Cox et at: Measurements of resistance and reactance in fish with the use of bioelectrical impedance analysis 
39 
Table 2 
Resistance ( R ) and reactance (X c ) mean values, significance levels, and percent difference for experiments to determine effects 
of needle location, covariates, different users, time needed for measurements, and stomach fullness on R and X r measures. Each 
category has adjustments that may be a source of error when compared to a standard protocol (control) found in Cox and Hart- 
man (2005). 
Category 
Error source 
Mean 
Significance 
% Difference 
R(£2) 
X c (Q) 
R (Q) 
X e (Q) 
R 
Needle location 
A (control) 
306.62 
75.44 
— 
— 
— 
— 
Al 
300.80 
73.12 
0.58 
0.54 
2 
1 
B 
301.12 
69.44 
0.60 
0.13 
2 
2 
C 
221.80 
48.80 
<0.01 
<0.01 
-28 
-35 
D 
160.18 
61.64 
<0.01 
<0.01 
-47 
-18 
Covariates 
Control 
261.48 
70.34 
— 
— 
— 
— 
Switched wires 
262.38 
71.02 
0.87 
0.81 
<1 
<1 
Salt 
226.04 
64.24 
<0.01 
<0.01 
-14 
-9 
Conductive board 
108.96 
56.30 
<0.01 
<0.01 
-58 
-20 
Needle size 
236.38 
64.04 
<0.01 
<0.01 
-10 
-9 
Needle depth 
288.24 
103.62 
<0.01 
<0.01 
10 
47 
Different users 
Control 
915.35 
155.73 
— 
— 
— 
— 
1 
969.00 
156.33 
0.09 
0.93 
6 
<1 
2 
929.03 
173.52 
0.66 
0.01 
1 
11 
3 
857.13 
142.20 
0.07 
0.04 
-6 
-8 
4 
795.27 
140.62 
<0.01 
0.02 
-13 
-10 
Time (h) 
0 
896.17 
166.77 
— 
— 
— 
— 
3 
924.45 
173.92 
0.34 
0.38 
3 
4 
6 
900.85 
174.62 
0.87 
0.34 
<1 
5 
9 
909.87 
178.03 
0.64 
0.17 
2 
7 
12 
919.90 
206.20 
0.42 
<0.01 
3 
24 
Stomach fullness 
Half-empty 
321.00 
89.00 
0.95 
0.99 
-3 
1 
Half-full 
311.60 
89.60 
Completely empty 
830.40 
204.60 
0.49 
0.14 
3 
9 
Completely full 
805.20 
185.20 
or X c (LME t 16 25 <1, P >0.55) (Fig. IB). Although dif- 
ferences were not statistically significant, differences 
in mean R values between location A (mean=306.62 Q) 
and locations Al and B were -5.82 Q (-2%) and -5.50 
Q (2%), respectively. Similarly, nonsignificant differ- 
ences in X c means between location A (mean=75.44 Q) 
and locations Al and B were -2.32 Z2 (-1%) and -6.00 
Q (-2%), respectively. The anatomical locations of C 
and D differed the most from A, whereas locations Al 
and B had the greatest similarity (Fig. 1A). Location C 
represented the entire length of the ventral region and 
D represented the forward half of the dorsal region of 
the fish. The location of Al represented the same area 
of the fish as in A, except there was a second puncture, 
and location B represented an area slightly below A, 
but was still within the dorsal musculature (Fig. 1A). 
Procedure deviations 
Some procedural deviations from those outlined in Cox 
and Hartman (2005) significantly affected R and X c 
measures. Specifically, changes in needle depth and 
size, and placing the fish on a conductive surface or 
on salt water caused significant changes in R and 
X c , (LME t 2 0 30 > ^’ B<0.001) (Table 2, Fig. 2). Ranked 
differences in R means and percentages from highest 
to lowest between the control and deviations were as 
follows: 1) -152.52 Q, -58% for fish placed on a conduc- 
tive board; 2) -35.44 Q, -14% for fish placed on salt 
water; 3) 26.76 12, 10% for shallow needle depth; and 
4) -25.10 12, -10% for the larger needle size (Table 2, 
Fig. 2). Ranked differences in X c means from highest 
to lowest and percent differences between the control 
and assorted covariates were as follows: 1) 33.28 12, 
47% for shallow needle depth; 2) -14.04 12, -20% for 
fish placed on a conductive board; 3) -6.30 12, -9% 
for the larger needle size; and 4) -6.10 12, -9% for 
fish placed on salt water (Fig. 2). Differences were 
not significant between the control means of R and 
X c and switched wires for either R (mean R=262.3812, 
LME t 20 30 <0.3, P>0.80), or X c (mean X c =71.02 12, LME 
1 2 0 3 o < 0.5, P>0.70) (Fig. 2). 
