The open-loop (i.e. blind) guiding of the MONET telescopes is, well...., capable of being improved! As the pointing model gets better, the blind pointing will get better, but a large steel telescope just isn't designed to be so exact (small high-tech amateur telescopes can be so good that they essentially don't need any additional help in guiding). The best solution is to place an additional camera near the main science camera just for guiding, but we don't have enough focal-plane space for this solution. One of these days we'll get the piggy-back telescopes working, but a non-aligned guider whose camera has to be rotated differently from the science camera is quite difficult to develop.
For projects where lots of science-images are being taken (e.g. photometry of an eclipsing variable, motion of an asteroid,...), we can use the short exposures to correct the guiding while the series is being taken. Dr. Tim-Oliver Husser, our magical MONET-programmer, is testing such a system on MONET/South. Here is an example of how well the telescope can track over an hour:
As soon as the system is fully tested, we will add science-image autoguiding to all observations automatically. Observations using just a few long-exposed images obviously won't profit from this service, but most MONET projects will.