A “ITEM 05 THE ECONOMIC RELEVANCY OF THE wmmmea use ZONE SYSTEM _ “1.8!: for? flu Dagn- of‘: DH. 0.; > MECHIGAN STATEUNWEKSITY > George R. .Gebhart ' 7 V » 196,6 fHESls This is to certify that the thesis entitled A TEST OF THE ECONOMIC RELEVANCY OF THE HOLDRIDGE LIFE ZONE SYSTEM presented by GEORGE RICHARD GEBHART has been accepted towards fulfillment of the requirements for Ph . D . degree in Economic S / Major professor Date November 10, 1966 0-169 MM MM. Mug ,o“ r! ABSTRACT A TEST OF THE ECONOMIC RELEVANCY OF THE HOLDRIDGE LIFE ZONE SYSTEM by George R. Gebhart The purpose of this study is to test the economic relevancy of a theory of climatic determinism prOposed by Dr. Leslie R. Holdridge of the TrOpical Science Center, San Jose, Costa Rica. Holdridge's thesis is that different combinations of the three climatic factors - heat, precipitation, and moisture - between areas will result in different ecological relationships for those areas. Ranges of the three climatic factors are established on a logarithmic base, and the concurrence of these ranges define what Holdridge calls ”life zones." Field work in the American Tropics has shown that these life zones correlate well with distinct changes in the natural vegetation. Given this "natural" relationship, Holdridge theorizes that many other variables in the social sciences and natural sciences will be influenced by these specific ranges of the three climatic parameters. Economic variables, especially those in the agricultural sciences, are expected to be related to these ranges in the climatic parameters. This study attempts to establish empirically the degree of this relationship for certain economic variables. It is hypothesized here that productivity, technology, cost of production, and land-use vary by life zone. To test this hypothesis two life zones are compared for differences in the above variables, and an attempt is made to connect the differences found to the George R. Gebhart variations in the climatic parameters. These comparisons are made on the basis of a sample of farms from each of the life zones which was extracted from the Costa Rican Census of Agriculture of 1963. An additional comparison is based on a sample of all farms in those political districts of Costa Rica falling within the two life zones. Productivity is measured in terms of the output per manzana of rice, corn, and beans. Both samples were compared by life zone for this analysis. Technology is defined as "the method of production," and is compared by life zone in the production of corn. Data on technology were collected by interviewing a sample of thirty farmers in each of the two life zones. Costs of production between life zones in the production of corn are compared using data collected during the same interviews. Land use is compared by life zone for the farms comprising the sample from the Census of Agriculture and the sample of political districts. A classification of land use which exhausts all possible uses of the land is the basis for this comparison. The variables productivity and costs of production are quanti- fiable and are compared for statistically significant variations. This consists of computing the t-statistic from the ratio of the difference in the mean to the standard error of the difference, and then either accepting or rejecting the null-hypothesis at the .05 level of probability. Differences in the Operations involved in the production of corn are considered as differences in technology, and variations in the percentages of land falling into the various categories are taken as indicative of differences in land-use. George R. Gebhart No statistically significant differences were found between life zones in.productivity. A distinct difference was found in the technology used in the two life zones, but this seemed to be related more to differences in topography than in the climatic parameters. The difference in the costs of production.was statistically signifi- cant, but was related to the difference in tapography. Differences in land-use, especially the difference in the percentages of the life zones planted in permanent craps, did seem to be related to the difference in heat between the two life zones.- The results of this test give very little support to the claims of economic relevancy for the Holdridge Life Zone System. A TEST OF THE ECONOMIC RELEVANCY OF THE HOLDRIDGE LIFE ZONE SYSTEM By A \( may 0 George R) Gebhart A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Economics 1966 ACKNOWLEDGMENT I would like to express my appreciation and gratitude to Dr. John Hunter for his assistance, encouragement and patience throughout the course of this work. 11 TABLE OF CONTENTS ACMOWLEMMENT O O O O O O O O O O O O O O O O O O O O O O O 0 LIST OF TABLES O O O O O O O O O O O O O O O O O O O O O O O O ' LIST OF ILLUSTRAT ION O O O O O O O O O O O O O O O 0 O O O O 0 Chapter I II III IV I NTRODUCT I ON 0 O O O O O O O O O O O O O O O O O O 0 LIFE ZONE ECOLOGY - THE LIFE ZONE SYSTEM . . . . . . General Description of the Life Zone System Basic Determinants of Life Zones . The Climatic Parameters Temperature Precipitation Meisture The Life Zone Chart The Vegetative Characteristics The Life Zone System . Definitions of the TrOpical Moist Forest and Tropical Premontane Wet Forest Life Zones The Life Zone System and Economic Deve10pment METHOWLmYooooooooooooooooooooo The Basic Hypothesis Selection of the Life Zones Selection of the Crap Selection of the Sample Description of the Department of Census Sample Selection of the Sample Employed Herein The Problem of the Small Farms THE LIFE ZONE SYSTEM AND PRODUCTIVITY . . . . . o . . The Problem of Isolating the Effects of the Climatic Parameters The Influence of Tepography on Productivity variation of TOpography Between the Two Life Zones Other Factors Affecting Productivity The Concept "ProductiVity" Test of Relationship Between Farm Size and Productivity Comparison of Productivity Between Life Zones - Total Sample iii Page . vii . vi . viii . l . 10 . 47 . 7O Chapter Page Comparison of Productivity Between Life Zones Census Districts Conclusions V FARM.TECHNOLOGY AND THE LIFE ZONE.SYSTEM . . . . . . . . 92 Definition of Farm Technology Ecologists Belief Concerning Relationship Between Farm Technology and the Life Zone System Description of Technology in use in the Trapical Premontane Wet Forest Life Zone - Labor Intensive Method Preparing the Land Planting the Seed Cultivating the Crap Harvesting the Crop Marketing the Crop Description of Technology in use in the Tropical Moist Forest Life Zone Capital Intensive Method Preparing the Land Planting the Seed Cultivating the Crap Harvesting the CrOp Marketing the Crop Comparisons of Specific Aspects of the Technology used Irrigation Fertilizer Herbicides, Fungicides, and Insecticides Time of Planting Rotation Systems Shading of Crops Selected or Hybrid Seed Use of By-Products Spacing and Depth of Planting Use of Tools and Machinery ' Use of Credit Size of Areas Devoted to Corn Percentages of Areas Devoted to Corn Conclusions VI COST OF PRODUCTION AND THE LIFE ZONE SYSTEM . . . . . . . .116 Nature of the Expected Variation by Life Zones in Cost of Production Relationship Between Technology and Cost of Production Measurement of the Cost of Production Cost of the Labor Input Cost of the Capital Input Cost of the Land Input iv Chapter Page Imputed Cost of Buildings and Improvements Cost of Other Direct Inputs Cost of Entrepreneurship Ignored Basis for the Variation in Cost of Production by Life Zone Comparisons of Five Cost Patterns Based on Cost Work Sheets Conclusions Based on this Comparison Statistical Comparison of Cost of Production by Life Zone Relation of Differences Found to the Climatic Parameters Conclusions VII LAND USE AND THE LIFE ZONE SYSTEM . . o . . . . . . . . 151 The Importance of Effective Land Use Nature of the Expected Difference in Land Use Between Life Zones The Nature of the Comparison Made Here of Land Use by Life Zones The Categories of Land Use to be Compared The Comparison of these Categories for the Sample of Farms Comparison of these Categories for the Census Districts 85 Per Cent Within the Life Zones Variations Between the Two Samples Conclusions VI I I CONCLUS IONS O C O C O O O C C O C O O O O O O O O O O O 1 65 APPENDIX C O O O O 0 O 0 O O O O O O O C O O O O O O O O O O O O 175 BIBLIOGRAPHY O O O O O O O O O O O O O O O O C O O O O O O O O O 192 Chapter Page Imputed Cost of Buildings and Improvements Cost of Other Direct Inputs Cost of Entrepreneurship Ignored Basis for the Variation in Cost of Production by Life Zone Comparisons of Five Cost Patterns Based on Cost Work Sheets Conclusions Based on this Comparison Statistical Comparison of Cost of Production by Life Zone Relation of Differences Found to the Climatic Parameters Conclusions VII LAND USE AND THE LIFE ZONE SYSTEM . . . . . . . . o . . 151 The Importance of Effective Land Use Nature of the Expected Difference in Land Use Between Life Zones The Nature of the Comparison Made Here of Land Use by Life Zones The Categories of Land Use to be Compared The Comparison of these Categories for the Sample of Farms Comparison of these Categories for the Census Districts 85 Per Cent Within the Life Zones Variations Between the Two Samples Conclusions VI I I CONCLUS IONS O O 0 O O O O O 0 O O O O O O O O O O O O O 1 65 APPENDIX 0 O O O O O O O O O O O O O O O O O O O O O O O O O O O 175 BIBIJIOGRAPHY o o o o e o o o e o o o o o o o o 0 O 0 O o O 0 o o 192 Table 1. 10. 11. LIST OF TABLES Minimum Limits for Separating Large and Medium Size Farms by Crop and Province . . . . . . . . . . . Classification of Farms into Large, Medium, and Small............o......... SlOpe variations Between the TrOpical Premontane Wet Forest and TrOpical Moist Forest Life Zones . . . Correlations Between Size of Farm and Output Per ManzanabyLierone....o.......... t-Statistics for the Tests of Significance of the Correlation Coefficients for the Correlation Between Farm Size and Productivity by Life Zone. Standard Errors of the Mean and the Ratios of these Standard Errors to their Means by Life Zone for Selected Craps . . . . . . . . . . . . . Farm #1 - Cost of Production of one Manzana of Corn with Macana Planting and no Plowing in the Tropical Premontane Wet Forest Life Zone . . . . Farm #2 - Cost of Production of one Manzana of Corn using Oxen Plowing with the Seed Planted by Stepping on it in the Tropical Premontane Wet Forest Life Zone . . . . . . . . . . . . . . . . Farm #3 - Cost of Production of one Manazana of Corn.with Macana Planting and no Plowing in the Trapical Meist Forest Life Zone . . . . . . . . Farm #4 - Cost of Production of Corn using Oxen Plowing with the Seed Planted by Stepping on it in the Trapical Moist Forest Life Zone. . . . . . Farm #5 - Cost of Production of one Manzana of Corn using Tractor Plowing with the Seed Planted by Stepping on it in the Tropical Moist Forest Life Zone. . . . . . . . . . . . . . . . . . . . v19 Page 56 57 74 84 84 87 131 133 135 137 139 Table Page 12. Summary Data Selected from the Five Cost werksheets in Tables 7 through 11 . . . . . . . . . . . 141 13. Components as a Percentage of Total Production Cost (excluding marketing cost) . . . . . . . . . . . . 142 14. Total cost of Production per Manzana, Output per Manzana of Corn, and Cost per Unit of Output by Li fe zone 0 O 0 O O O O O O O O O O O O O O O O O O O O 148 15. Percentages of Land in Various Land uses by Life Zone for the Sample Farms Producing Corn . . . . . . . . . . 156 16. Per Cent of Annual CrOp Land Planted in Various Craps by Life zone 0 O O O O O O O O O O O 0 O O O O O O O O 158 17. Percentage Distribution into Land Use Categories of the Total Area of Farms for those Districts at Least 85 Per Cent within one of the two Life Zones. . . 162 vii LIST OF ILLUSTRATION Page The Life Zone Chart 25 viii CHAPTER I INTRODUCTION The consanguinity of ecology and economics is implicit in much of the research undertaken.within either of these two disciplines, but there is a dearth of studies directed towards specifically examining this kinship. The purpose of this study is to investigate this kinship by applying economic analysis to variations on an ecological base. In 1947, Dr. Leslie R. Holdridge invented what has come to be known as the Holdridge Life Zone System.1 This system is a method of classifying groups of plant associations into higher order groupings so as to facilitate their identification and use. These life zones (or groupings of plant associations) are determined by certain ranges of the three climatic parameters: heat, precipitation, and moisture; and a given site can be placed within its correct life zone on the basis of data concerning these climatic parameters. A life zone can be defined as a geographical area having amounts of heat, precipitation, and moisture falling within given ranges. Thus, based on the Holdridge system, an area having 1) a mean annual biotemperature above 24 degrees centigrade (heat) 2) an average total annual precipitation of from 2000 to 4000 millimeters (precipitation), and 3) a potential evapotranspiration ratio of between 1.00 and 0.50 1Leslie R. Holdridge, Life Zone Ecolqu, Tropical Science Center, San Jose, Costa Rica. 2 (moisture), would be in the trepical moist forest life zone.2 The natural vegetation within a life zone is thought to be distinct from that in any other life zone; and the current research on the system by Dr. Holdridge and his associates-is directed towards establishing this difference. They hold that the differences in the three climatic parameters dictate different plant associations and groupings of plant associations in the various life zones, as well as different characteristic species and subspecies. Emanating from this proposition is a whole host of variations that could exist among these life zones. The ecologists believe that these differences do exist, and that they could be useful as a basis for policy-making on the part of governments or others interested in improving and conserving the resource productivity of the earth. Economics is one of the areas in which the ecologists are especially vociferous concerning important variations by life zones. They feel that these variations are especially numerous in agriculture, but do not restrict their expectations of variation to this area. Thus, they feel that production in industry, the cost of that production, the economics of plant location, transportation costs, and a host of other factors will vary by life zone, and that a knowledge of these variations could be useful in solving a multitude of problems that face industry. They see the Holdridge System as a method of classifying data in many disciplines in such a way that the work and decisions of that discipline ZA more complete description of this system can be found in Chapter II. This brief description is given to introduce the reader to the system and to illustrate how the climatic parameters delineate the life zones. 3 might be facilitated. In addition, they are convinced that Holdridge's three climatic parameters have a definite influence on many variables from seemingly unrelated areas, and that policies based on a knowledge of this influence can be used to improve both the performance of the economy and the culture of the people. It is within this context that the present study falls. The purpose of this study is to provide a partial test of the economic relevancy of the Holdridge Life Zone System. Almost all of the research to date on this system has had to do with its ecological and biological aspects. This study attempts partially to fill this gap by investigating the relationship between this system and certain economic variables. The economic variables with which this study is concerned are in the field of agriculture primarily because this is the area that has the most obvious and direct relationship to the climatic factors which form the basis of the Life Zone System. An additional reason for choosing agriculture is the urgency felt today to help underdeveloped countries produce more food to save their thousands from famine. The basic hypothesis of this study is that there are significant differences between life zones in economic variables, i.e., that measurable economic differences exist in various life zones. Definitely these are related to the factors that determine the life zones, and they have economic importance in that they determine what can be produced, the cost of production, the distinct technologies required, etc. If this hypothesis is substantiated, the questions of where to plant what, how much to plant, how much fertilizer to use, what methods to use and many 4 others could be answered by a thorough knowledge of the life zone within which the area falls. The Holdridge system could be very useful to governments in planning for agricultural develOpment and to farmers moving into virgin areas if a close connection could be found between these economic variables and the life zone system. The more specific hypothesis embracing the variables tested in this study is that there are differences in the productivity, in the technology needed, in the cost of production, and in the use made of the land between the tropical premontane wet forest life zone and the trapical moist forest life zone, as these life zones are defined in the Holdridge Life Zone System.3 Productivity is measured for three cr0ps -- rice, corn, and beans - in terms of the output per manzana (an area of approximately 1.7 acres). The second variable, technology, is defined very broadly as the methods used in planting, cultivating, harvesting, and marketing the crOp, and is compared by life zone in the production of corn. Thus, a difference in the method of preparing the land for planting or in the number of times the crop is cultivated makes for a difference in technology. The specific aspects of this definition will be discussed in the chapter on technology. For the third variable, the cost of production, corn is again the crap chosen for study. Cost of production includes all of the actual and imputed costs of planting, cultivating, harvesting, and marketing the corn. These costs are compared for each of the 3These life zones are defined more Specifically in Chapter II° 5 distinguishable methods of production in each of the two life zones, and the total cost of production for the various farms in the sample is compared by life zones to see if any differences are present. Cost of production is measured as the cost per unit of output and the cost per manzana. The last variable, land use, includes all uses to which the land is currently being put in the two life zones. These uses are grouped into appropriate categories reflecting general land use patterns, such as annual crops, permanent crops and pasture. An attempt is made to see if the use of the land by the farmers in their respective life zones differs, and to try to relate any difference found to those factors delineating the life zone. The methodology used for this study consisted of selecting a sample of farms within each of the two life zones, and then inter- viewing the farmers concerning their methods of production, the productivity, the use of inputs, and the land use. In addition, the Census questionnaires from the 1963 Census of Agriculture in Costa Ricaa for the farms in this sample were obtained and certain data extracted from them. A second sample was formed out of certain of the political districts of Costa Rica to supplement the analysis based on the primary sample. A district had to have at least 85 per cent of its land area within one of the two life zones to be included in this sampling. The data were then compared to see if discernible differences between life zones could be discovered in the four variables. 4This study was carried out in Costa Rica where the author spent a year on the staff of the Associated Colleges of the Midwest Central American Field Program. 6 The differences that were discovered between the life zones were then related to the parameters that define the life zones in an attempt to find a connecting link between the variation in the parameters and the difference in the variable being studied. In some cases it was possible to make such a connection, but in others an outside factor seemed better to explain the variation in the economic variable. The Holdridge Life Zone System is a relatively new and almost completely untested theory of climatic determinism. There have been attempts in the past to devise such a classificatory scheme as the Holdridge model, some of which Holdridge mentions in his book,5 but they have never proved of much value even to the ecologists and biolo- gists. The primary reason for this failure is that they have not been able to establish satisfactorily a natural base; that is, their boundaries have not coincided well with changes in the natural vegeta- tion. According to the ecologists and biologists working with the system, the Holdridge System has overcome this difficulty by the discovery of a logarithmic relationship of change in the natural vegetation which corresponds with a logarithmic change in the climatic parameters. This logarithmic base has its counterpart in other ecoloa gical relationships, and is the main factor differentiating the Holdridge System from other unsuccessful classificatory systems. The only research done specifically on the Holdridge Life Zone System from an other than natural science point of view is contained in several unpublished reports by students connected with this Associated Colleges of the Midwest Central American Field Program at BIdee, ppe 9-130 7 the Tropical Science Center in San Jose, Costa Rica. None of these, however, were concerned with agriculture, or even closely related to economics. Moreover, since the system devised by Dr. Holdridge has never received much support from the rest of the scientific community, there has been a paucity of research directly relating the climatic parameters of heat, precipitation, and moisture to productivity in agriculture, etc. There has been considerable research both in the United States and in other countries on what to plant in a given area, how much ferti- lizer to use, at what depth to plant, etc., but nothing has been done to relate the various combinations of these climatic parameters to economic variables in the way envisioned by Dr. Holdridge and his associates. For example, the Rockefeller studies on hybrid corn in Mexico discovered which types of corn grow best in the climate that is found in that area of Mexico where the studies were carried out, but there was no attempt to generalize the results of that research to other areas of the world with similar climate. The fact that some of the varieties that have proved best for that area have also been successful in other areas of the world is prima-facie evidence in favor of an approach along the lines proposed by the ecologists for the Holdridge System. The contribution that could be made to the science of agriculture by the relationship foreseen in the Holdridge System is very great indeed. If the research were carried out in sufficient detail, simply by knowing the values for the climatic parameters one would know what yields to expect from certain crops, what types of fertilizer to use and in what quantities, which insects and other plagues to expect, and one would have a host of other valuable information concerning production, 8 input and output, and related matters. Considering the role that agri= culture plays in developing nations, such information would be invaluable as a planning device. Its importance is echoed by Douglas H. K. Lee who writes: "There is even need for far more complete and particularized climatological information on tropical countries, in form and detail that will allow adequate analysis of the effect of these conditions on agricultural, pastoral, industrial, and social development."6 Another contribution that could be made by studies such as those anticipated here, would be toward the knowledge of the technology that has been developed by the farmers indigenous to the areas who have presumably adapted their practices as best they could to the climatic and cultural conditions as they have found them. This knowledge would provide a substantial base upon which to launch a program of improvement in agriculture, that would have a good chance of success. Then from the variety of practices being used in a given area those that show the most promise could be subjected to more careful scrutiny and promotion. Some students of the subject feel that this is a prerequisite for successful research in this area. Sufrin and Wolf write: "Special stress was placed in the foregoing analysis on the need to select from available technological alternatives those which would maximize productivity of existing capital and of new capital formation. The first requirement for implementing this approach is a selective study of comparative 6 Douglas H. K. Lee, Climate and Economic Development in the Tro ics, Council on Foreign Relations, (Harper and Brothers, N. Y., 1957), p. 174. 9 technologies."7 Such information would indicate which transplanted technology would be most likely to work in an underdeveloped country and which would not. The choice of technology actually used of course, would depend on input prices as well as the productiveness of the technology. Chapter II presents a more thorough description of the life zone system and its relationship to economic development, with emphasis upon both its theoretical and its practical aspects. Chapter III is a description of the methodology used in comparing the two life zones. Chapter IV takes up the relationship between productivity and the life zone system, and Chapter V deals with differences in technology by life zone. Chapter VI relates cost of production to life zones, and Chapter VII discusses variations in land use between life zones. Finally Chapter VIII presents the conclusions of the analysis, along with the limitations to them. Appendix A contains reproductions of the two questionnaires used in the interviews, and Appendix B contains the computations for the standard errors for the various statistical analysis. 7Sidney C. Sufrin, and Charles Wolf, Jr., Capital Formation and Forei n Investment in Underdevelo ed Areas; Maxwell School Series, (Syracuse University Press, 1958), p. 50. CHAPTER II LIFE ZONE ECOLOGY - THE LIFE ZONE SYSTEM1 The life zone system used as a basis for comparative analysis of economic variables in this study was invented by Dr. Leslie R. Holdridge, who first described it in Science2 in 1947. Since that time there has been considerable research done on the system, but there have been no major changes made in its basic thesis. Most of the research has involved testing the accuracy of the system, mapping various countries in Central America and northern South America with reference to the system, and refining certain measures used in delineating its parameters. Practically all of this research has been done by ecologists and geographers, and only recently have social scientists become interested. Thus, while it has been tested thoroughly (and proved of value) from an ecological point of view, almost nothing has been done to see if it has social and economic significance. At the present time there are a few studies of an economic and social nature being carried out on the system in Costa Rica.3 1This chapter draws heavily upon the following two works: Leslie R. Holdridge, Life Zone Ecolo , Central American Field Program, Asso= ciated Colleges of the Midwest, San Jose, Costa Rica, 1964 and, Joseph A. Tosi, Jr., "Climatic Control of Terrestrial Ecosystems: A Report on the Holdridge Model", Economic Geography, Vol. XL, N00 2 (1964)o 2Leslie R. Holdridge, "Determination of World Plant Formations from Simple Climatic Data", Science, 105 (2727), (1947), pp. 367=368. 3At the Tropical Science Center, Apartado 2732, San Jose, Costa RiCa, 10 AIIIIIIIIIlII-_______________________________________________________i444i 11 The purpose of this chapter is to describe as simply yet as thoroughly as possible the mechanics of the Life Zone System and to discuss the theoretical relationships between the system and economic development. The rest of the study seeks to eXplore these relationships and to test their validity. The life zone system is essentially a division of the climatic spectrum along lines that seem to correspond with natural divisions of the earth's vegetation. Dr. Holdridge theorized that each species- population should have evolved (1.6., should have become specialized through selective adaptations)... "to successfully compete and survive as a member of the natural community within only a limited sector of the earth's broad climatic spectrum."4 Therefore, he felt that the vegeta- tion within these limited climatic ranges... "should precisely reflect the integrated Operation of the climatic conditions prevailing over that vegetation."5 In other words, there should be distinctive characteristics of the vegetation.within each of these limited climatic ranges, and furthermore, these characteristics should have evolved over long periods of time due to the influence of certain climatic factors. Wherever the same climatic factors exist on the surface of the globe these distinctive characteristics of the vegetation should be found. The purpose of classifying the vegetative mass is the purely pragmatic one of providing a basis on which comparisons between areas can be made. Ecology has long suffered from the lack of a system for 4Joseph A. Tosi, Jr., "Climatic Control of Terrestrial Ecosystems: A Report on the Holdridge Model," Economic Geo ra h , Vol. XL, No. 2 (1964), p. 174. sIbid. 12 classifying vegetation, and while there have been sporadic attempts to devise such a system, not one has won universal acceptance by ecologists. Dr. Holdridge's system may fill this void, but much more research needs to be done to determine the full extent of its significance. BASIC DETERMINANTS OF LIFE ZONES To provide a system by which the tremendous masses of vegetation can be classified there must be established some order out of the complex vegetation, (i.e., some distinctive characteristics of the vegetation must be delineated), and some of the ecological factors which have a primary influence on the characteristics must be recognized. Only then will it be possible to correlate natural characteristics of the vegetation.with the ecological factors responsible for their evolu- tion. One approach toward establishing this correlation is to determine first the distinctive characteristics of the vegetation, and then to relate these to the various ecological factors. Ecologists generally agree that the plant association is the basic unit of vegetation.6 Due to the large number of distinct associations, however, groupings of these are necessary to facilitate the organization of data and for general comparative work. But this is not as easy as it might seem. Holdridge points out several reasons why it is impossible to construct 6"A plant association is a dominant community of plants which, in its natural state, has a physiognomy distinct from that of all other plant associations. Such an association may be characterized regionally or locally by certain indicator species of the community. There are four types of plant associations: climatic, edaphic, atmospheric, and hydric...Furthermore, there are various combinations of the latter three types." Ibid., p. 175. 13 these groupings by working upward from specific association descriptions. 7‘ The tremendous amount of work and expense involved precludes the collection and study of plant associations in a museum. In addition, man has been very active in altering these associations in the field so that at the present time, except in the most remote places, it is impossible to find them in an untouched state. In fact, in some areas the natural vegetation has long since been removed completely, and it would take centuries before nature could restore it. The large number of plant associations is another obstacle to this approach. Another possibility would be to group the plant associations scientifically on the basis of taxonomic species. However, the same associations on different continents may comprise two or more almost entirely distinct sets of species. In addition, man's past interference makes it difficult to determine whether a species has ever existed in a given area, or whether man's activity has eliminated it. Again the tremendous number of species is an obstacle. Since the above approaches presented serious problems, Holdridge looked to the possibility of organizing the various ecological factors into a system that permitted the grouping of the natural units of vegetation. In other words, he reversed the process by concentrating first on determining which ecological factors could be used in the system, and then relating some vegetational characteristics to the parameters of these factors. The various factors that could be used 7, . Leslie R. Holdridge, Life Zone Ecology, Central American Field Program, Associated Colleges of the Midwest, San.Jose, Costa Rica, (1964), p. 8. 14 include climatic, edaphic, and atmospheric, but since the system was intended for use on a global basis, only those factors which affect the entire area could be used. Only a few of the climatic factors satisfy this requirement, all other factors being local in extent. As Holdridge points out, ”...this does not mean that they (other factors) are not significant in defining the associations, but simply that they do not lend themselves to global categorization."8 Of the climatic factors, only heat, precipitation, and moisture seem to be world wide and capable of subdivision into equivalently valued groupings. These factors produce universal characteristics in the vegetation that are quite distinct from those produced by the more localized effects of the less extensive ecological factors. The solution was to divide the climatic spectrum into equally valued ranges with which vegetative characteristics could then be correlated. Thus, he deduced, "...that the general characteristics of climate in any given locality might be determined objectively from comparative observations of quantified vegetational parameters in the natural plant communities and, conversely, that weather station data might be employed to determine, in terms of these same parameters, what the vegetational climaxes had been and might again become in the absence of human disturbances of natural community equilibrium"9 Holdridge's system differs from other accepted classificatory systems in that, "...it is neither a classification of climate nor a classification of vegetation but is, rather, a classification of the relationship which Ibid. 9Tosi, 02. cit., p. 175. 15 exists between them. It differs in another important respect also: the classification was derived experimentally from comparative observa- tions of natural vegetation as related to climatic factors over a very wide range of geographical environments. Its bases have not been drawn arbitrarily but accord with observable phenomena in nature."10 Thus, the system is "natural" in the sense that the limits of the equivalently valued groupings of climatic data correspond to observable limits in certain parameters of the natural vegetation, and therefore, the limits to a life zone can be determined either from climatic data or from field observations of the vegetation. This "naturalness" is important, for it provides a basis for believing that other social, economic, and ecolo- gical variables may be related to life zones, whereas if this base did not exist - if the system were completely arbitrary - there would be no a priori reason why these other types of variables might be related to the climatic parameters. . . 11 The Climatic Parameters Holdridge's application of his theory in the field involved trying to set down the correct parameters of temperature and rainfall as boundaries between the major units of vegetation. In doing so he found that these parameters, and the related vegetational units, increased logarithmically. This logarithmic base found in nature is nothing new 10T051, 0 . cit., p. 175. 11The reader should refer to the life zone chart on page 25 in reading this section. 16 to the biological sciences,12 which again leads one to appreciate the "naturalness" of the Holdridge system. His three (climatic) factors - heat, precipitation, and moisture - which determine the boundaries of the vegetational units increase logarithmically in delineating those boundaries. Having made this discovery it was relatively easy to extend the parameters to encompass the range of climate found on earth. Recent work on the system by ecologists has been confined to proving the correlation between the vegetational units and the parameters, and to mapping large areas of the American tropics. Invariably the identi- fication of the life zone in the field for this mapping was confirmed by subsequent reference to climatic data. Temperature The measure of heat used in the Holdridge system is the sum of the average daily positiVe temperatures (Centigrade) divided by the number of days in the year, which is designated as the mean annual biotemperature. This differs from mean annual temperature in that only positive temperatures are used. In areas where the temperature never falls below 00 Centigrade mean annual temperature and mean annual biotemperature are the same, but where there are below 00 Centigrade temperatures, the two differ. This is an important distinction for the Holdridge system since at temperatures below 00 Centigrade vegeta- tive life is inactive, and this period of dormancy must be taken into 2 1 For example, it has been found that when an element is a limiting factor in plant nutrition, additions of the element up to the amount that could be utilized by the plant must be increased in logarithmic progression to obtain a sequence of equal increase in yield. 17 account in developing vegetative characteristics for correlation with the temperature parameters. As was mentioned earlier, these temperature values increase logarithmically to cover the range of temperatures found on earth. The earth's climate is divided into seven latitudinal regions from tropical to polar; tropical, low subtropical, temperate, cool temperate, boreal, subpolar, and polar. In addition, there exists a sequence of altitu- dinal belts rising above the basal, sea-level biotemperature belt of climates in each latitudinal region that contains mountains or plateaus reaching up into cooler and less-dense air. For example, in the tropical basal latitudinal region where there are mountains it is possible to find all six altitudinal belts: subtropical, lower montane, montane, subalpine, alpine, and nival. For each latitudinal belt away from the tropical toward the poles the number of altitudinal belts that can possibly exist is reduced by one. Thus, above the cool temperate latitudinal belt there can be only three altitudinal belts _ subalpine, alpine, and nival. What this means is that if you consider an area within the tropical latitudinal belt which is devoid of elevations much above sea-level, then it is impossible to have anything but the tropical basal life zones existing within that area. But if in the same area higher elevations do exist, it is possible to have any or all of the life zones existing within that latitudinal area. The existence of these altitudinal belts is important from a socio-economic point of view, as well as from an ecological point of View. It is wrong to conceive of the "tropics" as having a uniform, hot humid, wet climate, as has been done much too often in the past. There is a wide variety of climate within the latitudinal range 18 generally included in the definition of "tropic."13 In like manner, it follows that there is a broad spectrum of socio-economic differentiation based on, or related to, these differences in climate. For example, different climates certainly contribute to differences in demographic patterns since people have a preference for cooler, drier climates and avoid the hot, wetter climates. One other heat characteristic of importance to the Holdridge scheme is the presence (or absence) of killing frosts. In the field work of correlating vegetational changes with the climatic parameters, it was found that there were definite changes in vegetation when one moved from an area without killing frosts to an area with such frosts. Also, it was found that this change did not always occur at the same tempera- ture, nor did the critical temperature (killing-frost temperature) coincide with the logarithmically spaced climatic parameter. Moreover, in the wetter districts it was found that the line of vegetational change did not coincide with the line of killing frosts, but occurred at a lower level of elevation than the latter. Holdridge concluded that a combination of high humidity and pea; frost temperature could bring about a change in vegetation similar to that of a killing frost. Thus, the critical temperature line which provides the boundary between the two sub-life zones, subtropical (premontane) and lower montane (see chart on page 25 ), occurs over a range of biotemperature, and its 1%A recent conference on productivity and innovation in agriculture in the underdeveloped countries recognized this diversity to a limited extent when it considered four different climatic areas. The varying problems of these areas were studied, and recommendations made for them separately. (See David Hapgood (ed.), Policies for Promoti Agricultural Development, Center for International Studies, Massachusetts Institute of Technology, Cambridge, Massachusetts.) 1965. 19 position indicates an average value for frost occurrence. Precipitation Precipitation is the second major climatic factor determining life zones, and its measure is the mean annual total of water in millimeters which falls from the atmosphere as rain, snow, hail, or sleet. Dew and other water that condenses on the surface of the earth are excluded from this measure because standard precipitation measuring devices are not designed to measure them. As with the mean annual biotemperature values, the precipitation values delineating the life zones increase logarithmically, and include almost the entire range of precipitation found on earth. The exceptions are so rare that it was not considered necessary to include them. Moisture The third climatic factor determining life zones is humidity. Humidity is the relationship between temperature and precipitation, and one must be careful in trying to correlate either of these singly with humidity. As Holdridge points out, "There is often a bit of confusion in the linking of humidity directly with precipitation. Although within any given region or along a given temperature line, there is a direct correlation of humidity with precipitation, when the total world environment is considered, such is not the case. The same mean annual precipitation which gives rise to wet humidity conditions in the Subpolar Region or Alpine belt results only in arid conditions when it "14 falls in the lowland tropics. The reason for this is that whenever 14Holdridge, 0p. cit., p. 27. 20 the temperature is above freezing moisture is constantly being returned to the atmosphere by evaporation (from the soil and other surfaces) and by transpiration (the physiological return of water from plant tissues to the atmosphere). As Tosi reminds us, "...the supply of moisture available to plants is not exclusively a function of precipitation."15 Holdridge uses the potential evapotranspiration ratio as his measure of moisture (humidity). This ratio is the theoretical quantity of water that would be given up to the atmosphere from a zonal climate and a zonal soil by the natural vegetation of the area if sufficient but not excessive water were available throughout the growing season. It is determined by dividing the mean annual potential evapotranspira- tion in millimeters by the mean annual precipitation in millimeters. The mean annual potential evapotranspiration has been found to have a direct logarithmic relationship to mean annual biotemperature, and thus can be determined by multiplying the mean annual biotemperature of any 6 This ratio gives a reliable estimate of site by a constant, 58.93.1 the moisture conditions at any site, and makes possible comparisons of this factor at different sites. Again, as with the other factors, the limiting values of this factor increase logarithmically. There are nine humidity provinces delineated in the life zone system: semiparched, superarid, perarid, arid, semiarid, subhumid, humid, perhumid, and superhumid. 15Tosi, op. cit., p. 178. 16This logarithmic "base" is found very frequently in nature and seems to be a natural law. For example, the cell, which is the most basic unit of living things, divides itself logarithmically, and thus multiplies itself on this base. There are innumerable other examples. (See n. 10) 21 A potential evapotranspiration ratio of 1 indicates a situation where the amount of water evaporated and transpired back into the atmosphere is just equal to the amount of water made available in the form of precipitation. This is a rather ideal situation since there is neither a shortage of water with the resulting parched conditions, nor is there an over abundance of water leading to leached and eroded soils. In areas where this ratio is greater than one, the water made available is less than what is needed resulting in arid conditions, and where the ratio is less than one the water made available is more than is needed resulting in excessively wet conditions. These then are the global climatic factors that delineate the boundaries of the life zones. Obviously, any two of the three can determine the life zone. The following section on the life zone chart will show how these factors are integrated in determining life zone boundaries. The Life Zone Chart Figure l is a graphical representation of the three major climatic factors, and how they determine the 120 or so life zones on the planet earth.17 It should be kept in mind that this is a three- dimensional figure, which indicates latitudinal divisions, as well as altitudinal divisions. Considered latitudinally, the chart establishes life zone divisions as one moves from the equator to the poles in either hemisphere. Altitudinally the chart establishes life zone divisions as 17Holdridge and some of his associates believe that the system can be extended to other planets, as well as back in time on our own planet when climatic conditions were different from what they now are. 22 one moves to higher altitudes within a given latitudinal division. The mean annual biotemperature at sea level determines the latitudinal basal region. For example, if the mean annual biotemperature at sea-level is over 240 Centigrade, it indicates the tropical basal region, between 6° and 12°, the cool temperate basal region, etc. As the chart shows, the biotemperature values dividing the basal regions decrease logarithmically as one moves from the equator toward the poles, with the exception of the frost line with its related vegetational change. These same biotemperature values determine the altitudinal belts which are shown on the right hand side of the chart. A basal region can have only those altitudinal belts that lie above the minimum tem- perature limit for that region. Thus, one does not find the montane altitudinal belt in the cool temperate basal region, or the lower montane in the boreal basal region. The tropical basal region, of course, has all of the altitudinal belts associated with it. These altitudinal belts correspond with logarithmic changes in biotemperature as the chart indicates. A problem of nomenclature arises in connection with these altitudinal belts. The correct ordering of the life zone names is as follows: the basal region, the altitudinal belt, and then the humidity province. This presents no problem when the life zone to be named falls within the basal region for the area. For example, there is no problem in naming an area close to sea-level, with an average annual biotemperature of 100 Centigrade, an average annual rainfall of 750 millimeters, and a potential evapotranspiration ratio of .75. From the chart we can very quickly see that this is the cool temperate montane 23 moist forest life zone. However, if we take the same measurements of the climatic data, but at an altitude of 3000 meters, the basal region is no longer the cool temperate. It now becomes the trepical basal region, and the name of the life zone is tropical montane moist forest. As a rule of thumb, in determining the correct life zone, a correction factor of 60 Centigrade for every 1000 meters of elevation must be added to the mean annual biotemperature. For example, using the above data the correction factor would be 180 (3000/1000 x 60), which would give a sea-level biotemperature of 100 plus 18°, or 280 which falls within the tropical basal region. Hence, whenever measurements of the climatic factors are made at altitudes much above sea level, the tem- perature data must be corrected to sea-level to determine the correct basal region. Running diagonally downward from left to right on the chart are the humidity provinces, measured by the potential evapotranspiration ratio values. In the opposite direction, running diagonally from right to left, are the divisions based on rainfall, measured by the average total annual precipitation. Measurements of any two of the three climatic factors, along with knowledge of the altitude of the site, will permit the determination of the life zone. Using the same data as before for temperature (100) and rainfall (750 mm), one need only locate these values on the chart, and note their intersection to determine that this site is in the montane altitudinal belt and the moist forest humidity province. Then, with knowledge of the elevation, one can determine the correct basal region. The intersection of values for the potential evapotranspiration ratio and average total annual precipitation for a given site will determine 24 the altitudinal belt and the humidity province, flag will also determine the mean annual biotemperature which can then be corrected for elevation to determine the relevant basal region. A few moments study of the chart is all that is necessary to permit one to determine readily the life zone corresponding to data from a given site. The dotted line running horizontally across the chart at about 180 mean annual biotemperature should be noted. This is the frost line or critical minimum temperature line that was mentioned earlier. It should be remembered that this line can occur anywhere within the range of 120 to 24° Centigrade, depending upon humidity conditions, warm air currents, etc., which might affect a given site. The unity line of the potential evapotranspiration ratio should also be noted. Both of these have socio-economic importance which will be brought out later in this study. One other characteristic of the chart should be discussed. It will be noted from the chart that the representations of the life zones are hexagonal in shape, and that the guidelines for the average annual precipitation, the potential evapotranspiration ratio and the mean annual biotemperature are not coincident with the boundaries of the hexagons, but bisect them. This characteristic creates within each life zone a series of six small triangles. These triangles represent transi- tional zones, i.e., areas in which the life zones begin to transform themselves into the adjacent life zone. Within these areas the vegeta- tion begins to take on a different appearance, but the change is not great enough to remove the area from its life zone. It is of course, more difficult to distinguish the correct life zone in this area, but in moving from an area well within the life zone across these transi- tional zones, they are easily observable and can be readily mapped. g mluuzgoma ,>._._Q_2:I 3.9.5633 05:30 afarcuaam/ Blazcua / 0.2:: / 0.23:0...» / cisawm / o.¢< / 2:33.. I 2.319....» V ouxuccaiww/ :6 $5.3... .335 3.3. 2.0 3.0 3.. con . 8.. co.- 8... . .2 8.3 2 425.22 at» 0» >9. . . . . If :3 /IIIIIIIII/ IIIVIIII/IIIVIIIMIIIIVIV I / I/ I/ I/ I/I II IIIIII/IIIIIIII/IIII/III/IIII III/I / / // II/ // I/ I/I/I/VVVVVVVV//////////// / .04 .2338!» .38.. 8.: w o d a». 0/ a a I o a. co ac a. I... .e \\\\\\\\\\ \\ \&\\§\\§\§xx\xxxx\\.\\\\\\\\\\\\\\\ \ \ \ X x \ x \ x \ x: s s s \ I \ / s / \x I s / \s / x / \ \ mm m new :22 \ II :33. \ II :38 \ II .33.. \ II .33.. \ II 22233 \ .II :26 \ II :38 ¢?.ar .3213: Will.” u «3. I “W c_o¢\ I :3 \ to \ Ito €->\ I \ tea-o xx I /)/ ea c ”an. ..... an H «mm... lllllllllllll \\-.. In- I: ......... a..- {JuiIII ...... un..-!.!.-... H. m .2 I. r1.” 33.3533 \ . :...f I 3:: II \ :3... I x :28 I r/ 33:35.3... 30.: H V m 08. l.. M“ \\ so: .. Eel I \. to . I \ II ¢ ”W a I: l. C” 2.328. 533 ao\\. . . 1.. 55.3.3... www.mfimhrd: .:...:..) 955:3... as; ”M .n on. I H 00 \ =38 \ .33.; x I 3.8.x \ I :38 A/ .0 ”I M II \ so: . E... II A % In 4... 3. II Il.~.l llllllllllll \ lllll . lllllllll Al llllll lilllllloa.ll.m Wu“ H MH \\ .. , II » Hm I ”I \ III m u” I m.l| 2.320: s a. s. 225:3» .68 hm . n... a I s I a W a. II I t- o s \ IID E I. O \S I. a 23' III... IIIII ll lllllll l-.mu...l NI! II III lllll llllllilloellw .ou. I. an. x x I x .. I m x. l. W] .H \\\\§.2..muu.u II \s..oso.£. I 1H0. h I! 3.2.3...» n o\ e... :3: sauce- ]! I 00 \ Cs°¢ \ o I w \ x .I W I ell \ u M t. .| WI.» llllllllllllllllllll .ml: \l lllllllllllllll IJIIIIIIIIIBIIS n I w \ \\ I \ o | m M A. . x I, x I . I I a II a I um.e4< 0 Is 22:.» .32.... I 3......» 32:.» I 00 xdaoeoau I .: I MI W coo \\\ to: \ II :3 x .. :3: x I to 9%fla as IIM 9 .I x \ I \ .I x I . «I. W . alien nnnnnnnnnnnnnnnnnnnn 1M. lo unsure..--..L..U,.Il.n¢.-.,r.i- -...I. uuuuuuuuuuuuu --..------..._|. o \ . I . , . I II III. n a . .. .L I H ., r. I a I .. .w \ I I... . . .28. I: I: 44> . a 00 ac \ .I. .c. N 0* c 4 .- o a :.I ll- 0 s I . I l/ \ Ne II es II A/ 0/ a v \ M II \ II A/ O . M \ K II K II sf # .\ II es It s/ I \ o s II A 2... mm .3 ..... «I 928.. a \\ oat .0 I! 442535,. \ . .. I II, a. dessert... \ a \ \s or; ”a I\ .. o 0? \ )A/ .I .a on ..I. 33.2.: .c a. a. muzoN $3 45522 mo 39:28... Sin. osmo>> . co. 5.805320 9: c2 $960.5 . 25 26 THE VEGETATIVE CHARACTERISTICS In the Holdridge system, the climatic parameters specifically and precisely define the limits of the life zones. However, in order to be other than an arbitrary division of climate, the life zones must be "natural divisions of significance,"18 i.e., they must correspond as closely as possible to divisions made by nature in the earth's vegetation. Holdridge's basic hypothesis was that such natural divisions did exist, and subsequent work on the system has proved that there are such ecological differences between life zones. Anyone with a little training in the system, can learn to recognize the different life zones in the field. Large areas of the American tropics have been mapped on the basis of field observations,19 and where the availability of climatic data permitted and checks were run on the field work, it was invariably found that the areas had been correctly mapped.20 The vegetative characteristics used in discriminating between life Zones are more subjective than objective at the present time° The ecologists who are working with the system say it is the general look of the vegetation, or the "feel" of an area that indicates to them which 18Holdridge, op. cit., p. 9. 19Practically all of the work on the system has been carried out in the western hemisphere. However, after a recent reconnaissance trip through Africa and the Far East, Holdridge and Tosi believe that the system is equally applicable there. 201m fact in certain areas where the vegetation did not correlate with the climatic parameters, computational errors were discovered in the weather station data, which when subsequently corrected, provided the correct correlation. 27 life zone they are in. There are some quantifiable characteristics, however, that help one to arrive at this "feel" for the life zone. There are certain ”indicator species" for various life zones whose density of occurrence changes abruptly when one crosses the boundary of the life zone. The extensiveness and intensiveness of edaphic21 growth is another such factor, as is the measure of the cover heights of the dominant trees, bushes, and herbs. In areas where the natural vegetation has been completely cleared by man, similar characteristics of the cultivated vegetation indicate the life zone. Very recently, Holdridge and his associates have been introducing more quantitative measurements into their study of the vegetation. They have devised a "complexity" index which is a combination of measurements over various aspects of the vegetation. This index, along with some other measurements of relationships between certain characteristics, is now being correlated with the climatic parameters in order to determine a more quantifiable description of the vegetation within the various life zones. This data then will be analyzed statisti- cally to determine if the greater rigor is useful in describing the significant differences in the vegetation between life zones. The Life Zone System The life zone system then, is a method of classifying vegetation22 21An edaphic growth or plant is one that lives in parasitic fashion on another plant. 22This is not exactly correct since the proponents of this system contend that animal life and behavior, as well as that of humans, can be classified profitably on this basis. 28 on the basis of certain global ecological factors. These factors are the climatic factors of heat, precipitation, and moisture; and the vegetational units are groupings of plant associations. Extensive field work has shown this to be an effective and natural classificatory system. A life zone can be defined in terms of the relevant climatic factor values, but it may also be thought of as, "...a group of associations related through the effects of these three major climatic factors... These leave a definite mark on all the associations even though the group may comprise a quite diverse group of associations."23 Holdridge and his associates expect further research on the system to show that other natural science variables and social science variables are corre- lated with life zones. Costa Rica has been completely mapped on the basis of the Holdridge system. It is entirely within the tropical basal region, and contains the premontane (subtrOpical), lower montane, montane, and subalpine altitudinal belts. It is excellently suited for research on the life zone system because of the wide variety of climatic conditions existing in a relative restricted geographical area. Definitions of the Trapical Moist Forest and Trogical Premontane Wet Forest Life Zones The trOpical moist forest and tropical premontane wet forest life zones are the ones used for comparative purposes in this study. The trOpical moist forest life zone is in the tropical basal area and the moist forest humidity province. There is no altitudinal 23Holdridge, op. cit., p. 14. 29 belt associated with this basal region (as there are for all of the other basal regions) in the nomenclature, since it is impossible for this series of life zones to be found in an altitudinal belt correspond- ing to some lower altitude basal belt. The trOpical basal belt is at the upper limit of the biotemperatures found on earth (at sea-level), which precludes the existence of a hotter basal belt. In terms of the climatic parameters, the mean biotemperature of this life zone is above 24° Centigrade, it belongs to the humid humidity province with a potential evapotranspiration ratio between 1.00 and 0.50, and it has an average annual precipitation of between 2000 and 4000 millimeters. It is adjacent to the unity line of the potential evapotranspiration ratio, and at no time does this life zone feel the effects of a killing frost. It should be pointed out that the values for the climatic parameters given above describe the range of these parameters, and that there can be variations to the extent of this range between any two sites within the life zone. This is, of course, true of any of the life zones, and leads one to suspect that differences within life zones may be rather large from this factor alone. In Costa Rica, the tropical moist forest covers approximately 12,577 square kilometers, or 24.4 per cent of the total area of the country. This, of course, includes the transitional areas associated with this life zone, since the vegetative characteristics defining the life zone include these transitional areas. The tropical premontane wet forest life zone is in the tropical basal region, the premontane altitudinal belt, and the wet forest humidity province. Thus, the mean annual biotemperature for this life 30 zone is between 120 and 24° Centigrade, the average annual precipitation is between 2000 and 4000 millimeters, and the potential evapotranspira- tion ratio is between 0.50 and 0.25. The fact that the altitudinal belt is premontane (subtrOpical), and the basal region is trapical (and not low subtrOpical), means that this life zone exists at an elevation somewhat above sea-level. In Costa Rica the boundary between this altitudinal belt and the basal region generally occurs at an elevation of from 600 to 700 meters. As the chart shows, however, it is possible for the transitional area of this life zone (altitudinal belt) to extend downward into the basal region (i.e., at temperatures greater than 24° Centigrade), even to the extent of having this area reach sea-level. Thus, along the Atlantic coast of Costa Rica we find this transitional area at sea-level. In Costa Rica this life zone including transitional areas covers approximately 8,179 square kilometers, which is almost 16 per cent of the total surface area of that country. Thus, in Costa Rica these two life zones cover 40 per cent of the area of the country, indicating their importance to the economy of that country. The differential vegetative characteristics due solely to the effects of the climatic factors are difficult for the untrained, non» ecologist to discern, and there is practically no completed, systematic research on the differential animal and human behavior characteristics by life zone. Local factors play a big part in determining the structural and other characteristics of the vegetation that are used by the ecologists in recognizing a given life zone. In very general terms it is possible to describe the differential effects of the climatic factors on the various life zones, but it should be remembered that local 31 variations are very important and can have dominating influence on the Vegetation. In general, as one moves from the dry to the wet humidity provinces within a given basal region, the vegetation becomes more dense, the trees grow to greater heights, they contain more leaves, the stems are thicker, soils become more clayish and less sandy, etc. As one moves within the same humidity province to progressively lower temperatures, the vegetation becomes less dense, less tall, and takes on characteristics to permit it to withstand the colder temperatures. There are numerous of these characteristics that could be used to describe the differences due to climatic factors between life zones, and current research on the system is oriented to discovering what these characteristics are and how they differ by life zone. The most precise method of distinguishing between life zones is on the basis of the climatic parameters. These parameters delineate the life zones, and therefore can be used as a basis for correlating other variables with the life zones. In terms of the parameters, the two life zones chosen for this study are relatively similar, however9 there are some crucial differences. It was thought that for this type of study it would be best to have a modest amount of difference between the two life zones, i.e., not to choose the most similar, nor the most different life zones. The tropical moist forest and the tropical premontane wet forest life zones are similar in that they both have the same average annual precipitation of between 2000 and 4000 millimeters. They are different, however, in that the trOpical moist forest life zone has a mean annual biotemperature above 240 Centigrade, whereas the tropical premontane wet forest life zone has a mean annual biotemperature between 240 and / .l l/ - 32 (approximately) 180 Centigrade. This difference in biotemperature is thought by the ecologists to be of primary importance in distinguishing life zones. This difference in temperature, given the same mean annual precipitation, leads to a difference in humidity province - the trOpical moist forest being humid, and the trOpical premontane wet forest being perhumid. This difference in humidity province is reflected in different potential evapotranspiration ratios for the two life zones. The tr0pical premontane wet forest is one space removed from the unity line of potential evapotranspiration, whereas the trOpical moist forest is adjacent to it. What this means is that in the trOpical premontane. wet forest life zone the potential for evaporation and transpiration (from the vegetation and other surfaces) related to the amount of water made available (through rainfall, etc.) is much lower than in the tropical moist forest life zone. In other words, the evaporation and transpiration'gg precipitation ratio is lower in the tropical premontane wet forest life zone, and thus there would be more likelihood of excessive water conditions here than in the tropical moist forest, which has the same average annual precipitation. In addition, since the basal region for both of these life zones is "tropical", this means that the premontane wet forest life zone must occupy an area with an elevation somewhat above sea-level. Therefore, except for limited areas near sea-level along the Atlantic coast, this life zone will be found at altitudes ranging from 600 meters to close to 2000 meters. The fact that this life zone is an altitudinal belt above a basal region has important implications for farm technology, etc., as will be seen in a later chapter. .Another factor of some importance to the premontane wet forest 33 life zone is that it borders on the critical minimum temperature line - the frost line. There is a distinct change in vegetation where this line occurs, even though it is not logarithmically spaced from the other boundaries, and there is a relatively wide band of biotemperature over which the line ranges depending on more localized conditions. The ecologists working with the system like to combine this premontane wet forest zone with the lower montane wet forest zone, and to consider the entire area as one life zone. In effect they are considering the change in vegetation that occurs along the critical minimum temperature line to be of less importance than the changes that occur at the logarithmically spaced climatic parameters. They consider the change to be due more to local factors than to the global factors. Thus Holdridge says: . "Theoretically the best solution appears to be that of maintaining the complete hexagonal divisions as life zones and to consider any ecosystems within the Warm Temperate region or Tropical Lower Montane belt which are free from killing frosts as well as any portion of the basal tropical region with occasional freezing temperatures as atmospheric association only. Fundamentally, the change involved may be basically no more signi- ficant physiologically to plants than that of the monsoon or Mediterranean climate variations caused by differences in the annual rainfall pattern."24 He goes on to say that it is of practical importance to recognize this change in vegetation because of its seeming importance in many of man's activities, but he adds that as yet it is difficult to know how to handle this change in vegetation since very little research has been done on the effects of killing frosts.25 24Holdridge, op. cit., p. 22. 34 In choosing the life zones for this study it was thought that the changes resulting from the existence of killing frost could obscure or negate the changes that might otherwise be noted, and so it was decided to choose only that part of the life zone that was without frosts. Therefore, the premontane wet forest life zone was chosen, rather than the combination of this and the lower montane wet forest zone. THE LIFE ZONE SYSTEM AND ECONOMIC DEVELOPMENT The developers of the Life Zone System consider it to be an important classificatory system for classifying animal and human behavior as well as for classifying vegetative characteristics. Human behavior is thought by them to be classifiable by life zone from several different points of view, e.g., demographic, racial, political, economic, and social or cultural, some of which Holdridge discusses briefly in his Life Zone Ecology. This study is, of course, primarily interested in economic aspects of the Life Zone System about which very little research has been done.26 Holdridge and his associates believe that all economic variables are influenced to some degree by the climatic parameters that delineate the life zones. Moreover, they feel that the degree of this variation for most of these variables will be great enough that profitable 26However, see Joseph A. Tosi, Jr. and Robert F. Voertman, "Some Environmental Factors in the Economic Deve10pment of the Tropics, "Economic Geography, Vol. 40, No. 3, July, 1964. Also, Joseph A. Tosi, Jr., Zonas de Vida Natural en el Peru: Memoria Explicativa sobre el Mapa Ecologico del Peru, y Mapa Ecologico del PeruI Boletin Technico No. 5 del Institute Interamericano de Ciencias Agricolas de la O.E.A.Zona Andina, Programa de C00peracion Tecnica, Lima, Peru, 1960. 35 economic decisions can be made on the basis of them. They are rather strongly convinced that the life zone system has an important contri- bution to make in the area of agricultural economics, but they have no concrete evidence to support this conviction. A more specific dis- cussion of their position concerning the relationship between agriculture and the life zone system will be presented later in this study. Before doing so, however, there are other areas in the general field of economics which they feel to be related to the life zone system, and which should be mentioned here so as to capture the full breadth of the role they foresee for the Holdridge System. The ecologists who have been working with the Holdridge System think that the major immediate contribution it can make to economics is in the determination of correct land-use systems. Using an extremely long run and ecological concept of land use, they feel that the existence of life on earth for the next few decades depends upon man adapting more to the environment rather than adapting the environment to himself. They think that the first task facing man is to determine which land use is ”natural" for each area, i.e., which uses can the land support over an indefinite period of time without permanent damage to its productive powers. Thus, if a given area will perpetually support only the natural forest, then according to this concept of land use, it should be used only for this purpose. Agricultural land, land for grazing cattle, watershed areas, and other uses of the land would be determined on the same basis. All of these various land uses could be correlated with life zones, so that one could eventually come up with a classification of land use such that simply by knowing the values of the climatic parameters, one could determine the "correct" land use for an area. 36 This could then become the basis for policies and planning to insure an everlasting ecological system for the earth. Since a corollary to this proposition is that man is rapidly destroying large areas of natural forest that had best be left to (or exploited as) natural forest, a problem arises concerning how to deal with the mounting pressure of pOpulation on the land. If large areas that are now being forced into agricultural production are to be left as natural forest, one of the most important means of staving off widespread famine (the opening up of new lands to agricultural produc- tion) will be lost to many countries. There are three methods for solving this problem: 1) to limit population growth, 2) to increase agricultural production on "natural"27 agricultural land, 3) to change the dietary habits of the people so that they will consume those products "naturally" produced. Of these three, the first and third are stressed by the ecologists as solutions to the problem, while the second is relatively neglected by them. A major objection to these two approaches is that they age long run solutions, or rather solutions that can be implemented only in the long run, whereas the problem is one that requires a stop-gap short- run solution. Hence, the immediate objective should be to concentrate on the second of the three alternatives, while permitting whatever inroads into the natural forest might be necessary to stave off famine 27Natural agricultural land is land that can support intensive agricultural production over very long periods of time without permanently destroying the productivity of the land. This does not preclude the use of fertilizer to maintain this productivity. Also, this concept should not be confused with the natural vegetative cover that would exist on the land if it were left in an untouched state. 37 until solution number one, and possibly number three, can be implemented. This seems to be the only rational and politically feasible approach, since before the other two alternatives could be implemented there would be widespread famine and possibly anarchy. Nor can one arbitrarily assume short of a totalitarian system that either of the two solutions favored by the ecologist, will work. Population control has been talked about for a number of years but very little has been done in the way of initiating it. Only in recent years have attitudes started changing in the direction of acceptance of birth control, and there are still major power blocks aligned against it. Solution number three seems even more difficult to imple- ment. The argument is that there must be a change in dietary habits away from the grains, and toward the fruits, fibers, tubers, etc., that abound in the natural forest. This shift in taste would be very difficult to accomplish in a short period of time, and perhaps impossible even in the long run. The other alternative is to concentrate research on the possibia lity of increasing productivity on "natural" agricultural land, and to gain time by permitting agriculture to expand temporarily into the less favorable areas. This time could be used to introduce changes in technology that are necessary to high productivity agriculture in tropical areas. This proposal has several advantages. It could achieve results in a relatively short period of time, which is a crucial factor given the increasing pressure on food supplies. It could very well permit the type of increase experienced by agriculture in the United States when research was finally directed toward this end, and it could very well allow the reversion of land not permanently 38 suited to agriculture (and perhaps some of the marginal agricultural land) back to its original vegetative state. This plan is also much more feasible, politically as well as practically, than either of the other two alternatives. Thus, the ecologists fail to consider the market as the regulator of production. They favor an environmental determinism which in effect would subordinate the price system to an ecological system answering the questions of what, when, who, where, and how. There would, of course, still be a certain amount of land used for agricultural output, but the main burden of the ecologists approach rests on making much greater use of the foods and fibers that nature provides. In other words, let nature grow what it can naturally and concentrate on finding ways to consume and otherwise use this produc— tion. The life zone system should be able to provide us with a method of determining which areas to convert to agriculture, which to use for lumber production, which to leave with their original growth, etc. The ecologists feel that if production is not oriented on this base, the productive powers of the earth will soon be severely diminished. Demographic patterns are thought by the ecologists to be determined by life zones; that is, the climatic parameters affect the desire and the ability to live in certain areas. There should be variations in population pressures by life zone, with the accompanying variations in land tenure systems, unemployment and underemployment on farms and in the cities, public works and welfare programs, housing conditions, wage rates, rental rates, etc. Throughout the history of any given country the people should have populated first those areas -:-.._.__—.Q.b“ t} -._..-: ..-'“-“- a" 39 within the more favored life zones. Holdridge points out that early settlements in the western hemisphere were in the more favored life zones (cool dry life zones), and that most of the capital cities in that hemisphere lie in life zones adjacent to the unity line of the potential evapotranspiration ratio - the area most favorable for agricultural develOpment. Thus, he concludes that the people favor these areas because they are more comfortable to live in and they are easier to make a living in.28 This hypothesis could be tested in various ways, and this research could be used to develop policies for the newly develOping, still underpOpulated countries. Also, due to the deleterious effect of the climatic parameters (especially rainfall) on roads in certain life zones, the ecologists feel that transportation problems will be greater in these zones. This could have important implications for plant location, the distria bution of public funds for road construction, road maintenance, the private and/or social returns to private and/or social investment in the various areas, the technology of production along with the costs of production, the establishment of enterprises and subsidiary enter= prises in the transportation industry, the pricing of materials and products, and the many other variables too numerous to mention. Many of the less developed countries in the trOpical region are faced with certain areas of heavy rainfall which completely destroys the roads that do exist, and these usually are repaired by the central government only after a considerable lapse of time. In many of these areas the funds for road maintenance are channelled far too much into the urban 28Holdridge, 0p. cit., pp. 15-16. 40 areas, with continuing neglect of the poverty-stricken rural areas. A knowledge of the variation by life zone in this tranSportation vari- able would help in planning the distribution of these funds on a more equitable basis. A given industry or firm could be strongly affected by the climatic parameters governing the area within which it has located its plant or sells its product. Costs, prices, factor proportions (and therefore marginal productivities of the various factors), the necessity for and the type of maintenance, marketing times and methods, spoilage, storage, depreciation, etc., could all be related to life zones. For example, heavy rainfall and high temperatures could force a manufacturing firm to use a higher grade of construction material, resulting in higher costs of construction in a given life zone. Many other examples of similar influences come readily to mind. The price and availability of labor, could vary by life zone depending upon the degree to which people favor one or the other life zone, and assuming, of course, a free labor market. A policy to bring the quantity of each factor of production into line with the varying optimum factor proportions as determined by the life zone system would be of great value to some of the less developed countries that today have inefficient factor markets. These are some of the more important economic variables which the ecologists feel are influenced by heat, precipitation, and moisture, and they are subject to testing to determine if a significant variation exists in them by life zone. If significant variations between life zones were found, policies based on this discovery could prove useful for economic development. 41 The ecologists position on the relationship between differences in the climatic parameters and differences in variables surrounding agriculture is more specific and held more strongly. It is in this field that the climatic parameters determining the life zones probably have their most direct effect on economic variables. Here heat, precipitation, and moisture are important growth factors that deter- mine the basic productivity of the land, whereas in industry and commerce the effect is more indirect, working through such factors as depreciation, spoilage, and transportation. In addition, given the importance of agriculture to the less developed countries, this is one of the most important fields in which research on the life zone system should be conducted. The growing food crisis demands that more research be directed toward improving agricultural productivity, and the life zone system provides a possible means of doing so. Agricultural economists and ecologists would agree that there are many variables within the field of agri- culture that could be analyzed for variation by life zone, but they would probably not agree on the order of priority for the investiga- tions. A few of these which are mentioned most often in the writings and discussions of the ecologists will be discussed here to provide some idea of the type of relationship expected to exist between the climatic parameters and agriculture. Productivity, measured as the number of units of output per unit of land, is a very important variable which should differ by life zone. For each individual crop there is a theoretical optimum of the three climatic factors which would provide, other things equal, the highest average productivity per given unit of land. Since the climatic 42 parameters differ by life zone, the productivity per acre of corn (with all other factors, such as the amount of fertilizer, labor, seed, shade, held constant) should differ by life zone. Of course, these other factors that are held constant must not include ppy artificial additions to the climatic parameters (water or heat) as this would negate the purpose of the research which seeks to determine the effects of the climatic parameters as provided by nature. From a practical point of view this is the agricultural variable that should receive top priority for research. Extremely low producti- vity is characteristic of the underdeveloped trOpical countries, and any quick means of improving productivity would be greatly welcomed. Agricultural technology, or the method of producing, is another variable that should vary by life zone. The concept of technology includes the amounts of the factors of production that are used, and the way in which they are combined to produce a unit of output. The optimum amounts of the factors and the optimum combination could be expected to vary by life zone. It also includes the operations performed by these factors in producing a given crop, as well as the timing of these operations. There is a considerable amount of detail that could be investigated in depth between life zones, but time and foreseeable benefits limit these to relatively few. As a corollary to differences in technology, there should exist differences in costs of production between life zones. Costs of pro- duction per unit of output or per unit of a given input should vary by life zone because of the different amounts of the factors of pro- duction used which in turn are caused by the different amounts of heat, precipitation, and moisture, assuming relative prices of the 43 factors of production are the same in the life zones. For example, one life zone that is hotter and wetter than another might require shading or a drainage system, or more fertilizer, resulting in higher costs of production for that life zone. Irrigation might be required by another life zone receiving less rainfall, which would again increase costs. The point is that there should be differences in costs of production for a given crop by life zone which are measurable and of some use economically. Variations in the optimum amounts of fertilizer to use on a given crop per given land unit might be significant between life zones, as might the correct spacing of plants and rows for a given crop. Rotation systems, the possibilities of intercropping and mixed farming, the best growing pasture for cattle, and a host of other variables might profit- ably be compared by life zone. If it were found that there were significant differences between life zones in some or all of these variables, then planning and policy could be designed to capitalize on these differences. For example, if it were found that there were differences in productivity between life zones for the various crops grown in an economy, the government could encourage the production of that crOp that grows best in each separate life zone. This is simply allocating resources into those areas which can most economically use them, i.e., according to marginal principles. This assumes, of course, that it is necessary to so direct resources. Schultz29 feels that in traditional agriculture resources are 29Theodore W. Schultz, Transformipg Traditional Agricultural, (New Haven and London: Yale University Press, 1964), See Chapter 3, pp. 36-52. 44 efficiently allocated, given the effective decisions that are available to the farmers. This study is concerned with testing the Holdridge Life Zone System with respect to its relationship to certain economic variables. One must accept (through ignorance of the field) the biological vali- dity of the system, so that point is not in contention in this paper. What is tested is the hypothesis that there are significant variations in economic magnitudes among life zones. Before proceeding to the methodology to be used in these tests there are two points which need further discussion to set the stage for the analysis which follows. The first of these is the difficulty of holding certain factors constant when one wants to compare the relationship between only two variables. This is especially difficult in a study such as this where some of the variables to be compared are very general in application, whereas some of the variables to be held constant are of local applica- tion. The climatic parameters of the Holdridge System affect large areas, while some of the factors assumed to be constant (such as sloPe and wind) vary considerably within and among the life zones. It is much easier to compare the more local variables while holding constant the more general ones. For example, to determine the productivity effects of the climatic parameters (life zones), it would be best to have the same technology used in each of the life zones being studied. If this condition does not exist, differences in productivity could be due to the different technologies, rather than to the different clima- tic parameters. The same is true of any of the applied factors. If the amount of fertilizer applied differs by life zone, this could be responsible for any differences in productivity that might be discovered. 45 Thus, to obtain a true picture of the influence of the climatic para- meters, all other factors should be held constant. The same is not true, however, of factors that are themselves directly influenced by the climatic factors. The best example of this is the effect that rainfall and heat have upon soil. The soil in a given area will vary in fertility depending upon the amount of heat and rainfall it receives, as well as upon their periodicity. The greater the amount of rainfall and the more concentrated it is, the greater will be the leaching of the soil, and the lower will be its fertility. But this is a direct or natural effect of the climatic parameters, and should not be eliminated (or held constant) in measuring differences in productivity by life zones. The ecologists argue that to get a true picture of the effects of the climatic parameters in isolation, it is necessary to hold constant all exogenous factors, and would extend this to include differences in soils that were due to local conditions. Thus, if there had been recent volcanic activity within a life zone that affected the soils of only a part of the life zone, this would have to be taken into account. They would treat variations in topography in the same way. In their View, not to eliminate variations in t0pography, would be to falsify the differences found between life zones since these topographic differences are not caused by the climatic parameters. The end result of this view is to reduce the comparisons to very localized areas, and to conclude that little affective research on the life zone system itself is possible. The only possible solution is to proceed with the comparia sons on the assumption that these things are constant or that they offset one another when aggregated over the entire life zone. 46 The other matter to be disposed of before proceeding has to do with the fact that what might be a statistically significant difference in a variable might not be an economically significant one and vice- versa. In other words, statistical tests might indicate a difference is significant, but when the difference is translated into its effect upon the economy, it may have a negligible potential for improving the economy and thus be economically insignificant. Presumably, if a difference is statistically significant - (i.e., if it is not due to chance - it is a difference that can be eXpected to occur repeatedly over a considerable period of time, making it significant - in the use of resources for example) in the long run if not in the short run. Besides, where improvement is so needed and so hard to obtain, the smallest difference would become of importance. The point is that what one considers economically significant must be related to the need for improvement. Under certain conditions even the smallest variation can be important. CHAPTER III METHODOLOGY The basic hypothesis of this study is that in certain economic variables there are significant differences between life zones. This hypothesis includes practically all economic variables, not only those associated with agriculture where the influence of the life zones is most evident. The purpose of this study is to test in a limited way, the life zone system as a classificatory system for economic variables. Since the most direct effect of the climatic parameters on the economic system is in agriculture, this is the most logical area within which to begin research on the system. The system suggests that productivity and technology are influenced by the differences in the climatic parameters and that variations in these by life zone should be measurable. Since the hypothesis is all encompassing, it was necessary to limit the sc0pe of this study to keep it within the bounds of the resources available. Thus, it was decided to concen- trate on only one crop in two life zones. A stratified random sample of farms was selected from the two life zones, and data on productivity and technology were collected for these farms. The data for the two life zones were then compared and analyzed to determine if differences did or did not exist. In addition, several farms, some in each of the two life zones that were 47 48 growing corn, were visited and the farmers interviewed concerning their methods of production. These data and observations are also included in the analysis. Some agricultural extension agents and delegates from the Consejo Nacional de Produccidh were also inter- viewed concerning agricultural production and practices within their area. The word "significant" which appears above-in the statement of the hypothesis is used in two different senses. First, a "difference" in a variable between life zones, computed from a sample from each, is defined as being significant if it could not likely have happened by chance from two samples taken from a given pOpulation, but rather indicates that there are two distinct pOpulations from whence the samples came (each represented by one of the life zones). Thus, where it was possible to analyze the differences statistically, a .05 level of probability was accepted as determining if the differences were significant. In the many cases where the data did not lend itself to statistical analysis, they are analyzed more subjectively and presented on that basis. In this case, a difference is considered significant if it seems important enough that it could contribute to improving the performance of the economy. Selection of the Life Zones There were three primary considerations governing the selection of the life zones. The first was that the life zones chosen had to exist in Costa Rica where the research.was to be carried out. There 49 are twelve life zones in Costa Rica.1 The second consideration was that the life zones selected should have a moderate amount of difference between them. There is a continuum over which the differences in the various combina- tions of life zones range, i.e., between some of the life zones the difference is small and between others it is rather large; and there are intervening differences of varying degree. The ecologists position is, of course, that the variation in the variables being measured will correspond in degree to the extent of the difference between the climatic parameters. In other words, the greater the difference in the parameters delineating the life zones, the greater will be the difference in the variable being measured. This hypothesis would seem to be true, but needs to be verified. The choice of the amount of difference on which to base this study was made arbitrarily, for theoretically any two of the life zones could have been compared. Those life zones showing the greatest difference were not chosen so as to avoid the appearance of having "picked" the life zones in such a way that the possibilities of finding differences were maximized. The choice of two life zones with a minimum of difference was avoided for a similar reason. If these had been chosen, it would have appeared that the study had been designed so as to not find differences between life zones. Both such 1The twelve life zones that exist in Costa Rica are: tropical dry forest, tropical moist forest, tropical wet forest, tropical premontane moist forest, tropical premontane wet forest, tropical premontane rain forest, tropical lower montane moist forest, tropical lower montane wet forest, tropical lower montane rain forest, tr0pical montane wet forest, tropical montane rain forest, and tropical subalpine rain parame. 50 selections were excluded by the choice of life zones with a "moderate" amount of difference. The last, but by no means least important factor in choosing the life zones was their extensiveness in Costa Rica and in tropical areas generally. There are two reasons why the life zones chosen should cover large areas. First, to obtain the widest applicability for the results, the farms chosen should be representative of a large number of farms in the country; and second, the coverage of large areas would insure the representation of most of the geographical variation in the variables being measured. Thus, the conclusions would be important for reason of their applicability to a wide geographic area. In Costa Rica, the two life zones chosen (tropical moist forest and tropical premontane wet forest) account for 40 per cent of the total area. Moreover, these two life zones are important throughout Central America, covering approximately 54 per cent of the land area of that isthmus.2 There are also large areas of these two life zones in the South American countries. Since there are common characteristics for each of these life zones wherever they are found, any variations based on these characteristics should exist wherever the life zones exist. Thus, these variations and policies based on them are applicable in many different places on the globe. For example, if research on the life zone system were to show that in certain crops given factors were more productive in the tr0pical moist forest life zone as compared with the tropical premontane wet forest life zone using the same technology, the agricultural extension' 2Tosi and Voertman, op. cit., p. 194. 51 service could recommend to the farmers those crops best suited for these respective areas. Or if technology varies from one life zone to another for a given crop, these differences could be used to increase productivity. In other words, it would be possible to bring about a more efficient allocation of resources within and between countries if differences between life zones were found in productivity and technology. Selection of the Crop The vastness of the research needed to improve agriculture in the tropical areas is staggering, even from a purely technological point of view. If the economic aspects are included, the obstacles seem insurmountable, and where to begin the research becomes a serious problem. An argument can be made, however, for concentrating the earlier research on those crops that are a part of the basic diet of the pe0ple. With the rapidly rising world population it is impera- tive that food supplies be increased. I decided, therefore, to concentrate on the basic foods of the Costa Rican diet: rice, corn, and beans. Furthermore, the scarcity of resources required that, except for the measurement of productivity, the study be limited to only one of these. Corn was chosen because it is grown rather extensively in the two life zones, and because it has a better defined technology than either rice or beans, which facilia tates the measurement of that technology. For example, the various operations involved in growing corn occur in more distinct steps than is true of the Operations surrounding the growing of rice or beans, and the use of the factors of production (especially capital) is more 52 varied in the production of corn. Thus, in most of its aspects this study concentrates on corn with the exception that productivity by life zone is compared for each of the three basic crops - corn, rice, and beans. Of course, for the variable land-use, all possible uses of the land are considered. Corn is not a profitable crop as it is now grown in Costa Rica. Only about 40 per cent of the corn produced in that country in 1963 was sold, with about 52 per cent of the production being consumed on the farms by the producers and their families. The remainder was used on the farms for seed and forage. The technology in use in the production of corn in Costa Rica does not produce yields great enough to make corn commercially profitable when compared with some competi— tive crops. Some of what is sold is the result of the farmer selling the surplus production that his family does not need for consumption, or of his planting corn in a field to help keep the weeds down or the insects out. Corn also requires a relatively small cash outlay compared with some competitive crops, which makes it popular with some farmers even though it returns little in the way of profit. Many farmers are more interested in maximizing the difference between cash outlay and cash return rather than the difference between total cost and total revenue. Farmers with large amounts of land seldom devote more than one or two manzanas3 to corn. When interviewed both large and small farmers reported the lack of profit as the reason for their not growing more corn. Most of the large producers of corn in Costa / Rica produce for the Consejo Nacional de Produccion with whom they 3One manzana is approximately 1.7 acres. 53 have purchase contracts at relatively high prices, and for whom they produce seed corn. This seed is then sold by the Consejo to other farmers for planting. There is, however, some evidence that new technology would permit a substantial profit per acre. But this technology is not now "effectively" available to the farmer in the sense that he knows of it and can see even a remote possibility of using it. The problem is to get the small farmer to recognize these other opportunities and to provide him with the means of them. The lack of cash coupled with the high cost of introducing new technology and the absence of effective agricultural credit makes it difficult for the farmers to introduce new technology. The loss of foreign exchange due to the necessity of importing corn to meet domestic consumption requirements represents a real loss to the economy of Costa Rica. In 1963 there were approximately 2,170,000 bushels of corn produced on about 131,000 acres for an average production of around 16.6 bushels per acre. Most of this was grown on very small patches of land using very primitive technology. In addition to this production, Costa Rica imported 437,700 bushels of corn in 1964 which raised total consumption of this grain to around 2,607,000 bushels. Thus, imports were nearly 17 per cent of the total consumption, and represented an expenditure of over half a million dollars of valuable foreign exchange. Between the years 1951 and 1964, Costa Rica imported 3,603,336 bushels of corn at a total value of $6,640,850. While corn represents a relatively small percentage of the total imports of Costa Rica, productivity could be improved relatively easily to make the country self-sufficient in this crop. 54 Selection of the Sample The selection of the sample of farms was made relatively simple because a stratified random sample of farms producing rice, corn, and/ or beans was recently taken by the Costa Rican Department of Census4 for a study it was conducting. Their study was designed as a between census check on the data from the 1963 Census of Agriculture which theoretically entailed 100 per cent coverage. Except in certain small areas of Costa Rica, the expansion of the data obtained indicated that the Census of Agriculture of 1963 was accurate.5 DESCRIPTION OF THE DEPARTMENT OF CENSUS SAMPLE For the purposes of the Census study, the 36,417 farms producing either rice, corn, and/or beans were stratified on the basis of size to improve the reliability of the estimates. The farms were broken down into large, medium and small, with the boundary between medium and small falling at five manzanas (8.5 acres) regardless of crop or province. The upper limit of the medium-sized farms, or the boundary / 4Jose, G. Baptista and Mario Murillo M., Diseho de la Encuesta A ricola or Muestreo de Arroz, Mpiz y Frijol, 1965, Direcci n General de Estadistica y Censos, Republica de Costa Rica, December, 1965; and Jose G. Baptista and Mario Murillo M., Encuesta Agricola por Muestreo Arroz, Maiz y Frijol, 1965, Direccidn General de Estadfstica y Censos, Republica de Costa Rica, May, 1966. The first of these details the design of the study of these crops, and the latter gives the results. 5Underenumeration was anticipated in the 1963 census because of the lack of training and involvement of the enumerators, i.e. They lacked the initiative necessary to insure a complete enumeration. However, expansion of the data from this interim study indicated only one province in which there was some doubt that the area had been adequately covered in the 1963 census. 55 between the medium and the large, varied by province because of the variation by province in the average size of farms. This upper limit also varied by crOp; that is, the upper limit of the medium-size farms within a given province varied by crop. Thus in one province for rice, corn, and beans, this upper limit was 40, 30, and 15 manzanas (68, 51, 26 acres) respectively. This difference allows for the fact that the average size of planting for the various crops varies among the provinces. Table 1 shows the limits by province and by crop separating the large from the medium farms. Table 2 shows the breakdown into the three size categories of the farms producing these crops, and indicates the area planted in the crOps for each of the categories. The large farms comprise only .4 per cent of the total number of farms, but contain 19 per cent of the total area devoted to these crops, whereas the small farms represent 89.1 per cent of the farms, but only 44 per cent of the total area. The medium and large farms together are 11 per cent of all farms, with 56 per cent of the total acreage. After dividing the farms into large, medium, and small on this basis, Census officials decided to include in the sample all 159 farms designated as large, primarily because of the large percentage of the total area planted in the three crops accounted for by these farms. They adOpted a 10 per cent sampling fraction for the medium-sized farms because this seemed to give them the best balance between the cost of the survey and the validity of the results. A larger sampling fraction would have given them a more valid sample, but at a disprOportionate increase in the cost of the sample. They used a 5 per cent sampling fraction for the small farms for similar reasons. The methods used 56 by the Costa Rican Department of Census to extract these two samples were similar. TABLE 1 MINIMUM LIMITS FOR SEPARATING LARGE AND MEDIUM SIZE EARMS BY CROP AND BY PROVINCE (IN MANZANAS) Provinces Rice Beans Corn San Jose’ 20.0 20.0 20.0 Alajuela 40.0 40.0 20.0 Cartago 15.0 15.0 20.0 Heredia 40.0 15.0 30.0 Guanacaste 50.0 50.0 50.0 Puntarenas 50.0 30.0 50.0 Limo’n 20.0 15.0 50.0 Sources} José’G. Baptista and Mario Murillo M., Disefib de la Encuesta Agricola por Muestreo de Arroz, Maiz y Friidl, 1965, Direccidn General de Estad stica y Censos, Republica de Costa Rica, December, 1965; and Jose’ G. Baptista and Mario Murillo M., Encuesta Agricola por Muestreo Arroz, Maiz y Frijol, 1965 Direccion General de Estadiética y Censos, Republica de Costa Rica, May, 1966. 57 TABLE 2 CLASSIFICATION OF FARMS INTO LARGE, MEDIUM, AND SMALL Area under the Type of Farm Number of farms three crops (in manzanas) Absolute Percent Absolute Percent Large farms 159 .4 32.600 19.0 Medium farms 3.780 10.5 62.600 37.0 Small farms 32.478 89.1 73.200 44.0 TOTAL 36.417 100.0 168.400 100.0 Sources; José'G. Baptista and Mario Murillo My, Diseno de la I Encuesta A ricola or Muestreo de Arroz, M512 y Frijol, 1965, Direccion General de Estadlstica y Censos, Republica de Costa Rica, December, 1965; and Jose G. Baptista and Mario Murillo M., Encuesta Agricola por Muestreo Arroz, Maiz y Frijdl, 1965, Direccidh General de Estadistica y Censos, Republica de Costa Rica, May, 1966. For the medium-sized farms the first step in stratification was to obtain the provincial area under each crop. Then the provinces were classified according to the two crops most extensively cultivated in one of the following three groups: rice-maize, maizeubeans, and rice- beans. Next the farms were arranged in decreasing order of area under the most extensively cultivated crop, and divided into groups of 40 farms. The farms in each group were then arranged similarly for the second most extensively grown crop, and then divided into groups of 58 20 farms. From each of these groups two farms were chosen9 resulting in a 10 per cent sample of the medium—sized farms. For their sample of small farms, a multi—stage probability sample6 was designed, with the ultimate sampling units7 containing approximately 10 farms. In effect, the procedure here was to select randomly a number of small geographic areas with recognizable boundaries, each of which contained about 10 farms, and then to inter- view all the small farmers within those areas. There was a total of 163 ultimate sampling units selected, with an expected 1,624 farms within them to be enumerated. This resulted in a 5 per cent sampling fraction for small farms. The method used to select the ultimate sampling units was as follows: The political districts of Costa Rica were designated as primary sampling units,8 and in each province homogeneous groups of these were formed into strata based on the average area per farm for each of the three crops: rice, corn, and beans. In carrying out this process, a factorial design was adopted using the factors am, ar, and 6For a more detailed description of the method used in selecting this sample see Jose G. Baptista and Mario M. Murillo, Diséfio de la Encuesta Agricola por Muestreo de Arroz, Mgiz y Frijél, 1965, 0p. cit., p. 10. 7An ultimate sampling unit is a small geographic area containing approximately 10 farms, and the total of these make up the universe of the last strata from which a sample (in this case of 5 per cent) is selected. (see the following footnote) 8The political districts of Costa Rica were designated as primary sampling units. A sample of these units was selected, and then from this sample a sample of the ultimate sampling units was selected. In other words, the multi-stage probability sample method provides for the selection of a sample from a sample. 59 ab, (where am is the average area per farm for corn, ar is the average area per farm for rice; and a is the average area per farm for beans). b The final stratification did not appear with eight strata corresponding to the eight factor combinations since a restriction.was imposed on the size of the strata. This restriction was that each strata should con- tain at least fifty ultimate sampling units so that in selecting with a 5 per cent sampling fraction at least two sampling units per stratum would be in the sample. Thus, if any of the above eight strata had less than 500 farms (since 500 4.10 - 50 ultimate sampling units which is the minimum number possible under the limitation imposed), the corresponding districts were distributed among the other strata. In most cases only three strata were finally left in each province. Then within each province the selection of primary sampling units was done with probability prOportional to the number of small farms, and in such a manner that the sample of ultimate sampling units was proportionally allocated among the strata within the provinces. The selection of the ultimate sampling units was made independently within each sample district. The first step was to arrange census seg- ments into ultimate sampling units as nearly as possible. This meant that when a census segment had less than 10 farms it was grouped with one or more others to form a unit. These were then listed according to their geographic proximity by using the census district maps. Then the selection of segments or groups of segments within districts was made randomly with probability prOportional to the number of ultimate sampling units in the district. Once this was completed, one ultimate sampling unit per census segment was chosen at random. Some census segments contained only one ultimate sampling unit and thus there was no 60 selection problem. When there were more than one ultimate sampling unit within a census segment, the segment had to be sub-sampled. This usually involved dividing the segment into ultimate sampling units on the basis of recognizable boundaries on the census segment maps and randomly choosing one of them. Where there were no recognizable boundaries on the maps, the segments were divided into parts containing the expected number of farms and one of these was then selected. When even this was not possible, a complete list of the farms within the segment was made by a worker in the field, and a random sample of these taken with a number of farms chosen equal to the number in the other ultimate sampling units. In addition to the above sample, a small sample (1/25 sampling fraction) was taken of those districts not reporting these three crOps in the 1963 census. This was done as a check on faulty coverage by the 1963 census. The very simple method of listing the segments of these districts and then selecting the sample from this list was used. A random number between 01 and 25 was selected and then increased by 25 for a sufficient number of times to cover the total number of these segments. This, in brief, was the method used by the Costa Rican Departu ment of Census in selecting the sample for its study of the production of rice, corn, and beans. The present study used this sample as a starting point for its selection of a sample of farms producing corn. Selection of the sample employed herein A list of the medium and large-sized farms in the Department of Census sample producing either rice, corn, and/or beans was obtained 61 from the Department of Census, This list contained all of the 159 farms in Costa Rica defined as large by the Department of Census study, as well as its sample of 375 medium-sized farms. The next step was to determine in which life zones these farms were located. The list of farms contained the name of the owner or administrator, his address, the name of the administrator if this differed from the owner, and the address of the farm. This latter was used to determine in which life zone the farm was located. The Costa Rican Department of Census had available detailed maps of the entire country, and by comparing the address of the farm with these maps in most cases it was relatively easy to fix the geographic position of the farm. It was even possible in some cases to locate the name of the farm on the census segment maps. In addition, it was sometimes necessary only to determine in which census segment the farm was located, since the entire segment fell within a given life zone. The few farms impossible to locate precisely were eliminated from consi= deration, but this resulted in the elimination of fewer than ten farms that could possibly have fallen into one of the two life zones. The next step was to compare the geographic location of the farm, and the life zone map to determine in which life zones the farms were located. This was accomplished by taking the census segment maps with the locations of the farms marked on it, and comparing these with the life zone map. Again, in most cases it was easy to determine the correct life zone for each farm. Again where the life zone was questionable, the farm was eliminated from consideration. In this way the large and medium—sized farms from the Department of Census sample were put into the correct life zone, whereupon those 62 producing corn and falling into the tropical moist forest and tropical premontane wet forest life zones were extracted to form the sample for the present study. In all, there were 10 large farms and 108 medium- size farms in the trepical moist forest life zone, and 6 large farms and 54 medium-size farms in the trepical premontane wet forest life zone. The absolute number of farms is small, as is the percentage of the total estimated for these two life zones, but the percentage selected from the total estimate for each life zone is nearly the same namely, around 1 per cent. A question arose as to whether to include small farms as defined in the Department of Census study in this study. In terms of the percentage of the total farms producing corn these small farms are important. However, they represented less than half of the total output of corn, and less than half of the area planted in corn. A case could be made for excluding them. Since the purpose here is to measure differences by life zones on a limited number of variables, it could be argued that this could be done solely on the basis of the large and medium-sized farms as long as the definitions of these were consistent. The question.would-then become: are there differences between life zones for large and medium-siged farms in the particular variables? The fact that this sample might or might not be representa- tive of all farms in Costa Rica would be immaterial since the results would not be generalized to include all farms. For example, in the case of the more limited sample, the results might show that technology differs (or does not differ) between life zones for large and medium- sized farms. Since the purpose of this study is to distinguish, if possible, differences between life zones, it is not necessary to include 63 small farms, as it would be for example in an analysis of land tenure or productivity with the view of generalizing for all of the country. It is advantageous to concentrate only on the large and medium? sized farms from the practical viewpoint that the farms are more easily located and identified. The small farm sample taken by the Department of Census was made up of small geographical areas containing approxi- mately ten farms. In the first instance it would have been difficult to locate some of these areas totally within a given life zone, much more so than in the case of a single farm. In addition, it would have required locating the boundaries of the areas in the field, and then locating the various farms within the area without the advantage of having addresses for them. Another disadvantage of including these areas is that this would have meant a reduction in the number of the large and medium-sized farms that could be visited in the time avail= able, while still permitting only a very small number of small-farm visits. The obvious advantage of including the small farms (so defined) in the sample is that broader coverage of the formal producing corn is obtained. If the purpose of the study were to make generalizations concerning productivity, technology, etc., it would be necessary to include a sample of the 89 per cent of the farms that are defined as small. The practical advantage of limiting the study to the smaller sample greatly outweighs the gain from including some of these areas. A peculiarity in the method used by the Department of Census of Costa Rica partly obliterates the argument for including these areas of small farms in the sample, but at the same time it also dilutes the sample of large and medium-sized farms. This peculiarity involves the 64 classification of farms by size by the Costa Rican Department of Census. The farm sizes were determined by the number of acres planted in the various crops, with the boundaries as given earlier in this paper, and varying by province and crop. According to the method used, a farm was considered medium-sized if the farmer planted more than five manzanas of pay of the three crops. Therefore, if a given farm contained more than five manzanas of rice, but only two manzanas of corn, it was still considered a medium-sized farm; this designation being based on the five manzanas of rice. In other words, a farm did not have to have more than five manzanas of each of the three crOps to be classified as a medium-sized farm. Thus, when the present sample of medium-sized farms was taken from the Department of Census sample, some farms were included that had less than five manzanas of corn but more than five manzanas of either rice or beans, or possibly both. Thus, this sample includes farms of less than 5 manzanas of corn, and to this extent small farms are included. To be theoretically correct there are two ways of justifying this method. The first is to define the sample exactly as it has been chosen, i.e., that the sample is made up of those farms producing corn in any amount which were included in the sample taken by the method used by the Department of Census. If the sample is defined in this way, it must stand or fall on its own merits. Or secondly, one could make the assumption that the small farms so selected (i.e., all those farms in the sample with less than 5 manzanas of corn planted) are representa- tive of all the small farms in the country. The first of these seemed the better alternative and was followed. The result of this procedure was a sample of farms of all sizes. 65 When size was measured by the total extension of the farm they ran from 13 acres to over 4,012 acres, and when measured on the basis of the acreage planted to corn, they ran from 2 acres to 170 acres. A questionnaire was then devised to gather data on the pertinent economic variables related to the production of corn. The original objective was to interview fifty farmers in each life zone, but this proved impossible in the time available. The two life zones, tropical premontane wet forest and tropical moist forest, were scattered over the entire country, and the farms selected from the Department of Census sample were widely separated. Some of the farms were accessible only by horseback and reaching them would have required three or four days even if the exact location of the farm had been known. The extension agents did not know the locations of many of these farms and refused to make the trip to locate them. When it was discovered that many of the farms in the sample were impossible to get to except at a high cost in terms of the size of the sample only those farms (in the sample) that were readily accessible were visited. This permitted a sampling of 30 farms in each life zone as compared to 10 or 15 had I tried to adhered to the earlier approach. In order to cover all of the possible variations in technology existing in the production of corn in Costa Rica, the farms to be visited were selected from all of the major corn producing ggggg of that country. These areas were identified by looking at data from the Census of Agriculture of 1963 and through conversations with the extension agents and the delegates from the Costa Rican National 66 Production Council.9 In the final analysis, then, there are two distinct samples used in this study. The sample of 59 farms in the tropical premontane wet forest life zone and 114 farms in the trOpical moist forest life zone taken from the Department of Census sample is used for the analysis of productivity and land-use. This sample is a sample of farms producing corn, and includes farms of l manzana and over, For those farms with less than 5 manzanas of corn, there is the additional requirement that they be producing at least 5 manzanas of one of the other two crops, rice or beans. This means that small farms which do not have the required amounts of either of the other two crops are excluded from the sample. The randomness of the sample suffers to this extent. There are approximately 5,000 farms of less than 5 manzanas producing corn in all of Costa Rica which is less than 17 per cent of all farms producing corn in that country. In addition, about one-half of all farms producing corn in Costa Rica are in either the tropical moist forest or the tropical premontane wet forest life zone, and if the 5,000 farms mentioned above are distributed throughout the country in the same proportion, about 2,500 of these farms fall within these two life zones. 0f the remaining farms, all those with more than 5 manzanas of either rice or beans are represented in the sample. The remaining farms, of indeterminate number, were completely excluded from the possibility of being selected. In all probability then, less than 5 per cent of the farms producing corn in the two life zones 9The Consejo Nacional de Produccioh. 67 were summarily excluded. It is unlikely that the inclusion of a small sample of these farms would have materially changed the results of this study; still there is that possibility. The effect of this sampling procedure on the results of this study is to understate the influence of the particular characteristics of small farms on the variables being studied. It was discovered that there is no correlation between size of farm and productivity, thus, small farms are not significantly less nor more productive than large farms. In the comparison of productivity by life zones the exclusion of these small farms would have no substantial effect, unless for some reason these farms in one of the two life zones were extremely produc- tive or unproductive. This would mean a reversal of the lack of correlation between farm size and productivity and would be highly unlikely since there are some rather small farms included in the study for which there is no such reversal. The variable, land-use, was compared for all farms in the sample of 50 farms in the tropical premontane wet forest, and 114 farms in the tropical moist forest life zones. In addition, those political districts falling 85 per cent or more within the tropical premontane wet forest life zone were compared with their counterparts in the tropical moist forest in the search for variation in land-use. The data analyzed on the basis of this latter sample includes all farms from 1 manzana to over 5,000 manzanas, and thus the picture of land- use obtained from this analysis is fairly reliable. The sample that was used in the analysis of technology and cost of production was extracted from this larger sample, and consisted of 30 farms in each of the two life zones. All of these farms were in the fi-W 68 original list of farms taken from the Department of Census study. This is in no sense a random sample, but it does have the characteristic that all of the major corn-producing areas within these two life zones are included. In addition, several farms within the two life zones were visited which were not in the original list of farms, but which were producing corn. The methods of production for these farms corresponded well with those used on the 60 farms comprising the sample. The fact that these methods of production were representative of their respective life zones was confirmed by the extension agents as well as by the delegates of the National Production Council. For the analysis of the methods of production, the above sample, I believe, is adequate, but this is not the case for the analysis of the costs of production. The small size of the sample used for com- paring costs of production, the wide variability in the data, and the lack of a random sample make the validity of the results of this analysis highly uncertain. Cost of production is the area where the lack of a random sample would do the most damage to the validity of the study. It is possible that the small farms which were excluded from the sample might have substantially higher or lower costs of production. Since these farms do not grow much in the way of other annual crops, it is possible that they spend more of their time (labor, and capital) in the production of corn, thus raising their cost of production significantly above that of the rest of the sample. Or it is possible that these farms specialize in a permanent crop or in dairying, and grow corn solely as a means to keep the weeds down or to control insects. In this case the cost of production would be substantially lower than the rest of the sample. 69 The exact amount and direction of the bias introduced by excluding these small farms cannot be determined without additional research. Hence, the analysis for this variable should be taken as indica- tive of the costs of production rather than as an accurate representa- tion of them. The cost of production for those farms for which usable data was obtained is accurate, but the small size of the sample plus its lack of randomness precludes its being accepted as representative of all farms producing corn in the two life zones. This lack of randomness is a serious weakness of this study, but the choice was either to forego the random sample and visit a reason- able number of farms or to concentrate on randomizing the sample with the result that the number of farms actually visited would have been out severely. With the exception of the variable "cost of production", I consider the data presented to be reliable indicators of the universe. flu CHAPTER IV THE LIFE ZONE SYSTEM AND PRODUCTIVITY The relationship between the life zone system and productivity is very difficult to measure due to the problems involved in isolating the effects of the climatic parameters on this variable. Certain factors such as the texture of the soil, the existence or non- existence of nutrients in the soil, the degree of erosion, the rate of photosynthesis, and the rate of evaporation and transpiration, are directly related to the climatic parameters involved in the life zone system. The amount and seasonality of rainfall directly and strongly affects all of the above factors, and the amount and intensity of heat directly or indirectly affects all of them. Since the life zones are defined as differences in these two parameters, it follows that there should be differences in productivity between the life zones. But it is not that simple. There are many other factors which are not influenced by the climatic parameters but which do have an effect upon productivity. Of these perhaps the most important natural factor is topography. Productivity is affected by t0pography in several ways. Erosion is a much more serious problem in hilly or mountainous areas than it is in relatively flat areas. The pattern of rainfall in a given area is influenced by the existence of areas of higher elevation, which causes the rain to fall in one area rather than 70 71 another. The ability to use advanced technology which includes mechanization is affected by topography. In hilly regions it is often impossible to use tractors to pull plows, harrows, and planters, and many times the slope is too steep to permit the use even of oxen. In these areas when a crop is planted it is done by the crude method of simply dropping the seed into a hole poked into the ground with a stick. Cultivation is also made more difficult in these areas because of the inability to use mechanization. All of these factors are influenced by topography and affect productivity. This would make no difference to this study if there were no differences in topography between the two life zones. But there is a marked difference. From its very nature, the tropical premontane wet forest life zone is more hilly than the tropical moist forest life zone, because the former is an altitudinal belt within the tropical basal region, which necessarily means that it exists some distance above sea-level. Thus, wherever this life zone exists, it means that some of the land will be relatively steep. Of course, it could happen that a plateau might exist at just the right altitude so that a large and level area could be found within this life zone. If such were the case it would be much easier to compare the two life zones since the influence of gradient could be eliminated. But such a condition is highly unlikely, since not only must the altitude be correct but the amount of rainfall must also be such as to fall within the limits of that parameter for the life zone. Since mountains usually influence the pattern of rainfall, the likelihood of a combina- tion of the two requirements over a large level area is remote. In fact, there are no such areas in Costa Rica. 72 The tropical moist forest life zone on the other hand is a basal region life zone. This means that this life zone exists fairly near sea level and that any mountains that do exist are relatively low. Thus, this life zone is usually flat or perhaps rolling, but seldom is it mountainous. This is in fact the case for this life zone through- out Costa Rica. It is often flat or slightly rolling, but never mountainous. Thus, these two life zones are topographically distinct. The only evidence available on this aspect of the life zone system is contained in a study done for the United States Department of Defense.1 In this study sites were chosen from each life zone varying in size from 4 to 14 square kilometers, and mapped in detail for land use patterns, topography, and other characteristics. Two of these sites are in the tropical premontane wet forest life zone; one is in the tropical moist forest life zone and another is on the borderline between these two life zones. Table 3 shows the extensiveness of selected degrees of slope at the various sites. The table shows that for the two tropical premontane wet forest sites there are 27 and 29 per cent of the areas with a slope greater than 24 per cent. At the one site completely within the tropical moist forest life zone there is only 14 per cent of the area with a 24 per cent or more slope. The borderline site has 28 per cent of the area with slope greater than 24 per cent. If we consider a slope of 12 per cent or greater, we get 1Joseph A. Tosi, Jr., "Aerial Photographic Interpretation of Life Zones," Research on a Bio-Ecological Classification for Military Environments Found in Tropic Latitudes, Advanced Research Projects Agency, Army Research Office, OCRD, Department of the Army, (Arlington, Virginia, 1965.) 73 the same picture. The table shows that 72 and 75 per cent of the two tropical premontane wet forest sites have this slope or greater, while the one trapical moist forest site has only 41 per cent of the area in this category. It is interesting to note that while the borderline site has a relatively large per cent of its area with slape of 24 per cent or more, it has a relatively small per cent of its area with slope of 12 per cent or greater. This is explained by the fact that this site includes a relatively flat area that is quickly transformed into some rather rugged mountains. Thus, the approach to the mountains is relatively limited in extension. This gives further support to the hypothesis that the tropical premontane wet forest life zone must be quite different in topography from the tropical moist forest life zone. If this site had extended farther up the mountain side, it would have run into the tr0pical premontane wet forest life zone, and if it had extended in the other direction it would have encountered the tropical moist forest life zone. Since topography is not affected by the climatic parameters, but does affect productivity, its influence should be allowed for in the determination of the effect of the climatic parameters on productivity. Life zones are defined solely in terms of the climatic parameters; therefore, to test the hypothesis that there is a relationship between the life zones and productivity the influence of all extraneous factors should be eliminated. This would permit an accurate analysis of the influence of the life zones. Nevertheless, a case can be made from a practical point of view for taking the life zones as they are found - that is, with variations in t0pography, assuming that what has been said before about life zones 74 and tapography is true. So, if the tropical premontane wet forest life zone wherever found has more rugged terrain, then for practical pur- poses any analysis should include this characteristic as part of the life zone. In this study no adjustment was made for topography. TABLE 3 SLOPE VARIATIONS BETWEEN THE TROPICAL PREMONTANE WET FOREST AND TROPICAL MOIST FOREST LIFE ZONES SlOpe Greater than Greater than Site 24% slope 12% s10pe Site #1 14.1 per cent 40.6 per cent Tropical Moist Forest Site #2 28.7 per cent 37.2 per cent Tropical Moist Forest (borderline) Site #3 27.0 per cent 72.6 per cent Premontane Wet Forest Site #4 29.3 per cent 75.7 per cent Premontane Wet Forest Source: "Aerial Photographic Interpretation of Life Zones", Research on a Bio-Ecological Classification for Military Environments Found in TrOpic Latitudes, J. A. Tosi, Jr., Advanced Research Projects Agency, Army Research Office, Arlington, Virginia, 1965. Another factor which affects productivity but which is not related to the climatic parameters (at least to an extent great enough to be accurately measured) is the degree of technical knowledge on the part of the farmers. This, of course, is reflected in the methods used in the 75 various life zones, and includes not only the knowledge but the economic ability to implement this knowledge. Thus, if the knowledge (due to greater access to sources of information for example), or the economic ability (due to past successes in agriculture) of the farmers in one life zone surpasses that in another life zone, produc- tivity could differ by life zone from this factor alone. The ecolo- gists hold that productivity should differ by life zones, and moreover, that the tropical moist life zone should be more productive than the tr0pical premontane wet life zone. If this is true, in any given year it could reflect differences in past economic success or knowledge rather than the life zone parameters. There is no a priori reason why knowledge should be more advanced in one life zone in Costa Rica than in another, since the country is relatively small and both life zones are about equally dispersed with no obvious differences in their development of communications. But it is possible that a difference does exist between the two life zones in the ability to implement whatever knowledge does exist since the trOpical moist forest life zone would appear to be better suited to agricultural production, and thus more profitable. Nevertheless, it will be seen later in this chapter that this conclusion cannot be substantiated. There are other factors that could be mentioned which influence productivity but which are themselves not influenced by the climatic parameters. For example, the establishment of governmental or private institutions to assist the farmer, the extension of credit systems into various areas, the deve10pment of community projects for the farmer to become involved in the general health of the people, and the use of fertilizers and other applied inputs have an influence on productivity 76 but are only remotely influenced by the climatic parameters. Enough has been said to indicate the difficulty in isolating the effects of the climatic parameters on productivity, and also, to give some a priori expectation that the differences within life zones may be greater than the differences between life zones, since any of these variables could have a strong effect on production and could vary as much within the life zones as between them. Productivity is a concept often confusing and always difficult to measure. In economic theory productivity is generally thought of in two different senses. In one of these, productivity is used to refer to the output resulting from the application of one unit of a particular resource - all other resources being held constant. Thus, the marginal productivity of labor is the increase in output due to the addition of one more unit of labor with all other factors held constant. In the other sense, productivity is used to refer to the differences in output due to different combinations of resources. Thus, we say that one combination of resources is more productive than another in the production of a given product. In a theoretical sense both of these concepts are useful, but in a practical sense they present formidable problems of measurement. It is difficult to increase a given resource by one unit, and to hold all other resources constant. In teI‘ms of the various combinations of resources and their differing PITDductivity, measurement is equally difficult because of the infinite “lumber of possible combinations that could be studied. In addition, these problems are overshadowed by the difficulty of measuring the 11'11)th of certain factors — especially capital. The concept of productivity used in this study is that of a 77 combination of resources and is measured in terms of the output per manzana of land. The output per manzana for the farms in the one life zone is compared with the output per manzana for the farms in the other life zone. Productivity on the farms in the two life zones is measured giygp the combination of resources being used on the farms. No attempt is made to hold the methods of production constant, nor is there any attempt made to ascertain whether the most efficient combination of resources is being used. The assumption is made that the farmers have adjusted to that combination of resources that best fits their life zone, and that therefore, the technology used reflects the effects of the climatic parameters. This assumption is crucial to the following analysis. Thus, productivity might vary by life zone because the farmers must adjust their technology because of the differences in the climatic parameters. In other words, the assumption is that the differences in the climatic parameters have caused the farmers to use different technology, and that they always use the technology best suited to the climate within their life zone. Given this assumption one can be sure that any effect of the climatic parameters on producti- vity through technology will be registered. Productivity on all farms in those political subdivisions falling at least 85 per cent within the two life zones is compared on the basis of the output per manzana of corn. The same important assumption is made concerning the influence of the climatic parameters on the technology used in the two life zones. Before proceeding to a more detailed description of the method used to compare productivity by life zone and the results of that comparison, there is another ticklish matter to dispose of. It has often been assumed by economists that there is a positive relationship between 78 productivity and farm size and that there is an optimum farm size in terms of productivity. Very small farms are thought to be inefficient, with low productivity and high per unit costs, as are very large farms. The application of economic theory to agriculture leads one to the conclusion that economics of scale are possible up to a certain point, and that to have high productivity agriculture, farms must approximate the optimum size. The communist countries especially have been experi- menting with large scale agriculture, while in the United States farms are gradually becoming larger. The family-sized farm, which is majority of farms in the underdevelOped countries today, is no longer considered the ultimate in efficiency. The large corporate farm is now a preva- lent and efficient type of farming in this country. In many of the developing countries the trend is toward smaller and smaller farms, but for political rather than economic reasons. The usual procedure is to expropriate the large farms and to divide them into many small (usually too small) farms. In some cases this has led to a reduction in output, especially where the large farmer had originally been using the land relatively efficiently. Schultz suggests that there is no necessary relationship between farm size and efficiency. He states that: "In searching for the economic attributes of traditional agriculture, it became clear that small farms, or for that matter large farms, are not essential attributes... The size of farms may change as a consequence of the transformation (of traditional agriculture) they may become either larger or smaller - but changes in size are not the source of the economic ,_ __..._.- *— 79 growth to be had from this modernization process."2 For this study the matter of farm size is important, since if farm size does effect productivity, and farms are on the average larger in one life zone than in the other, productivity could differ for this reason alone. In order to provide for this possibility, a comparison is made between the two life zones to determine if there is a difference in the average size of farms, and in addition, tests are made to see if there is a correlation between farm size and productivity. The sample taken from the Costa Rican Department of Census is used to determine if there is a difference in the average size of farm. Size of farm for this purpose is defined in two ways: in terms of the total area of the farm, and in terms of the area of the farm planted in corn. Under the first definition, in the trOpical moist forest life zone the farms range from 8 manzanas to 2,360 Inanzanas, whereas in the trOpical premontane wet forest life zone, they range from 11 manzanas to 896 manzanas. Under the second definition, the farms in the trOpical moist forest life zone range from 1 manzana to 100 manzanas, and in the trOpical premontane wet forest life zone they range from 1 manzana to 32 manzanas. In terms of range, then, the trapical moist forest farms exceed the tropical premontane wet forest farms under both definitions. This does not mean that there is a difference in the average size farm, however, since a large number of smaller-sized farms could offset the few 2Theodore W. Schultz, Transformipg Traditional Agriculture, (Yale University Press, New Haven and London, 1964), p. 110. 80 larger farms in the trOpical moist forest life zone, or the opposite of this could mean a higher average in the trOpical premontane wet forest life zone. The method used to determine if these two samples of farm size are significantly different from each other is as follows: first, the null hypothesis is made that there is no difference between the two life zones in the average size of farm. Then to test this hypo- thesis the standard error of the difference of the means of the two samples is computed and compared with the actual difference in the two means. This value is distributed according to the t-distribution, however, in the computations made here the samples are large enough that the distribution closely approximates the normal curve. The .05 level of probability is used as the level of significance. The results of these computations,3 with farm size defined as ‘the total area of the farm, are interesting. In the trOpical moist :Eorest the standard error is 31.17 or 17.96 per cent of the mean, Vihile in the trOpical premontane wet forest it is 21.68 or 16.70 Iwer cent of the mean. This is somewhat of a measure of the accuracy ()1? the sample in describing the true mean of the pOpulation. The éstzandard error of the difference, which is the standard deviation of ii distribution of differences for a large number of pairs of random Samples of means independently drawn from the same population, is 37.97. The test statistic (t-value) is 1.15 which, since the number fo’ degrees of freedom is relatively large (171), proves not to be W1thin the area of rejection of the null hypothesis. Thus, the null 3The statistical analysis is presented in Appendix B. 81 hypothesis must be accepted, as there is probably no appreciable difference in the average size of farm for the two life zones. When farm size is defined as the area planted in corn, similar results are obtained. In this case the standard error of the mean for the trepical moist forest is 1.04 which is 13.16 per cent of the mean, and in the tropical premontane wet forest it is .768 which is 9.8 per cent of the mean. The standard error of the difference is 1.29, and the test statistic is .0775. Given the large number of degrees of freedom (171), we again accept the null hypothesis. That is, the evidence again is that there is no significant difference in the average size of farm between the two life zones. Under either definition of the size of farm the conclusion is the same. Nevertheless, the large standard errors in relation to the mean indicate some lack of reliability on the part of the four means. This suggests considerable variability in the distribution of sample means, and it is possible that the means deduced from our samples differ significantly from the true papulation mean. Therefore, the means of the samples of the sizes of farms, using either definition, may not give a valid picture of the average size of farm in the two life zones. To avoid the possibility of not detecting an influence of the size of farm on productivity because of the unreliability of the means of the farm sizes as described above, correlations are calcu- lated between the size of farm and productivity in each of the life zones. If such a correlation does exist, there will remain some doubt concerning the differences in productivity by life zone since the average size of farm could differ by life zone. The large standard 82 errors prevent us from making definitive statements on this latter point. If such a correlation does not exist, then any difference in farm size between the life zones will be inconsequential since the difference could have no effect on productivity. Correlations are calculated for both life zones between size of farm and output per manzana on the farms, with the size of farm defined in two ways, total area and area planted in corn, rice, and/or beans, whichever the case happened to be. Thus, in each life zone six correlations are run - one for each of the three cr0ps under each of the two definitions of farm size. This gives a total of twelve correlations in all. Table 4 shows the results of these computations. As can be seen, all of the correlations are very low; in fact, in five of the six correlations between area planted and output per manzana, the correlations are negative. The correlation between total area and output per manzana of corn in the trepical moist forest is the highest, but it still is far from being conclusive. To test the significance of these coefficients, the null hypo= thesis is made; that is, it is hypothesized that the coefficients are not significantly different from zero. The .05 level of probability is accepted as the level of significance, or in other words, if the probability is less than 5 in 100 that the coefficient could have happened by chance from a situation where there is actually no correla= tion between the two variables, the hypothesis is rejected. On the contrary, if the probability of obtaining this coefficient by chance out of an actual situation of no correlation between the variables is greater than 5 out of 100, the hypothesis is accepted. The tudistribu- tion is used since in several of the correlations the number of degrees 83 of freedom was fairly low. The values for the t-statistics are given in Table 5. Of the twelve coefficients of correlation only two, one in each life zone, are significantly different from zero as measured by this test. In the tropical premontane wet forest life zone the correlation between total area and output per manzana of rice is significantly different from zero; whereas the same is true of the correlation between total area and output per manzana of corn in the tropical moist forest life zone. Most of the other t-values are very low, and thus there is a very large probability of these coefficients of correlation occurring by chance when no actual correlation exists. The one correlation in the premontane wet forest life zone is significant at the .05 level of probability, and that in the trOpical moist forest life zone at the .02 level. The coefficient of correlation does not indicate cause and effect, but merely indicates that a relationship exists between the two variables being compared. Also, it is difficult to label the degree of relationship as being "low" or "high regardless of the value of the coefficient since a "low" value for the coefficient may indicate an important relationship between the variables. The null hypothesis test described above is one way, however arbitrary, of establishing the value of these coefficients. The conclusions to be drawn from the investigations of the relationship between the size of farm and productivity are as follows: First, there seems to be no difference in the average size of farm between the two life zones using either of the two definitions of size. Secondly, there is little evidence of a relationship between 84 5‘4”» w>¢.uo mmuv.uo. .vménu Educ NOAH» amonohpog oompdoeonnm Hdoaonfi e... .H u L wmm. u . So. u a cos. n a mom. u s Sim n s $0.8m $62 manage HZON Hand Wm WEE/HHUDQOMQ Q24 MNHm zmaflh ZHHBHHm ZOHH<1HHMHMOU HEB MOh wHZHHUHhhHOU ZOEUfiHHMMOO HEB MO HUZE are 20mm QHHOMAMm 49.3 erase/~23 NH HAMANH a: as: $3 .1. n H H N H mHmHnoHHZ an m... * om a sesame N v N s a. mwfiensm 2 2 ow. S mm 9.3 cm 0H. Q. Hi. mo Honda 2 M 3* Such 3w Each mew Such gm Such 2* Each uchOnHEOU mfiuow UsuocHom 05.3 new umoD GoHuodvoum H33 .Ho cwcucounan mmzoqoo zH Chou ozHHMMeEE UZHQDAUNMV HmOO ZOHH. UDQOMHnH A