Changes in the system of classifying the soils of Canada have resulted from the combined effects of international developments in concepts of soils and increasing knowledge of Canadian soils. From the time of the first surveys Canadian pedologists were influenced by the concept of soil as a natural body integrating the accumulative effects of climate and vegetation acting on surficial materials. This concept was introduced by Dokuchaev about 1870, developed by other Russian soil scientists, and proclaimed to western Europe by Glinka in 1914 in a book published in German. Marbut's translation of this book made the Russian concept of soil as a natural body easily available to the English-speaking world (Glinka 1927). The concept is of paramount importance in soil science because it makes it possible to classify soils on the basis of properties of the soils themselves rather than on the basis of geology, climate, or other factors. Classification systems based on the inherent properties of the objects classified are called natural or taxonomic systems.

Recognizing the relationships between soil features and factors of climate and vegetation was not limited to Russian scientists. In the United States, Hilgard had noted this association in a book published in 1860 (Jenny 1961), and Coffey had recognized soils as natural bodies by 1912 (Kellogg 1941). However, the Russians can be credited with developing the concept of soils as natural bodies with horizons that reflect the influences of soil-forming factors, particularly climate and vegetation.

Although classification was on the basis of texture in the earliest soil surveys undertaken in the Prairie Provinces during the 1920s, an increased awareness of soil zones and of the soil profile is evident in soil survey reports published during that decade. Preliminary soil zone maps of Alberta and Saskatchewan were presented by Wyatt and Joel at the first International Congress of Soil Science in 1927. They showed the broad belts of brown, black, and gray soils. The Congress and the associated field tours brought Canadian pedologists into close contact with international concepts of soil and systems of soil classification.

Developments in soil classification occurred independently in each province because surveys were carried out by university departments of soils or chemistry. For example, a numbering system indicating the soil zone, nature of parent material, mode of deposition, profile features, and texture was developed in Alberta. J.H. Ellis in Manitoba recognized the impossibility of developing a scientific soil taxonomy based on the limited knowledge of Canadian soils in the 1920s. Influenced by concepts of C.C. Nikiforoff in Minnesota, Ellis developed a field system of soil classification that was useful in soil mapping and endures in various revised forms to this day. The system identified "associations" of soils formed on similar parent materials and "associates" that differed according to topographic position within the association (Ellis 1932, 1971).

During the 1930s soil surveys proceeded in Ontario and the Prairie Provinces and were started in British Columbia in 1931, Quebec and Nova Scotia in 1934, and New Brunswick in 1938. Soil surveys began in Prince Edward Island in 1943, the Northwest Territories and the Yukon in 1944, and Newfoundland in 1949. The few soil surveyors employed permanently by federal and provincial departments of agriculture in the 1930s worked cooperatively with personnel of university soils departments. By 1936 about 15 000 000 ha (1.7% of the land area of Canada) had been surveyed, mainly in Alberta, Saskatchewan, and Ontario. Soil classification was limited by the fragmentary knowledge of the soils of Canada.

Canadian pedologists were influenced in the 1930s by Marbut's developing ideas on soil classification, Ellis's system of field classification (Ellis 1932), and the system of classification used in the U.S. Department of Agriculture (USDA) described by Baldwin et al. (1938). The latter system divided soils at the highest category among three orders:

1. Zonal soils, which are those with well-developed characteristics that reflect the influence of active factors of soil genesis such as climate, organisms, and particularly, vegetation (e.g., Podzol).
2. Intrazonal soils, which are soils having more or less well-defined characteristics that reflect the dominant influence of some local factor of relief or parent material over the normal effects of climate and vegetation (e.g., Humic Gleysol).
3. Azonal soils, which are soils without well-developed characteristics due either to their youth or to some condition of relief or parent material (e.g., alluvial soils).

To act as a coordinating body among the soil survey organizations in Canada supported by the Canada Department of Agriculture, provincial departments of agriculture, research councils, and departments of soil science at universities. Its stated functions include the following:

1. Improving the taxonomic classification system for Canadian soils and revision of this system because of new information.
2. Improving the identification of physical features and soil characteristics used in describing and mapping soils.
3. Reviewing the methods, techniques, and nomenclature used in soil surveys and the recommendation of changes necessary for either a greater measure of uniformity or for their improvement.
4. Recommending investigations of problems affecting soil classification, soil formation, and the interpretation of soil survey information.
5. Recommending and supporting investigations on interpretations of soil survey information for soil ratings, crop yield assessments, soil mechanics, and other purposes.
6. Cooperating with specialists in soil fertility, agronomy, agrometeorology, and other disciplines in assessing interrelated problems.

Developments in soil classification in Canada since 1940 are documented in reports of the meetings of the NSSC held in 1945, 1948, 1955, 1960, 1963, 1965, and 1968 and of the CSSC in 1970, 1973, 1974, and 1976. Soil classification was one of the main items on the agenda of the first meeting, and a report by P.C. Stobbe provoked a prolonged and lively discussion. He and his committee recommended a system of field classification similar to that developed in Manitoba by Ellis (1932). The proposed system was a hierarchical one with seven categories as follows:

Soil Regions - tundra, woodland, and grassland soils.

Soil Zones - broad belts in which a dominant kind of soil occurs, such as podzol or black soil.

Soil Subzones - major subdivisions of soil zones, such as black and degraded black.

Soil Associations or Catenas - the group of soils that occur together on the same parent material to form a land pattern.

Soil Series, Members, or Associates - the individual kinds of soils that are included in an association.

Soil Type or Soil Class - subdivisions of associations or of series based upon texture.

Soil Phase - subdivisions of mapping units based upon external soil characteristics such as stoniness and topography.

This proposal was for a field system of classification, or a system for classifying the units of soils mapped at various scales. The classes at all levels, phase to region, were segments of the landscape that included all the soil variability within the area designated. Thus, a soil zone was a land area in which a "zonal great soil group occurred as a dominant soil." The system was not intended to be a scientific or taxonomic one in which the classes at all levels had clearly defined limits based on a reasonably thorough knowledge of the properties of the entire population of soils in Canada. The proposed system was accepted for trial by the committee, which represented all provinces. Thus, an important step was taken in the development of a national system of classifying the units of soil mapped in soil surveys.

The first Canadian taxonomic system of soil classification was presented by P.C. Stobbe at the NSSC meeting in 1955. This system was a marked departure from the mapping or field classification system proposed in 1945. It probably resulted from the following circumstances:

• the greater knowledge of Canadian soils
• the desire to classify soils, even at the highest categorical level, based on the properties of the soils themselves
• the need for a taxonomic system better than the old USDA system (Baldwin et al. 1938) that focused unduly on "normal" soils. The Soil Conservation Service had begun to develop a new system in 1951, but the fourth approximation of that system was judged to be too complicated and too tentative for Canadian needs (National Soil Survey Committee 1955).

If the map is at a scale of 1:100 000 and the smallest area delineated is a square measuring 1 cm on each side, that area represents 100 ha. Such an area commonly includes upland and lowland positions in the landscape and the associated kinds of soils. The kinds of soils within that area are identified by digging pits at different topographic positions in the landscape. At each of these points the profile exposed usually has a rather narrow range of properties that reflects the influence of soil-forming factors at that point. Therefore, the soils at each point of observation can be classified as a single class in a taxonomic system. The area delineated on the map cannot be classified as a single class in such a system because it includes several kinds of soils. However, the area mapped can be classified as a kind of soil mapping unit such as a soil association in the system of Ellis (1932). Thus, the need was evident for both a taxonomic system to permit the naming and the ordering of information about specific kinds of soils, and a mapping system to permit the ordering of information about the areas delineated on soil maps and the naming of them.

The taxonomic system outlined in 1955, which is the basis of the system used today, had six unnamed categorical levels corresponding to the order, great group, subgroup, family, series, and type. The seven taxa separated at the order level were: Chernozemic, Halomorphic, Podzolic, Forested Brown, Regosolic, Gleisolic, and Organic. Taxa were defined only in general terms down to the subgroup level. Although this was inevitable because of the lack of sufficient information, it led to differences of interpretation of the taxa in various provinces and some lack of uniformity in the use of the system. The need for correlation was clearly recognized by senior Canadian pedologists.

Progress in developing the Canadian system of soil classification since 1955 has been toward more precisely defining the taxa at all categorical levels and increasingly emphasizing soil properties as taxonomic criteria. This is evident from the reports of NSSC meetings held in 1960, 1963, 1965, and 1968, at which the main topic of discussion was soil classification. Some changes in taxa at the order, great group, and subgroup levels were made at these meetings. For example, in 1963 the Meadow and Dark Gray Gleisolic great groups were combined as Humic Gleysol; in 1965 a system of classifying soils of the Organic order was presented and accepted; in 1968 the former Podzolic order was divided into Luvisolic (clay translocation) and Podzolic (accumulation of Al and Fe organic complexes) orders, and the concept and classification of Brunisolic soils were revised. Criteria of classification involving morphological, chemical, and physical properties became increasingly specific through this period. The bases of classifying soils at the family level were outlined and the series and type categories were defined more specifically.

Following the publication of The System of Soil Classification for Canada in 1970, topics other than soil taxonomy were emphasized at CSSC meetings. However, in 1973 a Cryosolic order was proposed to classify the soils with permafrost close to the surface, and some refinements were made in several orders. These changes, including the newly developed Cryosolic order, were incorporated into The Canadian System of Soil Classification published in 1978.

Between 1945 and 1970 little consideration was given at NSSC meetings to systems of naming and classifying soil mapping units. Between 1970 and 1978 a satisfactory taxonomic system was developed. In 1978 the Expert Committee on Soil Survey was formed, replacing the Canada Soil Survey Committee. Although the work of the Canada Soil Survey Committee focused on interpretation, mapping systems, and soil degradation, soil classification also formed an important part of its activities.

In 1980, after C. Tarnocai became chairperson of the Soil Classification Working Group, work began on a number of problems in soil classification that had been identified by soil scientists from various regions. The main items covered included the classification of Gleysols (McKeague et al. 1986), Folisols (Fox 1985, Fox et al. 1987, Trowbridge 1981, and Trowbridge et al. 1985), Podzols and the definition of contrasting horizons. At a meeting of the working group held in 1984, proposals were presented on these topics and solutions were formulated and presented to the Expert Committee on Soil Survey (Tarnocai 1985). The Second Edition of The Canadian System of Soil Classification published in 1987 incorporated the recommendations accepted by the committee.

In the late 1980s, as a result of an international soil correlation tour organized by the USDA Soil Conservation Service, attention was drawn to Vertisolic soils; soils of high clay content whose properties result from shrinking and swelling, caused by wetting and drying. Previously, Vertisolic soils were thought to occur only in temperate and tropical climates. However, when international experts attending this meeting examined high clay content soils in southern Saskatchewan, they found them to have the same properties as Vertisols occurring elsewhere and considered these soils to be Vertisols. This changed viewpoint initiated activities aimed at incorporating Vertisols in The Canadian System of Soil Classification. The Canadian Soil Classification Working Group formulated a number of proposals for classifying these soils. Between 1991 and 1994 field tours were held in the Prairie Provinces, British Columbia, and eastern Canada to test these proposals in the field. As a result of these activities, the working group recommended establishing a Vertisolic order to classify these soils and set up criteria for this new soil order.

This Third Edition of The Canadian System of Soil Classification includes minor revisions throughout the system and introduces the Vertisolic order. This edition also introduces the Vertic great group for Solonetzic soils and Vertic subgroups for the Chernozemic, Gleysolic, Luvisolic, and Solonetzic soils. It also provides a major revision of the Cryosolic order resulting from international activities, both in the United States and Russia.

Because soil occurs at the surface of the earth as a continuum with variable properties, it is necessary to decide on a basic unit of soil to be described, sampled, analyzed, and classified. Such a unit was defined by United States' pedologists (Soil Survey Staff 1960) and is accepted in Canada. Called a pedon, it is the smallest, three-dimensional body at the surface of the earth that is considered to be a soil. Its lateral dimensions are 1-3.5 m and its depth is 1-2 m. The pedon is defined more specifically in Chapter 2.

Both the theoretical and practical purposes of soil classification have been discussed in the literature (Cline 1949, 1963; De Bakker 1970). The general purpose of soil classification in Canada may be stated as follows:

• to organize the knowledge of soils so that it can be recalled systematically and communicated, and that relationships may be seen among kinds of soils, among soil properties and environmental factors, and among soil properties and suitabilities of soils for various uses.

The overall philosophy of the Canadian system is pragmatic; the aim is to organize the knowledge of soils in a reasonable and usable way. The system is a natural, or taxonomic, one in which the classes (taxa) are based upon properties of the soils themselves and not upon interpretations of the soils for various uses. Interpretations involve a second step that is essential if the information is to be used effectively. If the taxa are defined on the basis of soil properties, and if the boundaries of these classes, or of combinations of them, are shown on a map, interpretations can be made on the basis of properties implied in the class definitions.

1. It is a misconception that a good system results in the assignment of soils occurring close together to the same taxon, at least the higher categorical levels. This is neither possible, nor desireable, in some areas. Pedons a few metres apart may differ as greatly as pedons hundreds of kilometres apart within a climatic region.
2. Another common misconception is that a good national system provides the most suitable groupings of soils in all areas. This is not possible because criteria based upon properties of the whole population of soils in the country are bound to differ from those developed on the basis of properties of soils in any one region. Criteria developed for a national system will inevitably result in areas where most of the soils have properties that straddle the boundary line between two taxa.
3. The idea that if the system was soundly based there would be no need for changes every few years is erroneous. As new areas are surveyed, as more research is done, and as concepts of soil develop, changes in the system become inevitable to maintain a workable taxonomy.
4. Another unfortunate hope is that a good system will ensure that taxa at the order level at least can be assigned unambiguously and easily in the field. Actually, in a hierarchical system the divisions between orders must be defined just as precisely as those between series. With pedons having properties close to class boundaries at any taxonomic level, classification is difficult and laboratory data may be necessary.
5. The assumption is made by some that a good system permits soils occurring within mapped areas to be classified as members of not more than three series. Clearly, this is not reasonable because the number of taxononic classes occurring within a mapping unit depends upon the complexity of the pattern of soils in the landscape, on the scale of the map, and on the narrowness of class limits.
6. The idea that a good system is simple enough to be clear to any layman is erroneous. Unfortunately, soil is complex and, although the general ideas of the taxonomy should be explainable in simple terms, the definitions of taxa must be complex in some instances.
7. Another misconception is that a good system makes soil mapping easy. Ease of mapping depends more upon the complexity of the landscape, the access, and the predictability of the pattern of soils within segments of the landscape than upon taxonomy.

1. It provides taxa for all known soils in Canada.
2. It involves a hierarchical organization of several categories to permit soils at various levels of generality to be considered. Classes at high categorical levels reflect, to the extent possible, broad differences in soil environments that are related to differences in soil genesis.
3. The taxa are defined specifically so as to convey the same meaning to all users.
4. The taxa are concepts based upon generalizations of properties of real bodies of soils rather than idealized concepts of the kinds of soils that would result from the action of presumed genetic processes. The criteria chosen define taxa in accordance with desired groupings of soils. The groupings are not decided upon initially on the basis of arbitrary criteria.
5. Differentiae among the taxa are based upon soil properties that can be observed and measured objectively in the field or, if necessary, in the laboratory.
6. The system can be modified on the basis of new information and concepts without destroying the overall framework. Periodically, however, the entire framework of the system will be reevaluated.

Order Taxa at the order level are based on properties of the pedon that reflect the nature of the soil environment and the effects of the dominant, soil-forming processes.

Great group Great groups are soil taxa formed by subdividing each order. Thus, each great group carries with it the differentiating criteria of the order to which it belongs. In addition, taxa at the great group level are based on properties that reflect differences in the strengths of dominant processes, or a major contribution of a process in addition to the dominant one. For example, in Luvic Gleysols the dominant process is considered to be gleying, but clay translocation is also a major process.

Subgroup Subgroups are formed by subdividing each great group. Therefore, they carry the differentiating criteria of the order and the great group to which they belong. Subgroups are also differentiated on the basis of the kind and arrangement of horizons that indicate conformity to the central concept of the great group (e.g., Orthic), intergrading toward soils of another order (e.g., Gleyed, Brunisolic), or additional special features within the control section (e.g., Ortstein, Vertic).

Family Taxa at the family level are formed by subdividing subgroups. Thus, they carry the differentiating criteria of the order, great group, and subgroup to which they belong. Families within a subgroup are differentiated based on parent material characteristics, such as particle size, mineralogy, calcareousness, reaction, and depth, and on soil climatic factors.

Series Series are formed by subdividing families. Therefore, they carry all the differentiating criteria of the order, great group, subgroup, and family to which they belong. Series within a family are differentiated on the basis of detailed features of the pedon. Pedons belonging to a series have similar kinds and arrangements of horizons whose color, texture, structure, consistence, thickness, reaction, and composition fall within a narrow range. A series is a category in the taxonomic system; thus, it is a conceptual class in the same sense as an order.

A pedon is a real unit of soil in the landscape; a series is a conceptual class with defined limits based on the generalization of properties of many pedons. A particular pedon may be classified as a series if its attributes fall within the limits of those of an established series. However, it is not, strictly speaking, a series because the attributes of any one pedon do not encompass the complete range of attributes allowable within a series. Thus, it is not correct to study part of a pedon and to declare, "this is X series." Rather it should be stated, "this pedon has properties that fall within the limits of the X series," or "this pedon is classified in the X series."

The Podzolic order, for example, is defined based on morphological and chemical properties of the B horizon. However, these properties are associated with humid conditions, sandy to loamy parent materials, and forest or heath vegetation. Although the great groups within the order are defined on 7 the basis of the amounts of organic C and extractable Fe and Al in the B horizons, they have broad environmental significance. Humic Podzols are associated with very wet environments, high water tables, periodic or continuous reducing conditions, hydrophytic vegetation, and, commonly, a peaty surface. Ferro-Humic Podzols occur in areas of high effective precipitation, but they are not under reducing conditions for prolonged periods. Humo-Ferric Podzols generally occur in less humid environments than the other great groups in the order. An interrelation of climatic and vegetative factors, parent material, and relief determines the occurrence of the different classes of Podzolic soils. Similarly, general relationships exist between other orders, great groups, and soil environmental factors. However, these relationships are much less clear for some Regosolic and Brunisolic soils than they are for most soils of other orders. At lower categorical levels, in general, relationships between soil taxa and factors of the soil environment become increasingly close.

The Canadian system of soil taxonomy is more closely related to the U.S. system than to any other. Both are hierarchical, and the taxa are defined on the basis of measurable soil properties. However, they differ in several respects. The Canadian system is designed to classify only soils that occur in Canada and is not a comprehensive system. The U.S. system has a suborder, which is a category that the Canadian system does not have. In the Canadian system Solonetzic and Gleysolic soils are differentiated at the highest categorical level as in the Russian and some other European systems. These soils are differentiated at the suborder or great group level in the U.S. system. Perhaps the main difference between the two systems is that all horizons to the surface may be diagnostic in the Canadian system, whereas horizons below the depth of plowing are emphasized in the U.S. system. This may be because 90% of the area of Canada is not likely to be cultivated.