The analysis of 14 periodograms of EZ Lyn for the data spaced over 565 d in 2012--2014 (2-3.5 yr after 2010 outburst) yielded the existence of the stable signals around 100 c/d and three signals around 310 c/d, 338 c/d and 368 c/d (the corresponding
periods are 864 s, 279 s, 256 s and 235 s). We interpret them as independent non-radial pulsations of the white dwarf in EZ Lyn, but a possibility that a linear combination of frequency at 100 c/d and harmonic of orbital period could produce the frequency at 368 c/d also cannot be excluded. The signal at 100 c/d was detected during the first stay in the instability strip as a transient one. The period at 338 c/d, is a known non-radial pulsation EZ Lyn entered the instability strip after the 2010 outburst. We detected the signals around 310 c/d and 368 c/d for the first time. We applied the two-dimensional least absolute shrinkage and selection operator (Lasso) analysis for the first time to explore the behavior of these signals on the scale of hours for nightly runs of observations having duration of 6-12 hr. The Lasso analysis revealed the simultaneous existence of all three frequencies (310 c/d, 338 c/d and 368 c/d) for majority of nights of observations, but with variable amplitudes and variable drifts of frequencies by 2-6 percents on a time scale of ~5-7 hr. The largest drift we detected corresponded to 17.5 s in period in ~5 hours.
We carried out the photometric observations of the SU UMa-type dwarf nova ER UMa during 2011 and 2012, which showed the existence of persistent negative superhumps even during the superoutburst. We performed two-dimensional period analysis of its lig
ht curves by using a method called least absolute shrinkage and selection operator (Lasso) and phase dispersion minimization (PDM) analysis, and we found that the period of negative superhumps systematically changed between a superoutburst and the next superoutburst. The trend of the period change can beinterpreted as reflecting the change of the disk radius. This change of the disk radius is in good agreement with the predicted change of the disk radius by the thermal-tidal instability (TTI) model. The normal outbursts within a supercycle showed a general trend that the rising rate to maximum becomes slower as the next superoutburst approaches. The change can be interpreted as the consequence of the increased gas-stream flow onto the inner region of the disk as the result of the tilted disk. Some of the superoutbursts were found to be triggered by a precursor normal outburst when the positive superhumps appeared to develop. The positive and negative superhumps co-existed during the superoutburst. The positive superhumps were prominent only during four or five days after the supermaximum, while the signal of the negative superhumps became strong after the middle phase of the superoutburst plateau. A simple combination of the positive and negative superhumps was found to be insufficient in reproducing the complex profile variation. We were able to detect the developing phase of positive superhumps (stage A superhumps) for the first time in ER UMa-type dwarf novae. Using the period of stage A superhumps, we obtained a mass ratio of 0.100(15), which indicates that ER UMa is on the ordinary evolutional track of CVs.
The multi-site photometric observations of MN Dra were made over 77 nights in August-November, 2009. The total exposure was 433 hours. During this time the binary underwent two superoutbursts and five normal outbursts. During the course of first supe
routburst period of positive superhumps decreased with extremely large $dot P = -1.5 times 1.0^{-4}$ for SU UMa-like dwarf novae, confirming known behavior of MN Dra [1]. Between the superoutbursts MN Dra displayed negative superhumps. Their period changed cyclically around 0.096-day value.
