Survey of Period Variations of Superhumps in SU UMa-Type Dwarf Novae

We systematically surveyed period variations of superhumps in SU UMa-type dwarf novae based on newly obtained data and past publications. In many systems, the evolution of superhump period are found to be composed of three distinct stages: early evolutionary stage with a longer superhump period, middle stage with systematically varying periods, final stage with a shorter, stable superhump period. During the middle stage, many systems with superhump periods less than 0.08 d show positive period derivatives. Contrary to the earlier claim, we found no clear evidence for variation of period derivatives between superoutburst of the same object. We present an interpretation that the lengthening of the superhump period is a result of outward propagation of the eccentricity wave and is limited by the radius near the tidal truncation. We interpret that late stage superhumps are rejuvenized excitation of 3:1 resonance when the superhumps in the outer disk is effectively quenched. Many of WZ Sge-type dwarf novae showed long-enduring superhumps during the post-superoutburst stage having periods longer than those during the main superoutburst. The period derivatives in WZ Sge-type dwarf novae are found to be strongly correlated with the fractional superhump excess, or consequently, mass ratio. WZ Sge-type dwarf novae with a long-lasting rebrightening or with multiple rebrightenings tend to have smaller period derivatives and are excellent candidate for the systems around or after the period minimum of evolution of cataclysmic variables (abridged).


Introduction
This is a continuation of series of papers Kato et al. (2009), Kato et al. (2010), Kato et al. (2012a), Kato et al. (2013), Kato et al. (2014b), Kato et al. (2014a), Kato et al. (2015a) and Kato et al. (2016a) reporting new observations of superhumps in SU UMa-type dwarf novae. SU UMa-type dwarf novae are a class of cataclysmic variables (CVs) which are close binary systems transferring matter from a low-mass dwarf secondary to a white dwarf, forming an accretion disk [see e.g. Warner (1995) for CVs in general].
In SU UMa-type dwarf novae, there are two types of outbursts (normal outbursts and superoutbursts). Outbursts and superoutbursts in SU UMa-type dwarf novae are considered to be a result of the combination of thermal and tidal instabilities [thermal-tidal instability (TTI) model by Osaki (1989); Osaki (1996)].
During superoutbursts, semi-periodic variations called superhumps are observed whose period (superhump period, P SH ) is a few percent longer than the orbital period (P orb ). Superhumps are considered to originate from a precessing eccentric (or flexing) disk in the gravity field of the rotating binary, and the eccentricity in the disk is believed to be a consequence of the 3:1 resonance in the accretion disk [see e.g. Whitehurst (1988); Hirose and Osaki (1990); Lubow (1991); Wood et al. (2011)].
It has become evident since Kato et al. (2009) that the superhump periods systematically vary in a way common to many objects. Kato et al. (2009) introduced superhump stages (stages A, B and C): initial growing stage with a long period (stage A) and fully developed stage with a systematically varying period (stage B) and later stage C with a shorter, almost constant period (see figure 1).
It has recently been proposed by Osaki and Kato (2013b) that stage A superhumps reflect the dynamical precession rate at the 3:1 resonance radius and that the rapid decrease of the period (stage B) reflects the pressure effect which has an effect of retrograde precession (Lubow 1992;Hirose and Osaki 1993;Murray 1998;Montgomery 2001;Pearson 2006). As proposed by Kato and Osaki (2013) stage A superhumps can be then used to "dynamically" determine the mass ratio (q), which had been difficult to measure except for eclipsing systems and systems with bright secondaries to detect radial-velocity variations. It has been confirmed that this stage A method gives q values as precise as in eclipsing systems. There  Kato and Osaki 2013) have been more than 50 objects whose q values are determined by this method and it has been proven to be an especially valuable tool in depicting the terminal stage of CV evolution (cf. Kato et al. 2015a;Kato 2015).
In this paper, we present new observations of SU UMatype dwarf novae mainly obtained in 2016-2017. We present basic observational materials and discussions in relation to individual objects. Starting from Kato et al. (2014a), we have been intending these series of papers to be also a source of compiled information, including historical, of individual dwarf novae.
The material and methods of analysis are given in section 2, observations and analysis of individual objects are given in section 3, including discussions particular to the objects. General discussions are given in section 4 and the summary is given in section 5. Some tables and figures are available online only.

Data Source
The data were obtained under campaigns led by the VSNET Collaboration (Kato et al. 2004). We also used the public data from the AAVSO International Database 1 . Outburst detections of many new and known objects relied on the ASAS-SN CV patrol (Davis et al. 2015) 2 , the MASTER network , and Catalina Real-time Transient Survey (CRTS; Drake et al. 2009) 3 in addition to outburst detections reported to VSNET, AAVSO 4 , BAAVSS alert 5 and cvnet-outburst. 6 For objects detected in CRTS, we preferably used the names provided in Drake et al. (2014) and Coppejans et al. (2016). If these names are not yet available, we used the International Astronomical Union (IAU)-format names provided by the CRTS team in the public data release 7 Since Kato et al. (2009), we have used coordinatebased optical transient (OT) designations for some objects, such as apparent dwarf nova candidates reported in the Transient Objects Confirmation Page of the Central Bureau for Astronomical Telegrams 8 and CRTS objects without registered designations in Drake et al. (2014) or in the CRTS public data release and listed the original identifiers in table 1.
We provided coordinates from astrometric catalogs for ASAS-SN (Shappee et al. 2014) CVs and two objects without precise coordinate-based names other than listed in the General Catalog of Variable Stars (Kholopov et al. 1985) in table 2. We mainly used Gaia DR1 (Gaia Collaboration 2016), Sloan Digital Sky Survey (SDSS, Ahn et al. 2012), the Initial Gaia Source List (IGSL, Smart 2013) and Guide Star Catalog 2.3.2 (GSC 2.3.2, Lasker et al. 2007). Some objects were detected as transients by Gaia 9 and CRTS and we used their coordinates. The coordinates used in this paper are J2000.0. We also supplied SDSS g, Gaia G and GALEX NUV magnitudes when counterparts are present.

Observations and Basic Reduction
The majority of the data were acquired by time-resolved CCD photometry by using 20-60cm telescopes located world-wide. The list of outbursts and observers is summarized in table 1. The data analysis was performed in the same way described in Kato et al. (2009) and Kato et al. (2014a) and we mainly used R software 10 for data analysis.
In de-trending the data, we mainly used locallyweighted polynomial regression (LOWESS : Cleveland 1979) and sometimes lower (1-3rd order) polynomial fitting when the observation baseline was short. The times of superhumps maxima were determined by the template fitting method as described in Kato et al. (2009). The times of all observations are expressed in barycentric Julian days (BJD).
In figures, the points are accompanied by 1σ error bars whenever available, which are omitted when the error is smaller than the plot mark or the errors were not available (as in some raw light curves of superhumps).

Abbreviations and Terminology
The abbreviations used in this paper are the same as in Kato et al. (2014a): we used ǫ ≡ P SH /P orb − 1 for the fractional superhump excess. We have used since Osaki and Kato (2013a) the alternative fractional superhump excess in the frequency unit ǫ * ≡ 1 − P orb /P SH = ǫ/(1 + ǫ) because this fractional superhump excess is a direct measure of the precession rate. We therefore used ǫ * in discussing the precession rate.
The P SH , P dot and other parameters are listed in table 3 in same format as in Kato et al. (2009). The definitions of parameters P 1 , P 2 , E 1 , E 2 and P dot are the same as in Kato et al. (2009): P 1 and P 2 represent periods in stage B and C, respectively (P 1 is averaged during the entire course of the observed segment of stage B), and E 1 and E 2 represent intervals (in cycle numbers) to determine P 1 and P 2 , respectively. 11 Some superoutbursts are not listed in table 3 due to the lack of observations (e.g. single-night observations with less than two superhump maxima or poor observations for the object with already well measured P SH ).
We used the same terminology of superhumps summarized in Kato et al. (2012a). We especially call attention to the term "late superhumps". We only used the concept of "traditional" late superhumps when there is 10 The R Foundation for Statistical Computing: <http://cran.r-project.org/>. 11 The intervals (E 1 and E 2 ) for the stages B and C given in the table sometimes overlap because there is sometimes observational ambiguity (usually due to the lack of observations and errors in determining the times of maxima) in determining the stages. an ∼0.5 phase shift [Vogt (1983); see also table 1 in Kato et al. (2012a) for various types of superhumps], since we suspect that many of the past claims of detections of "late superhumps" were likely stage C superhumps before it became evident that there are complex structures in the O − C diagrams of superhumps (see discussion in Kato et al. 2009).

Period Analysis
We used phase dispersion minimization (PDM; Stellingwerf 1978) for period analysis and 1σ errors for the PDM analysis was estimated by the methods of Fernie (1989) and Kato et al. (2010). We have used a variety of bootstrapping in estimating the robustness of the result of the PDM analysis since Kato et al. (2012a). We analyzed 100 samples which randomly contain 50% of observations, and performed PDM analysis for these samples. The bootstrap result is shown as a form of 90% confidence intervals in the resultant PDM θ statistics. If this paper provides the first solid presentation of a new SU UMa-type classification, we provide the result of PDM period analysis and averaged superhump profile.

O − C Diagrams
Comparisons of O − C diagrams between different superoutbursts are also presented whenever available. This figure not only provides information about the difference of O − C diagrams between different superoutbursts but also helps identifying superhump stages especially when observations were insufficient or the start of the outburst was missed. In drawing combined O − C diagrams, we usually used E =0 for the start of the superoutburst, which usually refers to the first positive detection of the outburst. This epoch usually has an accuracy of ∼1 d for well-observed objects, and if the outburst was not sufficiently observed, we mentioned in the figure caption how to estimate E in such an outburst. In some cases, this E =0 is defined as the appearance of superhumps. This treatment is necessary since some objects have a long waiting time before appearance of superhumps. We also note that there is sometimes an ambiguity in selecting the true period among aliases. In some cases, this can be resolved by the help of the O − C analysis. The procedure and example are shown in subsection 2.2 in Kato et al. (2015a).

Individual Objects
3.1 V1047 Aquilae V1047 Aql was discovered as a dwarf nova (S 8191) by Hoffmeister (1964). Hoffmeister (1964) reported a blue color in contrast to the nearby stars. Mason and Howell (2003) obtained a spectrum typical for a quiescent dwarf nova. According to R. Stubbings, the observation by Greg Bolt during the 2005 August outburst detected superhumps, and the superhump period was about 0.074 d (see Kato et al. 2012b). The object shows rather frequent outbursts (approximately once in 50 d), and a number of outbursts have been detected mainly by R. Stubbings visually since 2004.
Although observations are not sufficient, visual observations by R. Stubbings suggest a supercycle of ∼90 d, which would make V1047 Aql one of ordinary SU UMatype dwarf novae with shortest supercycles.

BB Arietis
This object was discovered as a variable star (Ross 182,NSV 907) on a plate on 1926 November 26 (Ross 1927). The dwarf nova-type nature was suspected by the association with an ROSAT source (Kato,. The SU UMa-type nature was confirmed during the 2004 superoutburst. For more information, see Kato et al. (2014a).
The 2016 superoutburst was detected by P. Schmeer at a visual magnitude of 13.2 on October 30 (vsnet-alert 20273). Thanks to the early detection (this visual detection was 1 d earlier than the ASAS-SN detection), stage A growing superhumps were detected (vsnet-alert 20292). At the time of the initial observation, the object was fading from a precursor outburst. Further observations recorded development of superhumps clearly (vsnet-alert 20312, 20321). The times of superhump maxima are listed in table 4. There were clear stages A-C (figure 3). The 2013 superoutburst had a separate precursor outburst and a comparison of the O − C diagrams suggests a difference of 44 cycle count from that used in Kato et al. (2014a). The value suggests that superhumps during the 2013 superoutburst evolved 3 d after the precursor outburst.
The 2017 superoutburst was detected by P. Schmeer at a visual magnitude of 11.4 and also by the ASAS-SN team at V=11.82 on March 15. Single superhump was recorded at BJD 2457829.3171(2) (N=236). Although there were observations on three nights immediately after the superoutburst, we could neither detect superhump nor orbital periods.

OY Carinae
See Kato et al. (2015a) for the history of this well-known eclipsing SU UMa-type dwarf nova. The 2016 superoutburst was detected by R. Stubbings at a visual magnitude of 11.6 on April 2 (vsnet-alert 19676). Due to an accidental delay in the start of observations, the earliest time-resolved CCD observations were obtained on April 3 (vsnet-alert 19706). On that night, superhumps (likely in the growing phase) unfortunately overlapped with eclipses (figure 4, upper panel). Distinct superhumps were recorded on April 4 (vsnet-alert 19692; figure 4, middle panel). A further analysis suggested that stage A superhumps escaped detection before April 4 (due to the lack of observations and overlapping eclipses). At the time of April 4, the superhumps were already likely stage B (table 5, maxima outside eclipses). We could, however, confirmed a positive P dot for stage B superhumps (cf. figure 5), whose confirmation had been still awaited (cf. Kato et al. 2015a).
The combined data used in Kato et al. (2015a) and new observations, we have obtained the eclipse ephemeris for the use of defining the orbital phases in this paper using the MCMC analysis ): Min(BJD) = 2457120.49413(2) + 0.0631209131(5)E.
The epoch corresponds to the center of the entire observation. The mean period, however, did not show a secular decrease (e.g. Han et al. 2015;Kato et al. 2015a). It may be that period changes in this system are sporadic and do not reflect the secular CV evolution.

GS Ceti
This object (SDSS J005050.88+000912.6) was selected as a CV during the course of the SDSS (Szkody et al. 2005). The spectrum was that of a quiescent dwarf nova. Southworth et al. (2007) obtained 8 hr of photometry giving a suspected orbital period of ∼76 min.
Although there were no secure outburst record in the past, the object was detected in bright outburst on 2016 November 9 at V=13.0 by the ASAS-SN team (vsnet-alert 20328). Subsequent observations detected early superhumps (vsnet-alert 20334, 20342). Although the profile was not doubly peaked as in many WZ Sge-type dwarf novae (cf. Kato 2015), we consider the signal to be that of early superhumps since it was seen before the appearance of ordinary superhumps and the period was close to the suggested orbital period by quiescent photometry (figure 6). The object started to show ordinary superhumps on November 17 (vsnet-alert 20368, 20381, 20395, 20404; figure 7). The times of superhump maxima are listed in table 6. There were clear stages A and B. The best period of early superhumps by the PDM method was 0.05597(3) d. Combined with the period of stage A superhumps, the ǫ * of 0.0288(8) corresponds to q=0.078(2). Although the object is a WZ Sge-type dwarf nova, it is not a very extreme one as judged from the relatively large P dot of stage B superhumps and the lack of the feature of an underlying white dwarf in the optical spectra in quiescence (Szkody et al. 2005;Southworth et al. 2007). Although there were some post-superoutburst observations, the quality of the data was not sufficient to detect superhumps.

GZ Ceti
This object was originally selected as a CV (SDSS J013701.06−091234.9) during the course of the SDSS (Szkody et al. 2003). Szkody et al. (2003) obtained spectra showing broad absorption surrounding the emission lines of Hβ and higher members of the Balmer series. The object showed the TiO bandheads of an M dwarf secondary. A radial-velocity study by Szkody et al. (2003) suggested an orbital period of 80-86 min. There was a superoutburst in 2003 December and Pretorius et al. (2004) reported the orbital and superhump periods of 79.71(1) min and 81.702(7) min, respectively. Pretorius et al. (2004) reported the period variation of superhumps, which can be now interpreted as stages B and C. Pretorius et al. (2004) suggested that this object has a low mass-transfer rate. The same superoutburst was studied by Imada et al. (2006), who reported the superhump period of 0.056686(12) d. Imada et al. (2006) noticed the unusual presence of the TiO bands for this short-P SH object and discussed that the secondary should be luminous. Ishioka et al. (2007) obtained an infrared spectrum dominated by the secondary component. Ishioka et al. (2007) suggested that the evolutionary path of GZ Cet is different from that of ordinary CVs, and that it is a candidate of a member of EI Psc-like systems. EI Psc-like systems are CVs below the period minimum showing hydrogen (likely somewhat reduced in abundance) in their spectra (cf. Thorstensen et al. 2002;Uemura et al. 2002;Littlefield et al. 2013) and are consider to be evolving towards AM CVn-type objects. Superhump observations during the superoutbursts in 2009 and 2011 were also reported in Kato et al. (2009) and Kato et al. (2013), respectively.
The 2016 superoutburst was detected by R. Stubbings at a visual magnitude of 12.6 on December 18 (vsnetalert 20493). The ASAS-SN team also recorded the out-burst at V=12.66 on December 17. This superoutburst was observed in its relatively late phase to the postsuperoutburst phase (vsnet-alert 20594). There was also a post-superoutburst rebrightening on 2017 January 15 (vsnet-alert 20569). The times of superhump maxima are listed in table 7. The times after E=266 represent post-superoutburst superhumps. The maxima for E ≤54 were stage B superhumps and "textbook" stage C superhumps continued even during the post-superoutburst phase without a phase jump as in traditional late superhumps (figure 8).

AK Cancri
AK Cnc was discovered as a short-period variable star (AN 77.1933) with a photographic range of 14 to fainter  (Morgenroth 1933). Morgenroth (1933) detected two maxima on 48 plates between JD 2425323 and 2426763. Tsesevich (1967) classified this object to be a U Gem-type variable without a particular remark. Williams (1983) reported a G-type spectrum unlike for a CV. The identification was later found to be incorrect (Howell et al. 1990;Wenzel 1993b). The identification chart by Vogt and Bateson (1982) was correct. Amateur observers, particularly AAVSO and VSOLJ observers, made regular monitoring since 1986 and detected several outbursts. Time-resolved CCD observation by Howell et al. (1990) recorded a declining part of an outburst. Szkody and Howell (1992) obtained a spectrum in quiescence, which was characteristic to a dwarf nova. Wenzel (1993b) and Wenzel (1993a) reported observations using photographic archival materials and discussed outburst prop-erties. Wenzel (1993b) also gave a summary of confusing history of the identification of this object. Kato (1994) was the first to identify this object to be an SU UMa-type dwarf nova by observing the 1992 superoutburst. Mennickent et al. (1996) reported another superoutburst in 1995. The orbital period was spectroscopically measured to be 0.0651(2) d (Arenas and Mennickent 1998). Kato et al. (2009) provided analyses of the 1999 and 2003 superoutbursts. Kato et al. (2013) further reported observations of the 2012 superoutburst.
The 2016 superoutburst was detected at a visual magnitude of 13.5 by G. Poyner on April 5. The times of superhump maxima are listed in table 8. Due to the rather poor coverage, we could not determine P dot for stage B although the distinction between stages B and C was clear. Although positive P dot for stage B is expected for this P orb ,   Kato et al. (2014a). The 2013 superoutburst had a separate precursor outburst and the cycle count is different by 44 from that used in Kato et al. (2014a). The value suggests that superhumps during the 2013 superoutburst evolved 3 d after the precursor outburst. Since the start of the 2004 superoutburst was not well constrained, we shifted the O − C diagram to best fit the 2016 one.  it still awaits better observations (figure 9).

GZ Cancri
GZ Cnc was discovered by K. Takamizawa as a variable star (=TmzV34). The object was confirmed as a dwarf nova (Kato et al. 2001b;Kato et al. 2002a). Tappert and Bianchini (2003) obtained the orbital period of 0.08825(28) d by radial-velocity observations. The SU UMa-type nature was established during the 2010 (Kato et al. 2010). See Kato et al. (2014a) for more information.
The 2017 superoutburst was detected by R. Stubbings at a visual magnitude of 13.0 on February 2 and on the same night at 12.5 mag by T. Horie. Subsequent observations detected growing superhumps on February 3 and 4. Superhumps grew further on February 6 (vsnet-alert 20642). The times of superhump maxima are listed in table 9. Thanks to the early detection of the outburst, stage A superhumps were clearly detected (figure 10). The ǫ * for stage A superhumps [0.081(3)] corresponds to q=0.27(2).
The times of superhump maxima during the 2016 superoutburst are listed in table 10. Stage B with a positive P dot and a transition to stage C superhumps were recorded (see also figure 11).

V337 Cygni
V337 Cyg was discovered as a long-period variable (AN 101.1928). The dwarf nova-type nature was confirmed in 1996. The SU UMa-type nature was established during the 2006 superoutburst (cf. Boyd et al. 2007). See Kato  0 13  -------C  NSV 14681  The 2016 superoutburst was detected by M. Moriyama at an unfiltered CCD magnitude of 15.5 on November 17. Observations on a single night yielded three superhumps (table 11). The maximum E=2 suffered from large atmospheric extinction and the quality of this measurement was poor. The P SH is omitted from table 3 since there were observations with much more accurate values in the past.

V1113 Cygni
V1113 Cyg was discovered as a dwarf nova by Hoffmeister (1966). The SU UMa-type nature was identified by Kato et al. (1996b). See Kato et al. (2016a) for more history.
The 2016 superoutburst was detected by H. Maehara at a visual magnitude of 14.3 on July 27 (vsnet-alert 20003). A visual observation by P. Dubovsky on the same night and ASAS-SN detection on the next night indicated further brightening (vsnet-alert 20011, 20015). Thanks to the early detection and notification, growing superhumps were detected (vsnet-alert 20022). The times of superhump maxima are listed in table 12, which clearly indicate the presence of stage A superhumps (figure 12). It may be noteworthy that stage A lasted nearly 40 cycles (figure 12), which may be analogous to long-P orb SU UMa-type dwarf novae with slowly evolving superhumps (such as V1006 Cyg: Kato et al. 2016b;V452 Cas: Kato et al. 2016a). Since stage A superhumps were observed, a spectroscopic radial-velocity study is desired to determine q using the stage A superhump method.

IX Draconis
IX Dra is one of ER UMa-type dwarf novae (Ishioka et al. 2001). See Kato et al. (2014a) and Olech et al. (2004) for the

Object
Year   Otulakowska-Hypka et al. (2013) probably did not reflect the long-term trend well.

IR Geminorum
IR Gem was discovered as a U Gem-type variable star (AN S5423) by Popowa (1961). Although little was known other than outbursts with an interval of ∼75 d and amplitudes of ∼2.5 mag (Popova 1960;Meinunger 1976), 12 this object has been well monitored by AAVSO observers since its discovery. Several outbursts were already recorded in the 1960s (Mayall 1968). Bond (1978) obtained a spectrum typical for an outbursting dwarf nova. Burenkov and Voikhanskaia (1979) reported a dwarf nova-type spectrum in quiescence. Shafter et al. (1984) identified this object to be an SU UMa-type dwarf

Object
Year nova by detecting superhumps. Shafter et al. (1984) suggested a small mass ratio (either a massive white dwarf or an undermassive secondary) based on a radial-velocity study. Although Feinswog et al. (1988), Lázaro et al. (1990) and Lazaro et al. (1991) reported more detailed spectroscopic studies, the orbital period was not well measured. Observations of superhumps during the 1991 superoutburst were reported in Kato (2001). Kato et al. (2009) (vsnetalert 19645). The times of superhump maxima are listed in table 14. The observation started two days later than the announcement and stage A superhumps were not recorded.
The 2017 superoutburst was detected by K. Kasai on March 12 (vsnet-alert 20763) while observing KaiV36, an ellipsoidal variable star in the field of IR Gem. The outburst was detected early enough and stage A superhumps were observed (figure 14). The object was still in quiescence on March 10. The times of superhump maxima are listed in table 15. The observations were not long enough and P dot was not determined. The ǫ * for stage A superhumps is 0.068(11), whose errors mainly comes from the uncertainty in the orbital period [0.0684(6) d] (Feinswog et al. 1988). This ǫ * corresponds to q=0.22(4). Accurate determination of the orbital period is desired since the object is bright enough and its behavior during superoutbursts has been well documented.  -----C  MASTER J055845 2016 0.058070 0.000081 0 19  -------C2  MASTER J162323 2016 0.09013 0.00007 0  4  -------Ca  MASTER J165153 2017 0.071951 0.000079 0 31  -------C  3.14 NY Herculis NY Her was originally discovered by Hoffmeister (1949) as a Mira-type variable. Based on photographic observations by Pastukhova (1988) and the CRTS detection on 2011 June 10, the object was identified as an SU UMa-type dwarf nova with a short supercycle . For more history, see Kato et al. (2013).
The 2016 June superoutburst was detected by the ASAS-SN team at V=16.19 on June 28. Subsequent observations detected superhumps (vsnet-alert 19938, 19939, 19948). The times of superhump maxima are listed in table 16. There was a rather smooth transition from stage B to C. Since the 2016 observations was much better than the 2011 one, we provide an updated superhump profile in figure 15. It is noteworthy that the mean superhump amplitude (0.10 mag) is much smaller than most of SU UMa-type dwarf novae with similar P SH (or P orb ) (see figure 16). Such an unusual low superhump amplitude is commonly seen in period bouncers and it may be a signature that NY Her is in a different evolutionary location from the standard one with this P orb .
ASAS-SN light curve suggest that bright outbursts (likely superoutbursts) tend to occur in every ∼60-70 d (figure 17). We selected long outbursts (presumable superoutbursts) from the ASAS-SN and Poyner's observa-tions and listed in table 17. Note that we selected the brightest points of outbursts and they do not necessarily reflect the starts of the outbursts. These maxima can be well expressed by a period of 63.5(2) d with residuals less than 5 d. We consider that this period is the supercycle of this system. The entire durations of superoutbursts were less than 10 d, which are much shorter than those in ER UMa-type dwarf nova (cf. Kato and Kunjaya 1995;Robertson et al. 1995) but are similar to that of V503 Cyg with a supercycle of 89 d (Harvey et al. 1995). Although the supercycle is between ER UMa-type dwarf novae and ordinary SU UMa-type dwarf novae, it is not clear whether NY Her fills a gap between them since NY Her does not have intermediate properties between them. NY Her may be classified as an unique object with a short supercycle and a small superhump amplitude despite the relatively long P SH .

MN Lacertae
This object (=VV 381) was discovered by Miller (1971). Relatively frequent outbursts were recorded in Miller (1971) and the object was originally considered to be a Z Cam-type dwarf nova. T. Kato, however, noted a very faint quiescence during a systematic survey of I-band photometry of dwarf novae (1990, unpublished) and he   Approximate cycle counts (E) after the starts of outbursts were used. The 2015 superoutburst with a separate precursor outburst was shifted by 15 cycles to best match the others. Since the start of the 2016 superoutburst was not well constrained, the values were shifted by 45 cycles to best match the others. This shift suggests that the actual start of the 2016 superoutburst occurred 2 d before the initial detection. suggested that the outburst amplitude should be comparable to those of SU UMa-type dwarf novae.
Since this object was initially cataloged as a Z Camtype dwarf nova, Simonsen (2011) included it as a target for "Z CamPaign" project. As a result, the outburst behavior was relatively well recorded in the AAVSO database, particularly in 2010-2012. The possibility of an SU UMatype dwarf nova was particularly noted after a long outburst in 2011 June (vsnet-alert 13420, 13424). During this outburst, accurate astrometry was obtained confirming that the true quiescent magnitude is indeed faint (22nd mag or even fainter). There was another outburst in 2012 October, during which a call for observations of superhumps was issued (vsnet-alert 15063). Following this outburst, the object was withdrawn from the Z CamPaign project and it has not been observed as frequently as before.
The 2016 bright outburst was detected by the ASAS-SN team at V=15.32 on October 30. Single-night observations on October 31 indeed detected superhumps (vsnetalert 20283; figure 18). The times of superhump maxima were BJD 2457693.2873(15) (N=37) and 2457693.3684(8) (N=53). The best superhump period by the PDM method         Approximate cycle counts (E) after the peak of the superoutburst were used. Since the start of the 2016 superoutburst was very well defined, we used the peak of the superoutburst and redefined the cycle counts. The other outbursts were shifted to best match the 2016 one.         Miller (1971).

V699 Ophiuchi
This object was discovered as a dwarf nova (HV 10577) with a photographic range of 13.8 to fainter than 16.0 (Boyce 1942). Boyce (1942) recorded five outbursts between 1937 June 5 and 1940 July 5. The intervals of the first four outbursts were in the range of 320-390 d. Although Walker and Olmsted (1958) presented a finding chart, later spectroscopic studies have shown that the marked object is a normal star (Zwitter and Munari 1996; Kato et al. (2012a). We selected the range of −5 < E < 10 respect to the peak superhump amplitude to illustrate the maximum superhump amplitudes. The curve indicates a spline-smoothed interpolation of the sample in Kato et al. (2012a). The location of NY Her (reflecting the first night of observation; we consider that these observations were early enough to make a comparison in this figure) is shown by stars. The single point right to NY Her is a superhump of QY Per in 1999. The other superhumps of the same superoutburst had amplitudes larger than 0.2 mag and this measurement does not reflect the characteristic amplitude of superhumps in QY Per.    Liu et al. 1999). On 1999 April 16, A. Pearce detected an outburst (vsnet-alert 2877). Accurate astrometry and photometry of the outbursting object indicated that the true V699 Oph is an unresolved companion to a 16-th magnitude star (vsnet-alert 2878, vsnet-chat 1810, 1868). The first confirmed superoutburst was recorded in 2003. This outburst was preceded by a separate precursor and followed by a rebrightening, forming a "triple outburst". (Kato et al. 2009). The 2008 and 2010 superoutbursts were also reported in Kato et al. (2009) and Kato et al. (2010), respectively.
The 2016 superoutburst was detected by the ASAS-SN team at V=14.56 on May 15 and by R. Stubbings at a visual magnitude of 14.4 on the same night. Time-resolved photometric observations were obtained on two nights and

V344 Pavonis
This dwarf nova was discovered in outburst on 1990 July 21. The object was spectroscopically confirmed as a dwarf nova. There were two outbursts recorded in archival plates between 1979 May and 1984 September (Maza et al. 1990). Mason and Howell (2003) obtained a typical spectrum of a dwarf nova in quiescence. Uemura et al. (2004) studied the 2004 outburst and identified the SU UMa-type nature. The analysis was refined in Kato et al. (2009).  listed a global P dot in table 3. Observations in the earlier phase of the superoutburst are needed to characterize superhumps of this object better.

V893 Sco
V893 Sco was discovered as a variable star by Satyvoldiev (1972). The variable had been lost for a long time, and was rediscovered by K. Haseda (Kato et al. 1998). For more historical information, see Kato et al. (2014a). This object is an eclipsing SU UMa-type dwarf nova (cf. Bruch et al. 2000;Matsumoto et al. 2000. The 2016 superoutburst was detected by R. Stubbings at a visual magnitude of 12.8 on March 21. It once faded to V=13.64 on the same night and brightened to V=12.37 on March 25 (vsnet-alert 19652). The outburst on March 21 should have been a precursor. Our timeresolved photometry started on March 28 and detected superhumps (vsnet-alert 19661; figure 21). Since our observation started relatively late, we could record only the final part of the superoutburst. Later observations were dominated by the orbital humps and we could only extract a small number of superhump maxima outside the eclipses (table 21). We obtained the eclipse ephemeris for the use of defining the orbital phases in this paper using the MCMC modeling  using the data up to Kato et al. (2014a) and current set of observation.
The 2016 superoutburst was detected by T. Horie at a visual magnitude of 12.5 on June 5. It was pointed out by H. Maehara the outburst already started on June 1 (vsnetalert 19872). Time-resolved photometry was carried out on two nights, yielding superhump maxima in table 22. A comparison of O − C diagrams (figure 22) suggest that these observations recorded the early phase of stage C.

AW Sagittae
AW Sge was discovered as a dwarf nova by Wolf and Wolf (1906). The object was identified as an SU UMa-type dwarf nova during the 2000 outburst (Kato et al. 2009). See Kato et al. (2014a) for more history.  photometric observations were carried out on a single night and yielded the superhumps maxima: BJD 2457558.3859(5) (N=75) and 2457558.4606(9) (N=50).
The 2016 superoutburst was detected by the ASAS-SN Approximate cycle counts (E) after the start of the superoutburst were used. Since the start of the 2010 superoutburst was not known, we have shifted the O − C diagram to best fit the others.
team at V=13.52 on October 23. Subsequent observations detected superhumps (vsnet-alert 20267). The times of superhump maxima are listed in table 23. As in other typical long-P SH systems (cf. figure 4 in Kato et al. 2009), stage B was relatively short. A comparison of the O − C diagrams has confirmed that the superhumps recorded in 2008 were indeed stage C ones (figure 23). Although individual superhump maxima were not measured, a PDM analysis of the post-superoutburst data (4.5 d segment after BJD 2457697) detected a period of 0.08000(11) d. This value suggests that stage C superhump lasted even after the termination of the superoutburst.

SU Ursae Majoris
This object is the prototype of SU UMa-type dwarf novae. See Kato et al. (2015a) for the history. The 2017 superoutburst was detected by E. Muyllaert at a visual magnitude of 11.3 on February 23. Only single superhump at BJD 2457810.5647(3) (N=89) was observed.

HV Virginis
HV Vir was originally discovered by Schneller (1931) in outburst on 1929 February 11. The object was also given a designation of NSV 6201 as a suspected variable. Duerbeck (1984) and Duerbeck (1987) classified it as a classical nova and provided a light curve of the 1929 outburst based on his examination of archival plates. Amateur observers, particularly by the Variable Star Observers' League in Japan (VSOLJ), suspected it to be a dwarf nova and started monitoring since 1987 [i.e. following the publication of Duerbeck (1987)]. The object was caught in outburst by P. Schmeer on 1992 April 20 at a visual magnitude of 12.0 ). The 1992 outburst was extensively studied (Barwig et al. 1992;Leibowitz et al. 1994;Kato et al. 2001). It might be worth noting that Barwig et al. (1992) recorded low-amplitude variations with a period corresponding to the orbital period, their interpretation (originating from the hot spot as in quiescence) was strongly affected by Patterson et al. (1981). Although  reported the detection of superhumps, the detailed result has not been published. Leibowitz et al. (1994) reported the detection of historical outbursts in 1939, 1970 and 1981 in archival plates. Although Leibowitz et al. (1994) noted chaotic "early superhump variability", its period was not precisely determined. Leibowitz et al. (1994) recorded superhumps and reported a negative P dot , which was incorrect due to an error in cycle counts probably misguided by the received wisdom at that time that SU UMa-type dwarf novae universally show negative P dot (cf. Warner 1985;Patterson et al. 1993). Using additional observations and all available data, Kato et al. (2001) clarified that this object showed two types of superhumps (early superhumps and ordinary superhumps) and the P dot for ordinary superhumps was positive. Kato et al. (2001) proposed the close similarity to AL Com (cf. Kato et al. 1996a) and WZ Sge, giving a basis of the modern concept of WZ Sge-type dwarf novae (Kato 2015).
The object underwent another superoutburst in 2002. This outburst was also extensively studied by Ishioka et al. (2003), who established the positive P dot using a much more complete set of observations than in 1992. Patterson et al. (2003) also reported the superhump period of the same outburst and the orbital period of 0.057069(6) d from quiescent photometry. There was another superoutburst in 2008, which was reported in Kato et al. (2009).
The 2016 superoutburst was detected by the ASAS-SN team at V=12.0 on March 10 (cf. vsnet-alert 19571). Initial observations already detected early superhumps (vsnetalert 19573, 19576, 19589; figure 25). The object then developed ordinary superhumps (vsnet-alert 19581, 19599, 19633). The times of superhump maxima are listed in table 24. The data very clearly demonstrate the presence of stages A and B, although there was an observational gap in the middle of stage B. The superhump period of stage A was very ideally determined to be 0.05907(6) d (cf. figure  24). This period gives the fractional superhump excess of ǫ * =0.034(1), which corresponds to q=0.093(3). This value supersedes the earlier determination by the same method to be q=0.072(1) using the less extensive 2002 data. The period was determined for the 2002 data from single-night observations assuming that stage A continued up to the first observation of stage B while the present observation obtained an almost complete coverage of stage A (see figure 24). It was likely that the error was underestimated in the 2002 superoutburst. The outburst started rapid fading on March 29-30 and the entire duration of the superoutburst was at least 20 d. Despite dense observations, no post-outburst rebrightening was recorded.
A PDM analysis of the post-superoutburst observations yielded a period of 0.05799(2) d ( figure 26). This period corresponds to a disk radius of 0.33a assuming that the precession rate is not affected by the pressure effect. The value is in the range of 0.30-0.38a determined for well-observed WZ Sge-type dwarf novae (Kato and Osaki 2013).
The period of early superhumps [0.057000 (8)  observation (from the observations reported in Kato et al. 2009). The quality of past observations were lower: 0.057085 d (without error estimate) for the 1992 outburst , which was based only on published times of maxima, and 0.0569(1) d for the 2002 outburst (Ishioka et al. 2003). The current observations, combined with the 2008 data, established the period of early superhumps of this object to a precision directly comparable to the orbital period for the first time. The 2016 and 2008 periods were 0.13(2)% and 0.13(3)% shorter than the orbital period, respectively.

NSV 2026
This object was discovered as a variable star (=HV 6907) by Hoffleit (1935). The SU UMa-type nature was confirmed during the 2015 superoutburst. For more history, see Kato et al. (2016a).
There was a superoutburst in 2016 February (Kato et al. 2016a). Another superoutburst occurred in 2016 November, which was detected by J. Shears at an unfiltered CCD magnitude of 14.19 and by E. Muyllaert at a visual magnitude of 14.0 on November 25. The object was further observed to brighten to a visual magnitude of 13.2 on November 26. The times of superhump maxima are listed in table 25. These superhumps were likely stage B ones (figure 27). As judged from the interval of two superoutbursts in 2016 and the supercycle of ∼95 d (Kato et al.    2016a), two superoutbursts were likely missed between the two superoutbursts in 2016.

NSV 14681
NSV 14681 was discovered as a variable star (SVS 749) of unknown type with a photographic range of 14 to fainter than 14.5 (Belyavskii 1936). The CRTS team detected an outburst at an unfiltered CCD magnitude of 15.6 on 2007 June 13 and it was readily identified with NSV 14681 (Drake et al. 2014). The CV is a fainter component of a close pair (Kato et al. 2012b

1RXS J161659.5+620014
This object (hereafter 1RXS J161659) was initially identified as an X-ray selected variable (also given a name as MASTER OT J161700.81+620024.9), which was first de-tected in bright state on 2012 September 11 at an unfiltered CCD magnitude of 14.4 (Balanutsa et al. 2013). The dwarf nova-type variability was confirmed by analysis of the CRTS data (Balanutsa et al. 2013; see also vsnet-alert 16079, 16720).
The 2016 April outburst was detected by the ASAS-SN team at V=14.74 on April 22. Subsequent observations detected superhumps (vsnet-alert 19763, 19765, 19772; figure 29). The times of superhump maxima are listed in table 27. The nature of the humps for E ≥155 (postsuperoutburst) is unclear due to the gap in the observation. These humps may be either traditional late superhumps or the extension of stage C superhumps (if it is the case, the cycle count should be increased by one). We consider the latter possibility less likely, since this interpretation requires the period of stage C superhumps to be 0.07065(2) d, which appears to be too short (by ∼1%) shorter than that of stage B superhumps. We do not use these maxima in obtaining the periods in table 3.
The 2016 July outburst was detected by the CRTS team at an unfiltered CCD magnitude of 14.63 on July 10 (cf. vsnet-alert 19970). Although it was considered to be too early for a next superoutburst, subsequent observations detected superhumps (vsnet-alert 19996). The times of superhump maxima are listed in table 28. As in the superoutburst in 2016 April, the nature of maxima for E ≥112 (post-superoutburst) was unclear. A comparison of O − C diagrams between two superoutbursts is given in figure  30.
These observations indicate that the supercycle is only ∼80 d. We studied past ASAS-SN observations and detected outbursts (table 29). The outburst pattern became more regular since the 2015 July (it may have been due to the change in the variability in this system or the improvement of observations in ASAS-SN) and we obtained a mean supercycle of 89(1) d from five most recent superoutbursts (with |O − C| values less than 8 d). Despite the shortness of the supercycle, normal outbursts are not as frequent as in ER UMa-type dwarf novae (Kato and Kunjaya 1995;Robertson et al. 1995) or active SU UMatype dwarf novae, such as SS UMi Olech et al. 2006) and BF Ara (Kato et al. 2001a). The object resembles V503 Cyg with a supercycle of 89 d with a few normal outbursts between superoutbursts (Harvey et al. 1995). V503 Cyg is known to show different states (Kato et al. 2002b), which is now considered to be a result of the disk tilt suppressing normal outbursts (Ohshima et al. 2012;Osaki and Kato 2013a;Osaki and Kato 2013b). A search for negative superhumps in 1RXS J161659 would be fruitful.

ASASSN-13ak
This object was detected as a transient at V=15.4 on 2013 May 23 by the ASAS-SN team (Stanek et al. 2013). There was an independent detection by the MASTER network (Shurpakov et al. 2013b). The SU UMa-type nature was identified during the 2015 superoutburst (Kato et al. 2016a).
The 2016 superoutburst was detected by the ASAS-SN team at V=14.43 on August 2. We obtained time-resolved observations on two nights, yielding superhump maxima in table 30. The resultant period is longer than that in 2015 (Kato et al. 2016a) and these superhumps may have been stage A ones, despite that the amplitudes were already large since the 2016 observations were obtained in the earlier phase than in 2015.

ASASSN-13al
This object was detected as a transient at V=15.2 on 2013 June 1 by the ASAS-SN team (Prieto et al. 2013) (The ASAS-SN Transients page gave a magnitude of 16.03 with a "BADCAL" flag). The CV-type nature was confirmed by spectroscopy. 13 The 2016 outburst was detected by the ASAS-SN team at V=14.77 on October 9. There was a previous detection in the ASAS-SN data at V=14.75 on 2012 June 7. Subsequent observations detected superhumps (vsnetalert 20220; figure 31). The time of superhump maxima are listed in table 31. The period was not very well determined since the observations were undertaken only on a single night. The best superhump period by the PDM method is 0.0783(2) d.

ASASSN-13bc
This object was detected as a transient at V=16.9 on 2013 July 4 by the ASAS-SN team. A number of past outbursts were recorded in the CRTS data.
The 2015 outburst was detected by the ASAS-SN team at V=14.83 on July 30. Subsequent observations detected superhumps (vsnet-alert 18921, 18930). The times of superhump maxima are listed in table 32.
The 2016 outburst was detected by the ASAS-SN team at V=15.19 on May 24. Superhumps were also observed (vsnet-alert 19843, 19867). The object underwent a postsuperoutburst rebrightening at V=16.13 on June 9 (cf. vsnet-alert 19883). The times of superhump maxima are listed in table 33. The superhump profile is given for the better observed 2016 one (figure 32). A combined O − C diagram (figure 33) suggests that the 2015 observations covered the early phase of stage B and the 2016 ones recorded both stages B and C, although the later part of stage B was not well recorded due to the lack of observations.

ASASSN-13bj
This object was detected as a transient at V=16.2 on 2013 July 10 by the ASAS-SN team. Two superhump maxima were obtained during the 2013 superoutburst (Kato et al. 2014b).
The 2016 superoutburst was detected by the ASAS-SN team at V=14.98 on July 3. Subsequent observations detected superhumps (vsnet-alert 19957, 19965, 19975; figure 34). The times of superhump maxima are listed in table 34. There was a marked decrease in the superhump  period and we tentatively identified a stage B-C transition around E =22. The accuracy of the resultant periods was not sufficiently high since they were determined by only short baselines.

ASASSN-13bo
This object was detected as a transient at V=15.96 on 2013 July 13 by the ASAS-SN team. The 2016 outburst was detected by the ASAS-SN team at V=15.19 on August 1. The 2016 outburst was the brightest recorded one and a superoutburst was suspected. Subsequent observations detected superhumps (vsnet-alert 20053, 20068). There was a 3-d gap in the observations and alias periods are possible (figure 35). The period in table 3 refers to the one giving smallest residuals and it was determined by the PDM method. The object faded to ∼20 mag on August 13.

ASASSN-13cs
This object was detected as a transient at V=14.9 on 2013 September 2 by the ASAS-SN team. The object was spectroscopically identified as a dwarf nova in outburst. 14 The 2016 outburst was detected by the ASAS-SN team

ASASSN-13cz
This object was detected as a transient at V=14.9 on 2013 September 14 by the ASAS-SN team. The 2013 outburst turned out to be a superoutburst by the detection of superhumps (Kato et al. 2014a).
The 2016 superoutburst was detected by the ASAS-SN team at V=14.47 on July 27. Subsequent observations detected superhumps (vsnet-alert 20023, 20042). The times of superhump maxima are listed in table 37. We suspect that stages B and C were partially observed (cf. figure 37). We provide a superhump profile (figure 38), which was determined much better than in 2013.   figure 39). The times of superhump maxima are listed in table 38. The positive P dot for stage B superhumps is a common feature in many short-P SH systems.

ASASSN-15cr
This object was detected as a transient at V=14.9 on 2015 February 7 by the ASAS-SN team. The 2017 outburst was detected by the ASAS-SN team at V=14.73 on January 9. Subsequent observations detected superhumps (vsnet-

ASASSN-16da
This object was detected as a transient at V=16.1 on 2016 March 8 by the ASAS-SN team. The outburst was confirmed and announced on March 12, when the object was at V=15.5. The brightness peak was on March 10 at V=15.1. The object was identified with an g=21.5 mag SDSS object. The large outburst amplitude received attention.
The object showed double-wave early superhumps on March 13 and 14 (vsnet-alert 19579, 19592; figure 41). On March 15 (5 d after the brightness peak), the object started to show ordinary superhumps (vsnet-alert 19598, 19617, 19653; figure 42). The times of superhump maxima are listed in table 40. The epochs for E ≤2 and E ≥203 were apparently those of stage A and C superhumps, respectively. If we consider that stage A just ended at E=10 (which may not be a bad assumption as compared with O − C diagrams of well-observed objects), the period of stage A superhumps was 0.05858(10) d. The resultant ǫ * of 0.042(2) corresponds to q=0.12(1). This relatively large q for a WZ Sge-type dwarf nova is consistent with the appearance of stage C superhumps, short duration of stage A, relatively large P dot in stage B [+7.5(0.9) × 10 −5 ], and relatively early appearance of ordinary superhumps. This object is probably close to the borderline of WZ Sge-type dwarf novae and ordinary SU UMa-type dwarf novae.

ASASSN-16dk
This object was detected as a transient at V=16.4 on 2016 March 21 by the ASAS-SN team. The outburst was confirmed and announced on March 24, when the object was at V=15.1. Subsequent observations detected superhumps. The times of superhump maxima are listed in table 41. Since the observations were apparently obtained during the late phase of the superoutburst, the superhump stage was probably C. The lack of period variation is consistent with this interpretation.

ASASSN-16dz
This object was detected as a transient at V=15.0 on 2016 April 2 by the ASAS-SN team. The outburst was announced after the observation on April 3 at V=14.2. The object had been listed as an emission-line object IPHAS2 J064225.58+082546.7 (Witham et al. 2008). Although superhumps were detected, the period was not well determined due to short runs and the limited coverage only on two nights (figure 45). The times of superhump maxima

ASASSN-16ez
This object was detected as a transient at V=14.3 on 2016 May 6 by the ASAS-SN team. The large outburst amplitude suggested an SU UMa-type superoutburst (cf. vsnetalert 19804).
On May 12, this object started to show large-amplitude superhumps (vsnet-alert 19823, 19829). The times of superhump maxima are listed in table 44. These superhumps were stage B ones and there was little period varia- tion. The object was still in outburst on May 23 (17 d after the outburst detection and 11 d after the first detection of superhumps). Although there were some variations before May 12, we could not confidently detect stage A superhumps (probably due to a 1.5-d gap in the observation). Although the small P dot might suggest a small q (cf. Kato 2015), the relatively large amplitude of superhumps (figure 46) and the lack of long duration of stage A do not support this interpretation. The seemingly small P dot may be a result of a relatively short observational coverage of 4 d.

ASASSN-16fr
This object was detected as a transient at V=16

ASASSN-16fu
This object was detected as a transient at V=13.9 on 2016 June 5 by the ASAS-SN team. The large outburst amplitude received attention (cf. vsnet-alert 19864). On June 14 (9 d after the outburst detection), this object showed fully developed superhumps (vsnet-alert 19899, 19901; figure  48). The times of superhump maxima are listed in table 46. Although the maxima for E ≤1 were stage A superhumps, the period of stage A superhumps was not determined due to the unfortunate 2 d gap before the full growth of the superhumps. The relatively small P dot in stage B [+4.6(0.6) × 10 −5 ] suggests a relatively small q [q ∼0.08(1) according to equation (6) in Kato (2015)]. This expectation is consistent with the small amplitude of the superhumps (figure 48). A retrospective analysis of the data before ordinary superhumps appeared (BJD before 2457553) detected possible early superhumps with a period of 0.05623(3) d (figure 49). We consider that this object is a WZ Sge-type dwarf nova, which appears consistent with the small q as inferred from the small P dot .

ASASSN-16gh
This object was detected as a transient at V=15.5 on 2016 June 16 by the ASAS-SN team. The outburst was announced on June 18, when the object further brightened to V=14.3 No strong superhumps were detected up to observations on June 24. Growing superhumps were recorded on June 28 (12 d after the outburst detection; vsnet-alert 19935). Further development of superhumps were observed (vsnet-alert 19943, 19952; figure 50). Although we could not detect early superhumps, the object may be a WZ Sge-type dwarf nova since the waiting time (12 d) of the growth of the superhumps was long. If this identification is true, the object appears to be a candidate for a period bouncer. The relatively small amplitude of the superhumps (figure 50) and the large outburst amplitude (there was no known quiescent counterpart) would favor this possibility.

ASASSN-16gj
This object was detected as a transient at V=13.3 on 2016 June 18 by the ASAS-SN team (cf. vsnet-alert 19905). The  object was also detected on June 17 by the MASTER network (independent detection, Balanutsa et al. 2016a). The object was undetected on June 12 (Balanutsa et al. 2016a) and June 11 (ASAS-SN data). Observations on June 21 already recorded superhumps (vsnet-alert 19927). The superhumps grew further (vsnet-alert 19934, 19936, 19953; figure 51, figure 52). It took, however, some time to establish the superhump period since nightly observations were short. The times of superhump maxima are listed in table 48. The maxima for E ≤23 were most likely stage A superhumps as judged from the growing amplitudes and O − C values (figure 52). The period of stage A superhumps, however, was not convincingly determined due to the shortness of nightly observations.
The object showed a likely plateau-type rebrightening (figure 52; the plateau-type rebrightening was favored by the lack of rising/fading trends in the nightly light curves in the rebreightening phase). Although we could not observe early superhumps, we consider that this object is a WZ Sge-type dwarf nova as judged from the long rebrightening (cf. Kato 2015). The duration of the phase of early superhumps, if it was present, must have been shorter than 9 d. This shortness suggests that this object may not be an extreme WZ Sge-type dwarf nova. There was, however, the case of the 2015 superoutburst of AL Com (Kimura et al. 2016), in which no early superhumps were observed despite that the same object showed long phases of early superhumps in previous superoutbursts. It may be premature to draw any conclusion about the evolutionary state of ASASSN-16gj only from the present observation.

ASASSN-16gl
This object was detected as a transient at V=14.8 on 2016 June 19 by the ASAS-SN team. The outburst was announced on June 22, when the object faded to V=15.4. Subsequent observations detected superhumps (vsnetalert 19940, 19942, 19949; figure 53). The times of superhump maxima are listed in table 49. The object was still in outburst on July 11, and the entire duration of the superoutburst exceeded 22 d. Although our initial observation was already 8 d after the outburst detection, the possibility that we observed stage C superhumps may be rather small considering the long duration of the superoutburst (stage B-C transitions usually occur very late in such systems, e.g. SW UMa in figure 1).

ASASSN-16hi
This object was detected as a transient at V=15.5 on 2016 July 15 by the ASAS-SN team. Subsequent observations detected superhumps (vsnet-alert 20002, 20018; figure 54). The times of superhump maxima are listed in table 50. The observed maxima well illustrate typical stages B and C. Although the outburst was rather well recorded, the faintness (around 16 mag) made the quality of the averaged superhump profile rather poor.

ASASSN-16hj
This object was detected as a transient at V=14. We give the possible signal of early superhumps in figure 58. Although the signal was close to the detection limit, the period appears to be consistent with the superhump period and the profile is also consistent with that of early superhumps. The period with the PDM method was 0.05499(6) d. The ǫ * for stage A superhumps was 0.034 (7), which corresponds to q=0.09(2).

ASASSN-16ia
This object was detected as a transient at V=14.6 on 2016 August 1 by the ASAS-SN team. The object was also de- tected by Gaia (Gaia16azd) at a magnitude of 16.71 on August 7. 15 The coordinates of the object were taken from this Gaia detection. The object once faded to V=17.1 on August 5. It was observed bright (16.0 mag) again on August 7 and showed strong early superhumps (vsnetalert 20055, 20069, 20076). The object was followed until August 15, when early superhumps were still present. A transition to ordinary superhumps was not observed since the object became too faint. gree of the contribution of the eclipsing component. Since the object is expected to have a very high inclination, detailed observations in quiescence are desired to determine the system parameters.
It is noteworthy that a precursor outburst was apparently present before the phase of the early superhumps. This is probably the first case in WZ Sge-type dwarf novae and the reason why the cooling wave started during the initial peak needs to be clarified.

ASASSN-16ib
This object was detected as a transient at V=14.2 on 2016 August 5 by the ASAS-SN team. Subsequent observations detected growing superhumps (vsnet-alert 20066, 20083). The times of superhump maxima are listed in table 52. During the epochs for E ≤14, the amplitudes of superhumps grew, and these superhumps can be safely identified as stage A superhumps. The distinction of stages B and C was unclear. We listed a value for 47 ≤ E ≤133 as stage B in table 3. The mean profile of the superhumps is shown in figure 61.

ASASSN-16ik
This object was detected as a transient at V=15.26 on 2016 August 6 by the ASAS-SN team. The object further brightened to V=13.9 on August 8. The object started to show superhumps on August 11-12 (vsnet-alert 20082; figure  62). The times of superhump maxima are listed in table 53. The data very clearly show stages A (growing superhumps) and B. The object showed a rebrightening on August 25 (vsnet-alert 20109), which faded rapidly. During this rebrightening phase, a weak superhump signal was detected with a period of 0.0649(3) d.

ASASSN-16is
This object was detected as a transient at V=14.9 on 2016 August 9 by the ASAS-SN team. Initial observations detected double-wave modulations attributable to early superhumps (vsnet-alert 20078, 20084; figure 63). The period of early superhumps was 0.05762(2) d. The object started to show ordinary superhumps at least on August  figure 64). The times of superhump maxima are listed in table 54. The superoutburst plateau was terminated by rapid fading on August 28. The object is confirmed to be a WZ Sge-type dwarf nova.

ASASSN-16iu
This object was detected as a transient at V=15.3 on 2016 August 4 by the ASAS-SN team. The object once faded to fainter than V=17.6 on August 6 and brightened again to V=15.2 on August 9. The detection of the outburst was announced after this brightening. Superhumps were soon detected on August 11 (vsnet-alert 20075). The amplitudes of superhumps decreased and they became less prominent on subsequent nights. They became detectable again on August 15 (figure 65). The times of superhump maxima are listed in table 55. Due to the long period of undetectable superhumps, the P dot for stage B superhumps is very uncertain. The period for stage C given in table 3 is very approximate due to the short baseline.

ASASSN-16iw
This object was detected as a transient at V=13.9 on 2016 August 10 by the ASAS-SN team. There was a faint (g=21.9) SDSS counterpart (there were 17 measurements in the SDSS data with a range of 21.8-22.2 in g) and the large outburst amplitude suggested a WZ Sge-type dwarf nova.
The object started to show superhumps on August 17 (vsnet-alert 20086, 20091, 20100; figure 66). The times of superhump maxima are listed in table 56. The superhumps grew slowly and it took at least 47 cycles to reach the full superhump amplitude. Based on O − C variations, we have identified E ≤32 to be stage A superhumps ( figure 67).
An analysis of the early part of the light curve detected a possible signal of early superhumps (figure 69). Although the signal was weak (the amplitude was smaller than 0.01 mag) and the profile was not doubly humped as expected for early superhumps, we suspect that this is a candidate period of early superhumps since the period excess was close to what is expected for a WZ Sgetype dwarf nova. The period was 0.06495(5) d. The ǫ * of 0.029(1) for stage A superhumps corresponds to q=0.079(2). This q value is not as small as expected for a period bouncer at this orbital period. The relatively large P dot for stage B superhump may also be suggestive for a relatively large q. The object may be similar to WZ Sgetype dwarf novae with multiple rebrightenings with relatively large q, such as MASTER OT J211258.65+242145.4 and MASTER OT J203749.39+552210.3 (Nakata et al. 2013).

ASASSN-16jb
This object was detected as a transient at V=13.3 on 2016 August 18 by the ASAS-SN team. The object was caught on the rise to the maximum. The object was initially suspected to be a Galactic nova (cf. vsnet-alert 20092). The  object was confirmed to be blue, confirming the dwarf nova-type nature (vsnet-alert 20094). Subsequent observations detected early superhumps (vsnet-alert 20098, 20095; figure 70). The object started to show ordinary superhumps (figure 71) on August 25 and showed behavior similar to a short-period SU UMa-type dwarf nova rather than an extreme WZ Sge-type dwarf nova (vsnetalert 20112, 20125, 20154). The times of superhump maxima are listed in table 57. All stages (A-C) are clearly seen. The period of stage A superhumps listed in table 3 was determined by the PDM method.
The period of early superhumps was 0.06305(2) d ( figure 70). The fractional superhump excess of stage A superhumps ǫ * was 0.0321(5), which corresponds to q=0.088(1). This value is larger than what is expected for a period bouncer having this orbital period. The O − C behavior (positive P dot for stage B and the appearance of stage C) is also consistent with an object having an intermediately low q.
The object was also detected in outburst by ASAS of the observing season). The presence of earlier outbursts also seems to exclude the possibility of a period bouncer.
The identification in AAVSO VSX with UGPS J175044.95−255837.2 appears to be doubtful considering its red color (J − K=2.7). This supposed identification likely came from the initial proposed classification as a classical nova. We adopted coordinates by the ASAS-SN team. Although the maxima for E <16 are clearly stage A superhumps, the period was not determined due to the shortness of the segment. Although there was apparent stage B-C transition around E=146, the period of stage C superhumps was not well determined. The fractional superhump excess ǫ * for stage A superhumps was 0.0213(16), which corresponds to q=0.056(5). The small q and an orbital period significantly longer than the period minimum suggest that this object is a period bouncer. Although the mean superhump amplitude is larger than those of period bouncer candidates, this may have been due to the high orbital inclination as suggested by the strong early superhumps.

ASASSN-16kg
This object was detected as a transient at V=16.1 on 2016 September 7 by the ASAS-SN team. The object brightened to V=15.2 on September 8 and the outburst was announced. There was no quiescent counterpart recorded in previous plates, and the large outburst amplitude received attention. Subsequent observations detected superhumps (vsnet-alert 20182; figure 77). The superhump period was around 0.10 d, which was not expected from the large outburst amplitude. Since there was a 3-d gap in the observation and individual runs were compara- ble to one superhump cycle, it was impossible to select the alias uniquely. The candidate periods by the PDM methods (figure 78) are 0.09676(4) d, 0.10013(4) d, 0.10373(5) d, 0.107610(5) d and 0.11178(5) d. Other aliases can be likely rejected because they give large O − C scatters. Among them, we have selected 0.10013(4) d which gives the smallest θ in the PDM analysis to make cycle counts in table 62. One should note that there remains cycle count ambiguities due to the ambiguity in the alias selection. The object, however, is certainly located in the period gap. It might be worth noting that such a largeamplitude dwarf nova exists in the period gap.
The object was detected in outburst at an unfiltered CCD magnitude of 17.27 on September 26 by the CRTS team (=CSS160926:213630−251348). Since the object had already faded to 18.0 mag on September 20, this CRTS observation appears to have detected a post-superoutburst rebrightening.     (N=147). The best period determined by the PDM method was 0.0808(13) d ( figure 80). The object faded to fainter than 17.5 mag on October 14.

ASASSN-16lo
This object was detected as a transient at V=14.   period uniquely due to the faintness. We selected the most likely alias in calculating the O − C values in table 67. The superhump amplitudes were growing on the first two nights, and the period reported here may refer to that of stage A superhumps. The small amplitudes (figure 86) may also support this stage identification. The quiescent counterpart was originally proposed to be a very faint (g=22.9) object SDSS J050500.40+605455.3. This object, however, has a red color (u − g=+2.6) and it is unlikely a CV. The true quiescent counterpart should be fainter than g=23. We adopted the coordinates by the ASAS-SN team. rebrightening on December 23 (figure 88), which is relatively rare for such a long P SH system.

ASASSN-16nr
This object was detected as a transient at V=15.1 on 2016 November 26 by the ASAS-SN team. The outburst was announced after the observation on November 27 at V=15.1. The object showed superhumps (vsnet-alert 20420, 20432; figure 89). The times of superhump maxima are listed in table 69. Although a large outburst amplitude (∼7 mag) was suggested, the object was not a WZ Sge-type dwarf nova since it showed well-developed ordinary superhumps immediately after the outburst detection.

ASASSN-16nw
This object was detected as a transient at V=15.6 on 2016 November 23 by the ASAS-SN team. The outburst was announced after the observation on November 27 at V=16.1 and November 29 at V=16.3. Superhumps were

ASASSN-16ob
This object was detected as a transient at V=14.3 on 2016 November 28 by the ASAS-SN team. The outburst was announced after its further brightening to V=13.8 on November 30. Although the object was initially identified with a B=18.4 mag star in USNO catalog, B. Monard obtained outburst astrometry which indicated that the true counterpart is much fainter (vsnet-alert 20438). The object was also detected by Gaia (Gaia16bzl) 16 at a magnitude of 13.84 on November 30 and was announced (with an identification with ASASSN-16ob) on December 7. The large outburst amplitude suggested a WZ Sge-type dwarf nova. On December 11, low-amplitude superhumps were detected (vsnet-alert 20464, 20465, 20481), which grew to full superhumps (vsnet-alert 20484, 20498  didate periods (figure 91), we consider them false signals due to low signal-to-noise ratio since an analysis restricted to better observation quality gave a single period (figure 92; one-day aliases can be safely ruled out by O − C analysis). By using the data before BJD 2457734, we could not detect early superhumps. The upper limit of the amplitude of early superhumps was 0.01 mag. Although we could not detect early superhumps, we consider that this object belongs to WZ Sge-type dwarf novae based on its long waiting time (13 d) for ordinary superhumps to appear and the large outburst amplitude. Using the empirical relation between q and P dot for stage B superhumps [equation (6) in Kato 2015], the expected q is 0.069(3) (the error reflects the error in P dot ).

ASASSN-16oi
This object was detected as a transient at V=13.4 on 2016 December 3 by the ASAS-SN team (vsnet-alert 20443). Low-amplitude early superhumps were detected (vsnetalert 20466; figure 93). The object subsequently showed  ordinary superhumps (vsnet-alert 20466, 20485; figure  94). The object is confirmed to be a WZ Sge-type dwarf nova. The times of superhump maxima are listed in table 72. The best period of early superhumps with the PDM method is 0.05548(7) d. The fractional superhump excess ǫ * for stage A superhumps is 0.033(2), which gives q=0.091 (7). The relatively large q is consistent with a relatively large P dot for stage B superhumps and the relatively large amplitude of superhumps.

ASASSN-16os
This object was detected as a transient at V=13 A PDM analysis of the early part of the data yielded a signal which may be early superhumps (figure 97). This period was close to that of ordinary superhumps and we checked a possible contamination of ordinary superhumps by testing different segments. Although the test suggested that the period was not from a contamination of ordinary superhumps, we were not very confident about the reality of the signal since the amplitude was small and mostly only low time-resolution observations were obtained. If the detected period, 0.05494(6) d, is that of early superhumps, the fractional superhump excess for stage A superhumps is ǫ * =0.018(1). Although this value corresponds to q=0.047 (3), it needs to be treated with caution due to the limitation of the quality of observations. The overall behavior, however, suggests that this object is a rather extreme WZ Sge-type dwarf nova.

ASASSN-16ow
This object was detected as a transient at V=13.9 on 2016 December 13 by the ASAS-SN team. Since the object was near the Galactic plane, it was also suspected to be a nova. In contrast to many long-P SH SU UMa-type dwarf novae, this object showed a post-superoutburst rebrightening on Decmber 29-31 (vsnet-alert 20571). Although modulations were detected during this rebrightening, we could not detect a secure signal of superhumps.

ASASSN-17aa
This object was detected as a transient at V=13.9 on 2017 January 2 by the ASAS-SN team (vsnet-alert 20527). On January 11, superhumps were finally observed (vsnetalert 20562; figure 99). Although these superhumps were originally suspected to be stage A ones (vsnet-alert 20572, 20573), the large superhump amplitudes suggest that they were already stage B ones. The cycle numbers in table 75 follows this interpretation. There were possibly lowamplitude early superhumps (figure 100) with a period of 0.05393(3) d. These properties suggest the WZ Sge-type classification.

ASASSN-17ab
This object was detected as a transient at V=13.4 on 2017 January 2 by the ASAS-SN team (vsnet-alert 20527). The object was already in outburst at V=13.1 on January 1. Subsequent observations detected superhumps (vsnetalert 20531; figure 101). The times of superhump maxima are listed in table 76. Although the maxima for E ≤2 were stage A superhumps, we could not determine the period. The object was also detected by Gaia (Gaia17aep) 17 at a magnitude of 17.33 on January 18.

ASASSN-17az
This object was detected as a transient at V=14.4 on 2017 January 19 by the ASAS-SN team. Subsequent observations detected superhumps (vsnet-alert 20626; figure 102).  perhumps was recorded (vsnet-alert 20646; figure 103). The times of superhump maxima are listed in table 78. Although individual superhumps were not very well covered (most of them had only 10 observations or even less; the maxima could be reasonably determined since the object was bright), the overall O − C diagram indicates the clear presence of stage A and stage B with a positive P dot . The stage A-B transition was rather uncertain due to the lack of observations in the initial part.
An analysis of the early part of the superoutburst yielded a possible signal of early superhumps ( figure  104). Although there was a stronger signal around 0.0563 d, we consider it a false alias since it does not match the period of ordinary superhumps (the period of early superhumps should be shorter than that of ordinary superhumps). The suggested period by the PDM method was 0.05467(5) d. The ǫ * of stage A superhumps determined using this period was 0.0235(9), which corresponds to q=0.062(3). This value, however, could have a larger uncertainty since both the orbital period and the period of stage A superhumps were determined from insufficient   observations. The small q value, however, appears to be consistent with the WZ Sge-type behavior, the small amplitude of ordinary superhumps and the small P dot for stage B superhumps.

ASASSN-17bm
This object was detected as a transient at V=15.9 on 2017 January 25 by the ASAS-SN team. The outburst was announced after confirmation on January 27. Subsequent observations detected superhumps (vsnet-alert 20627; figure 105). The times of superhump maxima are listed in table 79. The period in table 3 was determined by the PDM method since individual maxima were not very well determined.

ASASSN-17bv
This object was detected as a transient at V=15.0 on 2017 January 31 by the ASAS-SN team. The outburst was announced after confirmation at V=14.9 on February 1. Subsequent observations starting on February 3 detected superhumps (vsnet-alert 20634; figure 106). The times of

ASASSN-17ce
This object was detected as a transient at V=14.6 on 2017 February 13 by the ASAS-SN team. Subsequent observations detected superhumps (vsnet-alert 20668, 20676; figure 107). The times of superhump maxima are listed in table 81. The maxima for E ≤6 likely correspond to a short stage B usually seen in long-P orb systems (Kato et al. 2009).
The object faded close to 18 mag on February 26.

ASASSN-17ck
This object was detected as a transient at V=16.6 on 2017 February 15 by the ASAS-SN team. The object was already in outburst at V=16.5 on February 13. Singlenight observations on February 17 detected superhumps (vsnet-alert 20680, figure 108). The times of maxima were BJD 2457801.5269(7) (N=25) and 2457801.6099(17) (N=21). The superhump period by the PDM analysis was 0.083(1) d.                   the outburst detection) and the large outburst amplitude ( > ∼ 9 mag) suggest that the object is a rather extreme WZ Sge-type dwarf nova.

ASASSN-17cx
This object was detected as a transient at V=16.4 on 2017 February 21 by the ASAS-SN team. The object was already in outburst at V=16.6 on February 18 and the outburst was announced after an observation at V=16.7 on February 23. Single-night observations on February 24 detected superhumps. The maxima were BJD 2457809.0621(9) (N=52), 2457809.1406(6) (N=48) and 2457809.2151(14) (N=50). The superhump period by the PDM method was 0.0761(7) d.

ASASSN-17dg
This object was detected as a transient at V=13.8 on 2017 March 7 by the ASAS-SN team. The object was already in outburst at V=13.8 on March 6. The last negative observation was on February 23. There is an ROSAT X-ray counterpart 1RXS J160232.8−603240. There was also an outburst with a maximum of V=13.03 on 2002 September 25, which lasted at least for 6 d in the ASAS-3 data. The actual maximum may have been even brighter since ASAS-3 did not observe this field for 6 d before this detection. The object has a bright (J=13.68) and blue (J − K=+0.10) 2MASS counterpart, indicating that the object was in outburst during 2MASS scans.
Observations started on March 9 and superhumps were immediately detected (vsnet-alert 20760; figure 112). The object started fading rapidly already on March 11. It was most likely the true maximum was missed by ASAS-SN observations for more than ∼5 d. The times of superhump maxima are listed in table 83. These superhumps were most likely stage C ones since observations were performed in the final phase of the superoutburst. The low amplitudes of superhumps (vsnet-alert 20760) suggested that superhumps were already decaying.
There was one post-superoutburst rebrightening on March 19 (V=14.3, vsnet-alert 20809). A PDM analysis of the post-superoutburst data (BJD 2457825.7-2457842.9) yielded a period of 0.06655(5) d, which is likely a contin- uation of stage C superhumps.

ASASSN-17dq
This    detected superhumps (vsnet-alert 20118, 20128, 20153; figure 115). The times of superhump maxima are listed in table 86. The maxima for E ≥137 were post-superoutburst ones. There was most likely a phase jump between E=80 and E=137 and the humps for E ≥137 were likely traditional late superhumps. The transition between stages B and C was rather smooth as in other relatively long P SH systems. probably refers to stage C one. We listed well-defined superoutbursts in the ASAS-SN data since 2014 in table 88. These superoutbursts can be well expressed by a supercycle of 108(1) d with maximum |O − C| values of 10 d. There have been typically two normal outbursts between superoutbursts. These features closely resemble those of V503 Cyg (Harvey et al. 1995) (vsnet-alert 20386). V503 Cyg, however, sometimes showed frequent normal outbursts (e.g. Kato et al. 2002b) and these alternations between phases of different number of normal outbursts have been considered to be a result of a disk tilt, which is considered to suppress normal outbursts (see the subsection of 1RXS J161659, subsection 3.27). Detection of negative superhumps is expected in CRTS J033349.

CRTS J044636.9+083033
This object (=CSS130201:044637+083033, hereafter CRTS J044637) was detected by the CRTS team at an unfiltered CCD magnitude of 17.70 on 2013 February 1. There was a bright outburst at V=15.12 on 2017 January 12 de- tected by the ASAS-SN team. Subsequent observations detected two superhumps on a single night (vsnet-alert 20580). The maxima were BJD 2457768.9773(11) (N=97) and 2457769.0716(14) (N=98). A PDM analysis yielded a period of 0.093(1) d. Although there were single-night observations 4 d later, the observing condition was not sufficient to detect superhumps.
The past data suggested that there was a bright (I=14) outburst in the past (vsnet-alert 10009). There was a bright outburst at unfiltered CCD magnitudes of 14.14-14.73 on 2008 November 20, detected by the CRTS team (cf .vsnetalert 10717). The outburst was suspected to be a superoutburst. Subsequent observations detected a superhump with a period of ∼0.08 d (vsnet-alert 10723; figure 119). There have been eight outbursts (up to 2016) in the CRTS database. The SDSS colors in quiescence suggested an orbital period of 0.08-0.12 d (Kato et al. 2012b).

CRTS J085603.8+322109
This object (=CSS100508:085604+322109, hereafter CRTS J085603) was detected by the CRTS team at an unfiltered CCD magnitude of 16.20 on 2010 May 8. There is a g=19.6mag SDSS counterpart and its colors yielded an expected orbital period of 0.067(1) d (Kato et al. 2012b). The 2016 outburst was detected by the ASAS-SN team at V=16.62 on November 26. Subsequent observations detected superhumps (vsnet-alert 20428; figure 121). The times of superhump maxima are listed in table 89.
The ASAS-SN data indicate that past outbursts occurred rather regularly. We listed outburst maxima (they are likely superoutburst as judged from the brightness) in table 90. These maxima can be expressed by a supercycle of 232(10) d, with the maximum |O − C| of 33 d. It was likely that the peak of the 2016 superoutburst was not covered by ASAS-SN observations.

CRTS J164950.4+035835
This object (=CSS100707:164950+035835, hereafter CRTS J164950) was detected by the CRTS team at an unfiltered  According to the ASAS-SN data, this object showed relatively regular superoutbursts (table 94). These superoutburst can be expressed by a supercycle of 128(2) d with maximum |O − C| of 20 d. The interval between the 2016 September and 2017 March superoutbursts was 166 d, which was rather unusually long for this object.

DDE 48
DDE 48 is a dwarf nova discovered by D. Denisenko (vsnet-alert 20146) in the vicinity of the dwarf nova MASTER OT J204627.96+242218.0 (Shumkov et al. 2016a). N. Mishevskiy monitored this object in 2016 and detected a bright outburst at V=15.5 on November 1 (vsnet-alert 20290). Subsequent observations detected a superhump (vsnet-alert 20291). Although this superhump was recorded only on a single night, the profile suggests a genuine superhump (figure 126). The superhump maximum was at BJD 2457694.2662(6) (N=43). The superhump became undetectable on two nights during the same superoutburst. This object shows frequent outbursts (cf. vsnet-alert 20291; figure 127). The shortest interval of outbursts was 3 d. The initial long outburst in figure 127 was also likely a superoutburst recorded in its the final phase. If it is indeed the case, the supercycle is around 62 d. The object may belong to ER UMa-type dwarf novae (Kato and  Kunjaya 1995; Robertson et al. 1995). (see also vsnet-alert 20291). Future continuous observations to determine the outburst characteristics, duty cycle and superhump period are desired.
3.99 MASTER OT J021315.37+533822.7 This object (hereafter MASTER J021315) was discovered as a transient at an unfiltered CCD magnitude of 16.8 mag on 2013 November 1 by the MASTER network (Yecheistov et al. 2013). The 2016 outburst was detected by the ASAS-SN team at V=16.39 on October 2. The ASAS-SN also detected the 2013 outburst and its duration was long (at least 8 d). During the 2016 outburst, long-period superhumps were detected (vsnet-alert 20218; figure 128). The period indicates that the object is in the period gap. The times of superhump maxima are listed in table 96. The period markedly decreased with a global P dot of −205(35) × 10 −5 . As recently recognized in many long-P orb objects, such a large period decrease is most likely a result of stage A-B transition (cf. V1006 Cyg and MN Dra: Kato et al. 2016b; CRTS J214738.4+244554 and OT  Kato et al. 2016a). The case is also likely since the initial observation of MASTER J021315 started only 1 d after the outburst detection. We gave values in table 3 following this interpretation. The ASAS-SN data suggest that outbursts in this system were relatively rare (only two were known with a separation of ∼3 yr). The object should have a low mass-transfer rate.

MASTER OT J030205.67+254834.3
This object (hereafter MASTER J030205) was discovered as a transient at an unfiltered CCD magnitude of 13.7 mag on 2016 December 4 by the MASTER network (Balanutsa et al. 2016b). Although initial observations suggested the presence of early superhumps of the WZ Sge-type dwarf nova (vsnet-alert 20447), they were later identified as developing superhumps (stage A) with double maxima (vsnet-alert 20449, 20451). Further development of superhumps were reported (vsnet-alert 20456, 20471; figure 129). The times of superhump maxima are listed in table 97. Stages A and B can be recognized and the P dot of stage B superhumps is positive, which is expected for  this P SH . The period of stage A superhump in table 3 was determined by the PDM method for the data before BJD 2457728.7. Short-term periodic oscillations were reported (vsnetalert 20452, 20457). An analysis of the entire data confirmed the presence of a coherent signal with a period 0.0035420(2) d [306.03(2) s] as originally reported (vsnetalert 20457) (figure 130). Given the sharpness (high coherence) of the signal, it may be an intermediate-polar (IP) signal rather than quasi-periodic oscillations (vsnet-alert 20458). Since IPs are relatively rare in SU UMa-type dwarf novae [see table 1 in Hameury and Lasota (2017); the only well-established SU UMa-type dwarf nova (not including WZ Sge-type one) is CC Scl (Kato et al. 2015b)], further confirmation of the signal in this system is desired.

MASTER OT J042609.34+354144.8
This object (hereafter MASTER J042609) was discovered as a transient at an unfiltered CCD magnitude of 12.9 on 2012 September 30 by the MASTER network ). The SU UMa-type nature was confirmed during this superoutburst (Kato et al. 2014b). This object is 3.102 MASTER OT J043220.15+784913.8 This object (hereafter MASTER J043220) was discovered as a transient at an unfiltered CCD magnitude of 16.8 on 2013 December 12 by the MASTER network (Shurpakov et al. 2013a). The 2017 outburst was detected by the ASAS-SN team at V=16.29 on January 25. The object was already seen in outburst at V=16.52 in the ASAS-SN data. Subsequent observations detected superhumps (vsnetalert 20614; figure 133). The times of superhump maxima were BJD 2457780.4941(5) (N=64) and 2457780.5584(5) (N=64). A PDM analysis yielded a period of 0.0640(6) d.

MASTER OT J043915.60+424232.3
This object (hereafter MASTER J043915) was discovered as a transient at an unfiltered CCD magnitude of 15.7 on 2014 January 21 by the MASTER network (Balanutsa et al. 2014). The SU UMa-type nature was confirmed during this superoutburst (Kato et al. 2015a). Fore more history, see Kato et al. (2015a).
The 2016 superoutburst was detected by the ASAS-SN 3.104 MASTER OT J054746.81+762018.9 This object (hereafter MASTER J054746) was discovered as a transient at an unfiltered CCD magnitude of 16.7 mag on 2016 October 12 by the MASTER network (Shumkov et al. 2016b). Subsequent observations detected superhumps (vsnet-alert 20227; figure 135). The times of superhump maxima are listed in table 100. The best superhump period with the PDM method is listed in table 3.
3.105 MASTER OT J055348.98+482209.0 This object (hereafter MASTER J055348) was discovered as a transient at an unfiltered CCD magnitude of 16.5 mag on 2014 March 13 by the MASTER network (Vladimirov et al. 2014). The 2017 outburst was detected by the ASAS-  (vsnet-alert 20681). Although observations on two nights were reported, neither data were of sufficient quality to determine the superhump period (the object already faded below 17 mag). The period used to calculated epochs in table 101 was one of the possibilities giving smallest O − C residuals. Other candidate aliases were 0.0784(1) d and 0.0720(1) d ( figure 136). The period of 0.0666 d reported in vsnet-alert 20681 could not express the second-night observation.                                                (Yecheistov et al. 2014). During the 2014 superoutburst, single-night observations detected superhumps (likely stage C ones) with a period of 0.0563(4) d (Kato et al. 2015a).
The 2016 outburst was detected by the ASAS-SN team at V=15.24 on September 14. The object was on the rise at V=16.51 on September 7. Rather queerly, the object was also detected at V=14.16 on August 30. There were no observations between August 30 and September 7. Threenight observations starting on September 16 detected superhumps (vsnet-alert 20186). During these observations, the object brightened from 15.4 mag (September 16) to 15.2 mag (September 18). Superhumps were recorded on the first two nights (table 102) This object (hereafter MASTER J064725) was discovered as a transient at an unfiltered CCD magnitude of 13.2 mag on 2013 March 7 by the MASTER network (Tiurina et al. 2013). Subsequent observations detected superhumps (Kato et al. 2014b).
The 2016        This object (hereafter MASTER J150518) was discovered as a transient at an unfiltered CCD magnitude of 15.5 mag on 2017 February 8 by the MASTER network (Gress et al. 2017). This transient was also detected by the ASAS-SN team (ASASSN-17cb) at V=15.4 on the same night, but the announcement was made after confirmation at V=15.1 on February 11. Although only the late course of the outburst was observed, superhumps were recorded (vsnetalert 20660). Since observations only recorded one superhump maximum on each night, the one-day alias could not be resolved. We selected one of them to minimize the θ of the PDM analysis to make cycle counts in table 105. Although the large negative global P dot may have reflected stage B-C transition, the quality of the data were insufficient to confirm it. This object (hereafter MASTER J151126) was discovered as a transient at an unfiltered CCD magnitude of 14.0 mag on 2016 March 18 by the MASTER network (Popova et al. 2016). Subsequent observations detected superhumps (vsnet-alert 19614, 19630; figure 142). The times of superhump maxima are listed in table 106. We interpreted that most of our observations recorded stage B as judged from a positive P dot expected for this P SH . The outburst faded on April 3. The duration of the outburst was at least 16 d.

MASTER OT J162323.48+782603.3
This object (hereafter MASTER J162323) was detected as a transient at an unfiltered CCD magnitude of 13.2 mag on 2013 December 9 by the MASTER network (Denisenko et al. 2013a). The 2013 superoutburst was well observed (Kato et al. 2014a).
The 2015 superoutburst was detected at V=13.49 on August 10 by the ASAS-SN team. Double-wave modulations were recorded on August 22 (vsnet-alert 19004).    These variations disappeared 6 d later (see figure 143). Since these observations covered only the last (and likely post-superoutburst) phase of the superoutburst and the nature of humps is unclear, we did not use these data for comparison with other superoutbursts.
There was an outburst at V=13.46 on 2016 April 22 (ASAS-SN detection). Subsequent observations did not detect superhumps (observers: Shugarov team and Akazawa). There was another outburst at V=13.34 on 2016 September 26 (ASAS-SN detection). Three superhumps were recorded during this superoutburst (table  107). These superhumps were likely obtained around transition from stage A to B (figure 144) and the period [0.09013(7) d, PDM method] is not listed in table 3.

MASTER OT J165153.86+702525.7
This object (hereafter MASTER J165153) was detected as a transient at an unfiltered CCD magnitude of 15.9 on 2013 May 23 by the MASTER network (Shurpakov et al. 2013b).  This object (hereafter MASTER J174816) was discovered as a transient at an unfiltered CCD magnitude of 15.6 mag on 2013 June 28 by the MASTER network (Denisenko et al. 2013b). The object has a blue SDSS counterpart (g=17.59). At least nine outbursts were recorded in the CRTS data. The object has a bright (J=15.55) 2MASS counterpart, suggesting that the object was in outburst during 2MASS  was initially considered to be an SS Cyg-type object from the low outburst amplitude (cf. vsnet-alert 20189), timeresolved photometry detected superhumps and eclipses (vsnet-alert 20190, 20196, 20206; figure 148, figure 149). We obtained the eclipse ephemeris using the MCMC analysis : This ephemeris is not intended for long-term prediction of eclipses. The epoch refers to the center of the observation. The times of superhump maxima outside the eclipses are listed in table 110. Although the O − C diagram suggests stage B-C transition, the periods and P dot may have not been well determined since the actual start of the outburst was much earlier than the detection announcement and determination of superhump maxima should have been affected by overlapping eclipses and orbital humps in the late epochs. A large positive P dot , however, is usual for such a short-P orb SU UMa-type dwarf nova. The small outburst amplitude (∼4.5 mag) was probably a result of the high orbital inclination. Since the object has deep eclipses and apparently shows a large positive P dot , it surely deserves further detailed observations to clarify the origin of increasing P SH during stage B. Observations of the early phase of a superoutburst are also desired to determine q by the stage A superhump method.

SBS 1108+574
This object (hereafter SBS 1108) was originally selected as an ultraviolet-excess object during the course of the Second Byurakan Survey (SBS, Markarian and Stepanian 1983). An outburst detected by CRTS on 2012 April 22 (=CSS120422:111127+571239) led to an identification as an SU UMa-type dwarf nova having a period below the period minimum (cf. Kato et al. 2013). Littlefield et al. (2013) studied this object by spectroscopy and found He I emission of comparable strength to the Balmer lines, indicating a hydrogen abundance less than 0.1 of ordinary hydrogen-rich CVs but still at least 10 times higher than that in AM CVn stars. The object received special attention since it is considered to be a candidate progenitor of an AM CVn system (also known as EI Psc-type objects) (Littlefield et al. 2013).
The 2016 outburst was detected by the ASAS-SN team at V=15.44 on March 17. Although subsequent observations detected superhumps (vsnet-alert 19615, 19674), the 2016 outburst was not as well observed as in 2012 and superhumps were detected only on two nights (table 111). We could not make a comparison of O − C diagrams between the 2012 and 2016 observations due to the insufficiency of observations in 2016.     colors suggested an object below the period gap (Kato et al. 2012b   This object (hereafter SDSS J115207) was originally selected as a CV by the SDSS (Szkody et al. 2007). Although Szkody et al. (2007) suspected an eclipsing system, its nature was established by Southworth et al. (2010), who determined the orbital period of 0.06770(28) d and an mass ratio of 0.14(3). Savoury et al. (2011) obtained further observations and refined the values to be 0.067721356(3) d and 0.155(6), respectively.
The object was confirmed to be an SU UMa-type dwarf nova by the detection of superhumps during the 2009 superoutburst (Kato et al. 2010). Due to the poor coverage of the 2009 superoutburst and the limited knowledge of the orbital period at that time, we could not determine superhump and orbital periods precisely in Kato et al. (2010).
We noticed that the ephemeris by Savoury et al. (2011) could not express our eclipse observations and found that the period 0.0677497 d satisfy all the data Savoury et al. 2011;Kato et al. 2010 (6)      This object (hereafter SDSS J153015) was originally selected as a CV by the SDSS (Szkody et al. 2009). The dwarf nova-type variation was confirmed by CRTS observations (Drake et al. 2014     3.124 SDSS J155720.75+180720.2 This object (hereafter SDSS J155720) was originally selected as a CV by the SDSS (Szkody et al. 2009   nova hunt). 20 The object showed a number of outbursts in the ASAS-3 data in the past.
The 2016 outburst was detected by R. Stubbings at a visual magnitude of 11.8 on April 17. Subsequent observations detected superhumps (vsnet-alert 19751; figure 159). The times of superhump maxima are listed in table 121. The period variation was rather smooth and we gave a global P dot rather than giving stages. An analysis of the post-superoutburst data did not yield a superhump period.

TCP J01375892+4951055
This object (hereafter TCP J013758) was discovered by K. Itagaki at an unfiltered CCD magnitude of 13.2 on 2016 October 19. 21 There is a blue SDSS counterpart (g=19.85 and u − g=0.15) and also a GALEX UV counterpart. The object was suspected to be a dwarf nova. Subsequent observations detected growing superhumps (vsnet-alert 20238). Superhumps with a stable period were observed 2 d after the discovery (vsnet-alert 20242, 20268; figure  160).   tered CCD magnitude of 12.2 on March 16. 22 Multicolor photometry by S. Kiyota showed a blue color, suggesting a dwarf nova-type outburst. A spectroscopic study by K. Ayani on March 20 showed Balmer absorption lines (Hβ to Hδ). The Hα line was not clear probably because the absorption is filled with the emission. The spectrum indicated a dwarf nova in outburst. Subsequent observations detected superhumps (vsnet-alert 19635, 19646, 19662, 19684, 19703; figure 162). The times of superhump maxima are listed in table 123. There were clear stages A-C, with a positive P dot for stage B, characteristic to a short-P SH SU UMa-type dwarf nova ( figure 163).
The outburst faded on April 8 and the duration of the total outburst was at least 23 d. The object showed a single post-superoutburst rebrightening on April 25 at 14.5 mag (figure 163).

Statistics of objects
Following Kato et al. (2015a) and Kato et al. (2016a), we present statistics of sources of the objects studied in our surveys (figure 164). Although ASAS-SN CVs remained the majority of the objects we studied, there have also been an increase in MASTER CVs and CRTS CVs. The noteworthy recent tendency is the increase of objects in the Galactic plane discovered by ASAS-SN. This region had usually been avoided by the majority of surveys (the best examples being SDSS and CRTS) and we can expect a great increase of dwarf novae if the Galactic plane is thoroughly surveyed by ASAS-SN. This increase of CV candidates in the Galactic plane, however, has made it difficult to distinguish dwarf novae and classical novae. Several dwarf novae studied in this paper were also flagged as "could also be a nova" on the ASAS-SN Transients page. Some objects in this paper (ASASSN-16jb and ASASSN-16ow) were initially suspected to be Galactic novae. Although they have not been a serious problem in studying dwarf novae, the supposed nova classification might cause a delay in timeresolved photometry to detect superhumps in the earliest phase, and observers should keep in mind the dwarf nova-type possibilities of nova candidate in the Galactic plane.

Period distribution
In figure 165, we give distributions of superhump and estimated orbital periods (see the caption for details) since Kato et al. (2009). For readers' convenience, we also listed ephemerides of eclipsing systems newly determined or used in this study in table 124. When there are multiple observations of superoutbursts of the same object, we adopted an average of the measurements. This figure can be considered to be a good representation of the distribution of orbital periods for non-magnetic CVs below the  Kholopov et al. (1985) in the latest version and objects named in New Catalog of Suspected Variable Stars (NSV: Kukarkin et al. 1982). The categories CRTS, MASTER, ASAS-SN represent objects which were discovered in respective surveys. A fraction of objects discovered by these surveys are already named in GCVS and are included in the category GCVS. period gap, since the majority of CVs below the period gap are SU UMa-type dwarf novae. The following features reported in Kato et al. (2016a) are apparent: (1) the sharp cut-off at a period of 0.053 d and (2) accumulation of objects ("period spike") just above the cutoff.
We determined the location of the sharp cut-off (period minimum) by using the Bayesian approch. We assumed the following period distribution D(P orb ): D(P orb ) ∝ c 1 , (P orb ≤ P min ) 1/(P orb − c 2 ), (P orb > P min ).
(6) P min is the cut-off and c 1 , c 2 are parameters to be determined. We defined the likelihood to obtain the entire sample of our SU UMa-type dwarf novae by using this distribution (the distribution is normalized for a range of 0.01-0.13 d). We obtained the parameters by the MCMC  (14) Psh ( (Lower) distribution of orbital periods. For objects with superhump periods shorter than 0.053 d, the orbital periods were assumed to be 1% shorter than superhump periods. For objects with superhump periods longer than 0.053 d, we used the calibration in Kato et al. (2012a) to estimate orbital periods. The line is the model distribution to determine the period minimum (equation 6, see text for the details). method as follows: c 1 = 1.93(25), c 2 = 0.0471(7) and P min = 0.052897(16). The value of P min is insensitive to the functional form above P min . The resulting distribution is drawn as a line in the lower panel of figure 165.
Although the model does not properly reproduce the location of the period spike, the numbers of dwarf novae are lower than the best fit curve above P orb ∼ 0.09 d. This appears to correspond to the period gap, contrary to our finding in Kato et al. (2016a). Figure 166 represents updated relation between P dot for stage B versus P orb . We have omitted poor quality observation (quality C) since Kato et al. (2016a) and simplified the symbols. The majority of new objects studied in this paper follow the trend presented in earlier papers.

Mass ratios from stage A superhumps
We list new estimates for q from stage A superhumps (Kato and Osaki 2013)  An updated distribution of mass ratios is shown in figure 167 [for the list of objects, see Kato and Osaki (2013), Kato et al. (2015a) and Kato et al. (2016a) (2017b) classified SDSS J105754 as a period bouncer and we have two objects (ASASSN-16dt and ASASSN-16js) near the location of SDSS J105754. Both objects are likely identified as period bouncers and these detections demonstrate the efficiency of the stage A superhump method. The present study has strengthened the concentration of WZ Sge-type dwarf novae around q = 0.07 just above the period minimum, as reported in Kato et al. (2015a) and Kato et al. (2016a).

WZ Sge-type objects
WZ Sge-type dwarf novae are a subclass of SU UMa-type dwarf novae characterized by the presence of early superhumps (Kato et al. 1996a;Kato 2002;Ishioka et al. 2002; see a recent review Kato 2015). They are seen during the earliest stages of a superoutburst, and have period almost equal to the orbital periods.
These early superhumps are considered to be a result of the 2:1 resonance (Osaki and Meyer 2002). These objects usually show very rare outbursts (once in several years to decades) with large outburst amplitudes (6-9 mag or even more, Kato 2015) and often have complex light curves (Kato 2015). The WZ Sge-type dwarf novae are of special astrophysical interest for several reasons. We list two of them: (1) From the point of view of outburst physics, the origin of the complex light curves, including repetitive rebrightenings, is not well understood. They  (3) are also considered to be analogous to black-hole X-ray transients which often show rebrightenings (cf. Kuulkers et al. 1996) and there may be common underlying physics between WZ Sge-type dwarf novae and black-hole X-ray transients.
(2) From the point of view of CV evolution, they are considered to represent the terminal stage of CV evolution and they may have brown-dwarf secondaries. Studies of WZ Sge-type dwarf novae are indispensable when discussing the terminal stage of CV evolution, such as the period minimum and period bouncers (e.g. Knigge 2006;Knigge et al. 2011;Patterson 2011;Kato 2015). We used the period of early superhumps as the approximate orbital period (Kato et al. 2014a;Kato 2015; labeled as 'E' in table 3). In table 127, we list the parameters of WZ Sgetype dwarf novae (including likely ones).
It has been known that P dot and P orb are correlated with the rebrightening type [starting with figure 36 in Kato et al. 2009 and refined in Kato et al. (2009)-Kato et al. (2015a and Kato (2015), and updated in Kato et al. (2016a)]. The five types of outbursts based on rebrightenings are: type-A outbursts [long-duration rebrightening; we include type-A/B introduced in Kato (2015)], type-B outbursts (multiple rebrightenings), type-C outbursts (single rebrightening), type-D outbursts (no rebrightening) and type-E outbursts (double superoutburst, with ordinary superhumps only during the second one). In figure 168, we show the updated result up to this paper. In this figure, we also added objects without known rebrightening types. These objects have been confirmed to follow the same trend, which we consider to represent the evolutionary track [see subsection 7.6 in Kato (2015)]. The respectively. The filled circles, filled squares, filled stars, filled diamonds represent q values from a combination of the estimates from stage A superhumps published in four preceding sources (Kato and Osaki 2013;Nakata et al. 2013;Kato et al. 2014b;Kato et al. 2014a;Kato et al. 2015a;Kato et al. 2016a and references therein), known q values from quiescent eclipses or radial-velocity study, q estimated in this work and dwarf novae in the Kepler data (see text for the reference), respectively. The objects in "this work" includes objects studied in separate papers but listed in table 1. We are not aware whether such a large number of detections were by chance or as a result of the recent change in detection policies of transients such as ASAS-SN. The most notable common fea-tures of these objects are the small number of normal outbursts. Since the short supercycle reflects the high masstransfer rate (cf. Osaki 1996), the small number of normal outbursts is unexpected.
A likely solution to this apparent inconsistency is the disk tilt, which would prevent the accreted matter accumulating in the outer edge of the disk and it would suppress normal outbursts (Ohshima et al. 2012;Osaki and Kato 2013a;Osaki and Kato 2013b). It has been also demonstrated that the prototypical example V503 Cyg (supercycle 89 d) showed negative superhumps (Harvey et al. 1995), which are considered to be a consequence of a disk tilt (e.g. Wood and Burke 2007;Montgomery and Bisikalo 2010). The temporary emergence of frequent normal outbursts in V503 Cyg (Kato et al. 2002b) also suggests that normal outbursts were somehow suppressed, most likely by a disk tilt. More recent examples in Kepler dwarf novae V1504 Cyg and V344 Lyr in relation to transiently appearing negative superhumps were discussed in Osaki and Kato (2013a), Osaki and Kato (2013b) and it has become more evident that the state with negative superhumps (i.e. the disk is likely tilted) is associated with the reduced number of normal outbursts.
It has been proposed that a high mass-transfer rate is prone to produce a disk tilt in a hydrodynamical model (Montgomery and Martin 2010). If it is indeed the case, the large number of SU UMa-type dwarf novae with few normal outbursts but with short supercycles may be a result of easy occurrence of a disk tilt in high-mass transfer systems and may not be surprising. A search for negative superhumps in the four systems reported in this paper is recommended. Long-term monitoring is also encouraged to see whether these objects switch the outburst mode as in V503 Cyg.
We should make a comment on another SU UMa-type dwarf nova V4140 Sgr with a short supercycle (80-90 d, Borges and Baptista 2005). Borges and Baptista (2005) and Baptista et al. (2016) used the eclipse mapping method and derived a conclusion that the distribution of the disk temperature in quiescence is incompatible with the diskinstability model and they interpreted that the outbursts in this object are caused by mass-transfer bursts from the secondary. We noticed that, despite its short supercycle, this object rarely shows normal outbursts (for example, there were no outburst between 2017 March 12 and April 16, observations by J. Hambsch). The object appears to be a V503 Cyg-like one and we can expect a disk tilt. The apparent deviation of the distribution of the disk temperature may have been caused by an accretion stream hitting the inner parts of the disk when the disk is tilted and may not be contradiction with the picture of the disk- instability model. Since no other V503 Cyg-like objects are eclipsing, we have had no observation about the structure of the disk in V503 Cyg-like objects. We propose to study V4140 Sgr more closely for detecting negative superhumps, and searching for a switch in the outburst mode to test the possibility of the V503 Cyg-like nature.

Summary
We provided updated statistics of the period distribution. We obtained the period minimum of 0.05290(2) d and confirmed the presence of the period gap above P orb ∼ 0.09 d. We refined the P orb -P dot relation of SU UMa-type dwarf novae, the updated evolutionary track using stage A superhumps and refined relationship between P orb -P dot versus the rebrightening type in WZ Sge-type dwarf novae. We also provide basic observational data of superoutbursts we studied for SU UMa-type dwarf novae.
The objects of special interest in this paper can be summarized as follows: • Four objects (NY Her, 1RXS J161659, CRTS J033349 and SDSS J153015) have supercycles shorter than 100 d. These objects do not resemble ER UMa-type dwarf novae but show infrequent normal outbursts as in V503 Cyg. We consider that these properties may be caused by a tilted disk and we expect to detect negative superhumps in these systems.
• DDE 48 is likely an ER UMa-type dwarf nova. NSV 2026 also has a short supercycle but with frequent normal outbursts.
• ASASSN-16ia showed a precursor outburst prior to the WZ Sge-type superoutburst. This is the first certain case of such a precursor outburst in a WZ Sge-type superoutburst.
• ASASSN-16js has a low mass ratio and is most likely a period bouncer. ASASSN-16gh is also a candidate for a period bouncer.
• ASASSN-16iw is a WZ Sge-type dwarf nova with five  168. P dot versus P orb for WZ Sge-type dwarf novae. Symbols represent the type of outburst: type-A (filled circles), type-B (filled squares), type-C (filled triangles), type-D (open circles) and type-E (filled stars) (see text for details). On the right side, we show mass ratios estimated using equation (6) in Kato (2015). We can regard this figure as to represent an evolutionary diagram.
• MASTER J021315 is located in the period gap. This object likely showed long-lasting phase of stage A. Outbursts in this system were relatively rare and it should have a low mass-transfer rate.
• ASASSN-16kg, ASASSN-16ni, CRTS J000130 and SDSS J113551 are also SU UMa-type dwarf novae in the period gap. ASASSN-16ni is possibly an SU UMa-type dwarf nova in or above the period gap.
• MASTER J030205 showed a likely spin period and it is likely a rare intermediate polar among SU UMa-type dwarf novae.
• Five objects OY Car, GP CVn, V893 Sco, MASTER J220559 and SDSS J115207 are eclipsing and we present refined eclipse ephemerides for some of them. Technology (MEXT) of Japan. This work was also partially supported by Grant VEGA 2/0008/17 and APVV-15-0458 (by Shugarov, Chochol, Sekeráš), NSh-9670.2016.2 (Voloshina, Katysheva) RFBR 17-52-175300 (Voloshina), RSF-14-12-00146 (Golysheva for processing observations data from Slovak Observatory) and APVV-15-0458 (Dubovsky, Kudzej). ASAS-SN is supported by the Gordon and Betty Moore Foundation through grant GBMF5490 to the Ohio State University and NSF grant AST-1515927. The authors are grateful to observers of VSNET Collaboration and VSOLJ observers who supplied vital data. We acknowledge with thanks the variable star observations from the AAVSO International Database contributed by observers worldwide and used in this research. We are also grateful to the VSOLJ database. This work is helped by outburst detections and announcement by a number of variable star observers worldwide, including participants of CVNET and BAA VSS alert. The CCD operation of the Bronberg Observatory is partly sponsored by the Center for Backyard Astrophysics. We are grateful to the Catalina Real-time Transient Survey team for making their real-time detection of transient objects and the past photometric database available to the public. We are also grateful to the ASAS-3 team for making the past photometric database available to the public. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France. This research has made use of the International Variable Star Index (VSX) database, operated at AAVSO, Cambridge, Massachusetts, USA.

Supporting information
(In the PASJ verision): Additional supporting information can be found in the online version of this article: Tables.  Figures. Supplementary data is available at PASJ Journal online.