17. Two-current choice flumes for fish chemosensory behaviour – method validation and limited effects of high pCO2

Fredrik Jutfelt (1), Josefin Sundin (2), Graham D Raby (3), Timothy D Clark (4)

1 Norwegian University of Science and Technology, Trondheim, 7491, Norway
2 Uppsala University, Uppsala, 751 24, Sweden
3 Great Lakes Institute for Environmental Research, University of Windsor, Windsor, ON, N9C1A2, Canada
4 University of Tasmania and CSIRO Agriculture Flagship, Hobart, 7000, Australia

Background
Ocean acidification has been suggested to disturb fish behavioural responses to chemosensory cues. Most experiments have investigated coral reef species, while the effects on species from other parts of the world are largely unknown. The methods used to quantify chemosensory behaviour in fish are variable and in need of standardisation.

Methods
Two-current choice flumes were constructed in a range of sizes to accommodate fish from 3 to 200 mm length. The flumes carry two parallel laminar water flows through an arena where the experimental animal can choose between the two flows, and the two flows can be manipulated (e.g. hypoxia, hypercapnia, prey/predatory cues). The methodology was validated using a range of flow, dye and pilot testing. To eliminate bias in behavioural observation we used automated video analysis.

Findings
Stable laminar flow can be difficult to maintain and can be disrupted by deviations in flow. Flow controllers, baffles and multiple layers of honeycomb collimators are vital components. Long observation times are needed to quantify side preference. Subconscious biases make objective observation difficult and blinded observation or automated video analysis is therefore needed.

Using the flumes we show that predator avoidance was high in both control and high CO2 (1000 μatm) exposed Atlantic cod. Furthermore, Atlantic cod strongly avoided high CO2 water, even after one month of acclimation to high CO2. In goldsinny wrasse, the predator avoidance was slightly reduced by high CO2 exposure compared to control.

Conclusions
We present a reliable flume methodology for measurements of chemosensory behaviour that provides stable laminar flow and unbiased behavioural quantification. We show that high CO2 appears to have limited effects on temperate fishes. We encourage the use of these approaches in all future studies to enable a comprehensive and robust understanding of any CO2 effects on the chemosensory behaviour of fish.