Progress in replacement has been restricted by certain plausible, but untenable, assumptions, summed up as the high fidelity fallacy.
One of the most important and challenging sections in The Principles of Humane Experimental Technique is Russell and Burch’s discussion of models and their proper use. Research on human health and disease, and tests on the effects of substances and pathogenic organisms, frequently involve establishing a model (a laboratory animal) of the system to be investigated (the human organism), then studying the model in ways which would not be possible with the system being modelled (for practical or ethical reasons).
By their very nature, models must differ from what is being modelled, and the importance and consequences of this difference depend on two major factors, fidelity and discrimination.
For a pdf of the full article click here.