Experimental Studies of Electroshock Effects on Fish Require In-Water Measurements and Fish Threshold Observations to Achieve Electrofishing Context: Response to Comment

James B. Reynolds’ comment (Reynolds 2015, this issue) and the answers to the questions that he poses could make our study (Stewart and Lutnesky 2014) more relevant to an electrofishing (EF) context. However, that was not the purpose of our study per se. The purpose was to test whether sublethal electroshock (ES) could influence reproduction in a species of fish. We accomplished this and provided the relevant data for replication of our study. We did not, however, design our study from an EF perspective with the application of power density and determination of power density thresholds of response, nor do we think we should have in this first step in understanding that ES can affect reproduction. While we think that a power transfer approach has merit, it also has shortcomings if one has to make major assumptions (see below). Testing these assumptions to properly use this approach was beyond the scope of our paper, just as taking blood samples to test for acidosis, which is also known to be involved with EF (e.g., Mitton and McDonald 1994) and may be quite relevant to our observed retardation of reproduction, would be. We do agree that tests of the influence of ES on reproduction should move in both these directions, but our initial study was not designed to do this. Nonetheless, Reynolds makes it clear that he thinks the work is important and valid despite its shortcomings from an EF perspective.

Where possible, we are pleased to provide more information to make our work more relevant to EF objectives, but we realize that our laboratory study on a small, nongame species under controlled conditions makes it less relevant to field operations involving ES with game species. We used a two-factor design (Zar 2010) involving the factors of voltage (V [100 or 400]), frequency (Hz [60 or 120]), and their interactions to test for the effects of ES on reproduction in Red Shiners Cyprinella lutrensis (see Stewart and Lutnesky 2014 for the variables tested). All combinations of these factors elicited immobilization by design. We considered immobilization the point of using ES in EF (in most circumstances), but finding the minimal settings for a capture-prone fish response threshold (immobilization) was not one of our objectives. We purposely chose the combinations of settings to be over the threshold yet potentially minimal in effect (e.g., 100 V, 60 Hz; personal observation) or potentially harmful to reproduction (e.g., 400 V, 120 Hz), if an effect was to be found. At the same time, none of the combinations could obviously harm the fish (e.g., broken backbones, death, etc.). The fish had to “appear” unharmed, as is sometimes the case even when fish injury occurs. Note that all surviving fish went on to spawn and that only 6 of the 250 fish died after treatment (Stewart and Lutnesky 2014).

To make the paper more relevant to EF activities, Reynolds converts our treatment values to voltage gradients and calculates the power transferred to the fish (D_m). He concludes that the 100-V setting was near the threshold (we can only state that it was over the threshold), but that the 400-V setting resulted in a D_m value that was well above the injury threshold for game fishes and would probably not be realistic given our high ambient-water conductivity. While the D_m values calculated in the comment are a good first approximation, we view them as only a rough guess. They depend on q values that represent the ratio of the ambient conductivity of the water to the conductivity of the fish (see comment). Using a general value of 115 µS/cm (Miranda and Dolan 2003) or a recommended general value of 100 µS/cm (Reynolds and Kolz 2012) is likely a good starting point from which to estimate D_m for fish in general, e.g., to minimize error if one were modeling a general response to EF. However, such general values are dangerous assumptions in dealing with the responses of a particular species given the 56–204 (or more) µS/cm range known for fishes (Miranda and Dolan 2003; Reynolds and Kolz 2012; comment). Given an unknown conductivity for Red Shiners, this range can result in calculated D_m values that are different by more than a factor of two at the same voltage (Table 1), i.e., D_m values close to the thresholds for injury (Miranda 2005) or those thought to be unrealistic (see comment). While 115 µS/cm is a good starting point for a general value, Miranda and Dolan (2003) warn that it is based on only a few species and that more estimates of fish conductivity are needed to validate it as a working standard. Furthermore, error rates are exacerbated in higher-conductivity waters (Miranda and Dolan 2003). One could make a large error by not knowing the actual fish conductivity (i.e., the D_m values in Table 1 of the comment should be changed to “unknown” until we have measurements of conductivity for these fishes).

Regardless of the actual D_m value in our study, we know (de facto) that even the highest value was realistic because fish response was implicit in the design of our experiment. In the ES treatments, the fish had to spawn and then continue to spawn after ES to be included in the study (this was clear in the paper). A setting with an unrealistically high D_m value, as concluded in the comment, would have caused obvious injuries and mortality. Either (1) Red Shiners have low conductivity, so that even the high D_m value was only at or near the threshold of obvious injury for game fishes (Table 1) or (2) the D_m threshold for obvious injury is much higher for Red Shiners than it is for game fishes. Either makes the settings that we used realistic. In any case, the delays in reproduction were only on the scale of a few days even for the fish exposed to the higher D_m values, whatever they were.

So, what can be considered “unrealistic”? If unrealistic means that the D_m values were far above the D_t (immobilization) values, i.e., that we transferred more power to the fish than was necessary for an immobilization response, this was implicit in the design of the study by having more than one setting (even if in one treatment D_m = D_t, the other would necessarily result in D_m > D_t if both caused immobilization). We were interested in determining whether nontarget fishes, that is, fishes for which the settings are not optimized, would be harmed by ES; this was a major point in the paper. If unrealistic means that the D_m values were far above what would give obvious injuries and cause many deaths, this was simply not the case for the Red Shiners in our study. Finally, if it implies an unrealistic setting for a game fish, one should consider a proportional response. That is, what proportion of the D_m value that would cause obvious injuries and mortality was used? If game fishes are injured at lower D_m values, they may also have retarded reproduction at some fraction of those lower values.

We used a higher voltage, knowing that the lower-voltage treatment already caused immobilization, to see whether ES can have a negative influence on reproduction because settings are not always correct in the field or different techniques are used (e.g., Cooke et al. 1998). Furthermore, nontarget, nongame fishes that are at, above, or below their threshold may be in the electric field with a target species at its threshold. Until our study, no one knew whether any of these settings would impact reproduction in a fish species since this was the first study of its kind to address the potential influence of ES on reproduction; we needed to start somewhere.

We reported the percent changes in weight and length (e.g., growth) of the fish from the treatments, i.e., how they fared after ES exposure (see Figures 2 and 3 in Stewart and Lutnesky 2014), but not the absolute length values that are important from an EF efficiency and injury perspective (Reynolds and Kolz 2012). The pretreatment size range for the 121 males used for analysis was 32.4–59.2 mm SL (0.63–4.77 g), while that for the 123 females was 28.2–58.6 mm SL (0.45–3.82 g). Additionally, the duty cycle (percentage of time that the power is on during a cycle) of the frequencies that we used during ES exposure was constant in our experiment. The duty cycle of the Smith-Root Model 12 Backpack Electrofisher unit used in our study is nonadjustable, and the Model 12 that we used has a duty cycle of 25 ± ∼4% across all frequencies (A. Hendrix, Smith-Root, Inc., personal communication; verified by oscilloscope), a value that Reynolds and Harlan (2011) find to be in the optimal range for EF. We acknowledge that the use of electrofishers with fully adjustable output waveforms would allow the duty cycle to be varied and that duty cycle may be an important factor in future studies concerning the influence of ES on reproduction.

Finally, we would like to thank Reynolds, a major contributor to fisheries professionals’ knowledge of ES and EF and their effects on fishes, for his interest in our paper. We again raise more questions than we answer, and we anticipate that our paper, the comment, and the response will stimulate more work on the influence of ES and EF on reproduction in fishes.