The Turnover of Farm Population in Kansas
by James C. Malin
November 1935 (Vol. 4, No. 4), pages 339 to 372
Transcribed by lhn; digitized with permission of
the Kansas Historical Society.
ALTHOUGH there is much discussion of the improvement of farm conditions and of the stabilization of agriculture, there is remarkably little specific information of historical character about the behavior of farm population and of the factors which influence that behavior. This study of the turnover of farm population in Kansas presents only one of many phases of an investigation undertaken in that field.
The state of Kansas was divided into five belts, or zones, from east to west and townships were selected in each in sufficient number to make a fairly representative sampling of each area. Except for the third or central belt, the selection resulted in the inclusion in each division of upwards of one thousand farms after the belt was fully settled, the number varying, of course, from time to time. The method for determining the division of the state presented many problems. Which should be used: arbitrary rectangles, time of first settlement, type-of-farming areas of contemporary times, soil, topography, temperature, altitude or rainfall? Arbitrary division into rectangles, while frequently used for statistical purposes, did not appear to have any meaning for this study. From the standpoint of the frontier alone, the division on the basis of time of settlement would seem to be most desirable, but such an arrangement would have a limited relation to subsequent development. As settlement moved from northeast to southwest, the process did not conform with the natural geographical conditions. Type-of-farming areas are more suitable for investigations where time and change do not enter. Soil areas are not sufficiently definite and uniform. Temperature belts in Kansas run northeast and southwest, with the longest growing season in the southeast corner and the shortest in the northwest corner. Altitude belts are similar although they run more nearly north and south, but in this respect also Kansas faces the southeast rather than the east or northeast. Rainfall belts in the eastern part of the state run northeast. and southwest also, but
* This article was read in part at the annual meeting of the Mississippi Valley Historical Association at Cincinnati, Ohio, April 26, 1935.
340 THE KANSAS HISTORICAL QUARTERLY
near the middle of the state, they change directions to north and south. 
For the present purpose, rainfall has been chosen as the basis of division because of the close relation of rainfall to agriculture and because this division more nearly conforms to some of the other possible divisions; time of settlement, temperature and altitude without their exaggerated extremes. The first rainfall belt of 35 inches per year and upward runs from the northeast corner of the state southwestward. This region includes the section most heavily populated during the territorial period. The second rainfall belt of 30 to 35 inches includes Brown and Nemaha counties in the northeast and extends southwest into the east central section, including such cities as Emporia, Newton and Wichita. The third rainfall belt of 25 to 30 inches extends westward to a line nearly north and south through Ellsworth, Great Bend and Pratt. The fourth belt of 20 to 25 inches extends to a point slightly west of the one-hundredth meridian. The fifth belt of less than 20 inches rainfall includes the remainder of the state west to the Colorado line. 
The selection of the township samples presented its difficulties. The boundaries of the townships must remain unchanged through the years for which census data are available, or if divided, the subdivisions must include the original area. The sample townships must be strictly rural in character without being isolated. The presence of a small town is permissible, but a city of any size would introduce the suburban factor which is a problem in itself. The township should be large enough to be fairly representative, and foreign populations or other unusual influences must not be present in sufficient degree to dominate or distort the results. In practice it has been found all but impossible to find townships that have not
MALIN: TURNOVER OF FARM POPULATION 341
been influenced by foreign population to some degree at some time in their history.
The materials used are the original federal and state census records giving names and other data for all farm operators, at five-year intervals from 1860 to 1885 and for ten-year intervals thereafter until 1915, after which five-year intervals are resumed.  Lists of names of farm operators were compiled from each census for all townships analyzed and these lists compared with succeeding name lists to determine operators who were represented in the township in their owls right or through male descendants.  From resulting statistics the following types of data could be established; firstly, the total number of farm operators at successive census dates; secondly, the persistence of farm operators; and thirdly, the proportion of the farm operators of any period who are descendants from those of any prior period.
The whole number of farm operators, both for the townships taken separately as well as for them taken as groups, increased through the settlement period, frequently, if not usually, to a number in excess of what the land would support under the existing stage of economic development. The second phase was usually a recession in numbers accompanied by an increase in the size of the farm unit. Beyond that point few generalizations seem possible. When the numbers were plotted in graphic form the curves showed no uniformity of pattern. After the frontier or settlement period the townships took on characteristics of established communities, but not necessarily of stabilized communities. Only in the eastern part could the term stabilization be applied with any degree of accuracy, because only there has sufficient time elapsed for fairly adequate adjustments to environment to be completed. The peculiarities of agricultural problems on the plains require a longer period of adaptation than has elapsed since the original settlement. And furthermore, throughout the state, both east and west, the advent of power machinery disrupted much of the adaptation already supposedly achieved.
342 THE KANSAS HISTORICAL QUARTERLY
The period of depression between 1870-1875 recorded moderate losses of farm operators in the first and second belts. The next five years brought increases in population along with partial economic recovery for most of the state. The five years 1880-1885, generally prosperous, present a different picture. The first and fourth belts lost farm operators. In the fourth it was the result of reaction from the boom in the northwest counties. The decade 1885-1895, mostly one of national and world-wide depression, shows increases in the first, second and third belts, the older region, but decreases again in the fourth or younger part of the state. The decade 1895-1905 recorded decreases in four belts, but substantial increases in the fourth. This was the period in which the fourth belt was achieving relative stabilization on a basis of hard winter wheat farming, and succeeded in running counter to the trends of the country both to the east and to the west of it. The decade 1905-1915 was the first one in which all belts registered the same trend, a substantial increase, especially in the fifth. Decreases occurred during the next five years, the World War period, except in the fifth, and increases during the first half of the twenties, except in the second and third. The period 1925-1930, another period of national prosperity, brought declines in numbers in all belts. It is a period of rapid mechanization of agriculture and correspondingly enlarged farm units. In the depression years 1930-1935 the decline in numbers was reversed except in the second belt. The 1930-1935 change in direction was substantial, otherwise it might not be significant as census rolls were probably more complete in 1935 than in former years on account of the federal agricultural allotment policy. For emphasis it may be well to stress the fact that the number of farm operators increased between 1930 and 1935 even in the semiarid fifth rainfall belt, the so-called "dust-bowl."
Although only limited generalizations may be permissible from these variegated data, a few things stand out. Economic depression was usually associated with declining numbers of farm operators during the frontier or settlement stage of development of the country, but increasing numbers usually occurred in older parts of the state. On the other hand, national prosperity was associated with increasing numbers in all parts of the state between 1905 and 1915, and with declining numbers in most of the state between 1925 and 1930. 
MALIN: TURNOVER OF FARM POPULATION 343
These conclusions have an important bearing on the so-called safety valve theory of the frontier hypothesis. It has been rather generally assumed by the followers of F. J. Turner that unfavorable conditions in the east or older regions resulted in a flow of population westward to free or cheap land, thus affording relief to the east and providing opportunity to the migrants. The data collected in this study do not seem to bear out such a theory. On the frontier the number of farm operators declined more often than it increased during periods of general economic stress. On the other hand the increases occurred in most substantial numbers in the older counties and especially those containing a town of some size. The significance of the shift resulting from depressed economic conditions appears to lie therefore in urban to rural rather than oldcountry to frontier readjustment. This urban-to-rural movement was conspicuous while there still was an open frontier and it was conspicuous in the 1930-1935 period after the frontier was gone.
The study of the agricultural census rolls, name by name and farm by farm, reveals many changes which cannot be presented statistically. For the decade 1925-1935, some of these furnish significant background for interpretation of the data. During the twenties rapid mechanization and increased size of farms necessarily reduced the number of farm operators. Many of the less efficient were squeezed out and found it difficult to make a living at any other occupation. The towns received most of them and thereby added to marginal urban population. Much of the tradition of agricultural depression of the twenties was associated with these who were eliminated or who were on the borderline. More accurately these farmers were the victims of a revolutionary advance in agricultural technology. Also the period seems to have encouraged the early retirement of many older operators from active management of their land. The depression of the thirties seems to have reversed to some extent both of these tendencies. Near larger towns especially, there was subdivision of farms associated with the town-to-country movement. Another tendency seems to be an attempt on the part of the head of a family to provide for all members through subdivision of the farm. In other cases farmers who had retired appear to have returned to active operation of their land. In still others, instead of older farmers retiring outright, many seem to reserve a small plot of ground which they operate separately from the original farm. There was an increasing tendency also for the sons in a family to operate a farm jointly under
344 THE KANSAS HISTORICAL QUARTERLY
such a title as Jones Brothers, or a father and sons to handle the farm jointly. Obviously there are conflicting factors present in these cases of joint management. In some it seems to have been a substitute for subdivision of the land, while in others the situation suggests that the joining of forces was a means for carrying on large-scale operations. Subdivision, consolidation and preservation of the size of farm units went on at the same time. The effect of subdivision and consolidation is to cancel or offset each other in the statistics. The figures for the number of farms in any township or county, therefore, may be unchanged, but the farm situation may be changed radically.
The second phase of the problem of population turnover, the persistence of farm operators, affords more that is unusual. General conclusions are presented first. In all rainfall belts, the rate of turnover was high, but was declining during the first twenty-five years from the time of settlement. At about twenty to thirty years after settlement, the rate of turnover may be said to have become somewhat stabilized, although the word stabilized must again be used loosely. In some cases, instead of stabilization, there was an increase in the rate of turnover after that high point twenty-five years from settlement. After the World War persistence increased substantially.
For purposes of summarizing persistence of population by rainfall belts, the data on the several sample townships were added together for each belt, and the persistence was expressed in percentages of the total of persons included in each base census list who remained at successive later census periods. In the first, or eastern belt in 1860 there were 478 operators in the five sample townships. Five years later only 35 percent remained; at the end of ten years 26 percent; at the end of twenty-five years 20 percent; in 1920, or after 60 years, 10.6 percent, and in 1935, or after 75 years, 8.3 percent. Taking in succession the years 1865, 1870, 1875, 1880 and 1885 as base years, the percentage of persistence increased to a high point in 1885. There were 953 farm operators in 1885, and of these ten years later 51.4 percent remained, after another ten years 40.8 percent, in 1920 after thirty-five years, 24.6 percent, and in 1935, or after 50 years, 19 percent. The next base year, 1895, showed increasing instability; only 47.7 percent remained after ten years. It was not until after 1915 that the 1885 level again was reached. The last three base years, 1920, 1925 and 1930, showed
MALIN: TURNOVER OF FARM POPULATION 345
little variation from the high mark of about 66 percent after five years and 5,6 percent after ten years. 
The five different townships in the first rainfall belt varied quite widely. Doniphan is the northeast county of the state along the Missouri river. Center township includes the county seat, Troy. Its leading economic interests are corn, livestock and apples. The farm population was highly stabilized at 55 percent at the five-year point and 46 percent at the ten-year point for the 1865 base year and changed little until the 1915 base year, when it declined somewhat. After the World War the stabilization reached a high percentage of 70.8 for the period 1920-1925 and declined to 67.1 for the years 1930-1935, being the only township in this belt to decline in stability after 1920.
Alexandria township in Leavenworth county lies in the Stranger creek valley, just to the west of the city of Leavenworth. Much of the township is rough and in the early day was timbered. Water and wood made it especially attractive to early settlers. It is a general farming area. It did not reach a high degree of stability as early as Doniphan county, but the level rose steadily to a high percentage of 71.6 for the 1920 base year for the period 1920-1925. After irregularity for 1925-1930, it made a new high of 72.4 percent for 1930-1935.
Eudora township in Douglas county is mostly bottom land, settled by Germans in the north part and Quakers, the Hesper community, in the south. It is a general farming township. In level of stability it was between Doniphan and Leavenworth, with a percentage of approximately 66 for each five-year period after 1920.
The most irregular population movements of the five were found in Valley township of Linn county. This community occupied the north watershed of the Marais des Cygnes river on the Missouri border and contains the village of Trading Post, made notorious in territorial days by John Brown's "Parallels." The first high point of stability was reached at the 1875 base year with 51 percent at five years and 49 percent at ten years. The second high point was 1905 at about the same level as 1875. The third high was the 1925 base year with 58 percent for the five years 1925-1930, but with a low figure of 37 percent for 1930-1935. The 1930 base year showed a decline also for the 1930-1935 quinquennium.
The most stable township of the group was Kanwaka in Douglas,
346 THE KANSAS HISTORICAL QUARTERLY
a general farming community, lying on the ridge dividing the Wakarusa and the Kansas river valleys and between the historic towns of Lawrence and Lecompton. A very high stability was found for the base years 1865, 1870 and 1875 of 58 percent to 65 percent at the five-year point and 46 percent to 53 percent at the ten-year point. The second and third high levels were the 1885 and 1905 base years with 71.2 percent and 63 percent at the ten-year mark. Beginning with the 1915 base year the five-year level of 70 percent was practically unchanged for the succeeding base years.
The second rainfall belt, represented by six townships, started with 1860 also, as its first base year, and the curve of persistence was similar to the first rainfall belt, differing only in details. A high degree of stabilization occurred by 1885, and the curves for the base years 1895 to 1915 were almost identical with 1885. At the end of fifteen years the four stood close to 45 percent. At that point the 1915 curve diverged but the others continued close together. The level of the 1920, 1925 and 1930 lines rose to a high point in 1930 of 71 percent at the end of five years. 
Brown county, in this belt, is in the heart of the Kansas corn belt and lies just west of Doniphan county. The 1875 base year showed the highest percentage of stability until the postwar period, with 73 at five years, 61 at ten years and 47 at fifteen years. Thereafter there was some irregularity at lower levels until the 1905 base year, which opened a period of increasing stability to almost the 1875 level. The 1925 base year was definitely lower, but the 1930 base year was again high at 72 percent for 1930-1935.
Lyon county, lying in the blue-stem pasture region, contributed three townships to this group. Agnes City township is mostly pasture, Pike township is largely bottom land with more general farming and alfalfa. Reading township partakes somewhat of the characteristics of both. These townships were outstanding in showing an unusually high level of stability in the earliest year. Pike township maintained a higher level of stability for 1860, 1865 and 1870 than for any base years since. This may be accounted for in part by the fact that it contained a closely-knit Quaker community. The 1860 base year retained 60 percent at five years and 51 percent at ten years. The 1865 base year retained 68.8 percent at five years and 66.2 percent at ten years. The 1870 base year retained 67.1 percent at five years and 59.6 percent at ten years. The middle years 1875-1905 were highly stable, but at a lower level. By 1915
MALIN: TURNOVER OF FARM POPULATION 347
the stability had risen to 60 percent retained after five years and 50 percent after ten years and succeeding base years remained almost unchanged until 1930 which advanced to 64 percent at the five-year point. The record of Agnes City township was very similar except the level of stability was not so high in the early years. Reading township, which was given present boundaries by 1875, reached higher levels but was more irregular.
Harvey county is in the eastern part of the wheat belt. Macon township lies between two towns, Newton and Halstead. After a somewhat irregular beginning it achieved a very high stability by 1915. That base year retained 67 percent of its farm operators after five years and 58 percent after ten years. The postwar years continued the stabilization process until the 1930 base year achieved a high level of 75 percent retained in 1935. Alta township is about the center of a triangle formed by the cities of Newton, Hutchinson and McPherson. In its early years it was settled by Mennonites from Russia and Germany. The percentages of persistence are quite irregular, but are relatively high. In the early years, 1880 to 1905, inclusive, the Alta township level was higher than Macon township, but since that time Macon was more consistent. and retained higher percentages, except for the five-year figures on the 1920 and 1930 base years. The number retained after ten years was higher for Macon than for Alta even for these two exceptions.
The third rainfall belt is represented by four townships, but for early years two whole counties were used, Dickinson and Saline. The first base year was 1860, using Dickinson county alone, which gave a percentage of 58.3 percent retained at the end of five years and 42 percent at the end of ten years. The 1865 base, using both counties, retained 43 percent at the end of five years. By 1875 the township lines were sufficiently established to change to the township units and one township in Phillips county was introduced for 1875 and one from Kingman in 1880. The high point of persistence was the 1875 base year for which 57 percent remained after five years, 47 percent after ten years, 37 percent after twenty years, 21 percent after thirty years, and 11 percent after forty-five years. All base years from 1875 to 1895, inclusive, showed a lower rate of persistence. Beginning with 1905 the level of stability rose steadily to the last base year, 1930, with 73.5 percent after five years.
The third zone is in the east central wheat belt. Dickinson and Saline counties lie in the lower Smoky river valley, which in the seventies received the name the Golden Belt as descriptive of its
348 THE KANSAS HISTORICAL QUARTERLY
leading crop. Jewell county, on the Nebraska line, raised less wheat and more corn and livestock. Kingman county is predominantly a wheat country. The record of these counties was so nearly uniform that they need not be treated separately. From the time of settlement to the World War each base year retained 55 percent to 58 percent of its farm operators after five years, and 41 percent to 46 percent after ten years. Kingman county, the one farthest southwest, was highest in stability, closing in 1935 with 80.8 percent of the farm operators of 1930. 
The fourth rainfall belt started from an 1875 base losing in five years all but 38 percent, in ten years all but 24 percent, in twenty years all but 14 percent, and after sixty years there remained 4 percent. The 1880 base year followed closely the same curve at the ten- and twenty-year points, but held up to more than 6 percent at the fifty-five-year mark. The base years 1895, 1905 and 1915 reached a high point of stability for the prewar period at more than 47 percent. In the postwar period the level of persistence rose in each successive census until the 1930 curve reached 76.1 percent in 1935 or at the end of five years. 
The fourth belt is in the heart of the Kansas wheat region and in area it is the largest of the five rainfall divisions. The selection of ten townships was made from eight different counties. On the northern border two counties, Phillips and Decatur, produce corn and livestock as well as wheat. Ellsworth, Russell, Ellis and Ness counties include a good representation of cattle country. Edwards county is devoted almost altogether to wheat. Barber county produces cattle and wheat. Five of the individual townships, in Barber, Decatur, Edwards (Trenton), Ellsworth and Russell counties, were moderately irregular in turnover until 1905 or 1915, and thereafter increased consistently in stability to a high level of 70 percent to 80 percent for the five years 1930-1935. In the others the irregularity from base to base continued through their whole history, but all arrived at a level of 70 percent or more for the final five years. In 1905 the level of stability declined in the townships from Barber, Decatur, Edwards (Kinsley and Wayne), Ness and Phillips counties, but increased in the other four. In 1915 the decline occurred only in the townships from Barber and Decatur counties, and in 1920 only in Ness, Phillips and Russell counties.
The fifth rainfall belt, represented by two whole counties, and five townships from two others, was settled in the late eighties. As the
MALIN: TURNOVER OF FARM POPULATION 349
federal census for 1890 is closed to investigators, the first base year available is the state census of 1895. At the ten-year mark this belt retained 33 percent of its members, at twenty-five, 16 percent, and at forty years (1935) 8.8 percent. The succeeding base-year curves were consistently higher, and that of 1925 substantially higher-with 59.1 percent retained after five years and 50.5 percent after ten years. The record for 1930-1935 was only eight tenths of a point lower. For the region as a whole the record of stability for post-World War years is lower than for the belts farther east, but an analysis by separate counties presents a different view. 
Three of the counties represented lie on the west line of the state, Cheyenne in the Republican river valley, Wallace in the Smoky river valley and Hamilton in the Arkansas river valley. Gove county is the third county east from Wallace, in the Smoky river valley. The cattle industry was dominant in this region until the wheat boom under the influence of power farm equipment in the twenties. Throughout the whole history of these counties, however, there was a wide divergence between them in stability of population, but the record was quite consistent within each one. Cheyenne county was always most stable, Wallace next and Gove, farther east, was third. Hamilton county was substantially lower than the others, and as it turned out its numbers in the post-World War period had too much influence as against the four townships of Cheyenne county in the combined figures for the fifth rainfall belt. Cheyenne county, represented by four townships, not only had the highest level of stability in this belt, but it ranked near the top for any rainfall belt. Only four townships were higher in the fourth belt, those in Ellis, Ellsworth, Ness and Russell. Two were higher in the third belt, those in Dickinson and Kingman. Three were higher in the second belt, those in Brown and Harvey. Only Kanwaka township in Douglas county was higher in the first belt. The record for Wallace county would average well with townships in any part of the state.
A study of individual townships presents additional interesting data. Jaqua township, in the southwest corner of Cheyenne county, while somewhat irregular from year to year, achieved the highest level of persistence of any township in the state represented in this study, regardless of location." It had no near rival in the fourth rainfall belt. Vinita township in Kingman county was nearest to
350 THE KANSAS HISTORICAL QUARTERLY
it in the third belt, Macon township in Harvey county in the second, and Kanwaka township in Douglas county in the first. In spite of its low average, Hamilton county had one township, Bear Creek, with an exceptionally high stability which would place it favorably in any rainfall belt. 
Several factors enter into the situation in the fifth belt that are either absent or less pronounced farther east. In age of settlement it had scarcely passed the frontier period when the World War came, if the same time is allowed for that process as in the eastern belts. On the contrary it might be argued that modern industrialism had shortened the period necessary for frontier adjustment. At least, there were some very different influences at work, but there is little or no clue to what their effect should have been on stability of farm population. In connection with the later period, the World War stimulated somewhat the emphasis on wheat, but the wheat boom proper, associated with power farm equipment, did not come until the last half of the twenties and the early thirties. It was more extensive in the southwest counties, such as Hamilton, than in the northwest. The depression did not begin to make itself felt in a serious way until the winter of 1931-1932.
In connection with the wheat boom two unusual factors were introduced, the absentee farm operator (often called the suit-case farmer) and the farm corporation. Adequate treatment of these is not possible because complete information of a nature required for this kind of a study was not collected by the census enumerators and some of the names of these classes may not have been placed on the rolls. In Wallace county only a few absentee operators, who can be clearly identified as such, were listed and none was listed in the Cheyenne townships used, nor in Grainfield township in Gove county. In Hamilton county quite a number appeared.
The rolls for 1935 are probably most complete because of their use by the federal allotment administration. In Wallace county eighteen absentees were listed, or 4.3 percent of the farm operators. In Hamilton county twenty-five absentees were listed, or 6 percent of the operators of the county. In Lamont township in the latter county ten of the ninety-five farm operators, or 10.5 percent, were of this class in 1931 and in 1935 fourteen of ninety-one farm operators, or 15.4 percent.
The wheat-farming corporations were present in Wallace and Hamilton counties, but held the larger acreage in the former. In
MALIN: TURNOVER OF FARM POPULATION 351
1930 one corporation was listed in Hamilton county with 3,360 acres, or the equivalent of five 640-acre farms. By 1935 its holdings had been disposed of, but a second corporation which acquired acreage after 1930 still held 2,800 acres in 1935. In Wallace county in 1930 one corporation held 25,610 acres distributed through three townships, or an equivalent of forty 640-acre farms. By 1935 it still held 3,200 acres. The census rolls do not show how many farm operators were displaced when these corporations accumulated their acreage, nor how many new operators returned when the corporations were carrying out forced liquidation of their holdings under the requirements of the legislative act of 1931. There can be no question, however, that the net effect of both absenteeism and the corporation farming episode was to increase instability of farm operators, even though the extent of that influence cannot be determined.
The history of the turnover of farm operators seems to fall into three periods, except in the fifth belt; the settlement period of exceptionally rapid change, a middle period of relative stabilization at rather low levels, and the recent period of higher stability. During the settlement period exceptionally heavy losses of population are registered for the first and second base years in the first, second  and fourth rainfall belts, and relatively moderate losses in the third and fifth. The Civil War period occupied the four years following the 1860 census and might seem to account for the great losses in the first and second belts, but the same fact could not account for the opposite effect in the third belt.
In most of the curves the losses of population during the settlement period are especially heavy for the first ten years, and then the curve flattens out during the second decade. For the curves representing the period of relative stabilization, the losses are not so great during the first decade, and are relatively greater for the second decade than for the first base-year curves. In other words, these losses after stabilization are distributed more evenly over the first twenty years, rather than being concentrated in the first ten years as in the curves for the settlement years.
The second period has been characterized as one of relative stabilization. The rate of turnover was still high. Few townships retained more than 55 percent to 60 percent of their farm operators for five years or 45 percent to 50 percent of them for ten years. The period 1915 to 1920, the World War era, seems to mark a
352 THE KANSAS HISTORICAL QUARTERLY
division point for most townships between the second and third periods. With relative suddenness the percentage of persistence for five years increased by ten to fifteen points, many townships retaining between 65 percent and 82 percent. For the first time it could be said that the emphasis was on stability rather than change. In the older eastern part of the state this new development appeared in a few cases as early as 1905, but a number of instances are found in 1915 and by 1920 it was genera1.  The general trend was for stability to increase with the age of the community.
The outstanding fact to be derived from this analysis of the persistence of farm operators, however, is that the general pattern presented by the curves of persistence is very nearly the same for the five rainfall belts. The two extreme western belts show results only slightly lower on the whole than the eastern belts, although some of them are actually higher. In other words, the persistence of farm operators was a relatively constant factor, except for the immediate settlement period. While the total number of farm operators fluctuated, the rate of turnover was constant. When the total was declining, it meant only that the losses from the normal turnover were not being replaced by new arrivals, and when the total number rose, it meant that they were more than being replaced. In either case the losses from any particular base period were going on at a fairly constant rate.
Further analysis of the curves does not indicate any uniform reflection of the influence of economic cycles or of rainfall cycles. If anything, after the communities became established, periods of drouth and depression such as 1895 and 1930 when taken as a base tended to show a higher stability than some other periods. The same is true of the post-war depression in many townships using 1920 as a base. The periods of reputed prosperity, such as those beginning with 1905 or 1925, displayed an unusually high rate of turnover in many communities. The relation of soil and land tenure to turnover require further study. The foreign population was usually more stable than the native born, but not so much so as is usually supposed.  The second and later generations seemed to take on rather quickly much of the characteristics of the native born. When the combined data. for each of the rainfall belts is broken down into the individual township samples, the separate curves of persistence show wide fluctuation, but the fluctuations
MALIN: TURNOVER OF FARM POPULATION 353
within a rainfall belt, with a few exceptions, are as wide on the whole as between rainfall belts.
From the facts available, it would appear that the problem is primarily one of group behavior, apart from specifically assignable accidents of farm life in the separate communities or regions involved. In other words, under any given set of general conditions, the farm operators in all parts of the state reacted in much the same manner, the variations of local physical environment exercising only a secondary or minor influence. For any conclusions that may be drawn, that assumption may well be employed as the point of departure. An interesting suggestion in this connection may be derived from a study of persistence of students in college. The data on a freshman group entering the college of liberal arts of the University of Kansas in 1928 provides a curve of persistence over a period of four to six years identical in shape with the curves for Kansas farm operators over a period of twenty to thirty years. Whether the matter has any significance or not, the fact remains that the students as a group in their brief career in college behave in much the same manner as their parents in their career as farm operators. Unfortunately comparable data are not available for other social groups.
The third phase of the problem of turnover of farm operators is to determine the proportion of farm operators of particular periods who are descendants of those of an earlier period. This procedure makes the approach from the opposite direction from the second. Three base years were chosen, 1885, 1915 and 1935, for the eastern belts whose settlement dated from 1860. The absence of data from the federal census of 1910 made it necessary to choose a prior or later date. The year 1915 was taken because it represented more nearly the base used for comparative purposes in most of the postWorld War discussions of agriculture. The latersettled parts of the state presented other problems, and for part of the third rainfall belt and for the fourth belt two periods were taken, dividing the life of the communities as near the half-way point as possible. The fifth belt was handled similarly, only the mid-point fell at 1915 instead of 1905 as in the fourth. The tables report the results in detail and therefore only brief interpretations will be presented here.  In all parts of the state the original or early settlers and their descendants constitute an extremely small proportion of the later or contemporary community. Except for Kanwaka township,
354 THE KANSAS HISTORICAL QUARTERLY
8 percent is the highest representation the settlers of 1860 held seventy-five years later.  These facts run contrary to much of the tradition about the character of a community being determined by the people who settled it and established its original institutions. Obviously, the pioneers constituted too small a proportion of the later community to exercise a controlling influence. The proportion of a community that can be traced back to later base points rose rapidly as those dates approached the present. It is notable, nevertheless, that in few townships does the proportion of the community which traced its origin to 1915 or 1905 appear as high as might be expected from the percentages of persistence of farm operators indicated in the previous section. In many cases the older families were represented in the community by only one operator, while newer families might have two or more. The opposite is true, however, in a few cases where one or more prolific families came to constitute a large proportion of the community. In one particular case, Wheatland township, Ellis county, the male lines of two families constitute 35.4 percent of the operators of 1935.
A comparison of figures for the five rainfall belts shows quite similar percentages for the different belts, except for the fifth. If age of the community is recognized, however, the percentages there are much higher than for any of the eastern belts at a similar community age.
The fact of the high rate of turnover of rural population during the early period and the middle periods of Kansas history suggests numerous questions about the effects such instability has had on institutions; political, economic, and social. As Edwards county has been studied most intimately, some illustrations are chosen from there. During the frontier stage of development, scarcely a mention was made in the press concerning reform of local political institutions. The sole question at issue in elections was the county-seat ring against the field, for the maintenance of power, and incidentally, the money income from offices and county contracts. This was particularly important in the early years during hard times when public money, derived mostly from taxing the railroads, was about the only cash in circulation in the community. In 1880 an unusual
MALIN: TURNOVER OF FARM POPULATION 355
situation was presented when the old county ring, which called itself Republican, was crumbling, and the drouth entered its second year. Emigration of politicians was so extensive that when the time came to call the Republican county convention in the midst of a national campaign, no member of the county committee was a resident of the county. The people got a new deal in politics when a group of citizens assumed the responsibility for calling a mass meeting to reorganize the party. The net result, however, was just to establish a new ring on a basis similar to the old. Whenever the newer settlers in the county tried to get control, the county ring raised the cry of "carpetbaggers" trying to exploit the "old settlers." One editor denounced in language more vigorous than elegant such attempts "to show that unless a man ran wild with the buffalo years ago, he is not eligible to office."  A correspondent closed the incident with the remark that if the candidate in question was a "tenderfoot" then three fourths of the voters were also. A few years later, after the Populist reformers had been in office for some time, a disgusted member of the party protested the failure to reduce taxes and to reform the fee system. One of the officials made a formal reply in which he invoked that age-old political wisdom so dear to reformers as well as to old party men that discussion "may create dissension in our party," and that the writer of the protest "implies that it would greatly please him . . . for the present incumbents . . . to preach their own funeral sermons and proclaim themselves fools at one and the same time by taking less than the Republican statute makes it lawful for them to take." 
Certainly the instability of the frontier population, together with the bare breadand-butter existence of the community as a whole, retarded the progressive adaptation of local political institutions to meet the obligations expected of them as a result of rapidly changing economic and social conditions. This influence was not limited to the frontier, because the older eastern communities were demoralized by competition with western agriculture. And furthermore, even after the frontier had achieved the status of established rural communities, the high rate of turnover of population kept at a minimum the interest which this moving farm population generated for its changing, yet unchanged, local institutions.
The economic development of the farm communities was not promoted by the high rate of turnover of farmers. Possibly agri-
356 THE KANSAS HISTORICAL QUARTERLY
cultural methods did not suffer so seriously in the humid areas of the East as they did in the more arid country west of the Missouri river. Each new crop of Eastern farmers found that they must unlearn most of what they thought they knew about agriculture and adapt themselves to new methods and to new crops. A large proportion starved out or moved out for other reasons without ever learning. It is well to remember that the tillage methods and the varieties of wheat that have given outstanding distinction to the Kansas hard winter wheat belt have become standard only since 1905. Forty years is a long time to discover what crops to raise and how to grow them.
Social institutions suffered more seriously if anything than others. Again and again lyceums, debating societies, literary clubs, and dancing clubs were organized only to break up within a few months. Each new organization usually carried a large proportion of new names indicative of the rapidly changing population. Churches suffered along with other institutions. In Wayne township, Edwards county, the first religious organization was Methodist church, South. Why this should have been is difficult to explain, because there were scarcely any southern people in the community. Possibly a preacher on a nearby circuit was willing to add this community to his other charges as an additional source of income. During the cattle boom of the middle eighties the ranch element and large farmers organized a Protestant Episcopal church, and hauling stone from a distance built a little Gothic church amid the sand hills. It had its six gables and a cross, a vestry room, an organ room, paneled ceilings, stained glass memorial windows, and rented pews. With the crash in the cattle business, the ranches came into the possession of the 160-acre farmers and the loan companies, and aristocracy in religion disappeared with the cattlemen. The 160-acre farmers organized a Methodist Episcopal church and put up a little frame building. After 1900 Missouri immigrants came and with them the Christian church and the Baptist church. With the shifting of population both of the latter failed in a few years, leaving only the Methodists. Probably the Methodists survived only because of a strong centralized organization and an emotional religion which provided the psychological compensation necessary in the arid life of the plains. Across the river in the German communities, the Catholic church, also a strong centralized organization with a genius for reaching the masses, maintained its position. Churches organized on a relatively independent congregational basis
MALIN: TURNOVER OF FARM POPULATION 357
had little chance of survival amid an unstable and changing population, except in the larger towns. This discussion of churches suggests further that changing community characteristics are closely related to changing sources of interstate migration, but that is a problem of sufficient importance to require a separate treatment.
The foregoing discussion applied to the period prior to the World War, when instability was the outstanding fact. The recent period of relatively high stability may produce different results. In any event, if the higher level of stability persists, it provides a substantially different population environment within which economic, social and political institutions must function and develop. What the results may be, only time can tell.
Under prevailing conditions in agriculture it would be remarkable indeed not to recognize the question whether the conclusions reached concerning the turnover of farm population have any significance for current agricultural policies. On April 11, 1935, a mid-West economic conference held at Kansas City devoted a session to the subject of land utilization policies. The plan of the national resources board was outlined, explaining how the government planned to purchase seventy-five million acres of submarginal land, to extend grass areas as protection against erosion, to relocate farmers on more economically planned farms, and for other purposes. In the course of the discussion S. L. Miller, of the University of Iowa, was quoted in the press as saying: "I was brought up in western Kansas and I know you have a lot more failures where you fail to get rain. The problem is whether or not we intend to conserve our resources." The opposite side of the question was taken by E. S. Sparks, of the University of South Dakota, who was quoted as saying: "It has always been my experience that good farmers succeed almost anywhere you put them and poor farmers will fail on the best land in Iowa. Why shift them around? You still have the problem of farm management."
He developed his theme further by declaring that until the government knew more about what it was doing the program looked like folderol. 
It is clear that Sparks was looking at the problem from the standpoint of the farmer as an individual and as a member of a group whose behavior is determined by forces among which national land utilization and other economic policies and conditions are largely incidentals. The results of this investigation of the turnover of
358 THE KANSAS HISTORICAL QUARTERLY
farm population do not constitute proof, of course, but so far as they may contribute to enlargement of knowledge about the conditions within which agricultural policies must function, they tend to give support to the Sparks contention. Possibly one further observation might be added, that the whole discussion might well lead to a reconsideration of the time-tried National Grange dictum that the farmer is more important than the farm.
In certain respects the relation of farm population movements to agricultural policy is reasonably clear. The mere fact of a high degree of persistence or of mobility of farm operators does not necessarily mean either prosperity or failure. A highly mobile population may be prosperous and a highly stabilized community may be stagnant and backward. On the other hand, the reverse may be true in both cases. At least there is nothing in this study to the contrary except during the early frontier stage. Much that has been proposed in the way of agricultural policy implies either directly or indirectly that a causal relationship does exist. If policies designed to increase rural prosperity are expected to stabilize rural population there is little hope of success. If resettlement or land utilization projects require operators to remain over a period of years, they will not hold out much hope thereby of insuring prosperity. Attempts to stabilize population in such resettlement plans run counter to the group habits of Kansas farmers, and there is no reason to believe that they differ widely from other farmers of the major agricultural areas. It is vital to such policies to know first why farmers move as they have done and why a rapid stabilization occurred during the post-World War period, and whether there is reason to assume that the high level of stability will continue. Without a fairly exact analysis of the factors determining such movements any agricultural policy which directly or indirectly involves movement or resettlement of farm population is obviously a step in the dark.
CHARTS AND TABLES
The charts are divided into three parts, except for the fifth rainfall belt. The lower division presents the data on the early years, approximately the frontier period. The middle division presents the data for the middle period of stabilization at relatively low levels. The top division presents the data for the post-World War period of higher stabilization. The year 1915 is included in both the middle and top divisions, because it seems to be a transition year and its presence in both divisions serves as a guide for more effective comparison of the two periods.
MALIN: TURNOVER OF FARM POPULATION 359
Chart I.-Five townships are represented in this chart, all starting from 1860.
Chart II.-The 1860, 1865 and 1870 curves represent only two townships in Lyon county, Agnes City and Pike.
The 1875 and later curves represent six townships in three counties, Brown, Harvey and Lyon.
Chart III.-The 1860 and 1865 curves represent the whole of two counties, Dickinson and Saline. After 1870 the population became so large that it did not seem practicable to carry them further in that form.
The 1875 curve represents one township each from Dickinson, Jewell and Saline counties.
The 1880 curve represents one township each from Dickinson, Kingman and Saline counties. The census roll for Sinclair township in Jewell county is missing for that year.
The 1880 and later curves are based on one township each in the four abovenamed counties, except 1895 for which the census roll for Walnut township in Saline county is missing.
Chart IV.-The 1875 curve represents four townships, one each in Ellsworth and Barber counties, and two in Edwards county.
The 1880 curve represents eight townships, one each from the counties of Barber, Ellis, Ness, Phillips, and Russell, and three from Edwards. Ellsworth is not included because of changes in township lines.
The 1885 curve represents ten townships, the same ones named for 1880 with the addition of one from Ellsworth and one from Decatur counties. The remaining curves on this chart are based on the same ten townships.
Chart V.-This chart contains only two divisions because of the short period since its first settlement. The inserts at the right of the figures for the belt as a whole present the curves for individual counties, and for certain individual townships.
The bottom division of this chart presents a summary of the charts for the five individual belts in the form of a composite of all years for each belt arrived at by averaging the percentages of persistence for each base year. This procedure is open to criticism, but it is sufficiently accurate to assist in presenting general trends. As the data for the first three belts begins with 1860, the fourth with 1875 and the fifth with 1895 the instability associated with the frontier period has too much influence in the averages for the western belts.
360 THE KANSAS HISTORICAL QUARTERLY
CHART I. FIRST RAINFALL BELT.
361 MALIN: TURNOVER OF FARM POPULATION
CHART II. THE SECOND RAINFALL BELT.
362 THE KANSAS HISTORICAL QUARTERLY
CHART III. THE THIRD RAINFALL BELT.
363 MALIN: TURNOVER OF FARM POPULATION
CHART IV. THE FOURTH RAINFALL BELT.
364 THE KANSAS HISTORICAL QUARTERLY
CHART V. THE FIFTH RAINFALL BELT.
365 MALIN: TURNOVER OF FARM POPULATION
TABLES OF PERCENTAGES OF PERSISTENCE OF FARM OPERATORS IN KANSAS.
366 THE KANSAS HISTORICAL QUARTERLY
SECOND RAINFALL BELT, 30" TO 35": EAST CENTRAL KANSAS.
367 MALIN: TURNOVER OF FARM POPULATION
THIRD RAINFALL BELT, 25" TO 30": CENTRAL KANSAS.
368 THE KANSAS HISTORICAL QUARTERLY
FOURTH RAINFALL BELT, 20" TO 25": WEST CENTRAL KANSAS.
369 MALIN: TURNOVER OF FARM POPULATION
[chart of fifth rainfall belt 20''--: western Kansas]
370 MALIN: TURNOVER OF FARM POPULATION
PERCENTAGE OF FARM OPERATORS IN EACH RAINFALL BELT
REPRESENTING OPERATORS OF EARLIER PERIODS.
371 THE KANSAS HISTORICAL QUARTERLY
[Percentage of farm operators in each rainfall belt (continued).]
372 MALIN: TURNOVER OF FARM POPULATION
[Percentage of farm operators in each rainfall belt (Concluded).]
1. Maps showing types-of-farming areas, rainfall and growing seasons may be found conveniently in Hodges, J. A., et al., "Types of Farming in Kansas," Kansas Agricultural Experiment Station Bulletin 251 (August, 1930). A soil map is to be found in the Twenty-eighth Biennial Report of the Kansas State Board of Agriculture, opposite p 100, in conjunction with an article on soils of Kansas by R. I. Throckmorton, pp. 91-102.
2. GROUP 1, First Rainfall Belt.-Doniphan county, Center township; Leavenworth county, Alexandria township; Linn county, Valley township; Douglas county, Eudora township, Kanwaka township
GROUP II, Second Rainfall Belt.-Brown county, Walnut township; Lyon county, Pike township, Agnes City township, Reading township; Harvey county, Macon township, Alta township.
GROUP III, Third Rainfall Belt.-Jewell county, Sinclair township; Dickinson county, Buckeye township; Saline county, Walnut township; Kingman county, Vinita township.
GROUP IV, Fourth Rainfall Belt.-Phillips county, Long Island township; Ellsworth county, Lincoln township Russell county, Big Creek township; Ellis county, Wheatland township; Edwards county, Kinsley township, Trenton township, Wayne township; Barber county, Sun City township, Deerhead township, Turkey Creek township; Decatur county, Center township; Ness county, High Point township.
GROUP V, Fifth Rainfall Belt.-Gove county, Grainfield township; Cheyenne county, Jaqua township, Benkelman township, Calhoun township, Cleveland Run township; Wallace county, Harrison township, Morton township, North township, Sharon Springs township, Stockholm township, Vega township, Wallace township, Weskan township; Hamilton county Bear Creek township, Coolidge township, Kendall township, Lamont township, Liberty township, Medway township, Richland township, Syracuse township.
3. These state census records for the period 1860-1925, inclusive, are deposited permanently with the Kansas Historical Society. Those for 1930 are temporarily in the possession of the department of agricultural economics of Kansas State College of Agriculture and Applied Science and those for the current year 1935 are held temporarily by the federal agricultural statistician at Topeka. No federal census data were used for 1890 and since, because the more recent federal records are closed to the public.
4. The shift of farm operation from father to children is usually very small during the first ten years from any particular census date and only somewhat larger during the second ten years. By the end of twenty years relatively few of the families in question are represented in the township, as will be seen from the analysis of data later in the paper, so the element of error inevitable through inability to follow the female line is relatively small. If a family includes male children, the possibility of the male succeeding to the farm instead of the females tends to minimize this constant error.
5. In this discussion the words "result" or "cause" have been excluded and the phrase "associated with, or an equivalent, is used in order to avoid any implications of "cause-effect" relationships.
6. See Chart I at the end of this article.
7. See Chart II.
8. See Chart III.
9. See Chart IV.
10. See Chart V.
11. See Chart V, inserts.
12. See Chart V, insert.
13. The 1865 base year is an exception in the second rainfall belt.
14. The federal census of 1910, if open to research, would be of particular interest at this point.
15. These subjects will be treated in separate studies.
16. See the table at the end of this article.
17. These figures are too low, but they are the nearest possible, because the female lines of descent cannot be traced from the census rolls. Investigation of this problem in one township, through the aid of old settlers, points to a conclusion, however that the error is relatively small for the average community, because the extent of popufation movement was so nearly complete. Furthermore, if the farm continued in the hands of the family the probability was in favor of a son continuing rather than a daughter. For later base points this kind of error is probably greater than for the old-settler period. The amount of error of this kind is probably greater here, however, than it is in the previous section of this study.
18. Kinsley Weekly Mercury, November 3, 1887.
19. Kinsley Graphic, March 29, 1896.
20. Kansas City (Mo.) Times, April 12, 1936.