We have investigated possible (negative) feedback and (positive) feed-forward activity within the human male gonadotropic axis by measuring serum concentrations of LH, FSH, and testosterone in blood sampled frequently and for a prolonged interval (every 20 min for 19 h) simultaneously from the peripheral circulation and the left spermatic vein. Cross-correlation analysis with time lag was used to evaluate relationships among serial serum LH, FSH, and/or testosterone concentrations over time (i.e. consistency or dissociation of trends in concentrations). Separately, Cluster analysis was applied to identify discrete LH, FSH, and testosterone pulses, which were cataloged for possible peak coincidence. The hypergeometric probability distribution was then used to test the null hypothesis that LH, FSH, and testosterone pulses are randomly associated. Cross-correlation analysis revealed: 1) peripheral blood LH and testosterone concentrations correlate positively at lags of 40-120 min with LH increases preceding testosterone increases, viz., feed-forward (P < 0.001); 2) LH and FSH concentrations in peripheral blood are positively correlated in simultaneous blood samples, as well as when FSH lags LH by 20 min (P < 0.01); 3) unexpectedly, LH and FSH concentrations in peripheral blood are inversely related at a lag of 80-100 min (P = 0.002) and 0.004, respectively) where LH lags FSH; 4) LH and testosterone concentrations in the spermatic vein show strongly positive correlations at lags of 80, 100, and 120 min (P = 0.002, 0.004, and 0.021, respectively); 5) spermatic vein testosterone concentrations correlate negatively with peripheral blood LH concentrations 20 or 40 min later (P = 0.012 and 0.05, respectively), which indicates autonegative feedback; and 6) in contrast, testosterone levels in the spermatic vein correlate negatively with FSH values in the periphery 100 and 120 min later (P < 0.01), indicating more delayed negative feedback of testosterone on serum FSH concentrations. Discrete pulse coincidence analysis disclosed: 1) a total of 30 testosterone pulses in the spermatic vein and 25 testosterone pulses in peripheral blood, with 28 LH and 29 FSH pulses in the periphery; 2) individual LH and FSH peak concordance was significantly nonrandom for FSH pulse maxima lagging LH pulse maxima by 20 min (P < 0.05 vs. randomness), with 6 observed coincidences vs. 2.9 +/- 1.5 (SD) expected; 3) peripheral LH pulses and spermatic vein testosterone pulses were strongly nonrandomly coupled at an 80-min lag, with 8 events observed vs. 3.0 +/- 1.5 events expected (P = 0.004); and 4) LH peaks in peripheral blood followed testosterone peaks in the spermatic vein by 40 min in a nonrandom manner, specifically, n = 11 observed vs. 3.0 +/- 1.5 expected (P < 0.001), indicating possible LH escape from testosterone's negative feedback. In summary, physiological regulation of the human male LH, FSH, and testosterone axis comprises multidirectional interactions, consisting of both (positive) feed-forward and (negative) feedback coupling. Based on a concept of network integration, we propose that age and other pathophysiological factors might modulate and/or disrupt these dynamic within-axis multihormonal linkages.