340 
Fishery Bulletin 106(4) 
Table 2 
Composition of Chionoecetes bairdi (Tanner crab) and C. opilio (snow crab) populations tested and monitored in the field experi- 
ment. Values are means and ranges (in parentheses) for the nonrandom experimental population and do not represent the overall 
results of the fishing operations. Shell condition is a relative index of molt stage, ranging from molting or recently molted crabs 
with no hardening of the shell (0) to old and heavily encrusted shells (5). Reflex impairment index A is reported, representing the 
number of reflex actions that were entirely lost; this index can range from 0 to 6. Injury scores can range from 0 to 5, representing 
crabs with no visible exoskeletal damage to those with badly broken carapaces and limbs. 
Species and 
gender 
n 
Carapace width 
(mm) 
Shell 
condition 
Reflex 
impairment 
Injury 
scores 
Mortality 
(%) 
C. bairdi 
250 
29.2 
Female 
89 
79 (58-97) 
2.6 (2-4) 
0.94(0-6) 
0.85(0-5) 
34.8 
Male 
161 
98 (67-140) 
2.7 (1-4) 
0.73 (0-6) 
0.66(0-5) 
26.1 
C. opilio 
399 
24.3 
Female 
74 
63 (52-81) 
2.7 (2-4) 
1.40 (0-6) 
0.62 (0-5) 
23.0 
Male 
325 
83 (49-133) 
2.6 (0-4) 
1.42 (0-6) 
0.72 (0-5) 
24.6 
[Pacific cod], Atheresthes stomias [arrowtooth flounder], 
and other unidentified flatfishes, seastars, and crabs) 
was estimated and C. bairdi, C. opilio, and P. camts- 
chaticus were separated by species and gender into bas- 
kets. Air temperature during the crab handling process 
ranged 5° to 10°C. Sorting normally took <15 minutes 
and the sorting baskets were placed in fish totes with 
flowing seawater if the sorting and subsequent handling 
exceeded that time limit. The crabs were then tested 
individually for loss or weakness according to the six 
selected reflexes (loss=0, weak=l, strong=2) (Table 1), 
and notes were made on autotomy and obvious injuries 
such as broken legs, cracked carapace, or torn abdo- 
men. Later, injuries were scaled from 0 to 5, where 
0 = no injuries, l=newly autotomized legs, 2 = broken 
legs, chelae, or mouthparts, 3=minor carapace or abdo- 
men damage, 4=major carapace or abdomen damage, 
and 5=major damage to multiple parts (carapace, leg, 
and other parts). Crab gender, shell condition, and CW 
were recorded during the assessment period. Briefly, 
shell condition in Chionoecetes spp. was scored from 0 
to 5, where 0 represented molting or recently molted 
crabs with no hardening of the shell whatsoever, crabs 
with shell condition 1 represented a soft flexible shell, 
condition 2 represented full hardness, and scores 3 to 
5 showed increasing stages of discoloration and encrus- 
tation with shell age. Following assessment the crabs 
were either discarded overboard or tagged and held for 
monitoring mortality. Initially, all crabs were marked 
and held. However, it soon became apparent that a large 
proportion had relatively low reflex impairment. Thus, 
as the cruise progressed, emphasis was shifted to gear 
configurations producing greater damage to crabs, and 
we chose individuals with observable reflex impair- 
ment or injury for holding. As a result, the impairment, 
injury, and mortality rates reported in Table 2 do not 
represent the fishing conditions, only the crabs included 
in this analysis. 
Crabs held for monitoring were marked with uniquely 
numbered vinyl spaghetti tags tied securely but loosely 
around the basi-ischium of the third leg, or fourth leg 
(if the third leg had been autotomized). Tagged crabs 
were immediately moved to one of 12 large fish totes 
(98 x 110x85 cm deep; -900 liters) secured on the ship’s 
trawl deck. Each tote was supplied with a constant flow 
of seawater (>20 L/minute). Water temperature during 
8-11 day holding periods ranged from 2.1° to 5.9°C, 
and oxygen (monitored morning and evening in every 
tote) never fell below 100% saturation. Mortality was 
assessed and dead crabs were removed each afternoon 
for the first five days of holding, then every other day 
until the end of the experiment. On the last day of 
holding all of the remaining crabs were re-assessed for 
reflex scores. 
Reflex impairment indices and statistical procedures 
Scores for reflex actions were combined into impairment 
indices and used in the analysis. Composites provided 
robust indices of overall condition for the animal and 
provided the advantage of reducing the weight of any 
one reflex (Davis, 2007). Analyses described below were 
conducted with two different impairment indices. Reflex 
impairment index A was calculated as the total number 
of zero scores for individual reflex actions (i.e., lost 
reflexes), and reflex impairment index B was the total 
number of scores that were either 0 or 1 (i.e., lost or weak 
reflexes). Both index scores ranged from 0 to 6. 
Before analysis of the relationships between reflex 
impairment, injury, and mortality, we wanted to be 
certain that the holding period was sufficiently long to 
provide an accurate estimate for mortality. The time 
course for mortality was evaluated as a simple cumula- 
tive curve of deaths for C. bairdi and C. opilio shown 
as a function of time in days. We used simple linear 
regression to explore relationships among mortality-re- 
lated variables, injury scores, and reflex impairment. 
Logistic regression was used to model mortality, by 
using potential predictors and mediators including re- 
flex impairment, injury score, gender, size, and shell 
