ReadingTimeMachine/rtm-sgt-ocr-v1 · Datasets at Hugging Face

The absolute value of the specific binding energy of two stars of mass m. separation r and relative velocity ο is where we have introduced « for the gravitational energy of the pair.

The absolute value of the specific binding energy of two stars of mass $m$ , separation $r$ and relative velocity $v$ is where we have introduced $x$ for the gravitational energy of the pair.

The number density can then be written as n(r)xαν

The number density can then be written as $n(r) \propto x^{q}$.

With a Alaxwellian distribution of relative velocities with dispersion o. the number of potential binary partners à starsees in the range € to €|de 2..

With a Maxwellian distribution of relative velocities with dispersion $\sigma$, the number of potential binary partners a starsees in the range $\epsilon$ to $\epsilon + {\rm d} \epsilon$ \ref{1024semidist}.

Debris dises are flattened distributions of planetesimals and dust located at radii of 1-1000 AU around main-sequence stars (see (seeWyatt2008.forarecentreview)..

Debris discs are flattened distributions of planetesimals and dust located at radii of 1-1000 AU around main-sequence stars (see \citep[see][for a recent review]{wyatt08}.

The dust cannot be primordial since its lifetime in orbit is significantly less than the age of the host stars.

Instead. dust is replenished from a population of colliding km-sized planetesimals (Wyatt&Dent2002:ThébaultAugereau 2007).

Instead, dust is replenished from a population of colliding km-sized planetesimals \citep{wyatt02,the07}.

. Over time. the dust distribution is shaped by any planetary-sized bodies in the system (e.g..Dominik&Decin2003:Wyattetal.2007).

Over time, the dust distribution is shaped by any planetary-sized bodies in the system \citep[e.g.,][]{dom03,wyatt07}.

Therefore. resolved images of disces constrain models of the structure and evolution of planetary systems.

Therefore, resolved images of discs constrain models of the structure and evolution of planetary systems.

Far-infrared and submillimetre observations are the best way to search for dust around nearby stars due to the favorable contrast of the disc relative to the star.

At these wavelengths. the disc emission is optically thin and ts sensitive to the large (up to ~| mm). cold grains which dominate the disc’s dust mass.

At these wavelengths, the disc emission is optically thin and is sensitive to the large (up to $\sim 1$ mm), cold grains which dominate the disc's dust mass.

The Space Observatory offers three major advantages for the detection and characterization of debris dises: far-infrared sensitivity. angular. resolution and wavelength coverage.

The Space Observatory offers three major advantages for the detection and characterization of debris discs: far-infrared sensitivity, angular resolution and wavelength coverage.

With its 3.5 m mirror, its sensitivity at far-infrared wavelengths is superior to any previous instrument.

With its resolution of 6777 at 100jim... has the potential to resolve many debris dises. particularly toward nearby stars.

With its resolution of 7 at 100, has the potential to resolve many debris discs, particularly toward nearby stars.

Finally. with detectors at 100. 160. 250. 350 and 500um.. has the means to sample the spectral energy distribution (SED) of dise emission across the peak. meaing models can be better constrained even for discs which are not resolved.

Finally, with detectors at 100, 160, 250, 350 and 500, has the means to sample the spectral energy distribution (SED) of disc emission across the peak, meaning models can be better constrained even for discs which are not resolved.

DEBRIS (Disc Emission via a Bias-free Reconnaissance in the Infrared/Submillimetre) is an Open Time Key Programme which uses PACS (Photodetector Array Camera Spectrometer) and (for appropriate targets) SPIRE (Spectral and Photometric Imaging Receiver) to detect. resolve and characterize debris dises around a volume-limited sample of 446 A through M type stars.

DEBRIS (Disc Emission via a Bias-free Reconnaissance in the Infrared/Submillimetre) is an Open Time Key Programme which uses PACS (Photodetector Array Camera Spectrometer) and (for appropriate targets) SPIRE (Spectral and Photometric Imaging Receiver) to detect, resolve and characterize debris discs around a volume-limited sample of 446 A through M type stars.

The goals of DEBRIS include establishing the incidence and evolution of debris discs as a function of stellar type. age. multiplicity. ete.:

The goals of DEBRIS include establishing the incidence and evolution of debris discs as a function of stellar type, age, multiplicity, etc.;

the characterization of discs in terms of size. temperature. dust mass and morphology (where the dise asymmetries could indicate the presence of planetary companions); and the understanding of our own Solar System in the context of the larger debris dise population.

the characterization of discs in terms of size, temperature, dust mass and morphology (where the disc asymmetries could indicate the presence of planetary companions); and the understanding of our own Solar System in the context of the larger debris disc population.

Full details of the DEBRIS survey and goals will be presented in à forthcoming paper (B. Matthews et al.

Full details of the DEBRIS survey and goals will be presented in a forthcoming paper (B. Matthews et al.

2010. in preparation).

2010, in preparation).

Here. we present PACS observations toward three of the first targets of the DEBRIS survey.

Here, we present PACS observations toward three of the first targets of the DEBRIS survey.

We briefly summarize theobservations and targets in Sect. 2..

We briefly summarize theobservations and targets in Sect. \ref{obs}, ,

present the results in Sect.

3. and discuss three sources in detail in Sect. 4..

\ref{res} and discuss three sources in detail in Sect. \ref{disc}.

We summarize the paper in Sect. 5..

We summarize the paper in Sect. \ref{sum}.

DEBRIS is a flux-limited survey and as such it observes each target to a uniform depth (1.2 mJy beam"! at 100 jm). resulting in different mass limits for targets at different distances and of different stellar spectral types.

DEBRIS is a flux-limited survey and as such it observes each target to a uniform depth (1.2 mJy $^{-1}$ at 100 ), resulting in different mass limits for targets at different distances and of different stellar spectral types.

Here. we present 100 and 160 pphotometry observations toward three nearby stars (see Table 1) performed with the ESAHerschel Space Observatory (Pilbrattetal.2010) utilizing the PACS (Poglitsch.Waelkens&Geis2010) instrument.

Here, we present 100 and 160 photometry observations toward three nearby stars (see Table \ref{table}) ) performed with the ESA Space Observatory \citep{pil10} utilizing the PACS \citep{pog10} instrument.

The results presented here were taken during early testing phases or during the Science Demonstration Phase on (2009 -- Dec.) The images were obtained using two different observing strategies: point source chop/nod. and small scan-map modes (see the PACS Observers’ Manual!)).

The results presented here were taken during early testing phases or during the Science Demonstration Phase on (2009 - Dec.) The images were obtained using two different observing strategies: point source chop/nod, and small scan-map modes (see the PACS Observers' ).

For point-source mode observations. seven contiguous repeat chop/nod observations were performed.

For point-source mode observations, seven contiguous repeat chop/nod observations were performed.

Scan map observations had eight repeats in a single scan direction at a rate of 20"//s.

Scan map observations had eight repeats in a single scan direction at a rate of /s.

Four sscan legs were performed per map with a sseparation between legs.

Four scan legs were performed per map with a separation between legs.

The total observing times were 1072 and 1220 seconds. respectively. for each chop/nod and scanning observation.

The total observing times were 1072 and 1220 seconds, respectively, for each chop/nod and scanning observation.

Table 1 shows the survey (“UNS”) identifier for each target as well as the source and observing details.

Table \ref{table} shows the survey (“UNS”) identifier for each target as well as the source and observing details.

Phillipsetal.(2010) contains details of the development of the Unbiased Nearby Stars sample from which the DEBRIS targets are drawn.

\cite{phil09} contains details of the development of the Unbiased Nearby Stars sample from which the DEBRIS targets are drawn.

These data were reduced using the interactive processing environment (HIPEOtt2010).

These data were reduced using the interactive processing environment \citep[HIPE][]{ott10}.

. Maps were obtained via the default PACS natvve map-making nethods and in HIPE for the scanning and point source observing modes respectively.

Maps were obtained via the default PACS naïvve map-making methods and in HIPE for the scanning and point source observing modes respectively.

Scanned data were pre-filtered to remove low frequency (1/f)) noise using a boxcar filter with width equal to 1755.

Scanned data were pre-filtered to remove low frequency ) noise using a boxcar filter with width equal to 5.

All bright sources in the map were masked prior to filtering to avoid filter ringing type artefacts.

The chop/nod configuration meant that no equivalent filtering was required for the data obtained in point source mode.

All three targets presented in this paper are shared targets with the DUNES Key Programme (PI: EEiroa) which has science goals complementary to those of DEBRIS.

All three targets presented in this paper are shared targets with the DUNES Key Programme (PI: Eiroa) which has science goals complementary to those of DEBRIS.

Details on the distribution of targets will be discussed in a survey deseription paper (B. Matthews 2010. in preparation).

Details on the distribution of targets will be discussed in a survey description paper (B. Matthews 2010, in preparation).

Figure | shows the 100 and 160 images for the three targets.

Figure \ref{images} shows the 100 and 160 images for the three targets.

The rms levels achieved in each observation are summarized in Table ]..

The rms levels achieved in each observation are summarized in Table \ref{table}.

The higher noise levels associated with the point-source mode are evident.

The scan map noise levels were significantly lower for comparable observing times and. for 7 Corvi. meet the DEBRIS rms specifications.

The scan map noise levels were significantly lower for comparable observing times and, for $\eta$ Corvi, meet the DEBRIS rms specifications.

For8 Leo. the rms is higher by ~ δύο.

For $\beta$ Leo, the rms is higher by $\sim$ .

Theintegrated flux densities are estimated for each image (star + dise).

Theintegrated flux densities are estimated for each image (star + disc).

This is done with simple aperture photometry

If synchrotron self-absorption is the origin of the turnover. we can obtain. from Eqs. (3))

If synchrotron self-absorption is the origin of the turnover, we can obtain, from Eqs. \ref{abssync}) )

and (4). v,=1 GHz. and Fass40 mJy. the following r—B and r—y relationships: Since. the synchrotron emitting electrons must. be relativistic. y210. and therefore. r€10!" em and B<1G. In this case. having a constraint on +. the argument of the limited energy budget can be applied as well. allowing us to derive further restrictions on the parameters: B~107—1 G: yκ~I0-IO: and r~(4.5—10)xI0em few AU (recall Ας.

and \ref{rel}) ), $\nu_{\rm max}=1$ GHz, and $F_{\rm max}\approx 40$ mJy, the following $r-B$ and $r-\gamma$ relationships: Since the synchrotron emitting electrons must be relativistic, $\gamma\ga 10$, and therefore, $r\la 10^{14}$ cm and $B\la 1$ G. In this case, having a constraint on $r$, the argument of the limited energy budget can be applied as well, allowing us to derive further restrictions on the parameters: $B\sim 10^{-2}-1$ G; $\gamma\sim 10-10^2$; and $r\sim (4.5-10)\times 10^{13}\,{\rm cm}\sim$ few AU (recall $R\ga r$ ).

The combination of PO7 and (005 observations. show variability at 234 MHz of a at year timescales. too large to be attributed only to interstellar medium scintillation effects.

The combination of P07 and G08 observations show variability at 234 MHz of a at year timescales, too large to be attributed only to interstellar medium scintillation effects.

The flux at 614 MHz does not show significant changes.

The value of f, at 234 MHz Is: probably too long to explain the variability just. with changes of the electron injection.

The value of $t_{\rm sync}$ at 234 MHz is: probably too long to explain the variability just with changes of the electron injection.

Some level of variability around | GHz may be induced by changes in the stellar wind if free-free absorption produced the spectral turnover. but at the estimated distance for the free-free absorption case. synchrotron self-absorption should still atfect the radiation at 234 MHz due to the region compactness.

Some level of variability around 1 GHz may be induced by changes in the stellar wind if free-free absorption produced the spectral turnover, but at the estimated distance for the free-free absorption case, synchrotron self-absorption should still affect the radiation at 234 MHz due to the region compactness.

The optically thin. nature of PO7 data. and the fast variability. can be otherwise explained by advection with timescales shorter than ~| yr together with injection variations.

The optically thin nature of P07 data, and the fast variability, can be otherwise explained by advection with timescales shorter than $\sim 1$ yr together with injection variations.

Taking vay€10! em s!. ~| yr timescale restricts the size where the (high-state) 234 MHz radiation is produced torX3xI0 em.

Taking $v_{\rm adv}\la 10^{10}$ cm $^{-1}$, $\sim 1$ yr timescale restricts the size where the (high-state) 234 MHz radiation is produced to $r\la 3\times 10^{17}$ cm.

τι. has been taken here to be at most mildly relativistic to avoid the complexities of significant boosting effects (otherwise not expected: see Paredes et al. 2000. 2002)).

$v_{\rm adv}$ has been taken here to be at most mildly relativistic to avoid the complexities of significant boosting effects (otherwise not expected; see Paredes et al. \cite{paredes00,paredes02}) ).

In addition. the fact that this emission is not synchrotron absorbed in the PO7 data. and a flux at 234 MHz of =70 my.implies a size: Accounting for energy budget constraints. the |] yr variability. and the constraint y210 (re. electrons must be relativistic). to Eq. (9).

In addition, the fact that this emission is not synchrotron self-absorbed in the P07 data, and a flux at 234 MHz of $\approx 70$ mJy,implies a size: Accounting for energy budget constraints, the 1 yr variability, and the constraint $\gamma\ga 10$ (i.e. electrons must be relativistic), to Eq. \ref{rvar}) ),

we get B.-4x1070.6 G: y~10—400: and r~R104%—3x10" em.

we get: $B\sim 4\times 10^{-4}-0.6$ G; $\gamma\sim 10-400$; and $r\sim R\sim 10^{14}-3\times 10^{17}$ cm.

R has similar limits to those of +: if bigger. the source would have appeared as extended under the GMRT resolution of ~10 arc-seconds at 234 MHz.

$R$ has similar limits to those of $r$; if bigger, the source would have appeared as extended under the GMRT resolution of $\sim 10$ arc-seconds at 234 MHz.

Concluding. the high-state 234 GHz emitting region probably has a larger size and smaller magnetic field than the region producing the emission 2614 MHz and at 234 MHz in the low-state. which indicates that they are probably different.

Concluding, the high-state 234 GHz emitting region probably has a larger size and smaller magnetic field than the region producing the emission $\ge 614$ MHz and at 234 MHz in the low-state, which indicates that they are probably different.

The different spectral shape of PO7 data and data >614 MHz would also point to a different electron population.

The different spectral shape of P07 data and data $> 614$ MHz would also point to a different electron population.

Otherwise. despite the variability at 234 MHz. the quite steady 614 MHz flux makes the conclusions of Sect.

Otherwise, despite the variability at 234 MHz, the quite steady $614$ MHz flux makes the conclusions of Sect.

22. still valid concerning the break at +1 GHz.

\ref{ff} still valid concerning the break at $\approx 1$ GHz.

In Table l.. at the top. the results of all this section are summarized.

In Table \ref{tab1}, at the top, the results of all this section are summarized.

The core seen with VLBA at 5 GHz in LS 5039 is not point like. presenting a projected size of ~8 AU. and two additional south-east (SE) and north-west (NW) components at a projected distance of ~8 AU from the core center are also detected (PAO: ROS).

The core seen with VLBA at 5 GHz in LS 5039 is not point like, presenting a projected size of $\sim 8$ AU, and two additional south-east (SE) and north-west (NW) components at a projected distance of $\sim 8$ AU from the core center are also detected (PA0; R08).

The variation of the core flux at 5 GHz within 5 days of difference ts about a1056... whereas the changes in flux of the SE and NW components are of about a 44 and a260%.. respectively (ROS).

The variation of the core flux at 5 GHz within 5 days of difference is about a, whereas the changes in flux of the SE and NW components are of about a 44 and a, respectively (R08).

Assuming that the flux changes of the SE and NW component fluxes are showing power engine variations (unnecessarily affecting both components in the same way). plus the fact that the synchrotron cooling timescales of radio electrons are too long to explain such a variability. we can derive that v,€108 em s! to have a steady core. and =3x105 em s! for the variable SE and NW components.

Assuming that the flux changes of the SE and NW component fluxes are showing power engine variations (unnecessarily affecting both components in the same way), plus the fact that the synchrotron cooling timescales of radio electrons are too long to explain such a variability, we can derive that $v_{\rm adv}\la 10^8$ cm $^{-1}$ to have a steady core, and $\ga 3\times 10^8$ cm $^{-1}$ for the variable SE and NW components.

If shorter time variability in the SE and NW components were observed. it would increase the vag, lower limit to 3109(7/5days)! em s7!.

If shorter time variability in the SE and NW components were observed, it would increase the $v_{\rm adv}$ lower limit to $3\times 10^8 (t/5~{\rm days})^{-1}$ cm $^{-1}$.

The main properties of the VLBA radio emission are summarized in Table 1.. bottom.

The main properties of the VLBA radio emission are summarized in Table \ref{tab1}, bottom.

Since LS 5039 harbors a very bright star (COS). and produces TeV emission (Aharonian et al. 2005)).

Since LS 5039 harbors a very bright star (C05) and produces TeV emission (Aharonian et al. \cite{aharonian05}) ),

secondary pars can be efficiently produced in the stellar wind via gamma-ray absorption.

secondary pairs can be efficiently produced in the stellar wind via gamma-ray absorption.

This population of secondary pairs could be behind the radio core (e.g. Bosch-Ramon. Khangulyan Aharonian 2008a:;; BO8 hereafter). moving collectively with the stellar wind. hence the low inferred advection speed.

This population of secondary pairs could be behind the radio core (e.g. Bosch-Ramon, Khangulyan Aharonian \cite{bosch08}; B08 hereafter), moving collectively with the stellar wind, hence the low inferred advection speed.

The feasibility of this scenario is illustrated in Fig. 2..

The feasibility of this scenario is illustrated in Fig. \ref{plot2},

where we show the computed radio spectral energy distribution (SED) of the secondary synchrotron emission created by gamma-ray absorption in LS 5039.

The magnetic field in the stellar surface has been fixed to 250 G (expected to be reasonable: see Bosch-Ramon. Khangulyan Aharonian 2008b)). the injected TeV luminosity to 4xI0? erg s! (similar to that inferred by Khangulyan. Aharonian Bosch-Ramon 2008)). and the TeV emitter distance to the stellar companion to 3.5x015 em. not exactly inside the binary system (as proposed in Bosch-Ramon et al. 2008b)).

The magnetic field in the stellar surface has been fixed to 250 G (expected to be reasonable; see Bosch-Ramon, Khangulyan Aharonian \cite{bosch08b}) ), the injected TeV luminosity to $4\times 10^{35}$ erg $^{-1}$ (similar to that inferred by Khangulyan, Aharonian Bosch-Ramon \cite{khangulyan08}) ), and the TeV emitter distance to the stellar companion to $3.5\times 10^{12}$ cm, not exactly inside the binary system (as proposed in Bosch-Ramon et al. \cite{bosch08b}) ).

As seen in the Fig. 2..

As seen in the Fig. \ref{plot2},

the computed fluxes match the observed values. and most of the 5 GHz emission is predicted to come from a region of few AU size. similar to first order to the values discussed in the previous paragraph.

the computed fluxes match the observed values, and most of the 5 GHz emission is predicted to come from a region of few AU size, similar to first order to the values discussed in the previous paragraph.

The computed α around 5 GHz is = —-04. roughly the same as the value found by M98 for the VLA radio emission (which should be dominated by the VLBA core).

The computed $\alpha$ around 5 GHz is $\approx -0.4$ roughly the same as the value found by M98 for the VLA radio emission (which should be dominated by the VLBA core).

The computed low-frequency spectrum. similar to the observed one (008). suggests that this radiation may come also from secondary pairs. although

The computed low-frequency spectrum, similar to the observed one (G08), suggests that this radiation may come also from secondary pairs, although

huninosities.

luminosities.

The observed IRAC flux deusities at 3.6500 and £542 from GOODS aud are iu most cases — there are two exceptious — fully consistent with the expected fiux densities for the bursts observed in the UV.

The observed IRAC flux densities at $\mu\rm{m}$ and $\mu\rm{m}$ from GOODS and are in most cases – there are two exceptions – fully consistent with the expected flux densities for the bursts observed in the UV.

In addition. the galaxies have the same sizes iu the J aud II ατα, indicating that the spatial extent of the region from which the line emission originates roughly follows the stellar light.

In addition, the galaxies have the same sizes in the J and H bands, indicating that the spatial extent of the region from which the line emission originates roughly follows the stellar light.

Hence. there is no evidence for muderlving older stellar populations.

Hence, there is no evidence for underlying older stellar populations.

Ποπονο, we cannot rule out icr existence: maximally old stellar populations have nass-to-lieht ratios that are up to ~50 times larger than rose of the bursts. even in the near infrared.

However, we cannot rule out their existence: maximally old stellar populations have mass-to-light ratios that are up to $\sim50$ times larger than those of the bursts, even in the near infrared.

Tf we asstune a past star formation rate that is coustautafter averaging over 2100 Myr time scales we find upper iuüts for the mass in older stars that is —5« the burst uass.

If we assume a past star formation rate that is constantafter averaging over $>100$ Myr time scales we find upper limits for the mass in older stars that is $\sim 5\times$ the burst mass.

The implied total stellar mass upper limits are hen €5«105A£..

The implied total stellar mass upper limits are then $\lesssim 5\times 10^8~\msol$.

This caveat notwithstanding. we asstune in the remainder of this paper that there is no sjeuificaut population of older stars im these ealaxics. and that the observed bursts account for the total stellar mass.

This caveat notwithstanding, we assume in the remainder of this paper that there is no significant population of older stars in these galaxies, and that the observed bursts account for the total stellar mass.

However. the bottom line is that the total stellar masses of these objects are well below LO?AL... in the reeine of dwarf galaxies.

However, the bottom line is that the total stellar masses of these objects are well below $10^9~\msol$, in the regime of dwarf galaxies.

Calaxies with simular properties have previously beeu identified through broad-band photometrvat 5«—0.1 in the Sloan Digital Sky Survey (Cardamoueetal.2009).. xb have been shown by Αποetal.(2010) and Izotovetal(2011) to constitute the most stronely star-forming aud most metal poor tail of the well-known class of blue compact dwarf galaxies (e.e..Sargent&Searle1970:Thuan&Martin1981:Caiithetal. 2011).. which have very low inetallieities aud extremely high. spatially concentrated star-formation activity (Carmien 2008).

Galaxies with similar properties have previously been identified through broad-band photometryat $z<0.4$ in the Sloan Digital Sky Survey \citep{cardamone09}, and have been shown by \citet{amorin10} and \citet {izotov11} to constitute the most strongly star-forming and most metal poor tail of the well-known class of blue compact dwarf galaxies \citep[e.g.,][]{sargent70, thuan81, griffith11}, , which have very low metallicities and extremely high, spatially concentrated star-formation activity \citep{guzman98, overzier08}.

. Cowieetal.(2011). (alsoseeSearlataetal. 200933) recently studied the ἵνα properties of hieh-EW Πα enütters. providing a direct connection between searches of Lava (8...Ouchietal.2008:IIual. 2010).. and find Lvo. EWs raneing from 20A to 200A.

\citet{cowie11} \citep[also see][]{scarlata09}) ) recently studied the $\alpha$ properties of high-EW $\alpha$ emitters, providing a direct connection between higher-redshift searches of $\alpha$ \citep[e.g.,][]{ouchi08, hu10}, and find $\alpha$ EWs ranging from $\rm{\AA}$ to $\rm{\AA}$ .

Combining this with the flucines of Nilssonetal.(2011).. who show that Ίσα emitters at 2~ are objects with a very wide range in properties. it is clear that frou Lyra cluitters one cannot derive a complete description of star formation in low-mass ealaxies.

Combining this with the findings of \citet{nilsson11}, who show that $\alpha$ emitters at $z\sim 2$ are objects with a very wide range in properties, it is clear that from $\alpha$ emitters one cannot derive a complete description of star formation in low-mass galaxies.

Ou the other hand. Lva cluitters at ligher redshifts (2> 3) appear to be vouug. with small stellar masses (e.e..Fiukelsteiuetal.2009).. simular to the euission line galaxies studied here.

On the other hand, $\alpha$ emitters at higher redshifts $z>3$ ) appear to be young, with small stellar masses \citep[e.g.,][]{finkelstein09}, similar to the emission line galaxies studied here.

Narrow-hband surveysideutified ealaxics with strong [OTT aud Πο emission lines with EW~100/—1000A at redshifts 2.=0.3l (eg.Kakazuetal.2007).. demonstrated to be vouug and metal poor (αιetal. 2009).

Narrow-band surveysidentified galaxies with strong [OIII] and $\alpha$ emission lines with $\rm{EW}\sim 100-1000\rm{\AA}$ at redshifts $z=0.3-1$ \citep[e.g.,][]{kakazu07}, demonstrated to be young and metal poor \citep{hu09}.

. Most notably, Ateletal.(2010) pointed out the existence of a class of emissiou line galaxies at 5~1.5 with EW>10004 that would most likely be inchided in our sample as well.

Most notably, \citet{atek10} pointed out the existence of a class of emission line galaxies at $z\sim 1.5$ with $\rm{EW}> 1000\rm{\AA}$ that would most likely be included in our sample as well.

ILoxcever. so fax. their nature has uot been described and their cosmological relevance iu the context of galaxy formation has remained unclear.

However, so far, their nature has not been described and their cosmological relevance in the context of galaxy formation has remained unclear.

Therefore. let us now put these starburstiug chwart ealaxies iu a cosmological context.

Therefore, let us now put these starbursting dwarf galaxies in a cosmological context.

Our sample with redshifts 1.6<2.1.5 consists of 69 low-mass (107A£.). voung (0.5—1&10*yr), extreme starbursting. presuniablv metal-poor salaxies.

Our sample with redshifts $1.6 < z < 1.8$ consists of 69 low-mass $\sim 10^8~\msol$ ), young $\sim 0.5-4 \times 10^7~\rm{yr}$ ), extreme starbursting, presumably metal-poor galaxies.

Their coanoving muuber is 3.7«10!Mpe7 two orders of maguitude higher than that of nearby ealaxies with simular EWs (Cardamoueetal.2009).

Their co-moving number is $3.7\times10^{-4}~\rm{Mpc}^{-3}$, two orders of magnitude higher than that of nearby galaxies with similar EWs \citep{cardamone09}.

. The individual star formation rates and the number deusitv combine into 1.7«10%AL.vr|Mpe.ο,

The individual star formation rates and the number density combine into $1.7\times10^{-3} ~\msol ~\rm{yr}^{-1} ~\rm{Mpc}^{-3}$.

This is a: contribution to the total star-formation rate density at Dol.Y occurnnug in galaxies that contribute perhaps 0.1% to the total stellar mass density at that epoch (οιοι,Warietal.2011).

This is a contribution to the total star-formation rate density at $z\sim 1.7$ occurring in galaxies that contribute perhaps $\sim 0.1\%$ to the total stellar mass density at that epoch \citep[e.g.,][]{karim11}.