The exciting contributions made by radio astronomical observations over the last decade would not have been possible without the development of radio telescopes of increasing sensitivity and resolving power. In particular the study of the `radio galaxies' has depended very much on the availability of instruments capable of determining their positions with accuracy, investigating their detailed angular structure and their radio spectrum. These sources have a radio emission very much greater than that of our own Galaxy, and the mechanism by which this intense radiation is produced is of great astrophysical interest; the socalled `quasi-stellar' sources in which the whole radio emission originates in a region only about one hundredth of the dimensions of the galaxy, represent an extreme case involving unknown sources of energy.
By using radio telescopes of great sensitivity and angular resolving power, radio galaxies can be detected at very large distances; owing to the time taken for the radio waves to reach the earth, they are consequently observed as they were some thousands of millions of years ago, and such observations offer the possibility of exploring evolutionary changes in the Universe. In order to extend these types of observation, more powerful instruments became necessary, but the construction of larger steerable paraboloids of sufficient accuracy presents very great engineering difficulties. More economical instruments have been built by using fixed parabolic cr spherical reflectors, which can be supported from the ground at many points, but even with these instruments the cost becomes very great if an angular resolution better than five minutes of arc is sought.
Alternative methods of building large telescopes have been developed, in one of which advantage is taken of the fact that for most observations the sky can be regarded as unchanging over the period of observation; it is then not necessary to have the whole of the telescope present at the same time, and a large effective instrument may be `synthesized' by using two small elementary aerials arranged successively in all the positions present in the large area. The observations may be combined in a computer to produce a map of the sky having a resolution and signal-to-noise very much greater than that associated with the small elementary aerials.
Four large instruments have been built using this principle, two of which have been in use for six years. They have been particularly valuable both in enabling large resolving powers to be attained at long wavelengths in order to establish the radio spectra of the sources, and in extending the observations to the weakest sources yet observed.