The SHEVE data were obtained and processed as described in chapter 2. Table 7.1 gives a journal of the observation dates and the participating telescopes.
Figure 7.1 shows the image resulting from observations at 2.3 GHz. The source appears unresolved with the 21.5 
11.7 mas restoring beam. The total flux in the image is 4.7 Jy and a dynamic range of approximately 100:1 was realised.
Figure 7.2 shows the image from observations at 4.8 GHz. With the 2.6 2.3 mas beam the source is resolved into a double with separation 6.8 mas and position angle 135 degrees. The total flux density in the image is 4.3 Jy and a dynamic range of approximately 100:1 was realised. The sizes, PA's and flux densities of the two components (de-convolved from the beam) are given in Table 7.2, along with the corresponding estimates of brightness temperature, radio luminosity and total internal energy (calculated from equipartition considerations).
Figure 7.3 shows the results of observations at 8.4 GHz. The total flux density in the image is 3.4 Jy and the dynamic range approximately 100:1. The clean components resulting from both the 4.8 and 8.4 GHz data were convolved with a 6.4 mas FWHM circular Gaussian beam (corresponding to the major axis of the formal 8.4 GHz beam) and the spectral indices of the two components estimated (Table 7.2).
Table: Observation log for PKS 1718-649
Figure: Map peak, 2.0 Jy/beam. Contours, -1, 1, 2, 4, 8, 16, 32, 64% of peak. Beam FWHM, 2.6 2.3 mas @ -29.5 
.
Figure: Map peak, 3.7 Jy/beam. Contours, 1, 2, 4, 8, 16, 32, 64% of peak. Beam FWHM, 21.5 11.7 mas @ -83.0 .
Figure: Map peak, 1.8 Jy/beam. Contours, 1, 2, 4, 8, 16, 32, 64%. Beam FWHM, 6.4 3.8 mas @ -83.2 .
The small temporal difference between each of these epochs and the mismatch in frequency and resolution means that no estimate of the motion of components on the pc-scale can be made. The model derived by Preston et al. [1989] for the VLBI structure of PKS 1718-649 in 1982 has the two components at a significantly different position angle and at a much larger separation than the images in Figures 7.1, 7.2, and 7.3 indicate. These differences can be readily explained by the sparseness of the data available in the earlier experiment. Evolution in the source need not be invoked.
In addition to the data from 1982, observations of PKS 1718-649 were made with the SHEVE array at two other epochs, 1988 July 24 at 2.3 GHz and 1989 December 10, also at 2.3 GHz. These data have not been published. The 1988 and 1989 data were model-fit using the technique described in 
3.1 of chapter 5. Unfortunately, like the 1982 data, the 1988 and 1989 data were too sparse to sensibly constrain a model, although there is obvious structure on the long baselines at 2.3 GHz. No useful comparison could be made between the new set of images and the model-fitted data from 1982, 1988, or 1989.
Further imaging observations will be required before a useful estimate of the pc-scale evolution in PKS 1718-649 can be made.
As can be seen from Table 7.2 the north-west component is the brighter and more compact of the two components, although both components are comparable in size and marginally resolved by the beam. The north-west component has a steep spectrum and the south-east component has a flat spectrum. The components appear very similar at 4.8 GHz and neither strongly resemble the pc-scale jets seen in many other compact radio sources. The two components are also significantly misaligned in position angle, differing by approximately 49 , and neither are aligned along the position angle joining the two components.
The images in Figures 7.1, 7.2, and 7.3 were made without the data on baselines from the Australian antennae to Hartebeesthoek in South Africa. The two components are prominent on these long baselines at 4.8 and 8.4 GHz, indicating that there exists emission from both components from regions more compact than about 0.3 mas (0.08 pc).
Table: Radio properties of PKS 1718-649 at 4.8 GHz.