Negative superhumps are detected in the light curves of RW Tri collected in September 1994 (Smak 1995) and November/December 1957 (Walker 1963). New system parameters, obtained using $K_2$ and $V_{2,rot}\sin i$ (from Poole et al. 2003) and $q$ (estimated from $P_{nSH}$), are $M_1=0.60\pm 0.20M\odot$, $M_2=0.48\pm 0.15M\odot$, $A=1.13\pm 0.09\times 10^{11}$cm and $i=72.5\pm 2.5$.

The light curves of the permanent superhumper DW UMa are analyzed in order to determine the amplitudes of its superhumps, $A_{SH}$, and the amplitudes of the periodic light variations with the beat period -- the {\it irradiation amplitudes}, $A_{irr}$. The resulting values of $A_{SH}$ and $A_{irr}$, together with other values from the literature, turn out to be correlated thereby confirming the irradiation modulated mass transfer model for superhumps.

Light curves of AM CVn are analyzed by decomposing them into their Fourier components. The amplitudes of the fundamental mode and overtones of the three components: the superhumps, the negative superhumps and the orbital variations, are found to be variable. This implies that variations in the shape of the observed light curve of AM CVn are not only due to the interference between those components, but also due to the variability of their parameters.

Light curves of superhumps and their evolution during superoutbursts are analyzed by decomposing them into their Fourier components, including the fundamental mode and the first three overtones. The amplitudes of the fundamental mode are found to decrease significantly during superoutburst while those of the overtones remain practically constant. The phases of maxima of the fundamental mode {\it increase} systematically during superoutburst while those of the overtones -- systematically {\it decrease}. The combination of the two effects is responsible for the characteristic evolution of superhump light curves: the appearance and growth of the secondary humps and the spurious phase jumps in the (O-C) diagrams. Two intrepretations are possible. Either that instead of just one superhump period $P_{sh}$ there are four periods $P_k$ which resemble -- but are significantly different from -- the fundamental mode and the first three overtones of $P_{sh}$. Or -- more likely -- that those time-dependent phase shifts are genuine.

Quasi-periodic oscillation (QPO) of TT Ari are transient, short-living phenomena. They appear and disappear and their periods and amplitudes vary on a time scale as short as 1 hour. Consequently the periodograms covering longer intervals of time are generally meaningless.

Results of photometric observations of the permanent negative superhumper TT Ari in 1961/62 and 1966 are presented. Together with data from the literature they are used to discuss the negative superhump amplitudes $A_{nSH}$ and the amplitudes $A_{beat}$ of the modulation with the beat period $P_{beat}$. Both amplitudes are shown to vary considerably from one season to another. Three correlations are found: (1) between $A_{nSH}$ and $A_{beat}$, (2) between $A_{nSH}$ and $P_{nSH}$, and (3) between $A_{beat}$ and $P_{beat}$.

Further evidence is presented supporting the alternative interpretation of superhumps as being due to irradiation modulated periodically variable mass transfer rate. NZ Boo, HT Cas and PU UMa are added to the sample of high inclination dwarf novae showing -- during their superoutbursts -- modulation of the observed brightness of the disk with beat period. Simple model calculations confirm earlier hypothesis that this modulation is due to a non-axisymmetric structure of the outer parts of the disk, involving the azimuthal dependence of their geometrical thickness, rotating with the beat period. The modulation amplitude $A_{mod}$ is found to decrease during superoutbursts. In particular, it is found that during two superoutbursts of OY Car the rate of decline of the superhump amplitude $dA_{SH}/dt$ was correlated with the rate of decline $dA_{mod}/dt$. This leads to a simple explanation for the decreasing amplitudes of superhumps: it is a consequence of the decreasing modulation amplitude.

Osaki and Kato (2012) interpreted variations of the negative superhump periods, discovered by them in dwarf nova V1504 Cyg, as evidence in favor of the thermal-tidal instability model for superoutbursts. It is shown that their interpretation was incorrect. The observational evidence is recalled showing that superoutbursts are due to the enhanced mass transfer rate.

Eclipses of the hot spot in Z Cha and OY Car observed by many authors during their quiescence are re-analyzed. Distances of the spot from the center of the disk r_s are determined from phases of ingress and egress. In the case of several eclipses of Z Cha and nearly all eclipses of OY Car it is found that r_s(egress) < r_s(ingress). This implies that they are not representative for the radius of the disk r_d and is interpreted as being due to the stream overflow. The r_s(ingress) - dt relations (where dt is the time since last outburst) are improved when points affected by the stream overflow are omitted.

Contribution from the irradiated secondary component is detected in the light curves of five dwarf novae observed during superoutbursts. Their superhump light curves show that irradiation is modulated with the superhump phase. This strengthens the new interpretation of superhumps (Smak 2009) as being due to irradiation controlled mass transfer rate resulting in modulated dissipation of the kinetic energy of the stream.

Superhump amplitudes observed in dwarf novae during their superoutbursts depend on orbital inclination: the maximum amplitudes in systems with low inclinations are $A_\circ \approx 0.25$ mag., while at higher inclinations they increase from $A_\circ \sim 0.3$ to $A_\circ \sim 0.6$ mag. The mean maximum superhump amplitudes normalized to the average luminosity of the disk are: $<A_n>=0.34\pm 0.02$ in low inclination systems and only $<A_n>=0.17\pm 0.01$ in high inclination systems. This shows that at high inclinations the superhump lIght source is {\it partly} obscured by the disk edge and implies that it is located close to the disk surface but extends sufficiently high above that surface to avoid full obscuration. Superhump amplitudes in high inclination systems show modulation with beat phase ($\phi_b$), interpreted as being due to azimuth-dependent obscuration effects in a non-axisymmetric disk. In addition they show modulation with $2\phi_b$ which implies that the orientation of the superhump light source is correlated with the direction of the stream. The dependence of superhump amplitudes on orbital inclination and their modulation with beat phase eliminate the tidal-resonance model for superhumps. Instead they support an alternative interpretation of superhumps as being due to periodically modulated dissipation of the kinetic energy of the stream. Superhump amplitudes in permanent superhumpers are $<A>=0.12$, i.e. much smaller than the maximum amplitudes observed during superoutbursts.

The mass transfer rate in SS Cyg at quiescence, estimated from the observed luminosity of the hot spot, is log\dotM_tr=16.8+/-0.3. This is safely below the critical mass transfer rates of log\dotM_crit=18.1 (corresponding to logT^o_crit=3.88) or log\dotM_crit=17.2 (corresponding to the "revised value of logT^o_crit=3.65). The mass transfer rate during outbursts is strongly enhanced.

The origin of tilted disks in cataclysmic variables is explained in terms of a model involving the stream-disk interactions. Tilted, precessing disk causes periodically variable asymmetry in the irradiation of the two hemispheres of the secondary component, resulting in variable vertical component of the velocity of the stream. The following stream-disk interactions provide additional vertical acceleration to disk elements needed to produce and maintain disk tilt. Predictions based on this model compare favorably with observations.

Observational evidence is presented for periodically variable irradiation of secondary components. This results in strongly modulated mass outflow. Superhumps are then due to enhanced dissipation of the kinetic energy of the stream. Qualitative interpretation of superhump periods and their variations is also presented.

From the analysis of superhump eclipses of Z Cha it is found that the location of the superhump light source coincides with the overflowing parts of the stream. This implies that superhumps are due to modulated mass transfer rate resulting in periodically enhanced dissipation of the kinetic energy of the stream.

The evidence presented earlier by several authors for the substantial disk eccentricity in dwarf novae during their superoutbursts is shown to result either from errors or from arbitrary, incorrect assumptions. In particular, the evidence resulting from hot spot eclipses was based on the assumption that the spot distances are identical with disk radii which is not always correct.

Amplitudes of bolometric light curves produced by 2D and 3D simulations are used to determine the corresponding visual amplitudes. They turn out to be about 10 times lower than typical amplitudes of superhumps. This means a major failure of the tidal model of superhumps.