Yochum et ai. A comparison of methods for evaluating mortality of discarded Cancer magister 
135 
Female-hard 
03 
03 
o 
100% 
37 
111 
I 
• 
80% 
60% 
40% 
20% 
0% 
Female-soft 
515 436 724 712 
Male-hard Male-soft 
209 
0.15 
0.12 3 -. 
0.09 „ 
S3 
CQ 
0.06 3 
CD 
0.03 CL 
0.00 
Calendar month 
Figure 4 
Composition of Dungeness crabs (Cancer magister) that were 
tagged and released during commercial fishing trips off Oregon 
between October 2012 and April 2014 and for which tags were 
returned, by combination of sex and shell hardness (bars), and 
proportion of crabs for which tags were returned (dots) by the 
numeric calendar month in which the tagged crabs were re¬ 
leased (from December to July). The numbers above the bars 
indicate the total number of tags released by month. 
higher return rates for noninjured crab than injured 
(all combined: 11% noninjured, 3% injured). This was 
consistent for both sexes (males 12% vs. 4%, females 
7% vs. 0%); however, there were only 25 injured of 
tagged crab and the average reflex impairment score 
was higher for injured animals, suggesting that, as 
with the commercial fishery, the decrease in probability 
of return for injured crab was likely represented by the 
reflex impairment score. 
Relative short-term survival 
The ratios of short-term survival rates of condition 
2 (score greater than 0) to condition 1 (score-0) crab 
were highly variable among release events for the tag- 
return study (Fig. 6) but indicated minimal differences 
in survival between conditions. Some release events in¬ 
dicated that condition-1 crab had higher survival than 
crab classified as condition 2, and others the opposite. 
Regardless, for all release events for females and hard¬ 
shell males, the overlapping CIs indicated no statisti¬ 
cal difference between conditions. For soft-shell males, 
there was only one release event with enough data to 
calculate a relative survival rate. That event indicated 
survival was higher for condition-2 crab. 
Ratios of relative short-term survival of condition-2 
to condition-1 crab in the laboratory holding experi¬ 
ment indicated no difference between conditions in 
some cases and, in others, that condition -1 
crab had higher survival. For all fisheries, the 
relative survival ratio for hard-shell males was 
either equal to or very close to 1, indicating no 
differences in survival between conditions. For 
females, 95% CIs of the ratios overlapped with 
1 for the recreational fisheries, but the ratio 
for the commercial ocean fishery indicated 
higher survival for condition-1 crab. For both 
the commercial ocean and recreational bay by¬ 
boat fisheries, the ratios indicated higher sur¬ 
vival for condition -1 soft-shell male crab than 
for condition-2 soft-shell male crab. The 95% 
CIs of these ratios, however, included a value 
of 1. 
Relative long-term survival 
Model results indicated that, for all release 
events where there were adequate sample siz¬ 
es to complete the analysis, the intercept and 
variable coefficients were not significantly dif¬ 
ferent from zero. When back-transformed from 
the logit scale and plotted, however, the inter¬ 
cepts were consistently at or above 0.5 (i.e., 
the proportion of tag returns from condition-2 
crab were 50% or greater, even though there 
were higher numbers of tags released for con¬ 
dition-1 crab). The estimated slope coefficients 
for the majority of the release events, although 
not significant, were negative (i.e., the log-odds 
of a returned tag being from a condition-2 crab 
decreased over time). If significant, this would indicate 
a chronic difference in survival between conditions, 
namely that probability of a tag return, and therefore 
survival, for condition-2 crab could decrease over time 
relative to condition-1 crab. There were no consistent 
patterns, however, among release events that would 
signify an optimal monitoring duration beyond the 5 d 
“recovery period.” Also, the lack of significance in the 
estimated model coefficients means there were no de¬ 
tected changes over time in the relative probability of a 
tag being returned from a condition-1 or -2 crab. 
Evaluating the return to water 
For the first drop experiment, 21 crab were dropped 
from 1 m, 22 from 3 m, and 20 from 8 m. There were 
only 7 soft-shell males and 6 soft-shell females dropped 
for all heights combined, and 81%, 73%, and 70%, re¬ 
spectively, by drop height were male. By drop height, 
0% (1 m), 5% (3 m), and 45% (8 m) died within 5 d of 
holding. The one crab that died from the 3-m drop was 
a soft-shell female that had incurred both a broken leg 
and an autotomized leg from the drop. From 8 m, 6 of 
the 9 mortalities were male, and 3 were female (43% 
of males died, and 50% of females). All of the crab that 
died for this experiment, except one, had a cracked 
carapace as a result of the drop. 
For the second experiment, a total of 18 crab were 
