Science is never certain, and periodically researchers want to poke holes in accepted knowledge. One such hole-poking was a 2011 critique of estimates of 19th century North American wildlife by Yolanda Wiersma and John Sandlos. Sandlos is a historian, Wiersma an ecologist, both based in Newfoundland, Canada, at Memorial University.1 

They are especially concerned with the Canadian government's 'wood' bison restoration project in northern Canada. They suspect the effort is guided by the estimate of 168,000 'wood' bison in the far north during the 19th century. This number was derived in 1941 by a biologist who started with a local national park population and extrapolated it across the 'wood' bison assumed historical range, backed up with explorer accounts of buffalo herds. Wiersma and Sandlos suggest explorers may have seen ordinary buffalo at the edge of their range, instead. Primarily they note the environment is harsh and difficult, and question whether it could sustain so many animals.

Wiersma and Sandlos then cast criticism more broadly. They pointed to sloppyiness in estimating baseline prairie dog ranges (which could skew estimates by 2.4). Their biggest concern were estimated bison numbers across all of North America, by the naturalist Ernest Evan Seton in his 1927 multivolume "The Lives of Game Animals." 

Seton combined visual herd estimates of early observers, and data on horse and cattle carrying capacities, then extrapolated densities across the historical North American bison range. Wiersma and Sandlos do not recalculate numbers, but consider Seton's process wildly speculative. They claim he "painted a fantastic portrait of bygone North America teeming with wildlife." Their concern is again that this could produce "unrealistic goals within contemporary wildlife restoration projects."

Criticism of historical wildlife populations should not be governed by concerns over contemporary restoration project goals, any more than historical sources should be used uncritically. Seton's use of domesticated animal densities was reasonable, given the extensive, free-range livestock management of the early 20th century. These were not animals fed hay, no feedlots like today, only grazing. Further, Wiersma and Sandlos recommend extrapolating from densities of animals in contemporary protected areas that resemble historical landscapes. This would obviously reduce estimates, since today's protected areas (think Yellowstone) are profoundly altered, without anything like historic wildlife numbers. Ironically this was the strategy they rejected when used by the Canadian biologist in 1941, to estimate 'wood' bison populations.

Nor is the use of population observations by 19th century writers unwarranted, given their general agreement. Railroad worker estimates, for example, agreed with military and early naturalist reports. But Wiersma and Sandlos don't consider this, and misrepresent data. A group of researchers modelled the effects of culling bison in U.S. national reserves, and Wiersma and Sandlos report they determined that Native Americans would have kept bison at low levels once they obtained horses (starting in 1735.) This is not what the research found. According to their research, bison populations would have remained stable, because 

"arrival of the horse coincided with increased contact between Native Americans and Europeans, initializing the onslaught of numerous deadly plagues, including smallpox, which decimated tribes in the Dakotas. Thus, while the horse allowed more selective culling of bison, there were simply less Native Americans alive in the Dakotas to take part in the harvest."2

Efforts to poke holes in assumptions are useful, but not like this.

By the way, Seton's estimates, which are speculative, not imaginary, are that early 19th century North America contained 65 million bison, 45 million antelope, 40 million white-tailed deer, 10 million mule deer, 10 million elk, 2 million bighorn sheep, 1 million moose, and 1 million wolves. 

It's worth noting there were at least 15 million white-tailed deer in North America in 1982, and at least 5 million more today.3 Deer are the only large animal that, after their population was decimated in the late 19th century, has recovered a lot. If today's 20 million represents their total 200 years ago, perhaps Seton's estimates should be reduced by half. Others produce estimates in the range of 35-60 million bison, for example. Numbers still massive.

1 Wiersma, Y.F. & J. Sandlos, (2011) "Once There Were So Many: Animals as Ecological Baselines" Environmental History; 16(3):400-407

Millspaugh, J.J., Gitzen, R. A., Licht, D. S., Amelon, S., Bonnot, T. W., Jachowski, D. S., … Suedkamp Wells, K. M. (2008) "Effects of Culling on Bison Demographics in Wind Cave National Park, South Dakota" Natural Areas Journal, 28(3):240–250

Passmore, R.C., (1990) "White-Tailed Deer" Canadian Minister of the Environment, CW69-4/7-1990E