Optical observations provide convincing evidence that the optical phase of the Crab pulsar follows the radio one closely. Since optical data do not depend on dispersion measure variations, they provide a robust and independent confirmation of the radio timing solution. The aim of this paper is to find a global mathematical description of Crab pulsar's phase as a function of time for the complete set of published Jodrell Bank radio ephemerides (JBE) in the period 1988-2014. We apply the mathematical techniques developed for analyzing optical observations to the analysis of JBE. We break the whole period into a series of episodes and express the phase of the pulsar in each episode as the sum of two analytical functions. The first function is the best-fitting local braking index law, and the second function represents small residuals from this law with an amplitude of only a few turns, which rapidly relaxes to the local braking index law. From our analysis, we demonstrate that the power law index undergoes "instantaneous" changes at the time of observed jumps in rotational frequency (glitches). We find that the phase evolution of the Crab pulsar is dominated by a series of constant braking law episodes, with the braking index changing abruptly after each episode in the range of values between 2.1 and 2.6. Deviations from such a regular phase description behave as oscillations triggered by glitches and amount to fewer than 40 turns during the above period, in which the pulsar has made more than 2.0e10 turns. Our analysis does not favor the explanation that glitches are connected to phenomena occurring in the interior of the pulsar. On the contrary, timing irregularities and changes in slow down rate seem to point to electromagnetic interaction of the pulsar with the surrounding environment.

The Crab nebula pulsar was observed in 2009 January and December with a novel very fast optical photon counter, Iqueye, mounted at the ESO 3.5 m New Technology Telescope. Thanks to the exquisite quality of the Iqueye data, we computed accurate phase coherent timing solutions for the two observing runs and over the entire year 2009. Our statistical uncertainty on the determination of the phase of the main pulse and the rotational period of the pulsar for short (a few days) time intervals are $\approx 1 \, \mu$s and ~0.5 ps, respectively. Comparison with the Jodrell Bank radio ephemerides shows that the optical pulse leads the radio one by ~240 $\mu$s in January and ~160 $\mu$s in December, in agreement with a number of other measurements performed after 1996. A third-order polynomial fit adequately describes the spin-down for the 2009 January plus December optical observations. The phase noise is consistent with being Gaussian distributed with a dispersion $\sigma$ of $\approx 15 \, \mu$s in most observations, in agreement with theoretical expectations for photon noise-induced phase variability.

We observed the Crab pulsar in October 2008 at the Copernico Telescope in Asiago - Cima Ekar with the optical photon counter Aqueye (the Asiago Quantum Eye) which has the best temporal resolution and accuracy ever achieved in the optical domain (hundreds of picoseconds). Our goal was to perform a detailed analysis of the optical period and phase drift of the main peak of the Crab pulsar and compare it with the Jodrell Bank ephemerides. We determined the position of the main peak using the steepest zero of the cross-correlation function between the pulsar signal and an accurate optical template. The pulsar rotational period and period derivative have been measured with great accuracy using observations covering only a 2 day time interval. The error on the period is 1.7 ps, limited only by the statistical uncertainty. Both the rotational frequency and its first derivative are in agreement with those from the Jodrell Bank radio ephemerides archive. We also found evidence of the optical peak leading the radio one by ~230 microseconds. The distribution of phase-residuals of the whole dataset is slightly wider than that of a synthetic signal generated as a sequence of pulses distributed in time with the probability proportional to the pulse shape, such as the average count rate and background level are those of the Crab pulsar observed with Aqueye. The counting statistics and quality of the data allowed us to determine the pulsar period and period derivative with great accuracy in 2 days only. The time of arrival of the optical peak of the Crab pulsar leads the radio one in agreement with what recently reported in the literature. The distribution of the phase residuals can be approximated with a Gaussian and is consistent with being completely caused by photon noise (for the best data sets).

Tidal effects on clumps of material during random non-stationary accretion onto a black hole produce phenomena with distinct temporal characteristics in observed light-curves. During such non-stationary accretion events, the shape of the accreting object evolves in time, and observable quasi-periodic phenomena with variable quasi-periods are produced. A number of characteristic light-curves, obtained with numerical simulations, will be shown. Their relevance to observed phenomena will be briefly discussed.

The discovery that the Galactic centre emits flares at various wavelengths represents a puzzle concerning their origin, but at the same time it is a relevant opportunity to investigate the environment of the nearest super-massive black hole. In this paper we shall review some of our recent results concerning the tidal evolution of the orbits of low mass satellites around black holes, and the tidal effect during their in-fall. We show that tidal interaction can offer an explanation for transient phenomena like near infra-red and X-ray flares from Sgr A*.

We present the telltale signature of the tidal capture and disruption of an object by a massive black hole in a galactic centre. As a result of the interaction with the black hole's strong gravitational field, the object's light curve can flare-up with characteristic time of the order of 100 sec \times (M_{bh} / 10^6 M_{Solar}). Our simulations show that general relativity plays a crucial role in the late stages of the encounter in two ways: (i) due to the precession of perihelion, tidal disruption is more severe, and (ii) light bending and aberration of light produce and enhance flares seen by a distant observer. We present our results for the case of a tidally disrupted Solar-type star. We also discuss the two strongest flares that have been observed at the Galactic centre. Although the first was observed in X-rays and the second in infra-red, they have almost identical light curves and we find it interesting that it is possible to fit the infra-red flare with a rather simple model of the tidally disrupted comet-like or planetary object. We discuss the model and possible scenarios how such an event can occur.

Low Mass X-ray Binaries (LMXBs) with either a black hole or a neutron star show power spectra characterised by Quasi Periodic Oscillations (QPOs). Twin peak high frequency QPOs are characterised by frequencies that are typical for matter orbiting within 10 r_g from the compact object. We consider clumps of material orbiting a Schwarzschild black hole, that are deformed by tidal interaction. We present some preliminary calculations of corresponding light curves and power spectra. We were able to fit the simulated power spectra with the high frequency part of the power spectra observed in the LMXB XTE J1550-564 containing a black hole. Our numerical simulations reproduce the twin high frequency QPOs and the power-law. The lower peak corresponds to the Keplerian frequency, the upper one to the sum of the Keplerian and the radial frequency.

Tidal disruption by massive black holes is a phenomenon, during which a large part of gravitational energy can be released on a very short time-scale. The time-scales and energies involved during X-ray and IR flares observed in Galactic centre suggest that they may be related to tidal disruption events. Furthermore, aftermath of a tidal disruption of a star by super-massive black hole has been observed in some galaxies, e.g. RX J1242.6-1119A. All these discoveries increased the demand for tools for tidal disruption study in curved space-time. Here we summarise our study of general relativistic effects on tidal deformation of stars and compact objects.

The compact radio source Sagittarius A (Sgr A) at the centre of our Galaxy harbours a supermassive black hole, whose mass has been measured from stellar orbital motions. Sgr A is therefore the nearest laboratory where super-massive black hole astrophysics can be tested, and the environment of black holes can be investigated. Since it is not an active galactic nucleus, it also offers the possibility of observing the capture of small objects that may orbit the central black hole. We study the effects of the strong gravitational field of the black hole on small objects, such as a comet or an asteroid. We also explore the idea that the flares detected in Sgr A might be produced by the final accretion of single, dense objects with mass of the order of 10^20 g, and that their timing is not a characteristic of the sources, but rather of the space-time of the central galactic black hole in which they are moving. We find that tidal effects are strong enough to melt the solid object, and present calculations of the temporal evolution of the light curve of infalling objects as a function of various parameters. Our modelling of tidal disruption suggests that during tidal squeezing, the conditions for synchrotron radiation can be met. We show that the light curve of a flare can be deduced from dynamical properties of geodesic orbits around black holes and that it depends only weakly on the physical properties of the source.

[Abridged] We present new VLT ISAAC spectra for 30 quasars, which we combine with previous data to yield a sample of 53 intermediate redshift (z ~ 0.9 - 3.0) sources. The sample is used to explore properties of prominent lines in the Hbeta spectral region of these very luminous quasars. We find two major trends: (1) a systematic increase of minimum FWHM Hbeta with luminosity (discussed in a previous paper). This lower FWHM envelope is best fit assuming that the narrowest sources radiate near the Eddington limit, show line emission from a virialized cloud distribution, and obey a well defined broad line region size vs. luminosity relation. (2) A systematic decrease of equivalent width of [OIII] (from W ~ 15 to ~ 1 A) with increasing source bolometric luminosity (from log L ~ 43 to log L ~ 49). Further identified trends required discrimination between so-called Population A and B sources. We generate median composite spectra in six luminosity bins Pop. A sources show reasonably symmetric Lorentzian Hbeta profiles at all luminosities while Pop. B sources require two component fits involving an unshifted broad and a redshifted very broad component. Very broad Hbeta increases in strength with increasing L while the broad component remains constant resulting in an apparent "Baldwin effect" with equivalent width decreasing from W ~ 80 to ~ 20 A over our sample luminosity range. The roughly constant equivalent width shown by the Hbeta very broad component implies production in optically-thick, photoionized gas. The onset of the redshifted very broad component appears to be a critical change that occurs near the Pop. A-B boundary at FWHM Hbeta ~ 4000 km/s which we relate to a critical Eddington ratio (~ 0.2 +/- 0.1).

Low-mass satellites, like asteroids and comets, are expected to be present around the black hole at the Galactic center. We consider small bodies orbiting a black hole, and we study the evolution of their orbits due to tidal interaction with the black hole. In this paper we investigate the consequences of the existence of plunging orbits when a black hole is present. We are interested in finding the conditions that exist when capture occurs. The main difference between the Keplerian and black hole cases is in the existence of plunging orbits. Orbital evolution, leading from bound to plunging orbits, goes through a final unstable circular orbit. On this orbit, tidal energy is released on a characteristic black hole timescale. This process may be relevant for explaining how small, compact clumps of material can be brought onto plunging orbits, where they may produce individual short duration accretion events. The available energy and the characteristic timescale are consistent with energy released and the timescale typical of Galactic flares.

In recent years the case for the presence of 3-4 10^6 M_sun black hole in our Galactic Center has gained strength from results of stellar dynamics observations and from the detection of several rapid X-ray and IR flares observed in the Sagittarius A* from 2000 to 2004. Here we explore the idea that such flares are produced when the central black hole tidally captures and disrupts a small body - e.g. a comet or an asteroid.

We derive black hole masses for a sample of about 300 AGNs in the redshift range 0 < z < 2.5. We use the same virial velocity measure (FWHM Hbeta broad component) for all sources which represents a significant improvement over previous studies. We review methods and caveats for determining AGN black hole masses via the virial assumption for motions in the gas producing low ionization broad emission lines. We derive a corrected FWHM measure for the broad component of H-beta that better estimates the virialized line emitting component by comparing our FWHM measures with a sample of reverberated sources with H-beta radial velocity dispersion measures. We also consider the FWHM of the FeII 4570 blend as a potential alternative velocity estimator. We find a range of black hole mass between log M ~ 6 - 10, where the black hole mass M is in solar masses. Estimates using corrected FWHM, as well as FWHM(Fe II) measures, reduce the number of sources with log M > 9.5 and suggest that extremely large M values (log M >~ 10) may not be realistic. Derived Eddington ratio values values show no evidence for a significant population of super-Eddington radiators especially after correction is made for sources with extreme orientation to our line of sight. Sources with FWHM(Hbeta broad component) <~ 4000 km/s show systematically higher Eddington ratio and lower M values than broader lined AGNs (including almost all radio-loud sources).

We attempt to characterize the geometry of Broad Absorption Line (BAL) QSOs by studying a low redshift sample of 12 sources. We find that the majority of these sources are Population A quasars as defined in (broad H-beta FWHM > 4000 kms/s; Sulenticet al 2000). A possible correlation between terminal velocity and absolute V magnitude suggests that the bolometric luminosity to black hole mass ratio \lm is a governing factor with classical BAL sources showing the highest values. CIV 1549 emission in classical BAL sources shows a profile blueshift that supports a disk wind/outflow scenario with a half opening angle of < 50 deg. Observation of ``secondary'' mini-BAL features in the CIV 1549 emission profile motivates us to model BALs with an additional component that may be involved with the BLR outflow and co-axial with the accretion disk.

We present composite UV spectra for low redshift Type 1 AGN binned to exploit the information content of the Eigenvector 1 (E1) parameter space. Composite spectra allow a decomposition of the CIV1549 line profile - one of the strongest high-ionization lines. The simplest CIV decomposition into narrow (NLR), broad (BLR) and very broad (VBLR) components suggests that different components have an analog in Hb with two major exceptions. VBLR emission is seen only in population B (FWHM(Hb)>4000 km/s) sources. A blue shifted/asymmetric BLR component is seen only in pop. A (FWHM(Hb)<4000 km/s) HIL such as CIV. The blueshifted component is thought to arise in a wind or outflow. Our analysis suggests that such a wind can only be produced in pop. A (almost all radio-quiet) sources where the accretion rate is relatively high. Comparison between broad UV lines in radio-loud (RL) and radio-quiet (RQ) sources shows few significant differences. Clear evidence is found for a narrow CIV component in most radio-loud sources. We find also some indirect indications that the black hole (BH) spin, rather than BH mass or accretion rate is a key trigger in determining whether an object will be RL or RQ. We find a ten-fold decrease in EW CIV with Eddington ratio (decreasing from ~1 to \~0.01) while NV shows no change. These trends suggest a luminosity-independent "Baldwin effect" where the physical driver may be the Eddington ratio.

We present high S/N spectra of the H-beta region in 17 intermediate redshift (0.85 le z le 2.5) quasars. The spectra represent first results of our campaign to test the redshift/luminosity robustness of the so-called Eigenvector 1 (E1) parameter space as developed for low redshift AGN in Sulentic et al. (2000). The new quasars span the luminosity range -26 ge M_B ge -29 while most of our low redshift sample (n=215) involve sources in the range -19 ge M_B ge -26. The high redshift sources show E1 parameter values and domain occupation that are similar to our low redshift sample supporting earlier findings that E1 parameters are uncorrelated with source luminosity. Elementary accretion theory can account for a systematic increase of the minimum observed H-beta profile width with source luminosity. Narrow line Seyfert 1 sources with M_B = -28 show FWHM(H-beta) as much as 2000 km/s broader than those with M_B = -22. A possible change in the higher redshift/luminosity sources involves systematically weaker [OIII]4959,5007 narrow line emission.

We consider the properties of radio-loud (RL) AGN in the context of the Eigenvector 1 (E1) parameter space. RL sources show a restricted E1 parameter space occupation relative to the radio-quiet (RQ) majority. The Fanaroff-Riley II ``parent population'' of relatively un-boosted RL sources (median radio/optical flux ratio ~490) shows the most restricted occupation. RL sources have different broad line properties (and inferred black hole masses and Eddington ratios). FWHM H_beta for the broad line component in RL sources are at least twice as large as the RQ majority. The average broad FeII emission line strength is also about half that for RQ sources. Our sample suggests that the RL cutoff occurs near R_k=70 or logP(6cm)=32.0 ergs/s/Hz. Sources below this cutoff are RQ although we cannot rule out the existence of a distinct intermediate population. We show that the Doppler boosted core-dominated RL sources (median flux ratio ~1000) lie towards smaller FWHM(H_beta_bc) and stronger FeII in E1 as expected if the lines arise in an accretion disk. Our subsample of superluminal sources, with orientation inferred from the synchrotron self Compton model, reinforce this general E1 trend and allow us to estimate the role of source orientation in driving E1 domain occupation.

[Abridged] We compute the virial mass of the central black hole (M) and the luminosity-to-mass (L/M) ratio of ~ 300 low-z quasars and luminous Seyfert 1 nuclei. We analyze: (1) whether radio-quiet and radio-loud objects show systematic differences in terms of M and L/M; (2) the influence of M and L/M on the shape of the H-beta broad component line profile; (3) the significance of the so-called "blue outliers" i.e., sources showing a significant blueshift of the [OIII] 4959, 5007 lines with respect to the narrow component of H-beta which is used as an estimator of the quasar reference frame. We show that M and L/M distributions for RQ and RL sources are likely different for samples matched in luminosity and redshift. Line profile comparisons for median spectra computed over narrow ranges of M and L/M indicate that a Lorentz function provides a better fit for higher L/M sources and a double Gaussian for lower L/M values. A second (redshifted) Gaussian component at low L/M appears as a red asymmetry frequently observed in radio-loud and radio-quiet sources with broader (FWHM > 4000 km/s) H-beta broad component profiles. This component becomes stronger in larger mass and lower L/M sources. No specific influence of radio loudness on the H-beta broad component profile is detected, although equivalent widths of H-beta broad component and especially of [OIII] 4959,5007 are larger for radio-loud sources. We identify five more "blue outlier" sources. Since these sources are, on average, one magnitude brighter than other AGNs with similar mass, they are accreting at an Eddington ratio that is 2-3 times higher. We hint at evolutionary effects that explain some of these results and reinforce the "Eigenvector 1" correlations.

We investigated the optical FeII emission in a sample of about 215 low-redshift AGN (quasars and luminous Seyfert 1 galaxies). We find that a scaled and broadened FeII template based on the I Zw 1 spectrum can satisfactorily model the FeII emission in almost all sources in our sample. We confirm that FWHM H$\beta$ and FeII$\lambda$4570 are strongly correlated. The correlation is different for sources with FWHM(H$\beta$) greater than or less than $\sim$4000 km s$^{-1}$. Sources with FWHM H$\beta$$\leq$ 4000 km s$^{-1}$ (Population A) show no difference between FWHM H$\beta$ and FeII while sources with FWHM(H$\beta$) $\geq$ 4000 km s$^{-1}$ (Population B) show FWHM FeII that is systematically smaller than FWHM H$\beta$. This may be telling us that FeII emission in Pop. B sources comes from only the outermost part of the H$\beta$\ emitting region where the degree of ionization is lowest.

We investigate the radial velocity difference between the [OIII]5007,4959 and H-beta lines for a sample of 200 low redshift AGN. We identify seven objects showing an [OIII]5007 blueshift relative to H-beta with amplitude larger than 250 km/s (blue "outliers"). These line shifts are found in sources where the broad high ionization lines (e.g. CIV1549) also show a large systematic blueshift. Such blueshifts occur only in the population A region of the Eigenvector 1 parameter domain (that also contains NLSy1 sources). We suggest that [OIII] blueshifts are also associated with the high ionization outflow originating in these sources. This is a direct kinematic linkage between narrow and broad line emitting gas.

[LaTeX removed] Recent work has shown that it is possible to systematize quasar spectral diversity in a parameter space called ``Eigenvector 1'' (E1). We present median AGN spectra for fixed regions of the E1 (optical) parameter space (FWHM(H-beta) vs. equivalent width ratio RFE=W(FeII4570)/W(H-beta). Comparison of the median spectra for different regions show considerable differences. We suggest that an E1-driven approach to median/average spectra emphasizes significant differences between AGN, and offers more insights into AGN physics and dynamics than a single population median/average derived from a large and heterogeneous sample of sources. We find that the H-beta broad component line profile changes in shape along the E1 sequence both in average centroid shift and asymmetry. While objects with FWHM(H-beta)< 4000 km/s are well fitted by a Lorentz function, AGN with FWHM(H-beta)> 4000 km/s are well fitted if two broad line components are used: a broad (the "classical" broad line component) and a very broad/redshifted component.

[Latex markup removed] We recently discussed an Eigenvector 1 (E1) parameter space that provides optimal discrimination between the principal classes of broad-line Active Galactic Nuclei (AGN). In this paper we begin a search for the most important physical parameters that are likely to govern correlations and data point distribution in E1 space. We focus on the principal optical parameter plane involving the width of the H-beta broad component and the equivalent width ratio between the FeII blend at 4570 A and H-beta broad component. We show that the observed correlation for radio-quiet sources can be accounted for if it is primarily driven by the ratio of AGN luminosity to black hole mass (L/M proportional to Eddington ratio) convolved with source orientation. L/M apparently drives the radio-quiet correlation only for FWHM(H-beta) < 4000 km/s which includes Narrow Line Seyfert 1 (NLSy1) galaxies and can be said to define an AGN ``main sequence.'' Source orientation plays an increasingly important role as FWHM(H-beta) increases. We also argue that AGN lying outside the radio-quiet ``main sequence,'' and specifically those with optical FeII much stronger than expected for a given FWHM(H-beta), may all be Broad Absorption Line QSOs.

New observations of the broad-line quasar PG1416-129 reveal a large decline in its continuum luminosity over the past ten years. In response to the continuum change the ``classical'' broad component of Hbeta has almost completely disappeared (a x10 decrease in flux). In its place there remains a redshifted/redward asymmetric very broad emission line component. The significance of this change is multifold: (1) It confirms the existence of a distinct redshifted Very Broad Line Region (VBLR) component that persists after the demise of the broad component and that is frequently observed, along with the broad component, in radio-loud sources. (2) The smaller (x2) intensity change in the Hbeta very broad component supports the previously advanced idea that the VBLR is physically distinct and likely to arise in an optically thin region close to the central source. (3) The presence of a strong very broad component in the radio-quiet quasar PG1416-129 reinforces the notion that such ``population B'' quasars share similar spectroscopic (and hence geometrical and kinematical) properties to radio-loud sources. (4) AGN can show broad, very broad, or both line components simultaneously, making statistical comparisons of source profile widths difficult. (5) The interpretation, in reverberation studies, of the presence or lack of correlated response in broad line wings will be affected by this composite BLR/VBLR structure.

Recent observations have shown that the Fe Kalpha line profile of the Seyfert~1 galaxy MCG-06-30-15 is strongly variable. We attempt accretion disk model fits to the Fa Kalpha line profiles in a high, low and medium continuum luminosity phase of this source. During the monitoring by Iwasawa et al. (1996) a broad red-shifted component remained reasonably constant while a narrower component at $\approx$ 6.4 keV strongly responded to continuum changes. Physically consistent fits are possible if the index xi of the power-law emissivity changes from 0.7 (high phase) to 3.0 (low phase). The shape of the red-shifted component at low phase is crucial to the disk model interpretation. We suggest that the actual shape may be a broad redshifted Gaussian. Three lines of evidence support the interpretation of the Fe Kalpha line as multicomponent, beyond the lack of correlation in the response to continuum changes of the red and blue components in MCG -06-30-15. (1) We show that the strong concentration of narrow peak centroids at 6.4 keV is inconsistent with expectations of a random distribution of disk orientations. (2) The average Fe Kalpha profile for a sample of 16 mostly Seyfert~1's suggests a natural decomposition into two Gaussians one unshifted/narrow and the other redshifted/broad. (3) Evidence for emission in excess to the expectation of disk models on the high energy side of the Fe Kalpha profile is both a challenge for low inclination disk models and support for the two component decomposition.

We examine the properties of the Fe emission lines that arise near 6.4 keV in the ASCA spectra of AGN. Our emphasis is on the Seyfert 1 galaxies where broad and apparently complex Fe K alpha emission is observed. We consider various origins for the line but focus on the pros and cons for line emitting accretion disk models. We develop a simple model of an illuminated disk capable of producing both X-ray and optical lines from a disk. The model is able to reproduce the observed Fe K alpha FWHM ratio as well as the radii of maximum emissivity implied by the profile redshifts. The overall profile shapes however do not fit well the predictions of our disk illumination model nor do we derive always consistent disk inclinations for the two lines. We conclude that the evidence for and against an accretion disk origin for the Fe K alpha emission is equal at best. The bulk of the data requires a very disparate set of line fits which shed little light on a coherent physical model. We briefly consider alternatives to disk emission models and show that a simple bicone model can reproduce the FE line profiles equally well.

A systematic difference in alignment between the central black hole spin and the angular momentum of the accreting gas may help explain several differences found in the optical and UV HST spectra of radio loud and radio quiet AGN.

We obtained a new spectrum for the radio galaxy Pictor~A in December 1993. Comparison with previous data obtained in 1983 and 1987 shows that the broad component of the Balmer lines has increased considerably in strength and complexity. The Balmer lines now show a boxy, double peaked profile with FWZI$\sim$24000 \kms. Although the 1993 H\a\ profile is reminiscent of Arp~102B, accretion disk models face immediate difficulty because (1) the redshifted peak in the profile is considerably stronger than the blue one, and (2) the red and blue peaks varied at different epochs. The latter finding strongly favors a predominance of radial motion. We argue that Pictor~A, and the rare objects with similar extremely broad emission line profiles, are better understood if the orientation is such that the gas is predominantly moving along the line of sight (i.e.\ a geometrical extremum).

We present spectrophotometric observations of the Broad Line Radio Galaxy OQ 208 (Mrk 668, 1404+286) obtained between 1985 and 1991. We show that the Balmer line fluxes and profile shapes undergo remarkable changes. The ratio of intensities between the broad and narrow components of Hbeta increased monotonically from 15 in 1985 to 40 in 1991. The peak of the broad components of Hbeta and Halpha were known to be strongly displaced to the red. We have discovered a correlation between the amplitude of the broad peak displacement and the luminosity of Hbeta, in the sense that the displacement is larger when the line luminosity is higher. We suggest that the observations are not compatible with either a binary Broad Line Region model or one involving ballistic acceleration of the line emitting gas. Radiative acceleration of a system of outflowing clouds readily explains the correlation between line shift and luminosity as well as the peculiar line profiles. Furthermore, it seems that most or all of the Balmer emission originates from the inward face of the clouds. Theoretical line profiles suggest that the observed Hbeta profile is best fit assuming the contribution of an ensemble which might be spherical or confined in a thick disk in addition to a component emitted in a thin shell contained in a cone of half opening angle 12 degrees seen along its axis.