Fig. 4. Survey of a cross-cryosection (5 µm thickness) of nasal gland tissue stained with Hemalaun and Eosin (A), description of structural correlates (c.f. Butler et al., 1991) (B) and immunohistochemical detection of AQP1 (C,D) or AQP5 (E–H), respectively, in nasal gland tissue obtained from naïve (C,E) or osmotically stressed (D,G) ducklings. Specific signals at the sites of antibody binding are visible as brownish precipitates, which are products of the peroxidase-reactions with diaminobenzidine and hydrogen peroxide as substrates. The dark speckles within some of the capillaries are red blood cells (RBCs). Control experiments, in which cryosections were preincubated with (C1,D1) or without (C2,D2) 3% hydrogen peroxide-solution for 30 min at room temperature before immunostaining without primary antibody, revealed that endogenous peroxidases of RBCs reacted non-specifically with the peroxidase substrates. Note the presence of AQP1 in capillary endothelial cells in tissue of naïve animals (C). Epithelial cells of the secretory tubules or the ducts are devoid of AQP1-related signals. Note the expression of AQP5 in individual cells lining the primary and central ducts. Apical as well as basolateral plasma membrane compartments of ductal cells in glands of naïve ducklings are labelled (F). Much less immunostaining was observed in nasal gland cryosections obtained from osmotically stressed ducklings using AQP1-(D) or AQP5-specific (G,H) antibodies, respectively.