Studies of human gastrointestinal drug delivery are severely restricted by the complex ethical considerations which preclude many such investigations. In an attempt to overcome such restrictions we have utilised a chimeric mouse model, which is suitable for studying human gastrointestinal drug delivery under carefully defined experimental conditions. By means of subcutaneous transplantation of human intestine into scid mice, this host strain is able to accept and accommodate the xenografts for several months. Under these conditions, immature human fetal intestine differentiates onto a morphologically and functionally developed tissue which closely resembles normal human gut. The present work describes the use of this experimental system to study epithelial transcytosis of horseradish peroxidase, a model peptide, across human small intestine in vivo. The great potential for this model is that it lends itself to studies which are not feasible or unethical in patients and volunteers and, as such, provides a novel alternative means for gaining essential basic scientific information which is directly relevant to humans. This is especially important since experimental extrapolations made between species are often difficult, if not misleading.