RESUMEN
Non-target organisms in aquatic environments may experience lethal or sublethal effects following exposure to contaminants. Most protocols and regulations, however, are designed to provide protection from lethal effects and are thus based on conventional estimates of population lethality. The relative lack of reliable behavioral endpoints makes it challenging to implement regulations that are similarly protective against sublethal toxicity. The objective of this study was to quantify the avoidance behavior of Hyalella azteca when exposed to three insecticides-bifenthrin (B), chlorpyrifos (C), and permethrin (P)-at a range of estimated lethal concentrations. A two-choice behavioral arena was used for each chemical to quantify H. azteca activity and time spent in either uncontaminated sediment or sediment spiked at concentrations reflecting estimated 48-h lethal concentrations (LC50, LC25, and LC10). For all three insecticides, naïve H. azteca demonstrated a preference for the uncontaminated sediment over the contaminated sediment at the LC50 (B: 312 ng/gOC; C: 1265 ng/gOC; P: 5042 ng/gOC) and LC25 (B: 230 ng/gOC; C: 859 ng/gOC; P: 3817 ng/gOC), spending significantly more time in the uncontaminated side of the arena. H. azteca did not avoid sediment at LC10 (B: 204 ng/gOC; C: 609 ng/gOC; P: 1515 ng/gOC) levels, indicating the existence of a potential threshold of detection. Despite the lack of substrate preference at this exposure level, H. azteca were nevertheless more active (i.e., increased zone-switching) when exposed to bifenthrin at the LC10, suggesting a possible irritation response (e.g., movement after exposure) to this chemical. Our results provide evidence that H. azteca exhibit innate avoidance responses to sediments contaminated with common insecticides at concentrations below those represented by traditional toxicological endpoints (e.g., LC50). The sensitivity and ease with which this behavioral endpoint can be assayed demonstrates the potential utility of behavioral endpoints in toxicological assessments using model organisms.