4 Discussion

Throughout NGC1569 we find 4 to 5 RSNe and/or SNRs, located within an area of $\sim$300pc diameter around the clusters A, B, C where active star formation occurred until recently, and might still be going on. We discuss separately the region near cluster C and the associated molecular cloud, and the SSCs A, B and the intermediate-size clusters and their surroundings.

4.1 Non-thermal sources near star cluster C and the associated CO cloud

The cluster C=No.10 is associated with a bright H II region (No.2, Waller 1991) located at the edge of a CO cloud complex of $\sim$180pc extension and $1.4\times10^{6}\,M_{\odot}$ total mass (Taylor et al. 1999, their Figs. 8 and 9). The sources M-1,2(b,c,d), VLA-10 and VLA-11; the H II regions 1,2,3 (Waller 1991); and the CO clouds 1,2,3 (Taylor et al. 1999) form a structure which resembles the 30 Doradus (R136) and N160 and N159 region of the LMC (Cohen et al. 1988; Johansson et al. 1998; Bolatto et al. 2000), although less massive. Similar to the LMC, the dominant CO clouds 1,2,3 (Taylor et al. 1999) are located at the Western end of NGC1569's stellar bar (Waller & Dracobly 1993) and extend from there in perpendicular direction to the major axis of the bar. Star formation has started at the Northern edge of the CO cloud, producing the cluster No.10 (Hunter et al. 2000) and intermediate-size clusters. The structure around cluster C is difficult to assess; there are several tentative sources M-1,b,cd, but for VLA observations the region contains too much extended emission and confusion to allow a clear distinction and identification of sources. The source M-1 is clearly a thermal source.

Star formation is probably still progressing towards the South into the CO cloud. There are two non-thermal sources, M-2 and VLA-11, at the edge and $\sim$20-30pc outside the CO cloud complex 1,2,3 (Taylor et al. 1999), respectively. If we interpret these sources as SNRs, or RSNe, we may conclude that at these positions half-way along and at the edge of the CO cloud, some star formation has taken place, or is currently taking place, although not (yet) very efficiently. This holds in particular for the source M-2 since its surrounding contains the young, intermediate-size star cluster No.5 of Hunter et al. (2000).

There are two non-thermal sources to the North of cluster No.10. The source VLA-16, at $\sim$25pc North, is probably a SNR, and possibly associated with the region of star cluster No.10 (C) or its immediate surrounding (Table 4). The same holds for the source M-3 which is associated with the cluster No.8 of Hunter et al. (2000; Table 4). In this area lies also the tentative source M-a.

4.2 The environment of the SSCs A and B

The SSC-A consists of two components (O'Connell et al. 1995; de Marchi et al. 1997; Hunter et al. 2000; see Table 4) and shows evidence of WR stars (Gonzalez-Delgado et al. 1997; Buckalew et al. 2000). Under the assumption of being a single object, Ho & Filippenko (1996) derived a lower limit of the (dynamical) mass M(A) = (3.3$\pm$0.5)$\times$10 $^{5}~M_{\odot}$, based on the measured stellar velocity dispersion $\sigma_{*} = (15.7\pm1.5)$kms^-1 and the half-light radius of the cluster r(A) = 1.9$\pm$0.2pc (0.18''). Using the fact that A is double, de Marchi et al. (1997) and Sternberg (1998) obtain, for A1, a mass between 2.8$\times$10 $^{5}~M_{\odot}$ and 1.1$\times$10⁶$M_{\odot}$. The SSC(s)-A is located in a cavity, visible in 21cm-H I and X-ray emission as a $\sim$200pc diameter hole (Israel & van Driel 1990; Heckman et al. 1995), The cavity is assumed to be blown out by SN explosions and stellar winds. The SSC-B is located in some diffuse interstellar material.

In Fig. 1, the circles around the clusters delineate areas of $\sim$50pc diameter. They represent, approximately, the distance a star escaping from a cluster with the velocity dispersion $\sigma_{*}$ $\approx$ 20kms^-1 (Ho & Filippenko 1996) can traverse within $\sim$1-2Myr. The encircled fields delineate approximately the areas onto which a search for RSNe and SNRs, possibly associated with the SSCs, should concentrate. The areas should not be significantly larger since the SSCs (A) contain a large number of low mass stars and thus are likely to be gravitationally bound (Ho & Filippenko 1996; Sternberg 1998; Smith & Gallagher 2001) so that the probability that massive stars escape to larger distances is small, although the systems may not yet be fully relaxed.

 

 

Table 4: Positions of star clusters referred to in this study (Hunter et al. 2000; A1, A2: de Marchi et al. 1997).

Cluster	RA(2000)	Dec(2000)	MVa	Radiusa	Non-thermal	Distance of RSN, SNR
	[ $^{\rm h~ m~ s}$]	[ $\hbox{$^\circ$ }$ ' '']	[mag]	['' $\leftrightarrow$ pc]	Source	to Cluster [pc]
SSC- A	4 30 48.19	64 50 58.6	-14.1	1.14-12	[VLA-8: SNR	$\sim$30]
SSC- A1b			-13.6	$\sim$0.15-1.6
SSC- A2b			-12.3	$\sim$0.17-1.8
SSC- B	4 30 48.99	64 50 52.7	-13.1	1.34-14
C/No.10	4 30 47.26	64 51 02.3	-11.9	0.71-7.6	VLA-16: SNR	$\sim$25
No.5	4 30 46.67	64 50 54.4	-8.6	0.46-4.9	M-2: RSNc	15
No.6	4 30 46.89	64 51 00.6	-9.7	0.34-3.6	M-1: therm. source
No.7	4 30 46.96	64 50 59.4	-9.2	0.34-3.6	M-1: therm. source
No.8	4 30 47.04	64 51 06.6	-8.6	0.23-2.4	M-3: RSNc	2
No.18	4 30 48.07	64 50 57.3	-7.8	0.18-1.9	VLA-8: SNR	10
No.45	4 30 54.53	64 50 43.2	-6.9	0.50-6.0	M-6: SNR	35
a For a distance of 2.2Mpc.
b Components of SSC- A, separated by 0.18'' [2.2pc]; A1 is located to the $\sim$South-East of A2, Hunter et al. (2000).
c Or a small SNR.

The observations do not reveal a RSN or SNR in the immediate surrounding of the SSC A and B, except for the source VLA-8 assumed to be a SNR of $\sim$20pc diameter.

Besides the statistical argument brought forward to explain the absence of short-lived RSNe and RSNs in and near the SSCs, we believe that there exists also a valid kinematical argument for their absence. When extrapolating to SSCs Canto et al.'s (2000) calculation of the action of stellar winds of many massive stars in a cluster, combined with the action of several SN explosions, and when considering the influence of the cluster gravitational field on the propagation of the SN blast in a similar way as done for proto globular clusters (Shustov & Wiebe 2000), a violent and turbulent outflow of hot material is expected to occur which leaves little room for a quiescent development of SNRs. Using the radius r(A) and mass M(A) of the SSC-A mentioned above, the stellar mass concentration $\rho$ and the average distance <$\delta$> between the cluster stars is

$$\rho \approx {M({\bf A}) \over {[4/3\pi r({\bf A})^3]}} \appr... ...0~ {\rm pc}^3}} \approx 1 \times 10^4 ~M_\odot ~{\rm pc}^{-3},$$ (1)

and

$$\langle \delta \rangle \propto (1/\rho)^{1/3} \propto 0.05 {-} 0.1 ~{\rm pc},$$ (2)

if we assume that the average mass of the cluster stars is $\sim$1-3$M_{\odot}$ (Sternberg 1998). Stellar winds and material ejected in SN explosions extend to similar distances which makes a strong interaction of cluster-internal gas plausible. An example of this process is SN1993J in M81 which has a shell diameter of $\sim$0.1pc some 1300days after explosion and which expands with a velocity of $\sim$15000kms^-1 (Marcaide et al. 1997). The individual stellar winds and SN ejected hot material blow out of the SSC as a common wind, which diffuses through the interstellar medium. In the immediate surroundings of the SSCs the outflowing stellar winds and SN ejected material shock with the interstellar medium, forming shells and holes. The shock wave around SSC-A has probably created the H I hole and the shell seen in the observations of Israel & van Driel (1990) and Greve et al. (1996). In this picture it is not surprising that the non-thermal source (VLA-8), which may be an extended SNR, is located $\sim$30pc outside the cluster core where the interstellar matter is in less turbulent motion. This picture of the diffusion of SN ejected gas agrees with the fact that locally metal-enriched gas has not been found (Kobulnicky & Skillman 1997).

We do not find RSNe or SNRs in or near the many other star clusters (Hunter et al. 2000) and WR sources (Buckalew et al. 2000), respectively.

4.3 A final remark

The phenomenon of SSCs and SNRs is much more spectacular in the starburst galaxy M 82 than in the irregular galaxy NGC1569; however, the heavy obscuration of M 82 prevents us from obtaining a complete view of the relation between SSCs and SNRs. The post-starburst galaxy NGC1569 with locally recent star formation similar to that found in M 82, allows on the other hand an unobscured view of many young stars in the body of the galaxy, of the large number of intermediate-size star clusters, and - as an exception to other unobscured closeby irregular galaxies - of two SSCs A and B. The VLA detects the non-thermal sources VLA-8 and VLA-16 at $\sim$25pc distance from the SSC A and cluster C, respectively. We did not detect these sources with MERLIN. We interpret these sources as low surface brightness SNRs. Because of their distance of $\sim$25pc from the clusters we hesitate to attribute their origin to stars originally belonging to these clusters. The regions closer to the SSCs and closer to the intermediate-size clusters are devoid of RSNe and SNRs. Although a comparison between M 82 and NGC1569 on statistical arguments seems to provide a plausible explanation for the absence of SNRs near and in the clusters, some caution in the use of this argument is appropriate in view of the locally large number of stars (10⁵ to 10⁶) very recently formed in the SSCs and the intermediate-size clusters. The absence of SNRs in and very close to the clusters, in both M 82 and NGC1569, may - at least partially - be due to the hostile environment. Unfortunately, within $\sim$5Mpc distance there are no other unobscured galaxies containing many SSCs, allowing a similar investigation with MERLIN.

Acknowledgements

We thank the MERLIN staff, Jodrell Bank, for the observations, the help in data reduction, and the pleasant hospitality. We thank the referee for putting the astrophysical question into the correct context of star formation and supernova explosions, and for eliminating contradictions.