194 
Fishery Bulletin 117(3) 
Table 3 
Range of number of individuals caught per set, mean number caught per set, and number of sets identified as outliers for species 
caught on longline sets baited with Atlantic mackerel (Scomber scombrus) or northern shortfin squid (Ilex illecebrosus ) in the 
northern Gulf of Mexico during March-April 2017. Values provided in parentheses represent number of individuals caught in 
longline sets identified as outliers or P-values associated with % 2 values from chi-square tests. Note that no catches of Gulf smooth- 
hounds (Mustelus sinusmexicanus) or red snapper (Lutjanus campechanus ) were identified as outliers. Therefore, no associated 
expected ratios or % 2 values are presented for either species. SD=standard deviation. 
Species 
Bait type 
No. per set 
Mean + 3 SD 
No. of outliers 
Expected ratio 
x 2 
Gulf smoothhound 
Mackerel 
1-8 
9.81 
0 
Squid 
1-7 
0 
Atlantic sharpnose shark 
Mackerel 
Squid 
1-50 
1-31 
42.22 
2 (43, 50) 
0 
0.96:1 
12.14 (<0.01) 
Blacknose shark 
Mackerel 
Squid 
1-4 
1-5 
4.77 
0 
1(5) 
1.02:0 
4.92 (0.03) 
Blacktip shark 
Mackerel 
Squid 
1-7 
1-3 
5.90 
1(7) 
0 
0.98:1 
0.02 (0.87) 
Scalloped hammerhead 
Mackerel 
Squid 
1-5 
1-6 
5.15 
0 
1(6) 
1.02:0 
0.01 (0.92) 
Red snapper 
Mackerel 
Squid 
1-15 
1-18 
22.50 
0 
0 
- 
- 
sets and alternating-bait-type sets. A valid criticism of 
Driggers et al. (2017) is related to possible interaction 
effects when bait type is alternated on a single set, as 
suggested by Watson et al. (2005) and Foster et al. (2012). 
Although this concern is reasonable, if an interaction 
effect occurred, a statistically equal number of individu¬ 
als of a given species would be expected to be caught on 
each bait. The results of Driggers et al. (2017) indicate 
significant bait type preference for 5 shark species and a 
clear trend for 3 other species; therefore, we believe the 
concern is unwarranted. However, the results of our study, 
for which we used a single-bait approach, indicate that 
in the absence of choice, most, but not all, species that 
Driggers et al. (2017) found demonstrated a bait prefer¬ 
ence will ingest whichever bait is available. Therefore, we 
conclude that studies examining the effect of bait type on 
catch rates with longline gear should use both single-bait 
and alternating bait approaches to determine the degree 
of bait preference and if a specific bait affects catch rates 
in such a way that would warrant management measures. 
Further, bait comparison studies that employ a single-bait 
approach must demonstrate that all bait types were fished 
in comparable abiotic and biotic conditions. 
Results of our study indicate that it is important that 
bait type be accounted for when analyzing fishery- 
dependent and fishery-independent catch data rather than 
assuming there is no difference in catch rates among dif¬ 
ferent baits. Although this assumption would have little 
to no effect when examining a species with no bait prefer¬ 
ence, in certain cases, such as when examining trends in 
abundance of Atlantic sharpnose sharks, significant bias 
would be introduced. Additionally, there can be a cascad¬ 
ing effect whereby captured fish become bait themselves, 
therefore introducing unexpected biases for other species 
showing preferences. For example, during the Mississippi 
Laboratories survey, Atlantic mackerel has been the sole 
bait used since 1995. During this time, 26% of all bull 
sharks (C. leucas) have been caught on hooks occupied by 
captured Atlantic sharpnose sharks (senior author, unpubl. 
data). If a change were made from Atlantic mackerel bait 
to northern shortfin squid bait, on the basis of the results 
of this study, fewer Atlantic sharpnose sharks would be 
expected to be captured, and in turn fewer bull sharks, a 
species whose diet has been reported to consist to a large 
extent (35.4%) of smaller sharks (Cortes, 1999), would be 
captured. 
In addition to indicating the importance of considering 
bait type in analyses of catch data, the results of this study 
demonstrate that the use of squid as bait can reduce the 
catch of at least one shark species in the northern Gulf 
of Mexico and could be employed as a bycatch reduction 
measure. However, some protected species, such as the 
loggerhead (Caretta caretta ) and leatherback ( Dermoche- 
lys coriacea) sea turtles, have been shown to have higher 
bycatch rates on squid-baited hooks (e.g., Watson et al., 
2005). Therefore, the effect of a specific bait across all taxa 
that are directly or indirectly affected by a particular gear 
type must be assessed before adopting any bycatch reduc¬ 
tion measure. 
Acknowledgments 
We thank the crew of the NOAA Ship Oregon II, R Felts, 
E. Hoffmayer, N. Hopkins, C. Jones, L. Jones, K. Rademacher, 
C. Stepongzi, and T. Wallace for their field assistance. 
