Knowledge of the circadian patterns of serum IGF-I I and the large molecular weight IGF binding protein, IGFBP-3 might, apart from its physiological relevance, be of clinical interest, inasmuch as measurements of these parameters are being introduced into the evaluation of GH deficiency. We therefore evaluated the 24-h (08.00-08.00 h) patterns of serum IGF-II and IGFBP-3 in 8 GH-deficient patients who were studied during three periods when receiving 1. GH (2 IU) at 20.00 h; 2. GH (2 IU) at 08.00 h and 3. no GH. For comparison, 10 age- and sex-matched untreated healthy subjects were studied once under similar conditions. The serum IGF-II levels of the patients were relatively stable over the 24-h periods, yielding mean levels which were significantly lower during no GH: 553±78 (evening GH), 554±54 (morning GH), and 429±65 μg/l (no GH). The mean IGF-II level in the normal subjects was 635±29 μg/l, which was significantly higher than in either patient study. Similarly, stable 24-h levels of IGFBP-3 were recorded in all studies. The mean IGFBP-3 level of the patients was significantly lower when they received no GH, and the mean level in the healthy subjects was higher than in any of the patient studies: 1853±301 (no GH), 2755 ± 317 (evening GH), 2904±269 (morning GH), and 3856±186 μg/l (healthy subjects). However, minute but significant changes over time, characterised by slight decrements at night, were observed for both parameters in several of the studies. Nevertheless, since both IGF-II and IGFBP-3 display rather stable 24-h levels in the individual, it is concluded that measurements of these parameters in evaluation of growth retardation can be based on a single daytime sample.