Collateral damage to marine and terrestrial ecosystems from Yankee whaling in the 19th century

Abstract Yankee whalers of the 19th century had major impacts on populations of large whales, but these leviathans were not the only taxa targeted. Here, we describe the “collateral damage,” the opportunistic or targeted taking of nongreat whale species by the American whaling industry. Using data from 5,064 records from 79 whaling logs occurring between 1840 and 1901, we show that Yankee whalers captured 5,255 animals across three large ocean basins from 32 different taxonomic categories, including a wide range of marine and terrestrial species. The taxa with the greatest number of individuals captured were walruses (Odobenus rosmarus), ducks (family Anatidae), and cod (Gadus sp.). By biomass, the most captured species were walruses, grampus (a poorly defined group within Odontoceti), and seals (family Otariidae). The whalers captured over 2.4 million kg of nongreat whale meat equaling approximately 34 kg of meat per ship per day at sea. The species and areas targeted shifted over time in response to overexploitation of whale populations, with likely intensive local impacts on terrestrial species associated with multiyear whaling camps. Our results show that the ecosystem impacts of whaling reverberated on both marine and coastal environments.

In addition to the animals captured, technological and environmental limitations resulted in large numbers of whales that were harpooned but not landed, often dying in the process (Scarff, 2001) This exploitation had effects on the whales' population structure that are still visible today (Alter, Rynes, & Palumbi, 2007;Mesnick et al., 2011;Monsarrat et al., 2016;Roman & Palumbi, 2003;Ruegg et al., 2013).
Whaling voyages lasted from several months to over 5 years and covered tens of thousands of kilometers (Table 1). Because crews were typically paid in proportion to the total value of the catch, there was economic incentive to not return until the vessels' holds were full.
Subsequently, their voyages covered immense areas of open ocean (Smith et al., 2012). Whaling voyages represent some of the earliest, and in some cases the only, sources of historical ecological knowledge about the pelagic habits of these highly migratory animals, and the details within whalers' logbooks give insight into marine ecosystems in the 19th century (Clapham et al., 2004;Townsend, 1935). In this way, This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. a careful reading of logbooks can highlight how human perceptions of whale abundances have shifted over time (Pauly, 1995).
While large whales were the primary targets of the American fleet (the so-called Yankee whalers), they were not the only species targeted during these voyages. Infamously, 79 American whaling vessels captured over 13,000 Galapagos tortoises between 1831 and 1868 to serve as fresh meat on long voyages (Townsend, 1925). Similarly, Bockstoce and Botkin (1982) estimated that Yankee whalers killed over 200,000 walruses between 1848 and 1914. Thus, the ecosystem impacts of the American whaling fleet were not limited to the reduction in biomass and fixed carbon in the system due to the removal of large whales.
The capture of great whales can be viewed as individual captains opportunistically supplementing both the ship's oil holds and their pantries. Fresh meat was difficult to obtain along these voyages, and the chance to add new meat was rarely passed over. This gustatory enthusiasm for fresh meat even made its way into the most apocryphal of Yankee whaling tales, Moby Dick (Chapter 65: The Whale as a Dish. Melville, 1851). During the long periods between capturing large whales, other species would have provided the whalers a welcome diversion from preserved food and also occasionally additional sources of valuable oil. In particular, as whales became depleted, multiyear expeditions to more distant locales became necessary, requiring that overwintering whalers obtain provisions locally. Additionally, some species, such as walruses, were captured to provide additional income, through rendering to produce oil and the collection of tusks (Fay, Kelly, & Sease, 1989).
To fully understand the historical ecology of the marine ecosystems, we must rely on the data provided by the whalers themselves. While the history, ecology, and fisheries impacts of the large whale hunt have been well-documented elsewhere (Herman, 1979;Smith et al., 2012;Townsend, 1935), the diversity of the other species targeted as well as their spatial distribution has not been fully explored. Here, we describe and quantify the diverse array of organisms other than large whale species captured by the American whaling fleet during the latter half of the 19th century (ships leaving port 1847-1900). In doing so, we have two main hypotheses. First, that because these were economic voyages, the majority of the nongreat whale catch recorded will be of species with economic value and not simply food items. Second, because of localized resource exploitation and increases in technology over time, we will see shift toward targeting populations in increasingly remote areas or species that were inaccessible with technology readily available during the beginning of the study period.

| MATERIALS AND METHODS
We collected data from 79 digitized logbooks from the New Bedford Whaling Museum (NBWM) that cover a total of 74 voyages during the years 1846 to 1901 (Table 1). Logbooks from this period are not common, and the collections at the NBWM represent the largest collection of these documents. We focused on the latter half of the 18th century as it was during this time that the American Whaling fleet moved almost exclusively offshore from New England and the industry shifted from baleen to oil. It was during this time that the Arctic grounds were opened and American whaling was in its "golden era" (Dolin, 2008).
For each vessel, we recorded the unique logbook ID name, years active, home port, dates of departure and arrival, number of days at sea, and overall whaling grounds targeted. Within each logbook, we compiled records of the presence of nonwhale species captured. Exact longitude and latitude of each point of capture were recorded when possible, but many of the specific locality data were incomplete due to a lack of location observations during the examined period. In those circumstances, longitude and latitude coordinates were extrapolated from known locations within 10 days before or after the examined date, whenever possible (Table S1).
To quantify the level of exploitation, we listed the organisms captured to the most specific taxonomic resolution possible. When archaic terms were used, we used metadata such as geographic range, physical descriptions, or logbook illustrations to help refine taxonomic assignment. We calculated both absolute numbers of organisms caught and estimated approximate biomass of the total catch based on recorded average adult weights (Bigelow & Schroeder, 2002;Delacour, 1954;Nowak, 1999;Rice, 1998), although we used modern size data, we do note that species such as Cod (Gadus morhua, Hutchings & Baum, 2005) and Polar Bears (Ursus maritimus Rode, Amstrup, & Regehr, 2010) have undergone a recent reduction in size, and thus, our findings represent a conservative estimate of biomass. For species with extreme sexual dimorphism, we averaged between sexes as logbooks did not frequently differentiate (Prieto et al., 2013). For the taxonomic designation "grampus," we used the weight of Cuvier's Beaked Whale (Ziphius cavirostrus, but see discussion below for the taxonomy of grampus).
We searched the historical literature to determine which species were associated with market goods (e.g., furs, oil) to differentiate between species targeted solely for food from those targeted for both food and opportunistic income supplementation.
To test the second hypothesis, that the fishery expanded in space (as measured by days at sea), we used a Mann-Whitney-Wilcoxon test, to assess averaged numbers of days at sea and numbers of individuals caught binned into before and after the ending of the US civil war (voyages starting 1846-1864 and 1865-1900, respectively).
We chose this time to partition the data because after the US Civil war, there was an increase in well trained, and armed, men entering the fishery (Bockstoce & Botkin, 1982). Additionally, we calculated the diversity nonincidental (>10 individuals of a single species taken by a single vessel) catches by decade and analyzed spatial changes in nonincidental catch over time, which we associated with known changes in the abundance and availability of whales. Lastly, we calculated the total amount of contributions made to the total catch by strictly aquatic, semiaquatic, and strictly terrestrial animals.

| RESULTS
We collected data from 79 logs of which 56 (73.68%) reported catches of nongreat whale targets. These logs record the capture of 5,255 individuals of 32 different taxonomic designations (Table S1). The species with the greatest number of individuals caught were walruses (Odobenus rosmarus N = 2,283), ducks (Anatidae N = 949), and cod (Gadus sp., Table 2). The species with the most biomass caught were walruses, "grampus," and "seals" (Table 2) The temporal patterns showed a heterogeneous pattern of exploitation. First, significantly more exploitation of nongreat whales  (Table 3). As well, whaling voyages required overwinter stays to make trips profitable (Bockstoce, 1986). In response, the diversity of nonwhale catch increased (Table 4), reflecting a shift to subsistence hunting as whalers became reliant on local provisioning of locally abundant game species like ducks, deer, grouse, ptarmigan, and rabbit (Table 5).  (Table 2). Similarly, the numbers of walruses in the data resulted in the vast majority of biomass (~95%) recorded being from semiaquatic animals (Table 2).

| DISCUSSION
The collateral damage of the large whale hunts of 19th Century  Figure 2). While the walrus data were not surprising (Bockstoce & Botkin, 1982), what was unanticipated, was the diversity of terrestrial animals that were also captured by these ostensibly marine voyages.
Many of the terrestrial species were taken in northern latitudes (Table S1), while vessels were searching for more sought after whale species. For example, the bowhead whale, Balaena mysticetus, is a cold-water specialist and was highly prized by Yankee whalers (Smith et al., 2012). The seasonal migrations of the animals coincided with the increasing daylight and subsequent increase in primary productivity in Arctic waters (Braham, Fraker, & Krogman, 1980). Whalers arriving ahead of these migrations would heighten their capacity to capture the greatest number of whales. Thus, it was not uncommon for ships to arrive early and prolong their stay, to maximize exploitation of the resource. Due to the vagaries of northern storms, ships were occasionally trapped in sea ice. For example, in September 1871, 40 American ships were frozen in the ice off of Port Franklin, Alaska. Thirty-two of 40 ships (including the Henry Taber, the Navy the Seneca, and the John Wells whose logs we included in this study) were crushed in ice and lost (Starbuck, 1878).
During the times when the vessels were close to shore (or trapped in ice), away teams were sent out to provision the vessels. This provided American whalers the opportunity to capture terrestrial and coastal animals such as ducks, ptarmigan, fox, deer, bear, moose, and, at least on one occasion, two kangaroos. Sailors in the high Arctic targeted caribou, as they believed the meat could counteract scurvy (Hadley, 1915). While the local impacts on the local ecology could be severe (see discussion of Hershel Island below), it is unlikely that whalers captured enough individuals to have a substantive impact across the entire range ( T A B L E 3 Nonincidental catch, or ten or more individuals of one species taken by a single ship. Here, we report only nonincidental catch that was associated with a known location hypothesis-that technology and exploitation patterns will lead to shifts in the places and kinds of species targeted. In our data, there In addition to the need for provisioning, falling whale oil prices lead to the need to target species that could be of secondary commercial importance. The walrus boom of the mid-to late-1800s resulted in the taking of upwards of 235,000 walruses by the American fleet with 90% of that occurring between 1867 and 1883 (Table S1, Bockstoce & Botkin, 1982), a total that represents the approximate modern census size of all walrus populations (Lowry, Kovacs, & Burkanov, 2008).
Our data show 2,283 individual walruses being captured. Based on the 60%-70% capture efficiency presented in Bockstoce and Botkin (1982), the whalers in our data set killed a minimum of 3,192 walruses.
Several forces led to the start of this walrus boom. Access to walruses was improved after The United States purchased Alaska in 1867, obtaining legal claim over the walrus populations therein. This period also coincided with reductions in bowhead whale populations and a steady market for walrus products (Bockstoce & Botkin, 1982). Walruses

| Data limitations
One of the major limitations to this study, and indeed many historical ecology studies in general, is that modern researchers are restricted to the quality of the data within the historical record (Josephson, Smith, & Reeves, 2008;McClenachan et al., 2015). In this paper, this limitation has three manifestations. The second limitation centers on locality information. Often, the exact location of where the species were targeted was often not recorded. While we are able to record information at the scale of ocean regions or basin, more spatially explicit information was only recorded for a limited number of records (Table S1) and therefore we are unable to make more detailed analysis as to the spatiotemporal patterns of species capture.
The third limitation lies in trying to navigate the targeted species' taxonomy. The people recording the logs were not trained scientists, and while they had intimate knowledge of the behavior and ecology of the large whales, they were unencumbered with formalized spelling rules, consistent common names, or widely accepted taxonomy (Townsend, 1925 (NBWM 1988.6.195) Vessels like the one pictured here overwintered in Arctic waters to capture bowhead Whales. While waiting for the ice to melt, they sent hunting and trading parties onto the land with ecological and social impacts to the animals and people living in those areas allows us to differentiate between anthropogenic and climate driven cycles in abundance (Schwerdtner Máñez et al., 2014), to model ecosystem productivity (Rosenberg et al., 2005) and to reconcile past species distributions (Drew, Philipp, & Westneat, 2013). While we urge caution when dealing with conclusions drawn from incomplete historical data, in many cases these data represent the only insight we have into the less perturbed past of ecosystems (Hayashi, 2014;Schwerdtner Máñez et al., 2014). Ignoring these data runs the risk of setting the conservation bar too low.  (Fraker & Bockstoce, 1980; Figure 3). During the 1890s, the estimated population size of 1,500 people (Bockstoce, 1986). Our limited sampling of the total whaler efforts showed that crews of vessels captured 316 ducks, 158 "deer" (most likely caribou), 36 foxes, 11 grouse, 120 ptarmigan, 149 rabbits, 21 seals, and one bear from Herschel Island.

| Conservation implications
Similarly, Bockstoce (1980) suggested whalers took over 12,000 caribou from Herschel Island between the periods 1890 and 1908. Modern studies looking at how the ecosystem including the community ecology and nutrient cycling patterns of the region have changed over time needs to factor in the magnitude of biomass removal. Only by doing this will researchers be able to set adequate targets for restoration and conservation.
In contrast to localized terrestrial impacts, walruses faced massive declines across their ranges due to unregulated hunting from both opportunistic whalers and targeted walrus hunts. The harvest data indicate that current walruses have gone through at least three anthropogenic population declines (Fay et al., 1989) although these bottlenecks may have occurred too recently to be reflected in molecular analyses (Andersen et al., 2009). Modern distribution of walruses, and the associated high levels of population connectivity, may be a result of population expansion into areas that were defaunated by whalers (Wiig, Gjertz, & Griffiths, 1996).
Additionally, the impacts of the whaling and walrus hunting on the Indigenous cultures that were dependant on those species were not overlooked by contemporary authors. For example, Aldrich (1889) recounted that "Whaleman have practically driven the walruses from the shore, and greatly reduced the numbers of hair seals and whales. Thus, all the supplies of food have been curtailed." The loss of both the bowhead whale and the reduction in walrus populations had negative consequences on the Indigenous tribes, resulting in loss of food, shifts in harvesting and migration patterns and urbanization around trading centers such as the one established in Herschel Island (Foote, 1964;Hadley, 1915). The rapid transition of Herschel Island into a whaling center had at least two impacts on the Indigenous population. First, it changed their annual trading voyages and leads to a centralization of the population. With the establishment of a trading outpost on the island, the population had less reason to migrate, especially because the store offered processed food. The importance of this store was reflected in the native language with the word iglupûk meaning big house, or in the context of Herschel Island, the Hudson Bay Trading company (or on occasion, the police barracks- Stefansson, 1909;).
Second, the sailors would also commission the Indigenous people to hunt caribou, fish, and ptarmigan, often paying for those goods in flour, molasses, and canned meats (Hadley, 1915). This shift in dietary preferences portended current concerns of cardiometabolic health among Indigenous peoples of the high Arctic. (Ryman et al., 2015).
Our data show that Yankee whalers targeted a number of species, both marine and terrestrial during their search for whales. We also show the number of these nongreat whale targets changed over both time and space, and while locally intense, the take of terrestrial organisms was probably insufficient to cause range-wide declines in terrestrial animals. However, we did show that there were substantial impacts to commercially valuable semiaquatic organisms such as walruses, with impacts on both biological and cultural diversity in the far north. Our work shows that Yankee whalers had a wide-ranging impact on marine ecosystems in general but also on localized terrestrial ecosystems. Logbooks of 74 vessels covering 79 voyages contain a sample of the vivid splendor of past ocean ecosystems. When one extrapolates the take of nontarget species from our small sample of 79 voyages out to the entirety of the American Fleet, estimated at over 1,600 voyages (Townsend, 1935), it becomes clear that commercial whalers represented a nontrivial removal of nonlarge whale biomass from terrestrial and marine systems.

ACKNOWLEDGMENTS
This paper arose as a class project for W4115 Historical Ecology offered in the spring semester of 2015 at Columbia University. We would like to thank K. Amatangelo, M. Young, numerous anonymous referees, and M. Dyer for useful comments.