Geografie 1965, 70, 336-343
Population Structure in Cartographic Representation
The differences between rural and urban population represent the most important factors of present day society. Geographical studies should always bear in mind this basic duality, and that is why even the cartographic representation of population structure and development should improve. In the quantitative conception, structure is determined by the share af distinguishable parts of an organized whole. As far as population geography is concerned, districts may be considered as such a whole of the lowest order as they represent the smallest territorial units in which the enclosed cycle of a permanent flow of the population between country and town takes place. For closer studies a district represents the smallest unit also in view of the fact that statistical data concerning the population structure are seldom available for smaller units. The cartographic representation should therefore strive to depict these small units on one cartogram only in their duality even in case their structure is concerned, that is, for instance, the proportion of industrial population. The present practice so far depicted the urban population with a different symbol, circle oftenest, placed in the middle of a district. The proportion of industrial population would then be represented by means of shading or colour in a way of analogy as the remaining district area. The "remaining" area represented the "remaining" population, that is, also the rural one. There are three reasons explaining why such a procedure is not logical; these are as follows: a) the rural population is depicted according to real administrative boundaries while the urban population according to any arbitrarily chosen geometrical symbol; the urban population is, however, also actually counted within the limits of administrative boundaries in the same way as the rural population; b) neither is proportionally depicted the size of both the urban and rural population component; and yet this proportion substantially determines both the economic and demographic development of the district; c) a scientific cartogram should bear a proportional size of the circle. In countries, therefore, where there are big cities, the circle will mostly be too small to odepict the degress af observed property by means of shading. The author suggests a more logical way. He follows the presupposition that the whole area of a territorial unit under observation, that is, a district, is considered to be the total number of its population, therefore, 100 per cent. This presupposition is generally acknowledged in the cartogram theory. If, therefore, the whole area of a district represents the total number of its population P on the cartogram, the urban population U should be therefore represented by a proportionate share of the area as to size and shape. It is true, however, that the urban population lives on a relatively smaller area than the rural population but, on a cartogram, it is always necessary to sacrifice the geographical representation in favour of the statistical one. The total number of the population is ever represented by a whole area though the population occupies only a small part of the area. (If we count the population within the frame of a built-up area, it occupies, in Czechoslovakia for instance, only 1,4 per cent of the total area.) We therefore suggest that the total of urban population be represented by area U which is reduced by means of the equidistance from district boundaries. The difference between both areas, that is, more or less, a broad zone surrounding the area U, would then represent the total af rural population R. The width of such strips R is differentiated in 8 groups by 10 % of the total population. The network of districts thus divided, then represents a basis on which to depict the population structure. Different degrees af observed property intensity, both for the rural population and the urban population, will then be depicted either by hatching or colour on thus reduced areas. The enclosed cartogram illustrating the industrial structure of Czechoslovak population according to districts in 1961, shows how the task has been carried out. The above-mentioned cartogram does not, however, include all districts as decreed by state administration, as five of them are not comparable units, and would, as a result, interfere with the proper purpose and sense of any cartogram. Five largest cities represent, of course, a particular administrative unit like other district but do not correspond to a geographic conception of the district the substance of which is created by economic and social relations of the town wiht many surrounding communities. That is why three of the largest towns are, on the abovementioned cartogram, connected with the neighbouring district, and two centres of the most dense population, that is Prague, and Ostrava, with two neighbouring districts. Further requirement of comparability relates to urban population, and, as a result, the administrative and geographical limitation of a town. This problem cannot be dealt with in paper. I would like to mention only, that a new classification of towns has been carried out on the ground of a detailed research for population census purposes in Czechoslovakia in 1961. The classification takes account of 67 characteristics, referring not only to economic structure and density of population but also of how they are equipped with communal constructions and services, at all. The present distinction of urban rural population have been carried out according to 462 towns thus classified. In case the proposed method has been applied, a relation comes to life and is not usually taken into account when compiling a simple cartogram. The fact is that the width of R strips is to a high degree influenced by areas and shapes of the districts being observed. This is apparent when comparing the districts of Karlovy Vary, Jablonec and Náchod, which are on the cartogram marked with lettres K, J, N. District J has the same proportion of rural population as district K, while the marginal band of this district is twice as wide as that of district N because the latter is four times larger. As far as it concerns the shape of districts, the comparison of N and J districts may be particularly informative. The former district is twice as large as the latter both of them having the same proportion of rural population, and yet the size does not influence the width of band R because the shape of district N is extended to a high degree while the shape of district J is, on the contrary, relatively round. Such a dependence on the shape of a district is however being reduced with the growing size of a district: see the comparison of districts Rychnov and Banska-Bystrica in the Table. For 103 czechoslovak districts the differences in the width of equidistant strips R are summarized in two frequency series, the series M concerns the arithmetical mean, the series V variation span. With the exception of the first degree, that is below 20 %, the variation span is smaller than the average. If an elongated district is slightly narrowed in the centre, inside area U breaks into two parts especially if there is only a small ratio of the urban population. Nové Zámky, marked NZ on the enclosed cartogram, may serve as an example. Similar cases occur in Great Britain, such as, Berkshire, for instance, or Oxfordshire, or Banff, the size of which corresponds to Czechoslovak districts. We should, however, put up with such a shortcoming as no regularity required by the cartogram cannot be put in harmony with natural irregularity of districts formed by islands as the case is with the county of Shetland. As far as the districts formed artificially are concerned, our method proves expecially clearly that the administrative division sometimes corresponds very little with the conception of centrality which forms the substance of the district conception. The proposed cartogram of the population structure is based on the presupposition that the whole area of a territorial unit represents 100 per cent of its population. This identification of area and population provokes some discrepances if we reduce the area according to certain ratio of population. In case the proposed method were applied at the density of population, the district area would be reduced by means of the equidistance not according to the ratio of the urban population, but according to the ratio of the areas of urban communities. On such cartogram the width of R strips would be not as variable as on the enclosed cartogram because the urban areas do not change as fast as their population does. The cartograms of either type have been compiled and published in the National Atlas of Czechoslovakia, sheet No 24 and 26. The proposed method draws attention to the shape of a territorial unit being observed. The relation between its shape and the numeric expression of its characteristics, will be of a steadily increasing intrest for geografers as they will ever oftener apply the quantitative method.
