Figure 3a shows a plot of log (L_IR/L _B) versus log (H-a EQW) for all Sbc-Sdm spirals in the sample of Young et al. (1989) that are not HI gas depleted (Note: the H-alpha equivalent widths have not been corrected for internal extinction).

We can see from this figure that there is a tendency for (L_IR /L_B) to increase with increasing H-a EQW, although the correlation is weak (with a linear regression coefficient r = 0.58). Indeed, if it were not for the two points with the largest H-a EQW (i.e. the merger galaxies NGC 3310 and NGC 1614) it would be reasonable to conclude that there was little or no correlation.

Figure 3b is the corresponding plot of log (SB_IR) versus log (H-a EQW). This plot shows that there is a definite correlation between SB_IR and H-a EQW although, as in figure 3a, the correlation is still relatively weak (r = 0.62).

A different picture emerges however, if we sub-divide the galaxies plotted in figure 3a and 3b into two groups based on their infra-red excess. Figures 3c and 3d are a replot of figures 3a and 3b with those galaxies with log (L _IR/L _{H- a}) > 2.33 (the sample mean)designated as having a “high infrared excess” (i.e. high IRE) and those with log (L _IR/L_{H- a}) < 2.33 as havinga “low infrared excess” (i.e. low IRE).

We can see from figures 3c and 3d that restricting the range in infrared excess splits the galaxies into two distinct groups each of which shows a much stronger correlation between the two star formation parameters, log (L _IR /L _B) (r = 0.81) and
log (SB_IR ) (r = 0.80), and log (H-a EQW). The solid lines in figures 3c and 3d are the least squares fit to the points for the high IRE galaxies. The low IRE appear to be systematically shifted down and/or to the right compared to the high IRE galaxies in both plots

A similar systematic shift to that seen between high and low IRE galaxies is also evident in figures 3e and 3f were we have segregated the sample galaxies into those with ISM that are H ₂ rich (i.e. log (M_HI/M _H2) < 0.37 - the sample mean), and those that are H_2­ poor (i.e. log (M_HI/M_H2 ) > 0.37). The solid line in figures 3e and 3f are the least squares fit to the points for the H₂ rich galaxies. (Note: The atomic hydrogen masses (M_HI) of H ₂ rich galaxies are NOT systematically different from those of H₂ poor galaxies. What is markedly different between the two groups of galaxies is their the molecular hydrogen masses (M _H2 ) (Young et al 1989,1996). Hence, the sub-division of galaxies into H₂ rich and H ₂ poor based upon their log (M_HI /M _H2 ) values (Young et al. 1989,1996)). Figures 3g and 3h show that the systematic shift between high IRE, H ₂ rich galaxies and low IRE, H₂ poor galaxies is related to galaxy luminosity. In these plots we have segregated the sample galaxies by blue luminosity (L _B ) using the sample mean of 10.49. It is immediately evident that the low luminosity spirals are systematically shifted down and to the right of the (solid) least squares fit line to the high luminosity spirals.

Hence, it appears that the high IRE, H ₂ rich spirals are systematically more luminous in blue light than low IRE, H₂ poor spirals. This implies that we are dealing with two fundamentally different populations of galaxies and that the changes we see in IRE and the M_HI/M_H2 ratio in figures 3c through 3h are directly dependent on parameter(s) related to galaxy mass (e.g. metallicity, disk column density of HI etc.). Interestingly, Kenney and Young (1988) also concluded from their study ofSbc-Sm Virgo cluster spirals that a parameter(s) related to galaxy mass determines whether or not a galaxy’s ISM is H₂ rich or H₂ poor.

Persson and Helou (1987) have shown that the 40 - 120 m infrared emission of spirals has two main components: a warm component associated with dusted heated by OB stars in HII regions and young star-forming complexes, and a cooler component from dust in the diffuse, neutral ISM, heated by the general interstellar radiation field of the disk population.

Persson and Helou’s model provides us with two possible explanations for the systematic shifts that are evident between high IRE, H₂ rich galaxies and low IRE, H ₂ poor galaxies in figures 3c through 3h. If the infrared emission of late type spirals primarily comes from the hot component i.e. dust heat by OB stars, then the systematic shifts evident in figures 3c through 3h could simply be explained by a systematically higher extinction within the HII regions of high IRE, H ₂ rich galaxies compared to the HII region of the low IRE, H₂ poor galaxies.

In order to highlight this point, we have drawn a dotted line in figures 3c through 3h to show the effect upon the least squares fit line of de-reddening the H-a equivalent widths of the high IRE, H₂ rich galaxies by 0.7 magnitude. These plots show that it would only take a difference in the mean extinction levels of this magnitude in order to explain most, if not all, of the systematic shift between these two galaxy types.

Interestingly, this hypothesis could also explain why there is a greater split between the two galaxy groups in the plots of
log (L_IR /L _B ) versus log (H-a EQW) (i.e. figures 3c, 3e and 3g) compared to plots of log (SB _IR ) versus log (H-a EQW)
(i.e. figures 3d, 3f and 3h).

Since part of the blue flux of spirals must come from the OB stars found in star forming regions, any systematic difference in the extinction levels in the HII regions of the two galaxy groups, would not only suppress a galaxy’s H-a EQW but it would also increase its L_IR/L_B, while leaving its SB_IR unchanged. Hence, the de-reddening of data points for high IRE, H _2­ rich galaxies in figures 3c, 3e and 3g would shift these points to the right, to correct for the suppression of the galaxy’s H-a EQW, and then down, to correct for the increase in L_IR /L _B . (NOTE: The decrease in L_IR/L_B due to de-reddening would be noticeably less than the increase in H-a EQW because only part of the blue flux (i.e. that from the OB stars) is affected by extinction that is internal to the HII regions).

In contrast, the de-reddening of data points to correct the relatively higher extinction for the HII regions in the high IRE, H_2­ rich galaxies in figures 3d, 3f and 3h would only shift the data points to the right, to correct for the suppression of the galaxy’s H-a EQW. No correction would have to be made to the galaxy’s SB_IR .

Of course, attributing all of the shift between the two galaxy groups to excess extinction with the HII regions of high IRE, H₂ rich galaxies, assumes that the bulk of the infrared emission from late-type spirals comes from their HII regions. Perrsons and Helou (1987) have shown that this is probably only true for the blue galaxies with high current to integrated star formation rates. They predict that the cool infrared cirrus component would be a significant contributor to the overall 40 – 120 m infrared flux in the red galaxies with low current to integrated star formation rate galaxies.

I n figures 3i and 3j , we replot figures 3e and 3f with the sample galaxies segregated by (B – V) _To colour (de Vaucouleurs 1991). This is done so that we can identify which galaxies are redder than the sample mean of 0.52 and which are bluer. It is immediately evident from these plots that all but one of the red spirals are concentrated in the lower left hand corner of the data distribution. Thus, figures 3i and 3j allow us to clearly identify a population of red spirals with low current to integrated star formation rates that, according to Perrsons and Helou (1987), are the most likely to be those that have their 40 – 120 micron infrared fluxes contaminated by cirrus emission.

In summary, there are a number of important conclusions that can be made fromfigures 3c through 3j :

a) The data sample of Young et al. (1989,1996) can be sub-divided into two fundamentally distinct galaxy
groups i.e. high luminosity, high IRE, H₂ rich spirals (hereafter referred to as H₂ rich spirals) and low
luminosity, low IRE, H ₂ poor spirals (hereafter referred to as H₂ poor spirals).

b) There is a good correlation between both L_IR /L _B and SB_IR and H-a EQW for each of these galaxy groups.
This means that L_IR /L _B and SB_IR can be used to determine the (relative) current to integrated star
formation rate for Sbc-Sdm galaxies.

c) The simplest explanation for the systematic shift in the correlations betweenL_IR /L_B and H-a EQW, as
well as SB_IR and H-a EQW, in figures 3c through 3h, is that the HII regions of the H₂ rich spirals are
subject to a systematically higher extinction (of approximately 0.7 magnitudes) compared to the HII
regions of the H₂ poor spirals.

compared to H₂ poor spirals because of the enhanced extinction associated with the HII regions of the
H₂ rich galaxies. The magnitude of this error depends on how much of the blue light of H₂ rich galaxies
comes from star located in regions of recent star formation. Obviously, this error could become significant
for H₂ galaxies with high current to integrated star formation rates because a significant fraction of their
blue light may come from regions of recent star formation.

Figure 4a is a plot of SB_IR versus L_IR/L _B for all the Sbc-Sdm galaxies in the sample of Young et al. 1996 that are not HI gas depleted gas. We can see from this plot that there is a good correlation between the two parameters up to L _IR /L_B = 0.7 with the scatter about this trend increasing for values above L_IR /L _B = 0.7.

Figure 4b is a replot of figure 4a with all galaxies having a logarithmic isophotal axis ratio (i.e. log(R₂₅)) > 0.45 highlighted (de Vaucouleurs et al. 1991). Figures 4b shows that the scatter at high current to integrated star formation rates is cause by the galaxies with log(R₂₅)) > 0.45 i.e. those with inclination angles above 70^o.

It is very likely that the inclination corrections to L_B have been over-estimated for the edge-on spirals, making their L _IR/L_B ratios unreliable as an indicator of their (relative) current to integrated star formation rates. Consequently, these galaxies will be highlighted in all subsequent plots of SB _IR versus L_IR/L_B to emphasis the uncertainty in their L_IR/L_B ratios.