3 Results

3.1 CO

Figure 1: a) The distribution in MBM 32 of $\int$ $T_{\rm {R}}^*$dv for 12CO(J=1-0). We combined the 4$^\prime$ raster MRS and the 2$^\prime$ raster HRS data. The three panels show the emission in the velocity intervals 2 - 7 km s-1(top), -5 - 0 km s-1 (middle), and 0 - 2 km s-1 (bottom). Contour levels are 1 to 10 K km s-1 in steps of 1 K km s-1. The small crosses are the observed positions. The asterisk indicates the reference position (0$^\prime$,0$^\prime$), see text. The large dashed box is the area mapped in 12CO(2-1) (Fig. 1b), and the large full-drawn box is the area mapped with the HRS in 13CO(1-0) (Fig. 1c). The small full-drawn box is the area mapped in 13CO(2-1) (Fig. 2a) small dashed boxes are the areas mapped in 12CO(3-2) (Figs. 2b,c). b) The same as Fig. 1a, for the 12CO(J=2-1) observations with the MRS made at a 2$^\prime$ grid spacing in the central area of MBM 32. Contour levels are 1 to 7 K km s-1 in steps of 1 K km s-1. The drawn box in the bottom panel indicates the area observed in 13CO(2-1), and the dashed boxes those observed in 12CO(3-2). c) The same as Fig. 1b for the 13CO(J=1-0) observations with the HRS made at a 2-4$^\prime$ grid spacing in the central area of MBM 32. Contour levels are 0.25 to 1.5 K km s-1 in steps of 0.25 K km s-1

In Fig. 1a we show the distribution of the velocity integrated emission $\int$ $T_{\rm {R}}^*$dv of ¹²CO(1-0) in MBM 32, obtained from Gaussian fits to the lines. We used the high velocity resolution (HRS) 2$^\prime$ raster data where possible and at the other positions we used the lower velocity resolution (MRS) 4$^\prime$ data (put to the HRS intensity scale - see Sect. 2.1). We separately show the emission in three velocity intervals, separated by the LSR velocities from the Gaussian fits. The distribution of peak $T_{\rm {R}}^*$ is essentially the same as that of $\int$ $T_{\rm {R}}^*$dv(although details differ), i.e. line widths do not vary much within the region. The H I spectra east of $\alpha$-offset -12$^\prime$ clearly show three velocity components (see Sect. 3.2). Therefore, we also separated the CO emission in three velocity intervals, the main component of which is at velocities $V_{\rm {lsr}}$ between 2 - 5 km s^-1 (the maximum $T_{\rm {R}}^*$ is 4.8 K at (0$^\prime$, 4$^\prime$)), and weaker and smaller components in the western part of the region between -5 - 0 km s^-1(the maximum $T_{\rm {R}}^*$ is 2.6 K at (-16$^\prime$, 16$^\prime$)), and between 0 - 2 km s^-1 (the maximum $T_{\rm {R}}^*$ is 1.3 K at (-52$^\prime$, -12$^\prime$)). The offsets of these $T_{\rm {R}}^*$ maxima differ from peaks in Fig. 1a due to small differences in line widths and noise. The separation between the two smaller components in CO is less clear in the MRS data east of $\alpha$-offset -25$^\prime$, where the smallest component may continue to -1 km$\,$s^-1. All components show fragments of typically a few resolution elements in size, connected by extended emission. The length of the main component is about 1 $.\!\!^\circ$6 (2.8 pc).

Because of the relatively narrow lines, the high resolution data show higher $T_{\rm {R}}^*$ values than the MRS data. Maxima are $T_{\rm {R}}^*=6.5$ K [at (6$^\prime$, 2$^\prime$)] for the main component, 2.1 K [at (-12$^\prime$, 14$^\prime$)] for the weak component, and 3.2 K [at (-6$^\prime$, 14$^\prime$)] for the negative velocity component (the MRS peak at (-52$^\prime$, -12$^\prime$) was not observed).

These ¹²CO(1-0) data can be compared in Figs. 1b and 1c with the ¹²CO(2-1) and ¹³CO(1-0) emission in the same region. ¹²CO(2-1) peak positions are close to, but not coinciding (probably due to pointing errors and/or noise) with the ¹²CO(1-0) peaks and the lines are weaker: $T_{\rm R}^*=4.0$ K at (8$^\prime$, -2$^\prime$) for the main component, 1.8 K at (-14$^\prime$, 14$^\prime$) for the small component, and 1.9 K at (-20$^\prime$, 16$^\prime$) at $V_{\rm {lsr}}<0$ km$\,$s^-1. In ¹³CO(1-0) only the main cloud component was detected. It shows two equally strong (1.6 K) peaks at (0$^\prime$, 4$^\prime$) and (6$^\prime$, 20$^\prime$). Only smaller parts of this region were observed in ¹²CO(3-2) and ¹³CO(2-1) (see Fig. 2). In ¹²CO(3-2) the strongest lines are seen within the second ¹³CO(1-0) peak, which splits into three 2.2 K peaks at (5$^\prime$, 25$^\prime$), (8$^\prime$, 18$^\prime$), and (2$^\prime$, 16$^\prime$). Within the latter area the strongest ¹³CO(2-1) emission is near the middle peak [0.7 K at (9$^\prime$, 17$^\prime$)].

Figure 2: The same as Fig. 1, for a) the 13CO(J=2-1) observations with the MRS made at a 1-2$^\prime$ grid spacing in the central area of MBM 32. Contour levels are 0.15 to 0.75 K km s-1 in steps of 0.15 K km s-1. b,c) the 12CO(J=3-2) emission in the interval 2 - 7 km s-1 obtained with the MRS at a 1$^\prime$ grid spacing in two regions in the central area of MBM 32. Contour levels are 0.25 to 1.5 K km s-1 in steps of 0.25 K km s-1

All maps show much structure at all scales, which cannot easily be described. The general structure of the distribution is the same for all transitions, but in the details there are differences, which are possibly caused by excitation effects and/or pointing errors. A quantitative analysis is made in Sect. 4. The position of the two NH₃ maxima detected by Heithausen et al. ([1998a]) or the two H₂CO maxima found by Heithausen et al. ([1987]) are in the general region where there are also maxima seen in the different ¹²CO or ¹³CO transitions, but there is no clear correlation between the locations of any maximum. This confirms the anticorrelation found by Heithausen et al. ([1987]) of H₂CO with lower resolution CO data. Only a small part of the cloud was observed in either NH₃ or H₂CO, and it seems likely that clouds such as MBM 32 contain a large number of such cloudlets. It is unclear whether these cloudlets can form stars. Kun ([1992]) has searched for H$\alpha$ emission-line stars towards a sample of high latitude clouds, among them MBM 32. She found three objects with H$\alpha$ emission which are seen in the direction of the main CO component of MBM 32: K92 27 [at offset (7 $.\mkern-4mu^\prime$6, -7 $.\mkern-4mu^\prime$5)], K92 30 [at offset (13 $.\mkern-4mu^\prime$5, 5 $.\mkern-4mu^\prime$5)], and K92 33 [at offset (47 $.\mkern-4mu^\prime$4, 28 $.\mkern-4mu^\prime$4)]. Since no higher resolution spectra have been made of these stars, it is not yet known whether these objects are indeed T Tauri stars associated with MBM 32.

3.2 Velocity structure

Channel maps in intervals of 1 km s^-1 of the ¹²CO(1-0) emission are shown in Fig. 3. The main part of the cloud is at $V_{\rm {lsr}}= 3 - 4$ km s^-1, but both the central and eastern maxima show small velocity gradients. Gaussfits show that the velocities of the line peaks range from -4.0 to 5.3 km$\,$s^-1 for the MRS ¹²CO(1-0) data.

Figure 3: Channel maps in intervals of 1 km s-1 of the 12CO(1-0) emission in MBM 32 observed with the MRS at a 4$^\prime$ grid. Contour levels are 0.6 to 3.6 K km s-1 in steps of 0.6 K km s-1. The small crosses are the observed positions

The other transitions have not been mapped in the whole area of Fig. 1a and therefore show emission in a smaller velocity range. For the central region the velocity structure is shown in more detail in Fig. 4 for ¹²CO(2-1). The channel maps for the two small areas observed in ¹²CO(3-2) and in ¹³CO(2-1) (not shown) show structures which agree with those visible in the 1$^\prime$ resolution ¹²CO(1-0) map of a small (8$^\prime$ $\times$ 16$^\prime$) region by Magnani et al. ([1990]).

At most positions the CO line shapes are Gaussian within the noise and show no structure such as self absorption or wings. However there are some exceptions. This is best seen in the higher resolution ¹²CO(3-2) spectra: the main cloud shows at offsets around (7$^\prime$, 21$^\prime$) a stronger line at 3.5 km s^-1 and weaker emission at about 2.0 km s^-1. Northeast of this position the line shapes are asymmetric. We explain this by overlapping in this region of the two fragments of the main cloud of MBM 32 (see Fig. 4).

Figure 4: The same as Fig. 3 for the 12CO(2-1) emission in the central part of MBM 32 observed with the MRS at a 2$^\prime$ grid. Note that the velocity step used changes at $V_{\rm {lsr}}$ 2 km s-1. Contour levels are 0.5 to 2 K km s-1 in steps of 0.5 K km s-1( $V_{\rm {lsr}}<2$ km s-1) and 0.25 to 2 K km s-1 in steps of 0.25 K km s-1( $V_{\rm {lsr}}>2$ km s-1)

The line widths of the negative velocity cloud (and possibly also of the weak emission in the interval 0 - 2 km s^-1) are somewhat larger than those of the main cloud. This is consistent with the findings of de Vries et al. ([1987]). Within the main cloud, line widths at the west side ( $\Delta\alpha<+6$$^\prime$) are significantly smaller than in the rest of the cloud. This is seen in all transitions, but the effect depends on the velocity resolution of the different spectra. The 115 GHz HRS data show an average value of 1.12 $\pm$ 0.26 km s^-1 (135 positions) in the western part and of 1.49 $\pm$ 0.33 (249 positions) in the eastern part. This division is not clearly correlated with some clumps or peaks in the cloud. The negative velocity emission has an average line width of 1.92 $\pm$ 0.63 km s^-1 (108 positions). Except for the region NE of (7$^\prime$, 21$^\prime$) (see above), the differences in line widths cannot be explained by the presence of more than one velocity component along the line of sight.

3.3 Hi

The neutral hydrogen emission towards MBM 32 might be confused by an unknown amount due to fore- and background components at about the same velocity as the CO. However, Vladilo et al. ([1994]) concluded from optical data that in the direction of (among others) MBM 32 the local bubble extends to the assumed distance of MBM 32 (100 pc). If this is true, the amount of H I foreground emission will be low. Background emission was also found at velocities considerably different from that of MBM 32. The strongest H I lines found in the mapped region are 31.8 K at offset (88$^\prime$, 12$^\prime$), at the eastern border of MBM 32. This emission is at 2.9 km$\,$s^-1, near the velocity of the molecular cloud emission. The weakest line near the velocity of the main CO component is 7.2 K (at 4.7 km$\,$s^-1) at (-16$^\prime$, 16$^\prime$), in the region where is the negative velocity component of CO. Stronger H I emission at negative velocities is usually due to the wings of the main component emission. The weakest line at negative velocities (7.1 K) is at (-20$^\prime$, 40$^\prime$), north of the corresponding CO line, and at a minimum in the IRAS 60 and 100 $\mu$m maps (see Sect. 3.4). To compare CO and H I emission for MBM 32 we have plotted in Fig. 5 the H I and ¹²CO(1-0) spectra, summed for all positions where we detected ¹²CO(1-0). For the main CO component we distinguish between the two largest cloud fragments, east and west of $\Delta \alpha=+34$$^\prime$. It is seen that in each of the Figs. 5a to d, the H I spectra have a peak close to the CO velocity. Except for the smallest CO component (Fig. 5b), it is also the strongest feature in each H I spectrum, which in addition contains emission from both other components. In each spectrum there is also a very broad (>15 km s^-1) component, which is due to warmer, unrelated gas.

Figure 5: Comparison of H I (full drawn) and 12CO(1-0) (MRS; histograms) emission. All H I spectra (the broader lines) are on the same (TB) scale. The spectra were averaged for positions where there is CO emission with $V_{\rm {lsr}}<0$ km s-1 a), $0<V_{\rm {lsr}}<2$ km s-1 b), $V_{\rm {lsr}}>2$ km s-1 at $\Delta \alpha < 34$$^\prime$ c), or >34$^\prime$ d). The CO intensities ( $T_{\rm {R}}^*$) were multiplied by a factor of 10

Because the H I emission has a larger velocity range and intrinsic line width than the CO emission (see Fig. 5), and shows small differences in velocity compared to the CO (see below), it is not much of use to integrate the H I emission over the same velocity interval as the CO or to investigate the peak intensities. Instead we show in Fig. 6 the channel maps with intervals of 2 km s^-1 between -8 and +10 km s^-1. H I at negative velocities is mainly present in the western part of the cloud where there is also CO emission at $V_{\rm {lsr}}<0$ km$\,$s^-1. The main part of the H I emission is in the interval +1 to +4 km s^-1 where there is some correlation of the H I distribution with that of the strongest CO component. The strongest H I is just NE of the eastern CO peak. The negative velocity CO component, which it's approximately elliptical distribution shows a similar structure in H I at -4 to -2 km$\,$s^-1 and -2 to 0 km$\,$s^-1 which seems to be slightly outside the CO emission region.

The velocity structure of the H I gas associated with MBM 32 is shown from another perspective in Fig. 7, which contains position ( $\Delta\alpha$) - velocity diagrams at $\delta$-offsets from -24$^\prime$ to +44$^\prime$. The velocity components from Fig. 5 are clearly visible, along with many other structures, such as secondary maxima, holes, and velocity gradients, which cannot easily be modeled.

We have made Gaussian fits to the H I spectra using two to four velocity components: one weak (few K) and broad (typically 20 - 40 km$\,$s^-1) component and stronger but narrower (<10 km$\,$s^-1) components. The broad component shows little variation in intensity over the mapped region and we assume that it is not associated with MBM 32. In Fig. 8 we compare the velocities of the narrow components with the $V_{\rm {lsr}}$ of the MRS+HRS ¹²CO(1-0) data. Only in part of the main CO cloud (-5$^\prime$ $<\Delta \alpha <37$$^\prime$) the velocities of one of the H I components agrees with the CO velocities within 0.5 km$\,$s^-1. For the other CO components differences are in the range -1.5 - +1.5 km$\,$s^-1. At positions where there are such differences, the H I spectra do not show signs of self absorption due to cold foreground gas. These differences are less than the velocity dispersion of the molecular gas: the line widths of spectra of the sum of all ¹²CO(1-0) MRS data are 2.4 km$\,$s^-1 (main component) and 3.4 km$\,$s^-1(negative velocity component). It is possible that only the H I gas in a certain velocity range is converted into H₂, leaving stronger emission at higher or lower velocities. Or H I at one edge of the cloud has a different velocity due to external pressure. H I line widths of all components are about 4 - 5 km$\,$s^-1: the average value for the component associated with the main CO cloud is 4.95 $\pm$ 1.98 km$\,$s^-1; those of the three components at negative offsets are 4.35 $\pm$ 0.91 km$\,$s^-1 ( $V_{\rm {lsr}}<0$ km$\,$s^-1), 4.75 $\pm$ 1.75 km$\,$s^-1( $0<V_{\rm {lsr}}<4$ km$\,$s^-1), and 4.31 $\pm$ 1.20 km$\,$s^-1 ( $V_{\rm {lsr}}>4$ km$\,$s^-1). These values are a factor 2 - 4 larger than the CO line widths in the same region, which is less than expected from the difference in mass of H I and CO. It suggests that either the molecular gas is relatively more turbulent, or the H I gas is colder. We have used the H I data to derive masses assuming the emission is optically thin [N(H I) = $1.822\ 10^{18}\int T_B$dv cm^-2]. The resulting values (not including He) are indicated in Fig. 8 for the different velocity components. We also indicate the H₂ masses derived from CO (see Sect. 4.1). If all H I is associated with MBM 32, both masses are approximately equal for the main CO component. For the negative velocity component, the H₂ mass is somewhat larger than the H I mass (however this depends which region is considered).

3.4 FIR emission

The distribution of the dust emission associated with MBM 32 can be seen in the 60 and 100 $\mu$m IRAS (ISSA) maps in Figs. 9a and b, which have an angular resolution of about 5$^\prime$. For comparison the integrated ¹²CO(1-0) and H I distribution are shown in Figs. 9c and d on the same scale. The IRAS data are the maps produced from all HCONs, and were taken from the Sky Survey Atlas. To avoid negative values of flux densities in the 60 $\mu$m map we added here a constant offset to all pixels of 1 MJy/sr. There clearly is FIR emission associated with the molecular clouds, both at 60 and 100 $\mu$m. However the distribution is confused by emission from dust associated with the H I gas, which has a slightly different distribution. The H I maximum near offset (50$^\prime$, 40$^\prime$) outside the CO cloud shows as a peak in the 60 $\mu$m map suggesting the presence of relatively warm dust in this region. A combined image of 12, 60, and 100 $\mu$m emission in MBM 32 was produced also by Verter & Rickard ([1998]), however without interpretation.

Figure 6: Channel maps of H I emission in intervals of 2 km s-1 between -8 and 10 km s-1. Contour levels are 6 to 60 K km s-1 in steps of 6 K km s-1. For comparison in the upper left panel the 12CO(1-0) distribution for the three velocity intervals from Fig. 1, as grey-scale (Fig. 1s), full-drawn (Fig. 1b), and as dotted (Fig. 1c) contours

Figure 7: Position (Right Ascension offset) - velocity diagrams of H I emission at different $\delta$-offsets (indicated in the panels). Offsets are with respect to $\alpha (1950)=9^{\rm h}28^{\rm m}42^{\rm s}$, $\delta (1950)=+66$$^\circ$5$^\prime$. Contour levels are 2 to 30 K in steps of 2 K

Figure 8: $V_{\rm {lsr}}$ plotted against Right Ascension offset (with respect to $\alpha (1950)=9^{\rm h}28^{\rm m}42^{\rm s}$) using velocities from Gaussian fits to H I profiles (only narrow components) and 12CO(1-0) data (lower panel). Masses (without He contribution) for different velocity components are indicated

Before we can compare the CO, H I, and FIR emission, we have to convolve the CO and IRAS 60 and 100 $\mu$m data first to the same resolution as the H I (9 $.\mkern-4mu^\prime$2) data. Here we use only the ¹²CO(1-0) MRS map because it covers the largest area on the sky. Subsequently we sampled the FIR data at the same (4$^\prime$) raster as both other data sets. The correlation between FIR and H I emission was investigated by selecting positions where the convolved CO spectra showed no emission stronger than about 0.2 K (114 positions). We compared the 100 $\mu$m emission with the H I emission integrated over the whole velocity range where there is emission (-80 to 50 km$\,$s^-1), with the H I emission of the narrow components in the Gaussfits, and with all the local emission (including the broad component). It appears (using the bisector linear least squares fit, see Isobe et al. [1990]) that in the latter case the correlation is best (correlation coefficient 0.60), however the offset remains rather uncertain:

$$F_{\rm 100~\mu m} = (-3.438 \pm 0.524) + (0.034 \pm 0.003) \int T_{\rm b}({\rm HI}){\rm d}v. \eqno(1{\rm a})$$

Here, $F_{\rm 100~\mu m}$ is the flux density in MJy sr^-1. For 60 $\mu$m the correlation coefficient is 0.49:

$$F_{\rm 60~\mu m} = (-1.545 \pm 0.094) \!+\! (0.0059 \pm 0.0006) \int T_{\rm b}({\rm HI}){\rm d}v. \eqno(1{\rm b})$$

The correlation for 100 $\mu$m (see Fig. 10a) is better than found by Meyerdierks & Heithausen ([1996]) for the HLCs of the Polaris Flare, because in the case of MBM 32 there are no other branches with different slope, probably because towards the Polaris Flare a larger area of the sky was investigated containing regions with different properties. There is no significant difference in Figs. 10a and b between the distribution of the points located in the eastern (negative CO velocity) and western cloud parts.

The results (Eq. 1) were used to correct the FIR emission for dust associated with H I towards the 403 positions where there is CO emission, and subsequently we obtained the following relation between 100 $\mu$m emission and ¹²CO(1-0) integrated line intensities:

$$\Delta F_{\rm 100~\mu m} {=} (-0.737 \pm 0.062) + (0.618 \pm 0.021) \int T_{\rm {R}}^*(1-0){\rm d}v . \eqno(2{\rm a})$$

The correlation coefficient is 0.73. We distinguished between three parts of the cloud and obtained the following slopes: 0.857 $\pm$ 0.046 (for $\Delta \alpha <-10$$^\prime$), 0.601 $\pm$ 0.029 (for -10$^\prime$< $\Delta \alpha < 34$$^\prime$), and 0.596 $\pm$ 0.028 (for $\Delta \alpha >34$$^\prime$). In the latter case the correlation coefficient is best (0.79). These results are shown in Fig. 10c. The range in CO intensity is smaller for the eastern cloud part (containing the negative velocity components), where the slope is steepest.

For 60 $\mu$m our result is (correlation coefficient 0.60):

$$F_{\rm 60~\mu m} = (-0.144 \pm 0.011) + (0.088 \pm 0.004) \int T_{\rm {R}}^*(1-0){\rm d}v. \eqno(2{\rm b})$$

For the negative velocity CO component ( $\Delta \alpha <-10$$^\prime$) there is little correlation. The slope does not change when using only data with $\Delta \alpha >-10$$^\prime$: 0.086 $\pm$ 0.004. The results are shown in Fig. 10d.

The slope in Eq. (1a) translates into a ratio of H I column density and 100 $\mu$m flux density of N(H)/ $F_{\rm 100~\mu m} = 5.4 \ 10^{19}$ cm^-2 Jy^-1 sr. Comparing this with Eq. (2) of Meyerdierks & Heithausen ([1996]) we find that the $100~\mu$m emission per H atom is a factor 2.6 smaller in MBM 32 than in the Polaris Flare.

The slope in Eq. (2a), 0.618 MJy s (K km sr)^-1 is slightly steeper than the value found by Meyerdierks & Heithausen [0.5 MJy s (K km sr)^-1]. However their Fig. 7 essentially shows a scatterplot and no accurate value for the slope could be determined. From Eqs. (1a) and (2a) we obtain a conversion factor between $W_{\rm {CO}}(=\int T{\rm d}v(^{12}$CO(1-0)) and N(H₂) of (0.17 $\pm$  $0.02)\ 10^{20}$ cm^-2 (K km s ^-1)^-1, much lower than derived from our CO data alone ( $0.7\ 10^{20}$; see Sect. 4.1), but similar to values derived in the same way towards some parts of the Draco nebula by Herbstmeier et al. ([1993]). However, this value is a lower limit (possibly by a factor 2 or 3) if the dust temperature is lower in the inner parts of the cloud (see Meyerdierks & Heithausen [1996]).

Our results can also be used to derive the dust temperature and mass. A correlation between 60 and 100 $\mu$m emission associated with MBM 32 gives (correlation coefficient 0.92):

$$\Delta F_{\rm 100~\mu m} = (0.200 \pm 0.026) + (7.840 \pm 0.132) \Delta F_{\rm 60~\mu m}. \eqno(3)$$

The offset is mainly caused by the uncertain zero point in the subtraction of dust associated with background H I gas. From the slope in Eq. (3), we derive a dust temperature of 20 K, assuming a dust emissivity proportionally to $\lambda^{-2}$. The data are consistent with a constant dust temperature in the cloud which does not depend on CO intensity. To derive the mass of the dust associated with the CO from the 100 $\mu$m data (using the method outlined by Wood et al. [1994]), we have to correct the latter by a remaining offset. The value of 0.200 MJy sr^-1 from Eq. (3) is probably not correct because there might be in addition a remaining offset in the 60 $\mu$m emission. Instead we choose the value from Eq. (2a), (-0.737), giving the amount of dust associated with CO (there could be additional dust around the CO cloud). We used no lower limit for the optical depth at 100 $\mu$m. The total dust mass then is 0.073 $\pm$ 0.022 $M_\odot$, where the error is determined by an uncertainty in the offset of 0.5 MJy sr^-1. The dust mass in the western part of MBM 32 ( $\Delta \alpha <-10$$^\prime$) is 0.020 $M_\odot$. The maximum optical depth at 100 $\mu$m is 0.23 m Neper (corresponding to A_V=4.6 mag, using Eq. (5) from Wood et al. [1994]), at offset (8$^\prime$, 16$^\prime$). This value appears relatively high compared to optically derived values (0.6 mag; see Sect. 1). The A_V, derived from the cloud-averaged 100 $\mu$m flux density, is 1.49 mag, which is also higher than the optical values. The ratio of gas mass (derived from the CO data, see Sect. 4.1: 21.5 $M_\odot$), and dust mass (0.076 $M_\odot$) is 283, which is higher than the usually adopted values of 100, and suggests that part of the dust is colder than 20 K. However the fraction of such cold dust is less than in most other molecular clouds, where gas to dust ratios above 1000 are found (Wood et al. [1994]). Our data of MBM 32 do not suggest the presence in this cloud of an infrared excess such as found by Meyerdierks & Heithausen ([1996]) in the Polaris Flare (interpreted as H₂ gas without detected CO).