The morphological type of M33 has been classified on all the existing systems: Sc (37), Sc+ (38), Sc II-III (39), SA(s)cd (40), fS3 or fS4 (41a, b), and L; cd, Ha (42). Taken together, these indicate that M33 is a late-type right-hand ordinary (non-barred) spiral galaxy, with an apparent axial ratio of roughly 2: 1, with an indefinite number of long, moderately broad, patchy spiral arms that can be traced directly into the nuclear region (which has an F-type spectrum), and with a smooth underlying haze or halo. Vorontsov-Velyaminov (43) noted that the spiral arms could be traced to within 20" (75 pc) of the nucleus.
Whitford (44) measured M33 photoelectrically through a 54' circular diaphragm, and obtained mpg = 6.9 after correction for foreground stars and conversion to the international scale. Stebbins and Whitford (45a, b, c) included M33 in their six-colour photoelectric photometry of galaxies. In all three studies they used diaphragms of fixed sizes, centred on the nucleus; none of the diaphragms was large enough to include more than a small central portion of M33. As expected from the composite nature of the spectrum, the colour indices showed both violet and red colour excesses when compared with the colours of an individual star. Tifft's (46) four-colour photometry and McClure's and van den Bergh's (47) five-colour photometry also included only the nuclear region of M33. Although one of the primary objectives of multicolour photometry of galaxies is to provide information on the relative numbers of different types of stars contributing to the light of the galaxies, no photometric model of the stellar population of M33 has yet appeared in the literature.
A different type of photometric study is surface photometry in one or a few colours to determine the orientation of a galaxy and the distribution of light within it. Seyfert (48) found that, in spite of the very patchy appearance of M33 in photographs, most of the light followed a very smooth distribution. The spiral arms were barely visible on NS and EW tracings of photographs in red and blue light. The arms were somewhat bluer than the nucleus. Patterson (49) discovered that the gradient of photographic luminosity followed an exponential law - photographic magnitude a linear function of radius - except within 6' of the nucleus. From microphotometer tracings every 30° of position angle, she found a position angle of the major axis of 20°, an inclination of the galactic plane to the plane of the sky of 60°, and an integrated mpg of 6.0 out to 30'. Because Danver (50) worked only with the bright, inner arms of the galaxy, his results for position angle and inclination disagreed with those of Patterson and of more recent studies. However, he also noted the difference in pitch angle of the two main spiral arms, cf. (19), and he found that isophotes crossed the north-preceding half of the minor axis farther away from the galactic nucleus than they crossed the south-following half. Currently, the definitive photometric studies of M33 are the photographic photometry by Holmberg (38), (51a, b) on the international system and the photoelectric photometry by de Vaucouleurs (52) on the UBV system. De Vaucouleurs found a position angle of 23°, an inclination of 55° and a total mB of 6.27.
Vashakidze (53) measured 13 per cent mean optical polarization for M33, which was in line with the trend he found of percentage polarization increasing with advancing morphological type in a sample of 11 galaxies.
The existence of emission nebulae within M33 has been known at least since Pease (24) obtained a spectrogram of NGC 604. Detection of these emission regions has been accomplished primarily by comparison of blue plates (sensitive to [O II] 3727 Å) and/or red plates (sensitive to H) to yellow plates and/or infra-red plates. Catalogues of emission regions have been compiled by Haro (54), Aller (55a, b), Shajn (56), and Sérsic (57). By means of narrow H interference filters Courtés and Cruvellier (58a, b) found many emission regions in the central portion of M33, which is burned out on plates obtained by more classical methods. Within the errors of measurement, the H II regions in M33 have the same chemical composition as those in our own Galaxy (59a, b), and the Magellanic Clouds (60). Most of the emission nebulae are of low excitation (61), but some higher-excitation regions are found farther from the nucleus (62). The linear diameters of the largest H II regions are quite similar to those in the LMC. They have been used as calibrators for a method of distance determination to the more distant galaxies (63a, b) - (65). The masses of the largest H II regions have been calculated by Shajn and Haze (66a, b). Their measurements yield, after conversion to the presently accepted distance scale, a mass of the order of 3 × 106 M for NGC 604, the largest H II region in M33. (All distance-dependent quantities have been recalculated for consistency with a distance of 720 kpc (64).)
In addition to the emission lines in the H II regions, a diffuse background of line emission has been found at 3727 A (67) and at H and [N II] 6548, 6584 Å (68). The ratio of the [N II] lines to H indicates higher excitation for this diffuse component than for the classical H II regions.
Radio-continuum radiation from M33 has been observed at frequencies between 159 MHz and 3 GHz (69a - i). Fig. 1 is a graph of the published values of the integrated flux density from M33 as a function of frequency. The integrated flux densities can be represented approximately by a power-law spectrum, S = S0( / 0). The lines in the figure represent least-squares solutions for the power law: S0 (0 = 1 GHz) 5 flux units (5) and -1.0. The 4C catalogue (70) contains no source within 45' of the position of the galaxy, indicating that at 178 MHz the emission is smoothly distributed over a region significantly larger than 1' in extent. At centimetre wavelengths, the radiating region is ~ 0.5° in diameter.
Figure 1. The radio-continuum spectrum. The points are published integrated flux densities of M33, with mean errors indicated where known. The points are labelled with authors' initials and dates of publication (69a - i). The lines represent three different linear-least-squares regression curves of log S upon log . All points were used in the 'weighted' and 'unweighted' solutions. In the 'weighted' solution, the four values obtained by mapping the galaxy (BH at 159 MHz, V at 750 and 1410 MHz, and D at 2795 MHz) were given double weight. Only these four points were used in the 'high weight only' solution.
The neutral-hydrogen distribution was found to extend 'far beyond the structure visible on the available photographs' (71) even in the earliest 21-cm line observations of M33. However, an attempt to find an H I link between M33 and its nearest neighbour, M31, was unsuccessful (72). The angular resolution of the early data was insufficient to provide any details of the hydrogen distribution. But, more recently, Davies (73) and Burke (73) noted that their observations with the 250-foot and 300-foot telescopes, respectively, indicated a ring-shaped distribution of the neutral hydrogen in M33. The radius of this ring has been placed between 15' and 20' (73), (74a, b), (75). However, the centre of the ring is displaced by a few minutes of arc from the optical centre of the galaxy (76), and there are other departures from axial symmetry. The most striking of these are a pair of 'wings' on the north-preceding and south-following ends of the galaxy, nearly 1° from the galactic centre (75).
Because there is a difference of several magnitudes between the photographic luminosities of the brightest stars of Populations I and II, the first stars to be resolved in external galaxies are Population I objects. The first individual stars recognized in M33 were Duncan's (33) three variables and a nova. Hubble (34), (35) - who at the time was interested primarily in determining the distance to M33 - discussed this nova, Cepheids, and a luminosity function derived from star counts of selected fields within M33. Hubble and Sandage (77) found that the brightest variable stars (intermediate-type supergiants) formed a group with a small dispersion in luminosity, thus providing another distance criterion for galaxies. By comparing a colour-magnitude diagram of stars in M33 to a composite colour-magnitude diagram of young galactic clusters (Population I objects in our own Galaxy), Johnson and Sandage (78) rederived the distance to M33.
De Vaucouleurs (79) extended the star counts to cover the entire face of M33 down to 20th photographic magnitude (mB 20.2 or MB -4.4). On the basis of these counts, he found (80) the centre of the distribution of the brightest Population I stars to be displaced to the southwest of the optical centre. Vorontsov-Velyaminov (43), (81a, b) found that the supergiants in M33 tended to crowd together and form compact groups, 65 of which he catalogued. He stated that the dust in the spiral arms appeared to avoid close contact with such groups.
The first mention of the possible presence of Population II stars in M33 was Hubble's (82) discovery of 12-15 objects which he tentatively identified as globular clusters. He noted, however, that they were fainter and bluer than most globular clusters in M31 and our Galaxy Hiltner (83) performed UBV photometry on 23 clusters, including those found by Hubble. He suggested that they might be rich galactic clusters rather than globulars, but noted that, on the other hand, they might be counterparts of the faint blue globular cluster NGC 7492 in the Milky Way. These results were corroborated by Kron and Mayall (84), who included 4 clusters in M33 in their PV photometry. Vorontsov-Velyaminov (43) stated that the only trace of a Population II nucleus he could find for M33 was a feature no larger than a single globular cluster. Walker (85a, b) obtained an infra-red plate (103a-U + RG8) of M33, with a limiting magnitude between the apparent magnitudes of Populations II and I red giants, and a blue plate (103a-O + GG13) exposed to the same density for OB stars as the infra-red plate. Making a composite from the blue negative and an infra-red positive, Walker brought to light the existence of a smooth, elliptical distribution of Population II red giants in M33. From this evidence he argued that M33 must be just as old as M31, and that the difference in their morphological appearance was due solely to different initial conditions.
Although the rotation of M33 was discovered in 1916 (25), the first study of the rotation appeared more than two decades later (86a, b). This study by Mayall and Aller was based upon the radial velocities of 25 emission nebulae, determined from low-dispersion spectrograms. The rotation curve was roughly linear within 15 of the nucleus, reached a peak of ~ 105 km s-1 radial velocity (~ 125 km s-1 circular velocity) between 15' and 20', and decreased thereafter. Models of the mass distribution responsible for the observed rotation were discussed by various authors (87a - h) between 1941 and 1957. Because of the limited accuracy of the data, quite different models could yield satisfactory agreement. In contrast to the situation in M31 and the Milky Way, no rapidly rotating nucleus has been seen in M33 (88a, b). The observations of M33 were of limited usefulness in the discussions (89a, b, c), (90) of whether spiral arms lead or trail in galactic rotation because of the difficulty in determining which side of M33 is closer to the observer. After the controversy had been settled in 1958 (90), de Vaucouleurs' photometric study of M33 (46) showed that its arms trailed, like those of the rest of the galaxies for which the sense of rotation could be determined.
New kinematical data for M33 were added by the observations of Dieter (91) and Volders (92) in the 21-cm line of neutral atomic hydrogen. These observations yielded much better velocity resolution than did the optical data. However, the positional accuracy of the measurements was quite low, since the half-power beam widths of the telescopes were respectively 53' and 34'. The masses derived in the two radio studies (~ 2 × 1010 M) were both significantly larger than the masses found from the Mayall and Adler data. Burke, Turner & Tuve (93) obtained 21-cm observations with angular and velocity resolution both improved over those available to Dieter and Volders, but restricted their study to the major and minor axes of M33. They found that the systemic velocities obtained from the observations along the two axes disagreed by 5 km s-1, an amount which they stated 'is significant but is not understood'. They did not derive a mass for M33 from their observations.
Brandt (94) used 110 Å mm-1 spectrograms to determine radial velocities from the H emission of 24 H II regions in M33, mostly the same regions used by Mayall and Aller. Brandt's measurements indicated that the turnover point of the rotation curve lay beyond the outermost optical features within M33 (in disagreement with the earlier optical measurements). Because his measurements did not reach the turnover point of the rotation curve, he supplemented them with Dieter's and Volders' radio data, finding a total mass of 3.9 × 1010 M for M33. Also, by truncating his model at appropriate distances from the centre of the galaxy, he found 'remarkably consistent' agreement between his results and all of the previous mass determinations. Takase and Kinoshita (95) combined the radio data with the older optical velocity measurements and found a significantly smaller total mass.
The most detailed kinematical study of M33 available is by Carranza et al. (68). Using Fabry-Perot interferograms of the H emission from M33, they measured 1048 points in the galaxy at dispersions of 15-30 Å mm-1. They found an asymmetry between the northern and southern halves of the rotation curve; the mean of the two agreed with Brandt's measurements. Furthermore, in the southern half of the galaxy they found that the gas in the spiral arm appeared to be rotating more rapidly than the gas in the adjacent interarm region at the same distance from the nucleus.
The extensive 21-cm study by Gordon (75) with a beamwidth of 10' confirmed the location of the peak of the rotation curve outside the optical picture of the galaxy. It also confirmed the existence of the asymmetry of the rotation curve and found departures from circular motion, particularly in the outer parts of M33. Gordon's value for the total mass of the galaxy was 4.9 × 1010 M, the increase over Brandt's value arising primarily from extrapolating a greater fraction of the mass in the outermost fringes of the galaxy.
5 n1 flux unit = 10-26 W m-2 Hz-1. Back.