The article addresses the question whether the glandular epithelium of the mammary gland, or other mucous membranes, has the capacity to transfer selectively IgA and IgM into secretion in a manner similar to that for IgG1. The article adds that such a mechanism has been presumed to account for the high levels of IgA and IgM in colostrum (relative to those in serum) of cows. However, this presumption is invalidly based, as it does not take into account local production of immunoglobulin, which would have the effect of increasing concentrations in the interstitial fluid.