"... In PAGE 8: ... The object traversed many parallel paths, and tra- versed each path 25 times.The average localized po- sitions, and standard deviations for all 25 localized positions are given in Table2 . The average localized position was computed for each path, and these aver- ages agreed to within #060:01 mm.... ..."

"... In PAGE 4: ... The temperature distributions of the ve regions are shown in Figure 1. Some properties of the regions are characterized in Table1 , which gives the zone locations (x; z), height above the midplane, mean density, mean temperature, and the median pressure for the ve regions shown in Figure 1c. Bubbles 1 { 3 are dominated locally by hot gas and so should be most similar to the hot bubble around the Sun.... In PAGE 4: ... Note that Bubble 2 is also farther removed from the midplane (jzj = 230 pc, as opposed to Bubble 1, which is within 10 pc of the midplane, and Bubble 3 with jzj = 60 pc). The median pressures listed in Table1 for each of the ve positions show that the hot bubbles have the highest median pressure, with the cooler gas zones at a median pressure that is half an order of magnitude lower. This is consistent with an observational result (Bowyer et al.... In PAGE 5: ... There can be only one such distribution per location since the simulation has only two independent coordinates. The column density is calculated from the lower left-hand corner of the zone location ( Table1 ) and along a line with a constant 1 cm2 cross-section. We have computed these column densities for all latitudes with 0: 5 separation (720 lines of sight), and thereby generated a simulated strip scan.... In PAGE 5: ... The positional variation of the maxima (column 6 of Table 2) and the high quartile (column 5) model data follow di erent patterns from the variation of the lower statistical quantities. The physical signi cance of this di erence is revealed by comparing column 6 in Table 2 with the height above the midplane, jzj, which is given in column 3 of Table1 . The three positions with jzj = 10 pc all have roughly the same maxima (log NHI 22.... In PAGE 10: ...70 keV band are often larger than those in the 0.22 keV band for each of the three bubbles, particularly as seen from within Bubble 1 (see Figures 1b and 1c), which contains relatively cooler hot gas (T 5 105 K, see Table1... ..."

"... In PAGE 8: ... The object traversed many parallel paths, and tra- versed each path 25 times.The average localized po- sitions, and standard deviations for all 25 localized positions are given in Table2 . The average localized position was computed for each path, and these aver- ages agreed to within 0:01 mm.... ..."

"... In PAGE 5: ... We determine weights for the features with a modified version of the Generative Iterative Scaling algorithm (Curran and Clark, 2003). Templates for local features are similar to the ones employed by Ratnaparkhi (1996) for POS-tagging ( Table3 ), though as our input already includes POS- tags, we can make use of part-of-speech information as well. Long-distance features are simple hand- written regular expressions matching possible sites for EEs (Table 4).... ..."

"... In PAGE 5: ... We determine weights for the features with a modified version of the Generative Iterative Scaling algorithm (Curran and Clark, 2003). Templates for local features are similar to the ones employed by Ratnaparkhi (1996) for POS-tagging ( Table3 ), though as our input already includes POS- tags, we can make use of part-of-speech information as well. Long-distance features are simple hand- written regular expressions matching possible sites for EEs (Table 4).... ..."

"... In PAGE 5: ... We determine weights for the features with a modified version of the Generative Iterative Scaling algorithm (Curran and Clark, 2003). Templates for local features are similar to the ones employed by Ratnaparkhi (1996) for POS-tagging ( Table3 ), though as our input already includes POS- tags, we can make use of part-of-speech information as well. Long-distance features are simple hand- written regular expressions matching possible sites for EEs (Table 4).... ..."

"... In PAGE 5: ... We determine weights for the features with a modified version of the Generative Iterative Scaling algorithm (Curran and Clark, 2003). Templates for local features are similar to the ones employed by Ratnaparkhi (1996) for POS-tagging ( Table3 ), though as our input already includes POS- tags, we can make use of part-of-speech information as well. Long-distance features are simple hand- written regular expressions matching possible sites for EEs (Table 4).... ..."

"... In PAGE 5: ... We determine weights for the features with a modified version of the Generative Iterative Scaling algorithm (Curran and Clark, 2003). Templates for local features are similar to the ones employed by Ratnaparkhi (1996) for POS-tagging ( Table3 ), though as our input already includes POS- tags, we can make use of part-of-speech information as well. Long-distance features are simple hand- written regular expressions matching possible sites for EEs (Table 4).... ..."