Experiment 3: Lambs (n = 20/group), which were from ewes that were naïve to ovalbumin, had received one of the following treatments: 1) control injection (open squares, solid line; 1 mL of adjuvant + 1 mL of sterile isotonic saline) on d 1 of age + control injection on d 15 of age; 2) ovalbumin injection (open circles, solid line; 12 mg of ovalbumin + 1 mL of aluminum hydroxide gel adjuvant + 1 mL of sterile isotonic saline) on d 1 + ovalbumin injection on d 15; 3) control injection (closed squares, dotted line) on d 28 + control injection on d 42; and 4) ovalbumin injection (closed circles, dotted line) on d 28 + ovalbumin injection on d 42. Also, at an average age of 159 d, which was soon after weaning, lambs received either a control or an ovalbumin injection that was consistent with their original injection type. All injections were subcutaneous. Jugular blood samples were collected immediately before the d-159 injections and at weekly intervals for the next 4 wk to quantify antiovalbumin IgG (OV-IgG). Serum from each sample was diluted 1:100 in PBS; 100 μL of diluted sample was assayed; and data were expressed as optical density units (odu). Injection type (control vs. ovalbumin; P < 0.0001), but not initial injection schedule (d 1 and d 15 vs. d 28 and d 42 of age; P = 0.59), affected OV-IgG. The injection type × injection schedule interaction was not significant (P = 0.08). Time after injection (P < 0.0001) affected OV-IgG, and the injection type × time interaction was significant (P < 0.0001). Injection schedule × time and injection type × injection schedule × time were not significant (P = 0.86 and 0.68, respectively). Values are least-squares means, with a pooled SE of 0.013 odu.