Anxiety-related behaviors were evaluated across five tests in a sample of 277 rats from a genetically heterogeneous stock (N/Nih-HS rats), derived from an eight-way cross of inbred strains, and compared with the performance of RLA-I (high anxious) and RHA-I (low anxious) rats in the same tests. These tests either evoke unlearned (novel-cage activity (NACT), elevated "zero" maze (ZM), baseline acoustic startle response (BAS)) or learned (fear-potentiated startle (FPS), two-way active-shuttle box-avoidance acquisition (SHAV)) anxious/fearful responses. The results overall showed that unlearned anxiety responses/behaviors were predictive of behavior in learned fear (i.e. fear-potentiated startle) and conflict (i.e. two-way active avoidance acquisition) situations. Moreover, it was found that N/Nih-HS rats either resemble RLA-I rat anxiety/fear scores or fall in between those of the RLA-I (high anxious) and the RHA-I (low anxious) rat strains. An additional regression analysis (of N/Nih-HS rat data) showed significant positive influences of (unlearned) baseline startle response, risk assessment (i.e. stretch-attend) behavior and activity (5min) in a novel cage on SHAV acquisition, while baseline startle and entries into the open section of the elevated 'zero' maze test of anxiety were the main variables influencing FPS. This indicates that startle responses may have a facilitating role in the rat's active responses in the two-way active (shuttlebox) avoidance acquisition. The results of this behavioral evaluation of N/Nih-HS rats show that unconditioned anxiety (e.g. in the ZM test) predicts learned fear-related responses (e.g. FPS and SHAV) to some extent, while a positive association is also observed between BAS and SHAV. These findings are discussed in terms of their potential usefulness for present and future neurobehavioral and genetic studies of fearfulness/anxiety.