The objective of this research was to measure the load history effects of negative stress ratio loading on the high cycle fatigue (HCF) crack growth threshold of Ti-6Al-4V. Previous work has shown an increase in the HCF threshold that is dependent on the K_max during the low cycle fatigue loading that nucleated a crack. The increase in fatigue limit was attributed to an overload effect from the LCF loading and was quantified through the use of a simple overload model. On the contrary, evidence from R = -1 testing suggests that negative overloads, referred to as underloads, may reduce this K_max load history effect and even lower the threshold below long crack values. To investigate this further, smooth and notched Ti-6Al-4V specimens were pre-loaded in fatigue below their endurance limit at stresses expected to nucleate cracks in approximately 10 million and 100 000 cycles, respectively, and at stress ratios of -3.5 and -3, respectively. Although the smooth specimens could not be monitored for crack nucleation, the notched specimens allowed the use of an infrared damage detection system to monitor the localized region at the notch root for indications of crack nucleation. These cracked specimens were then heat tinted, and several were also stress relieved to remove load history effects. All of the preloaded specimens were then HCF step tested to determine the fatigue limit stress or threshold. Although the smooth bars showed little effect due to the preloading at negative R, the threshold results on the notched specimens that developed measurable cracks seem to show competing effects of underloading and K_max overloading, dependent somewhat on precrack length. The specimens with smaller cracks nucleated at R = -3 tend to have a reduced HCF threshold compared to the conventional long crack threshold. Short crack effects and load-history effects are quantitatively explained with the aid of a Kitagawa diagram with an El Haddad short crack correction.