William M. Davis


Converted in Spring 1996 by Lance Christian.

The Need of Comparing Opinions

One of the objects to which geographers may well direct their attention is the nature of the whole subject of geography, under whose broad shelter individual studies are carried on. If we work only in view of our chosen field within the whole area of geography, we lose something of breadth, although we may gain much in depth. Our work will become more serviceable to others if it is presented in such a manner that its relations to the whole subject are made clear. Not only so; we may often benefit ourselves by systematically setting forth the place of our individual studies in geography as a whole, for we may be thereby led to discover that the systematic sequence of parts is interrupted here and there by gaps, which can be filled by well- directed effort. 

 But it is manifest that, if we should attempt to make exposition of our ideas concerning the relation of our own studies to the whole subject, we must have previously gained a tolerably definite idea of the nature of the whole. Very likely we each have some such conception. If so, it would be profitable to institute a collection and a comparison of opinions, though it can hardly be doubted that they would show much diversity. 

Examples of Geographical Statements

 As a contribution towards such a collection of opinions, I propose to consider the nature or content of geography in this essay; but instead of presenting an abstract definition of the subject, the problem will be regarded from the other side. There will first be presented a number of sample statements that are presumed to be geographical because they are taken from geographical works; these statements will then be analyzed in order to discover their essential nature; finally, an attempt will be made to discover the nature of the whole subject of geography on the basis of the statements thus analyzed; that is, to discover the content of geography on the basis of an inductive study. The statements to be presented are taken from a variety of sources text-books, treatises, and journals; they are often condensed from the original. 

 Here, for example, is a first set of items or statements: There are three native states in the Himalaya which are independent of British rule. The rivers of southern Gascony flow in radiating courses. There are more than three hundred asteroids. The average man has about eighty-seven cubic inches of brain. The mean depth of the ocean is about two miles. A new map of France on a scale of 1:50,000 is in course of publication. Protozoa are the simplest forms of animals, consisting of a single cell. The magnetic poles do not coincide with the poles of rotation. Celtic is spoken in Wales, the Highlands of Scotland, and western Ireland. Cotton is produced in the southern United States. The fall of rivers is usually greatest in their upper courses. The Grand Duchy of Prussia became a kingdom in 1701. Cryptogams are flowerless plants. The Hindu Kush, the highest mountains of Persia, trend to the northeast. Australia has many marsupials, of which the kangaroo is the largest. Amundsen has successfully made the Northwest Passage. 

 If all these statements are truly of geographical quality, it follows that all similar statements would be of geographical quality also. We should then have, in a complete geography, accounts of all states, in the Himalaya or elsewhere, dependent or independent, of British or of other rule; descriptions of the course of rivers in all parts of the world; additional facts about other bodies in the heavens besides the asteroids; accounts of man's other organs than his brain; details as well as averages of ocean depths; descriptions of multicellular as well as of unicellular animals, and of flowering as well as of flowerless plants; reports on old as well as on new maps of France and of all other countries, whatever their scale; items concerning terrestrial magnetism as well as concerning the position of the magnetic poles; reports of all sorts of voyages; accounts of all changes of duchies to kingdoms and of kingdoms to duchies, and of other political changes as well; mention of the animals of other lands than Australia, and of the products of other countries than our southern states; announcements of voyages in all parts of the world in all centuries. 

The Complexity of Geography

 The first impression about so varied a collection of statements would be that the subject under which they are all properly included must be a sort of omniumgatherum, without any well-defined unity of substance. Such, indeed, geography has been thought to be by some critics. A closer examination of our materials must be made in order to see if we really have to do with a heterogeneous collection of incoherent facts, without continuity of thought or unity of discipline. It may be that some of the items above cited do not justly belong under geography, and that an enchainment not apparent at first may be found to bring the remaining items and their proper fellows into some systematic association. 

 On reviewing the statements given above it appears that some are concerned with matters remote from the earth and commonly associated with astronomy; that some are concerned with matters of a personal or technical nature; that a third group of statements contains accounts of inorganic features of the earth, and a fourth group contains accounts of organic inhabitants of the earth. Let us examine these several classes to see how far they may be justly considered geographical. 

 Excepting as an astronomical statement is found to be related to terrestrial matters in such a way as to throw light on them, it should not be considered directly or indirectly geographical, in spite of the mention of such a matter in a geographical textbook; for by common consent geography has to do primarily with the earth. Hence the mere statement of the number of asteroids should be excluded from our consideration. It may, however, be advisable to retain in elementary text-books certain statements concerning planets in so far as they serve to emphasize the globular form of the earth by showing that there are other bodies like it, and such mention of the sun and moon as is needed for the understanding of insolation in relation to climate and of gravitation in relation to tides; also such account of the stars as shall serve in the proof of the rotation of the earth and in the determination of the attitude of its axis. Nevertheless, all these items seem to be only indirectly associated with our subject; they do not aid us to discover the real nature of geography. 

 Technical and personal items may naturally be associated with geography if the things and events with which they are concerned are of a geographical quality, as maps and explorations certainly are. The announcement of the publication of a new map or of the completion of a remarkable voyage is therefore a piece of appropriate geographical news. Accounts of old maps, of old voyages, and of old ideas are very illuminating in showing how the geography of today has been developed, but they belong more with history than with geography when presented in order of time. 

Non-Geographical Statements

Many of the statements under the two remaining classes - which concern inorganic and organic matters on the earth - are properly geographical if we judge by the repeated occurrence of such statements in many geographical works. Yet we cannot believe that all statements about inorganic and organic matters on the earth belong under geography. It is easy to instance a number of such statements from geographical sources which can hardly be regarded as having any clear association with geography. Consider, for example, the following set of items, all taken from geographical text-books: 

The gastrula stage occurs just before maturity in the sponge. All substances transmit electrical energy. Crystallization is explained by supposing that the cohesive force of molecules is not exerted equally on all sides. After 1810, several independent republics were formed in South America. The Rosaceæ have regular flowers.
However interesting and useful these items of information may be, they can hardly be regarded as geographical, even under the very broadest interpretation of our subject. They have probably been added to the text-books from which they are quoted in the belief that they are worth knowing, and that the pupils in our schools will have little chance of learning them if they are not brought in under the popular subject of geography. I question the advisability of such a method of promoting popular education; but that is another story. Our problem is to distinguish these and many similar statements from a large variety of other statements concerning the inorganic and organic terrestrial matters, so as to know which are properly of geographical quality and which are not. 

Statements involving Relations

 The best solution that I have been able to find for this problem comes from a consideration of such statements as are included in the following third set of extracts, which, like the other two, have been taken from geographical sources: 

The people of the Dalmatian coast are largely engaged as fishermen on the Adriatic; here the Austrian navy obtains its best seamen. A number of gulfs and mountain ranges naturally divide the people of Europe into groups or nations. Some parts of Bengal are so favored with rivers that almost every cottage has a navigable stream at its door, and the Bengalese farmer keeps his boat, just as the English farmer keeps his gig. Northern Uganda is drier, rain often fails, grass is short, and dry-country animals, such as zebras and ostriches, abound. The level land near lake Titicaca is of small extent, and is occupied by alkaline marshes subject to overflow; for this reason the Indians cultivate the slopes in preference, converting them into narrow, terraced garden beds. The periodic or monsoon winds that blow alternately now in one direction and now in the opposite along certain coasts, as between India and East Africa, have long been favorable to navigation and trade. In northwest France the farm- houses on the chalk uplands are gathered into compact villages around one or more deep wells, because the ground water there is deep below the surface and difficult to obtain; but in the valleys the houses are scattered, because each family can easily procure its own water supply with little trouble.
Geography a Study of Relations

 These statements differ from some of those of the two earlier groups of citations in that they each contain at least two kinds of elements, one of which stands in a more or less distinctly causal relation to the other. The statements vary in that some of the relations suggested are relatively simple, while others are more complex, and in that some statements are rather empirical, while others are more fully explanatory; but they are all alike in that they involve a relation of cause and effect, usually between some element of inorganic control and some element of organic response. As such, these statements are examples of the most generally treated material that I can find; and hence I am disposed to say that any statement is of geographical quality if it contains a reasonable relation between some inorganic element of the earth on which we live, acting as a control, and some element of the existence or growth or behavior or distribution of the earth's organic inhabitants, serving as a response; more briefly, some relation between an element of inorganic control and one of organic response. The geographical quality of such a relation is all the more marked if the statement is presented in an explanatory form. There is, indeed, in this idea of a causal or explanatory relationship the most definite, if not the only, unifying principle that I can find in geography. All the more reason, then, that the principle should be recognized and acted upon by those who have the fuller development of geographical science at heart. 

Geography as the Study of Location or of Distribution

 There still exists in some quarters a tendency to limit geography by definition to the study of location, leaving the study of all the things that are located to other subjects. There is so little support for this narrow view of the subject to be found in modern geographical books that it need not be further considered. Another and more widely accepted definition treats geography as the science of distribution. This is particularly a British view of the subject, and at least one British geographer urges that the distributed things should not be regarded as belonging to geography at all; but his writings are broader than his definition. While location and distribution must always be important elements of geography, all geographical books give much attention to the nature of the things that are distributed, and all recent books give much weight to the relations in which the distributed things occur; hence relationship seems to me the primary principle of the two, and distribution takes a secondary rank. The thing must be known before its distribution can be serviceably studied. The division of the peoples of Europe into groups or nations in consequence of the division of Europe by gulfs and mountain ranges is a geographical relation, in which the unlikeness of the things distributed takes precedence of their distribution. 

 Indeed, if geography were only the science of distribution, - that is, the regional aspect of other subjects, - it would be hardly worth while to maintain the study of geography apart from that of the subject whose regional aspect it considers. Moreover, if geography is defined simply as the science of distribution, then the distribution of anything is a fit subject for geographical study. Under such a definition the distribution of hypersthene andesite, or of books of poetry, would be part of our responsibility; but there is nowhere any indication that geographers feel responsible for the distribution of such things. Again, if distribution is given first rank, the regional or spacial occurrence of all sorts of things is thereby given so great an importance that an insufficient place is left for systematic considerations; yet most books first present things of a kind together that is, they present the subject systematically before they take up the distributional or regional treatment of different kinds of things. Moreover, distributional treatment is too apt to take up one kind of thing at a time, and follow it wherever it is found, thus failing to give account of the natural occurrence of many kinds of things in their actual association which regional geography demands. 

 As a matter of fact, nearly every example that is presented in this essay as an example of a geographical relation might, if desired, be presented as an example of distribution, and nearly all statements of distribution may be turned around so that they shall enter into or constitute relations; hence the total content of geography would be much the same under either principle. The relations that could not be presented under the head of distribution are those which are the same everywhere; but it seems to me rather arbitrary to include relations that vary from place to place, and to exclude relations that are world-wide in their uniform occurrence. For example, the composition of the atmosphere or of the ocean, always accepted as an appropriate matter for mention in elementary texts, deserves no place in geography, treated as the science of distribution, until the minute variations of composition from place to place are considered. 

 In any case, location and distribution are fundamental elements of geography, and maps of the lands and charts of the oceans are essential in its every chapter. In view of the importance of these elements, some are disposed to attach an undue value to surveying as a part of geography. But surveying properly belongs in geography about where writing belongs in literature, and if it is given higher rank the student may, by misfortune, turn his chief attention to the art of mapping instead of to a study of the things that are mapped. 

Regional Geography

 It has been well maintained by certain German geographers that geography is concerned with the "material filling of space"; that is, geography has to deal with regions or districts, as occupied by all the kinds of things that there occur together. This, however, is only the regional aspect of geography, and before it can be treated, the geographer must have a good acquaintance with the different kinds of things that occur in various regions, that is, with systematic geography. It is with this second aspect of the subject that we are here chiefly concerned. 

Physiography and Ontography

 If the principle of explanatory relationship be adopted as a general guide to the content of geography, it follows that neither the inorganic nor the organic elements which enter into geographical relations are by themselves of a completely geographical quality; they gain that quality only when two or more of them are coupled in a relation of cause and effect, at least one element in the chain of causation being inorganic and one organic. When they are considered separately, but as if in preparation for an understanding of the causal relations in which they will later be presented in the study of geography proper, they may be considered as sub-geographical; and then the inorganic elements may be called physico-geographic, or physiographic, and the organic elements may be called ontographic. Common usage recognizes the first of these divisions, but not the second, as we shall see later. It is well to emphasize the clause "as if in preparation for an understanding of the causal relations in which they will later be presented in the study of geography proper for many items which, under this proviso, are of physiographic or of ontographic quality, may under another proviso belong in other departments of knowledge. 

Physiographic Matters

 This may be made plainer by citing particular cases in which a number of different topics, familiar from their occurrence in books on geography, will be seen to fall under other subjects when they enter into non-geographical relations. For example, the size and rotation of the earth, and the general movements of the tides, are undoubtedly physiographic elements, yet both may be treated appropriately under astronomy when they are considered in their relations to other planets. Sea water, when regarded as a liquid compound which holds various salts in solution, may be studied properly by the chemist with regard to its composition; nevertheless, sea water as the medium in which organic forms live is an indispensable subject for physiographic inquiry. The horizontal strata of plateaus are a fit topic for geological investigation, as regards their origin and history, but they are no less a subject for physiographic investigation as affecting present form. It therefore seems impossible to determine, merely by a consideration of the thing studied, whether it belongs to physiography or not. 

 A given object may belong under several different sciences, and may be treated in text-books on different subjects; it is the relation into which the object enters that determines its place. 

 There are, however, certain inorganic topics, commonly found in geographical books, which seldom, if ever, enter into relations with organic topics, and which would, therefore, under the strict application of the principle of relationships, be ruled out of physiography, and thus out of geography also. Such are cirrus clouds and halos, the crevasses and blue veins of glaciers, and the polar flattening of the earth. Nevertheless, most or all of these topics will probably hold their places in books on physiography, because they serve to complete the picture of the whole earth as the home of life. In any case, little is gained by a very strict or over-logical application of a useful principle of classification in a problem such as we are considering. 

Ontographic Matters

 The determination of the science under which a thing belongs by means of the relations into which the thing enters is an even more important guide in ontography than in physiography. For example, one may read in a certain textbook of geography that all forms of life consume food. In so far as the assimilation of food, and the organs by which it is accomplished, are concerned, the consumption of food belongs under physiology; but the consumption of food is an ontographic matter in so far as it brings an organism into contact with the rest of the world and thus causes it to enter into geographical relations. Commercial geography is largely concerned with relations that grow out of this element of ontography. Water is essential in organic processes of many kinds. This is, again, a physiological matter if it is examined with reference to the processes of circulation, but it is ontographic when it is found to lead to a relation with the sources of water supply: villages gathered around deep wells on the chalk uplands of northwestern France are examples of the geographical relation thus brought about. 

 Plants and animals tend to diffuse themselves, or to be diffused, over the earth. This is a fundamental fact, usually associated with the study of biology; but the limits of diffusion, and in many cases the means of diffusion, are determined by physiographic controls; hence the tendency to diffusion is an ontographic matter. 

 The need experienced by all forms of life to secure food, already instanced, leads to many other relations than those of commercial geography. The need of food is satisfied without going in search of it if the food is contained in a moving medium that surrounds the organism; hence those organisms that live chiefly upon food contained in one or the other of the two great mobile envelopes of the earth the atmosphere or the hydrosphere are often rooted or fixed: the air or water currents carry the needed food to the waiting plant or animal, and this is surely too important a geographical relation to be omitted from a broad consideration of our subject. Other organisms take advantage of the currents of air or water to be passively carried about by them, taking their food when they happen to come upon it. If they are land dwellers, they are so small that they can easily be wafted about by the winds; if they are dwellers in water, they may gain greater size by assuming about the same specific gravity as that denser medium, so as to float easily in its perpetual currents. Still a third class of organisms move of their own volition, and in connection with these there are all manner of geographical relations. Some of them involve the development of wings, whereby motion can be effected in the comparatively unsustaining air, as in the case of many insects and birds, of a few mammals, and formerly of some reptiles; some involve the development of fins to produce motion in the sustaining water. These examples are as good illustrations of organic responses to inorganic controls as are the canoes and the steamships of uncivilized and of civilized man. 

 It follows from the preceding paragraph that the more closely our standard geographical material is examined, the more clearly it appears that its ontographic as well as its physiographic elements may fall into other sciences when they are treated in other relations, and that they become most distinctly geographical when they enter into the causal relations of the kind set forth above. The rise and fall of the tides is a physiographic matter when it is seen to determine the distribution of certain forms of life, such as barnacles, or to influence the availability of harbors for the entrance of shipping; the occurrence of coal is a physiographic matter when it is found to influence the industry of a district and the commerce between nations; the small size of spores, pollen, and germs is an ontographic matter when it is seen to be related to their transportation by the thin air; the sensitiveness of organisms to temperature changes is an ontographic matter when it is shown to affect the distribution of plants and animals over the earth. Yet all these matters may be treated with entire justice under other sciences, than geography. It is, therefore, to my reading, of capital importance, in determining whether a statement is of geographical or of sub-geographical nature, to know how far it constitutes or enters into causal relations of the kinds that have here been considered. 

Is Ontography a Part of Geography

 It is perfectly true that many of the illustrations just given are not commonly regarded as belonging under geography, but it seems to me that their exclusion is illogical and arbitrary. They are practically all to be found in certain standard geographical works, and many more might be taken from such books as Ratzel's "Anthropogeographie" and Beddard's "Zoögeography." The general principle by which one should be guided in determining the relevancy of such matters is as follows: If a certain relation between an inorganic control and a responding organism is a geographical relation, then all similar relations are also geographical. For example, a well-known text-book makes the statement that water plants are supported by the relatively dense medium in which they grow, and hence do not need strong, woody stalks such as many land plants have. This is an excellent example of a geographical fact: it involves a relation between an organic growth and an inorganic medium. But the flight of birds, the small size of germs, the essential agreement between the specific gravity of fish and of water, the universal habit of breathing oxygen, - all involve similar relations. The first example seems to me undeniably geographical; the others are no less so. To exclude the latter from geography while including the former would be to set very arbitrary limits to our subject. 

 It may, perhaps, be objected that flight and breathing are processes of too ancient origin to be considered as geographical, but inasmuch as they have been maintained to the present time by inheritance through persistent conditions of environment, they have the same right to a place under geography as is enjoyed by such examples as the prevalence of the fishing industry on the Dalmatian coast, or the custom of the French farmers on the chalk uplands of living in compact villages; for these habits, also, are not independently originated by each man who follows them, but are continued by inheritance through persistent conditions of environment. 

 There are certain matters frequently encountered in text-books of physical geography which belong better, it seems to me, under the head of ontographic relations. Such are the distribution of plants and animals, and the races of man. The association of such topics with physical geography is probably the result of conceiving all the rest of the subject as contained under political geography. The contrast of physical with political conditions may serve well enough in elementary books, where the distribution and behavior of man are the chief subjects in political geography, and where plants and animals are therefore thrown in with physical geography; but in the more advanced and general treatment of the subject such an arrangement is not satisfactory. It does, however, seem legitimate to introduce as often as desired ontographic responses in a physiographic text, in order to show at once the kind of response that certain controls call forth, and thus to impress the fact that the physiographic items are really related to ontographic items; a similar introduction of physiographic items is appropriate in chapters on ontography. This practice is followed by certain writers. 

 There are three definite positions and many indefinite positions that might be taken with respect to the attention that should be given by geographers to organic considerations. The narrowest position limits geography almost entirely to the inorganic features of the earth, that is, to physical geography, or physiography. This is the view of geography held by some historians, who take unto themselves practically all the human element that is so commonly encountered in political geography. An intermediate position would include physiography and the more manifest relations into which it enters with various forms of life, and particularly with man, but would not accept responsibility as to the less manifest responses of various living things. This seems to be the position taken by many geographers, more or less consciously. The third position would treat ontography as thoroughly as physiography, and would search for all the geographical relations of physiographic controls and ontographic responses. This is certainly the broadest of the three positions, being, as many would feel, too broad, and involving too much overlapping upon other subjects. For my own part, there seems to be so manifest a necessity of gaining a responsible knowledge of ontography, at least of elementary ontography, before geography proper can be successfully treated, that ontography should come to be regarded as a part of it. The analytical and inductive methods of this paper, therefore, lead me to adopt the third position; and I believe that this position is essentially consistent with the opinion of writers who, like Ratzel and Reclus, have cultivated the most advanced or matured stage of geographical science. 

The Importance of Explanatory Relations

 Although various facts which may make parts of relations between inorganic controls and organic responses, or which are met with in preparation for an understanding of such relations, thereby gain a most characteristic geographical flavor, a very brief review of geographical books will suffice to show that many statements there included do not explicitly possess this flavor. In the first place, in the older books the idea of relationship had no distinct recognition. That was the time of memorized names of capes, of empty boundaries, of unexplained lists of products. Almost anything about the earth or its inhabitants was then accepted if it reached a satisfactory degree of what is called "importance." In the newer books the principle of explanatory relationship is very generally acted upon; there are, nevertheless, many relations stated so empirically that the pupil may fail to gain from them a full appreciation of the best essence of geography. Consider, for example, the following statement from a good textbook: "Some form of Celtic is spoken in Wales, the Highlands of Scotland, and western Ireland." This is nothing more than an empirical statement of a fact of distribution. Many a student might memorize such an item without gaining any clear insight into its geographical meaning. The real point is that the early inhabitants of Great Britain, being attacked by invaders from the continent on the east, survived only where pursuit was difficult, as in the rough ground of Wales, in the rough and distant ground of the Highlands, or in the distant and isolated ground of Ireland ; and hence only in those localities is the early language still preserved. The fact that Celtic, as Celtic, is spoken in certain parts of Great Britain belongs to philology as well as to any other subject; the fact that Celtic is the language that was once spoken generally through Great Britain might come under history as well as under any other subject; but the fact that Celtic is still spoken in rough, distant, or isolated parts of Great Britain, because of their roughness, distance, or isolation, is a local example of an important class of relations between controls and responses, and as such it belongs distinctly under rational geography. 

The Expansion of Geography

 Geography has still much progress to make. There is not only much to be done in the way of exploration of little-known lands and seas, but many of the more civilized countries still merit closer study than they have yet received if we may judge by the notable incompleteness of the best handbooks and treatises. One method of carrying the subject forward consists in outdoor observation, and this method cannot be too highly recommended to those who wish to contribute to the fuller development of our subject. When we realize that we have no modern and maturely developed account of the geography of such a state as Virginia, or Ohio, or Colorado, or California, it becomes evident that abundant opportunity for exploration lies near home. But there is a second method by which geography may be promoted; that is, by thinking about what we see, and thus expanding every example of a geographical relation that we find to its farthest legitimate extension. Take, for example, the common case of a road which runs through a notch in a high ridge; such a detail of location is a response of the ontographic element, movement, to the physiographic control, gravity; hence all such responses to gravity should be searched out so that they may be systematically treated as to kind, and regionally treated as to distribution. Such matters are surely as properly of a geographical quality as are various responses to sunshine which are always found among the standard matters of our books. How many more responses to the universal and persistent force of gravity there may be, we do not yet know. Again, take such an example as that afforded by the habit of working by day and summer and resting by night and winter. These are responses to the controls of light and heat in contrast to darkness and cold, which result from the illumination and warmth of the rotating and revolving earth by an external sun. This suggests that all other responses to light and to darkness should be sought out and studied: for it leaves the limitations of our subject very vague if we make a beginning with such relations and do not carry them to an end; it makes the limitations very arbitrary if we make a beginning of such relations and stop with the first or the second or the third example, instead of pressing on to the very last. 

 It is an application of this same principle that makes it impossible to limit the organic side of geography to man; for the habits which man has developed in his search for food, clothing, shelter, and so on, are in very many ways closely analogous to those developed by animals in a similar search. In the same way the sensitiveness of man and animals to climatic conditions by which limits of distribution are so largely determined is paralleled by the sensitiveness of plants to similar conditions. Life is a unit; if one form of life comes under the study of geography because it responds to physiographic controls, then all forms of life come under geography. 

 The expansion of geography through time is likewise inevitable unless it is most arbitrarily limited to the "present day." Precisely the same principles have been embodied in the relations between physiographic and ontographic elements in the past as are embodied in them now. It is therefore most illogical to think of geography as a science that deals only with today. There has always been a geography, all through the geological ages; geology is the integration of all its momentary or geographical differentials. We may never know very much about the successive geographies of the past, but all the fragments that have thus far been learned assure us that it was of essentially the same order as the geography of the present. The recognition of this principle is of the greatest importance to geography, and to geology as well. 

The Subdivisions of Geography

 It is but natural that the different phases of human geography should have been more fully developed than the other branches of our science. Political geography, frequently overrunning civics, economics, and history, has long been familiar as an elementary subject; but with us it has seldom been carried into the higher reaches of education. Commercial geography is rapidly gaining an important place in our schools, and is meeting the danger of becoming almost as empirical as the old-time lists of products of the several states. Biogeography has several parts. Anthropogeography, as expanded by Ratzel, seems destined to become an important subject in the universities, because of the greater insight that one gains through it into history. Zoögeography and phytogeography are in my opinion, as a rule, too strictly limited to facts of distribution alone; these divisions of the subject should be expanded so as to include for animals and plants a consideration of what would correspond to the political and commercial geography of man. Paleogeography is occasionally treated, but it must always be a fragmentary subject, because it is based on fragmentary records; it will, however, be better treated by geologists in proportion as they have had geographical training. 

 The growth of explanatory treatment, which makes so characteristic a difference between the content of geography in books of a hundred years ago and of today, is chiefly due to the different amounts of general knowledge in stock then and now, and consequently to the different philosophies then and now prevailing. The subject has thus gained greatly in strength, in disciplinary value, and in living interest. At the same time geography has come to cultivate more and more - some would say, to trespass more and more upon - the fields that are also cultivated by other subjects. If the trade winds are not simply described as to region of occurrence, direction of blowing, weather conditions therein prevailing, and so on, but are also explained as parts of an extensive convectional circulation between equator and poles, modified by the deflective effect of the earth's rotation, all this explanatory matter has a strong flavor of physics. If the occurrence of plants of a certain kind in a given region is not merely asserted, but is shown to be the result of climatic conditions to which the plant responds, owing to its sensitiveness to temperature and moisture, this closely resembles certain chapters of botany, and the same may be said regarding the relation of animal distribution to zoology. If the boundary of a state, the location of a city, or the industry of a district is rationally explained instead of empirically stated, the explanation is of a kind likely to be found in books on history and economics. Shall we then, in view of this, relinquish explanatory treatment to other subjects and content ourselves with empirical statements? Shall we adopt the limitation to the location of things, as above suggested, and thus avoid duplication with other subjects? No. Duplication is unavoidable; and, moreover, it is reciprocal. The historian, as well as the botanist and the zoologist, must borrow from the geographer all manner of facts regarding location, extent, distribution, climate, form, movements, products, populations, and so on; the geologist can hardly make a step into the realm of the past without having made preparatory study of the present. Overlapping and duplication are unavoidable. We must each of us try as far as possible to concentrate upon his own subject; but we must at the same time borrow and quote with the utmost freedom from any other subject that will give us aid in the consideration of our own. 

Examples of Helpful Duplication

 One of the most interesting fields that I have run across in geographical research is that open common where geography and philology overlap. It has often been remarked that the Arabs of the desert have many terms for sand dunes, evidently because they are familiar with the many forms that sand dunes there assume; their language has developed in a peculiar direction as an ontographic response to their peculiar physiographic environment. In the same way the people of the Alps have various terms by which to name mountain summits of different shapes; this is another ontographic response to a peculiar physiographic environment. But there are other less manifest examples of a somewhat different kind. We treasure as a fine geographical example the long preservation of ancient forms of speech in remarkable purity in Iceland, an out-of-the-way island; and, by way of contrast, we like to mention Malta, an island that is very much in the way, where the mingling of peoples has resulted in the mixture of Arabic and south-European tongues. We have already called attention to the diffusion of organic forms as a fundamental ontographic fact, and we know that variation is an ontographic consequence when diffusion leads to a physiographic environment that involves separation as a result either of distance or of isolation. As examples of this sort of geographical relation we may point either to the several species of cassowaries on the islands of Australasia, or to the several races of man on the larger divisions of the lands, - where barriers formed by the oceans are supplemented by a desert barrier in the Sahara and by a mountain barrier in the Himalaya, or to the several nationalities in Europe, where gulfs and barriers are so important. But why not continue this line of inquiry, and instance, as another example of variation following on diffusion and separation, the differentiation of the Romance languages from Rumania to Portugal, and thus bring in the interesting story of the division of the langue d'oc and the langue d'oui by the central highlands of France? Why not include the even more extraordinary story of the words "pecuniary" and "fee," which have come to look so unlike on their arrival in our composite language because of their different paths of travel from their common source? This is of just as good geographical quality as is the contrast between the whites and the blacks in our southern states; they began alike somewhere long ago, and have come to be different through long ages of separation before they are again found together, and separation is an important element of physiographic control. 

 Not only speech but figures of speech are affected by physiographic environment. Two examples of this are found among far-separated maritime peoples who were so impressed by the manner in which a boat is guided by its rudder that they both came to use a word, which primarily meant a stick and secondarily a stern oar, in a later figurative sense meaning the guidance or control of a whole people. One of these examples is found among the Scandinavians, where the same word is used in "the helmsman steers the boat" and in "the king steers the people"; here the original and the figurative meanings of the word both survive today. The other example began with the people of the eastern Mediterranean and is now spread through all the Romance languages and the Romance element of English. French still preserves the original sense of the word in gouvernail, for "rudder"; but we have lost that sense, unless perhaps in a passage in the Bible concerning the "governor" (helmsman or captain) of a ship, and only know the figurative sense: the word "govern" is therefore a long-lasting response to an early maritime environment. In contrast to these maritime examples is one from the Arabian desert: a missionary has told me, - in reply to the question, "What is the Arab word for 'govern,' and what is its origin?" - "It is a word that means 'guide,' and is derived from the word meaning 'to guide a horse.'" In all these cases it is as legitimate to instance the effect of environment on language as to instance its effect on industry. 

The Practical Value of Defining the Content of a Subject

 In occasional conferences with different geographers on such questions as have been treated in this essay, I have gained the impression that they attached relatively little value to abstract considerations, and that it sufficed them to go on with their work without inquiring particularly into the general content of the subject under which it belonged, and without attempting to develop what may perhaps be called the more philosophical view of the subject as a whole. There was a time when I shared this indifference to abstract inquiry - a time when I was, as it were, overwhelmed with the great quantity and variety of material with which I had to make myself more or less familiar, and when there seemed to be no more need than there was occasion of bringing it all under an orderly and systematic scheme. But that time is now a good many years ago, and since then I have passed out of the stage of life in which, we are assured, our original work is to be completed, and have entered well upon the later stage in which the contemplation and arrangement of work previously done is, we are told, more attractive than the accomplishment of new work. It is, perhaps, for some such reason that I have devoted this essay to the content of geography as a whole, because I am persuaded that there is a practical value in abstract considerations such as I have presented, even for younger men, and that if a general scheme of work in accordance with some broad and philosophical view of one's chosen science is formulated by a young geographer early in his career, he will profit greatly from it; for he will thus be led more surely and directly to detect all the facts that are pertinent to any inquiry he may enter upon. Such a scheme is always open to modification as experience increases. If the geographer undertakes field study, as I hope he may, either at home, where there is plenty of field work to do, or abroad, where there is still more to be done, it will serve him well to know as definitely as possible the essential quality of the work before him. If he wishes to become an all-round geographer and to give a thorough geographical account of the region of his field work, he will be greatly aided in keeping his eyes open to the facts before him by bearing in mind the systematic content of the science as a whole, a part of which he proposes to study in the region he has selected.