We study the size of BLRs of low luminosity Active Galactic Nuclei, also called 'dwarf AGN', defined as having ($L_{H\alpha}\le10^{41}{\rm erg\cdot s^{-1}}$)). We more than double the sample size analyzed previously (Wang & Zhang 2003, hereafter Paper I). In this study we first confirm our previous result that the sizes of BLRs of low luminosity AGN are larger than the ones expected from the empirical relation $R_{BLRs} - L_{H\alpha}$ valid for 'normal' AGN: Seyfert 1s and quasars, except for the objects with accretion rate $\dot{m_{H\alpha}}>10^{-5.5}$. Second, we find a positive correlation between the line width of the narrow emission line (as tracer of velocity dipersion and thus bulge and black hole mass) and the size of BLRs for both normal and low luminosity AGN. In this paper we find a non-linear dependence of the BLRs sizes of low luminosity AGN on BH masses. We also show that their sizes of BLRs are more strongly dominated by the 'specific accretion rate' $\dot{m_{H\alpha}}$ defined as $\dot{m_{H\alpha}} = L_{H\alpha}/L_{Edd}$, than by the masses of their cetral black holes. As an expected result, the distance of emission regions of low-ionization broad H$\alpha$ of NGC 4395 should be consistent with the value from the empirical relation of $R_{BLRs} - L_{H\alpha}$, according to the high accretion rate.

We report on the radio quiet quasar SDSS J2125-0813 which obviously emits optical FeII emission lines and double-peaked broad Balmer emission lines. Using the accretion disk model for double-peaked broad low-ionization emission lines, we reproduce the composite line spectra at the optical band between 4100$\AA$ and 5600$\AA$. Broad FeII emission lines can be fit simultaneously with the broad H$\beta$ and MgI lines, such that all broad lines have an elliptical disk profile with the same disk parameters. This results indicates that the optical FeII emission lines originate from the accretion disk near the central black hole which produces the double-peaked broad Balmer emission lines. Furthermore, we find that the object has dimensionless accretion rate $\dot{m}\sim0.4-0.6$ which is much larger than accretion rate for ADAF mode, and that the energy budget of the accretion disk is enough to power the double-peaked broad Balmer emission lines.