1. Based on observational data on evolution of quasars and galaxies of different types along with the results of numerical simulations we make a conclusion that on low redshifts ($z<0.5$) QSOI/II objects in massive elliptical and spiral galaxies with
classical bulges cannot be in late single activity event (be primordial). Instead of it they have had events of activity earlier in their evolution. It means that their presence on low redshifts is connected with the recurrence phenomenon, sequential wet minor mergings, because timescale of the activity does not exceed several units of $10^7$ years. 2. We define a new class - AGN III as active galactic nuclei in isolated late-type spirals with low-mass rapidly rotating pseudobulges. We also state that only such objects can be in the primordial phase of activity at low redshifts. Black holes in such galaxies have masses $M_{BH}<10^7M_odot$ and also, probably very high spin. Such properties can explain their peculiar emission spectra. A good representative of AGN III might be the galaxies with narrow (${rm FWHM}(Hbeta)leq1200$ km/s) broad permitted emission lines - NLS. It is believed that their black hole masses are less than $M_{BH}<10^7M_odot$ and their host galaxies have pseudobulges instead of the classical ones. Because host galaxies of NLS have pseudobulges and BLS (Broad-Line Seyfert galaxies) have classical bulges these two types of objects cannot have evolutionary connection. Presumably, the parent population of NLS are the quasars of population A.
Using standard cosmological model we show that the cores of ultra-compact radio sources observed with ground-based Very Long Baseline Interferometry (VLBI) on the angular scales of milliarcseconds cannot be used as a reasonable standard unit of linea
r size. Luminosity - linear size correlation obtained by many authors for ultra compact radio sources has different origin than that for the radio galaxies and quasars on the angular scales of arcminutes. It is just the manifestation of the fact that ground-based VLBI networks are unable to resolve VLBI-cores and the Malmquist bias presents. Thus, the cores of compact radio sources cant be used as standard rods at least with resolution offered by ground-based VLBI. This conclusion is illustrated on 15 GHz VLBA sample of radio sources.
