [20], those galaxies where one-parameter still seem to contain a factor of two more matter than ac- MOND fits do worse than halo fits often display compli- tually observed in all known forms [10]. The optimist will cating factors that could mitigate its less than striking stress that MOND has alleviated the discrepancy with- performance. In disk galaxies MOND is unquestionably out even once overcorrecting for it; the pessimist [27] more economical, and thus more falsifiable, than the DM will view this finding as damaging MOND’s credibility. paradigm. But one should keep in mind that clusters may contain When applied to many galaxies, the MOND fits men- much invisible matter of rather prosaic nature, either baryons in a form which is hard to detect optically, or tioned serve to determine a0 with some accuracy; the −10 −2 massive neutrinos [28, 29]. However, the option that clus- value a0 =1.2 10 ms is often used [21]. It is sig- × ters contain non-baryonic cold DM between the galaxies, nificant that a0 as deduced from RC fits agrees well with −1 while logically possible, seems hardly justifiable in view the value for a0 obtained by comparing (Ga0Υ) with of MOND’s overall philosophy. the empirical coefficient in the TF law: the roles of a0 are different in the two subjects, and they are tied together only in MOND. Milgrom [22] notes a few additional con- ceptually distinct roles played by a0 in extragalactic as- 4. What is behind the MOND formula? trophysics. An early example is that a0 sets a special scale of mass surface density Σm = a0/G, and wherever What is the physical basis of MOND’s success in unify- in a system the actual surface mass density drops be- ing a lot of extragalactic data? There are certain things low Σm, Newtonian gravitational behavior gives way to the MOND formula cannot be. First, it is unlikely to be 5 just a recipe for the way DM is distributed spatially in outer planet ephemerides to an intolerable extent [36]. astronomical objects, as sometimes proposed [30]. Such (An ephemerides is the calculated position of a celestial proposals strive to explain how the scale a0 enters into body as a function of time, past or future.) By contrast, spiral galaxy properties through some regularity in the Milgrom’s nonlocal theory does predict different modifi- cosmogony of DM halos. But strong arguments exists cations of Newtonian dynamics for radial (Pioneer) and against an intrinsic scale of the halos [31]. And it is hard circular (planetary) orbits, though details have yet to be to see how such an explanation could give an account of worked out [37]. the multiple roles of a0 in different systems [22]. Milgrom has also speculated [37, 38] that modified in- The MOND formula, conceived as exact, cannot be a ertia may have its origin in an effective interaction of generic modification of the inertia aspect of Newton’s sec- bodies with the vacuum. This is motivated by the well ond law ma = f (here f is the sum of all forces acting known fact that an object swimming through a fluid has on the particle including the Newtonian gravity force), its inertial mass increased by interacting with fluid de- an alternative proposed originally by Milgrom [6] (see a grees of freedom. Of course, by local Lorentz invariance latter elaboration in Ref. 32). To see why consider a inertial motion through the vacuum is possible as usual. binary system with unequal masses m1 and m2 evolv- However, accelerated motion is distinguished from in- ing under mutual gravity alone. The time derivative of ertial motion (even before we come to inertia) by the m1v1 +m2v2, as calculated from Eq. (1), does not vanish perception that the accelerated system is immersed in a in general if at least one of the accelerations is compara- thermal bath whose temperature is proportional to the ble to or smaller than a0, since the µ’s will generally not acceleration (Unruh radiation [39]). Accordingly, Mil- be equal. So in the proposed interpretation, the MOND grom suggests that accelerated motion with respect to formula does not conserve momentum. By the same to- the vacuum may shape the inertial characteristics of ob- ken it does not conserve angular momentum, nor energy. jects in a way compatible with MOND. An alternative Milgrom sidesteps the problem by stipulating that the speculation [37, 38] is that MOND style modified iner- MOND formula is only valid for test particles moving on tia and its particular scale a0 may be generated by a a given background, e.g., stars moving in the collective symmetry in a higher dimensional spacetime, just as we gravitational field of a galaxy. We may inquire more gen- think of ordinary (Newtonian or Einstenian) inertia as erally, does the MOND formula, or some closely related reflecting Lorentz symmetry in Minkowski spacetime. In one, represent a modification of inertia in test particle one implementation of this program, with the extremely motion? To comply with the conservation laws we likely symmetric deSitter spacetime playing the role of the em- want the formula to arise from a Lagrangian. The kinetic bedding spacetime, a relation between the cosmological constant and a0 surfaces, which is actually approximately part of the Lagrangian should give us the µ( a /a0) a part with whatever “corrections” are required. However,| | satisfied if the observed acceleration of the Hubble expan- it is a theorem that no MOND-like dynamics exists that sion is interpreted as reflecting a nonzero cosmological constant. simultaneously has a Newtonian limit for a0 0 (all accelerations are large), is Galilei invariant, and→ is deriv- able from a local action [32]. (“Local” means that the 5. MOND as modified gravity relevant Lagrangian can be written as a single integral over volume.) This prohibition is even more stringent from a relativistic standpoint [33]. Accordingly Milgrom What if, instead of the modified inertia interpretation, introduced a nonlocal action, i.e. one which is a func- we follow Milgrom’s alternative proposal [5] to regard tional of complete orbits, but cannot be reduced to an MOND as a modification of Newtonian gravity? This integral over a Lagrangian [32]. This approach does not entails a change in the r.h.s. of ma = f, with the new quite reproduce formula (1) generically, but that formula gravitational force possibly depending nonlinearly on its is recovered for the special case of circular orbits. sources. Can such implementation be protected from vi- olation of the conservation laws? Of course circular stellar orbits a la MOND is really all An affirmative answer to this query is readily avail- that one needs to analyze data on spiral galaxies. How- able by starting from a modification of the Lagrangian of ever, it is known that elliptical galaxies comprise highly Newtonian gravity with Milgrom’s a0 playing the role of radial stellar orbits too, yet elliptical galaxies also seem characteristic scale, to wit to be well described by MOND [34], and so the nonlo- 2 2 cal modified inertia approach might be found wanting a0 ∇Φ 3 here. On the plus side one may mention the Pioneer L = F | 2| + ρΦ d x. (3) − Z 8πG a0 anomaly [35]. The Pioneer 10 and Pioneer 11 spacecraft, h i as they travelled almost radially in the outer parts of Here F is some positive function, ρ is the matter’s mass the solar system, were found to be subject to an anoma- density and Φ, the field in the theory, is to be iden- lous sunward acceleration of order 8 10−10ms−2 (not tified with the gravitational potential that drives mo- × far from a0) which does not fall off measurably with dis- tion, i.e. a = ∇Φ. The Lagrangian’s kinetic part is tance from the sun. Were such an acceleration to reflect the most general− one that depends only on first deriva- a generic gravitational field, it would have affected the tives of Φ, and is consistent with the isotropy of space 6
(only ∇Φ appears). This Aquadratic Lagrangian the- regarded as structureless. AQUAL predicts that the cen- ory (AQUAL)| | [40] reduces to Newton’s in the limit ter of mass (CM) of a collection of particles, no matter F (X) X, when it readily leads to Poisson’s equation what the internal gravitational field, will follow the orbit for Φ.→ More generically it yields the equation determined by a = ∇Φ with Φ the external potential coming from the AQUAL− equation [40]. As mentioned, ∇ [µ( ∇Φ /a0)∇Φ] = 4πGρ, (4) this acceleration will approximate that predicted by the · | | µ(√X) F ′(X). (5) MOND equation. AQUAL does indeed supply a physical ≡ basis for MOND. One identifies the µ function here with MOND’s µ. A system with sub-a0 internal gravitational field im- Comparison of the AQUAL equation (4) with Poisson’s mersed in an external field ge > a0 is predicted by for the same ρ shows that AQUAL to exhibit quasi-Newtonian internal dynam- ics [40]. Originally Milgrom conjectured this “external µ( ∇Φ /a0)∇Φ= ∇Φ + ∇ h, (6) | | N × field effect” to explain why no acceleration discrepancy is evident in the loose open star clusters residing in the where h is some calculable vector field. Now in highly Milky Way [5]: the galaxy’s field, still stronger than a0 in symmetric situations (spherical, cylindrical or plane sym- its inner parts, suppresses the MOND behavior expected metry), use of Gauss’ theorem with the AQUAL equation in light of the weak internal fields. By contrast, if the said shows that ∇ h has to vanish. In light of the relation system is immersed in an external field g < a0, which a = ∇Φ, AQUAL× then reproduces the MOND for- e is, however, stronger than its internal fields, the internal mula− (1) exactly. However, the curl term does not usu- dynamics is predicted to be quasi-Newtonian, but with ally vanish in situations with lower symmetry (Milgrom the effective gravitational constant G/µ(g /a0) [40]. and Brada exhibit a curious exception [41]). We conclude e that the MOND formula is generically an approximation A useful product of AQUAL is a formula for the force to the solution of the AQUAL equation. between two masses in motion about each other, e.g. a How good an approximation? Work by Milgrom and binary galaxy. The MOND formula cannot be used di- Brada [41, 42] has shown that often the curl is just a rectly here because, as mentioned in Sec. 4, it would sug- 5-10% correction. This explains how the MOND for- gest that the binary’s center of mass should accelerate mula manages to be so successful while not constituting (nonconservation of momentum). Instead, Φ has to be ∇ by itself a physically consistent theory. Now almost all computed from the AQUAL equation, and then ρ Φ MOND fits to observed RCs are calculated using that has to be integrated over one galaxy’s volume to obtain formula exclusively. It should thus be clear that modest the force on it, as effectively done by Milgrom [42]. For point masses m1 and m2 separated by distance r, the departures of empirical RCs from MOND’s predictions 2 force is (Gm1m2/r )f(m2/m1,r/rt) with rt [G(m1 + do not constitute evidence against the paradigm. The 1/2 ≡ era of RC modeling with the full AQUAL Eq. (4) is still m2)/a0] ; Milgrom gives the graph of f. In the New- tonian limit (r r ), f 1. In the deep MOND limit in its early stages [43]. ≪ t ≈ I mentioned in Sec. 4 that the MOND formula is in- (r rt), Milgrom [26] shows analytically that the force ≫ 1/2 −1 3/2 3/2 3/2 consistent with the conservation laws. The AQUAl the- is (2/3)(Ga0) r [(m1 + m2) m1 m2 ]. This − − ory is free of this problem. This needs no special proof is most elegantly computed by exploiting the discovery (although explicit proofs exist [40]): a theory based on that in that limit the AQUAL equation is conformally a Lagrangian which does not depend explicitly on coor- invariant, an approach that also yields the exact virial dinates, directions and time (all true for AQUAL’s La- relation in the deep MOND regime [44]. grangian) automatically conserves momentum, angular AQUAL permits a study of galaxy disc stability in the momentum and energy. At the level of the first integral MOND paradigm. I mentioned in Sec. I that Newto- (6) of the AQUAL equation, it is the curl term which nian disk instabilities motivated the idea that DM halos protects AQUAL gravity from violating the conservation surround disk galaxies. In the MOND paradigm there laws. is not DM, so how do a fraction of the spiral galax- AQUAL consistently embodies the weak equivalence ies in the sky avoid forming bars? Using a mixture of principle (all bodies, regardless of inner structure, and MOND formula and AQUAL equation arguments, Mil- starting from the same initial conditions, follow the same grom [45] showed analytically that MOND enhances lo- path in a gravitational field). As remarked by Mil- cal stability of disks against perturbations as compared grom [5], the MOND formula is ambiguous in this re- to the Newtonian situation. Brada and Milgrom [46] spect. In a star accelerations of individual ions greatly combined an N-body code with a numerical solver for exceed a0 in magnitude. Thus for each ion µ should be the AQUAL equation to verify this and to demonstrate very close to unity, and we would guess that µ 1 for that MOND enhances global stability. Both works con- the star as a whole. Accordingly, if regarded as≈ a col- clude that the degree of stabilization saturates in the lection of ions, the star would follow a Newtonian orbit deep MOND regime; this means no disk in MOND is ab- in the galaxy’s field. This conclusion is entirely at vari- solutely stable. Christodoulou [47] and Griv and Zhyt- ance with the evidence from the flat RCs, or from the nikov [48] reach similar conclusions with N-body simula- dictate of the MOND formula when applied to the star tions based on just the MOND formula. Recently Tiret 7 and Combes [49] studied evolution of bars in spiral disks the core of Fornax has a flat density profile, in contradic- using N-body simulations built on the AQUAL equa- tion to the DM simulations. In fact, there is other grow- tion. They find that bars actually form more rapidly ing evidence negating DM density cusps [4, 55]. Now, in MOND than in DM halo Newtonian models, but that some of the dwarf spheroidal satellites of the Galaxy in- MOND bars then weaken as compared to those in Newto- cluding Fornax are in the deep MOND regime. Following nian models. The long term distribution of bar strengths Binney and Ciotti [50], S´anchez-Salcedo, Reyes-Iturbide predicted by MOND is in better agreement with the ob- and Hernandez [56] point out that the mentioned speedup served distribution than that obtained from Newtonian of settling by dynamical friction in AQUAL should long halo models. ago have caused the globular clusters in Fornax to amal- AQUAL is also important for studies of two-body re- gamate at its center. AQUAL thus has a problem ex- laxation in stellar systems. Two-body relaxation is the plaining the observations of Fornax; however, as we see process whereby distant gravitational encounters of pairs the DM alternative has its problems too. of stars in a galaxy or a cluster redistribute the system’s Can MOND give a picture of the late stages of galaxy energy and help it to approach some sort of equilibrium. formation? A step in this direction was taken by Nipoti, In Newtonian theory two-body relaxation is excruciat- Londrillo and Ciotti [57] on the basis their earlier work ingly slow. But as Binney and Ciotti show [50], two-body with N-body codes incorporating the full AQUAL equa- relaxation in the framework of AQUAL proceeds faster. tion [43]. They simulate the dissipationless collapse of a In fact, if the system is deep in the MOND regime, the cloud of particles which ends up in the MOND or deep speedup factor is the square of the acceleration discrep- MOND regimes. They note that phase mixing (the dis- ancy. appearance of structure in the initial conditions due to A related effect is dynamical friction, whereby a mas- commingling of particle orbits) proceeds slower than in sive object moving through a collection of lighter bodies, the counterpart Newtonian simulations, and that the fi- e.g. a globular star cluster coursing among the stars of nal velocity distribution is anisotropic and favors radial its mother galaxy, loses energy by virtue of its kicking orbits. Current thought in cosmogony is that elliptical the background objects gravitationally as it sweeps by and spheroidal galaxies are the outcome of anisotropic them. As a result the heavy object settles in the grav- collapse after star formation is complete. Nipoti et al. itational well confining it. In AQUAL for systems deep note that the end-products of some of their simulations in the MOND regime, dynamical friction is speeded up resemble, in their surface density and velocity dispersion over the Newtonian value by one power of the accelera- profiles, the dwarf-elliptical and dwarf-spheroidal galax- tion discrepancy [50]. ies in the sky. But other final states have trouble fitting in Both of the above phenomena open a new window all the accepted empirical correlations between luminos- onto the MOND-Newtonian gravity dichotomy. Suppose ity, radius and velocity dispersion of elliptical galaxies. AQUAL is the correct nonrelativistic description of grav- ity. In regard to RCs of spiral galaxies, Newtonian the- ory with DM could be “saved” by imagining the DM in 6. Can MOND be made relativistic? each galaxy to be distributed in just such a way that the resulting total ΦN is just the Φ generated in AQUAL Why is a relativistic version of MOND important? Al- theory by the baryonic (visible) matter. Admittedly, though most astronomical systems with significant ac- the required DM distribution might be peculiar (nega- celeration discrepancy are nonrelativistic in the extreme, tive density somewhere [51], or incompatible with cos- there are two important exceptions. Cosmology, which mogonic simulations). But leaving these options aside, is important in itself and as a framework for the process observational uncertainties might make it difficult to tell of galaxy formation, is widely regarded as requiring the the two theories apart, as recent fits of RCs of nearby presence of much DM on the largest scales, and cosmol- galaxies based on high quality data testify [52]. Yet con- ogy is universally believed to require relativistic treat- sideration of the mentioned dissipative effects might re- ment. Gravitational lensing by galaxies and clusters of move the effective degeneracy between the two models of galaxies has become a salient astronomical tool in the a galaxy offered by the two theories. last two decades, and, since it discloses acceleration dis- In view of the generic prediction of Newtonian cos- crepancies, it is commonly regarded as supporting the mogonic simulations that the DM distribution in a galaxy need for DM. Since light propagates with speed c, non- rises to a cusp at the center (see Sec. I), Newtonian dy- relativistic MOND cannot even begin to formulate the namical friction should have caused the globular clus- problem of gravitational lensing. ters of many a dwarf spheroidal galaxy to already have The above cases drove the relativistic formulation merged at its center and there lost their integrity [53]. of AQUAL [40], which I shall refer to as RAQUAL. While globular clusters are indeed absent from several RAQUAL retains Einstein’s equations of GR as a tool such dwarf companions of the Milky Way, e.g. Draco, for deriving the spacetime metric gαβ from the energy- Sextans and Carina [54], the Fornax dwarf spheroidal has momentum tensor of the matter (one of whose compo- five globular clusters near, but not at, its center. Goerdt, nents represents matter density). RAQUAL further stip- Moore, et al. [53] regard this as showing that the DM near ulates that matter and radiation play out their dynamics, 8
2 2ψ/c not in the arena of gαβ, but in that ofg ˜αβ = e gαβ, nonrelativistic limit inherits the good traits of MOND by where ψ denotes a scalar field with dimensions of squared way of AQUAL. velocity. Since all matter is treated alike in this respect, However, on the relativistic front RAQUAL hit two the exquisitely tested weak equivalence principle is safe- roadblocks. Already the original paper [40] remarks that guarded, but since the metric governing gravitational perturbations of ψ from a static background for which field dynamics is different from the one experienced by ∇ψ a0, e.g. ψ waves travelling in the reaches of a matter, the strong equivalence principle fails. This is es- galaxy,| | ≪ propagate superluminally, i.e., outside the light sential, for otherwise we should end up back with GR cone ofg ˜αβ. Nowadays one hears the view that superlu- as the relativistic gravity theory, and GR in the nonrel- minality by itself is not sufficient reason for causal prob- ativistic limit leads to Newtonian theory, not to MOND. lems [58]. This is by no means the standard view, and so How is ψ determined in RAQUAL? One takes the La- the specter of acausality motivated disillusionment with grangian for ψ to be the covariant version of that for RAQUAL [59, 60], and formulation of a two-scalar field 2 Φ in AQUAL, Eq. (3). This means replacing ∇Φ theory, phase-coupled gravity (PCG) [59, 61] intended αβ 3 1/2 3 | | 7→ g ∂αψ ∂βψ as well as d x ( g) d x, where g is the to prevent superluminality. In the PCG Lagrangian the 7→ − determinant of the matrix of components of gαβ. But in new scalar σ contributes a quadratic kinetic part, just as RAQUAL one does not include a term in lieu of ρΦ; the does ψ; there is also a simple coupling between σ and coupling between ψ and matter, which is to provide the the kinetic term for ψ which gives the theory its MOND source of the equation for ψ in accordance with its La- character. Although PCG generates an intolerable drift 2 grangian, is automatically generated by the eψ/c factor of the Kepler constant a3/P 2 in the solar system, and in the metricg ˜αβ with which we built the matter’s La- marginally contradicts the observed precession of Mer- grangian. For time independent situations with nonrela- cury’s perihelion [59], the theory has merits for both tivistic matter sources, the ψ equation reduces to Eq. (4) galaxy dynamics and cosmology [62, 63]. However it may with Φ ψ. However, it would be a mistake to conclude be, RAQUAL and PCG were both finally disqualified as that in7→ RAQUAL ψ is the gravitational potential. consistent representations of gravity by their inability to In GR for weak gravity and nonrelativistic motion, the cope with observations of gravitational lenses. potential in which matter moves, ΦN , is related to the Discovered in 1979, gravitational lenses have been in- 1 2 metric by ΦN = (gtt + c ) (t is the time coordinate). tensively studied ever since. In strong gravitational lens- − 2 Because, in this linearized theory, the most relevant Ein- ing a cluster of galaxies (but sometimes a single galaxy) stein equation reduces to Poisson’s, ΦN coincides with forms a few images of a single quasar in the background the usual Newtonian potential. In RAQUAL gαβ still by bending the incoming light rays with its gravita- satisfies Einstein-like equations, so the above relation ap- tional field. This provided a novel tool for determina- plies as well. But the stipulation that matter’s dynam- tion of masses of large astronomical systems by an ele- ics go forward in the metricg ˜αβ determines the nonrel- mentary use of GR. By the late 1980’s it was apparent ativistic potential in which matter moves, now called Φ, that masses of clusters so determined were larger than 1 2 to be Φ = 2 (˜gtt + c ). Writing in linearized theory the observed baryonic mass in them, and commensurate − 2 2 g˜tt = (1+2ψ/c ) gtt, we have Φ = ΦN gttψ/c . Now with the masses determined from motions of the galaxies 2 − to first approximation gtt = c ; thus to leading order comprising them. The novelty was that the acceleration − discrepancy was now put in evidence by a relativistic Φ=ΦN + ψ. (7) phenomenon. Many pounced upon the finding as added support for DM’s presence in clusters. Consequently, in RAQUAL the gravitational potential RAQUAL cannot deal with the new evidence; the rea- traced by the matter’s motions is the sum of the Newto- son is instructive. Maxwell’s equations are known to nian potential and the AQUAL scalar field, both gener- be conformally invariant: replacement of the metric gαβ γ ated by the same baryonic matter. The ψ does the job with which they are formulated by f(x ) gαβ does not of the DM’s potential in conventional approaches, but its change the form of the equations. It follows that study source is ordinary matter. of light propagation cannot distinguish between the met- Let us compare RAQUAL with the successful AQUAL. rics gαβ andg ˜αβ of RAQUAL. Accordingly, although We take the µ in Eq. (4) with Φ ψ to be approximately Maxwell’s equations are supposed to be written with 7→ equal to its argument for ∇ψ a0, and to level off g˜ , the metric for studying gravitational lensing might | | ≪ αβ at a value µ0 when ∇ψ a0. Then for strong ∇ψ just as well be g . But this last obeys Einstein’s equa- | | ≫ | | αβ Poisson’s equations gives ψ = ΦN /µ0 so that the true tions whose sources are the baryonic matter’s energy- potential is Φ = (1 + 1/µ0)ΦN . This means gravity is momentum tensor and that of ψ. However, for systems Newtonian with effective gravitational constant GN = as distended as clusters or galaxies, this last is negligi- (1+1/µ0)G. For ∇ψ a0, the RAQUAL equation tells ble. The light bending, then, is predicted by RAQUAL | | ≪ us that ∇ψ is much stronger than ∇ΦN . Consequently, to be nearly the same as predicted by GR without DM. by Eq. (7), ∇Φ ∇ψ so that in the weak field regime Yet there is observational evidence that the inner parts RAQUAL’s gravitational≈ field is close to AQUAL’s. Thus of clusters gravitationally lense light more strongly than in RAQUAL we have a relativistic theory which in the would be expected from GR with no DM. The same prob- 9 lem recurs in PCG, in which the two metrics are, again, guarantees that the weak equivalence principle will be conformally related in the same manner as in RAQUAL. obeyed precisely. The conventional GR Einstein-Hilbert It is thus quite clear that if a scalar field is to be the Lagrangian is used to give dynamics to gαβ, which then vehicle for MOND effects, then regardless of the form trivially induce dynamics forg ˜αβ. The scalar field φ is of its dynamics, the relation between the metricsg ˜αβ provided with a RAQUAL type Lagrangian. For conve- and gαβ must be non-conformal. I attempted [64] to nience this is couched in terms of a quadratic form in the construct such a coupling out of scalar field alone, to wit 4-gradient of φ in interaction with a second scalar field,
2 σ, which also occurs in a potential-like term. (However, 2ψ/c g˜αβ = e (Agαβ + B ψ,α ψ,β), (8) no derivatives of σ enter, so σ is not dynamical, and if eliminated, the explicit aquadratic Lagrangian for φ αβ with A and B functions of the invariant g φ,α φ,β. How- shows up.) The mentioned quadratic form is not the ever, Sanders and I found that avoidance of propagation αβ α β usual one, but rather (g )φ,α φ,β; it is intro- of gravitational waves which are superluminal with re- duced to forestall the superluminal−U U propagation that af- spect to the matter metricg ˜αβ requires that B < 0, while flicts RAQUAL. A > 0 is required so that both metrics have Lorentzian The kinetic part of the α Lagrangian is quadratic in signatures [65]. We then calculated that for equal sources its first derivatives, and exactlyU analogous to that for a of Einstein’s equations, the light bending in the proposed gauge field. This is not the only quadratic form possible, modification of RAQUAL is weaker than that in GR. As but it is the one in which the Einstein equations for gαβ mentioned, in systems of the scale of galaxies and clus- will not have second derivatives of α in their sources. As ters of galaxies, the scalar field provides very little energy U mentioned, it is imperative that α be a timelike vector. or momentum density as compared to the matter. Thus In TeVe S this is accomplished byU introducing, alongside in regard to light ray bending, the new theory performs the kinetic term, a Lagrange multiplier term of the form worse than RAQUAL. α λ( α + 1) which forces the norm of α to be negative Actually the above conclusion could be avoided if ψ,α is (andU U to be of unit length as a bonus).U Uα a timelike 4-vector. However, we would expect that deep Whereas RAQUAL is a one parameter (a0) theory with inside or near a galaxy, or a massive cluster of galaxies one free function (µ), TeVe S has one free function, F (σ2) which is certainly virialized, the scalar field, whose prin- [distinct from the F in Eq. (3)] that determines the shape cipal source is the matter therein, should be quasistatic. of the potential-like term, and three parameters in addi- Then ψ,α is expected to be spacelike, i.e. to give lensing tion. One of these is a scale of length ℓ in the φ La- weaker than needed. The problem is thus still present. grangian that sets the strength of the potential, while To overcome it Sanders [66] proposed to replace ψ,α in a dimensionless one, k, determines the scale of its argu- its above role by a nondynamical vector α which in an ment. The third parameter, K, also dimensionless, sets isotropic cosmological model points preciselyU in the time the strength of the α Lagrangian. One can form the direction, and approximately so in the presence of mass acceleration scale U concentrations. He further related the metrics by √ 2 2 2 2 3kc g˜ = e−2φ/c g (e2φ/c e−2φ/c ) . (9) a0 = , (10) αβ αβ − − Uα Uβ 4πℓ Sanders further supposed the action for the scalar field which parallels the role of Milgrom’s constant in MOND φ, which supersedes ψ, to be of RAQUAL form, or more and AQUAL. It is found that the limit of TeVe S with generally with the function F depending also on the k 0, K k and ℓ k−3/2 amounts to GR. In fact α → ∝ ∝ scalar φ,α. The timelike nature of the vector is con- a0 0 in this limit, so that all accelerations are to be duciveU to enhancement of the lensing, and such a theory, regarded→ as strong, and gravitation as conventional. named stratified theory, can cope with the observations of How about nonrelativistic motion? Let φc be the co- gravitational lensing. However, it was clear to Sanders eval value of φ in the cosmological model in which the that this is just a toy theory since a globally timelike nonrelativistic system is embedded. This will be the vector field, such as α, determines a universal preferred asymptotic boundary value of φ for any local solution frame of reference, anU aether. Additionally, since the vec- of the φ equation. From a linearized version of Einstein’s tor field is supposed constant, the theory is not covariant. equations, and use of the relation (9), it follows [67] that the effective nonrelativistic gravitational potential is
7. TeVeS and other relativistic MOND theories Φ = ΞΦN + φ, (11) 2 Ξ (1 K/2)−1 e−2φc/c (12) The mentioned problems are avoided by making the ≡ − vector field a dynamical one, as first done in TeVe S, (this corrects a sign error in Refs. 67 and 68). Here ΦN my tensor-vector-scalar theory of gravity [67, 68, 69]. comes from the usual Poisson equation while φ is found TeVe S retains the relation (9) between a gravitational from the scalar equation; in both the source is the bary- metric gαβ and a physical (or observable) metricg ˜αβ. onic mass density ρ. Thus far all studies have assumed The matter Lagrangian is built exclusively withg ˜ ; this that 0 < K 1 and φ c2, in which case Eq. (11) αβ ≪ | c| ≪ 10 essentially coincides with Eq. (7) in RAQUAL. For this emerges an Einstein-Aether like theory in which the met- same reason any time variation of Ξ [70] should be neg- ricg ˜αβ also satisfies Einstein-like equations. However, in ligible. This is of particular significance for the above contrast to orthodox Einstein-Aether theories, the vec- derived value of a0 which, strictly speaking, includes a tor kinetic action in the theory in question is a generic factor Ξ [for example Eq. (6.7) of Ref. 68]. quadratic form in the vector’s derivatives contracted into For a static situation the equation for φ takes a form polynomials of the vector’s components. This form of like Eq. (4), with µ (essentially σ2) expressed in terms of TeVe S is more complex than the original one, but there ∇φ through TeVe S’s free function F (σ2). By a suitable are some circumstances in which it may be more conve- |choice| of F one can reproduce any µ function that would nient in ferreting out consequences of the theory. be relevant in AQUAL, with the scale thereon agreeing In TeVe S in either form the vector is normalized with with a0 in Eq. (10). By analogy with our discussion respect to gαβ, notg ˜αβ. Zlosnik, et al. have also proposed of RAQUAL we see that TeVe S can have MOND phe- a variant tensor-vector theory with a timelike vector nor- nomenology and also a Newtonian limit. The transition malized with respect tog ˜αβ, and which also has MOND between Newtonian and MOND regimes takes place as like behavior [79]. This is constructed by taking the vec- ∇φ sweeps through the value a0, near the point where tor’s action as a function of , the quadratic form in | | ∇Φ a0. the derivatives of the vectorF fieldK of orthodox Einstein- | |≈ Actually for systems strongly departing from spheric- Aether theories. The Zlosnik, et al. theory has four pa- ity, the situation is not as clear, but it seems that at rameters: a length scale and three dimensionless parame- the nonrelativistic level TeVe S implements the MOND ters. The form of the free function ( ) can be deduced paradigm overall. When it comes to details, my origi- approximately from the requirementF thatK MOND arise in nal choice for F [67] can be criticized: while it leads to the nonrelativistic quasistatic limit, and from the stipu- flat outer RCs, it does not give a satisfactory account of lation that a cosmology built on this theory shall have an the transition part of the RCs of several well measured early inflationary period and an accelerated expansion at galaxies, including our own [71, 72]. By contrast, the late times. MOND formula with the “simple” variant of Migrom’s In summary, the relativistic implementations of µ (FIG. 3) leads to very good fits for many RC’s [70]. MOND involve either one or three free functions; those Forms of the TeVe S function F which would do a cor- with one free function have either three or four free pa- respondingly good job have been proposed by Zhao and rameters. Famaey [73], Sanders [72] and Famaey, et al. [70]. Sanders has also proposed a variation on TeVe S in- volving a second scalar σ with PCG-like dynamics in- stead of RAQUAL ones [74]. This bi-scalar tensor vec- 8. TeVe S and gravitational lensing tor theory (BSTV) has three free functions and a free parameter. The theory is not especially needed to obvi- ate superluminal propagation since TeVe S seems to do As stressed earlier, the particular structure of TeVe S well on this [67]. However, BSTV, like PCG, is a more reflects the desire to encompass in the MOND paradigm appropriate frame for generating cosmological evolution the observation that what passes, from a dynamical point of a0. Since numerically a0 cH0, it is often argued of view, for DM in galaxies and clusters of galaxies also ∼ that a0 must be determined by cosmology, and should lenses light to a commensurate degree. How does TeVe S thus vary on a Hubble timescale [5, 10]. As mentioned, measure up to the task? Now the measured gravitational a0 in TeVe S is essentially set by the parameters ℓ and k; lensing by galaxies and clusters of galaxies takes place it should be nearly constant in the expanding universe. over cosmological distances. The light rays in TeVe S It is otherwise in BSTV. Discrimination between con- are null geodesics of the observable metricg ˜αβ; to com- stant and evolving a0 may be possible with good RCs of puteg ˜αβ one needs Uα, φ and gαβ. Isotropic cosmologi- disk galaxies at redshifts z = 2—5. Such curves are just cal models in TeVe S closely resemble the corresponding now coming into range. In fact, the data in Ref. 75 put cosmological models of GR; they sport a Uα pointed pre- the z = 2.38 galaxy BzK-15504 right on the MOND de- cisely in the time direction and feature very slow change rived Tully-Fisher law with the standard value for a0 [76]. of φ [67]. A consequence is thatg ˜αβ inaTeVe S isotropic Thus the meagre data available today are consistent with cosmological model differs little from the metric in the no a0 evolutions on the Hubble timescale. corresponding GR cosmology. Thus the cosmological Recently Zlosnik, Ferreira and Starkman [77] have facet of gravitational lensing is very much like in GR. clarified the relation between TeVe S and the so called The second facet is the local bending of light rays in Einstein-Aether gravity theories in which a timelike unit a mass’ vicinity. As in isotropic cosmological models, so vector field (but no scalar) plays a role alongside the met- in static situations, the solution of the vector’s equation ric [78]. They do this by re-expressing TeVe S solely has Uα pointed precisely in the time direction. To com- in terms of the observable metricg ˜αβ and the vector pute the bending in linearized theory, the approximation field. The φ is eliminated with the help of the constraint commonly used in the business, one also needs the scalar imposed through the Lagrange multiplier term. There field φ and the metric gαβ, both to first order in ΦN . The 11
final result is that the line element takes the form well correlated with the luminosities of the galaxies. The corresponding mass-to-light ratios are found to be in the ds˜2 = (1+2Φ/c2)dt2 + (1 2Φ/c2)(dx2 + dy2 + dz2) − − normal range for stellar populations, with some excep- (13) tions. with Φ given by Eq. (11). Note that the same poten- How frequently should strong lensing occur? This tial Φ appears in both terms of this isotropic form of the question is taken up in the context of TeVe S by Chen and line element. Hence light ray bending, which leans on Zhao [84]. Again modeling the mostly elliptical galaxies both to equal degree, measures the same gravitational with Hernquist profiles, they compute the probability of potential as do dynamics which are sensitive only to the two images occurring as a function of their separation. temporal part of the line element. This mirrors the sit- Their prediction falls somewhat beneath the frequency uation in GR whose line element is obtained by sending observed in the lensing surveys, though they consider this Φ ΦN in Eq. (13). Thus half of the problem that still acceptable. The predictions are sensitive to the as- plagues7→ RAQUAL and similar theories is overcome: the sumed mass profile, as well as to the assumed shape of acceleration discrepancy makes itself felt equally through the µ function, which Chen and Zhao assume to switch the gravitational lensing as through the dynamics. brusquely from linear in the argument to unity. The second half of the problem revolves about the mass distribution that generates Φ. In GR Φ is all there is, and its Laplacian, as determined from the lensing obser- vations or from the dynamics, will give the total (bary- onic plus dark) mass distribution directly. Because DM is not seen directly, this prediction can only be judged by the plausibility of the derived distribution of DM. By contrast in TeVe S the measured Φ is to be decomposed into two parts in the manner of Eq. (11), with each gen- erated by the same baryonic mass density ρ, one part through Poisson’s equation, and the second through a highly nonlinear AQUAL type equation. Evidently, in a TeVe S model of a gravitational lense, the baryonic mat- ter will be distributed differently from the total matter in a GR model of the same lense. And when the lensing system is not spherically symmetric, the centers of the two distributions may be offset. FIG. 6: The colliding clusters 1E0657-56. The bullet clus- ter (right) rammed through the cluster on the left. Hot gas Chiu, Ko and Tian [80] have explored light ray bend- stripped off both clusters is colored red-yellow. Green and ing by a pointlike mass M in TeVe S. They note that white curves are level surfaces of gravitational lensing conver- the deflection angle in the deep MOND regime [impact 1 2 gence; the two peaks of this do not coincide with those of the parameter b b0 (k/4π)(GM/a0) / ] approaches a ≫ ≡ gas which makes up most of the visible mass, but are skewed constant, as might have been expected from naive argu- in the direction of the galaxy concentrations. The white bar ments, but is less predictable in the intermediate regime corresponds to 200 kpc. Figure reproduced from Ref. 85 by b b0. Thus calculations of lensing based on a mixture permission of the American Astronomical Society. of∼ MOND and GR motives [81, 82] can easily mislead. Chiu et al. work out the lens equation in TeVe S, which Gravitational lensing by the colliding galaxy clusters controls the amplifications of the various images in the 1E0657-56 has been claimed to give theory independent strong lensing of a distant source, and remark that for proof of DM dominance at large scales [82]. In this sys- two images the difference in amplifications is no longer tem (FIG. 6) a smaller cluster, the “bullet”, has crashed unity as in GR, and may depend on the masses. With a through a larger one and the intracluster gas of both has photometric survey like the Sloan Digital Survey it may been stripped by the collision, the bullet’s gas trailing be- be possible to check for this effect. Finally, these authors hind its galaxy component. Weak lensing (distorted but work out the gravitational time delay in TeVe S, which unsplit images) mapping shows the lensing mass to be could be of use in interpreting differential time delays in concentrated in the two regions containing the galaxies, doubly imaged variable quasars. rather than in the two clouds of stripped gas which con- More phenomenologically oriented, Zhao, Bacon, Tay- tains the lion’s share of baryonic mass [82, 85]. Collision- lor and Horne [83] compare TeVe S predictions with a less DM would indeed move together with the galaxies. large sample of galaxy strong lenses which each produce Hence the inference that much DM continues to accom- two images of a quasar. They model the galaxy baryonic pany the bullet. Angus, Famaey and Zhao [86] (see also mass distribution either by point masses or by the pop- Ref. 73) point out that in the very asymmetric system ular Hernquist profile. Galaxy masses are estimated by 1E0657-56, MOND, AQUAL and TeVe S all predict sub- comparing observations both with predicted image posi- stantially different gravitational field distributions (com- tions and with predicted amplifications ratios; the two pare Eqs. (1) and (6)), a situation which confuses Clowe methods are found to give consistent results, themselves et al.’s “theory independent” inference. 12
Whereas Angus, Famaey and Zhao consider it pos- back mechanism involving the scalar field perturbations. sible to explain the lensing with a reasonable purely Dodelson and Liguori [92] have independently calculated baryonic matter distribution, a later paper by Angus, perturbation growth, and stressed that it is rather the Shan, et al. [87] concludes that dark matter is needed vector field in TeVe S which is responsible for growth of after all. This is hardly surprising; as we saw in Sec. 3, large scale structure without needing DM for this [92]. It pure MOND does not fully account for the acceleration thus seems there is potential in TeVe S to do away with discrepancy in the dynamics of quiescent galaxy clus- cosmic DM, as well as with DM in galaxies. ters [10]. But DM models of the bullet clusters within GR Much attention is commanded today by the mystery are not without their problems. Farrar and Rosen [88] of the “dark energy”, the agent responsible for the ob- note that the relative velocity of the clusters is too high served acceleration of the Hubble expansion. Such agent as compared to those seen in DM simulations of struc- is obligatory in the context of GR cosmological models. ture formation. To remove the contradiction they pro- Quite naturally many have wondered whether TeVe S pose that a non-gravitational attraction of a new sort could obviate this need. Diaz-Rivera, Samushia, and Ra- acts only between clumps of DM. But is assuming ex- tra [93] have found exact deSitter solutions of TeVe S istence of DM together with a new interaction specific cosmology which can represent either early time infla- to it more parsimonious than a modification of standard tion epochs or the late time acceleration era. In another gravity such as MOND? TeVe S study, Hao and Akhoury [94] have concluded that with a suitable choice of the TeVe S function F , the scalar field can play the role of dark energy. And Zhao [95] 9. TeVe S and cosmology maintains that with Zhao and Famaey’s choice of F [73], cosmological models can be had that evolve at early times With my original choice of F , TeVe S cosmologi- like those of standard cold dark matter cosmology, and cal models with baryonic matter content alone can be display late time acceleration with the correct present very similar to the corresponding GR models by virtue Hubble scale, all this without assuming DM or dark en- of the scalar φ’s energy density remaining relatively ergy. For the related Einstein-Aether theory, Zlosnik et small [67, 80, 83]. In particular Chiu et al. [80] note that al. [79] remark that with suitable choice of their theory’s , the vector field can both drive early inflation as well these TeVe S models give a reasonable relation between F redshift and angular distances, and are thus as effective as double for dark energy at late times. as GR models in providing the scaffolding for the analysis of cosmologically distant gravitational lenses. How would changing F affect the evolution of a cosmological model? 10. Assessment of relativistic MOND This is studied exhaustively by Bourliot et al. [89] follow- ing an earlier exploration by Skordis et al. [90]. Bourliot It may be unnecessary to go far from Earth to tell et al. display a large set of variants of my F for which, TeVe S, BSTV and similar theories apart from GR; the while the scalar energy density may not be negligible, it solar system can be turned into a sieve for the correct tracks or mimicks the behavior of another energy compo- modified gravity. Between any two bodies in the so- nent of the model, e.g. the radiation’s during radiation lar system there is an extremum point, strictly a saddle dominance. They also characterize shapes of F which point, of some relevant field. In AQUAL this would be can lead to future singularities in cosmology, and which the Φ potential and in TeVe S the φ field. Near such presumably should be avoided. a point the gradient of the said field is small, and de- MOND critics have always held up the complicated partures from Newtonian gravity are significant (µ in power spectrum of cosmological perturbations (back- AQUAL falls short of unity, and the derivative of F ground radiation or baryons) as a proof that DM is in TeVe S is driven away from the pole which signals needed on the cosmic level; after all the spectrum is said Newtonian behavior). A detailed study by Magueijo and to be well fit by the “concordance” DM model of the uni- me [96] with TeVe S shows that the anomalous regions, verse. This argument took for granted that MOND could e.g. between Sun and Jupiter and Earth and Moon, are never measure up to the test. With TeVe S on the scene small, but gives some hope that with the fine guidance of one can face the question technically. space probes like ESA’s LISA-Pathfinder (scheduled for In a massive work Skordis has provided the full co- liftoff in October 2009), it may be possible to diagnose variant formalism for evolution of cosmological pertur- MOND-like effects. bations in TeVe S [91]. And using this Skordis et al. [90] Well away from the saddle points, but still in the re- have shown that, without invoking dark matter, TeVe S gion occupied by the planets, TeVe S predicts small de- can be made consistent with the observed spectrum of partures of φ from 1/r form, i.e. small departures from the spatial distribution of galaxies and of the cosmic mi- Newtonian behavior as encoded in Φ [67]. Similar pre- crowave radiation if one allows for contributions by mas- dictions issue from BSTV [72]. These are constrained by sive neutrinos (in the still allowed mass range) and the the observed constancy of Kepler’s constant a3/P 2 out cosmological constant. In this approach the role of dark to the orbits of the major planets, and by the absence matter in standard cosmology is taken over by a feed- of significant departures from GR’s predictions for the 13 precessions of the perihelia of Mercury and the asteroid delay. Sanders [72] has provided a simple argument that Icarus. According to Sanders, TeVe S can satisfy these β and γ are always unity in a class of theories (including constraints with choices of the function the F proposed BSTV) which start from the disformal relation (9). by Refs. 72 and 73, which, as mentioned in Sec. 7, also Next in order of relevance are the preferred frame pa- lead to correct galaxy RCs. BSTV can do just as well rameters α1, α2, and α3. These vanish in GR; after all, when its extra parameter, ǫ, is chosen properly [72]. GR has no preferred frames. In TeVe S or BSTV at least Both TeVe S and BSTV with the above choices pre- one of the αs should be nonvanishing because the vector dict that beyond 100 astronomical units (AU) from the α establishes a locally preferred frame, with the vector sun, there is a fairly r-independent attractive component pointingU out the time direction in that frame. Sanders of the sun’s gravitational field of strength 0.3a0. The argues heuristically that the αs should be strongly sup- ∼ Pioneer anomaly, as measured at distances beyond 20 pressed in both theories [72]. But explicitly computing AU, is also r-independent but stronger. Is this evidence the α’s for relativistic MOND theories should be a high for MOND? Unfortunately, the interpretation of the Pio- priority because they are subject to tight experimental neer anomaly as an almost constant sunward acceleration bounds. seems to clash with limits on the variation of Kepler’s The odyssey in search of a relativistic embodiment constant set with spacecraft at the distance of Uranus of the MOND paradigm has led, not to one relativis- and Neptune [72, 97] and with the latest ephemerides tic MOND, but to many. TeVe S by itself exemplifies of the outer planets [36]. It thus seems prudent to sus- a family of theories. First there is the freedom in the pend judgment until a prosaic origin for Pioneer anomaly, choice of the function F , which is still only modestly e.g. drag by unknown matter, can be excluded with high constrained [71, 72, 73]. Next, the coefficient of the sec- probability. αβ α β ond term in the factor (g )φ,αφ β in the scalar’s The above focuses on the nonrelativistic limit. How Lagrangian can be changed−U to anyU other negative num- do relativistic MOND theories fare with regard to the ber. In TeVe S the Lagrangians for φ and α are formu- lowest order relativistic effects in the solar system? As U lated upon the metric gαβ; a different but related theory for any metric theory of gravity, these effects can be cal- emerges if one uses instead the background ofg ˜αβ. Ad- culated from the post-Newtonian (PPN) parameters of ditionally one can replace the aquadratic Lagrangian for the theory of choice. For any metric theory it is possi- φ by a PCG-style Lagrangian for two scalars, as indeed ble to parametrize the first and second order departures done in Sanders’ BSTV. Finally, one can altogether dis- of the metric from Minkowski form in terms of a set of pense with the scalar fields and go the way of the Zlosnik- ten intuitive looking potentials. The Newtonian poten- Ferreira-Starkman theory [79]. tial ΦN figures in the list, and occurs in the corrections While the lack of uniqueness of relativistic MOND is both linearly as well as squared; two others are a nuisance, it does add needed flexibility to the search ′ ′ ′ ′ ρ(r )ΦN (r ) 3 ′ ρ(r )vi(r ) 3 ′ for a “final” fundamental theory with which to underpin Φ2 = d r , V = d r , Z r r′ i Z r r′ the MOND paradigm. It is already clear that some of | − | | − | (14) the above mentioned theories may not be in full accord where v denotes the fluid velocity in the system. The with the facts. For example, the published TeVe S has an post-Newtonian (PPN) parameters are the dimensionless exact MOND limit at low accelerations, yet as mentioned coefficients multiplying the diverse potentials in the cor- in Sec. 3, MOND cannot handle the dynamics of clusters rection. Some examples will clarify this. of galaxies without invoking additional unseen matter. The correction to the space-space part of the metric, An interesting escape is suggested by Sanders [100]. In gij , is 2γΦN δij ; it defines the PPN parameter γ. The BSTV cosmology, just as in the PCG one [63], the σ field correction to the temporal metric component, gtt, starts can undergo oscillations which generate bosonic particles with the terms 2Φ 2βΦ 2, which define the PPN pa- early on. If the BSTV parameters are right, these bosons N − N rameter β, and also includes a Φ2 term whose coefficient can be trapped by clusters (but not by galaxies), and brings in a further PPN parameter, ζ1. Each time-space can thus comprise the additional unseen matter. The metric component gti gets corrections proportional to the charm of this resolution is that this new cold dark matter Vi and wiΦN (w denotes the velocity of the chosen co- emerges from the modified gravity theory itself, and is ordinate system with respect to the cosmological matter not a separate invention. comoving frame). With these last terms come two ad- Relativistic MOND as here described has developed ditional PPN parameters, α1 and α2, which are called from the ground up, rather than coming down from the preferred frame parameters. A third one, α3, is associ- sky: phenomenology, rather than pure theoretical ideas, 2 ated with a correction of the form w ΦN to gtt. There has been the main driver. Actually a large industry flour- are four more PPN parameters for a total of ten. ishes on the sidelines with imaginative ideas from first The parameters β and γ are both unity for GR and principles regarding the essence of MOND. I have not for some of it competitors. They have also been com- touched here on these motley approaches because they puted to be unity in TeVe S [67, 99], which thus fares have given so little that is observationally viable. Neither as well as GR in reference to gravitational light bending, have I dwelt here on modified gravity theories that are perihelia precessions of the planets, and the radar time not MOND motivated or oriented. But the time may be 14 ripe for turning to a more deductive approach to MOND. so, this would both ameliorate the common feeling that Now that we have some idea of what constitutes a viable TeVe S is unduly complicated, and point the way to a relativistic MOND theory, it should be easier to single lode for theories which might do better justice to the out theoretical frameworks which might yield a promis- observations in the spirit of the MOND paradigm. ing candidate for the fundamental MOND theory either as a limiting case, as an effective theory, after dimensional reduction, etc. To give an example, let us recall that Brans-Dicke mod- Acknowledgments ified gravity plus Maxwellian electromagnetism in the real world can both be recovered by dimensional reduc- tion of 5-D Einstein gravity theory. TeVe S has the same I thank Mordehai Milgrom, Bob Sanders and Stacy number of degrees of freedom as that pair of theories; McGaugh for many useful remarks on the original specifically, its vector field has three degrees of freedom manuscript, and the last two as well as Douglas Clowe on account of the normalization condition, and the elec- for providing figures. Research on this subject has been tromagnetic vector potential also has three on account of supported by grant 694/04 from the Israel Science Foun- gauge freedom. Might some variant of TeVe S arise from dation established by the Israel Academy of Sciences and dimensional reduction of a pure gravity theory in 5-D? If Humanities.
[1] Ostriker, J. P. and Peebles, P. J. E., 1973, Astroph. [22] Milgrom, M., 2002, Astroph. Jour., 571, L81. Jour., 186, 467. [23] McGaugh, S. S. and de Blok, W. J. G., 1998, Astrophys. [2] See sundry reviews in 1987, Dark Matter in the Uni- J., 499 66. verse, editors G. R. Knapp and J. Kormendy, (Dor- [24] McGaugh, S. S., 2006, in Mass Profiles and Shapes drecht: Reidel). of Cosmological Structures, edited by G. Mamon, F. [3] Khalil, S. and Mu˜noz, C., 2002, Contemp. Phys. 43, 51. Combes, C. Deffayet and B. Fort, (EAS Publication se- [4] Bertone G., Hooper D. and Silk J., 2004, Phys. Rep., ries, Vol. 20), ArXiv astro-ph/0510620 405, 279. [25] Scarpa, R., 2006, in First Crisis in Cosmology Con- [5] Milgrom, M., 1983, Astroph. Jour., 270, 365. ference, edited by E. J. Lerner and J. B. Almeida, [6] Milgrom, M., 1983, Astroph. Jour., 270, 371. (AIP Conference Proceedings, Vol. 822), p.253, ArXiv [7] Milgrom, M., 1983, Astroph. Jour. 270, 384. astro-ph/0601478. [8] Rubin, V. C., Ford, W. K., et al., 1982, Astroph. Jour., [26] Milgrom, M., 1994, Astrophys. J., 429, 540. 261, 439; Bosma, A., 1981, Astron, Journ., 86, 1791. [27] Pointecouteau, E. and Silk, J., 2005, Mon. Not. Roy. [9] Sofue, Y. and Rubin, V., 2001, Ann. Rev. Astron. As- Astron. Soc., 364, 654. trophys., 31, 137. [28] Sanders R. H., 2003, Mon. Not. Roy. Astron. Soc., 342, [10] Sanders R. H. and McGaugh, S. S., 2002, Ann. Rev. 901. Astron. Astroph., 40, 263. [29] Pointecouteau, E., 2006, ArXiv astro-ph/0607142. [11] McGaugh, S.S., 2006, ArXiv astro-ph/0606351. [30] Kaplinghat, M., and Turner, M. S., 2002, Astroph. [12] Tully, R. B. and Fisher, J. R., 1977, Astron. Astroph., Jour., 569, L19. 54, 661. [31] Sanders, R. H., and Begeman, K. G., 1994, Mon. Not. [13] Sanders, R. H., 1996, Astroph. Jour., 473, 117. Roy. Astron. Soc., 266, 360. [14] Bahcall, J. N. and Cassertano, S., 1985, Astroph. Jour. [32] Milgrom, M., 1994, Ann. Phys. 229, 384-415. 293, L7. [33] Domokos, G., 1996, preprint JHU-TIPAC-96009. [15] Zwicky F., 1933, Helv. Phys. Acta, 6, 110. [34] Milgrom, M. and Sanders, R. H., 2003, Astroph. Jour., [16] Zwicky F., 1957, Morphological Astronomy (Berlin: 599, L25-L28. Springer). [35] Anderson, J. D., Laing, P. A., et al. 1989, Phys. Rev. [17] Finzi, A., 1963, Mon. Not. Roy. Ast. Soc., 127, 21; Letters, 81 , 2858. Tohline, J. E., 1982, in Internal Kinematics and Dynam- [36] Iorio, L. and Giudice, G., 2006, New Astron. 11, 600, ics of Galaxies, edited by A. Athanassoula (Dordrecht: ArXiv gr-qc/0610050. Reidel), p.205; Sanders, R. H., 1984, Astron. Astroph., [37] Milgrom, M., 2006, in Mamon et al., Ref. 24; ArXiv 136, L21; and 1986, Astron. Astroph., 154, 135; Gold- astro-ph/0510117. man, I., 1986, Astron. Astroph., 170, L1; Kuhn, J. R. [38] Milgrom, M., 1999, Phys. Letters A, 253, 273. and Kruglyak, L., 1987, Astroph. Jour., 313, 1; and [39] Unruh, W. G., 1976, Phys. Rev. D, 14, 870. others. [40] Bekenstein, J. D. and Milgrom, M., 1984, Astroph. [18] Bell, E. F. and de Jong, R. S., 2001, Astroph. Jour., Jour., 286, 7. 550, 212. [41] Brada, R. and Milgrom, M., 1995, Mon. Not. Roy. As- [19] Mc Gaugh, S. S., 2005, Astroph. Jour., 632, 859. tron. Soc., 276, 453. [20] Barnes, E. I., Kosowsky, A. and Sellwood, J. A., 2006, [42] Milgrom, M., 1986, Astrophys. J., 302, 617. ArXiv astro-ph/0701535. [43] Ciotti, L., Londrillo, P., and Nipoti, C., 2006, Astro- [21] Begeman, K. G., Broeils, A. H. and Sanders, R. H., phys. J., 640, 741. 1991, Mon. Not. Roy. Astron. Soc., 249, 523. [44] Milgrom, M., 1997, Phys. Rev. E, 56, 1148. 15
[45] Milgrom, M., 1989, Astrophys. J., 338, 121. 2006, Phys. Rev. D74, 044037 (2006). [46] Brada, R. and Milgrom, M., 1999, Astrophys. J., 519, [78] Jacobson, T. and Mattingly, D., 2001, Phys. Rev. D64, 590. 024028. [47] Christodoulou, D., 1991, Astrophys. J., 372, 471. [79] Zlosnik, T. G., Ferreira, P. G. and Starkman, G. D., [48] Griv, E. and Zhytnikov, V. V., 1995, Astroph. Sp. Sci., 2006, ArXiv gr-qc/0607411. 226, 51. [80] Chiu, M. C., Ko, C. M. and Tian, Y., 2006, Astroph. [49] Tiret, O. and Combes, F., ArXiv astro-ph/0701011. Jour., 636, 565. [50] Binney, J. and Ciotti, L., 2004, Mon. Not. Roy. Astron. [81] Mortlock, D. J. and Turner, E. L., 2001, Mon. Not. Roy. Soc., 351, 285. Ast. Soc., 327, 557. [51] Milgrom, M., 1986, Astrophys. J., 306, 9. [82] Clowe, D., Gonzalez, A. and Markevitch, M., 2004, As- [52] Corbelli, E. and Salucci, P., 2006, ArXiv troph. Jour., 609 , 596. astro-ph/0610618. [83] Zhao, H. S., Bacon, D. J., Taylor, A. N. and Horne, K., [53] Goerdt, T., Moore, B., et al., 2006, Mon. Not. Roy. 2006, Mon. Not. Roy. Astron. Soc., 368, 171. Astron. Soc. 368, 1073. [84] Chen, D. M. and Zhao, H. S., 2006, Astroph. Jour., 650, [54] Mateo, M. L., 1998, Ann. Rev. Astron. Astroph., 36, L9. 435. [85] Clowe, D., Badrac, M., et al., 2006, Astroph. Jour., 648, [55] Evans, N. W., 2001, in IDM 2000: Third International L109. Conference on the Identification of Dark Matter, edited [86] Angus, G. W., Famaey, B. and Zhao, H. S., 2006, Mon. by N. Spooner and V. Kudryavtsev (Singapore: World Not. Roy. Astron. Soc. 371 138. Scientific), p. 85. [87] Angus, G. W., Shan, H. Y., Zhao, H. S. and Famaey, [56] S´anchez-Salcedo, F. J ., Reyes-Iturbide, J. and Hernan- B., 2006, ArXiv astro-ph/0609125. dez, X., 2006, Mon. Not. Roy. Astron. Soc. 370, 1829. [88] G. R. Farrar and Rosen, R. A., 2006, ArXiv [57] Nipoti, C., Londrillo, P. and Ciotti, L., ArXiv astro-ph/0610298. astro-ph/0701418. [89] Bourliot, F., Ferreira, P.G., Mota, D. F. and Skordis, [58] Bruneton, J.-P., ArXiv gr-qc/0607055. C., ArXiv astro-ph/0611255. [59] Bekenstein, J. D. 1988, in Second Canadian Con- [90] Skordis, C., Mota, D. F., Ferreira, P. G. and Boehm, ference on General Relativity and Relativistic Astro- C., 2006, Phys. Rev. Lett. 96, 011301. physics, edited by A. Coley, C. Dyer and T. Tupper [91] Skordis, C, 2006, Phys. Rev. D74, 103513. (Singapore: World Scientific), p. 68. [92] Dodelson, S. and Liguori, M., 2006, Phys. Rev. Lett. [60] Bekenstein, J. D. 1990, in Developments in General Rel- 97, 231301. ativity, Astrophysics and Quantum Theory, edited by. [93] Diaz-Rivera, L. M., Sasmushia, L. and Ratra, B., 2006, F. I. Cooperstock, L. P. Horwitz and J. Rosen (Bristol: Phys. Rev. D73, 083503. IOP Publishing), 156. [94] Hao, J. G. and Akhoury, R, 2005, “Can Relativistic [61] Bekenstein, J. D., 1988, Physics Letters B, 202, 497. MOND Theory Resolve Both the Dark Matter and Dark [62] Sanders, R. H., 1988, Mon. Not. Roy. Ast. Soc., 235, Energy Paradigms?”, ArXiv astro-ph/0504130. 105. [95] Zhao, H. S., 2006, ArXiv astro-ph/0610056. [63] Sanders, R. H., 1989, Mon. Not. Roy. Ast. Soc., 241, [96] Bekenstein, J. D. and Magueijo, J, 2006, Phys. Rev. D, 135. 73, 103513. [64] Bekenstein, J. D. 1992, in Proceedings of the Sixth Mar- [97] Anderson J.D., Lau, E. L., et al., 1995, Astroph. Jour. cel Grossman Meeting on General Relativity, edited by 448, 885. H. Sato and T. Nakamura (Singapore: World Scientific), [98] Will, C., 1986, Theory and Experiment in Gravitational p. 905. Physics (Cambridge: Cambridge University Press). [65] Bekenstein, J. D. and Sanders, R. H., 1994, Astroph. [99] Giannios, D., 2005, Phys. Rev. D, 71, 103511. Jour., 429, 480. [100] Sanders, R. H., 2006, in Third Aegean School: The In- [66] Sanders, R. H., 1997, Astroph. Jour., 480, 492. visible Universe: Dark Matter and Dark Energy, ArXiv [67] Bekenstein, J. D., 2004, Phys. Rev. D 70, 083509; erra- astro-ph/0601431. tum 2005, Phys. Rev. D, 069901(E). [68] Bekenstein, J. D., 2004, Proc. Sci. JHW2004, 012, ArXiv astro-ph/0412652. Jacob Bekenstein received his Ph. D. (1972) [69] Bekenstein, J. D. and Sanders, R. H., 2006, in Mamon from Princeton University. He worked at the et al., Ref. 24; ArXiv astro-ph/0509519. Ben Gurion University in Israel, where he be- [70] Famaey, B., Gentile, G. Bruneton, J.-P. and Zhao, H.-Z, came full professor in 1978 and the Arnow 2006, ArXiv astro-ph/0611132. Professor of Astrophysics in 1983. In 1990 he [71] Famaey, B. and Binney, J., 2005, Mon. Not. Roy. As- moved to the Hebrew University of Jerusalem tron. Soc. 363 603. where he is the Polak Professor of Theoretical [72] Sanders, R. H., 2006, Mon. Not. Roy. Astron. Soc., 370, Physics. A member of the Israel Academy of 1519. Sciences and Humanities, of the World Jewish [73] Zhao, H. S. and Famaey, B., 2006, Astroph. Jour., 638, L9. Academy of Sciences, and of the International [74] Sanders, R. H., 2005, Mon. Not. Roy. Astron. Soc., 363, Astronomical Union, Bekenstein is a laureate 459. of the Rothschild prize and of the Israel Na- [75] Genzel, R. et al., 2006, Nature 442, 786. tional Prize. His scientific interests include [76] Milgrom, M., private communication. gravitational theory, black hole physics, rela- [77] Zlosnik, T. G., Ferreira, P. G. and Starkman, G. D., tivistic magnetohydrodynamics, galactic dy- 16 namics, and the physical aspects of informa- tion theory.