We perform a stacking analysis of the HI spectra from the Arecibo Legacy FastALFA (ALFALFA) survey for optically-selected local galaxies from the SloanDigital Sky Survey (SDSS) to study the average gas fraction of galaxies atfixed stellar mass ($M_*$) and star formation rate (SFR). We first confirm thatthe average gas fraction strongly depends on the stellar mass and SFR of hostgalaxies; massive galaxies tend to have a lower gas fraction, and activelystar-forming galaxies show higher gas fraction, which is consistent with manyprevious studies. Then we investigate the morphological dependence of the HIgas mass fraction at fixed $M_*$ and SFR to minimize the effects of theseparameters. We use three morphological classifications based on parametricindicator (S\'{e}rsic index), non-parametric indicator (C-index), and visualinspection (smoothness from the Galaxy Zoo 2 project) on the optical image. Wefind that there is no significant morphological dependence of the HI gas massfraction at fixed $M_*$ and SFR when we use C-index. In comparison, thereexists a hint of diminishment in the HI gas mass fraction for "smooth" galaxiescompared with "non-smooth" galaxies. We find that the visual smoothness issensitive to the existence of small-scale structures in a galaxy. Our resultsuggests that even at fixed $M_*$ and SFR, the presence of such small-scalestructures (seen in the optical image) is linked to their total HI gas content.

We compare the relations among various integrated characteristics of ~25,000low-redshift (z<1.0) compact star-forming galaxies (CSFGs) from Data Release 16(DR16) of the Sloan Digital Sky Survey (SDSS) and of high-redshift (z>1.5)star-forming galaxies (SFGs) with respect to oxygen abundances, stellar massesM*, far-UV absolute magnitudes M(FUV), star-formation rates SFR and specificstar-formation rates sSFR, Lyman-continuum photon production efficiencies(xi_ion), UV continuum slopes \beta, [OIII]5007/[OII]3727 and[NeIII]3868/[OII]3727 ratios, and emission-line equivalent widthsEW([OII]3727), EW([OIII]5007), and EW(H\alpha). We find that the relations forlow-z CSFGs with high equivalent widths of the H\beta emission line,EW(H\beta)>100A, and high-z SFGs are very similar, implying close physicalproperties in these two categories of galaxies. Thus, CSFGs are likelyexcellent proxies for the SFGs in the high-z Universe. They also extend togalaxies with lower stellar masses, down to ~10^6 Msun, and to absolute FUVmagnitudes as faint as -14 mag. Thanks to their proximity, CSFGs can be studiedin much greater detail than distant SFGs. Therefore, the relations between theintegrated characteristics of the large sample of CSFGs studied here can provevery useful for our understanding of high-z dwarf galaxies in futureobservations with large ground-based and space telescopes.

Using the Large Binocular Telescope (LBT)/Multi-Object Dual Spectrograph(MODS), we have obtained optical spectroscopy of one of the most metal-poordwarf star-forming galaxies (SFG) in the local Universe, J2229+2725. Thisgalaxy with a redshift z=0.0762 was selected from the Data Release 16 (DR16) ofthe Sloan Digital Sky Survey (SDSS). Its properties derived from the LBTobservations are most extreme among SFGs in several ways. Its oxygen abundance12+logO/H=7.085+/-0.031 is among the lowest ever observed for a SFG. With itsvery low metallicity, an absolute magnitude Mg=-16.39 mag, a low stellar massMstar=9.1x10^6 Msun and a very low mass-to-light ratio Mstar/Lg~0.0166 (insolar units), J2229+2725 deviates strongly from the luminosity-metallicityrelation defined by the bulk of the SFGs in the SDSS. J2229$+$2725 has a veryhigh specific star-formation rate sSFR~75 Gyr^-1, indicating very activeongoing star formation. Three other features of J2229+2725 are most striking,being the most extreme among lowest-metallicity SFGs: 1) a ratioO32=I([OIII]5007)/I([OII]3727)~53, 2) an equivalent width of the Hbeta emissionline EW(Hbeta) of 577A, and 3) an electron number density of ~1000 cm^-3. Theseproperties imply that the starburst in J2229+2725 is very young. Using theextremely high O32 in J2229+2725, we have improved the strong-line calibrationfor the determination of oxygen abundances in the most metal-deficientgalaxies, in the range 12 + logO/H<7.3.

Using a currently most representative sample of 477 late-type galaxies within11 Mpc of the Milky Way with measured star-formation rates ($SFR$s) from thefar ultraviolet ($FUV$) and H$\alpha$ emission line fluxes, we select galaxieswith the extreme ratios: $SFR(H\alpha)/SFR(FUV) > 2$ and $SFR(H\alpha)/SFR(FUV)< 1/20$. Each subsample amounts to $\sim5$\% of the total number and consistsof dwarf galaxies with the stellar masses $M^*/M_{\odot} = (5.5 - 9.5)$~dex. Inspite of a huge difference in their $SFR(H\alpha)$ activity on a scale of$\sim10$~ Myr, the temporarily "excited" and temporarily "quiescent" galaxiesfollow one and the same relation between $SFR(FUV)$ and $M^*$ on a scale of$\sim100$~Myr. Their average specific star-formation rate $\log[SFR(FUV)/M^*] =-10.1\pm0.1$ (yr$^{-1}$) coinsides with the Hubble parameter $\log(H_0)=-10.14$ (yr$^{-1}$). On a scale of $t \sim10$~Myr, variations of $SFR$ have amoderate flash amplitude of less than 1 order above the main-sequence and afading amplitude to 2 orders below the average level. In general, bothtemporarily excited and temporarily quiescent galaxies have almost similar gasfractions as normal main-sequence galaxies, being able to maintain the currentrate of star-formation on another Hubble time scale. Ranging the galaxiesaccording to the density contrast produced by the nearest massive neighborexhibits only a low average excess of $SFR$ caused by tidal interactions.

For years, the extragalactic community has divided galaxies in two distinctpopulations. One of them, featuring blue colours, is actively forming stars,while the other is made up of "red-and-dead" objects with negligible starformation. Yet, are these galaxies really dead? Here we would like to highlightthat, as previously reported by several independent groups, state-of-the-artcosmological numerical simulations predict the existence of a large number ofquenched galaxies that have not formed any star over the last few Gyr. Incontrast, observational measurements of large galaxy samples in the nearbyUniverse suggest that even the most passive systems still form stars at someresidual level close to $sSFR\sim10^{-12}~\text{yr}^{-1}$. Unfortunately,extremely low star formation poses a challenge for both approaches. We concludethat, at present, the fraction of truly dead galaxies is still an importantopen question that must be addressed in order to understand galaxy formationand evolution.