1 Introduction

The neutral hydrogen atom (H I) is the most abundant of the atomic constituents known to exist in the interstellar medium (ISM). Its hyperfine transition at $\lambda\sim$21-cm ($\nu$ = 1420.40576 MHz), predicted on theoretical grounds by Dutch astronomers (van de Hulst [1945]), was first successfully detected by an American team (Ewen & Purcell [1951]). Shortly afterwards, this line was also observed by Dutch (Muller & Oort [1951]) and Australian (Pawsey [1951]) groups, respectively.

Taking into consideration the excitation conditions prevailing in the ISM it was realized (van de Hulst et al. [1954]; Purcell & Field [1956]; Field [1958]), that this line could be a powerful tool to directly explore the structure of the Milky Way at large. Indeed, the $\lambda\sim$21-cm emission is so ubiquitous that its brightness temperature, column density and velocity field provide key information on a variety of physical circumstances which can be found in the ISM. Furthermore, for a few phenomena, like for most of the high velocity clouds (HVC), the 21-cm line remains the only tracer as yet identified.

Therefore, it is understandable that right from the beginning a lot of effort went into carrying out surveys of the H I 21-cm emission line, which cover huge areas of the sky, with a sensitivity good enough to be suitable for general investigations of the galactic interstellar medium (ISM). The main characteristics of the major southern sky H I surveys are summarized in Table 1. Bearing in mind their limitations in sensitivity, velocity resolution and both velocity and spatial coverage, and considering that the present day technology of spectral analyzers and low noise receivers allows a much better velocity coverage and resolution, as well as a much lower rms noise, it was highly desirable to undertake a long-term observational project with the aim of surmounting the restrictions inherent to the available H I surveys.

 

 

Table 1: Relevant parameters of the southern H I surveys

Survey	HPBW	rms noise	Velocity	Velocity	Grid	Region
			coverage	resolution	spacing	surveyed
	($^\prime$)	(K)	( $\rm\, km\, s^{-1}$)	( $\rm\, km\, s^{-1}$)	(degrees)
Cleary et al. (1979)						$\delta\leq$ $-30^\circ$
Intermediate latitude	48	0.3	-148 $\leq$ V $\leq$ +300	7	$\bigtriangleup\delta$ = 1	$\mid b \mid~\geq 10^\circ$
						b $\geq$ $-25^\circ$
High latitude	48	0.3	-230 $\leq$ V $\leq$ +218	7	$\bigtriangleup\delta$ = 2	$\delta~\leq$ $-30^\circ$
						b < $-25^\circ$
Kerr et al. (1986)	48	0.3	-150 $\leq$ V $\leq$ +150	2.1	$\bigtriangleup$ l = 0.5	240$^\circ$ < l < 350$^\circ$
					$\bigtriangleup$ b = 0.25	$\mid b \mid~\leq 10^\circ$
Colomb et al. (1980)	30	0.8	-50 $\leq$ V $\leq$ +50	2	$\bigtriangleup\delta$ = 1.0	$\delta\leq$ -25$^\circ$
						$\mid b \mid~\geq 10^\circ$
This Survey	30	0.07	-450 $\leq$ V $\leq$ +400	1.27	$\bigtriangleup$ l = 0.5/cos b	$\delta\leq$ -25$^\circ$
					$\bigtriangleup$ b = 0.5

From the beginning it was clear that to succeed in this undertaking, the observing instrument should meet two main requirements, namely: a) a system temperature against cold sky in the range of 30-40 K, and b) a back-end that should be able to provide both a large velocity coverage, at least 1000 $\rm\, km\, s^{-1}$, and a velocity resolution of the order of 1 $\rm\, km\, s^{-1}$. As a back-end we ended using the "old'' Arecibo 1008-lags autocorrelator, which, at that time, by mid 87, had already been replaced by a new one. The computer controlling "on-line'' the data-taking process was a $\mu$VAX II. Since the autocorrelator interfaces and the driving program were originally written for a Harris Computer, we had to rebuild the hardware interfaces and rewrite the data-taking software. As for the receiver a dual polarization L-band cooled receiver was built by a IAR staff member at the Max-Planck-Institut für Radioastronomie (MPIfR), Bonn, Germany, with the technical support of their radio-frequency laboratory, which was at that time developing a similar receiver for the 100-m Effelsberg radiotelescope. The details of the autocorrelator hardware and software, and the automation of the antenna movements will be published elsewhere.

An extra motivation was brought into scene, when we became aware of the characteristics of the northern hemisphere Leiden/Dwingeloo H I survey (Hartmann [1994]; Hartmann & Burton [1997]). The observational parameters intended for this survey (see Table 1) were so similar to the one we were planning to use, that the idea of having, as a result of both projects, a whole sky high sensitivity H I survey of uniform spatial coverage, derived naturally in an agreement with Drs. Burton and Hartmann. What we describe here is the IAR survey itself and its principal merits, as compared to the others listed in Table 1. These merits lie in the improved sensitivity, which is a factor of four to ten better, the sky coverage, the sampling, and in the fact that it will be corrected for the effects of stray radiation.