Maschner et al.: A 4500-year time series of Gadus macrocephalus size and abundance 
389 
Table T 
Comparisons of mean fork lengths of Pacific cod ( Gadus macrocephalus) caught by Aleuts from 1540 AD to 2550 BC compared to 
the mean fork lengths of fish captured in 2005 by longline surveys in the Gulf of Alaska. Comparisons were made by using L tests, 
and values in bold indicate no significant difference at the P<0.01 level. Sample size n is the number of measured bone elements 
from the prehistoric eras. The site numbers reflect individual winter village sites on Sanak Island from which the archaeological 
samples were derived. 
Mean fork length (cm) 
Site number 
Era 
Prehistoric 
Modern 
Lvalue 
df 
P 
n 
111 
2550 BC 
67.5 
67.5 
0.003 
26 
0.997 
27 
054 
1750 BC 
70.7 
67.5 
5.932 
573 
<0.001 
467 
036 
840 BC 
72.1 
67.5 
5.471 
340 
<0.001 
313 
061 
595 BC 
64.1 
67.5 
-1.181 
9 
0.267 
10 
058 
80 BC 
69.7 
67.5 
3.801 
601 
0.012 
507 
056 Upper 
520 AD 
65.5 
67.5 
-2.011 
175 
0.046 
166 
056 Lower 
1030 AD 
66.6 
67.5 
-0.981 
257 
0.327 
241 
110 
1540 AD 
62.8 
67.5 
-6.223 
515 
<0.001 
464 
Figure 2 
Box plots showing the fork-length (mm) distributions of Pacific 
cod (Gadus macrocephalus) caught in the Gulf of Alaska since 
2550 BC. The prehistoric distributions were reconstructed from 
skeletal elements (premaxilla and centra from trunk verte- 
brae) by using allometric relationships established by Orchard 
(2003). The modern length distributions were compiled from 
data collected from longline surveys conducted in the Gulf of 
Alaska by the Alaska Fisheries Science Center between 1978 
and 2006 (Thompson et al., 2006). 
Results and discussion 
Measuring changes in Pacific cod size 
and abundance 
The mean size of Pacific cod has varied consid- 
erably in the Gulf of Alaska over the last 4500 
years (Fig. 2; range 62.8-72.1 cm; F { 8 5503) =11.97, 
P<0.001). However, modern mean lengths did not 
differ significantly from the mean lengths of fish 
caught around 2550 BC, 595 BC, and 1030 AD 
(Table 1 ; Ltests P>0.05), or when compared with 
lengths of those taken around 80 BC and 520 AD 
(Table 1; Ltests P>0.01). 
Fecundities of Pacific cod (estimated from body 
length distributions) varied dramatically, ranging 
between 2.1 and 3.9 million eggs per individual 
(Fig. 3). Average fecundity varied consistently 
with changes in mean length as expected, but 
the apparent trend shown in Figure 3 towards 
decreasing fecundity over 4500 years was not 
significant (r=-0.42, P=0.35). 
The Al measures revealed changes in the rela- 
tive abundance of Pacific cod over the past 4500 
years (Fig. 4). All of the AIs were highly variable 
between 2550 and 2580 BC, but were synchro- 
nized after 520 AD. They indicated high relative 
abundance of Pacific cod in middens spanning 
2550-1750 BC and showed an opposite trend in 
the abundance of salmon, which was consistent 
with the inverse relationship previously noted 
for this region and time period (e.g., Tews, 2005; 
Misarti, 2007). From 840 BC to 80 BC another 
period of variability occurred when salmon Al 
was opposite that of the groundfish taxa. During this 
time, the AIs for Pleuronectidae and Hexagrammidae 
indicated a slight decline in relative abundance of Pa- 
cific cod, whereas the Cottidae index indicated a greater 
decline in Pacific cod abundance. Taken in tandem, 
however, all of the groundfish taxa recovered from the 
