40 
Fishery Bulletin 109(1 ) 
User experience 
Untrained users produced different and more vari- 
able R and X c measurements than the experienced 
user. Pairwise tests indicated that one untrained user 
obtained a significantly different mean for R (LME 
t 2 o 30 >3, P<0.009) than the trained user (mean=915.35 
42) with means differing by -120.08 £2 (-13%) (Fig. 
3). Although differences were not significant, two of 
the untrained users had mean differences greater 
than 5% from the control. Significance was not found 
because of the high variance obtained from untrained 
users (mean deviation = 46.75) and when compared to 
the trained user, standard deviations from untrained 
users were 4.6 times larger. Reactance values were 
significantly different for the trained user compared 
to those for the three untrained users (LME t 20 30 >2, 
P<0.04) with differences ranging from -15.12 to 17.78 
£2 (-10% to 11%). Variability of X c standard deviations 
was greater (1.3x) in three of the untrained users 
when compared to variability in the control, but was 
not as great as the R variability (1.3x vs. 4.6x) (Table 
2, Fig. 3). 
Deviation 
Figure 2 
Boxplots of repeated resistance and reactance measure- 
ments taken from five pink salmon (Oncorhytichus gor- 
buscha ) (control) and when an additional variable was 
added to the procedure (deviation). Specific deviations 
were 1) switched wires: the signal and detector leads 
were switched; 2) salt: two cups of seawater (31 practical 
salinity units) were poured under the fish; 3) conductive 
board: the fish was placed on a stainless steel conduc- 
tive board; 4) needle size: the 28 gauge needles were 
replaced with larger 14 gauge hypodermic needles; and 
5) needle depth: BIA electrode needles were placed in the 
fish in the same orientation as that used in the control, 
except needles were inserted to a depth of 1 mm. Open 
circles (O) represent outliers determined by a Grubbs 
test. Closed circles (•) represent mean values. Different 
symbols indicate differences determined by the statisti- 
cal tests applied. 
Time 
The time period between death and BIA measurements 
did not affect R or X c measures on iced fish from 0 to 
72 h and from 0 to 9 h, respectively (Fig. 4, A and B). 
In this study, R differences between 0 and 72 h were 
not significant, LME (t 55 66 <2, P>0.16) (Table 2, Fig. 
4A). Although during the first 72 h, significant differ- 
ences were not detected, non-\ significant differences 
in mean R between 0 h (mean = 896.16 £ 2 ) and sub- 
sequent times (3-60 h) ranged from to 28.28 £2 (3%) 
at 3 h to -41.48 £2 (-5%) at 60 h (Fig. 4A). The mean 
deviation of all grouped R values was 47.81. Mean 
values of X c were not significantly different between 
0 and 9 h (LME t 55 66 <2, P> 0.17) (Table 2, Fig. 4B). 
Although differences were not detected, nonsignificant 
differences in X c means between 0 h and subsequent 
times (3, 6, and 9 h) were 7.15 £2 (4%), 7.85 £2 (5%), 
and 11.26 £2 (7%), respectively. Starting at 12 h, mean 
values of X c were significantly different from 0 h LME 
(f 55 66 >4, P> 0.001) and increased from 166.76 £2 (mean 
at 6 h) to 206.19 £2 (24%) (mean at 12 h). The mean 
deviation for all grouped X c values was 12.81. 
Temperature 
Temperature affected R and X c measurements. As 
temperature increased, R and X c decreased and slopes 
were not equal to zero (Fig. 5, A and B). Individual 
regressions of X c and R with temperature were sig- 
nificant and all individual regressions had r 2 >0.92. 
Individual slopes from each of the five R regressions 
were negative (-12.19, -11.64, -11.61, -10.95, and 
-12.01) and the mean slope was -11.65 (Fig. 5A). In 
the R regressions, there was no evidence of differences 
between slopes (L = 3.82, P=0.43). In the X c regres- 
sions, significant differences were found between slopes 
(L = 93.65, P<0.05) (Fig. 5B). Slopes from the each of 
the five X c regressions were negative (-4.86, -2.77, 
-2.87, -2.53, and -1.59). In X c regressions, the maxi- 
mum and minimum slopes were possible outliers and 
represented the only juvenile fish (closed circle symbol 
in Fig. 5B, slope=-4.86) whose temperature went in 
the opposite direction to that of the rest of the fish 
(i.e., cold to warm rather than warm to cold) and an 
adult fish (circle symbol in Fig. 5B, slope = -1.59) with 
an initial temperature 3°C higher than the others. 
When these two fish were removed from the regression 
analysis, the three remaining X c slopes were not found 
to be different from each other (L = 3.97 and P=0.14) 
(Table 2, Fig. 5B). 
Stomach fullness 
Stomach fullness did not affect R or X c measures in 
either half- or whole-body measures (Fig. 6, A and 
B). Differences in R measures were not significant 
between fish with full or empty stomachs for half- 
body measurements (Tukey HSD, P= 0.95) or full-body 
measurements (Tukey HSD, P=0.49) (Fig. 6A). Mean 
