Draft (Jan. 18, 2010) Executive Summary

Дата канвертавання24.04.2016
Памер0.62 Mb.
1   2   3   4   5   6   7   8   9   ...   12

4.4Star formation in the local universe and at high-redshift

How stars form in the Universe is a key unsolved question in astrophysics. We witness star formation both in the local universe and at high-redshift, in quiescent disk galaxies such as our own Milky Way and in violent interacting and merging galaxies. In the local universe, TMT, with its unique capabilities, will be able to study the initial mass function down to much lower masses, and probe the dynamics of star formation with its sensitive infrared instruments. At moderate redshift, TMT can be used to trace the star formation rate with a new indicator, Paschen line emission (see section 4.4.2) in ultra-luminous infrared galaxies (ULIRGs). Extensive observations of star formation by TMT at the key epoch z ~ 2 will directly probe how the stellar assembly occurs in galaxies and how they are connected with active galactic nuclei activities. These observations will also address how several fundamental scaling relations in disk and elliptical galaxies (the Tully-Fisher relation and the fundemental plane) evolve as a function of time.

4.4.1Star formation in the local universe

Di Li (JPL), Qizhou Zhang (CfA)
Key questions:

  1. What are the initial mass functions (IMFs) of local star formation regions? Are IMFs universal?

  2. What are the key dynamical processes in star formation regions?

  1. Initial mass function: connectioning dark universe to visible matter

The formation of stars is the key process to construct most of the structures, including galaxies, stellar clusters, and planets, in the universe. Heavy elements and dust are believed to have originated in stellar nucleosythesis. The enrichment of matter further accelerates star formation. Thus, understanding the cycle of star formation, especially, formation of massive stars given their relatively short lifetime, is essential for understanding why we are living in the environment as known to human today.
One of the holy-grail in studying star formation is understanding the initial mass function of stars (IMF). Coupled with star formation efficiency, IMF is a crucial descriptive law for star formation, which specifies what will be made out of how much matter. To connect cold dark matter simulations to actual observable structures and to explain galaxy evolution, the state of art treatment of star formation in these simulations is still largely a parameterized extrapolation based on galactic IMF. Our understanding of galactic IMF, however, is far from certain neither in terms of observational facts nor of underlying physics. There are abundant evidences for a relatively well accepted uniformity of IMF (Parvel 2002). There are also evidences for a flattening of IMF in special environment, such as the high density regions near galactic center (e.g. Figer et al. 1999). Since the majority of star formation in early universe is likely to occur in environments with enhanced density (e.g. mergers), it is essential to correctly measure and understand IMF under different conditions. Therefore half a century after its discovery by Salpeter, the IMF is still at the forefront of star formation research. TMT will have a direct and maybe game changing impact in IMF characterization.
The prime example of a top heavy IMF is the Arches cluster. Source confusion and large extinction variation are main problems for converting JHK photometry data to the correct stellar population. The image in Figure 4.1 represents the current state of art AO system on a 10 meter class system. With TMT, the expected PSF core of about ~ 0.02” will be a factor of 4 improvement in resolution. The sensitivity will expected to increase by 2 orders of magnitude. TMT will provide a dramatic more comprehensive and more accurate count of stars in different star forming conditions (density, metalicity, etc.) in the local universe.

Table 4.1 Limiting K-magnitudes and corresponding lower mass limit in Arches like clusters. This table is based on radio profiles of the Arches cluster and adopted from TMT Detailed Science Case (2007).

The resolving power and the sensitivity of TMT will allow for an unprecedented calibration of IMF under different conditions. In the galaxy, the stellar population census will be complete down to brown dwarf or even Jupiter mass objects. The characterization of IMF will be a major step in understanding star formation, and in turn, provides essential parameters for simulations of large-scale structures formation and galaxy evolution.

  1. Dynamics and impact of star formation

Star formation is a complex process involving gravity, magnetic field, turbulence, etc. To move beyond descriptive knowledge of laws of star formation (e.g. IMF), a direct probe of the dynamics of the star forming cores is highly desirable. At R up to 120,000, TMT/MIRES achieves a 3 km/sec spectral resolution in the mid-infrared bands. Such a spectral resolution along with the high spatial resolution will be suitable to study the dynamical process of collapsing cores and protostars.

Compared with existing mid-infrared spectrometer, in particular, the Spitzer IRS instrument in the same bands (see Figure 4.2), MIRES will increase sensitivity by about two orders of magnitude while increasing the spectral resolution by 3 orders of magnitude. JWST will be slightly more sensitive than MIRES, albeit at ~100 times worse spectral resolution. The capability of TMT will be unique for the foreseeable future.
A primary example of spectral tracers are rotational transitions of H2, the most abundant molecule in the universe. In mid infrared bands, warm molecular hydrogen produces S(1) transition at 17.04 m, s(2) at 12.28 m, and s(3) at 9.67 m. These lines traces gas with upper level energy of hundreds to a few thousands K and both ortho and para population. The population ratio between rotational levels and the ortho/para ratio are powerful diagnostic tools of the physical condition and the evolutionary history of molecular gas.
The resolution element of TMT will be about 100 Au at a distance of 1 kpc. A comprehensive survey of protostellar envelopes in this local volume will produce a systematic view of the dynamical processes in star formation including collapses accretion, and jet/outflow.

Figure 4.2 Averaged spectra toward positions in NGC 1333 obtained using Spitzer IRS in SH mode (Maret et al. 2009). Note that H2 lines are not resolved in these spectra.


  1. Figer, D.F., Kim, S.S., Morris, M., Serabyn, E., Rich, R.M., & McLean, I.S. 1999, ApJ, 525, 750

  2. Kroupa, P. 2002, Science, 295, 82

  3. Maret, S., et al. 2009, ApJ, 698, 1244

4. Espinoza, P., Selman, F.J., & Melnick, J. 2009, A&A, 501, 563

4.4.2Star formation at high redshift

Jiasheng Huang (CfA), Lin Yan (Caltech)
“ORIGIN” is a key program identified by NASA and the international astronomy community for the future space exploration, which includes studies of origin of earth, planets, stars and galaxies in our universe. Studies of galaxy formation and evolution are an important part of the program. The critical epoch in our universe is at z=2 when both the total star formation and AGN activities are at their highest level. At present, astronomers began to observe formation process of galaxies with 8m class telescopes and Hubble Space Telescope (HST). HST detected fine structures with a scale down to 0.8kpc in galaxies at z=2. The new Near-infrared IFU, SINFONI, on VLT measures velocities of H emission lines to study the dynamics of galaxies at 110 solar masses at a resolution of 1 kpc with adaptive optics and 4-5kpc with natural seeing. The average size of galaxies at z~2 is rather small, thus a larger telescope with a better resolution and higher sensitivity is needed to study the whole galaxy population. There are a number of questions that TMT can make fundamental contributions.
1. Disk settling and star formation at z ∼ 2:
Key questions:

  1. Are disks present at z ∼ 2 but thus far largely unseen because rest-UV images are dominated by dust and young stars? If disks are in place, what are their radii, thicknesses, star formation rates, and mass distributions? If disks are not present, is there an alternative, simple description for the distribution of stars in these galaxies?

  2. Did disks at z~2 have the Tully-Fisher relation? If so, how does it differ from the local one?

  3. How does the star formation rate change during the formation of disks?

Theoretical models suggest that gas falling onto galaxies should settle quickly into disks, with radii scaling with the size of the universe. Current hydrodynamic models produce disks readily, but they are predicted to be very gas-rich and turbulent due to strong “cold flows” that feed them (e.g., [1]). This suggests that the disk morphologies are more irregular and clumpier than present-day disks, which is corroborated by the UV morphologies of the few disks that have been seen before z ∼ 1.5 ([2, 3]). More observation on NIR morphologies for galaxies at z~2 will be done with the HST WFC3 imaging survey. Dynamic evidence for disk formation at z=2 are lacking. Kassin et al. [4] measured the Tully-Fisher relation at z~1, confirming that disks were already formed. SONFONI observations reveal that only <1/3 star forming galaxies at 1
2. The emergence of massive spheroids at z ∼ 2:
Key questions:

  1. What causes star formation to shut down, and what is the spectroscopic evidence?

  2. Were spheroids formed in a single event, or did they grow slowly over time?

  3. How are the spheroidal bulges of disk galaxies and E/S0’s related? How did E/S0 galaxies move onto or along the fundamental plane?

Our current picture is that spheroidal galaxies – today’s E’s and E/S0’s – are the end-state in the evolution of all massive galaxies (e.g., [5, 6]). Extremely low star formation rates make them look “red-and-dead”, while dynamically “hot” kinematics give them their spheroidal shapes and smooth brightness profiles. The shut-down of star formation leads to the build-up of a distinctive “red sequence” [7] as colors redden after quenching.

A number of candidate mechanisms have been proposed for turning off star formation in galaxies. Some are dynamical, such as quenching by mergers, which scramble disks into spheroids dynamically and quench star formation by triggering a starburst or AGN which then drives out gas ([8, 9]). Others predict quiescent quenching, e.g., low-luminosity black hole feedback that prevents gas (i.e., star formation fuel) from cooling onto galaxies (see [10-12] for an alternative scenario). In the merger model, galaxies that are evolving red-ward are recent merger remnants with disturbed or asymmetric morphologies [13], while red-and-dead spheroids should be old remnants with smooth and symmetric profiles. In more quiescent quenching models, quenched galaxies can evolve without merging, need not look disturbed, and may retain disks they had before quenching.

A slow build-up of stellar spheroids is suggested by NICMOS images of red galaxies at z ∼ 2, which reveal sizes up to 5 times smaller than similar-mass spheroids today ([14-16], but see also [17]). A series of gas-poor mergers may cause a gradual growth in size [18, 19], implying continuous assembly of massive, quiescent galaxies over time. WFC3-IR will soon provide unparalleled data for testing these theories. The Dynamic study of red galaxies will be able to determine how red galaxies evolve From z=2 to z=0. Spectroscopy of red galaxies at z=2 are extremely hard even with 8m class telescopes. Kriek et al. [20] performed ultra-deep exposure spectroscopy for red galaxies at z~2 with VLT, could only measure the 4000A jump for their redshift determination. A 30m telescope such as TMT is clearly needed for this study to address the key questions in spheroidal formation.

3. The role of AGN at the peak of the QSO era:
Key questions:

  1. Are mergers the main trigger of BH growth at z ∼ 2?

  2. Does BH growth precede or lag star formation or the build-up of a spheroid?

  3. Does BH feedback cause quenching in spheroidal galaxies?

The close connection between massive black holes (BHs) and spheroids (e.g., [21]) is puzzling because only a few percent of galaxies appear to be building BHs at any instant – how is this connection established and maintained? HST observations do not support dominant merger-driven BH growth at z ∼ 1 (e.g., [22,23]) even though most models of AGN/galaxy co-evolution have mergers as an essential component. These mergers may be occurring at higher redshifts, but very little is known about the structure of AGN hosts when luminous optical QSO activity was at its peak. The deep multi-wavelength data available in our fields will enable new tests of co-evolution scenarios when WFC3 imaging is added.

QSOs at their birth are usually in very dusty environment, such as ULIRGs and

Dust Obscured Galaxies [24-27], which make their optical counterparts too faint for 8m class telescopes. TMT will be able to identify faint AGN signature spectroscopically in the dusty environment. By combining TMT spectroscopy and ALMA imaging for high-z Dusty galaxies and ULIRGs, we will be able to study the birth of QSOs.

On the other hand, the relation between black hole mass and AGN hosting galaxies at high redshifts are unclear at their dawn, while the local relation is well established. TMT will be able to measure the dynamic masses of hosting galaxies which are impossible at all for 8m class telescopes.

Possible TMT programs:

  1. A kinematic survey of disk galaxies at z~2 selected from other imaging surveys to study the structure, dynamics and evolution of the Tully-Fisher relation.

  2. A survey of spheroidal galaxies at z~2 to study their formation, and the evolution of the fundamental plane.

  3. A systematic study of selected QSOs to understand the interplay of star formation and the formation of black hole.

China’s strength and weakness in this area:
China has several groups of people working on star formation at high-redshift at PMO and TJNU, using multi-wavelength data, particularly in the radio. Several oversea Chinese astronomers are playing leading roles in studying star formation using infrared data, e.g. from large ground-based telescopes and space probes from SPITZER. However, closer collaboration between astronomers inside China and overseas may be a fruitful way to gain access to current state-of-the-art ground and space telescopes and further improve international visibilities in this area.

  1. Dekel, A., Sari, R., & Ceverino, D. 2009, ApJ, 703, 785

  2. Law, D.R., Steidel, C.C., Erb, D.K., Pettini, M., Reddy, N.A., Shapley, A.E., Adelberger, K.L., & Simenc, D.~J. 2007, ApJ, 656, 1

  3. Elmegreen, D.M., Elmegreen, B.G., Marcus, M.T., Shahinyan, K., Yau, A., & Petersen, M. 2009, ApJ, 701, 306

  4. Kassin, S.A., et al. 2007, ApJL, 660, L35

  5. Bell, E.F., et al. 2004, ApJ, 608, 752

  6. Faber, S.M., et al. 2007, ApJ, 665, 265

  7. Strateva, I., et al. 2001, AJ, 122, 1861

  8. Sanders, D.B., Soifer, B.T., Elias, J.H., Madore, B.F., Matthews, K., Neugebauer, G., & Scoville, N.Z. 1988, ApJ, 325, 74

  9. Hopkins, P.F., Hernquist, L., Cox, T.J., Di Matteo, T., Robertson, B., & Springel, V. 2006, ApJS, 163, 1

  10. Croton, D.J., et al. 2006, MNRAS, 365, 11

  11. Sijacki, D., & Springel, V. 2006, MNRAS, 366, 397

  12. Dekel, A., & Birnboim, Y. 2006, MNRAS, 368, 2

  13. Barnes, J.E., & Hernquist, L. 1996, ApJ, 471, 115

  14. Zirm, A.W., et al. 2007, ApJ, 656, 66

  15. van Dokkum, P.~G., et al. 2008, ApJL, 677, L5

  16. Damjanov, I., et al. 2009, ApJ, 695, 101

  17. Mancini, C., et al. 2009, MNRAS, 1721

  18. Hopkins, P.F., Bundy, K., Hernquist, L., Wuyts, S., & Cox, T.~J. 2009, MNRAS, 1635

  19. Naab, T., Johansson, P.H., & Ostriker, J.P. 2009, ApJL, 699, L178

  20. Kriek, M., et al. 2007, ApJ, 669, 776

  21. Tremaine, S., et al. 2002, ApJ, 574, 740

  22. Grogin, N.~A., et al. 2005, ApJL, 627, L97

  23. Pierce, C.~M., et al. 2007, ApJL, 660, L19

  24. Dey, A. 2006, HST Proposal, 10890

  25. Yan, L., et al. 2005, ApJ, 628, 604

  26. Yan, L., et al. 2007, ApJ, 658, 778

  27. Huang, J.-S., et al. 2009, ApJ, 700, 183

4.4.3Near-Infrared Emission Line Studies of ULIRGs

Zhong Wang (CfA)
Key question:

How do we measure the star formation rate independently in interacting galaxies?

Ultra-luminous infrared galaxies (ULIRGs), with their bolometric luminosities on the order of 1012 Lsun and beyond, represent some of the most active star forming sites on galactic scale, and have significant implications to galaxy population and evolution. In addition to rapidly forming stars, two other fundamental properties are now well-established for these systems: 1) most, if not all, of them are interacting or merging galaxy pairs undergoing a drastic transition phase; and 2) nearly all contain a large amount of gas and dust, often (but not always) near its central region (Sanders and Mirabel 1996; Genzel et. al 2001; Tacconi et. al 2008).

Because of the high extinction in ULIRGs, conventional star formation probes in the optical tend to be ineffective; longer wavelength measurements such as mid- and far-IR data, on the other hand, are complicated by the presence of AGNs and black holes in the nuclei. The near-infrared Paα emission is potentially one of the most useful for this purpose, and it is known to be very bright. Unfortunately it cannot be used in ground-based observations of many well-known local ULIRGs, due to the severe absorption of a telluric water-vapor absorption feature.
However, the K-band sensitivity of a TMT-class telescope can easily extend the near-IR emission line study to a large number of more distant galaxies (those beyond z=0.15), whose redshifts move the Paα line into a range where the water-vapor absorption is far less contaminating. This potentially opens up an excellent window of opportunity for us to learn a great deal more about individual ULIRGs in terms of their star formation activities and internal kinematics.
Paschen alpha line: a unique probe of star formation in dusty galaxies:
The Paschen line emission is of particular interest to us for a number of reasons. First, being a near-infrared line it provides a powerful probe of deeply embedded star forming clusters that are the hallmarks of ULIRGs. It not only promises to reveal more underlying physics than the optical/UV

lines, but can also yield kinematic information on sub-angular scale (the average seeing in the NIR for the TMT is expected to be around 0.3''). This is especially useful for the moderately distant (z >= 0.15) ULIRGs, of which the existing longer wavelength (e.g., mid-IR) probes only measure global properties. Secondly, basic atomic physics tells us that Paschen α has a fixed intrinsic flux ratio to H which (along with the Hα line) is the conventional yardstick to gauge normal star forming galaxies (Calzetti et. al 2007; Kennicutt et. al 2009). Therefore, by measuring this line we will be able to make direct comparisons between the star forming activities in moderate to high redshift ULIRGs and those in better studied local galaxies, while taking a more accurate account of the average reddening in their central regions (Genzel et. al 2008).

Among the potentially observable hydrogen transitions beyond Hα line, Paα is also by far the brightest. It is at least a factor of 10 brighter than Br the next brightest near-IR line observable (at 2.166 micron) from the ground (and often used as a substitute). Because of the strong atmospheric absorption at its rest wavelength, unfortunately, it has been impossible to observe local objects using Paα line from the ground. Nevertheless, its brightness and usefulness for studying star-forming galaxies have been demonstrated with space-borne observations, in particular, those of NICMOS aboard the Hubble (e.g., Scoville et. al 2000).
So far most ground-based NIR spectroscopic studies of local (U)LIRGs have concentrated on the stellar absorption features in the H band. For example, Dasyra et. al (2006a, b) have carried out a Keck-VLT legacy program to measure the velocity dispersion and stellar rotations of more than 50

local (U)LIRGs, using the CO bandhead absorption features. While this is an important and worthwhile effort, many questions remain as to what extent these absorption features are representative of the active star forming component at the nuclei of typical ULIRGs. Since most of the ULIRGs we are studying have redshifts beyond 0.15, which moves the Paα line out

of the water-vapor absorption feature and into K band where a larger, NIR-optimized ground-based telescope such as the TMT would have sufficient sensitivity up to z=1-2, we suggest that in these cases, it would actually be far more efficient to measure the emission lines instead. [In fact, in one of the earlier works of the Dasyra group, Paα line at z=0.164 was clearly detected in at least one of the ULIRG's spectra with the Keck (Genzel et. al 2001). But they did not pursue this line of inquiry, choosing to focus instead on the H band and targets which have lower redshifts (and thus brighter)].
ULIRG samples for observation with a large ground-based telescope
In collaboration with the GOALS (Great Observatory All-sky LIRG Survey) project (Armus et. al 2009), we have been studying a large number of LIRGs in the infrared and mm/submm wavelengths. The GOALS sample starts with the Revised Bright Galaxy Sample (RBGS, Sanders et. al 2003), selecting infrared bright ones based on a criteria of IRAS 60 micron flux of greater than 5~Jy. A much more extensive collection, which includes largely ULIRGs (LFIR > 1012 Lsun as opposed to LIRGs (LFIR > 1011 Lsun, is the 1 Jy sample of Kim and Sanders (1998). Because the NIR emission line study focuses on the ULIRGs, we can draw sample galaxies primarily from the latter. Limiting the sample to those ULIRGs with redshift beyond 0.15 would ensure detection and a reliable calibration of the Paα line.
Besides the redshift range, we can rank the candidate galaxies based on total infrared luminosity (LFIR), which for this sample ranges from approximately 1011 to 1013 Lsun. With the top-ranked objects, we would further give higher priorities to those ULIRGs already studied in the other wavelengths. This is done such that the emission-line study may be used to systematically compare with the results of complementary studies and draw useful conclusions with regard to the underlying physics from different probes.
There is a general consensus that reaching into a large population of ULIRGs at beyond z=0.15 is of critical importance to the study of galaxy interactions and their impact on cosmic evolution. With the powerful NIR capabilities the 30-m class TMT will allow precisely this type of quantitative measurements. The rationale in ranking the sample ULIRGs primarily on total infrared luminosity

is that other measurements (e.g., visual magnitudes measured at shorter wavelengths) have little bearing on the actual level of star forming activities in these extraordinary systems, given the high level of average extinction. This is also consistent with our study of GOALS galaxies with Spitzer and the radio telescopes. Our recent observations with the JCMT and CSO telescopes, for example, have shown that when ranked in FIR luminosities, nearly all of the top 100 galaxies in the GOALS sample turns out to be molecular gas and dust-rich interacting pairs, mergers, or post-merger remnants.

Overall scientific goals:
Emission lines from ionized regions are the tell-tale signs of star forming activities in galaxies, and, provided that the extinction factor can be reasonably evaluated, the most consistent and reliable measure of the amount of star formation taking place (e.g., Kennicutt et. al 2009). In the case of

ULIRGs, the NIR emission lines are especially meaningful because of the heavy extinction. So our first goal is to obtain a reliable measure of star formation rate, and compare that with the far-IR luminosity, as well as the molecular gas content. This will provide not only overall rates but also

the star formation efficiencies in a large population of true ULIRGs.
The second goal is to measure the structural parameters of the star forming component in these galaxies. With the expected spectral resolving power and a ~0.3'' seeing, we should be able to obtain the position-velocity diagram from the Paα line, and thus to evaluate the rotational velocities near the center of many ULIRGs. This can then be directly compared with the rotation curves derived for local LIRGs. These kinematic information are very important in modeling the formation and evolution of ULIRGs, widely recognized as a transitional phase from merging disk galaxies to large ellipticals with central AGNs and black holes (Hopkins et. al 2007, 2008).
In many cases, we also expect to be able to resolve the Paα emission well enough to separate the nuclear and disk contributions, thus assess the distribution of the starburst activities in these systems. Because of the presence of AGNs in many of these systems, there has always been a debate concerning the contribution of the non-thermal nuclear activities to the far-IR luminosity. Since many of the longer wavelength diagnostics are only measuring global properties, being able to spatially resolve the different components for a representative population of ULIRGs would be critical to address such controversies.
Finally, the NIR emission line study of these moderately distant ULIRGs helps to bridge a gap between the relatively well-studied nearby LIRGs and the large population of ULIRGs at truly high redshifts. While the more familiar examples such as Arp 220 and NGC 6240 are indeed spectacular and offer a close-up view of the merger-in-progress, questions remain regarding to what extent these resemble the prominent ULIRGs around z=1-2 (for example, those from the 24-micron selected sample) discovered with Spitzer (Le Floch et. al 2005, 2009; Yan et. al 2007). After all, it is the latter population, formed when the universe was far younger and much more efficient in processing its baryonic masses through star formation, whose evolutionary path would determine the overall appearance of the universe of galaxies as we see today.
Possible TMT program:

  1. Select a sample of ULIRGs both locally and at intemediate redshift and obtain their spectra using TMT/IRMS to detect the Paα emission and measure their star formation rates.

China’s strength and weakness in the area:

This is an emerging new method where China can take lead in collaboration with oversea Chinese astronomers through joint PhD student programs, particularly the pre-doctoral program at CfA.


  1. Armus, L. et. al 2009, PASP, 121, 559

  2. Calzetti, D. et. al 2007, Ap.J. 666, 870

  3. Dasyra, K. M. etal 2006a, Ap. J. 638, 754

  4. Dasyra, K. M. et. al 2006b, Ap. J. 651, 835

  5. Hopkins, P. F. et. al 2008, Ap.J. Suppl. 175, 356

  6. Hopkins, P. F. et. al 2009, MNRAS, 397, 802

  7. Genzel, R. et. al 2001, Ap.J. 563, 527

  8. Genzel, R. et. al 2008, Ap.J. 687, 59

  9. Kennicutt, R. C. et. al 2009, Ap.J. 703, 1672

  10. Kim, D. -C. and Sanders, D. B., 1998, Ap.J. Suppl. 119, 41

  11. Le Floch, E. et. al 2005, Ap.J. Lett. 632, 169

  12. Le Floch, E. et. al 2009, Ap.J. Lett. 703, 222

  13. Sanders, D. B. and Mirabel, I. F. 1996, ARA&A, 34, 749

  14. Sanders, D. B. et. al 2003, AJ, 126, 1607

  15. Scoville, N. Z. et. al 2000, AJ. 119, 991

  16. Tacconi, L. J. et. al 2008, Ap.J. 680, 246

  17. Yan, L. et. al 2007, Ap.J. 658, 778
1   2   3   4   5   6   7   8   9   ...   12

База данных защищена авторским правом ©shkola.of.by 2016
звярнуцца да адміністрацыі

    Галоўная старонка