2D Spectroscopy of Candidate Polar-Ring Galaxies: I. The Pair of Galaxies UGC 5600/09
L.V. Shalyapina, O.A. Merkulova, V.A. Yakovleva, E.V. Volkov
aa r X i v : . [ a s t r o - ph ] F e b Astronomy Letters, 2007, vol. 33, No.8, pp.520-530Translated from Pis'ma v Astronomi heskii Zhurnal, 2007, vol. 33, No.8, pp.585-5972D Spe tros opy of Candidate Polar-Ring Galaxies:I. The Pair of Galaxies UGC 5600/09L.V. Shalyapina1, O.A.Merkulova2, V.A.Yakovleva, and E. V. VolkovSobolev Astronomi al Institute, St. Petersburg State University,Universitetskii pr. 2, Petrodvorets, 198904 RussiaAbstra tObservations of the pair of galaxies VV 330 with the SCORPIO multimode instrumenton the 6-m Spe ial Astrophysi al Observatory teles ope are presented. Large-s ale velo -ity (cid:28)elds of the ionized gas in H α and brightness distributions in ontinuum and H α havebeen onstru ted for both galaxies with the help of a s anning Fabry Perot interferometer.Long-slit spe tros opy is used to study the stellar kinemati s. Analysis of the data ob-tained has revealed a omplex stru ture in ea h of the pair omponents. Three kinemati subsystems have been identi(cid:28)ed in UGC 5600: a stellar disk, an inner gas ring turnedwith respe t to the disk through ∼ ◦ , and an outer gas disk. The stellar and outer gasdisks are non oplanar. Possible s enarios for the formation of the observed multi ompo-nent kinemati gala ti stru ture are onsidered, in luding the ase where the large-s alevelo ity (cid:28)eld of the gas is represented by the kinemati model of a disk with a warp. Thevelo ity (cid:28)eld in the se ond galaxy of the pair, UGC 5609, is more regular. A joint analysisof the data on the photometri stru ture and the velo ity (cid:28)eld has shown that this is prob-ably a late-type spiral galaxy whose shape is distorted by the gravitational intera tion,possibly, with UGC 5600.Key words: galaxies, groups of galaxies, intera ting galaxies kinemati s, stru ture.1 Introdu tionPolar-ring galaxies (PRGs) onstitute a rare lass of dynami ally pe uliar systems in whi ha ring or a disk of gas, dust, and stars rotates around the main body approximately in thepolar plane (Whitmore et al. 1990). The formation of a polar ring is widely believed tobe related to the intera tion between galaxies or even to their mergers. Attempts to modelsu h pro esses have been made in several works (see Reshetnikov and Sotnikova 1997; Bekki1lshalastro.spbu.ru2olga_merkulovalist.ru 1998; Bournaud and Combes 2003). Despite the progress in interpreting the morphology ofthe observed obje ts, PRGs, a hieved in the above works, it should be re ognized that there ismost likely no single universal me hanism that would explain the polar-ring formation in ea hspe i(cid:28) ase. Therefore, it is appropriate to perform numeri al simulations to reprodu e theobserved morphology and velo ity (cid:28)eld for ea h of the andidate PRGs.Mu h observational data on PRGs have been a umulated to date, but many questionsrelated to the polar-ring formation, stability, and age still remain debatable. New, more om-plete and a urate properties of the stellar population and the interstellar medium, and the starformation pro esses is needed to solve these and several other problems. Opti al data on themotion of the gas and stars are usually obtained with long-slit spe trographs. The results ofsu h observations are not always interpreted unambiguously, parti ularly in the ase of omplexmulti omponent systems.Analysis of the velo ity (cid:28)elds gives mu h more information. The velo ity (cid:28)elds are on-stru ted mainly from observations of neutral hydrogen or mole ular gas (mostly in CO lines)in the radio frequen y range and from 2D or 3D spe tros opy in the opti al range. 2D spe -tros opy makes it possible to study the motions of the gas and stellar subsystems in galaxies ingreatest detail and to obtain data on a qualitatively new level. However, su h data for PRGsand similar obje ts are still very s ar e. For example, the velo ity (cid:28)elds for the best knownPRGs, su h as NGC 2685 (Shane 1980) and NGC 4650A (Arnaboldi et al. 1997), and someother obje ts were onstru ted from 21- m observations, but the spatial resolution of these datais low. Until re ently, the velo ity (cid:28)elds in the opti al range were obtained only for the entralregions of two galaxies, NGC 2685 and IC 1689 (Sil' henko 1998).The atalog of PRGs byWhitmore et al. (1990) in ludes 157 galaxies and the existen e oftwo roughly orthogonal kinemati systems ( lassi al PRGs) has been on(cid:28)rmed observationallyonly for 15(cid:21)20 of them. Obviously, the dis overy of ea h new obje t and a detailed studyof galaxies from Whitmore's atalog, parti ularly by means of 2D spe tros opy, are of greatinterest.In 2000, a program to investigate PRGs and related obje ts from the atalog of Whitmoreet al. (1990) by means of 2D spe tros opy using a multipupil (cid:28)ber spe trograph (MPFS) and as anning Fabry Perot interferometer (FPI) of the 6-m Spe ial Astrophysi al Observatory (SAO)teles ope was initiated at the Astronomi al Institute of the St. Petersburg State University. Oneof the (cid:28)rst interesting results of this program was the dete tion of a superwind from the galaxyNGC 6286 (Shalyapina et al. 2004a). At the same time, two almost perpendi ular gas systemswere found in the galaxy NGC 7468 (Shalyapina et al. 2004b) and the existen e of an inner polarring was assumed on the basis of a joint analysis of 2D spe tros opy and photometry. MPFSdata on the stellar and gas kinemati s in the entral regions of andidates PRGs on(cid:28)rmed the2xisten e of two almost orthogonal kinemati systems in the entral region of the galaxy UGC5600 (Shalyapina et al. 2002). Two stellar kinemati systems were dis overed in the galaxyUGC 4892, whi h suggested the presen e of a satellite (Hagen-Thorn et al. 2003). The velo ity(cid:28)eld in H α for the galaxy NGC 2685 was onstru ted from FPI observations (Hagen-Thorn etal. 2005).In this paper, we present FPI observational data and new results of long-slit spe tros opy forthe isolated pair of galaxies VV 330 Vorontsov(cid:21)Velyaminov 1959, 1977). An R band image ofthis pair is shown in Fig. 1. The s ale is 0.2 kp per 1 ′′ if the distan e to one of the omponents,UGC 5600, is assumed to be 44.6 Mp at a gala to entri velo ity V gal = 2897 km/s and H = 65 km/s/Mp . Both galaxies have lose line-of-sight velo ities and the separation betweenthem in the plane of the sky is 1 ′ .4. One of the galaxies, UGC 5609, with a fairly omplexstru ture, has not been studied in detail previously. The other galaxy, UGC 5600, is known asthe most likely andidate for PRGs (Whitmore et al. 1990). Our previous study of the stellarand gas kinemati s in the galaxy based on MPFS observations (16 ′′ × ′′ (cid:28)eld) in the greenand red spe tral ranges and on long-slit spe tros opy in the red spe tral range onsisted largelyin analyzing the velo ity (cid:28)eld of the entral region in this obje t (Shalyapina et al. 2002).In this paper, we established that the galaxy has a omplex kinemati stru ture: the starsrotate around the galaxy's minor axis, while the gas, ex ept for the entral 3 ′′ region, rotatesaround its major axis. This led us to the on lusion about the existen e of an inner polar ring.However, the pi ture of the gas motion in the outer regions remained un lear.Figure 1: R -band image of the pair of galaxies VV 330 obtained with the 6-m SAO teles ope.That is why the main obje tive at this stage of our study was to obtain information about3he velo ity (cid:28)eld on the s ales of the entire galaxy, in luding its outermost parts, to ompare thenew observational data with the results for the entral region, and to bring all data to a uni(cid:28)edpi ture. In onne tion with this formulation of the problem, the main emphasis was shifted toan investigation with FPI. The results of studying the stellar kinemati s at a distan e up to18 ′′ from the enter also turned out to be very useful at the (cid:28)nal stage. Our more thoroughand omprehensive investigation showed that UGC 5600 is not just the main galaxy with aninner polar ring, but is a more omplex dynami al system and what was previously taken asthe inner polar ring is one of the omponents of the overall omplex stru ture. This issue isdis ussed in detail in Se tion (cid:16)Kinemati s of the Gas and Stars(cid:17). Until then, we will use theterm (cid:16)inner ring(cid:17) following Shalyapina et al. (2002).2 Observations and data redu tionThe spe tros opi observations of the galaxies UGC 5600/09 were performed at the prime fo usof the 6-m SAO teles ope with the SCORPIO fo al redu er (Afanasiev and Moiseev 2005). A logof observations for this pair of galaxies is given in the Table 1, whi h also provides informationabout the R -band image. Parameters of the fo al redu er in interferometri observations aregiven in Moiseev (2002). Premono hromatization was made using a narrowband (cid:28)lter with F W HM = 20(cid:6)A. The FPI spe tral resolution was about 3(cid:6)A ( ∼
130 km/s). Sin e the (CCDTK1024) dete tor readout was performed in 2 × × ′′ .56 per pixel was obtained in ea h spe tral hannel.Table 1: Log of observationsObje t Instrument, Exposure, Field Seeing Spe tral .À.date se region, (cid:6)AUGC5600/09 IFP 05.03.2002 32 ×
180 5 ′ × ′ ′′ H α UGC5600/09 SCORPIO 05.04.2002 540 6 ′ × ′ ′′ RUGC5600 Slit 23.02.2006 11 × ′′ × ′ ′′ ◦ UGC5609 Slit 21.02.2006 4 ×
900 1 ′′ × ′ ′′ ◦ The interferometri observations were redu ed with software developed at SAO (Moiseev2002). After the primary pro edures (the subtra tion of night-sky lines and the redu tionto the wavelength s ale), the observational data onstitute (cid:16)data ubes(cid:17) in whi h ea h pointin a 512 × F W HM equal to 1.5 hannels and a 2D Gaussian4moothing in spatial oordinates with
F W HM = 2 pixels, was performed using the ADHOCsoftware pa kage.3 Gaussian (cid:28)tting of the H α emission line pro(cid:28)les was used to onstru t thevelo ity (cid:28)elds and mono hromati images. The measurement errors of the line-of-sight velo itiesfor lines with symmetri pro(cid:28)les were ∼
10 km/s. We also onstru ted an image in ontinuumnear H α .We used the method of in lined rings (Begeman 1989; Moiseev and Mustsevoi 2000) toanalyze the velo ity (cid:28)eld. This method allows us to determine the positions of the dynami al enter and the dynami al axis, to re(cid:28)ne the in lination of the galaxy to the plane of the sky,and to onstru t the rotation urve. Analysis of the dependen e of the position angle of thedynami al axis and the in lination on radius provides information about the gas motion in thegalaxy. The MPFS data obtained previously by Shalyapina et al. (2002) were used to onvertthe H α (cid:29)uxes to the absolute energy s ale.The slit-spe trograph observations were arried out with the same SCORPIO instrument inFebruary 2006. The slit length and width are about 6 ′ and 1 ′′ , respe tively, the s ale along theslit is 0 ′′ .36 per pixel, and the spe tral resolution is 5(cid:21)6(cid:6)A. The spe tral range ontained the H β ,[OIII℄ 4959, 5007 (cid:6)A emission lines and absorption lines of the old stellar population: MgI 5175(cid:6)A,FeI + Ca 5270(cid:6)A, and others. The dete tor was a 2048 × B , V , R c olor bands, Karataeva et al. (2001) suggestedthat the galaxy ould be most likely lassi(cid:28)ed as a late-type (S d) spiral. The presen e of su hdi(cid:27)erent estimates of the galaxy type ia most probably explained by the fa t that UGC 5600 is3The ADHOC software pa kage was developed by J. Boulesteix (Marseilles Observatory) and is freely avail-able on the Internet. 5 multi omponent obje t in whi h, as will be demonstrated here, several kinemati subsystemsare learly identi(cid:28)ed.Figure 2 shows the ontinuum (near H α ) and H α brightness distributions for the binarysystem VV 330. Note that both galaxies are ri h in gas and the H α emission is tra eable inthem up to the outermost ontinuum isophotes. Along with di(cid:27)use H α emission, both galaxiesexhibit numerous bright knotes that are probably star-forming regions. The entral region ofUGC 5600 is distinguished by a parti ularly intense emission. Let us onsider in more detailthe brightness distribution for ea h galaxy.UGC 5600. Comparison of Fig. 1 with Figs. 2a and 2b shows that the ontinuum andH α brightness distributions are di(cid:27)erent, parti ulary in the entral region. The main bodyof the galaxy, whi h is an in lined disk whose major axis has a size of ∼ ′′ ( ∼ ∼ ◦ , is identi(cid:28)ed in ontinuum. In the disk, the isophotes are nearlyellipti al in shape with small twists n the southern side. Protrusions are noti eable on theisophotes in the E(cid:21)W dire tion almost perpendi ular to the major axis of the galaxy's mainbody at a distan e of ∼ ′′ from the enter. It is this photometri feature that Whitmore et al.(1990) alled the inner ring. The main body of the galaxy is surrounded by an extended outer b) a) Figure 2: Brightness distributions for VV 330 obtained with FPI: (a) in a narrow ontinuum nearH α and (b) in H α . The gray s ale orresponds to the brightness in arbitrary units for Fig. 2a and inerg/s/ m /ar se for Fig. 2b.round envelope with a lumpier stru ture on the side fa ing the ompanion. Su h features inthe B, V, R brightness distribution were also pointed out by Karataeva et al. (2001); they alsopointed to a de rease in (cid:29)attening for isophotes fainter than 23.5 m /ar se in the B band and6o a turn of the major axis of the ellipses (cid:28)tted to these isophotes through ∼ ◦ to the west.These authors also onsidered the photometri pro(cid:28)les of the galaxy's main body whose shapedoes not show the presen e of a noti eable bulge, while the brightness distribution along themajor axis in the range from 3 ′′ to 18 ′′ is well represented by an exponential disk with thefollowing parameters: µ ,B = 19.7 m /ar se and h B = 5 ′′ (1 kp ).The H α image learly shows an elongation of the isophotes in the E(cid:21)W dire tion in the entralpart of the galaxy. The existen e of this preferential dire tion seems to be not a idental andto be related to the previously noted feature of the ontinuum isophotes (the (cid:16)inner ring(cid:17)).The details of the stru ture of the entral region are better represented in the MPFS data andwere des ribed by Shalyapina et al. (2002). Previously, we already noted that the luminousgas is tra eable roughly to the same boundaries as the outer ontinuum isophote, but thestru ture of the outer gas disk is more omplex and ontains numerous knots of di(cid:27)erentbrightness and sizes. Parti ularly bright knots with sizes up to 5 ′′ (1 kp ) are lo ated inthe southern part of the galaxy, with some of them forming hains. These are most likelystellar omplexes or giant HII regions. The H α (cid:29)ux in the entral region of the galaxy (r ≤ ′′ (3 kp )) is 2.5 × − erg/s/ m and the total (cid:29)ux within the 10 − erg/s/ m /ar se isophoteis 3.8 × − erg/s/ m . The lower limit for the star formation rate (without absorption)obtained using a relation from Kenni utt (1998) for the entral region of UGC 5600 is 0.4 M ⊙ /yrand SFR = 0.7 M ⊙ /yr for the entire galaxy.Whitmore et al. (1990) suggested that the faint outer envelope in UGC 5600 is a ring. Amore detailed study of this outer stru ture (Karataeva et al. 2001) showed that this may be nota ring but two spiral arms wounding ounter lo kwise around the galaxy's main body. To testthis suggestion, we used a spirality riterion based on the expansion of the surfa e brightnessdistribution into a Fourier series of the azimuth angle in the gala ti plane followed by analysisof the shape of the lines of the maxima of individual Fourier harmoni s (Moiseev et al. 2004).For our analysis, we used a deeper (approximately by 0.5 m /ar se ) R -band image that weobtained with SCORPIO (Fig. 1) than that in Karataeva et al. (2001). The amplitude ofthe se ond harmoni turned out to be largest; a two-armed stru ture is satisfa torily (cid:28)tted tothe observed brightness distribution at the following parameters: the position angle (PA) ofthe outer stru ture 160 ◦ and the in lination ( i ) of the disk plane to the plane of the sky 25 ◦ .Similar values of PA and i for the outer isophotes in the B band were obtained by Karataevaet al. (2001).UGC 5609. The ontinuum and R -band images of the galaxy (Fig. 2a and Fig. 1) showa faint, nearly ellipti al disk with a major axis a = 62 ′′ ( ∼ ◦ , and an axial ratio of 0.5. A slight in rease in brightness is observed in the outer partof the disk and we get the impression that a faint envelope with a more blurred inner edge7s present. No rise in brightness is noti eable near the geometri al enter of the isophotes.Only at a distan e of ∼ ′′ to NE is a bright ompa t feature with nearly spheri al isophotesdistinguished. Another prominent feature that stands out against an amorphous disk is lo atedto the west of the enter. This is a fairly bright extended region elongated in the N(cid:21)S dire tion.It has no lear boundaries and its approximate size is 10 ′′ × ′′ (2(cid:21)3 kp ). Sin e the stru tureof this region is omplex, the isophotes in it are highly irregular in shape.The infrared image of the galaxy in the J, H, K bands (2MASS) di(cid:27)ers sharply from thepreviously des ribed ontinuum image. Instead of the amorphous disk, only two bright regions oin ident with the features noted in ontinuum are seen in all three bands. A ording toour estimates of the 2MASS data (we used the alibration from Jarrett et al. 2000), the K magnitude of the bright ompa t feature is ∼ m and its olor index J − H ≈ m .9. Itsspe trum turned out to be typi al of a late-type M1(cid:21)M3? star with a radial velo ity of about(cid:21)50 km/s. Thus, this is a Gala ti foreground star that is proje ted onto the disk of UGC 5609by han e. The surfa e brightness of the se ond, more extended region ranges from 18 m /ar se at the enter to 20 m /ar se on the periphery. The emission in the J, H, K bands is produ edmainly by an old stellar population, whi h is usually grouped in galaxies symmetri ally aboutthe dynami al enter. Therefore, we assume that this extended region is probably the main bodyof the galaxy. This on lusion is also on(cid:28)rmed by the analysis of the kinemati s performedbelow.The ring envelope, the absen e of a bright enter, and the bright knot on one side of the ringall resemble lassi al ollisional ring galaxies, for example, II Zw 28 (Appleton and Marston1997). However, if we examine in more detail the images of this galaxy in di(cid:27)erent olor bands(e.g., Fig. 1 and the blue image from SDSS), then we an noti e that UGC 5609 more likelyresembles a spiral galaxy whose stru ture is distorted by tidal intera tion, espe ially in the blueband. The bright knot is the main body of the galaxy from whi h an ar -shaped spiral armor tidal tail goes away to the north. The se ond arm goes to the east and then turns southat a distan e of about 10 ′′ from the geometri al enter, forming the SE part of the luminousenvelope.The H α image of the galaxy (Fig. 2b) also has a roughly ellipti al shape but with a moreblurred outer edge than that in ontinuum. Numerous knots of various sizes and brightnesses,whi h are probably HII regions, are superimposed on the faint underlying galaxy along thear -shaped stru tures mentioned above. A onsiderable number of bright knots is also observedin the main body of the galaxy; the H α (cid:29)ux from it is ∼ × − erg/s/ m . At the sametime, the total (cid:29)ux within the 10 − erg/s/ m /ar se isophote is 1.3 × − erg/s/ m .8 Kinemati s of the gas and starsUGC 5600. Figures 3b and 3 show the large-s ale line-of-sight velo ity (cid:28)eld for the galaxyobtained in H α with FPI and the velo ity dispersion map. The distribution of the lines of equalvelo ity (isovels) shows regular deviations from the normal pattern of rotation of a (cid:29)at gala ti disk. A kinemati subsystem rotating around the galaxy's major axis, whose existen e wasestablished from MPFS observations (Shalyapina et al. 2002), is learly identi(cid:28)ed in the regionof the inner ring. For the onvenien e of analyzing the total velo ity (cid:28)eld of the galaxy, Figs.3d and 3e present the MPFS velo ity (cid:28)elds for the inner ring and the entral region obtainedfrom stars and ionized gas. As we see from the (cid:28)gures, the MPFS and FPI velo ity (cid:28)elds inH α are in good agreement. In parti ular, the run of the isovels in the nu lear region (r ≤ ′′ ) inboth Fig. 3b and Figs. 3a and 3e oin ides with the dire tion of stellar rotation (Fig. 3d).Previously (Shalyapina et al. 2002), we noted that the system velo ities found from the starsand gas di(cid:27)er. However, our new long-slit observations to be dis ussed below did not on(cid:28)rmthe di(cid:27)eren es. Therefore, we on e again thoroughly analyzed the line-of-sight velo ities of thestellar omponent determined from MPFS data and on luded that they should be in reasedby 30 km/s. Figure 3 shows the orre ted velo ity (cid:28)eld for the stellar omponent; the systemvelo ity is V sys = 2770 ±
10 km/s, whi h is equal, within the error limits, to the system velo itydetermined from emission lines. The velo ity (cid:28)eld of the stellar omponent learly points torotation around the galaxy's minor axis; the northern side approa hes us, while the southernside re edes from us. Previously (Shalyapina et al. 2002), we estimated the position angle(PA dyn = 182 ◦ ). It turned out to oin ide with the position of the photometri axis of thegalaxy's main body (Karataeva et al. 2001). Analysis of the total velo ity (cid:28)eld for the gas omponent (Fig. 3b) shows that the gas in the outer parts rotates around an axis in linedwith respe t to the rotation axis of the stellar disk (for a more detailed dis ussion, see below).Moreover, the rotation has su h a dire tion that the gas in the NW part of the outer envelopere edes from us, while the gas in its SE part approa hes us.Sin e the gas observations on the s ales of the entire galaxy revealed a more omplex stru -ture than that followed from the observations of the entral region, the following question arose.How the stars behave at distan es from the enter larger than those measured with MPFS? Itis with this goal that we performed the long-slit (PA = 0 ◦ ) observations in the green spe tralrange. Figure 4b presents the line-of- sight velo ity urve for the stars; for omparison, twose tions of the velo ity (cid:28)elds for the stellar (MPFS) and gas (FPI) omponents are also shownhere. As we see from the (cid:28)gure, the long-slit and 2D spe tros opi data are in lose agreement,within the error limits, and the pattern of stellar motion is retained in regions farther from the enter. At the same time, there is a signi(cid:28) ant ontrast in the stellar and gas kinemati s at9 b a ed Figure 3: Line-of-sight velo ity (cid:28)elds for UGC 5600 in H α : (a) from three MPFS frames and (b) FPIdata; ( ) velo ity dispersion distribution. The velo ity (cid:28)elds for the galaxy's entral region from MPFSdata: (d) stellar omponent and (e) ionized gas, H α . The step between the isovels in all (cid:28)gures is 20km/s.distan es 10 ′′ ≤ r ≤ ′′ from the enter. The line-of-sight velo ities of the gas obtained fromboth permitted H α lines (FPI) and forbidden lines (long slit, Fig. 4a) are in good agreement.10 ) H β +[OIII]S N b) Slit: starsMPFS: starsS N c) Figure 4: Long-slit line-of-sight velo ity urves for UGC 5600 at PA = 0 ◦ : (a) ionized gas in H β and[OIII℄ 5007(cid:6)A, (b) stellar omponent, the dashed lines in both (cid:28)gures indi ate the FPI data in H α ,( ) rotation urve for the stellar disk (the dashed lines indi ate the model rotation urve of a thinexponential disk).Figure 4 presents the rotation urve for the stellar omponent; the theoreti al rotation urveof a thin exponential disk with a s ale fa tor h disk = 1.2 kp is indi ated here by the dashed11ine. This value is lose to that obtained by Karataeva et al. (2001) from B -band photometri data. The mass of the galaxy within a radius r ≤ ∼ M ⊙ .Our analysis of the large-s ale line-of-sight velo ity (cid:28)eld by the method of in lined rings hasshown the following. In the region 3 ′′ ≤ r ≤ ′′ , the behavior of the isovels orresponds tothe gas rotation around the major axis of the stellar disk. In the outer parts at r ≥ ′′ , thedynami al axis of the gas is along the NW(cid:21)SE line. Thus, it o upies an intermediate positionbetween the dynami al axes of the stellar disk and the inner gas ring. This suggests that thereare two kinemati ally de oupled gas subsystems: an inner ring and an outer disk. The entersof these subsystems oin ide, within the error limits. In the region 10 ′′ ≤ r ≤ ′′ , the stru tureof the line-of- sight velo ity (cid:28)eld is very omplex and it annot be reliably represented by amodel of ir ular motion. Note that the photometri data reveal a sharp de line in brightness, a hange in isophotal (cid:29)attening, and a turn of the major axis of the ellipses (cid:28)tted to the isophotesin this region.Therefore, we onstru ted two separate models of ir ular motion for the inner subsystem(3 ′′ ≤ r ≤ ′′ ) and the outer disk (20 ′′ ≤ r ≤ ′′ ). As we have already noted above, thesystem velo ity in both models is the same and equal to 2770 ± i dyn ∼ ◦ ± ◦ and PA dyn ∼ ◦ ± ◦ .As in our previous paper (Shalyapina et al. 2002), the angle between the planes of the innergas ring and the stellar disk was found to be ∼ ◦ ± ◦ or ∼ ◦ ± ◦ . On the periphery20 ′′ ≤ r ≤ ′′ , the model of gas motion has the following parameters: PA dyn ≃ ◦ ± ◦ and i dyn = 30 ◦ ± ◦ . The angle between the planes of the stellar disk and the outer gas disk isestimated to be 70 ◦ ± ◦ and 40 ◦ ± ◦ .Below, we make several remarks on the velo ity dispersion distribution for louds of ionizedgas. As we see from Fig. 3 , the maximum dispersion is observed in the nu lear region (morethan 100 km/s). As we re ede from the enter, the dispersion de reases and is ∼
50 km/s inthe outer parts. In the region of the inner ring, the dispersion varies between 50 and 100 km/s,but the stru ture of its distribution is highly nonuniform and lumpy. It should be noted thatthere exists a problem of a urately measuring the velo ity dispersion if its value is smallerthan the FWHM of the FPI instrumental pro(cid:28)le,
F W HM = 130 km/s, whi h orresponds toa dispersion σ gas ≈
55 km/s. Estimations of the stellar velo ity dispersion along the galaxy'smajor axis showed that it is lower than the velo ity dispersion of the ionized gas at the enter( σ ⋆ ≤
50 km/s) and in reases to 80 km/s at a distan e of 10 ′′ from the enter.UGC 5609. Figure 5a presents the line-of-sight velo ity (cid:28)eld onstru ted from FPI datain H α . At (cid:28)rst glan e, it appears regular and similar to the rotation of the gas disk around theminor axis. A detailed analysis of the (cid:28)eld showed that its stru ture is more omplex; numeroustwists that an be aused by various fa tors are noti eable on the isovels. First, the stru ture12f UGC 5609 is probably severely distorted either by tidal intera tion or by galaxy ollision.Se ond, the measurement errors of the line-of-sight velo ities depend on the brightness level,while the brightness distribution in H α is highly nonuniform. Therefore, the behavior of theisovels is related to di(cid:27)erent errors. Third, to the NE of the geometri al enter of the isophotesat a distan e of ∼ ′′ , the distortions in the velo ity distribution are produ ed by the absorption omponent (from a foreground star) in the red H α wing. Sin e the line-of-sight velo ity rangefor this galaxy is small, these fa tors will a(cid:27)e t the shape of the isovels. c) b) a) Figure 5: UGC 5609: (a) observed velo ity (cid:28)eld, (b) residual velo ity (cid:28)eld, ( ) model (solid line) andobserved (squares) rotation urves. The step between the isovels in all (cid:28)gures is 20 km/s.As we noted above, on the one hand, UGC 5609 is similar to ollisional ring galaxies and,on the other hand, this an be a spiral galaxy whose stru ture is distorted by gravitationalintera tion. Our analysis of the velo ity (cid:28)eld for this galaxy was also performed by the methodof in lined rings, but we onstru ted di(cid:27)erent models: both with allowan e made only forthe ir ular velo ity ve tor omponent and with the assumption of nonzero radial velo ities(expansion of the outer ring stru ture). As the dynami al enter, we took either the enterof the western knot, whi h we believe to be the galaxy's main body, or the geometri al enterof the ellipses (cid:28)tted to the isophotes. Next, we onstru ted the model velo ity (cid:28)elds and ompared them with the observed (cid:28)eld. Note that the residual velo ities for most modelsare low, 10(cid:21)20 km/s. Sin e the velo ity is almost onstant in the entral region along theminor axis of the ellipses, the system velo ity is determined unambiguously and its value is2740 ± i = 55 ◦ . If the dynami al enter oin ides with the geometri al enter of the ellipses, then the maximum positive residualvelo ities are observed in the region of the main body. Near the geometri al enter (r ≤ ′′ ), theresidual velo ities are opposite in sign and do not ex eed 10(cid:21)15 km/s, but a strong depression isnoti eable in the initial segment (r ≤ ′′ ) of the rotation urve. The same pi ture is observedin the models with a nonzero ring expansion velo ity. The residual velo ity rea hes +20 km/s13n the periphery of the galaxy in the northern and southeastern parts of the ring. Thus, we (cid:28)ndno eviden e for the existen e of an expanding envelope (ring) at the level of the measurementerrors.If the dynami al enter is displa ed by about 10 ′′ to the west to the enter of the main body,then the entire velo ity (cid:28)eld is satisfa torily des ribed by the model of ir ular motion. Figure5b presents the residual velo ity map. The rotation urve an be represented by an exponentialdisk (Fig. 5 ) with h = 4 kp ; in this ase, a small hange in the position of the dynami alaxis from 15 ◦ to 30 ◦ is observed and the residual velo ities in the region of the northern spiralrea hes their maximum value of +30 km/s. If this is assumed to be a tidal stru ture, then it an be warped with respe t to the plane of the gala ti disk. The total mass of the galaxy isestimated to be 1.6 × M ⊙ .5 Dis ussionPreviously, we have already mentioned that the andidate PRGs from the list by Whitmoreet al. (1990) are the obje ts of our study. However, as we see from the previous se tion, no lassi al polar ring is observed in the galaxy UGC 5600, as, for example, in the lassi al PRGsNGC 4650A, NGC 2685, and IC 1689. In this ase, we most likely en ounter the phenomenon ofwhat was alled kinemati ally de oupled omponents in the literature (see Bertola and Corsini1999). The phenomenon of ounterrotation in several galaxies (Rubin 1994; Galletta 1996), the ir umnu lear polar rings (Sil' henko et al. 1997; Corsini et al. 2002), and the lassi al polarrings (Whitmore et al. 1990) also fall under this name. Despite the di(cid:27)eren es in photometri stru ture and the kinemati pe uliarities, all of them are probably united by the origin: theyare the result of (o asionally multiple) galaxy mergers or mass a retion onto the galaxy fromother galaxies. Let us onsider the possible s enarios that unite the observed pe uliarities ofVV 330.Sin e the observed pi ture for UGC 5600 is very omplex, let us summarize the mainobservational fa ts before passing to a dis ussion.(1) The main body is a stellar disk with a diameter of ∼ dyn ∼ ◦ , and i dyn ∼ ◦ (Shalyapina et al. 2002). The presen e of an ionized gas rotating in the same way as the starswas established in the entral region (r ≤ ′′ ).(2) The inner gas ring rotating in a plane lose to the polar plane of the stellar disk hasPA dyn ∼ ◦ and i dyn ∼ ◦ . The pattern of kinemati s in the region of the inner gas ringis ompli ated by the fa t that we possibly see the total emission from the ionized gas thatbelongs to the ring and the gala ti disk, but we annot separate them be ause the instrumentalpro(cid:28)le has a large FWHM. 143) The outer gas disk or ring of low surfa e brightness is tra eable at distan es approx-imately from 4 to 8 kp from the gala ti enter. For this omponent, PA dyn ∼ ◦ and i dyn ∼ ◦ . Thus, the outer gas disk rotates in a plane that is appre iably in lined (either by70 ◦ to by 40 ◦ ) to the plane of the galaxy's main body.Taking into a ount the fa ts listed above, we onsidered the following ases.It an be assumed that UGC 5600 is the proje tion of two galaxies: one is a gas-poor polar-ring galaxy seen at a large angle to the plane of the sky and the other is a gas-ri h late spiral.Sin e the enters of all the observed stru tures oin ide, within the error limits, and sin e thesystem velo ities of the stellar and gas omponents are lose, this ase seems unlikely, althoughit annot be ruled out ompletely.The next explanation for the observed features of UGC 5600 is the merger of two galaxies;di(cid:27)erent s enarios are possible in this ase. For example, the existen e of stellar and outer gasdisks rotating in di(cid:27)erent planes an result from the merger of two (late and earlier-type) galax-ies that produ ed the above independent kinemati subsystems. The pi ture is ompli ated bythe presen e of an inner ring. To explain this feature, we an attempt to use a me hanismsuggested by Friedli and Benz (1993) that is related to the presen e of a bar in the galaxy.However, no signatures of an extended bar were found in the velo ity (cid:28)eld and the photometri stru ture for UGC 5600. Re urrent mergers an serve as an alternative explanation.In the last ase that we will onsider, a warped polar ring is assumed to be present inUGC 5600. One of the formation me hanisms for su h rings is mass a retion from a gas-ri hgalaxy onto a neighboring galaxy. A warped polar ring or disk is formed around the latteras a result of mass transfer (Bournaud and Combes 2003). The se ond omponent of thepair, UGC 5609, may be onsidered as the donor galaxy. Let us assume that the entire gasin UGC 5600 onstitutes a single system and that the turn of the dynami al axis from the entral regions to the periphery is related to the warp of the gas disk. In this ase, the observedfeatures of the velo ity (cid:28)eld an be the result of proje tion e(cid:27)e ts. To test this hypothesis, we onstru ted a model of a warped ring. We used the density distribution in a thin exponentialdisk whose rotation urve is shown in Fig. 4 . We assumed that PA dyn and i dyn hanged linearlyfrom the inner edge of the ring to its outer edge. As the (cid:28)rst approximation, at the ring edges,we took the angles obtained by analyzing the velo ity (cid:28)eld by the method of in lined rings(see the previous se tion) for the inner ring and the outer gas omponent. We al ulated thevelo ities in the ellipses at given position of the dynami al axis and in lination. The boundaryvalues of the angles were then varied to a hieve the best agreement with the observed velo ity(cid:28)eld. The model velo ity (cid:28)eld shown in Fig. 6a was onstru ted at the following parameters:PA dyn = 240 ◦ and i dyn = 45 ◦ for the inner boundary; PA dyn = 300 ◦ and i dyn = 30 ◦ for the outerboundary. The angle between the normals to the surfa e at the inner and outer boundaries is15 ) b) Figure 6: UGC 5600: (a) rotation model with a disk warp and (b) residual velo ities. ∼ ◦ . As we see from the residual velo ity map (Fig. 6b), the model velo ity (cid:28)eld des ribessatisfa torily the observed one, ex ept the region at distan es of 10 ′′ (cid:21) 15 ′′ from the enter tothe west. The feature in the velo ity (cid:28)eld is also observed at this lo ation.At this stage of our study, we do not have su(cid:30) ient information to hoose between thesuggested s enarios.In on lusion, let us dis uss the data obtained for UGC 5609. Based on the ontinuum andH α images of the galaxy and on their omparison with the image in the blue band, we madethe following assumptions: this is either a ollisional galaxy or a spiral galaxy whose shapewas distorted by the gravitational intera tion, possibly, with UGC 5600. Our analysis of thevelo ity (cid:28)eld revealed no expanding envelope. Based on the entire data set, we on luded thatUGC 5609 is most likely a late-type spiral galaxy with two ar shaped spiral arms; as a result,the illusion of an envelope is reated. One of the arms may be a tidal tail. The high brightnessof this feature in H α , the R band, and the blue range and its low brightness in ontinuum nearH αα