Single-Epoch Measurements of Broadband Radio Continuum Spectra A Cast of Summer Students1 Jo Ann Eder Tapasi Ghosh, Chris Salter. NAIC/AO, P. O. Box 995, Arecibo, Puerto Rico 00613. April 13, 1999 Abstract The ability of the upgraded Arecibo 305-m telescope to produce \quasiinstantaneous" radio continuum spectra covering over a decade of frequency has been investigated, the study being undertaken as an Arecibo Observatory2 summer-student observing project. Within the limits of early post-upgrade instrumentation and telescope performance, it was found to be relatively easy for inexperienced observers to obtain the measurements needed to achieve the above objective. Good-quality spectra were produced for three quasars (J1609+266, J2115+295 and J2203+317) which exhibited mutually dierent spectral shapes. The planetary nebula, G064.7+05.0, was also included in the target list. This is shown to be optically thick at 1.4 GHz, while only an upper limit to its ux density could be determined at 430 MHz.

1 Introduction Radio continuum spectra are usually \synthesised" from already available ux density measurements made at many dierent epochs. Much more of a rarity are spectra compiled from ux densities measured quasi-simultaneously at a number of frequencies. However, such spectra have been obtained using the VLA (i.e. for extragalactic radio supernovae, stellar variables, some AGNs, etc.), and by University of Michigan (UMRAO) observers for a sample of extragalactic variable Angel Alejandro Qui~nones (UPR-Humacao), Monique Aller (Wellesley), Yira Cordero Lebron (UPR-Humacao), Ingrid Daubar (Cornell), Simon DeDeo (Harvard), David Kaplan (Cornell), Dale Kocevski (U. of Michigan), Felix Mercado Cortes (U. Metropolitana), Benjamin D. Oppenheimer (Harvard), Celia Salmeron (U. of Houston) 2 The Arecibo Observatory is part of the National Astronomy and Ionosphere Center, which is operated by Cornell Univ. under a cooperative agreement with the National Science Foundation. 1

1

sources between 5 and 15 GHz. The upgraded Arecibo telescope, with its ability to switch between frequencies in a matter of seconds, would seem potentially well suited to such work, providing both speed and \photometric" performance. To investigate the present quality of such Arecibo measurements, and also to serve as \hands-on" observing experience, the 1998 Arecibo summer students made measurements of a number of continuum sources at four dierent frequencies spanning over a decade in frequency on two nights in July 1998. Sources of diering spectral types were chosen as targets, while two steep-spectrum ( > 0:5; S/  ,) objects of moderate strength served as ux-density calibrators. The summer students were split into two groups of ve, with no overlap between the members present on each of the nights. The observations were overseen by Jo Ann Eder and Chris Salter. The students themselves reduced much of the data, their measured values being checked and re ned by Chris Salter and Tapasi Ghosh, who also planned the measurements.

2 The Sample Pending realignment work scheduled for late 1998, the Arecibo telescope is known to have greater, less symmetric, gain variations with azimuth and zenith angle than speci ed in its upgrade design goals. Hence, we selected sources within a relatively small declination range (29 < Dec < 33 ), separated in Right Ascension by about an hour each. This was to permit observing within a rather limited range of azimuth and zenith angle, without too much detailed scheduling being demanded of the student observers. The selected steep-spectrum calibrator sources, J1831+291 and J0040+331, have spectral indices of about 0.6 and 0.8 respectively (Kuhr et al. 1981). In the NVSS survey made with the VLA in its D-array con guration (Condon et al. 1998), both are listed as being smaller than 15 arcsec, meaning that they are essentially point sources to Arecibo, even at the highest frequency presently available of 5 GHz. The \target" sources were as follows;

 Two at, or inverted, spectrum quasars, J2115+295 and J2203+317, where

the rst source is a few times weaker than the latter.  The planetary nebula, G064.7+05.0 (J1934+305). This source was selected from the catalog of planetary nebulae from NVSS (Condon & Kaplan 1998). Together, Condon & Kaplan and the full NVSS catalog show the source to be smaller than 15 arcsec, relatively unconfused, and of about 0.25-Jy ux density at L-band.  On the rst observing night, the observations began suciently early that a test source, J1609+266, was used to check out the observing procedures 2

Table 1: Target Sources

Source Planetary Neb G064.7+05.0 AGNs J1609+266 J2115+295 J2203+317 Calibrator J1831+291 J0040+331

RA (J2000)

Dec S(1.4 GHz) Other Name (J2000) (mJy)

19 34 45.2 30 30 59

245

16 09 13.3 26 41 29 21 15 29.4 29 33 38 22 03 15.0 31 45 38

4775 975 2879

CTD93

18 31 14.9 29 07 10 00 40 55.1 33 10 08

2924 3224

4C29.56 3C19

4C31.63

at 430, 1405 and 2380 MHz before moving on to the rst program source (J1831+291). J1609+266 is a GHz-peaked spectrum quasar which shows \compact-double" structure in VLBI images (though see Shaer & Kellermann 1998). As these test measurements were recorded in the data le, they too were reduced as per the other targets, despite the declination of J1609+266 being outside the 29 { 33 range. A full source list is presented in Table 1.

3 The Observations The observations were made on the nights of July 9 { 10 and 14 { 15, 1998. A late start on July 14 { 15, plus nding that from an earlier S-band radar run the protective cover had been left over the 6-cm horn, necessitating a trip to the feed turret, meant that the calibrator source, J1831+291, was only observed at 1405 MHz on that occasion. The students cabled up the square-law detectors, the oscilloscope display and the radar interface (continuum data logger) as per instructions (Perillat, private communication). They also checked that the recorded values were relative to a true zero level. At L-Band, the \L-Narrow" receiver was used. All receivers responded to circular polarization, except the C-band receiver which was set up for linear polarization. Data-taking with the radar interface was initiated, and the observations on each source were cycled through L-band, C-band, 430 MHz and S-band in that order. The students made the nal scheduling of the sources such as to minimize the range of azimuth and zenith angle involved (see Section 2 above). Once the telescope was pointed to a given target, the operations performed at each frequency were; 3

1. the feed turret was positioned for the appropriate receiver. 2. The correct IF/LO con guration was set up. The chosen center frequencies in observing sequence were 1405, 5000, 430 and 2380 MHz. A pair of 5MHz I.F. lters centered at 260 MHz were used for all frequencies except C-band, where a pair of \quasi-50 MHz" lters were employed. This use of wider bandwidth at C-band was useful given the lower telescope eciency there pending upcoming dome alignment work. 3. The I.F. attenuators were adjusted to give an appropriate signal level into the radar interface to avoid saturation. 4. A pair of orthogonal 2-minute cross scans were made in \great-circle" coordinates, rst in \cross zenith angle", then in zenith angle (using the telescope drive procedure \crossgc"). The lengths of the arms of the crosses were 20, 6, 60 and 12 arcmin respectively for the frequency order given above. These represent about 5{6  HPBW at each frequency. 5. After the cross scans were completed, the telescope was stopped. When the pointing had moved o the source, the noise source was red for 30 sec. The injection of these steps of time-invariant signal power permitted highly accurate intercalibration of the dierent measurements at each frequency. 6. In real time, the students monitored all stages of the observations from the Observing Room using the Analyz reduction package. 7. Full notes were taken, this task being rotated between the individual students. The complete data taking was run by the students, each taking their turn to act as team coordinator for the above operations. A number of instrumental problems were encountered during the runs. All were subsequently xed by the electronics division. On the night of July 9 { 10, the polarization-B channel of the S-Band system dropped out intermittently, and data for J1831+291 and G064.7+05.0 were only acquired with a single circular polarization. On the night of July 14 { 15, only the polarization-A channel of the C-Band receiver received the noise diode signal; calibration for the other channel was made assuming a constant value of Tsys (see below). At 430 MHz, only a single receiver channel was available on both nights.

4 Data Reduction Data reduction was made using the Analyz analysis package. For each cross scan, the data was baselined by subtracting a straight line tted to selected minima on either side of the transit. A Gaussian was then tted to the central section of 4

Table 2: Azimuth and Zenith Angle Ranges

Freq (MHz) 430 1405 2380 5000


Az (Jul 9) 163 { 201 192 { 225 151 { 173 181 { 217


ZA (Jul 9)
Az (Jul 14) 11.6 { 14.8 143 { 167 13.7 { 16.6 156 { 190 11.0 { 15.6 132 { 158 12.7 { 15.3 149 { 177


ZA (Jul 14) 14.4 { 17.2 11.3 { 15.1 16.2 { 19.2 12.7 { 16.0

the main beam response. Next, after baselining the appropriate data section, the mean and rms of the calibration de ection was computed. During the analysis, the following information was tabulated; 1. 2. 3. 4. 5. 6.

Azimuth and elevation of scan. The o-source intensity due to the total \system temperature". The tted Gaussian height. The time of transit (yielding the pointing oset). The HPBW. the noise-source intensity.

The spread of azimuth and zenith angle for the calibrators and main targets, i.e. excluding J1609+266, are given in Table 2. A number of parameters related to system performance are detailed in Tables 3 { 6. The mean HPBW's on the two observing days are given in Cols. 2 and 3 of Table 3. Col. 4 contains the ratio of the main beam area (1.133 HPBWAZ  HPBWZA) at 1405 MHz to that at the other frequencies as normalized by the square of the ratio of frequencies. The early post-upgrade deterioration in performance with increasing frequency is clearly seen at the two higher frequencies. This is also illustrated by the maximum observed sidelobe level as exempli ed by observations of J0040+331 on July 10 (Col. 5). As the frequency increases, the principal sidelobe increasingly has the form of a \coma lobe". The mean pointing osets, and the rms pointing about these, at the dierent frequencies are given in Table 4. Although the range of azimuth/zenith angle covered was rather small, the pointing at L-band and above was found to be excellent. The values at 430 MHz re ect the situation that little pointing eort has yet been expended at this frequency. The System Equivalent Flux Densities (SEFDs) adjacent to the calibrator sources are tabulated in Table 5. The ux densities used for J1831+291 and J0040+331 were computed from the formulae given for these sources by Kuhr et al. (1981), 5

Table 3: Telescope Beam Parameters

Freq (MHz) 430 1405 2380 5000

HPBW (Jul 9) HPBW (Jul 14) Beam Area Max. sidelobe (AZ  ZA) (AZ  ZA) (Normalized) (dB) 00 00 00 00 650  732 624  741 1.029 13.2 (ZA=14.9) 20200  22700 19900  22400 { 12.4 (ZA=16.8) 00 00 00 00 125  144 124  148 0.865 10.9 (ZA=14.9) 63.700  76.500 63.000  72.300 0.758 9.2 (ZA=15.4) Table 4: Pointing Parameters for each Frequency

Freq (MHz) Az Oset
Azrms ZA Oset
ZArms 430 +96.100 11.800 +7.600 22.000 00 00 00 1405 +5.5 4.1 +0.6 2.600 2380 +3.100 6.200 +0.000 2.200 00 00 00 5000 {4.6 5.7 {1.2 3.600 and are on the scale of Baars et al. (1977). It should be remembered that at 430 MHz, J1831+291 and J0040+331 are seen against celestial backgrounds of Tb  34:5 and 22 K respectively, these values being extrapolated from the 408MHz survey of Haslam et al. (1982). This is expected to add about 3.8 and 2.4 Jy to the respective SEFDs. We also note that the 430-MHz receiver was uncooled on July 14 { 15, (and swapped to the other polarization channel!) In Table 6, Cols. 3 and 5, we present the mean value of Tsys in units of the noise source intensity, with the rms's of all individual measured values in Cols. 4 and 6. The Tsys values at 430 MHz correlate well with the celestial background temperatures (Fig. 1), explaining the much higher rms's at this frequency. It is of interest that the best- t straight line to Fig. 1 gives an extrapolated Tsys of 44 K for zero sky brightness. This yields a 430-MHz point-source response for the telescope of 9.0 K/Jy, close the value expected at this frequency.

5 Results The directly measured Gaussian ts for the sources had their peak de ections corrected for the pointing errors estimated from the orthogonal scan of each pair. Next, they were normalized by the noise-source de ections, and then converted into Jy using the ux densities for the two calibrators from Kuhr et al. (1981). The derived ux densities for the target sources are given in Table 7. In this table, the 1405-MHz ux densities are the average of both observing days, while only the results from July 9 { 10 are given for the other frequencies, these representing 6

Table 5: System Equivalent Flux Density at each Frequency

Freq Date Source Az El Tsys(Ch A) Tsys(Ch B) (MHz) (deg) (deg) (Jy) (Jy) 430 Jul 9 J1831+291 201 11.7 { 8.77 Jul 9 J0040+331 183 14.8 { 7.78 Jul 14 J0040+331 167 15.2 10.49 { 1405 Jul 9 J1831+291 227 15.5 3.97 3.65 Jul 9 J0040+331 206 16.6 4.19 3.82 Jul 14 J1831+291 156 12.0 3.84 3.49 Jul 14 J0040+331 190 15.1 4.10 3.74 2380 Jul 9 J1831+291 170 11.0 4.19 Jul 9 J0040+331 173 15.0 4.59 4.13 Jul 14 J0040+331 156 16.5 4.84 4.12 5000 Jul 9 J1831+291 217 13.9 10.62 10.95 Jul 9 J0040+331 194 15.3 14.54 14.66 Jul 14 J0040+331 177 14.8 17.78 16.87

Table 6: Tsys /Tcal

Freq Date Tsys/Tcal (MHz) (1998) (Ch A) 430 Jul 9 { Jul 14 1.81 1405 Jul 9 18.60 Jul 14 18.13 2380 Jul 9 4.17 Jul 14 4.42 5000 Jul 9 2.88 Jul 14 2.90

RMS { 0.14 (7.6%) 0.30 (1.6%) 0.36 (2.0%) 0.02 (0.4%) 0.04 (0.9%) 0.02 (0.6%) 0.03 (1.0%)

7

Tsys/Tcal (Ch B) 1.44 { 16.19 15.65 3.64 3.66 4.74 {

RMS 0.16 (10.9%) { 0.22 (1.4%) 0.46 (2.9%) 0.04 (1.2%) 0.05 (1.3%) 0.02 (0.5%) {

Figure 1: The measured 430-MHz signal versus the celestial brightness temperature, TB , at that frequency as derived from Haslam et al. (1982). Table 7: Flux Densities of Target Sources

Source

Freq Flux Density (MHz) (Jy) J1609+266 430 3.379 1405 5.063 2380 3.502 5000 { G064.7+05.0 430 <0.038 1405 0.265 2380 0.475 5000 0.587 J2115+295 430 0.464 1405 0.792 2380 0.904 5000 1.118 J2203+317 430 3.216 1405 2.597 2380 2.441 5000 2.436 8

Error (Jy) 0.132 0.125 0.147 { { 0.008 0.023 0.071 0.026 0.021 0.040 0.136 0.132 0.064 0.102 0.300

Table 8: Derived Flux-Density Dierences for the Two Epochs

Source

Freq (MHz) J2115+295 430 1405 2380 5000 J2203+317 430 1405 2380 5000 G064.7+05.0 430 1405 2380 5000 J0040+331 430 1405 2380 5000


S (Jy) {0.002 {0.017 {0.070 0.144 {0.347 {0.018 {0.285 0.650 { {0.012 {0.100 0.137 { { { {


S (%) {0.4 {2.1 {7.7 12.9 {10.8 {0.69 {11.7 26.7 { {4.7 {21.1 23.3 { { { {

ZA (deg) 11.7/14.4 16.3/11.3 11.7/16.7 14.3/12.7 14.1/17.3 13.7/14.5 15.6/18.7 13.4/16.0 { 15.1/13.2 12.5/19.3 12.7/14.6 14.8/15.2 16.6/15.1 15.0/16.4 15.3/14.9

the data for which observations of both calibrators were available. The errors given in Col. 4 of the table are the quadratic sum of the error derived from the disagreement between calibration factors derived independently from the two calibrators, and the internal spread of the values derived from individual measurements. No error is included for the likely deviation of the adopted ux densities of the calibrators from their true Baars et al. (1977) values. This is believed to be about 5% for each calibrator. To permit meaningful inter-comparison, the measurements on July 9{10 and 14{ 15, 1998 were both reduced using only the source J0040+331 as calibrator, this being common to all frequencies on both days. The dierence between the derived

ux densities on the two days are listed in Table 8. It should be remembered that while the sources were observed within a relatively small range of azimuth and zenith angle, no correction was applied for the residual azimuth/zenith-angle dierences. In particular, the telescope gain at (and presumably below) 1.4 GHz is known to be relatively constant for ZA<15 , but to decrease monotonically beyond this. An example of this is the  {10% dierence in the ux densities for J2203+317 derived at 430 MHz for the two epochs. If we correct the gain for the high zenith angle of the source relative to the calibrator on the second epoch using more recent 430-MHz observations, this dierence is completely accounted for. In fact, taking this into account, the values between the two epochs at 430 and 1405 MHz agree to better than 5% for all sources. 9

Figure 2: The 4.8-GHz variability of J2203+317 over the past 17 yr, derived from the UMRAO database monthly averages.

Figure 3: The 4.8-GHz variability of J1609+266 between 1983 and 1986, derived from the UMRAO database monthly averages. 10

Telescope commissioning observations have established that the system gain becomes an increasingly complicated function of both azimuth and zenith angle as the frequency increases. This is also demonstrated by the dierences between our two epochs at 2380 and 5000 MHz, which show a progressive increase in magnitude with frequency, and are systematically negative and positive respectively. This latter phenomenon is presumably due to systematic eects from the J0040+331 observations. However, even at these higher frequencies, the mean dierences of 13% and 21% suggest that the uncertainties in the uxdensity values at these two frequencies in Table 7 due to this complex gain behavior are 6.5% and 10.5% respectively, using the two calibrators observed on the rst epoch. These values are, in fact, consistent with the calibration dierences found for the two calibrator sources at the rst epoch, and which are included in the errors given in Table 7. Of course, we cannot a priori exclude some contribution to the measured dierences at the two epochs from actual source variations. However, monthly 4.8-GHz averages from the UMRAO database (http://www.astro.lsa.umich.edu:80/obs/radiotel/radiotel.html; Aller et al. 1985) for the quasar, 2203+317 (Fig 2) suggest that, while this source has large long-term variations, these variations occur on the time scale of months rather than days. Similarly, J1609+266 is found from the UMRAO monthly averages between 1983 and 1986 (the period over which it was monitored) to be essentially non-variable at 4.8 GHz (Fig. 3), and it is therefore unlikely to have varied signi cantly over 5 days at our highest observed frequency of 2.38 GHz. The UMRAO database contains only a few measurements of 2115+295 between 1982 and 1985. These are at 8.0 and 14.5 GHz, but suggest that the source is not strongly variable on the time scale of a month.

6 Discussion All the spectra derived from our measurements (the lower panes in Figs. 4 { 7) are seen to be smooth within the errors. This con rms the power of the telescope, even in its early post-upgrade incarnation, to produce the quality broadband single-epoch spectra which were the objective of the present project. We will now brie y discuss the astronomical signi cance of the results obtained.

6.1 The Three Quasar Targets The spectra derived from published data are shown in the upper panes of Figs. 4 { 6. These suggest that while the GPS source, J1609+266 is not strongly variable over the years covered by these measurements, the large spreads in the data for the sources J2115+295 and J2203+317 indicate that these vary signi cantly on the time scale of years. (For the latter source, this is clearly seen in the UMRAO 4.8-GHz measurements of Fig. 2.) 11

The new Arecibo ux densities provide quasi-snapshot spectra for these sources. The GHz peak in the spectrum of J1609+266 is clearly de ned by the three ux densities in the lower pane of Fig. 4. For the two variable quasars, the spectral \snapshots" show smooth spectra over more than a decade of frequency. That of J2115+295 (lower pane of Fig. 5) is the more interesting, showing a signi cantly inverted spectrum, with a spectral index of 5000 1405  ,0:30 above 1405 MHz, and an apparent steepening to 1405  , 0 : 45 below this. In contrast, J2203+317 430 displays a classic at spectrum at the higher frequencies, with a mildly positive spectral index below 1 GHz (lower pane of Fig. 6). In respect of these spectral dierences, it may be of signi cance that although both sources show structure on the milliarcsec scale, recent images from VLBI observations made between 2.3 and 15 GHz (Fey & Charlot 1997; Kellermann et al. 1998) show J2115+295 to be considerably the more compact of the two. This would suggest that opticaldepth eects may be more pronounced for this source, consistent with its inverted spectrum. In fact, J2203+317 displays a very beautiful core-jet morphology at 2.3 GHz on the scale of a few tens of milliarcsec.

6.2 G064.7+05.0 The planetary nebula, G064.7+05.0, is found to be signi cantly optically thick below 5 GHz (Fig. 7). From our measured ux densities, we derive an optical depth of   2:5 at 1.4 GHz for a pure brehmstrahlung spectrum. The predicted ux density at 2.38 GHz then agrees very well with the measured value at that frequency in Table 7. The predicted ux density of 27 mJy at 430 MHz is consistent with the upper limit of 38 mJy given in Table 7, as is the non-appearance of the source in the 327-MHz WENSS survey (http://www.strw.LeidenUniv.nl/%7Edpf/wenss/). We note that the 430-MHz upper limit was derived from the actual scans, but the value of 38 mJy is similar to the predicted rms confusion for Arecibo at 430 MHz of 40 mJy obtained from the formula of Condon (1987).

7 Conclusions The present investigation was observed entirely by a group of undergraduate summer students at Arecibo Observatory, many having their rst observational experience. The success of the project in demonstrating that quality \quasisnapshot" spectra covering over a decade of frequency can be made with the rebuilt instrument, also demonstrates the ease with which relatively complicated experiments can now be observed (as it does the quality of the students!) The moderate gain changes with both azimuth and zenith angle present for measurements at the higher frequencies during the summer of 1998 should soon be reduced to the design values in ZA, and hopefully eliminated in respect of azimuth, following the adjustment of the Gregorian dome attitude in late 1998. As 12

only a small gain-ZA dependency should be present subsequently, with only a relatively simple correction being needed to account for this, very high quality spectral measurements should then be possible, eventually spanning a frequency range from 300 MHz to 10 GHz. In addition, a continuum correlation polarimeter is presently nearing completion at Arecibo, and full Stokes-parameter information should soon be available at no extra cost but data volume for all continuum observations. The possibility of measurements across a 30:1 frequency range for studies of total-intensity spectra, Faraday rotation and depolarization is highly exciting.

Acknowledgements

The summer students are grateful for support from the NSF Research Experience for Undergraduates Program, the Univ. Metropolitana, and the Univ. of Houston. We wish to thank Phil Perillat without whom the project would not have been possible. We are very grateful to the Arecibo telescope operators who cheerfully helped guide us through the observing sessions.

References

Aller, H. D., Aller, M. F., Latimer, G. E. and Hodge, P. E. 1985 Astrophys. J. Suppl. 59, 513.

Baars, J. W. M., Genzel, R., Pauliny-Toth, I. I. K. and Witzel, A. 1977 Astron. & Astrophys. 61, 99. Condon, J. J. 1987 \Scienti c Bene ts of an Upgraded Arecibo Telescope", Eds. Taylor, J. H. and Davis, M. M., NAIC, P89. Condon, J. J., Cotton, W. D., Greisen, E. W., Yin, Q. F., Perley, R. A., Taylor, G. B. and Broderick, J. J. 1998 Astron. J. 115, 1693. Condon, J. J. and Kaplan, D. L. 1998 Astrophys. J. Suppl. Ser. 117, 361. Fey, A. L. and Charlot, P. 1997 Astrophys. J. Suppl. Ser. 111, 95. Haslam, C. G. T., Salter, C. J., Stoel, H. and Wilson, W. E. 1982 Astron. & Astrophys. Suppl. Ser. 47, 1. Kellermann, K. I., Vermeulen, R. C., Zensus, J. A. and Cohen, M. H. 1998 Astron. J. 115, 1295. Kuhr, H., Witzel, A., Pauliny-Toth, I. I. K. and Nauber, U. 1981 Astron. Astrophys. Suppl. 45, 367. Shaer, D. B. and Kellermann, K. I. 1998 IAU Colloquium No. 164, Eds. Zensus, J. A., Taylor, G. B. and Wrobel, J. M., A.S.P. Conf. Ser. No. 144, P191.

13

Figure 4: Continuum spectra for J1609+266 (CTD93) between 300 MHz and 10 GHz, a) for data taken from Kuhr et al. (1981), the NVSS Survey (1998), White & Becker (1992), the GB6 Catalog, the UMRAO database and the Texas Catalog (http://utrao.as.utexas.edu/txs.html) and, b) for the new Arecibo data contained in Table 7.

14

Figure 5: Continuum spectra for J2115+295 between 300 MHz and 10 GHz, a) for data taken from Kuhr et al. (1981), the NVSS Survey (1998), the WENSS Survey (1998), White & Becker (1992), the GB6 Catalog, the UMRAO database and the Texas Catalog (http://utrao.as.utexas.edu/txs.html) and, b) for the new Arecibo data contained in Table 7.

15

Figure 6: Continuum spectra for J2203+317 between 300 MHz and 10 GHz, a) for data taken from Kuhr et al. (1981), the NVSS Survey (1998), the WENSS Survey (1998), White & Becker (1992), the GB6 Catalog and the Texas Catalog (http://utrao.as.utexas.edu/txs.html) and, b) for the new Arecibo data contained in Table 7.

16

Figure 7: Continuum spectra for Planetary Nebula G064.7+05.0 between 300 MHz and 10 GHz, a) for data taken from the Einline database, the NVSS Survey (1998), the WENSS Survey (1998), White & Becker (1992), and the GB6 Catalog and, b) for the new Arecibo data contained in Table 7. The values plotted as crosses from the WENSS survey at 327 MHz (top gure), and Arecibo at 430 MHz (bottom gure), represent upper limits.

17