¹ To estimate the number of Canadian-bom persons living in a state at the end of a given decade who had migrated to the state prior to that decade, we must have knowledge of mortality rates and of the age distribution of Canadian immigrants. Although no specific data on the age distribution of Canadian immigrants exist, an approximation of the age distribution of all immigrants may be obtained from the age distribution of all foreign-born persons in a state at the first census after the beginning of substantial immigration of foreign-born persons to the state. It will be assumed that the age distribution of Canadian settlers at the first census after their migration is the same as that of all foreign-born persons at the census following the first substantial settlement. To this age distribution the 1901 American life tables are applied to estimate the proportion of settlers that would be expected to be living ten years, twenty years, and thirty years following the first census after their migration. With the use of these survival proportions, the number of persons entering a state prior to a given decade who would be expected to be alive at the end of that decade can be computed.

The age distributions of foreign-born persons following initial settlement in the six states are similar enough so that the proportion expected to survive subsequent decades does not differ significantly between states. Therefore, the same survival proportions are used for all states. To take account of the slightly higher mortality rates during the decades prior to 1901 than in that year, a downward adjustment of just under two percentage points is made in the survival proportions based on 1901 data. The survival proportions that are used in these migration estimates are as follows: 86 per cent of the Canadian born persons living at the first census after their migration will survive the next decade, 82.5 per cent of those surviving the first decade will survive a second decade, and 75 per cent of those surviving two decades will survive a third. The estimates of migration of Canadian born to Manitoba are derived by applying the same survival proportions to Manitoba's non-prairie-born Canadian population.

² The proportion of Canadian-born settlers that did not survive until the first census year after their migration would depend on the concentration of migration between the earlier and latter parts of the decade, and could range from 2 to 10 per cent of the total migration. Because the concentration of immigration cannot be accurately estimated for each state in each decade, an over-all average of 5 per cent is used, which in some cases may be an overestimate or underestimate.

³ For a discussion of the early development of the Red Valley, River see Briggs, Harold E., Frontiers of the Northwest: A History of the Upper Missouri Valley (New York, 1940).Google Scholar Also Crawford, Lewis E., History of North Dakota, 2 vols. (Chicago, 1931)Google Scholar; and Shannon, Fred A., The Farmer's Last Frontier: Agriculture 1860–1897 (New York, 1945).Google Scholar

⁴ Manitoba Department of Agriculture and Statistics, Report, 1882 (Winnipeg, 1883), 33–4.Google Scholar

⁵ Manitoba Department of Agriculture and Immigration, The Story of Manitoba's Agriculture (Winnipeg, 1958)Google Scholar; and US Commerce Department, Statistical Abstract of the United States, 1891 (Washington, 1892), 302.Google Scholar

⁶ For estimates of frost damage, the prevalence of various wheat varieties, and an evaluation of various cultural practices see Manitoba Department of Agriculture and Immigration, Crop Bulletin, nos. 1–60 (1883–1899).Google Scholar

⁷ To make this adjustment, it will be assumed that the wheat yield response to summer-fallow in our selected areas on both sides of the border for 1924–60 was the same as it was in the North Dakota areas during the seven years 1956–62, for which the USDA has estimated the yields for both wheat planted on land summer-fallowed and on land cropped the previous year. To estimate the number of bushels by which planting part of the wheat acreage on summer-fallow has raised the long-term average yield of an area above what it would have been if all wheat had been planted on land cropped the year before, the USDA estimates of the yield increases resulting from summer-fallow are multiplied by the average proportion that the summer-fallow acreage was of the wheat acreage for 1924–60. This increase is then subtracted from the long-term average to arrive at the wheat yield on continuously cropped land. It is assumed that virtually all the summer-fallow acreage is sown to wheat. The proportion that the summer-fallow acreage of an area was of the wheat acreage is derived from census returns.

⁸ While the Manitoba farmer was favoured by both the wheat price at the terminal market and freight rates, most of the price differential resulted from terminal market prices; about 2.8 of the 3.4 cent differential was in terminal market prices.

⁹ That wheat prices were higher in Manitoba than in North Dakota during this period but lower following 1900 may be explained by the primary dependence of Canadian wheat prices on domestic market forces before 1900 and on world market forces afterward. Net Canadian wheat exports did not consistently top 10 million bushels until after 1895.

¹⁰ North Dakota Department of Agriculture and Labor, “Cost of Producing Wheat in 1893,” Biennial Report, no. 3 (1894), 215–332 Google Scholar, and USDA, Report of the Statistician (Washington, 1894).Google Scholar

¹¹ It must be remembered that costs differ widely from farmer to farmer as tillage practices, the equipment used, and managerial ability vary. The cost estimates of these surveys, if accurate, can only represent the range of costs of a substantial number of farmers. Our estimates based on these surveys might be thought of as the range of costs for widely used equipment and cultural practices. These same qualifications will apply to our subsequent estimates of the differential in the profitability of farming between Manitoba and North Dakota.

¹² For ten counties in northern North Dakota, there were an average of nine respondents per county, ranging from two in a western county to twenty-three in an eastern one.

¹³ These cost estimates are based on the use of a drill for planting and a binder for harvesting. The other two counties in the North Dakota survey had average costs of $4.40 and $5.28 per acre.

¹⁴ Coulter, John L., “Industrial History of the Valley of the Red River of the North,” North Dakota State Historical Society, Collections, III (1910), 607.Google Scholar

¹⁵ Bennett, M. K., “Average Pre-War and Post-War Farm Costs of Wheat Production in the North American Spring Wheat Belt,” Wheat Studies of the Food Research Institute, Stanford University, I, no. 6 (05 1925).Google Scholar These estimates include a charge for marketing the wheat; the estimates above did not.

¹⁶ The crop acreage per farm in Manitoba in 1891 averaged 56 acres compared to 88 acres in northern North Dakota. The progress of Manitoba farmers when compared to that of farmers in other areas settled during the late nineteenth century appears neither exceptionally rapid nor slow. While the crop acreage per farm in Manitoba was less than in northern North Dakota whose settlement occurred concurrently with Manitoba's, it was larger than in western Kansas where settlement had begun five years before that in Manitoba.

¹⁷ Canadian Pacific Railway, Annual Report, 1894, 16.Google Scholar

¹⁸ A ten-year contract, on our assumptions, would have required a net cash outlay over homestead land of $68.

¹⁹ It should be pointed out that in this respect the higher prices created by tariffs were especially onerous to the new settler.

²⁰ Had the settler had more capital, there undoubtedly would have been much less emigration during this period, for CPR land was reasonably priced in relation to its economic advantage over the homestead land of northern North Dakota.

²¹ While it is certainly true, as Professor Mackintosh has pointed out, that price-cost relationships became more favourable for wheat production on the prairies after 1896, there is little reason to believe that development before the turn of the century could not have been more extensive, especially when the capabilities of the Canadian prairies are compared to those of some American frontier areas that were successfully developed.

²² This does not imply a reduction in the amount of land granted to railroads, but rather suggests a different policy in locating granted land which would have left as much land as possible open to homestead in southern Manitoba and southeastern Saskatchewan. The authority for the large majority of the land granted to railroads was enacted before 1880, but most of the land had not been located by that date. For a discussion of locating CPR land, see Hedges, James B., Building the Canadian West (New York, 1939), 32–61.Google Scholar

²³ The Southern International Boundary Reserve gave the CPR the odd-numbered sections still held by the government from the Red River on the east to the Dirt Hills of Saskatchewan on the West and from the 48-mile belt south to the international border.

²⁴ The farm size expansion, which was accomplished through the purchase of CPR land, could have been arranged by the purchase of land vacated by fanners who departed for various reasons, as it was in North Dakota.

²⁵ If a branch line served farmers for a distance of ten miles on each side of the track, each mile of track would serve twenty square miles. The density of population in these areas, when settled, was about seven persons per square mile; there would then be a population of 140 per mile of track. If tariff revenues were $10 per capita per year, there would be $1,400 per year per mile available to amortize the subsidy after the area had been settled. In 1885 tariff collections in Manitoba amounted to $16 per capita.