¹ Durham, Lord, Report on the Affairs of British North America, Carleton Library abridged (Toronto, 1963), p. 27Google Scholar.

² Jones, Robert Leslie, “French-Canadian Agriculture in the St. Lawrence Valley, 1815–1850,” Agricultural History, 16 (07 1942), 137–48Google Scholar. The most notable French-Canadian contributions are Ouellet, Fernand, Histoire economique et sociale du Quebec (Montreal, 1966)Google Scholar, and Seguin, Maurice, “Nation Canadienne” et I'agriculture, 1760–1850 (Trois-Rivieres, 1970)Google Scholar. The book by Seguin is a cently published doctoral dissertation written in 1947. The author was very influential in the development of the “national” school of French-Canadian history that traces many of the ills of modernday Quebec to the consequences of the conquest of New France by Britain. The “inefficient” characterization of the French-Canadian farmer continues to prevail. Ouellet, Witness Fernand, Le Bas-Canada, 1791–1840: Changements structuraux et Crise (Ottawa, 1976)Google Scholar, and Isbister, J., “Agriculture Growth in Canada Since 1850,” Economic Development and Cultural Change, 25 (07 1977), 673Google Scholar. A lone note of dissent has been sounded by Pacquet, Gilles and Wallot, Jean-Pierre in “Crise Agricole et Tensions Socio-Ethnique Dans le Bas-Canada, 1802–1812,” Revue d'Histoire de I'Amerique Fran caise, 26 (09 1972), 185–237Google Scholar. Their main concern is to contest the argument of Ouellet that there reject out of hand the view that the French-Canadian farmers pursued an inefficient, irrational husbandry. Pacquet and Wallot, however, neither contest the arguments put forward by other writers nor present any counter evidence.

³ Lower Canada had a sizable minority of English-speaking settlers. Few of these actually were English but were Irish, Scottish, and Americans from New England. For convenience we refer to them in this paper as English.

⁴ Lord Durham, quoted in Jones, R. L., “French-Canadian Agriculture,” p. 148Google Scholar.

⁵ Ouellet in many ways has been a severe critic of Seguin, yet his overall characterization of French-Canadian agriculture is very similar. In his first of several books on the issue he wrote “En somme, depuis 1785, il existent dans le Quebec deux types d'agriculture. L'une, traditionelle et can-adienne-francaise, occupait la masse; l'autre, progressive et britannique, rallait les Britanniques et une elite canadienne-francaise,” Histoire Economique, p. 461Google Scholar.

⁶ Gates, Paul, The Farmer's Age: Agriculture, 1815–1860 (New York, 1963), p. 35Google Scholar.

⁷ In citing censusfigures,previous writers have overlooked problems with the data that cast French fanners in a disadvantageous light. Comparisons of crop yields are made between English and French regions without recognizing that for the French regions thefiguresactually are minots per ar-pent, not bushels per acre. Since 1 minot equals 1.107 bushels and 1 arpent equals 0.845 of an acre, 1 minot per arpent is more than 30 percent greater than 1 bushel per acre. Also, comparisons are made of average farm size without adjusting for the inclusion in many French districts of a large number of small (that is, no more than one arpent) garden plots that were not really farms.

⁸ The Scottish agriculturalist J. F. W. Johnston often is cited for his comparisons of French- and English-Canadian husbandry in Lower Canada. Yet Johnston makes clear that his impressions of Lower Canada were based on a most cursory look and he was repeating what was told to him by others. The comparison he actually makes is between common practice and the accomplishments of three identified British immigrant farmers in the vicinity of Montreal. See his Notes on North America: Agricultural, Economical and Social (Edinburgh, 1851)Google Scholar.

⁹ Jones, along with most other writers, quotes Pierre de Sales Laterriere to the effect that the habitant farmers are “far from adventurous, they cling with pertinacity to the little piece of land which in the division of the family property, has fallen to their share.” Laterriere, de Sales, A Political and Historical Account of Lower Canada (London, 1830), p. 120Google Scholar. Jones does not share with his readers, how ever, statements from the same source to the effect that, in 1830, the agriculture of Lower Canada produced “a good level of subsistence, equally distributed.… Only in the United States does one find any people better off...” (p. 119), and that “The comforts of the people, if compared with any other nation, are wonderfully great” (p. 127). Indeed, de Sales Laterriere insightfully explains how Canadian fanning may appear “slovenly and unskillful to the farmer of England but is nevertheless rational” (p. 123). Jones also neglects Robert Russell, who is well known to American readers for his severe criticism of the way livestock were managed in the U.S. South, yet does not comment unfavorably on the animal husbandry of the French-Canadians and is otherwise complimentary about their farming. See Russell, R., North America, Its Agriculture and Climate (Edinburgh, 1857)Google Scholar.

¹⁰ Many examples could be cited. One rather colorful one may be sufficiently indicative of the general problem. It is widely asserted that the French-Canadian farmers were so backward that they did not know enough to spread the manure on the fields but instead hauled it onto the ice of the river to be washed away in the spring ( , Jones, “French-Canadian Agriculture,” p. 141Google Scholar). The story goes back to the anonymous eighteenth-century author of American Husbandry, Carman, Harry J., ed. (Port Washington, 1964), p. 26Google Scholar. What was actually said was that, unlike in Europe, in the towns of Quebec and Montreal the nightsoil was not carted into the countryside to be used as fertilizer but was disposed of in the river. The story was repeated often, notably in 1863 by Yind, H.Y., Eighty Years Progress of British North America (Toronto, 1863), p. 33Google Scholar, so Jones prefaced his description with “In fact, as late as 1860,…” as though Hind were describing current practice.

¹¹ There are many aspects of the agricultural situation in Lower Canada that are in need of re-examination. We proceed in small steps, however, and wish to emphasize the limited focus of the present study.

¹² Other prominent contributors to the literature have thought it appropriate to highlight the French-English comparison for this date. See especially Hamelin, Jean and Roby, Yves, Histoire Economique du Quebec, 1851–1896 (Montreal, 1971), p. 8Google Scholar.

¹³ The parishes and townships of Lower Canada were quite detailed localities, the smallest units for which census data were tabulated. The distinction between parish and township was neither cultural nor linguistic—there were French settlers in the townships and we include a few parishes of English farmers—but related to the survey and the form of land tenure. The parishes were subdivisions of the old seigneurial area; the townships were surveyed under British rule. For brevity we use district to mean either parish or township.

¹⁴ An alternative approach would be to utilize the abundance of manuscript census data that are available to compare individual French and English farms within the same locality. Undoubtedly this should be done but difficult problems stand in the way. First, there were in fact very few districts that contained a substantial mixture of both French and English farmers. Second, there is a sticky problem of how to handle interfarm exchanges of feed crops. Most seriously, though, there is the difficulty of finding a basis on which to allocate hired labor to individual farms.

¹⁵ Comparability undoubtedly is weakest within Region III. The districts of Region III are included in the study primarily because they constitute the main area of English and French farming in Lower Canada. This weakens the comparisons made later of French and English districts within Region III, but it gives substance to the overall comparison of French and English farms.

¹⁶ Although this region looks most extensive on the map, it actually was thinly settled and contained only a small number of farms.

¹⁷ The assumptions of elasticities of substitution equal to one and equal factor shares are restrictive but do not seriously impinge on the results. The reason is that we derive estimates using a wide range of parameter values (labor's share, for example, is varied over a range from 0.355 to 0.700; see Table 2). That should accommodate possible differences by ethnicity in factor shares that might arise from different factor input proportions combined with non-unitary elasticities of substitution.

¹⁸ Fogel, Robert W. and Engerman, Stanley L., “The Relative Efficiency of Slavery: A Comparison of Northern and Southern Agriculture in 1860,” Explorations in Economic History, 8 (Spring 1970), 353–67Google Scholar; , Fogel and , Engerman, “Explaining the Relative Efficiency of Slave Agriculture in the Antebellum South,” American Economic Review, 67 (06 1977), 275–96Google Scholar.

¹⁹ It is said that in the hillier parts of the English-settled Eastern Townships wheat could not reliably be grown. At the time the same was true, but for a different reason (the wheat midge), in the French districts. There were locational differences between British and French districts with respect to distances from markets. These are not easily interpreted, however. The British were generally farther away from the only important urban markets in Montreal and Quebec, but they were more favorably situated with respect to the New England market upon which they are claimed to have depended. Despite these differences, we apply the same product prices to all districts, as is discussed below. Farmers of both cultural groups were engaged in a largely subsistence agriculture, producing mainly for their

²⁰ We can perceive some differences in product mix by ethnicity. The English produced more beef and perhaps more dairy products; the French more wheat and pork.

²¹ Neither poultry nor egg production was reported in the census and there is no plausible basis for estimating this output. Traditionally, however, they were the responsibility of the farm wife and we have not included her labor in our measure of labor input. Forest products (timber, firewood, shingles, staves and ashes) are also excluded but, again, we make some allowance for this omission by adjusting labor input, as we discuss in a later section. A number of minor farm products such as honey, wax, tallow and domestically produced textiles are also omitted.

²² This is the methodology used in well-known historical estimates of net agricultural production for the United States. See Gallman, Robert E., “Commodity Output, 1839–1899,” in Conference on Research in Income and Wealth, Trends in the American Economy in the Nineteenth Century, Studies in Income and Wealth, vol. 24 (Princeton, 1960), pp. 13–71Google Scholar; Strauss, Frederick and Bean, Lewis H., Gross Farm Income and Indices of Farm Production and Prices in the United States, 1869–1937, Technical Bulletin no. 703, U.S. Department of Agriculture (Washington, D. C., 1940)Google Scholar.

²³ For example, converting feed into dairy products required more labor than converting feed into beef.

²⁴ Suitable compilations of Montreal prices are not readily available; we obtained our data mainly from newspaper reports of prices in the Montreal wholesale market.

²⁵ This might give the appearance of assuming away the very productivity differentials we seek to examine. As explained below, however, we compensate for this in the way we calculate the deduction for feed.

²⁶ After consulting a wide range of literature on animal feeding practices, including some that relate to the early nineteenth century, we settled on using the recommendations reported in one quite comprehensive source: Bailey, L. H., ed., “Animals,” Cyclopedia of American Agriculture, vol. 3 (New York, 1908), 100–01, 107–08, 432Google Scholar.

²⁷ We should not overemphasize the value added by animals to feed. Respondents to a questionnaire circulated in 1850 by a select committee of the Legislative Assembly of Canada were divided over whether it was more profitable to sell coarse grains and hay or to feed them to animals. See the Report of the Special Committee on Agriculture, Canada, Legislative Assembly, Journals 1851, Appendix J.

²⁸ A further source of bias for which the method does not adjust would be in systematic differences between districts in the efficiency with which feed grains are converted into livestock products. We would have failed to prevent the estimate from being biased in favor of the French if the English farmers followed practices that allowed them to get larger outputs of animal products from given inputs of feed. This could result from their having better varieties of animals, or having better knowledge of proper timing in the feeding of animals. It is not at all clear that such differentials would always have been in favor of the English. For example, in an age when cattle were generally low grade stock, the French-Canadians were known to have kept a notably thrifty variety of milk cow. Its milk output was low but the milk was high in butterfat content and the animals required less feed than the general purpose varieties of cattle kept elsewhere in North America.

²⁹ Land “under cultivation” was specified as land in.crop, pasture, garden, or orchard and was equivalent to all land not designated as “wild or woodland.”

³⁰ These values are shown in Appendix Table 2, available on request.

³¹ The value of farm implements was not reported in the Canadian census of 1852. In 1861 the value of implements amounted to only 5 percent of the value of capital in the form of livestock. We have assumed that in 1851 the relative importance of capital was about the same as a decade later and have augmented the value of livestock in each district by 5 percent to provide a crude allowance for implements.

³² This fraction is the proportion of agricultural workers over 14 years of age multiplied by 0.286, the weight derived by Ransom and Sutch for juvenile agricultural labor in the U.S. post-bellum South. See Ransom, Roger and Sutch, Richard, “The Impact of the Civil War and of Emancipation on Southern Agriculture,” Explorations in Economic History, 12 (01 1975), 23CrossRefGoogle Scholar. There were differences between districts in the coverage of reported occupations, primarily owing to variations in the treatment of the youngest and the oldest workers. We impose a degree of uniformity by assuming that 90 percent of the male population aged 15 to 70 had an occupation and adjust the number in each of the five occupational categories proportionally.

³³ For a very few counties there exist no district-level data on occupations. In those cases the number of agricultural workers is computed at the county level from published census data, in the same manner as the parish-level approximation described above. This estimated county agricultural work force is then allocated to districts within the county in proportion to the number of farms of more than 10 acres extent. Almost all of the recorded “farms” of less than 10 acres are residential garden plots of an acre or less, occupied by persons reporting an occupation other than farmer. The frequency of these small plots varies widely from district to district.

³⁴ Minor exceptions would be maple sugar, which was a significant component of output in only a o few districts, and wool, since sheep would sometimes have been run on wild land.

³⁵ Although draft animals were used on uncultivated land, we make no adjustment to the capital input. The reason is that we base our estimate of the number of draft animals solely on the number of cultivated acres (see the Appendix to this paper, available on request).

³⁶ This specification is consistent with Cobb-Douglas production functions for output produced on each of improved and unimproved land, and is premised on a rational allocation by farmers of their resources such that they equate marginal value products of inputs applied to either improved or unimproved land. It follows that the output elasticities of factors will equal the factor income shares. For example:

where Q_i is output by type of land; T_i is land input; r_i is the return on land; and w is the wage rate; α_i and γ_i are the output elasticities of labor and land respectively; i = 1 for improved land and 2 for unimproved land. The prices of output on improved and unimproved land are normalized to one. It then follows that:

³⁷ We explored the possibility that the ratio of unimproved to improved land in 1852 was inversely related to land quality since that could also account for a significant b coefficient. We ran two separate regressions, one for districts shown by twentieth-century soil surveys to have higher proportions of poor quality land and one for the remaining districts. Both regressions yielded statistically significant b coefficients of about the same magnitude as the one estimated from the entire sample. It should be added that according to modern soil surveys there was no difference in the quality of land between districts according to the ethnicity of their occupants in 1852.

³⁸ Note that a = Q/L since, if τ_u = 0, all output is the product of improved land.

³⁹ The second restriction is imposed to force the coefficients on land and capital to assume “reasonable” values. The restriction embodies two assumptions: first, that the ratio of the value of land to the value of capital was the same in 1851 as in 1861; second, that the rate of return on land, per dollar invested, was the same as on capital.

⁴⁰ The relatively high F-statistic indicates that the estimated coefficients are biased. Because of that and our concern that the rather low estimated value of labor's share may be favoring the more labor-intensive French agriculture, we change our procedure slightly. The subsequent discussion relies on assumed values of factor shares, introduced as weights in our productivity equation, and covers a wide range of values.

⁴¹ It may be worth noting that this result is fairly robust to modifications in the measures of input and output. In the process of carrying out the research we had occasion to revise several times what we thought was the best way to measure the input and output series. In no case did we get a result remarkably different from that reported here.

⁴² This differential is greater than that estimated in equation (6), mainly as a consequence of variations in the size of the districts. In estimating equation (6) all districts were treated as representing the same number of farms. It happens that a few large English districts were also among the most efficient. These get a greater weight in the calculation underlying Table 2 than they do in the regression estimation. The difference might be phrased as involving that between a comparison of farms rather than a comparison of districts.

⁴³ The value of labor's share (0.355) in equation (6) is toward the lower end but certainly not outside the range of values obtained in other agricultural production function studies.

⁴⁴ The result for Region II (Table 2, line 3) is particularly interesting. This region, just to the south of Montreal, was the one most frequently visited by nineteenth-century travellers and commentators on the backwardness of French-Canadian agriculture. It was there that the habitant farmers were claimed to have been especially resistant to adopting the superior farming practices of their English neighbors.

⁴⁵ By far the most efficient farming was to be found in the English townships of the county of Stanstead. This was an upland, cattle-grazing district, located in the southeastern corner of the province, remote from the Montreal market but with commercial links to New England. There was no French district that was comparable in terms of location, land characteristics and market ties. At the other end of the productivity scale were a few French parishes in the counties of Nicolet and Yam-aska. These can be seen in Figure 1 as the outlying French districts in the western part of Region IV. They were quite recently settled districts with unusually small farms in an area of rather poor soil. They were remote and quite different from any of the English districts of Region IV.

⁴⁶ Further tests of the relative efficiency of French and English farming should be carried out. Estimates could be made for other census years to check the consistency of our results. The manuscript census data might be used to examine individual farms in regions where French and English were farming in close proximity. If these conclusions hold up it will be necessary to look more closely at the agriculture of Lower Canada to discover just why the French-Canadian farmers were performing better than so many commentators believed.

⁴⁷ It could be argued that the seasonal nature of agriculture allowed work on unimproved land to be done in off-peak periods with little or no reduction in labor time available for application to the production of improved land. Certainly the clearing of land and cutting of timber was predominantly a winter occupation, but then so was the threshing and hauling to market of the crops grown on cultivated land.