Maschner et al.: A 4500-year time series of Gadus macrocephalus size and abundance 
391 
Figure 4 
Abundance indices (e.g., AI=Pacific cod / [ comparison taxon + 
Pacific cod]), calculated from the number of identified specimens 
(bone fragments) for each taxon recovered from nine archaeologi- 
cal deposits on Sanak Island. The index ranges between 0 and 
1, with 0 indicating the complete absence from the middens and 
1 indicating a complete dominance of Pacific cod ( Gadus macro- 
cephalus). Sample sizes from the deposits ranged between 426 and 
8339 identified bone specimens. All indices indicate a decline in 
Pacific cod abundance beginning at 520 AD, except for the Gadi- 
da e-Oncorhynchus spp. index, which is included to show the differ- 
ence between riverine ( Oncorhynchus spp.) and nearshore marine 
fish (the other families of fish represented in the graph) during 
this period. This decline in abundance continues throughout the 
period until 1030 AD, after which we see a recovery in Gadidae 
abundance at ca. 1540 AD. The correspondence in the multiple 
measures negates the mathematical conundrum of closed arrays 
and indicates real changes in the relative abundance of Pacific 
cod in Sanak middens. 
single preserved Pacific cod bone dating from 
AD 1100 to 1300 in the Sanak archaeological 
data has been recovered despite the presence 
of village sites. The period between AD 1100 
and 1300 were the only centuries during the 
last 5000 years of extensive global warming 
that occurred before the modern era. Pacific 
cod returned in the middens only during the 
hemispheric cooling associated with the Little 
Ice Age. Interestingly, regardless of periodic 
regime shifts in ocean climate, Pacific cod 
populations appeared to have returned to the 
western Gulf of Alaska with approximately 
the same length structure as that before they 
disappeared. 
The perturbations in the body lengths of 
Pacific cod did not correlate with changes in 
their frequencies of occurrence in the middens 
(Fig. 4). For example, the most extreme shift 
in mean Pacific cod length occurred between 
2550 BC and 80 BC, yet Pacific cod abun- 
dances varied minimally during this time. It 
is noteworthy that this entire period occurred 
during the Neoglacial, a period of generally 
cooler and wetter oceanic conditions. This re- 
lationship may indicate that perturbations 
in abundance, at least during cold periods, 
were not related to systemic fluctuations in 
the length or fecundity structure of the Pacific 
cod populations in the Gulf of Alaska. 
At the end of the temporal sequence ca. 
1030 AD and 1540 AD, an inverse relationship 
is noted between mean length and abundance 
of Pacific cod. The mean lengths increased 
during the Medieval climatic anomaly and 
decreased during the Little Ice Age, whereas 
overall relative abundances decreased during 
the Medieval climatic anomaly and increased 
during the Little Ice Age. This finding would indicate 
that Pacific cod were more abundant but smaller during 
cool periods and less abundant but larger during warm- 
er periods. Although the inverse relationship between 
length and abundance is intriguing, it is uncertain how 
they are related. The shift in mean Pacific cod length 
between these periods was small (less than 3 cm), yet 
abundances fluctuated widely. 
Climatic regime shifts appear to have had minimum 
effects on the length structure of Pacific cod populations 
in the Gulf of Alaska, but have had dramatic effects 
on their numbers and possibly geographic range. The 
apparent drop in Pacific cod numbers during periods of 
warming may reflect a shift in Pacific cod distribution 
(e.g., Charnov et al., 1976), and may explain why the 
Aleut word for Pacific cod is the “fish that stops.” 
General observations regarding 
the prehistoric Pacific cod data 
Additional insights about Pacific cod can be made from 
the analyzed bones, which show, for example, that Pacific 
cod have long-term, millennial-scale population dynam- 
ics that appear to have a complex relationship with cli- 
matic shifts. Pacific cod seem to be vulnerable to major 
climatic regime shifts, particularly warming conditions, 
and appear to rapidly recover from major perturbations 
to their environment. This conclusion would signify that 
oceanic regime shifts have no long-term (centennial- to 
millennial-scale) ramifications on the overall structure 
of Pacific cod populations. Furthermore, Pacific cod 
stocks appear to have largely maintained their popula- 
tion structure over some 4500 years despite the pres- 
sures of commercial harvesting in the modern era. 
The trends uncovered for Pacific cod in the North Pa- 
cific are significantly different from those in the North 
Atlantic. Swain et al. (2007) concluded that length dis- 
tributions can be a key indicator of genetic changes in 
Pacific cod as a result of overfishing, and that these 
changes can persist for decades despite low harvest pres- 
sure. In particular, they found that rapid evolutionary 
changes can result from size-selective overfishing, and 
these changes can far outpace the levels expected for 
natural mortality as a result of disease or regime shifts. 
