The top portion of Fig. 1 (click here) shows our new 232-MHz map of the vicinity of G76.9+1.0, while the bottom portion shows a similar image at 2695 MHz (Fürst et al. 1990a; Reich et al. 1990b), with the sixteen source positions marked. The 232-MHz image is a weighted mean of the five self-calibrated images. The remaining artefacts in the 232-MHz map due to Cygnus A are very evident.
Figure 2 (click here) presents the portion of the 232-MHz map at the position of
G76.9+1.0, again illustrating the limited dynamic range of the new
observations. The dotted ellipse represents the approximate expected half-power
size of the SNR, a circle 4
in diameter convolved with the telescope
response.
Table 4 (click here) lists the approximate positions (to an accuracy of about 15''),
sizes (to an accuracy of about 1'), and identifications for all
sixteen sources studied in this paper. Two of the sources, Nos. 2 and 4,
actually consist of two components at the higher frequencies. The point
companion to CTB 87, E 2013+3702 (Morsi & Reich 1987), is probably
unrelated to the SNR. On the other hand, the two bright components of
the ON 2 source (No. 4) are probably part of the same HII complex.
Table 5 (click here) presents the flux densities used to establish
the spectra for the three comparison sources, as well as the 232-MHz
flux densities deduced from these spectra. The 232-MHz flux densities
for the comparison sources that were derived from the analysis of the
five 232-MHz images were: Source 6, 8.4 
0.6; Source 15, 3.9 
0.4;
Source 16, 5.3 
0.7. Thus flux densities derived directly from the
232-MHz images should be scaled
by a factor of 
. The three comparison sources have
non-thermal spectra, and are probably extragalactic.
Table 6 (click here) lists flux densities obtained for the other thirteen sources, including G76.9+1.0. A comparison of flux densities derived in this study with published flux densities, in cases where such lists exist, e.g. the list of Reich et al. (1990a) for the 1410-MHz data and the list of Fürst et al. (1990b) for the 2695-MHz data, indicated no significant differences, except for a few cases where the choice of background could greatly influence derived flux densities.
There were systematic variations in the derived 408-MHz flux densities from the
five DRAO sets of observations. The systematic scale errors that were
determined in this study, expressed as the ratio of flux density
derived for an individual 408-MHz observation divided by the derived average
flux density, are: Obs. 1, 1.10 
0.06; Obs. 2, 0.95 
0.09; Obs. 3,
0.83 
0.04; Obs. 4, 1.07 
0.05; Obs. 5, 1.12 
0.06. (The observation
numbers refer to Table 2 (click here)). These are, of course, based on only a few sources
distributed in differing sectors of the individual fields of view, and often
at low levels in the primary beam. Therefore they cannot be considered
as good estimates of the reliability of DRAO 408-MHz data in general.
The spectra of these thirteen sources, plus those of the calibration
sources, are shown in Fig. 3 (click here)a. In this figure, the dotted lines connect
points on a best-fitting quadratic function in the log-log plots, and
serve only to highlight spectral trends. They do not correspond to any
fitted physical radio-emission spectrum. The spectrum of G76.9+1.0 is included
in the figure, and is presented in more detail in Fig. 3 (click here)b where the
weighted best-fit linear spectrum for frequencies greater than 1 GHz
is indicated by the solid line. This has a spectral index of 0.64 
0.06,
essentially the same as the value found by Landecker et al. (1993).
It is clear that the 232-MHz flux density lies significantly below the value suggested by the
extrapolation of the spectrum at frequencies above 1 GHz. The 232-MHz
flux density predicted by the higher-frequency fit plotted in Fig. 3 (click here)b is
4.4 
0.6 Jy. The observed value differs from this by over 
.
This fact, and the low flux densities at both 327 MHz and 408 MHz
suggest a spectrum which is flat or turns over below 
1 GHz.
This conclusion is, of course, heavily dependent upon the reliability of
the re-calibrated 327-MHz data. If the 327-MHz data point is ignored,
the best-fitting linear spectrum from 200 MHz to 5 GHz is shown by the
dashed line in Fig. 3 (click here)b. This has a spectral index of 0.48 
0.02. The
232-MHz value is still a factor of two below this fit (still a deviation
of more than 
), and the 1.41-GHz point is high. As was noted
by Landecker et al. (1993), G76.9+1.0 is superimposed on a ridge of thermal
emission which may contribute to deduced flux densities from Gaussian
fitting or surface integration, and the ``somewhat high'' 1.4-GHz flux
densities found in that paper were attributed to this. In the present
work, all the flux densities have been derived in the same fashion, and
one might expect any thermal background contamination to be similar for
all the data points above 1 GHz, except that the 1.41-GHz beam size is
larger than the others.