We have been studying the properties of the canine parvovirus (CPV), a new virus of dogs that was first recognized in 1978 when it spread world-wide, causing serious disease and killing many millions of dogs. The CPV appears to be a naturally derived host range variant of a virus that is similar to the feline panleukopenia virus (FPV) which infects domestic cats and many related animals including large cats, raccoons, and foxes. CPV differs from FPV in only a small number of sequence in its genome, and we have shown that the ability of the virus to infect dogs resulted from only a small number of changes in its capsid – the virus particle which carries the viral DNA between cells and animals. We are now seeking to understand in detail how the virus infects cells and animals, and to determine how viral host range varies to allow it to infect many different types of animals. As part of this study we are examining the atomic structures of the parvovirus capsids and defining point mutants that control host range, and as part of that study we are examining the flexibility of the capsid using specific probes and antibodies. Studies of a series of variant viruses indicates that the capsids of viruses from different animals are very precisely designed for the infection of their hosts, and that most modifications of the optimal structures result in viruses that cannot infect dog cells. This points to a very close interaction between the virus capsid and the processes of infection of the host cells and of animals.