Two major antibody classes operate in saliva: secretory IgA (SIgA) and IgG. The former is synthesized as dimeric IgA by plasma cells (PCs) in salivary glands and is exported by the polymeric Ig receptor (pIgR). Most IgG in saliva is derived from serum (mainly via gingival crevices), although some is locally produced. Gut-associated lymphoid tissue (GALT) and nasopharynx-associated lymphoid tissue (NALT) do not contribute equally to mucosal PCs throughout the body. Thus, enteric immunostimulation is an inadequate mode of stimulating salivary IgA antibodies, which are poorly associated with the intestinal SIgA response, for instance after enteric cholera vaccination. Nevertheless, the IgA response in submandibular/sublingual glands is better related to B cell induction in GALT than the parotid response. Such disparity is suggested by the elevated levels of IgA in submandibular secretions of AIDS patients, paralleling their highly upregulated intestinal IgA system. Moreover, in patients with active celiac disease, IgA antibodies to disease-precipitating gliadin are reliably represented in whole saliva but not in parotid secretion. Parotid SIgA may be more consistently linked to immune induction in palatine tonsils and adenoids (human NALT), as supported by the homing molecule profile of NALT-derived B cell blasts. Also several other variables influence the levels of antibodies in oral secretions. These include difficulties with reproducibility and standardization of immunoassays, the impact of flow rate, acute or chronic stress, protein loss during sample handling, and uncontrolled admixture of serum-derived IgG and monomeric IgA. Despite such problems, saliva remains an interesting biological fluid with great scientific and clinical potentials.