Experiments were conducted on the grand Banks of the north Atlantic ocean during 2002-2003 to evaluate the effect of circle hooks and mackerel bait on pelagic longline catches and bycatch, and compare the treatments to the industry standards at the time. Circle hooks were 18/0, and both non-offset and 10° offset were used as separate treatments and compared to the 9/0 J-hook control hooks with 10°-30° offset. Bait treatment was mackerel (Scomber scombrus linnaeus, 1758), which was compared to the squid control (Illex spp.). We analyzed the effects of hook (one J- and two circle hooks), bait (mackerel and squid), temperature, soak time, and animal length on anatomical hooking location for seven fish species and two sea turtle species. We also analyzed the effects of the same variables, inclusive and exclusive of hooking location, on the odds of boating a dead fish. We found that hook was one of the most important variables in predicting anatomical hooking location, and that soak time and hook and/or anatomical hooking location were important in predicting the odds of observing a dead animal boatside. The importance of the other variables differed by species, and for several species no models were significant for predicting hooking location or for predicting observed mortality.