3 Observations

3.1 The Nançay observations

In total, new observations of 437 flat galaxy candidates were undertaken using the Nançay Decimetric Radio Telescope and 1024-channel autocorrelator spectrometer between May 1996 and January 1999.

The Nançay telescope is a meridian transit-type instrument with an effective collecting area of roughly 7000 m². At 21-cm, the FWHP beam size is $3\hbox{$.\mkern-4mu^\prime$ }6$ E-W$\times\ge$ $22\hbox{$.\mkern-4mu^\prime$ }0$ N-S. Because of the design of the Nançay telescope, the N-S beam diameter changes as a function of the declination of the source (see Fig. 1).

Observations were obtained in total power mode, using consecutive pairs of two-minute on-source and two-minute off-source integrations. Off-source integrations were generally taken $\sim20'$ E of the target position. Tracking was limited to 0.75-3.0 hours per source per day. Due to scheduling constraints, total integration times for each target varied, but were typically a few hours per galaxy. Typically system temperatures were $\sim40$ K.

Two different autocorrelator configurations were used to obtain the Nançay spectra. The bulk of the observations were obtained in a single (horizontal) polarization mode, with the autocorrelator divided into four banks of 256 channels each and a small overlap in frequency between consecutive banks. The total bandpass was $\sim23$ MHz, yielding a channel spacing of $\sim5.3$ km s^-1, for an effective resolution of $\sim6.3$ km s^-1. Galaxies observed with this set-up were initially observed over a search range from 341-5299 km s^-1. Most targets were also observed over higher-velocity search ranges, typically from 5023-10138 km s^-1. For a small number of targets, observations were obtained with a bandpass of $\sim$12 MHz and channel spacings of $\sim$2.6 km s^-1. In these cases the effective search ranges were 324-2900 km s^-1 or 5006-7664 km s^-1.

   
3.1.1 Reduction and analysis of the Nançay data

Our Nançay spectra were reduced using the SIR spectral line reduction packages available at the Nançay site. With this software we fitted and subtracted baselines (third-order or lower polynomials) and applied a declination-dependent conversion factor to convert from units of $T_{\rm sys}$ to flux density in millijanskys. The $T_{\rm sys}$-to-mJy conversion factor is determined through regular monitoring of strong continuum sources by the Nançay staff. This procedure yields a relative calibration accuracy of $\sim$15%. In addition, we applied a scaling factor of 1.26 to all Nançay spectra to derive our absolute flux scale (see Matthews et al. 1998a).

Peak fluxes, integrated fluxes, velocity widths, and radial velocities were measured for all detected sources using IDL software written by one of us. Our measured parameters are presented in Table 1. Spectra of all the detected galaxies are presented in Fig. 2.

In order to increase signal-to-noise to a level suitable for high-quality measurements, in most cases we smoothed the data to a resolution of $\sim$16 km s^-1 before performing the line measurements. In general smoothing permits a better determination of the systemic velocity, although it results in a systematic overestimation in the measured linewidths (see e.g., Giovanelli et al. 1997). For this reason, we have derived a correction to our measured linewidths to account for instrumental resolution and smoothing (see below).

We determined peak fluxes in the detected lines using the mean of the two strongest channels, after smoothing. Velocity widths were measured interactively, by moving the cursor outward from the profile center. Radial velocities were defined to be the centroid of the two 20% peak maximum points on the line profile. Total fluxes were measured by integrating the area under the line profile on the baseline-subtracted spectrum.

In a number of observations obtained at Nançay during January 1997 spurious features with velocity widths of $\sim$200-300 km s^-1 intermittently appeared in our spectra near $V_{\rm h}\sim1300$ km s^-1. Their cause is unknown. These features generally had intensities of a few tens of millijanskys, and thus mimicked the appearance of a real extragalactic source. As a result, we generally treat apparent detections near $V_{\rm h}\approx1300$ km s^-1 as marginal unless they were subsequently confirmed by independent observations (see Notes to Tables 1 & 2 in Appendices A & B).

3.2 The Green Bank observations

Observations of 54 FGC targets were carried out in October 1997 using the NRAO 140-ft telescope located in Green Bank, West Virginia. The primary goal of the Green Bank observations was to observe high-priority FGCE targets (e.g., objects of large angular size) located between declinations $-38^{\circ}$ and $-44.5^{\circ}$ (i.e., which were too far south to be observed at Nançay) or which otherwise could not be observed at Nançay due to scheduling constraints. In addition, we observed at Green Bank a few sources that we had previously detected at Nançay in order to provide an external check on flux calibration.

Observations at Green Bank were carried out using the 1.3-1.8 GHz receiver and Mark IV 1024-channel autocorrelator spectrometer. The autocorrelator was divided into two orthogonally polarized banks of 512 channels, each with a 20 MHz total bandpass, but with center velocities offset by $\sim$500 km s^-1. Typically, search ranges were approximately 600-4700 km s^-1 and 1100-5200 km s^-1 in the respective banks, although a few targets were observed over slightly different velocity ranges (see Table 3). Channel widths were $\sim$8.4 km s^-1, for an effective velocity resolution of $\sim$10 km s^-1.

Observations were obtained in total power mode using a series of six-minute on-, and six-minute off-source integrations. The off-source observations were obtained at locations offset $6^{\rm m}42^{\rm s}$ west of the target position. Total integration times were typically 1.25 hours per source. Observations were made after sunset to avoid solar interference in the sidelobes. At 1400 MHz, the 140-ft telescope has a FWHP beamwidth of $\sim$21'. System temperatures ranged from 18 K for observations near zenith, to 28-35 K for observations at $\delta<-40^{\circ}$. Absolute flux calibration was accomplished through the monitoring of both line calibrators (from Davies et al. 1989 and van Zee et al. 1997) and continuum calibration sources (selected from the VLA Calibration Manual) several times per night. These two methods generally agreed to within better than 5%. The mean derived calibration factor varied by $\sim\pm$3.5% over the course of 8 nights.

Table 2. Marginal and questionable detections

Galaxy Name	$\alpha$(B1950.0)	$\delta$(B1950.0)	Type	a$\times$b	rms	$F_{\rm max}$	W20	W50	$V_{\rm h}$	$\sigma(V)$	S	$\sigma(S)$	S/N	Notes
ESO 243-016	00 55 01.8	-42 56 24	Scd	1.29$\times$0.11	12.8	29	239	192	3985	12	2.9	1.8	2.3	G
FGCE 124	01 03 37.9	-19 08 17	Sd	0.71$\times$0.09	3.43	9.4	180	146	7625	24	0.88	0.66	2.8
UGC 778	01 11 25.5	+49 57 39	Sd	1.10$\times$0.11	3.34	11	266	224	6867	22	1.4	0.7	3.3
FGC 242*a	02 03 33.2	+30 52 50	Sd	1.03$\times$0.11	3.92	12	172	166	3891	10	1.2	0.8	3.1	*
FGC 242*b	...	...	...	...	3.69	11	142	126	6051	15	0.79	0.62	3.0	*
FGC 284	02 20 04.6	+17 35 17	Sd	1.01$\times$0.10	7.5	35	222	200	4128	8	4.2	1.2	4.7	G,*
FGC 392	03 08 28.4	+34 49 59	Sd	1.18$\times$0.11	3.81	9.9	310	231	4566	-	1.3	0.9	2.6
FGCE 454	04 59 29.6	-22 19 37	Sd	0.65$\times$0.09	3.00	8.0	289	275	1385	16	1.3	0.8	2.7
FGC 511	04 49 26.2	-06 01 51	Sd	0.83$\times$0.08	3.99	21	166	124	2749	14	2.3	0.8	5.2	d,*
FGCE 527	05 44 08.4	-17 24 10	Sd	1.19$\times$0.12	3.42	6.7	172	152	5738	26	0.74	0.71	2.0
FGCE 554	06 00 41.0	-28 04 07	Sd	0.76$\times$0.08	2.96	6.1	118	93	6417	27	0.56	0.56	2.1
FGC 567	06 24 31.2	+56 13 32	Sd	1.10$\times$0.10	2.92	8.5	284	201	5293	35	1.3	0.7	2.9	*
FGC 598	07 03 28.6	+44 44 36	Sd	0.95$\times$0.11	2.76	12	239	181	5867	20	1.8:	0.7	4.2	*
UGC 3716*	07 06 58.0	+39 47 13	Sd	0.91$\times$0.10	2.34	16	398	312	6359	15	3.4	0.7	6.8	c,p,*
FGC 613	07 13 20.5	+33 02 37	Scd	1.03$\times$0.13	4.12	11	186	173	6392	15	1.1	0.8	2.6	*
FGC 647	07 36 35.6	+62 45 55	Sd	0.83$\times$0.10	3.88	9.9	314	155	6577	55	1.3	0.9	2.6
UGC 4068	07 49 33.4	+40 18 14	Sd	1.48$\times$0.18	2.35	4.9	159	138	8263	25	0.54	0.49	2.1
FGC 781	08 37 46.8	+57 49 27	Sd	0.72$\times$0.09	5.38	12	119	97	2340	24	1.1	1.0	2.2
FGC 904	09 29 32.5	-16 27 05	Sd	1.49$\times$0.12	2.90	20	287	261	2180	8	3.8	0.8	6.9	c,p,*
FGCE 768*a	09 41 49.0	-24 58 20	Sd	0.86$\times$0.12	4.05	11	174	111	2504	32	1.1	0.8	2.8
FGCE 768*b	...	...	...	...	4.05	9.4	202	190	3775	17	1.4	1.0	2.3
ESO 498-023	09 42 15.4	-23 43 21	Sd	1.12$\times$0.11	3.07	8.6	242	182	5778	31	0.97	0.65	2.8
UGC 5550	10 14 03.7	+64 38 19	Sdm	1.01$\times$0.10	5.47	14	237	199	4507	27	1.7	1.2	2.5	*
FGC 1136	10 46 04.6	+19 24 05	Sd	0.90$\times$0.10	4.37	14	71	57	1856	12	0.67	0.55	3.2
FGC 1248	11 24 22.8	+70 45 15	Sd	0.92$\times$0.11	3.86	11	260	251	6934	12	1.6	1.0	2.8	b,*
FGC 1359	12 02 32.4	-03 35 52	Sd	1.01$\times$0.13	2.84	6.4	278	236	5712	33	1.1	0.7	2.2	*
UGC 7553*	12 24 29.6	-01 14 27	Sdpec	1.18$\times$0.11	2.11	12	371	332	8820	13	2.3	0.6	5.5	c,*
FGC 1563	13 05 36.3	-15 58 19	Sdm	0.95$\times$0.12	3.10	11	185	149	2790	19	1.1	0.6	3.5	i,*
ESO 576-047	13 21 28.8	-17 38 24	Sd	1.25$\times$0.10	3.91	17	...	...	$\ge$6801	...	...	...	4.4	e,*
UGC 8538	13 31 07.7	+46 05 33	Sd	1.41$\times$0.11	4.14	21	$\ge$133	$\ge$116	1327	9	$\ge$2.4	0.9	5.0	i,*
FGC 1647	13 35 47.3	+08 26 30	Sd	0.95$\times$0.10	2.99	6.4	84	56	2662	28	0.51	0.53	2.1
FGC 1660	13 41 07.0	+22 20 51	Sdm	1.08$\times$0.10	3.09	9.6	320	275	8134	24	1.2	0.7	3.1
FGC 1680	13 50 39.7	+68 37 15	Sdm	0.84$\times$0.09	6.31	18	191	165	3865	20	1.6	1.2	2.8
FGC 1793*a	14 38 25.6	-17 25 20	Sd	0.65$\times$0.09	3.16	32	78	63	3422	4	2.2	0.5	10.3	*
FGC 1793*b	...	...	...	...	3.16	36	240	106	4140	11	5.0	0.7	11.5	*
ESO 580-019	14 42 50.6	-22 14 57	Sd	0.95$\times$0.10	4.26	10	163	131	5638	-	0.96	0.82	2.4
FGC 1865	15 12 02.2	+37 31 14	Sdm	0.95$\times$0.10	2.78	10	246	235	8988	10	0.94	0.53	3.7
FGC 1869	15 12 34.2	+06 06 52	Sdm	0.83$\times$0.09	2.55	8.1	83	36	6870	24	0.48	0.39	3.2
FGC 1903	15 25 07.1	+66 23 16	Sdm	0.93$\times$0.11	4.56	33	207	146	3459	12	4.5	1.2	7.2	*
FGC 2028	16 20 17.5	+63 14 13	Sd	1.01$\times$0.08	4.60	9.1	207	174	3075	33	1.4	1.1	2.0
FGC E1446*	20 12 04.9	-20 00 23	Sd	0.73$\times$0.07	4.68	22	234	225	4914	7	3.6	1.2	4.6	c,*
ESO 596-026	20 20 13.8	-21 21 14	Sd	1.03$\times$0.11	1.83	7.2	285	...	8393	46	0.76	0.37	4.0
ESO 342-044	21 22 36.3	-40 20 25	Scd	1.14$\times$0.11	17.4	50	468	449	5034	12	12.6	4.0	2.9	G,*
ESO 407-012	23 12 27.5	-33 31 30	Irr	1.10$\times$0.11	2.92	10	153:	80:	5999	27	0.65	0.46	3.5	*
FGC 2506	23 29 18.6	-01 06 05	Sdm	0.99$\times$0.11	4.00	14	...	...	$\ge$5200	...	...	...	...	e,*

Unfortunately our Green Bank run was plagued by recurrent radio frequency interference (RFI) at a range of frequencies throughout our observed spectral range. Approximately 15% of our scans in both polarizations were affected. Most of the interference caused either high-amplitude ringing throughout a large fraction of the bandpass, or else produced strong, broad spectral features mimicking astronomical signals. Some of the interference was traced to on-site sources (which were located and eliminated), while other sources remained unidentified. Effects appeared to be most severe during observations of targets near the horizon.

In addition to the interference problems, during a night following a substantial rainfall, intermittent baseline structures occurred in a number of our spectra, which could not be removed by fitting sinusoids or low-order polynomial baselines. As a result, a significant number of additional scans had to be discarded. In total, $\sim$25% of our Green Bank scans were unusable due to interference and/or bandpass structure. As a result, the rms noise in our Green Bank spectra was typically $~\rlap{$>$ }{\lower 1.0ex\hbox{$\sim$ }}$3 times higher than in our Nançay data.

3.2.1 Reduction and analysis of the Green Bank observations

The Green Bank data were reduced using the line version of the NRAO Unipops spectral analysis package. Baselines (typically first or second-order polynomials) were subtracted from individual scans and the scans for each object were then averaged. Scans contaminated by significant interference or strong residual baseline structure were not included in the final averages. Peak fluxes, line widths, radial velocities, and integrated fluxes were measured for detected sources in the manner described for the Nançay data, but using routines written in Unipops. All measurements of the Green Bank data were performed on unsmoothed spectra.