Effects of different durations of cold stress (CS) and the time point of immunization relative to the CS of 3 chicken lines were studied. The first 2 chicken lines were divergently selected for high and low antibody responses, and the third was a random-bred control line. In 2 experiments, 26-d-old growing chicks of the 3 lines were feed restricted at 80% of ad libitum and subjected to CS of 10 degrees C for 7, 2, or 0 d. Birds were immunized with keyhole limpet hemocyanin (KLH) at -1, +1, +3, +5, or +7 d relative to the end of the CS treatment. Specific antibodies to KLH were determined. In addition, in vitro lymphocyte proliferation responses to concanavalin A (ConA) and KLH as measures of cell-mediated immunity, production of zymosan-induced reactive oxygen intermediates as a measure of phagocytosis, and BW gain as a measure of a production trait were determined. Significantly higher antibody responses to KLH were found in the high line as compared with the other 2 lines. Specific antibody responses to KLH were not significantly affected by the CS treatments. CS had a delayed effect on in vitro mitogen responses to ConA. In vitro lymphocyte proliferation responses to ConA were higher in the low line birds than in the other 2 lines. In general, 7 d of CS significantly enhanced cellular immunity to ConA in vitro, whereas the 2-d CS treatment had differential effects on lymphocyte proliferation to ConA, depending on the line of bird and the time of immunization. KLH-specific lymphocyte proliferation was enhanced by 2 d of CS at 28 d after immunization. Effects of various CS treatments and the time of immunization on the production of reactive oxygen intermediates were inconsistent. In addition, BW gain was negatively affected by CS. We concluded that the innate part of the immune system (phagocytes) responded immediately to CS with an as yet unexplained variability, irrespective of the genetic background. When CS was prolonged, the cellular adaptive immune response and, to some extent, the specific humoral immune response were also affected. The lack of line-by-treatment interactions suggested that the genetic background was a prominent factor for the magnitude of the specific immune response. Our data confirmed earlier studies that, under restricted feeding with simultaneous stress (energy demand for thermoregulation and growth), cellular immunity is more sensitive than humoral immunity. A negative correlation between BW gain and cellular immunity suggest a trade-off between these 2 life traits.