Use as many different inoculation doses as possible spanning the range where an effective infection should happen. Use a dose-response curve analysis to determine the ED90 (or ED95). If you can use a single animal that gets increasing doses, it's sufficient to use 3 animals. However, there must not be a carry-over effect of previous inoculations to give valid data. If you must use a differet animalfor each dose, you need 1 animal per dose. if you have no idea of a good dose or if you have to determine the dosequite precisely, you may plan a 2-stage experiment:in stage 1, use a few doses (4-5) to span a large range and identify a narrower interesting range, and in stage 2 test 4-5 doses within this range. (Note: just 1animal per dose, no need to sacrifice several animals at a particular dose!)

Thank you! This aligns with what I was thinking. I'm aware of doses that will cause infection and disease for my virus, but it's unclear if they are too high. And the virus, when it makes a successive infection, is fully-immunizing. So unless I work up from a very low dose, if the animal develops any antibodies (perhaps without clinical disease), I won't be able to use the same animal again. To be safe, it sounds like separate animals per dose is the way to go.