Local nasal immunotherapy (LNIT) was administered in a double-blind study to 67 subjects. Twenty-three received an unmodified ragweed extract (RW), 24 received a glutaraldehyde polymer of ragweed extract (PRW), and 21 received placebo. Serum ragweed-specific IgE (S-IgE), ragweed-specific nasal secretory (NS-) IgE, secretory IgA (SIgA) and IgG, and NS-albumin were measured. RW therapy caused a significant increase in ragweed-specific S-IgE (p less than 0.005) and NS-SIgA (p less than 0.05). PRW therapy caused a significant rise in ragweed-specific NS-SIgA (p less than 0.001). NS-IgE (p less than 0.05), and NS-IgG (p less than 0.01). Ragweed-specific S-IgG was not affected by any of the treatments. There was no consistent correlation between NS-antibody levels and symptom/medication scores.