Appendix B: An account of all X-ray bursts seen from globular cluster bursters

In this appendix we give an account of all X-ray bursts reported previously, as well as those analysed by us. It is organized as follows: the first paragraph of each subsection describes the results from the literature, whereas in the second paragraph we describe our RXTE/PCA and/or BeppoSAX/WFC results (if present). At the end we summarize in Table B.1 the bolometric peak fluxes for all the X-ray bursts; the peak fluxes used in our investigation are indicated with a $\surd$. The RXTE/PCA peak fluxes quoted from the literature are not corrected for the absolute RXTE/PCA calibration (Sect. 3.5). The correction has been applied to our derived peak fluxes from the RXTE/PCA data; in Table B.1 we give between brackets also the uncorrected peak fluxes.

 

 

Table B.1: Observed peak fluxes for the twelve globular cluster X-ray bursters. Peak fluxes from RXTE/PCA observations derived in this paper have been corrected, see Sect. 3.5; the uncorrected values are given between brackets. Also given are the instrument with which the events were seen (BSAX refers to BeppoSAX), whether photospheric radius expansion (RE) was reported or not, and comments. X-ray bursts indicated with a $\surd$ are used to derive the average weighted bolometric peak flux of radius expansion bursts, $\overline {F_{\rm bb,peak}}$, or the maximum observed bolometric peak flux for normal X-ray bursts, $F_{\rm bb,max}$, see Table 2. References to the relevant paper are given in brackets in the sixth column: [1] Cominsky (1981), [2] this paper, [3] Swank et al. (1977), [4] Grindlay et al. (1980), [5] Guainazzi et al. (1998), [6] Molkov et al. (2000), [7] Marshall et al. (1979), [8] Inoue et al. (1981), [9] Pavlinsky et al. (2001), [10] Makishima et al. (1981), [11] in 't Zand et al. (1999), [12] Sztajno et al. (1987), [13] in 't Zand et al. (2002, in preparation), [14] Vacca et al. (1986), [15] van Paradijs & Lewin (1987), [16] Damen et al. (1990), [17] Strohmayer & Brown (2002), [18] in 't Zand et al. (1998), [19] Mukai & Smale (2000), [20] Swank et al. (1976), [21] Cominsky et al. (1977), [22] Hoffman et al. (1980), [23] van Paradijs et al. (1990), [24] Smale (2001).

X-ray burster	peak flux	satellite/	RE?	$\overline {F_{\rm bb,peak}}$	ref	# of bursts + comment
	(10-8 erg cm-2 s-1)	instrument		or $F_{\rm bb,max}$
MX 0513-40	0.9-1.7	SAS-3/HTC	no		[1]	range of 4
	1.26 $\pm$ 0.05 (1.52 $\pm$ 0.06)	RXTE/PCA	no		[2]	1
	1.99 $\pm$ 0.20	BSAX/WFC	yes	$\surd$	[2]	1; strong RE
	2.05 $\pm$ 0.39	BSAX/WFC	yes	$\surd$	[2]	1; weak RE
4U 1722-30	$\simeq$6.2	OSO-8	yes		[3]	1; strong RE
	$\simeq$5.2	Einstein/MPC	yes		[4]	1; weak RE
	$\simeq$5.0	BSAX/MECS	no		[5]	1
	6.23 $\pm$ 0.15	BSAX/WFC	yes	$\surd$	[2]	24; all strong RE
	7.29 $\pm$ 0.06	RXTE/PCA	yes		[6]	1; strong RE
	6.11 $\pm$ 0.10 (7.59 $\pm$ 0.12)	RXTE/PCA	yes	$\surd$	[2]	1 (same as above); strong RE
MXB 1730-335	$\simeq$1.7	SAS-3/HTC	no		[7]	average of many X-ray bursts
	0.9-1.7 (1.1-2.1)	RXTE/PCA	no		[2]	4
	1.66 $\pm$ 0.08 (2.06 $\pm$ 0.10)	RXTE/PCA	no	$\surd$	[2]	1 (of the above 4)
XB 1733-30	7.4 $\pm$ 1.0	Hakucho/FMC	no	$\surd$	[8]	1
	6.3 $\pm$ 0.8	Granat/ART-P	no		[9]	1
XB 1745-25	2.0-6.1	Hakucho/CMC/FMC	no		[10]	14
	6.1 $\pm$ 0.7	Hakucho/FMC	no		[8]	1 (of the above 14)
	4.74 $\pm$ 0.18 (5.89 $\pm$ 0.22)	RXTE/PCA	yes	$\surd$	[2]	1; weak RE
	4.75 $\pm$ 0.17 (5.90 $\pm$ 0.21)	RXTE/PCA	yes	$\surd$	[2]	1; weak RE
MX 1746-20	1.77 $\pm$ 0.14	BSAX/NFI	no	$\surd$	[11]	1
4U 1746-37	0.50 $\pm$ 0.17	SAS-3/HTC	no		[1]	1
	0.75 $\pm$ 0.27	SAS-3/HTC	no		[12]	1 (same as above)
	0.46 $\pm$ 0.23	SAS-3/HTC	no		[12]	1
	1.0 $\pm$ 0.1	EXOSAT/ME	yes	$\surd$	[12]	1; weak RE
	0.9 $\pm$ 0.1	EXOSAT/ME	yes	$\surd$	[12]	1; weak RE
	0.4-0.6 (0.5-0.75)	RXTE/PCA	no		[2]	8
GRS 1747-312	1.71 $\pm$ 0.06	RXTE/PCA	yes	$\surd$	[13]	1; weak RE
4U 1820-30	6.5 $\pm$ 1.3	ANS/HXX	no		[14]	1
	$\simeq$6.0	SAS-3/RMC	no		[14]	22
	$\simeq$5.2	SAS-3/RMC	no		[14]	5 (of the above 22)
	4.2 $\pm$ 0.4	SAS-3/RMC	yes	$\surd$	[14]	6 (of the above 22); all strong RE
	5.28 $\pm$ 0.19	EXOSAT/ME	yes	$\surd$	[15]	7; all strong RE
	$\simeq$4.65	EXOSAT/ME	yes		[16]	7 (same as above); all strong RE
	$\simeq$6.5	RXTE/PCA	yes		[17]	1 (superburst); strong RE
	5.26 $\pm$ 0.12 (6.53 $\pm$ 0.15)	RXTE/PCA	yes	$\surd$	[2]	1; strong RE
	5.9 $\pm$ 0.3	BSAX/WFC	yes	$\surd$	[2]	15; all strong RE
H1825-331	$\simeq$0.8	BSAX/WFC	no		[18]	2
	$\simeq$0.2	ASCA/SIS	no		[19]	1
	2.87 $\pm$ 0.07 (3.57 $\pm$ 0.09)	RXTE/PCA	no	$\surd$	[2]	1
A1850-08	$\simeq$6	OSO-8	no		[20]	1
	6.0 $\pm$ 0.2	SAS-3/HTC	yes		[22]	1; weak RE
	5.2 $\pm$ 0.5	SAS-3/HTC	yes	$\surd$	[1]	1; weak RE (same as above)
4U 2129+12	4.2 $\pm$ 0.1	Ginga/LAC	yes	$\surd$	[23]	1; very strong RE
	$\simeq$5.0	RXTE/PCA	yes		[24]	1; strong RE
	3.81 $\pm$ 0.07 (4.68 $\pm$ 0.09)	RXTE/PCA	yes	$\surd$	[2]	1 (same as above); strong RE
	3.9 $\pm$ 0.8	BSAX/WFC	no		[2]	1

B.1 MX 0513-40/NGC 1851

The first X-ray burst from MX 0513-40 was discovered by Uhuru (Forman & Jones 1976). It had a peak flux of about $0.5 \times 10 ^{-8}$ erg cm^-2 s^-1, but the start of the X-ray burst was not seen. Five X-ray bursts were seen by SAS-3/HTC (see Li & Clark 1977), four of which had bolometric peak fluxes ranging from $0.9 \pm 0.3$ to $1.7 \pm 0.3 \times 10 ^{-8}$ erg cm^-2 s^-1 (Cominsky 1981). More recently, an X-ray burst was seen with the Chandra/HRC-S (Homer et al. 2001). So far no radius expansion bursts had been reported.

The RXTE/PCA observed one short ( $\mbox{$\la$ }$25 s) X-ray burst from MX 0513-40. It did not exhibit photospheric radius expansion. The BeppoSAX/WFCs observed eleven X-ray bursts from this source. The longest one is shown in Fig. 4. This X-ray burst showed a clear strong (factor $\sim$10) photospheric radius expansion phase; its bolometric peak flux of about $2 \times 10 ^{-8}$ erg cm^-2 s^-1 is reached near touch-down (Fig. 5). Another X-ray burst showed some evidence for (weak; factor $\sim$2) photospheric radius expansion. This is the first time radius expansion bursts are reported from this source.

B.2 4U 1722-30/Terzan 2

A long ( $\mbox{$\ga$ }$10 min) X-ray burst was observed from the direction of 4U 1722-30 by OSO-8 (Swank et al. 1977). It reached a peak flux of about $6.2 \times 10 ^{-8}$ erg cm^-2 s^-1. During the first 20 s of the event the black-body temperature was lowest, while the inferred black-body radius was 100 ⁺⁶⁰_-20 km (at 10 kpc). During the remainder of the X-ray burst the inferred black-body radius was $15 \pm 4.5$ km (at 10 kpc). This indicates that it was a (strong) radius expansion burst. Other strong (factor  $\mbox{$\ga$ }$10) radius expansion bursts were seen by Granat/ART-P and RXTE/PCA (Molkov et al. 2000), whereas a weak radius expansion burst (factor $\sim$2) was reported from the Einstein/MPC (Grindlay et al. 1980). The RXTE/PCA event reached a bolometric peak flux of about $7.29 \pm 0.06 \times 10 ^{-8}$ erg cm^-2 s^-1, whereas the event seen with the Einstein/MPC reached a bolometric peak flux of about $5.2 \times 10 ^{-8}$ erg cm^-2 s^-1, both near touch down. Another X-ray burst was reported from BeppoSAX/NFI data; no clear photospheric radius expansion was observed. The peak flux was $\simeq$5. $0 \times 10 ^{-8}$ erg cm^-2 s^-1 (Guainazzi et al. 1998). The BeppoSAX/WFCs observed fourteen strong radius expansion bursts in the first three years of monitoring the Galactic Center region (Cocchi et al. 2000b). Another ten were seen in later observations, see below. The peak intensities of four of the fourteen X-ray bursts were consistent with each other, with a weighted mean of $1.01 \pm 0.03$ Crab, which was extrapolated to a bolometric flux of $3.2 \pm 0.1 \times 10 ^{-8}$ erg cm^-2 s^-1. One of these X-ray bursts was analysed in more detail; from their spectral fits we infer that a bolometric peak flux of about $3.5 \times 10 ^{-8}$ erg cm^-2 s^-1 was reached (note that Cocchi et al. 2001 quote a peak flux of $4.1 \pm 0.4 \times 10 ^{-8}$ erg cm^-2 s^-1).

Our fit parameter values for the X-ray burst from 4U 1722-30 seen by RXTE/PCA (Fig. 3) are in agreement with those reported by Molkov et al. (2000). It is a very strong (factor >10) radius expansion burst; the maximum bolometric flux was reached near touch-down (Fig. 3). The BeppoSAX/WFCs observed a total of 24 X-ray bursts. Most of them have a "precursor'' event in the lightcurves (see e.g. Fig. 4), indicating strong photospheric radius expansion (see Fig. 5). All the X-ray bursts are radius expansion bursts. We found maximum bolometric peak fluxes between $5.4 \pm 0.6$ and $8.4 \pm 1.2 \times 10 ^{-8}$ erg cm^-2 s^-1, with a mean value of $6.23 \pm 0.15 \times 10 ^{-8}$ erg cm^-2 s^-1 (see also Sect. 4.1). For the X-ray burst analysed in detail by Cocchi et al. (2001; see above) we found a bolometric peak flux of $6.2 \pm 0.6 \times 10 ^{-8}$ erg cm^-2 s^-1. We attribute the difference with the lower flux reported by Cocchi et al. (2001) to an improved response matrix (in 't Zand 1999; see also in 't Zand et al. 2001a). Also, our values for the peak fluxes are comparable with those derived for other instruments. We therefore do not quote the earlier results obtained by the BeppoSAX/WFCs in Table B.1.

B.3 MXB 1730-335/Liller 1

MXB 1730-335 (The Rapid Burster) is a source which receives a lot of attention. It is a transient source; in the first weeks of an outburst it shows a phase during which only type I X-ray bursts are seen. The rest of the outburst is dominated by type II X-ray bursts, with occasionally a type I X-ray burst (Guerriero et al. 1999). Most authors focus on the phenomenology of the type II X-ray bursts. No radius expansion burst has so far been reported. The average type I X-ray burst peak flux is about $1.7 \times 10 ^{-8}$ erg cm^-2 s^-1 (SAS-3; Marshall et al. 1979). For a more detailed account of the properties of MXB 1730-335 we refer to Lewin et al. (1993), Guerriero et al. (1999), and Masetti et al. (2000), and references therein.

We selected various (mostly) bright X-ray bursts observed by RXTE/PCA from MXB 1730-335, as reported by Guerriero et al. (1999) and seen in later outbursts. None of them showed photospheric radius expansion. The brightest X-ray bursts show typical bolometric peak fluxes in the range $\simeq$1- $2 \times 10 ^{-8}$ erg cm^-2 s^-1, consistent with values reported previously. The fit results for one of the brightest X-ray bursts are shown in Fig. 2.

B.4 XB 1733-30/Terzan 1

XB 1733-30 has been off for some time now (see e.g. Guainazzi et al. 1999). Only three short ( $\mbox{$\la$ }$25 s) X-ray bursts have been reported in the past. Two X-ray bursts were discovered by the Hakucho/FMC (Makishima et al. 1981). They reached an average maximum peak flux of $7.4 \pm 1.0 \times 10 ^{-8}$ erg cm^-2 s^-1 (Inoue et al. 1981). Another X-ray burst was observed by Granat/ART-P; 3-5 s into the X-ray burst a bolometric peak flux of $6.3 \pm 0.8 \times 10 ^{-8}$ erg cm^-2 s^-1 was found (Pavlinsky et al. 2001). No photospheric radius expansion phase was reported.

B.5 XB 1745-25/Terzan 5

Fourteen X-ray bursts from the transient XB 1745-25 were discovered by Hakucho (Makishima et al. 1981). The maximum peak flux was $6.1 \pm 0.7 \times 10 ^{-8}$ erg cm^-2 s^-1 (Inoue et al. 1981), but peak fluxes down to about $2.0 \times 10 ^{-8}$ erg cm^-2 s^-1 were reported (Makishima et al. 1981). During a recent outburst fifteen X-ray bursts were seen by the RXTE/PCA (Markwardt et al. 2000). No photospheric radius expansion and only modest cooling were reported.

We find that two of the X-ray bursts seen by the RXTE/PCA were much brighter than the other thirteen. Our analysis shows (for the first time) that they both have a weak (factor of $\sim$2) photospheric radius expansion phase and clear cooling during the decay (see Fig. 2). Maximum bolometric fluxes ($\simeq$4. $75 \times 10 ^{-8}$ erg cm^-2 s^-1) are reached near touch-down. The weaker X-ray bursts did not show evidence for photospheric radius expansion.

B.6 MX 1746-20/NGC 6440

Observations with the BeppoSAX/WFC and NFI of the transient MX 1746-20 during its second ever recorded outburst revealed four X-ray bursts (in 't Zand et al. 1999). The bolometric peak flux during the first 12 s interval of the $\sim$100 s long X-ray burst seen with the NFI was $1.77 \pm 0.14 \times 10 ^{-8}$ erg cm^-2 s^-1. No evidence for photospheric radius expansion was seen.

B.7 4U 1746-37/NGC 6441

Various X-ray bursts have been seen from the location of 4U 1746-37, i.e. by SAS-3/HTC (Li & Clark 1977; Cominsky 1981), EXOSAT/ME (Sztajno et al. 1987), Ginga/LAC (Sansom et al. 1993), BeppoSAX/NFI (Parmar et al. 1999) and RXTE/PCA (Jonker et al. 2000). One of the two X-ray bursts seen with the SAS-3/HTC had a peak bolometric luminosity of $0.50 \pm 0.17 \times 10 ^{-8}$ erg cm^-2 s^-1(Cominksy 1981). The two X-ray bursts seen with EXOSAT/ME both showed weak (factor $\sim$3) photospheric radius expansion. Their peak fluxes were $1.0 \pm 0.1$ and $0.9 \pm 0.1 \times 10 ^{-8}$ erg cm^-2 s^-1(Sztajno et al. 1987). Sztajno et al. (1987) also performed a re-analysis of the two SAS-3/HTC X-ray bursts. They found $0.75 \pm 0.26$ and $0.46 \pm 0.23 \times 10 ^{-8}$ erg cm^-2 s^-1, respectively.

Eight of the twelve X-ray bursts reported by Jonker et al. (2000) were bright. Our analysis shows that none of the X-ray bursts showed photospheric radius expansion. An example of the fit parameter results of one of the eight bright X-ray bursts is given in Fig. 2. The eight bright X-ray bursts reached bolometric peak fluxes in the range $\simeq$0.4- $0.6 \times 10 ^{-8}$ erg cm^-2 s^-1.

B.8 GRS 1747-312/Terzan 6

X-ray bursts from the transient source GRS 1747-312 were only recently found during an outburst (in 't Zand et al. 2002, in preparation). One of these showed photospheric radius expansion of a factor of $\sim$5 with a bolometric peak flux of $2.10 \pm 0.07 \times 10 ^{-8}$ erg cm^-2 s^-1.

B.9 4U 1820-30/NGC 6624

4U 1820-30 was the first source seen to exhibit X-ray bursts (Grindlay et al. 1976). One of these two X-ray bursts seen by the ANS/HXX had a peak flux of about $6.5 \times 10 ^{-8}$ erg cm^-2 s^-1 (1-15 keV; re-analysis by Grindlay, private communication in Vacca et al. 1986). The 22 X-ray bursts seen by SAS-3/RMC (Clark et al. 1977) had a mean peak flux of $6.0 \times 10 ^{-8}$ erg cm^-2 s^-1 (re-analysis by G.W. Clark, private communication in Vacca et al. 1986). A subset (five) of these X-ray bursts (Clark et al. 1976) showed a mean peak flux of $5.2 \times 10 ^{-8}$ erg cm^-2 s^-1 (re-analysis by G.W. Clark, private communication in Vacca et al. 1986). Another subset (six) of the SAS-3/RMC X-ray bursts were analysed in more detail by Vacca et al. (1986). They concluded that the bolometric peak fluxes, which ranged from $3.9 \pm 0.2$ to $5.3 \pm 0.4 \times 10 ^{-8}$ erg cm^-2 s^-1, were the same within the errors. The weighted mean peak value was $4.2 \times 10 ^{-8}$ erg cm^-2 s^-1, with an estimated uncertainty of about 10%, allowing for systematic errors. All six X-ray bursts showed photospheric radius expansion. EXOSAT/ME saw seven radius expansion bursts (Haberl et al. 1987). Individual flux measurements (0.25 s or 0.5 s resolution) during the photospheric radius expansion phase ($\sim$3 s duration) ranged from 3 to $8 \times 10 ^{-8}$ erg cm^-2 s^-1, whereas the average peak luminosity during this first 3 s ranged from 4.1 to $5.7 \times 10 ^{-8}$ erg cm^-2 s^-1, with a mean of $5.1 \times 10 ^{-8}$ erg cm^-2 s^-1. (Note that van Paradijs & Lewin (1987) used the results presented by Haberl et al. (1987) to derive a maximum flux during the photospheric radius expansion phase of $5.28 \pm 0.19 \times 10 ^{-8}$ erg cm^-2 s^-1.) A re-analysis of the EXOSAT/ME X-ray bursts showed that the bolometric flux at touch down ranges between 4.3 and $5.0 \times 10 ^{-8}$ erg cm^-2 s^-1, with a mean of $4.65 \times 10 ^{-8}$ erg cm^-2 s^-1(Damen et al. 1990). More recently, the RXTE/PCA observed a short ($\sim$25 s) X-ray burst (Zhang et al. 1998), which was also a radius expansion burst (Franco & Strohmayer 1999; see also Strohmayer & Brown 2002). Another very long (>2.5 hr) X-ray burst (so-called superburst), also seen with RXTE/PCA, showed strong photospheric radius expansion and reached a bolometric peak flux of $6.5 \times 10 ^{-8}$ erg cm^-2 s^-1 (Strohmayer & Brown 2002).

Our analysis of the normal radius expansion burst seen by RXTE/PCA showed indeed strong photospheric radius expansion, by a factor of $\sim$20, see Fig. 2. Note the strong similarity with the X-ray burst seen from XB 1745-25, except for the brief stronger expansion in 4U 1820-30 (Fig. 2). The BeppoSAX/WFCs have seen many X-ray X-ray bursts from 4U 1820-30 (see. e.g. Fig. 4). From our analysis of 21 X-ray X-ray bursts we identified 15 radius expansion bursts (see e.g. Fig. 5). The maximum peak fluxes of all the radius expansion bursts do reach the same value of $5.9 \pm 0.3 \times 10 ^{-8}$ erg cm^-2 s^-1with a $\chi ^2_{\rm red}$ of 0.82 for 14 d.o.f.

B.10 H1825-331/NGC 6652

Two type I X-ray bursts were first observed by the BeppoSAX/WFC from the transient H1825-331 (in 't Zand et al. 1998). A bolometric peak flux of about $0.8 \times 10 ^{-8}$ erg cm^-2 s^-1was found over the first 16 s of each X-ray burst, which had exponential decay times of about 16 s and 27 s (2-8 keV). During a serendipitous ASCA observation Mukai & Smale (2000) found another X-ray burst. It reached a peak flux of about $0.2 \times 10 ^{-8}$ erg cm^-2 s^-1during the first 30 s of the X-ray burst. No photospheric radius expansion was seen in both cases.

The RXTE/PCA observed a bright and a weak X-ray burst from H1825-331. No clear evidence for photospheric radius expansion is found, but we note that near maximum of the bright X-ray burst the black-body temperature shows a short ($\sim$1 s) and small dip, while the apparent black-body radius shows a small hump (Fig. 3). This might indicate that the event reached the Eddington limit, but that it was not strong enough to cause the photosphere to expand appreciably.

B.11 A1850-08/NGC 6712

An X-ray burst with a peak intensity of about 1.5 Crab was observed from the vicinity of A1850-08 with OSO-8 (Swank et al. 1976). We note here that the peak intensity of the long X-ray burst from 4U 1722-30 observed by OSO-8 (see Sect. B.2) also reached 1.5 Crab (Swank et al. 1977). Assuming the X-ray burst spectra are similar during the peak (which may not necessarily be the case since the event from 4U 1722-30 showed photospheric radius expansion), one may infer a peak flux of about $6 \times 10 ^{-8}$ erg cm^-2 s^-1. An X-ray flare was reported by Cominsky et al. (1977) using data from Uhuru. It increased by a factor of 40 in less than 5.5 min. The next 16.7 min the intensity decreased with an exponential decay time of $3.1 \pm 0.4$ min. It may have persisted for up to 78 min. The observed peak flux was about $0.6 \times 10 ^{-8}$ erg cm^-2 s^-1. Three X-ray bursts were seen with SAS-3/HTC, of which one showed a weak (factor $\sim$2) photospheric radius expansion (Hoffman et al. 1980; see also Cominsky 1981). Hoffman et al. (1980) found a bolometric peak flux of $6.0 \pm 0.2 \times 10 ^{-8}$ erg cm^-2 s^-1, whereas Cominsky (1981) report $5.2 \pm 0.5 \times 10 ^{-8}$ erg cm^-2 s^-1 for the same X-ray burst. The bolometric peak flux we use in our investigation, as indicated with a $\surd$ in Table B, is the value derived by Cominsky (1981). The fluxes quoted by the Hoffman et al. (1980) are calculated from conversions from count rates to photon rates using a Crab-like spectrum, whereas those quoted by the former derive conversions using black-body spectra (see Cominsky 1981 for a more detailed discussion).

B.12 4U 2129+12/NGC 7078

A very strong (factor $\mbox{$\ga$ }$100) radius expansion burst was observed by Ginga/LAC from 4U 2129+12 (Dotani et al. 1990; van Paradijs et al. 1990). The bolometric peak flux was about $4.2 \times 10 ^{-8}$ erg cm^-2 s^-1, with a conservative error of $0.1 \times 10 ^{-8}$ erg cm^-2 s^-1. A less strong (factor $\sim$10) radius expansion burst was recently seen by RXTE/PCA, which reached a bolometric peak flux of $5 \times 10 ^{-8}$ erg cm^-2 s^-1 (Smale 2001). Many more X-ray bursts were identified in the RXTE/ASM data (Charles et al. 2002). A Chandra ACIS-S/HETG observation revealed that 4U 2129+12 is in fact composed of two bright sources, separated by only 2 $.\!\!^{\prime\prime}$7 (White & Angelini 2001). One source, NGC 7078 X-1, is associated with AC 211, the other is identified as NGC 7078 X-2. NGC 7078 X-2 is the most likely source for the X-ray bursts.

Our reanalysis of the event seen by RXTE/PCA showed similar results (Fig. 3) as those reported by Smale (2001). We found a bolometric peak flux of $\sim$ $4 \times 10 ^{-8}$ erg cm^-2 s^-1 near touch-down. The BeppoSAX/WFC also observed one X-ray burst (Figs. 4 and 5). However, no evidence for photospheric radius expansion was found, although it reached a comparable bolometric peak flux ($\sim$ $4 \times 10 ^{-8}$ erg cm^-2 s^-1) as seen during the strong radius expansion burst observed by Ginga/LAC and RXTE/PCA. Note that our method of subtracting the persistent pre-burst emission automatically also removes any contribution of AC 211/NGC 7078 X-1.