Neurotoxicity of tricresyl phosphates (TCPs) and jet engine oil (JEO) containing TCPs were evaluated in studies conducted in both rat and hen. Results for currently produced samples ("conventional" and "low-toxicity") were compared with published findings on older samples to identify compositional changes and relate those changes to neurotoxic potential. Finally, a human risk assessment for exposure by oral ingestion of currently produced TCPs in JEO at 3% (JEO + 3%) was conducted. TCPs and certain other triaryl phosphates administered as single doses inhibited brain neuropathy target esterase (B-NTE; neurotoxic esterase) in the rat and the hen (hen 3.25 times as sensitive), and both species were deemed acceptable for initial screening purposes. Neither rat nor hen was sensitive enough to detect statistically significant inhibition of B-NTE after single doses of IEO + 3% "conventional" TCP. Subacute administration of 2 g/kg/d of JEO + 3% "conventional" TCP to the hen produced B-NTE inhibition (32%), which did not result in organophosphorus-induced delayed neurotoxicity (OPIDN). Subchronic administration of JEO + 3% TCP but not JEO + 1% TCP at 2 g/kg/d produced OPIDN. Thus, the threshold for OPIDN was between 20 and 60 mg "conventional" TCP/kg/d in JEO for 10 wk. The current "conventional" TCPs used in JEO and new "low-toxicity" TCPs now used in some JEO are synthesized from phenolic mixtures having reduced levels of ortho-cresol and ortho-xylenols resulting in TCPs of very high content of meta- and para-substituted phenyl moieties; this change in composition results in lower toxicity. The "conventional" TCPs still retain enough inhibitory activity to produce OPIDN, largely because of the presence of ortho-xylyl moieties; the "low-toxicity" TCPs are largely devoid of ortho substituents and have extremely low potential to cause OPIDN. The TCPs produced in the 1940s and 1950s were more than 400 times as toxic as the "low-toxicity" TCPs produced today. Analysis of the doses required to produce OPIDN in a subchronic hen study suggests that the minimum toxic dose of "conventional" TCP for producing OPIDN in a 70-kg person would be 280 mg/d, and for JEO containing 3% TCP, 9.4 g/d. Food products could be inadvertently contaminated with neat "conventional" TCP but it is unlikely that food such as cooking oil would be contaminated with enough JEO + 3% TCP to cause toxicity. Further, at the dosage required for neurotoxicity, it would be virtually impossible for a person to receive enough JEO + 3% TCP in the normal workplace (or in an aircraft) to cause such toxicity. There is no record of a JEO formulated with the modern "conventional" TCP causing human neurotoxicity.