source
stringlengths 1
2.05k
⌀ | target
stringlengths 1
11.7k
|
---|---|
The peak in IL, between 2 = 90 and 130 AU is due to the heavy depletion of molecules (hat would normally destroy it CCO and No. | The peak in $_3^+$ between $z$ = 90 and 130 AU is due to the heavy depletion of molecules that would normally destroy it CO and $_2$. |
The variation in D/II ratio with z for IL, translates into variations in the D/II ratio of other molecules τα: < 50 AU. but at z > 50 AU is the dominant isotopomer. | The variation in D/H ratio with $z$ for $_3^+$ translates into variations in the D/H ratio of other molecules $^+$ $^+$ $>$ 1 at $z$ $<$ 50 AU, but at $z$ $>$ 50 AU $^+$ is the dominant isotopomer. |
Figure I2. shows the fractional abundances at 250 AU as calculated in Model C which includes both photodesorption aud CRIL ( | Figure \ref{fig:frac_250_c} shows the fractional abundances at 250 AU as calculated in Model C which includes both photodesorption and CRH. ( |
Model D. which includes only photodesorption has a similar abundance distribution to Model C [or z > 60 AU. but with much lower eas phase abundances in the midplane.) | Model D, which includes only photodesorption has a similar abundance distribution to Model C for $z$ $>$ 60 AU, but with much lower gas phase abundances in the midplane.) |
Photodesorption provides an efficient means of returning accreted species to the gas. especially important for those molecules which are too stronely bound. to be affected. by thermal desorpüon. | Photodesorption provides an efficient means of returning accreted species to the gas, especially important for those molecules which are too strongly bound to be affected by thermal desorption. |
It. produces higher fractional abundances than Model Bin the molecular laver. where the UV field is high enough that molecules can be removed from the grains but low enough that they can still survive in the eas. | It produces higher fractional abundances than Model B in the molecular layer, where the UV field is high enough that molecules can be removed from the grains but low enough that they can still survive in the gas. |
Photodesorption also increases the depth of the molecular laver. | Photodesorption also increases the depth of the molecular layer. |
At & = 250 AU. photodesorption affects the vertical distribution of most molecules. even those that can thermally desorb. | At $R$ = 250 AU, photodesorption affects the vertical distribution of most molecules, even those that can thermally desorb. |
Without photodesorption. elements are gracually removed from the gas bv being incorporated into less volatile molecules ccarbou is removed from CO and converted. into hydrocarbon ices. and oxvgen and nitrogen. form IIO. ancl NIL, ices respectively. | Without photodesorption, elements are gradually removed from the gas by being incorporated into less volatile molecules carbon is removed from CO and converted into hydrocarbon ices, and oxygen and nitrogen form $_2$ O and $_3$ ices respectively. |
When phliotodesorption is included. we see large increases in the extent of the CO and No lavers. | When photodesorption is included we see large increases in the extent of the CO and $_2$ layers. |
The most dramatic differences are for the less volatile species. | The most dramatic differences are for the less volatile species. |
For example. the peak fractional abundance of ICN increases to 2.5 x * in Model C from 6 x !* in Model D. NIL; and H35O show dramatic increases in both peak fractional abundance and column densitv. | For example, the peak fractional abundance of HCN increases to 2.5 $\times$ $^{-8}$ in Model C from 6 $\times$ $^{-11}$ in Model B. $_3$ and $_2$ O show dramatic increases in both peak fractional abundance and column density. |
All of these molecules can form rapidly on the grain surface. where deuteration is also efficient and so high column densities of ΠΟ. D5O. NIISD aare also seen. | All of these molecules can form rapidly on the grain surface, where deuteration is also efficient and so high column densities of HDO, $_2$ O, $_2$ D are also seen. |
Similarly. N(HSCO) is increased by grain formation and desorption. although to a lesser exlent than some of the other molecules. | Similarly, $_2$ CO) is increased by grain formation and desorption, although to a lesser extent than some of the other molecules. |
wavelength points. | wavelength points. |
Figure 2 shows two spectra, before they are decomposed with BSS. | Figure \ref{fig:spectra} shows two spectra, before they are decomposed with BSS. |
There is a clearly defined separation between the 11.0 and 11.2 wm emission features as well as distinct features at 12.0, 12.7, 13.5, and 14.1 um with additional features at longer wavelengths. | There is a clearly defined separation between the 11.0 and 11.2 $\mu$ m emission features as well as distinct features at 12.0, 12.7, 13.5, and 14.1 $\mu$ m with additional features at longer wavelengths. |
Among the features at longer wavelengths are the PAH 16.4 jum feature, the blended PAH and C, feature at 17.5 jum, and the pure Ceo feature at 18.9 pum, which are further discussed in ?.. | Among the features at longer wavelengths are the PAH 16.4 $\mu$ m feature, the blended PAH and $_{60}$ feature at 17.5 $\mu$ m, and the pure $_{60}$ feature at 18.9 $\mu$ m, which are further discussed in \citet{2010ApJ...722L..54S}. |
The relative intensities of the H» lines vary with position in the nebula, representing a non-linear component in our spectra. | The relative intensities of the $_2$ lines vary with position in the nebula, representing a non-linear component in our spectra. |
Another non-linear aspect of the spectra is the onset of the dust grain continuum caused by heating from the source star. | Another non-linear aspect of the spectra is the onset of the dust grain continuum caused by heating from the source star. |
These non-linear components cannot be analyzed by BSS methods, since the method demands a linear combination of signals. | These non-linear components cannot be analyzed by BSS methods, since the method demands a linear combination of signals. |
Therefore, we have chosen to exclude the emission at wavelengths greater than 15 um where the continuum from dust grains is present. | Therefore, we have chosen to exclude the emission at wavelengths greater than 15 $\mu$ m where the continuum from dust grains is present. |
We have also clipped the H5 lines everywhere in the cube by hand and replaced them by a linear interpolation. | We have also clipped the $_2$ lines everywhere in the cube by hand and replaced them by a linear interpolation. |
The resulting spectra were then analyzed using the NMF algorithm from ? with both divergence and Euclidian distance optimization, see ? for details. | The resulting spectra were then analyzed using the NMF algorithm from \citet{NMF} with both divergence and Euclidian distance optimization, see \citet{Olivier} for details. |
We investigated the possibility of 3, 4, 5, and 6 component source signals. | We investigated the possibility of 3, 4, 5, and 6 component source signals. |
Irrespective of the particular minimization technique, when attempting to separate 4, 5 and 6 sources, there are 2 or more signals which are very similar (linear combinations of each other) and at least one signal that is pure noise. | Irrespective of the particular minimization technique, when attempting to separate 4, 5 and 6 sources, there are 2 or more signals which are very similar (linear combinations of each other) and at least one signal that is pure noise. |
Therefore, we can conclude that there are 3 significantly different spectral components responsible for the AIBs in NGC 7023-NW. | Therefore, we can conclude that there are 3 significantly different spectral components responsible for the AIBs in NGC 7023-NW. |
This result confirms the findings of ? that there are only 3 source signals in this PDR. | This result confirms the findings of \citet{Olivier} that there are only 3 source signals in this PDR. |
The final extracted spectra are shown in Figure 3.. | The final extracted spectra are shown in Figure \ref{fig:weightfactors}. |
To increase confidence in these results, and ensure that this solution is not a random local minimum, the same analysis was repeated 100 times using different random initializations. | To increase confidence in these results, and ensure that this solution is not a random local minimum, the same analysis was repeated 100 times using different random initializations. |
These 100 spectra shared the same general shape, but varied in intensity, especially in the 11.0 - 11.3 wm region. | These 100 spectra shared the same general shape, but varied in intensity, especially in the 11.0 - 11.3 $\mu$ m region. |
The average spectra of the 100 iterations is plotted with a red line in Figure 3,, and will be used for the remainder of our analysis as the final BSS extracted signals (H matrix). | The average spectra of the 100 iterations is plotted with a red line in Figure \ref{fig:weightfactors}, and will be used for the remainder of our analysis as the final BSS extracted signals (H matrix). |
We can also estimate the error at each point in the spectrum using these results (Figure 3)). | We can also estimate the error at each point in the spectrum using these results (Figure \ref{fig:weightfactors}) ). |
The BSS method has the most difficulty separating the signals in the 11.0 to 11.3 um range due to the strong changing spectral gradients there. | The BSS method has the most difficulty separating the signals in the 11.0 to 11.3 $\mu$ m range due to the strong changing spectral gradients there. |
This results in large errors in this range (see Appendix AppendixA: for discussion on unmixing artifacts). | This results in large errors in this range (see Appendix \ref{sec:appendixa} for discussion on unmixing artifacts). |
Since X«WH, we can estimate W by minimizing ||X—WH]| using a standard least squares minimization. | Since $X \approx WH$, we can estimate $W$ by minimizing $\left \| X-WH \right \|$ using a standard least squares minimization. |
Figure 4 compares the observations in X and the final reconstruction of these observations with WxH. | Figure \ref{fig:residuals} compares the observations in $X$ and the final reconstruction of these observations with $W \times H$. |
The reconstruction is in good agreement with the observations. | The reconstruction is in good agreement with the observations. |
Using the weighting factors that come as a resultant matrix of the above reconstruciton (W), we can map the spatial distribution of each source signal separately (Figure 3)). | Using the weighting factors that come as a resultant matrix of the above reconstruciton $W$ ), we can map the spatial distribution of each source signal separately (Figure \ref{fig:weightfactors}) ). |
The spatial distribution shows clear variation for the three emitting components. | The spatial distribution shows clear variation for the three emitting components. |
Signal 1 is most abundant in the middle of the PDR. | Signal 1 is most abundant in the middle of the PDR. |
Signal 2 has its highest concentration closest to the source star (located at the bottom left of this image) and Signal 3 appears to trace the edge of the PDR farthest from the star. | Signal 2 has its highest concentration closest to the source star (located at the bottom left of this image) and Signal 3 appears to trace the edge of the PDR farthest from the star. |
The well defined regions where each signal is most concentrated implies a physical cause and gives further confidence that these results are not random. | The well defined regions where each signal is most concentrated implies a physical cause and gives further confidence that these results are not random. |
In this section, we will compare our results to the results of the low resolution study of the same region (??) to gain insight about the three extracted signals. | In this section, we will compare our results to the results of the low resolution study of the same region \citep{Olivier,Olivier10} to gain insight about the three extracted signals. |
Creating spatial contours of intensity for each signal allows us to compare the spatial distribution of our signals to the distribution of the three signals from the study of ? of the 5 - 15 um low resolution spectra. | Creating spatial contours of intensity for each signal allows us to compare the spatial distribution of our signals to the distribution of the three signals from the study of \citet{Olivier} of the 5 - 15 $\mu$ m low resolution spectra. |
The contours are created from the IRS-SH spatial distribution maps (Figure 3)) and overlaid with the spatial distributions (represented incolor) of the IRS-SL results (Figure 5)). | The contours are created from the IRS-SH spatial distribution maps (Figure \ref{fig:weightfactors}) ) and overlaid with the spatial distributions (represented incolor) of the IRS-SL results (Figure \ref{fig:contours}) ). |
The three signals extracted here show a strong spatial correlation to the PAH*, PAH®, and VSG maps of ?.. | The three signals extracted here show a strong spatial correlation to the $^+$, $^0$, and VSG maps of \citet{Olivier}. |
The spatial distribution and the results of ? seem to suggest that Signal 2 traces the distribution of PAH cations, Signal 1 the neutral PAH distribution, and Signal 3 the distribution of VSGs. | The spatial distribution and the results of \citet{Olivier} seem to suggest that Signal 2 traces the distribution of PAH cations, Signal 1 the neutral PAH distribution, and Signal 3 the distribution of VSGs. |
Although the spatial distributions of Signal 1, Signal 2, and Signal 3 correlate well with PAH?, PAH*, and VSGs of ?,, there are some small discrepancies, in particular, for the PAH? map. | Although the spatial distributions of Signal 1, Signal 2, and Signal 3 correlate well with $^0$ , $^+$, and VSGs of \citet{Olivier}, there are some small discrepancies, in particular, for the $^0$ map. |
As discussed in ?,, the degradation of spatial or spectral resolution always implies a loss in the quality of the NMF efficiency. | As discussed in \citet{Olivier10}, the degradation of spatial or spectral resolution always implies a loss in the quality of the NMF efficiency. |
Since ? have a higher spatial resolution, while here we have a higher spectral resolution, none of the data-sets can be considered "better" and small discrepancies between the results of NMF are expected. | Since \citet{Olivier} have a higher spatial resolution, while here we have a higher spectral resolution, none of the data-sets can be considered “better” and small discrepancies between the results of NMF are expected. |
Figure 6 compares the low-resolution source signal spectra of ? to the high-resolution source spectra obtained here (Figure 3)). | Figure \ref{fig:spec_comp} compares the low-resolution source signal spectra of \citet{Olivier} to the high-resolution source spectra obtained here (Figure \ref{fig:weightfactors}) ). |
The low resolution extracted spectra share all the major features with the high resolution spectra, specifically the broad 11.3 um emission feature in the VSGs, the 11.2 jum and 12.7 um emission features in the neutral PAHs, and the 11.0 pum and broad 12.7 uum emission features in the PAH cations. | The low resolution extracted spectra share all the major features with the high resolution spectra, specifically the broad 11.3 $\mu$ m emission feature in the VSGs, the 11.2 $\mu$ m and 12.7 $\mu$ m emission features in the neutral PAHs, and the 11.0 $\mu$ m and broad 12.7 $\mu$ m emission features in the PAH cations. |
Although evidence for the 11.0 uim feature was present in the IRS-SL observations of ?,, it was notimmediately attributed | Although evidence for the 11.0 $\mu$ m feature was present in the IRS-SL observations of \citet{Olivier10}, , it was notimmediately attributed |
shotoionization code Cloudy 8.00 (Ferland2002). to build a erid of models covering a large space of nebular parameters. | photoionization code Cloudy 8.00 \citep{ferland02}
to build a grid of models covering a large space of nebular parameters. |
In hese models. a stellar cluster was assumed to be responsible for he ionization of the nebulae. with a spectral energy distribution (SED) obtained using Sfarburs/99 (Leithereretal.1999). | In these models, a stellar cluster was assumed to be responsible for the ionization of the nebulae, with a spectral energy distribution (SED) obtained using $Starburst99$ \citep{leitherer99}. |
.. We built models with stellar cluster formed by instantaneous burst with Salpeter initial mass function (v= —2.35). lower and upper stellar mass limits of O.1 AZ, and 100 Μι. respectively. and age of 2.5 Tyr. | We built models with stellar cluster formed by instantaneous burst with Salpeter initial mass function $\alpha = -2.35$ ), lower and upper stellar mass limits of 0.1 $M_{\odot}$ and 100 $M_{\odot}$, respectively, and age of 2.5 Myr. |
Other papers that have considered stellar clusters as ionizing sources in order to reproduce strong forbidden lines of region (Dors&Copetti2005:Dopitaetal.2000:StasinskaIzotov2003:Bresolinetal.1999:Copetti1985). have derived about the same age for star-forming regions (i.e. 1-3 Myr). | Other papers that have considered stellar clusters as ionizing sources in order to reproduce strong forbidden lines of region \citep{dors05, dopita00, stasinska03, bresolin99, copetti85} have derived about the same age for star-forming regions (i.e. 1-3 Myr). |
Similar ages have also been found from optical photometric data of giant regions (e.g. Mayya&Prabhu 19969). | Similar ages have also been found from optical photometric data of giant regions (e.g. \citealt{mayya96}) ). |
Selection ettects may explain this limited range of ages. | Selection effects may explain this limited range of ages. |
regions younger than about | Myr are ditheult to be detected in the optical. because they are generally embedded in dusty molecular clouds which cause considerable optical extinction. | regions younger than about 1 Myr are difficult to be detected in the optical, because they are generally embedded in dusty molecular clouds which cause considerable optical extinction. |
Nebulae older than about 5 Myr are also ditheult to be observed because their original massive stars have cooled or are dead (Dopitaetal.2000:Garefa-Vargas1996:Copettietal. [985). | Nebulae older than about 5 Myr are also difficult to be observed because their original massive stars have cooled or are dead \citep{dopita00, garcia96, copetti85}. |
. We used the stellar evolution models from the Geneva group with high mass-loss rates and without stellar rotation (Meynetetal.19943. | We used the stellar evolution models from the Geneva group with high mass-loss rates and without stellar rotation \citep{meynet94}. |
.. The non-LTE atmosphere model of Pauldrachetal.(2001). was assumed in the models. | The non-LTE atmosphere model of \citet{pauldrach01} was assumed in the models. |
If LTE atmosphere model is assumed instead of non-LTE model a lower ionization degree is produced in the hypothetical nebulae (Dors&Copetti2003:StasifiskaSchaerer 1997). | If LTE atmosphere model is assumed instead of non-LTE model a lower ionization degree is produced in the hypothetical nebulae \citep{dors03,stasinska97}. |
. This would attect mainly ionization parameter rather than metallicity determinations via strong-line methods. | This would affect mainly ionization parameter rather than metallicity determinations via strong-line methods. |
The models were built having ionization parameter ranging from logU=—1.5 to —3.5 (with a bin size of 0.5 dex). metallicities (traced by the oxygen abundance) Z- 0.04. 0.02. 0.008. 0.004. 0.001. plane-parallel geometry. and electron density of NM.=200em?. | The models were built having ionization parameter ranging from $\log U =-1.5$ to $-3.5$ (with a bin size of 0.5 dex), metallicities (traced by the oxygen abundance) $Z$ = 0.04, 0.02, 0.008, 0.004, 0.001, plane-parallel geometry, and electron density of $N_{\rm e}= 200 \: \rm cm^{-3}$. |
This electron density value is typical of not evolved regions (Copettietal.2000). | This electron density value is typical of not evolved regions \citep{copetti00}. |
The abundances of heavy metals in the nebula ἰς scaled. linearly to the solar. metal. composition fiough the comparison. of the oxygen abundances. with the exception of the N abundance. which was taken from t15 relation N/Oj2log(0.0324--120OO/H). of. Vila-Costas. (1993). | The abundances of heavy metals in the nebula is scaled linearly to the solar metal composition through the comparison of the oxygen abundances, with the exception of the N abundance, which was taken from the relation O/H) of \citet{vila93}. |
The solar composition (Z= 0.02) refers to AIendePrieto(2001) and correspond to |2+log(O/H)= 8.69. | The solar composition $Z$ = 0.02) refers to \citet{allende01} and correspond to 12+log(O/H)= 8.69. |
T1e presence of internal dust was considered and the grain abundances (vanHoofetal.2001) were also linearly scaled with the oxygen abundance. | The presence of internal dust was considered and the grain abundances \citep{hoof01} were also linearly scaled with the oxygen abundance. |
To take depletion of refractary elements onto dust grains into account the abundances of the elements Mg. Al. Ca. Fe. i. and Na were reduced by a factor of 10. and Si by a factor of 2 (Garnettetal.1995). relative to the adopted abundances in each model. | To take depletion of refractary elements onto dust grains into account the abundances of the elements Mg, Al, Ca, Fe, Ni, and Na were reduced by a factor of 10, and Si by a factor of 2 \citep{garnett95} relative to the adopted abundances in each model. |
The solar metallicity for the stars from the Geneva evolutionary tracks. which corresponds to the old solar oxygen abundance value. [l2+log¢(O/H)= 8.87 | The solar metallicity for the stars from the Geneva evolutionary tracks, which corresponds to the old solar oxygen abundance value [12+log(O/H)= 8.87 |
The solar metallicity for the stars from the Geneva evolutionary tracks. which corresponds to the old solar oxygen abundance value. [l2+log¢(O/H)= 8.87. | The solar metallicity for the stars from the Geneva evolutionary tracks, which corresponds to the old solar oxygen abundance value [12+log(O/H)= 8.87 |
Dressler Gunn (1983: 1992) discovered galaxies with mysterious spectra while investigating high redshift galaxy clusters. | Dressler Gunn (1983; 1992) discovered galaxies with mysterious spectra while investigating high redshift galaxy clusters. |
The galaxies had strong Balmer absorption lines with no emission in [OIL]. | The galaxies had strong Balmer absorption lines with no emission in [OII]. |
These galaxies are named "EA" galaxies since their spectra looked like a superposition of that of elliptical galaxies (Mgsj75. Fez»zy and Cazo31.315s absorption lines) and that of A-type stars (strong Balmerabsorption)!. | These galaxies are named “E+A” galaxies since their spectra looked like a superposition of that of elliptical galaxies $_{5175}$ , $_{5270}$ and $_{3934,3468}$ absorption lines) and that of A-type stars (strong Balmer. |
. Since the lifetime of A-type stars is about | Gyr. the existence of strong Balmer absorption lines shows that these galaxies have experienced starburst within the last Gyr. | Since the lifetime of A-type stars is about 1 Gyr, the existence of strong Balmer absorption lines shows that these galaxies have experienced starburst within the last Gyr. |
However. these galaxies do not show any sign of on-going star formation as non-detection in [OIL] emission lines indicates. | However, these galaxies do not show any sign of on-going star formation as non-detection in [OII] emission lines indicates. |
Therefore. E+A galaxies are interpreted as a post-starburst galaxy. that is. a galaxy which has undergone a truncated starburst (Dressler Gunn 1983. 1992: Couch Sharples 1987: MacLaren. Ellis. Couch 1988; Newberry Boroson Kirshner 1990; Fabricant. MeClintock. Bautz 1991: Abraham et al. | Therefore, E+A galaxies are interpreted as a post-starburst galaxy, that is, a galaxy which has undergone a truncated starburst (Dressler Gunn 1983, 1992; Couch Sharples 1987; MacLaren, Ellis, Couch 1988; Newberry Boroson Kirshner 1990; Fabricant, McClintock, Bautz 1991; Abraham et al. |
1996). | 1996). |
The reason why they underwent a starburst then abruptly stopped still remains one of the mysteries in galaxy evolution. | The reason why they underwent a starburst then abruptly stopped still remains one of the mysteries in galaxy evolution. |
Since starburst-driven galaxy transformations are an important part of galaxy formation theories (e.g. Somerville et al. | Since starburst-driven galaxy transformations are an important part of galaxy formation theories (e.g. Somerville et al. |
2001). it is Important to understand what triggered the starburst in these galaxies and. perhaps more importantly. why star formation subsequently ceased so abruptly. | 2001), it is important to understand what triggered the starburst in these galaxies and, perhaps more importantly, why star formation subsequently ceased so abruptly. |
At first. E-A galaxies were found in cluster regions. both in low redshift clusters (Franx 1993; Caldwell et al. | At first, E+A galaxies were found in cluster regions, both in low redshift clusters (Franx 1993; Caldwell et al. |
1993. 1996; Caldwell Rose 1997; Castander et al. | 1993, 1996; Caldwell Rose 1997; Castander et al. |
2001: Rose et al. | 2001; Rose et al. |
2001) and high redshift clusters (Sharples et al. | 2001) and high redshift clusters (Sharples et al. |
1985: Lavery Henry 1986: Couch Sharples 1987: Broadhurst. Ellis. Shanks 1988: Fabricant. McClintock. Bautz 1991; Belloni et al. | 1985; Lavery Henry 1986; Couch Sharples 1987; Broadhurst, Ellis, Shanks 1988; Fabricant, McClintock, Bautz 1991; Belloni et al. |
1995: Barger et al. | 1995; Barger et al. |
1996; Fisher et al. | 1996; Fisher et al. |
1998: Morris et al. | 1998; Morris et al. |
1998: Couch et al. | 1998; Couch et al. |
1998: Dressler et al. | 1998; Dressler et al. |
1999). | 1999). |
Therefore. a cluster specific phenomenon was thought to be responsible for the violent star formation history of E+A galaxies. | Therefore, a cluster specific phenomenon was thought to be responsible for the violent star formation history of E+A galaxies. |
A ram-pressure stripping model (Spitzer Baade 1951. Gunn Gott 1972. Farouki Shapiro 1980; Kent 1981: Abadi. Moore Bower 1999; Fujita Nagashima 1999; Quilts. Moore Bower 2000: Fujita 2004: Fujita Goto 2004) may first accelerate star formation of cluster galaxies and later turn it off as well as tides from the cluster potential (e.g.. Fujita 1998) and the evaporation of the cold gas (e.g.. Fujita 2004). | A ram-pressure stripping model (Spitzer Baade 1951, Gunn Gott 1972, Farouki Shapiro 1980; Kent 1981; Abadi, Moore Bower 1999; Fujita Nagashima 1999; Quilis, Moore Bower 2000; Fujita 2004; Fujita Goto 2004) may first accelerate star formation of cluster galaxies and later turn it off as well as tides from the cluster potential (e.g., Fujita 1998) and the evaporation of the cold gas (e.g., Fujita 2004). |
However. recent large surveys of the nearby universe found many E+A galaxies in the field regions (Goto 2003: Goto et al. | However, recent large surveys of the nearby universe found many E+A galaxies in the field regions (Goto 2003; Goto et al. |
2003a. GO3 hereafter: Quintero et al. | 2003a, G03 hereafter; Quintero et al. |
2004). | 2004). |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.