Cox et al.: Measurements of resistance and reactance in fish with the use of bioelectrical impedance analysis 
37 
User experience 
To determine if untrained users produce different and 
more variable R and X c measurements compared to an 
experienced user, R and X c measurements were com- 
pared between a user with training and four users 
without training. In this experiment, the single trained 
user had taken over 5000 BIA measurements on fish and 
the untrained users had no experience or previous train- 
ing with BIA. Thirty juvenile coho salmon (n = 30, mean 
weight=9.7 g, SD=2.1) were killed and randomly split 
into five groups of six fish, placed in plastic bags, and 
covered with ice. Before the experiment, four untrained 
users were introduced to the concepts of BIA and also 
to the protocols used by Cox and Hartman (2005). They 
were allowed to observe the trained user take R and X c 
measurements on all six fish in a group. Each person 
was then randomly assigned a bag of fish. Immediately, 
each of the four inexperienced persons took R and X c 
measures on all six fish within a group. During the 
time of taking measurements, untrained users were not 
allowed to ask for assistance. All fish from all groups 
were measured within 1 hour. 
Time 
We examined the effect of the time between death of the 
fish and BIA to determine how long dead fish can be held 
on ice before R and X c measurements are compromised. 
Juvenile coho salmon (n = 60, mean fork length =99. 2 mm, 
SD=7.7 mm, and mean weight=10.3 g, SD=2.3 g) were 
killed and groups of six fish were randomly placed in plas- 
tic bags and placed on shaved ice. At 0, 3, 6, 9, 12, 24, 36, 
48, 60, 72, and 96 h, a bag of fish was randomly removed 
and all six fish were measured for length, weight, R and 
X c . It was assumed that because of the small size of the 
fish, the temperatures equilibrated within one hour 
of placement in ice and remained stable and therefore 
would minimize the effect of temperature on measure- 
ments. Measurements of R and X c were taken according 
to the procedures found in Cox and Hartman (2005). 
Temperature 
The effect of temperature on R and X c measurements 
was examined by taking repeated measurements on 
individual freshly killed fish over a range of tempera- 
tures (~0° to 12.5°C). The length of the experiment 
was <3 hours (the time it took the fish to freeze) and 
data in this study indicated that R or X c do not change 
significantly within that short period and should nul- 
lify confounding effects of time on body condition after 
death. Regression analysis was used to test whether 
slopes and intercepts differed from 0 for regressions 
of R and X c measurements on temperature. Adult pink 
salmon (n= 4, mean fork length-550 mm) and juvenile 
(« = 1, fork length-100 mm) Chinook salmon were killed 
and connected to a BIA Quantum-II Desktop by using 
standard needle electrodes and orientations as described 
by Cox and Hartman (2005). The Quantum-II was set 
to record impedance every 5 minutes for 12 hours. An 
ibutton thermometer (Maxim Integrated Products Inc., 
Sunnyvale, CA) was placed 1 cm (for juvenile) or 3 cm 
(for adults) inside the dorsal musculature of the fish 
and was set to record temperatures every 5 minutes. 
The juvenile fish was brought from cold (0.5°C) to warm 
(8.0°C) and the remaining adults were brought from 
warm (ambient water temperature) to cold (freezing). 
One adult had a starting temperature of ~12.5°C and the 
remaining three had a starting temperature of ~8.0°C. 
The automated Quantum II and the ibutton thermom- 
eters were synchronized to start recording at the same 
time. Each transferred fish was placed on a 4-in stand in 
the empty freezer compartment of a standard refrigera- 
tor. After 12 h, the fish was removed from the freezer, 
and R and X c measurements and temperature logs were 
downloaded onto a computer. The one juvenile fish was 
removed from an outside cold tank, killed, and placed 
on a standard laboratory bench at room temperature. 
Initial temperature for the juvenile fish was ~0.5°C. 
For regression analysis, only impedance measurements 
taken when the fish temperature was >0°C were used. 
Significance tests were performed on each fish to test for 
slopes = 0 by using a standardized major axis (SMA) test. 
The Bartlett-corrected likelihood ratio test (L) was used 
to test for differences between slopes of regression lines. 
Stomach fullness and electrode location on live fish 
To determine if R and X c measurements in fish are 
affected by stomach fullness and electrode location, R 
and X c values were taken and compared from locations 
on whole-body lengths and half-body lengths and with 
both full and empty stomachs. Measurements were taken 
across the length of the whole body (Al) and half its 
length (D) (Fig. 1A). Half- body length refers to the elec- 
trode orientation; one set of electrodes was placed toward 
the head region and the second was placed around mid- 
point of the fish (see D in Fig. 1A). Whole-body length 
refers to the second orientation that followed methods in 
Cox and Hartman (2005), where one set of electrodes was 
placed towards the head region and the second set was 
placed towards the tail region (see A in Fig. 1A). A two 
factor analysis of variance (ANOVA) was used to simul- 
taneously test for differences in half-body and whole- 
body R and X c measurements for brook trout with full 
and empty stomachs. If significance was found, a Tukey 
multiple comparison test was used to identify like values 
of R and X c . Brook trout (n= 20, length range=110-130 
mm) were randomly split into four groups: A) half-body 
length full-stomach; B) half-body length empty-stomach; 
C) whole-body length full-stomach; and D) whole-body 
length empty-stomach (i.e., five replicates for each of the 
four combinations). All sets of fish were starved for three 
days before the experiment to ensure that the stomach 
was empty. Within 2 hours of the start of the experiment, 
the full-stomach group was fed fly larvae ( Sarcophaga 
bullata) (Grubco Inc., Hamilton, OH), to satiation while 
those with the empty-stomach designation remained 
unfed. Before R and X c measurements were taken, fish 
