COMPLEXITY OF STRUCTURAL DESIGN DP _, .;~*_'.—‘5* ii DAYTIME DRESSES, AND ITs RELATIONSHIP-To :f. :gi’ f fi ~ EDUCATIONAL AND OCCUPATIONAL LEVELS-QR, j ..: : , - . - WOMEN FROM. 1860 T111940 ‘TheSisfoT the DegTee bf’MIA, A ' 5 :1 a; MICHIGAN STATE UNIVERSITY JOYCE LEONARD ALLRED -' 1971 u —. A I. 1- 3 LIJRAR}; a? IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII « 3 ’I-uoq ‘ ; 7 { 113;, } MAGIC 2 ABSTRACT COMPLEXITY OF STRUCTURAL DESIGN OF DAYTIME DRESSES AND ITS RELATIONSHIP TO EDUCATIONAL AND OCCUPATIONAL LEVELS OF WOMEN FROM 1860 TO 1940 BY Joyce Leonard Allred The major purpose of the study was to develop and to test the reliability and validity of a quantitative index of the complexity of structural design of clothing. The index was used to compare the complexity of structural de- sign of women's daytime dresses to changes in fashion from one costume period to another from 1860 to 1940, and to determine whether or not a relationship exists between the complexity of structural design of women's daytime dresses and the educational and occupational levels of women between 1860 and 1940. The index of complexity of structural design was the number of meters of stitching per square meter of fabric in the garment. The variables included in the index of com- plexity of structural design were the area of garment fabric, the area of supportive fabric, the area of trim, and the total length of stitching in the garment. Trim Joyce Leonard Allred was defined as decoration which stands away from garment fabric and constitutes a separate structural area of the garment. The index of the educational level of women used in the study was the percentage of women in the United States of school age who were attending school. The occupational level of women was defined as the percentage of women in the United States of working age who were employed outside the home. The sample was composed of daytime dresses randomly selected from the Costume Collections of the College of Human Ecology and the Museum at Michigan State University. A total of eighty dresses, ten from each of eight costume periods were included in the sample. A one-way analysis of variance was used to test for significant differences between the means of the complex- ity of structural design scores for each costume period. Post hoc comparisons were used to identify the exact differences between the means. Correlations were computed to establish whether or not relationships exist between the complexity of structural design of women's daytime dresses and the educational and occupational levels of women across the eight costume periods included in the study. The findings of the study showed that significant differences between mean scores for complexity of Joyce Leonard Allred structural design were present in only three of twenty- seven possible comparisons among the eight costume periods. Definite trends in the data were noted, however, indicating that the lack of significance may have been due to factors such as costume periods spanning too many years, too broad a definition of a daytime dress, the presence of garments combining characteristics of two costume periods, and the small number of dresses from each costume period. No significant relationships were found between the complexity of structural design of women's daytime dresses and the educational and occupational levels of women. Additional, serendipitous findings also evolved from this study. The shape of bodice back pieces appears to be a reliable indicator of the costume period in which a gar- ment was worn. The scores of complexity of structural design of garments worn during the 1870's and 1880's did not differ appreciably from those of the 1920's. This is contrary to the common assumption that women's clothing became simpler in design in the years from 1870 to 1930. Another assumption which appears to be contradicted by the findings of this study is that the invention of the sewing machine was responsible for the lavish ornamentation of dresses worn during the 1870's. The greatest amount of stitching in the garments in the sample from the 1870's, that to finish and attach the trim, was done by hand. Joyce Leonard Allred The costume periods with the highest means and greatest variability on the index of complexity of struc- tural design coincided with the beginning of Young's silhouette cycles and possibly the introduction of a new fashion. COMPLEXITY OF STRUCTURAL DESIGN OF DAYTIME DRESSES AND ITS RELATIONSHIP TO EDUCATIONAL AND OCCUPATIONAL LEVELS OF WOMEN FROM 1860 TO 1940 BY Joyce Leonard Allred A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Human Environment and Design 1971 / / / I 1‘ ‘. yi,fif 01 ACKNOWLEDGMENTS The writer gratefully acknowledges her advisor, Dr. Anna M. Creekmore, for her encouragement and guidance throughout this study; her committee, Dr. Norma S. Bobbitt, Dr. Susan K. Kilborn, and Dr. Robert R. Rice, for their helpful suggestions; Mr. Val R. Berryman of the Michigan State University Museum for his assistance in obtaining the Museum garments for the sample; and her parents, Mr. and Mrs. Grant I. Leonard, and especially her husband, W. James Allred, whose faith and encouragement made this study possible. ii TABLE OF CONTENTS Chapter Page I 0 INTRODUCTION 0 O O O O O O O O O O 1 II. REVIEW OF LITERATURE . . . . . . . . 5 Methods of Analyzing Costume . . . . . 5 Relationship of Clothing to Changing _/x Status of WOmen. . . . . . . . . (l4 ‘ III. STATEMENT OF THE PROBLEM . . . . . . . 19 Definition of Terms . . . . . . . . l9 Hypotheses . . . . . . . . . . . 23 IV 0 PROCEDURE 0 O O O O O O O I O O O 2 5 DevelOpment of Measuring Instruments . . 25 Complexity of Structural Design . . . 26 Educational Level of Women . . . . . 31 Occupational Level of Wbmen. . . . . 33 Selection of the Sample. . . . . . . 34 Collection of the Data . . . . . . . 38 Analysis of the Data. . . . . . . . 40 V. RESULTS AND DISCUSSION. . . . . . . . 41 Description of the Sample . . . . . . 41 Relationship of Complexity of Structural Design of Daytime Dresses to Fashion Change. . . . . . . . . . . Relationship of Complexity of Structural Design of Daytime Dresses to Educational and Occupational Levels of WOmen . . . 52 . 44 VI. SUMMARY AND CONCLUSIONS . ,. . . . . . 54 iii Chapter VII. RECOMMENDATIONS AND IMPLICATIONS . . . Recommendations for Improvement of the Index of Complexity of Structural Design . . . . . . . . . . Implications for Further Research . . BIBLIOGRAPHY . . . . . . . . . . . . APPENDICES Appendix A. Sketches of Dresses in Sample . . . . B. Identifying Numbers and Sources of Garments in Sample . . . . . . . C. Illustration of Measuring Technique . . iv Page 59 59 60 66 69 79 82 LIST OF TABLES Table Page 1. Mean Scores, Standard Deviation, and Range of Scores for Complexity of Structural Design by Costume Periods . . . . . . 46 2. Differences in Means of Complexity of Structural Design for Costume Periods . . 47 3. Mean Scores for Complexity of Structural Design of Daytime Dresses and Educational and Occupational Levels of WOmen in the United States by Costume Period . . . . 52 4. Correlations Relating Complexity of Struc- tural Design of Women's Daytime Dresses and Educational and Occupational Levels of WOmen from 1860 to 1940. . . . . . 53 CHAPTER I INTRODUCTION WOmen's daytime dresses worn between 1860 and 1940 exhibited great variability in construction, decoration, and fit. The construction ranged from simple silhouette seams to numerous intricate pieces; the decoration from a single row of binding to yards and yards of ruffles, ribbons, and lace, while the fit varied from the tightly boned bodices of the nineteenth century to the loosely hanging dresses of the 1920's. The extent of the change can be detected by viewing the evolution of dresses from the crinoline silhouette of the 1860's, through the ornate, hustled designs of the 1870's and 1880's, to the boyish styles of the 1920's. Changes in women's clothing have been documented by descriptions of characteristic garments of time periods, by division of centuries into costume periods, by analysis of silhouettes, and by tracing the evolutionary develOp- ments of a specific part of a costume. Attempts have ‘been made to classify, quantify, and record the relative frequency of design features of women's clothing. Investigations in this area have yielded varying degrees of success, leaving it a fruitful one for further study. During the 1930's Agatha Brooks Young, noted author of Recurring Cycles of Fashion, identified fashion cycles and fashion silhouettes based on the frequency of design features in illustrations and photographs of garments.1 A. L. Kroeber statistically analyzed "stylistic changes" in women's evening and formal dress by measuring design features.2 His study which was completed in 1919, was followed by a more comprehensive investigation carried out jointly with Jane Richardson in the late 1930's.3 Researchers of the past have concerned themselves with costume silhouettes but have overlooked the qualities of the line and space divisions which result from the shape and arrangement of the separate pieces of fabric which create the garment design. Silhouette is actually an ele- ment of structural design and has been the subject of in- vestigations but the total concept of structural design 1Agatha B. Young, Recurring Cycles of Fashion (New York: COOper Square Publishers, Inc., 1966), P. 553. (Hereinafter referred to as Cycles of Fashion.) 2A. L. Kroeber, "On the Principle of Order in Civilization as Exemplified by Changes of Fashion," American Anthropologist, XXI (July, 1919), 235-63. 3Jane Richardson and A. L. Kroeber, "Three Centuries of Women's Dress Fashions: A Quantitative Analysis," Anthropological Records, V, No. 2 (1940). (Hereinafter referred to as "Quantitative Analysis".) including silhouette, has not been investigated. Struc- tural design is closely related to the line and space divisions within the garment in that the more numerous the divisions of garment design, the more complex the structural design of the garment. The main objective of this study was to develop a method to quantitatively mea- sure the complexity of structural design of women's cloth- ing, and to compare scores assigned to the structural design of women's daytime dresses to changes in fashion through the costume periods from 1860 to 1940. A second objective was to investigate the status changes for women as reflected in educational and/or occu- pational levels of women in the years from 1860 to 1940, and the relationships existing between the status changes and the complexity of structural design of women's daytime dresses. The possibility of the existence of this re- lationship is indicated by the changes in the cultural ideal of women paralleling the changing design of women's clothing. The cultural ideal of the middle nineteenth century appears to have been the beautiful, refined, gentle lady with little formal education, but by the 1920's the ideal seems to have changed to the worldly, sophisticated "flapper." This change in cultural ideal seems to reflect a change in attitudes toward women in general and may be indicative of a change in status. The change in attitudes from the middle nineteenth century to the "flapper" era of the 1920's seems to parallel a status change from homemaker or lady of leisure to financially independent working woman, as well as a change from complex to simple design in women's clothing. CHAPTER II REVIEW OF LITERATURE Since the validity of any study is largely dependent upon the theoretical framework upon which it builds, the following review of pertinent literature presents the basis for the present investigation. Methods of analyzing cos- tume are discussed in the first section while the second section is devoted to the relationship of clothing to the changing status of women. Methods of Analyzing Costume The various methods used in the analysis of costume range from the dichotomous concepts of straight lines versus curved lines, to the more complex statistical analysis of silhouettes by Richardson and Kroeber. Harriet Mchmsey used the simple classification of costumes based on the degree of straightness or curvature of basic design lines. Garments with a predominance of curved lines are considered soft in character and are often associated with garments typically worn by women while garments with straight lines are considered to be more severe and are associated with clothing typically worn by men.1 The concepts of straight and curved lines were used by Margaret WOod Hager in her cross-cultural analysis of clothing on the basis of line factors. Costumes from four cultures including ancient Egypt, ancient Greece, Eliza- bethan England, and Victorian England were analyzed accord- ing to the effect on the body, the type of line, and the direction of line. The first step in the analysis of a particular garment was to determine its effect on the apparent size and shape of the body area it covered. The three classifications were "minimizes, makes natural, or over-emphasizes."2 A distinction was also made as to whether or not the body was misshapen. After a garment was categorized according to its effect on the body it was further classified as to type of line, whether straight, curved, or a combination of both straight and curved. Each of these categories were broken down into classifications as to diagonal, vertical, or horizontal line.3 29“” i. 1‘1""; 1Harriet T. Mchmsey, Art in Clothing§election (New York: Harper & Row, Publishers, 1963), p. 159. (Herein- after referred to as Clothing Selection.) 2Margaret Ruth Wood Hager, "The Costume of Four Selected Cultures, Analyzed on the Basis of Line Factors" (unpublished Master's thesis, Pennsylvania State University, 1963), PP. 14-15. 31bid. The classifications of tailored and draped silhouettes are attributed to Grace Morton. Tailored silhouettes, those with plain, flat surfaces, straight, sharp edges, and shap- ing created by cutting and stitching fabric pieces, are divided into severely or formally tailored and soft or dressmaker tailored. Draped silhouettes, those with soft, curved lines created by hanging fabric in folds, include straight or soft-hanging, bias-draped, and crisp.4 The uni-form-multi-form, determinate-indeterminate system of analysis developed by Marilyn DeLong deals with the total visual effect of costume with its background. A costume is considered to be uni-form if it is viewed as a totality, while a multi-form costume is a unit viewed in terms of its separate parts. The determinate-indeterminate factor concerns the clarity or definition of the surfaces of a costume. A two-dimensional scheme was used to place garments on a continuum from uni-form to multi-form and from determinate to indeterminate.5 The concepts of structural and decorative designs were presented by Harriet and Vetta Goldstein. Structural design was defined as "the design made by the size, form, 4Grace Margaret Morton, The Arts of Costume and Per- sonal A earance (n.p.: University of Nebraska Foundation, 1955), p. 208. 5Marilyn Ruth Revell DeLong, "Analysis of Costume Organization: Development of a Model for Considering Style" (unpublished Ph.D. dissertation, Ohio State Uni- versity, 1967), p. 64. color, and texture of an object, whether it be the object itself, in space, or a drawing of that object worked out on paper."6 They considered decorative design to be the "sur- face enrichment of a structural design."7 Bernice Chambers defines structural design as “that which depends on form and not superimposed ornamentation."8 Harriet Mchmsey considered decorative design to be "ornamentation or sur- face enrichment by applied design."9 The methods of analysis used by Morton and Mchmsey concerned the type of line in costume. Although Mchmsey associated the type of line with clothing typically worn by men or by women while Morton identified draped and tailored silhouettes, both were concerned with the straightness or curvature of basic design lines. Hager's cross-cultural analysis of costume included type of line but added the dimension of direction of line. Each of the three methods provide a narrow analysis of the design of costume in that they deal only with line which is only one element of design. 6Harriet Goldstein and Vetta Goldstein, Art in Every- day Life (New York: The Macmillan Company, 1954), p.45} 71bid. 8Bernice G. Chambers, Color and Design (New York: Prentice-Hall, Inc., 1951). P. 33. 9Mchmsey, Clothing Selection, p. 117. The concepts of structural and decorative design provide a more comprehensive analysis of costume in that they include line as well as shape or form and space divisions within the garment and the garment silhouette. Structural design is the design of a garment created by the form of the garment and the assembling of garment parts. Decorative design or superimposed ornamentation includes still another element of design which is texture. This method of analysis of costume provides a more com- prehensive system of analysis of costume than that which deals solely with line. The uni-form-multi-form, determinate-indeterminate analysis developed by DeLong is even more inclusive than the structural design and decorative design method in that it includes all elements of design except color and is based on the total visual effect of the costume with its background. DeLong's method encompasses the methods used by Morton, Mchmsey, and Hager as well as those of decora- tive and structural designs in that it includes line, form and space, texture, and light and dark. Agatha Brooks Young's theory that the three basic types of silhouettes in women's clothing recurr in cycles was based on a study of silhouettes and design features of typical daytime dresses worn each year from 1760 to 1937. The typical dress for each of the 178 years included in the study was selected by recording the relative fre- quency of characteristics found in fashion illustrations 10 and photographs. The dress combining the most frequently occurring or typical characteristics of a given year was judged to be the typical dress for that year. Illustrations of typical dresses were arranged in chronological order and when carefully examined revealed the slow, continuous change through the three basic silhouettes.lo Young also completed a more detailed analysis of fashion trends by using pictures in fashion magazines and computing monthly mean values of skirt length, skirt width, and shoulder width measured in terms of waist widths. Values for the years from 1931 through 1937 were graphed and when smooth curves were drawn over the raw data the curves were found to reveal trend lines in costume sil- houettes.ll Jane Richardson and A. L. Kroeber studied fashion plates for the years from 1605 to 1936 and quantitatively analyzed stylistic changes in women's evening dresses by measuring garment dimensions. The six basic measurements taken were length of skirt or dress, length of waist, length or depth of decolletage, width of skirt, width or 12 thickness of waist, and width of decolletage. The ratio of each measurement to the total length of figure was loYoung, Cycles of Fashion, pp. 3, 7. llIbid., pp. 158-59. 12 p. 112. Richardson and Kroeber, "Quantitative Analysis," ll computed and used for analysis. Values such as the mean for each year for each measurement, five- and ten—year running averages, and the variability coefficients for year-to-year and five- and ten-year averages were analyzed. Findings showed basic dimensions of garments to alternate between maximum and minimum widths in cycles of approxi- mately fifty years.13 Despite the differences in the purposes and methods used by Young and by Richardson and Kroeber, both studies resulted in remarkably similar conclusions. These con— clusions were: (1) basic qualities underlying the changes in fashion change slowly and recurr in cycles, and (2) more superficial qualities change more rapidly and there- 14 Both studies fore more often than the basic qualities. identified the slowly changing quality to be garment sil- houettes. Young concluded that three basic silhouettes including the bell silhouette, the backfullness silhouette, and the tubular silhouette, recurred in orderly cycles of 15 about one-third of a century each. Richardson and Kroeber found that basic dimensions of women's dresses swung from 13Ibid., p. 148. 14Young, Cycles of Fashion, pp. 3-4; Richardson and Kroeber, "Quantitative Analysis," p. 148. 15Young, Cycles of Fashion, p. 21. 12 maximum to minimum to maximum width in a hundred years' time.16 Richardson and Kroeber identified only two cycles per century while Young reported three per century. This difference is understandable when the methods used are taken into consideration. Young visually analyzed illus- trations of fashionable garments while Richardson and Kroeber were concerned with quantitatively analyzing dimensions of full-face illustrations only. Obviously the use of full-face illustrations would make the back- fullness silhouette described by Young undetectable and therefore result in the more simple classification of maximum and minimum garment dimensions. The second set of characteristics of fashion change which make this year's dress distinguishable from last year's dress was recognized but not identified by Young or by Richardson and Kroeber. Neither of the studies were concerned with these qualities other than to acknowledge their existence and to state that they change more rapidly than the more basic silhouette upon which they build. These characteristics were the focus of the present in- vestigation. A study of the 178 typical daytime dresses selected by Young indicated that the difference in the typical gar- ment from one year to the next was actually a combination 16 p. 148. Richardson and Kroeber, "Quantitative Analysis," 13 of changes, such as a different position for the waistline or armscye seam, a change in the shape of a collar, or the type and amount of fitting in the waistline of the garment. These changes were found to be nothing more than changes in structural design. If silhouettes could be successfully quantified by Richardson and Kroeber, then perhaps struc- tural design could be also. Richardson and Kroeber used evening dresses as the basis for their study since they were primarily interested in fashion as an art form.17 Young was interested in fashion change as a phenomena worthy of study for its own value. She chose daytime dresses which she defined as the kind of dresses women wore for daytime calls and for shop- ping, for her study for the following reasons: (1) daytime dresses are worn by more women than are evening gowns, (2) evening gowns tend to be more extreme styles than daytime dresses, and (3) fashion trends in evening dresses are more short lived than are trends in daytime dresses. For these same reasons, the daytime dress was judged to be best suited to the present investigation. Various methods have been devised for use in the analysis of costume and the three most pertinent to this investigation are the concepts of structural and decorative design and the methods of Young and of Richardson and Kroeber. The two major conclusions of the studies by 17Ibid., p. 111. 14 Young and by Richardson and Kroeber were: (1) certain qualities of costume change slowly and recurr in regular cycles repeating themselves every century, and (2) other undefined qualities of costume change from year—to—year making fashionable dresses for each year distinguishable from those of other years. The second of these conclusions was the basis for this study. After a preliminary exami- nation of the year-to-year changes, these changes were found to manifest themselves in the form of changes in structural design. The successful quantification of the first quality of costume, the silhouette, by Richardson and Kroeber served as an impetus to attempt to quantify structural design assuming it could be the second quality of costume mentioned by Young and by Richardson and Kroeber. Relationship of Clothing to Changing Status of*Women A number of researchers have suggested relationships of varying strength between women's clothing and the status of women. Some use very specific points to illustrate the relationship while others make broad, general statements implying the existence of a relationship. Characteristics of garments, actual garments, and parts of garments have been associated with the status of women at various points in time. 15 Barbara Allan found the general trend toward sim- plicity in design of women's clothing between 1910 and 1920 to be directly related to women's achievement of "equal rights" with men and their movement from the home into the world of business.18 Muriel COOper concluded that complicated designs in women's clothing became sim- plified with the needs of women in the business and pro- fessional worlds.19 Leota Alton supported this idea and added that clothing changed to meet the needs of women's way of life, thereby becoming useful as well as beauti- fu1.20 Women's clothing during the decades from 1900 to 1910 and from 1920 to 1930 was related to women's employment patterns during the same periods of time, according to Marjory Joseph. More specifically, as women entered fields of employment primarily filled by men, women began to wear clothing more like men's clothing in detail, cut, fabric, 18Barbara Allan, "The Flapper: A Study of the Evo- lution of an American Fashion Stereotype from 1910 to 1929" (unpublished Master's thesis, Michigan State University, 1960), p. 78. 19Muriel Sieving COOper, "The Growth of the Commercial Pattern Industry and Its Contribution to Contemporary Dress" (unpublished Master's thesis, State University of Iowa, 1961): PP. 15-16. 20Leota R. Alton, "A Study of Women's Fashions in the United States Between 1860 and 1900" (unpublished Master's thesis, State University of Iowa, 1949), p. 41. l6 and sewing. This affected both the employed and the non— employed women's clothing.21 Sandra Strand considered the costumes worn by women in the 1870's and 1880's which imitated the lines of men's clothing as an indication of women's struggle for equality 22 --their banner of emancipation. The gaudiness and dis- harmony of women's clothing in the mid-nineteenth century was related to the confusion over the role of women during that period, according to Elaine Lumbra.23 In some cases, specific characteristics of garments are associated with the status of women. Elinor Nugent saw the wide bell-shaped skirts of the 1860's as an indication 24 that "woman's place is in the home." Strand, on the 21Marjory Lockwood Joseph, "Changes in Women's Day- time Dress as Related to Other Selected Cultural Factors During the First and Third Decades of the Twentieth Cen- tury" (unpublished Ph.D. dissertation, Pennsylvania State University, 1962), p. 263. 22Sandra Lou Strand, "The Psychology of Dress and Its Effects Interpreted in Nineteenth Century Costume of Women" (unpublished Master's thesis, University of Washington, 1964), p. 224. (Hereinafter referred to as "Psychology of Dress".) 23Elaine Lumbra, "Everyday Women's Dresses Worn in the United States During the Period 1870-1900" (unpublished Master's thesis, University of Indiana, 1966), p. 29. (Hereinafter referred to as "Dresses During 1870-1900".) 24Elinor Roth Nugent, "The Relationship of Fashion in WOmen's Dress to Selected Aspects of Social Change from 1850 to 1950" (unpublished Ph.D. dissertation, Louisiana State University, 1962), pp. 217-18. (Hereinafter referred to as "Relationship of Fashion".) 17 other hand, equated the growing dimensions of women's skirts during the 1850's and 1860's with the symbolic im- portance of women's place in the world.25 According to Nugent, the helplessness of women in the bustled dresses of the 1870's and 1880's was an indication of women's resignation to their dependence upon men; and the change in attitude of the late 1880's was reflected in the free 26 flowing skirts of the period. Lumbra attributed the decrease in the restrictiveness of women's skirts of the late nineteenth century to the emancipation of women.27 Nugent considered the broadened shoulder line of women's clothing of the 1940's symbolic of women's accept- ance of the responsibilities of men during World War II.28 Strand associated the evolution of the walking dress of the 1860's with the beginnings of higher education for women.29 James Laver has stated that the clothing worn by women of any period reflects their relative status during that time. More specifically, he associated periods of 25Strand, "Psychology of Dress," p. 224. 26Nugent, "Relationship of Fashion," pp. 217-18. 27Lumbra, "Dresses During 1870-1900," p. 54. 28Nugent, "Relationship of Fashion," p. 221. 29Strand, "Psychology of Dress," p. 224. 18 female emancipation with garments having straight lines, pale coloring, and waistlines not in the natural position. His point was illustrated by use of the early nineteenth century and the 1920's as examples of such periods.30 The discussion above leads to the belief that some relationship exists between the status of women and women's clothing. A portion of this study was an attempt to vali- date a limited aspect of that relationship. 3oJames Laver, Women's Dress in the Jazz Age (London: Hamish Hamilton, Ltd., 1964), p. 13. CHAPTER III STATEMENT OF THE PROBLEM The primary purpose of the study was to deve10p and to test the reliability and validity of a quantitative technique to measure the complexity of structural design of clothing and to compare scores assigned to the struc- tural designs of women's daytime dresses to changes in fashion from one costume period to another from 1860 to 1940. The secondary purpose was to use the method evolved in determining whether or not a relationship exists between the complexity of structural design of women's daytime dresses and the educational and occupational levels of women between 1860 and 1940. Definition of Terms Commonly used terms often develop vague meanings which are subject to variability in interpretation by the reader. The meanings intended by the writer do not always coincide with those assumed by the reader. The following section dealing with the definition of terms used in this investigation was included to prevent such misunder- standings. l9 20 A costume period is a span of time in which typical garments worn exhibit several particular characteristics in common. These characteristics are not present in the same combination during any other time and make garments worn during a given period recognizable as different from those worn during any other period. A daytime dress is a dress suitable for wear outside the home during the day within a particular costume period, but because of design and/or fabric would not be suitable for a formal evening occasion. It differs from an evening or ball gown in that it usually is more modest in terms of body coverage, is less elaborate in design, and is made from fabric with duller, rougher surface texture. Surface enrichment is design created by the addition of design elements such as color and texture to the fabric from which a garment is made. Types of surface design in- clude sequins, beads, embroidery, cut work, appliques, fagoting, trapunto, all tucking and shirring completed prior to the cutting of garment pieces, and all fabric such as ribbon applied smoothly over the garment fabric. Trim is fabric decoration which stands away from garment fabric or is inserted into garment fabric so that it constitutes a structural area of the garment. Examples of trim are ruffles, edging, and lace inserts. 21 Structural design is the design of a garment created by stitching the separate pieces of garment fabric and trim 'f‘l" together. It includes both the silhouette and the line and space divisions within the garment. “_. --.G .l The complexity of structural design of a garment is ham_____1 . - sun. the number of meters of stitching per square meter of fabric in that garment. hi “Mn-..“ ' I P“ The educational level of women is the ratio of the '. ._ number of women in the United States attending school to the total number of women in the United States of school age during a given period of time. The occupational level of women is the ratio of the number of women in the United States employed outside the home to the total number of women in the United States be- tween sixteen and fifty-nine years of age during a given period of time. Costume periods. Due to the fact that decades and costume periods do not always coincide with each other, the following costume periods were defined for the pur- poses of this investigation: 1. 1860's refers to the years from 1855 to 1867. Garments typically worn during this period had fitted bodices with waistlines in the natural position and full skirts, worn over hoops or crinolines, with the fullness nearly evenly distributed from front to back. 22 2. 1870's actually encompasses the years from 1868 through 1879. Dresses of this period were highly decorated with ruffled or pleated fabrics. Skirts often had over- skirts which added to the fullness of the bustle worn under- neath, and were often lengthened in back to form a train. Bodices were tightly fitted and extended slightly below the natural waistline in front and back. Sleeves were often decorated at the lower edge. 3. 1880's includes the years from 1880 through 1889. The dresses of this period had bodices which extended further below the waistline than those of the previous period. The skirts of this period were worn over a bustle and the excessive decoration of the 1870's was replaced by drapery. Sleeves were long and fitted at the lower edge with slight fullness at the top. 4. 1890's refers to the years encompassed in the decade of the 1890's. The extreme fullness at the top of the sleeve and the full gored skirt shaped to fit at the waistline without gathers were characteristic of the typical garment worn during this decade. 5. 1900's includes the years from 1900 through 1912. The dresses of this period had high collars, sleeves with slight fullness at the top or bottom or both, and gathered skirts which hung without support of substructures. Garments of this costume period were often made of soft fabrics and were fitted with soft gathers rather than darting. 23 6. 1910's refers to the years from 1913 through 1920. The outstanding characteristics of this period were the kimono sleeve, the loosely fitted one piece dress, and skirts which were ankle length or shorter. 7. 1920's includes that period of time extending from 1921 through 1927. This costume period was dis— tinguished by garments with straight lines, shortened skirts, dropped waistlines or no defined waistlines, and often without collars or sleeves. 8. 1930's includes the time Span from 1928 through 1938. Dresses of this period were longer than those of the previous period and had fitted or semi-fitted waist- lines placed in the natural position. These garments were often cut on the bias and had short sleeves with fullness at the top. Hypotheses Young and Richardson and Kroeber conclude that un— defined qualities of costume change from year-to-year within the cycle of a given silhouette. Since variations in the structural design of garments appear to account for these yearly changes in women's dresses, structural design warrants consideration in the definition of these quali- ties. Based on this observation, the following hypotheses were proposed: 1. A quantitative measure assigning a numerical score to the complexity of structural design of women's daytime dresses worn in a given 24 period results in similar scores for all such dresses worn during that period. The mean score of the complexity of the struc- tural design of women's daytime dresses worn during each of the costume periods between 1860 and 1940 will differ from the mean score for typical dresses worn during every other costume period from 1860 to 1940. Since fashion change is a continuous process rather than a series of isolated incidents, and division of time into costume periods is artificial, the following hypothe- sis was proposed: 3. Scores of complexity of structural design of garments worn during a given costume period will differ more from nonadjacent periods than from adjacent ones. A relationship is thought to exist between the status of women and women's clothing. Tate and Glisson state that the changing status of women is reflected in their educational and occupational levels.l Researchers have suggested that women's clothing becomes simpler as women move into the business and professional worlds. 4. The complexity of structural design of women's daytime dresses will be inversely related to the educational levels of women in the United States for the decades between 1860 and 1940. The complexity of structural design of women's daytime dresses will be inversely related to the occupational levels of United States' women for the decades between 1860 and 1940. 1 Mildred Tate and Oris Glisson, Family Clothing (New York: John Wiley & Sons, Inc., 1961), p. 12. CHAPTER IV PROCEDURE The first portion of the description of procedure deals with the development of the measure of complexity of structural design and the other indexes used in this study. The selection of the sample used to test the hypotheses of this investigation is discussed in the second section of this chapter. The two remaining portions of the chapter are devoted to the collection of the data and the analysis of the data. Development of Measuring Instruments The process involved in the development of the method of quantifying structural design is described below. Since quantitative indexes of the educational and occupational levels of women were desired for comparison with the scores of complexity of structural design and none were available, indexes were devised for this study. The development of these indexes is discussed in a later section. 25 26 Complexity of Structural Design The first step in the development of the measure of complexity of structural design was to enumerate as many as possible of the variables which might be related to structural design. Those variables listed were: the number, shape, and grainline of pieces of fabric composing a garment; the amount and type of hand and machine stitch- ing; the amount and type of trim; the amount and type of fitting to body contours; the type of seams; the area and weight of supportive fabrics; and the area of garment fabric. Each of these factors were studied singly and in combination with other factors. Grainline of garment pieces was eliminated from the index since it is a function of the accuracy with which a garment is cut, the quality of the fabric, and the skill of the dressmaker. The type of stitching and the type of seam were also deleted from the list since they are too closely related to the quality of construction of a par- ticular garment, the skill of the dressmaker, and the technological developments at the time the dress was made. Also eliminated from the list of variables were the amount and type of fitting necessary to mold the fabric to the body form. The fitting in any garment is as closely related to the size and contours of the body of the wearer as to the structural design of the garment. A larger body with more prominent body curves requires more fitting and 27 often a different type of fitting than a small body with more angular lines. The number and shape of pieces of fabric in a garment were excluded from the measure since they were found to be unreliable estimates of structural design as defined for this investigation. Two garments might have identical numbers of fabric pieces but the pieces of one garment could have intricate fitting and darting within the pieces which would make it more complex in terms of structural design. Likewise, the varying shapes of garment pieces would be accounted for by the relationship of their area (area of garment fabric) to their perimeter (the length of stitching necessary to attach it to other portions of the garment). In other words, if two pieces of a garment with equal area, one having smooth edges and one with irregular edges are compared, the perimeter of the smoothly shaped piece is less than that of the irregularly shaped piece. Therefore, the amount of stitching necessary to encircle the irregularly shaped piece would be greater and the ratio of the length of all stitching to the area of fabric would be greater. The complexity of the structural design would then be greater for the irregularly shaped piece and would be accounted for by the use of area of fabric and length of stitching. Amounts of garment fabric, supportive fabric, trim, and stitching were evaluated to be most closely related 28 to the structural design of a garment. Amount of garment fabric was included because bulk of fabric increases the lines and space divisions within a garment and contributes to the structural design of that garment. For example, a garment made of three yards of fabric is less complex than a garment of the same size made of ten yards of fabric, just by virtue of the additional amount of fabric. The supportive fabrics in a garment were included in the index since the qualities added to the garment fabric by suppor- tive fabrics affect the garment silhouette which is a part of structural design. Stitching is the means by which the design is executed and therefore an essential component of an index of structural design. A more complex structural design requires more stitching for seaming and fitting than does a simple design. Some types of applied decoration were included in the index of complexity of structural design even though they are commonly considered decorative rather than struc- tural design. Applied decorations fall into one of two categories, they are applied smoothly over the garment fabric or they stand away from the garment fabric and constitute a separate structural area of the garment. Those types of decoration such as sequins, beads, em- broidery, cut work, appliques, fagoting, and trapunto are applied smoothly over the fabric as surface enrich- ment and are comparable to the pattern printed on or woven into the fabric and are not part of structural 29 design. The other type, referred to here as trim, includ- ing decorations such as a ruffle or an edging constitute a structural area of the garment and therefore should be considered structural design. Trim may contribute to the silhouette of a garment and since silhouette is part of structural design, trim should be also. By way of illus- tration a band of lace applied smoothly over the garment fabric would enrich the surface of the fabric and would not be included in the index of complexity of structural design for that garment. If, however, the lace is gathered or is inserted into the garment fabric, it is part of the structural area of the garment and would be included. Like— wise, the stitching to attach the latter would be included in the index of complexity of structural design while that to attach the former would not. The four variables included in the index of complex- ity of structural design were area of garment fabric, area of supportive fabric, area of trim, and length of stitch- ing. The possibility of weighting garment fabric scores more than those for supportive fabric and trim was evalu- ated but found to be unnecessary in that all three factors contribute equally in prOportion to their area to the structural design of a garment. Garment fabric was weighted more heavily in the sense that in most cases it constitutes a larger area of the garment than does the trim or the supportive fabric. In some garments the three 30 variables overlapped to the extent that trim or supportive fabric could not be distinguished from garment fabric. A tentative decision was made to simply add the scores for each of the four variables described above to get the total score for complexity of structural design. Although this combination was more satisfactory than others evaluated, it did not allow for variability of garment size. A correction factor which accounted for differences in waistline circumference and height of the wearer was developed but found to be of little use in garments with no waistline shaping. The great variability in the dis- tance from the hemline of the dress to the floor further complicated the use of such a correction factor. The final form of the index of the complexity of structural design was the total number of linear meters of stitching in a garment, excluding the exceptions listed below, divided by the number of square meters of fabric, including all garment fabric, supportive fabric, and trim in the garment. More simply, the structural design was measured by the amount of stitching per square meter of garment fabric. The following were considered exceptions not related to the structural design of a garment and were not in- cluded in the index: area of seam allowances and darts, stitching required for preliminary construction such as stay-stitching, stitching for seam finishes and button- holes, stitching required for piecing fabric, reinforcement 31 stitching, top stitching, area of fabric and stitching in pockets and figure improving pads, and stitching used to apply surface enrichment as well as surface decoration. Reliability of the index was a function of the accuracy with which the garments were measured. The in— vestigator measured two different garments of varying complexity on three different occasions and compared the three scores obtained for each. Differences among the three scores for a garment were so small as to be negli- gible. To test the validity of the index, a panel of five graduate students in Textiles and Clothing were asked to rank a group of ten dresses in order of complexity of structural design as defined for this investigation. Their reliability as judges was evaluated by asking them to repeat the ranking procedure with the same garments one week later. The ten garments were then measured and the complexity of structural design was found to be con- sistent with the order established by the panel. Educational Level of WOmen The best quantitative measure for this investigation of the educational level of women would have been the total number of women in the United States completing more than eight years of formal education divided by the total number of women in the United States. This measure was found to be inOperable since information of the type 32 needed was not available as far back as the 1860's, there- fore, a compromise was made. The most accurate data avail- able were in the United States Census Reports for each dec- ade from 1860 to 1940. The total number of women in the United States attending school was divided by the total num- ber of school-age females in the United States. School age was defined as between the ages of five and twenty years of age. This index of the educational level of women in the United States does not account for the level or type of for- mal education received by women in relation to that received by men and does not distinguish between completion of edu- cational requirements and mere school attendance. In addi- tion, possible changes in compulsory education laws and vari- ation in the accessibility of educational facilities are not reflected in this index. Despite these limitations, the index was selected as adequate for the scope of this par- ticular study. Although the collection of data for the United States Census for each decade began in the first year of that decade, collection was not completed until several years later. The 1860 census, for example was not com- 1 pleted and reported until 1864. Likewise, the 1920 cen- 2 sus was not reported until 1923. On the basis of the fact that census data were not reported until several 1U.S., Department of Interior, Bureau of the Census, Eighth Census of the United States, 1860. 2U.S., Department of Commerce, Bureau of the Census, Fourteenth Census of the United States, 1920. 33 years into the decade, the assumption was made that the figures of each census are valid for the entire decade. Occupational Level of WOmen The best quantitative index of the occupational level of women was evaluated to be the percentage of women in the United States working outside the home with some adjustment or weighting factor related to the various types of occupations of women as compared to the relative status of the types of occupations of men. The development of a truly reliable and valid measure including those variables was judged to be beyond the scope of this investigation. The variable of relative status presented problems in that the relative status of various occupations may well have changed in the years from 1860 to 1940. In addition, occu- pational status is composed of elements such as the exact duties and responsibilities involved in a particular job, the size and relative status of the employing organization and the amount of training and/or skill necessary to ful- fill the responsibilities of the job. Each of these factors must be carefully studied and weighted in the context of each decade. The best alternative for the purposes of this study was judged to be the total number of women in the United States working outside the home divided by the total number of women in the United States of working age,between the ages of sixteen and fifty-nine years. This index of the 34 occupational level of women does not account for the rela- tive status of women's occupations as compared with that of men's occupations. Changes in the economy relative to employment-unemployment are also excluded. However, with full awareness of these limitations, the measure was deemed acceptable for use in this investigation. Selection of the Sample Although the study by Young and the one by Richard- son and Kroeber utilized pictures and illustrations as their source of data, the present study necessitated information in greater detail that could be obtained only from actual garments for each period. Young cautioned against the use of historic costumes alone since she felt the type of garments found in costume collections are often not representative of garments actually worn. The reason they have been preserved may well be related to their exceptionalness.3 She was, however, concerned with the single most typical dress for each of 168 years and the probability of exactly the most typical dress appearing in a costume collection would be relatively small. This study deals with the variety of fashionable dresses worn in a given costume period. A garment cannot be judged to be unrepresentative just by virtue of the fact that it is found in a costume collection. 3Young, Cycles of Fashion, p. 3. 35 Representativeness must be judged on the basis of more extensive information. Karen Marie Basralian, in her study of the day dress of the 1890's, concluded that a firm relationship existed between the dresses worn by upper and middle class women (the type found in costume collections) and those illustrated in fashion journals.4 A minimum of five issues each of two or more fashion magazines for each costume period were studied carefully to gain proficiency in judging the representativeness of specific garments. Young's 178 typical daytime dresses and five historic costume books were also reviewed. The investigator tested the reliability of her judgment by examining other illustrations with specific dates and about which judgments as to representativeness had been made by others such as fashion editors or authors of his- toric costume books. Garments used in this investigation were only those which were judged to be representative in terms of silhouette and design. The assumption was then made that garments representative in silhouette and design would be representative in all other aspects related to the study. The limited quantity of garments available and the ambiguity of assigning exact dates to historic costumes 4Karen Marie Basralian, "The American Women's Day Dress, 1890 Through 1900, as Reflected in the American Fashion Magazines" (unpublished Master's thesis, Uni- versity of Maryland, 1969), p. 157. 36 made a study of year-to-year changes impossible. Since a costume period is a period of time in which garments worn are similar enough to each other to be identified as a group, it was assumed that the variability in the struc- tural design of garments worn within a given costume period would be less than that between different periods. For this reason, the costume period was the unit of analysis rather than the year as used by Young and by Richardson and Kroeber. Richardson and Kroeber found the variability within a given period was not significantly different from the variability between periods.5 This is understandable and does not pose a serious problem in this study since their investigation dealt exclusively with evening dresses which show great variability by virtue of their extreme styles and their rapid changes in design. Richardson and Kroeber defined the periods in their investigation strictly on a chronological basis with no consideration of costume periods. They concluded that fashion change does not occur at a uniform rate and therefore a division of time into five- or ten-year periods cannot be expected to coin- cide with fashion changes. Garments used in the study were those found in the Costume Collection of the College of Human Ecology and in the Museum at Michigan State University. The eight costume 5 p. 149. Richardson and Kroeber, "Quantitative Analysis," 37 periods included in the investigation were selected on the basis of the dresses available. Since the years from 1860 to 1940 include some part of each of three of Young's silhouette cycles, and follow the invention of the sewing machine, the eight costume periods were evaluated as satisfactory for the study. A sample size of ten garments for each costume period was judged to be adequate and realistic in terms of garments available and sufficient for statistical analysis. The increase in precision which would have been gained by increasing the sample size beyond ten garments per costume period was not great enough to warrant such an increase. Equal sample size was desirable because it allowed the use of a non-conservative test in post hoc comparisons in the statistical analysis. Garments for each costume period were reviewed and evaluated by the investigator in terms of condition, defi- nition of a daytime dress, and similarities to garments judged as representative of the particular costume period in which the garment was worn. Dates on the garments were not used in placing them in specific costume periods, since many were approximate and appeared to have been assigned from memory. Those garments which were in poor condition with pieces missing or badly torn, and those which showed any evidence of remodelling were eliminated. The broad definition of the daytime dress necessitated by the limited 38 number of garments available placed definite limitations on the study. However, in reviewing the dresses for inclusion in the study, any dress which was not definitely a daytime dress according to the definition for this study was elimi- nated from the study. Each of the garments which met the criteria mentioned above were assigned a number. The sample of dresses in- cluded in the study were randomly selected by use of a table of random numbers. The sample was composed of a total of eighty dresses, ten from each of the eight cos- tume periods. Collection of the Data The eighty garments selected for inclusion in the sample were assigned a number. Dresses representing the 1860's were assigned numbers ranging from one through ten, those representing the 1870's eleven through twenty, and so on. The numbers from one through eighty were then randomly ordered using a table of random numbers. Gar- ments were measured according to the randomly determined order. This procedure was followed to compensate for the possible confounding variable of increased skill and/or accuracy in the process of measuring eighty dresses. The method of measuring the area of fabric consisted of dividing each of the garment pieces into geometric figures. The areas of each of the separate geometric 39 figures were then added together to give the total area of that particular garment piece. Garment pieces were divided into small rectangles, triangles, and trapezoids using fabric grainline as a guide. Dressmaker pins were used to define the perimeters of the figures. Fabrics too fragile for pins were divided into figures by laying contrasting colored thread across the fabric. Curved edges such as sleeve caps and necklines were divided into series of straight lines with corresponding geometric shapes. Sharper curves required shorter lines. Experimentation with curved shapes drawn on squared paper indicated this to be the most accurate method of measuring curved areas. Single large figures were used in preference to several small figures whenever possible in an attempt to keep error in measurement and rounding to a minimum. Appendix C contains an illustration of the divisions used in mea- suring the area of a basic sleeve. Data were recorded so totals for garment fabric, supportive fabric, and trim could be figured separately. All measurements were rounded to the nearest one-half centimeter. After calculations combining the areas of all figures were completed, scores were changed from centimeters to meters rounded to the nearest hundredth. Garments were measured in a systematic manner in order to gain efficiency and to reduce possible error of excluding any area of the garment or including an area 40 more than once. Scores for garments which were symmetri- cal were computed by measuring the right half of the gar- ment and doubling the score for that half. This procedure was followed to keep handling and possible damage to the garments at a minimum. Area of garment parts was measured in the following order: sleeves, collar, bodice front, bodice back, skirt front, skirt back, and any remaining portions of the garment such as belts, bows, and over- skirts. Length of stitching was measured in the same manner after area of fabric was computed. The amount of time necessary to completely measure a garment ranged from twenty minutes to four hours. The average amount of time spent measuring the dresses was approximately two and one-half hours. Analysis of the Data A one-way analysis of variance was used to test for significance in the differences between the means of the complexity of structural design for the eight costume periods. Tukey post hoc comparisons were used to detect significant differences between each period and all other periods. The relationships between the complexity of structural design mean for each period and the educational and occupational levels of women were examined by compu- tation of Pearson product-moment correlation coefficients. A significance level of .05 was judged as adequate in all analyses. CHAPTER V RESULTS AND DISCUSSION The following report of findings includes a de- scription of the sample, an analysis of the index of complexity of structural design, and a discussion of the relationship of complexity of structural design of women's daytime dresses to the educational and occupational levels of women. Description of the Sample The sample consisted of eighty daytime dresses, ten from each of the eight costume periods from 1860 to 1940. Sketches and identifying numbers of garments in the sample are included in Appendices A and B respectively. Forty- four of the garments were from the costume collection of the College of Human Ecology. The remaining thirty-six dresses were from the Michigan State University Museum. Labels inside some of the garments indicated they origi- nated in Paris, London, and New York, while others were of local origin. Most of the dresses, however, did not have labels and their origin therefore remains unknown. 41 42 Forty-five of the dresses were walking dresses in that they were more casual than the other thirty-five which tended to be more like afternoon or tea dresses. A t-test of the mean scores for the two types of daytime dresses showed no significant difference between the two groups in terms of complexity of structural design. At least four of the eighty dresses were known to be wedding dresses. According to Cunnington, it was not uncommon for a bride of the mid-nineteenth century to be married in a daytime dress.1 The scores for complexity of structural design for the wedding dresses were not significantly different from those which were not wedding dresses. Silk was by far the most common garment fabric for the dresses since forty-four of the eighty were made of silk. Cotton was the second most frequently used fiber, accounting for twenty dresses. Nine of the dresses were of rayon, four of wool, one of linen, and two were of undetermined fiber contents. Three of the dresses worn during the 1860's were stitched together entirely by hand. Garments of the 1870's and 1880's had the main seams sewn by machine while all of the details and trim were stitched by hand. This would tend to contradict somewhat the claim that the invention of the sewing machine was responsible for the 1C. Willett and Phyllis Cunnington, Handbook of En lish Costume in the Nineteenth Century (London: Faber an Faber, Ltd., 1966), p. 522. 43 lavish ornamentation of the dresses of the 1870's.2 The dresses in the sample with excessive amounts of ornamen- tation had only a few seams sewn by machine, while the majority of the stitching, that used to finish and attach the trim, was done by hand. The amount of hand stitching in the dresses seemed to decrease as the years progressed. Perhaps the most outstanding difference between garments with large amounts of machine stitching and those with the majority of the stitching done by hand was that a single row of hand stitching often served several functions at the same time. It was often used to gather a ruffle, attach the ruffle to the garment, as well as hold the garment hem in place at the same time. A single row of hand stitching served the same functions as three rows of machine stitch- ing found in present day garments. Definite trends in methods of fitting and construct- ing garments were noted. For example, supportive fabrics were used in most bodices and some skirts during the 1860's. By the 1870's supportive fabrics were found in bodices and skirts of all garments and continued to be found in both skirts and bodices through the 1890's. Supportive fabrics were not used as extensively during the 1900's and by 1920's only a few garments contained any type of supportive fabric. Those garments which did, 2Harald Hansen, Costume and Styles (New York: E. P. Strutton and Co., 1956), p. 15 . 44 had silk linings which were different from the cotton underlinings used previously. Another general trend noted was that the shape of fabric pieces used to construct bodices ranged from very simple geometric shapes in the 1860's to more complicated shapes during the remainder of the nineteenth century, back to more simple shapes during the 1910's and 1920's. The writer found this characteristic of garments reliable in placing garments into a costume period. The shape and construction of skirts also showed definite similarities in construction and design within each costume period. Relationship of Complexity of Structural Desi n of Da time Dresses to Fasfiion Change The variables in the index of complexity of struc- tural design were area of garment fabric, area of suppor- tive fabric, area of trim, and length of stitching in a garment. Trim was defined as fabric decoration which stands away from garment fabric or is inserted in garment fabric so it constitutes a structural area of the garment. Scores for complexity of structural design were determined by measuring the total length of stitching in a garment and dividing that total by the area of fabric, including garment fabrics, supportive fabric, and trim in the gar- ment. Not included in the score were: area of seam allowances and darts, stitching required for preliminary construction such as stay-stitching, stitching for seam 45 finishes and buttonholes, stitching required for piecing fabric, reinforcement stitching, top stitching, area of fabric and stitching in pockets and figure improving pads, stitching to apply surface enrichment, and area of surface enrichment. Hypotheses l, 2, and 3 prOposed that complexity of structural design scores would be similar within a costume period, that mean scores for complexity of structural design would differ from one costume period to another, and that the mean score for complexity of structural design would be more similar to scores for adjacent periods than to scores for nonadjacent periods, respectively. These three hypotheses were tested using an analysis of variance technique and Tukey post hoc comparisons. A one-way analysis of variance indicated that the mean of at least one costume period was significantly different at the .01 level, from the mean of at least one other period. The mean score, the standard deviation, and the range of scores for each costume period are given in Table 1. Tukey post hoc comparisons between all possi- ble pairs of means were computed to identify the location of the significant differences. The results of the post hoc comparisons are shown in Table 2. The mean of the complexity of structural design of daytime dresses worn during the 1860's was significantly different from the mean of the 1900's and that of the 46 TABLE l.--Mean scores, standard deviation, and range of scores for complexity of structural design by costume periods. 3232.“? Mean 32323233.. Range 1860's 4.15 2.50 1.16- 9.47 1870's 6.96 3.38 2.81-12.55 1880's 6.44 2.37 2.98-11.18 1890's 5.61 3.15 2.13-12.18 1900's 9.59 3.32 5.56-16.61 1910's 6.19 2.48 3.20-10.22 1920's 6.37 2.17 2.75- 9.62 1930's 8.24 2.46 5.70-14.64 Total for all eighty garments 6.69 3.12 1.16-16.61 47 Hm>ma ac. um uchAHAcmHm4. Hw>ma mo. um unmoAMAcmHm. III mmm.a meo.~ ~mm.H s~m.~ mme.a oem.a .mmo.e omafl III oma. ma~.m ems. meo. mam. e-.~ ommfl II- mmm.m «mm. mem. mes. eeo.~ oama II- .mem.m oma.m -e.~ .«Hee.m coma III mmm. emm.a ~mv.a omma III mum. Hm~.~ owed III mam.m onwa ommH omma oama coma Gama omwa chmd coma .moowumm wasumoo you :mwmmo Hausuosuum mo muwonmEoo mo msmoa ca moosmuommw911.~ wands 48 1930's. The levels of significance were .01 and .05, re- spectively. The only other significant difference was between the means of the 1890's and the 1900's at the .05 level. The low number of significant differences in the mean scores for the different costume periods prompted further analysis using slight modifications of the index of the complexity of structural design. Computations were completed including only the garment fabric and trim rather than garment fabric, supportive fabric, and trim in the area of fabric in a garment. This modification yielded results less significant than the original index and therefore was abandoned. The scores for length of stitching, area of garment fabric, and total area of fabric in the garment were each used singly in an analysis of variance to determine if the means for each of the eight costume periods were signifi- cantly different. None of the three variables alone pro- vided results as significant as those of the original index of complexity of structural design. The scores for dresses within each period classified as walking dresses were compared with those of the afternoon or tea dresses of the same period. Again, no significant differences were found. The means for the complexity of structural design for dresses worn during the 1870's and the 1880's were 6.96 and 6.44, respectively. The dresses worn during 49 these two costume periods were most difficult to identify. The differences in the characteristics of garments worn during these two periods seemed to be more a matter of degree than type. The same problem arose in placing gar- ments in the costume periods of the 1910's and the 1920's. The mean scores of these two periods were also very close in numerical value, 6.19 and 6.37. These similarities of scores appear to indicate that the structural design of garments worn in costume periods with similar character- istics will also be similar in terms of complexity. The garments with scores at the extreme ends of the range for a given period were found to be garments which had some characteristics of garments worn in a later or previous period. The dress with the score of 9.47 in the 1860's has an overskirt which is characteristic of the 1870's even though the rest of the garment has character- istics of the 1860's. The dress with the score of 2.98 in the 1880's is also a transitional garment between the 1880's and the 1890's in that the sleeves have only a slight amount of fullness at the top placing it in the 1880's, but it has a skirt requiring only a small bustle and with no draping which are characteristic of the 1890's. It appears that more significant results might be found if the sample was more strictly defined to include only garments which possess characteristics unquestionably placing it within a given costume period. Another 50 alternative would be to increase the number of costume periods in the same span of time thereby making finer distinctions as to characteristics of garments. It should also be noted that the period with the lowest mean contained eight dresses which might be cate- gorized as walking dresses. The 1900's, the period with the highest mean, contained only three walking dresses and seven afternoon dresses which were more characteristic of garments worn for formal afternoon occasions, and would therefore be eXpected to be more elaborate in design and more complex in terms of structural design. The scores for the walking dresses did not differ significantly from those for the afternoon dresses across the eight costume periods. When comparisons between the two types of day- time dresses within a costume period were made, differences were also not significant. There appeared, however, to be a relationship between the magnitude of the score for com- plexity of structural design and the degree of formality of the garment. With only ten dresses for each costume period, the sample size within a period may not have been sufficiently large to detect the apparent difference since the smaller the size of the sample, the larger the minimal detectable difference. By limiting the definition of a daytime dress to that type appropriate for a specific day- time activity such as shopping and maintaining sample size at ten garments per period, more significant differences between means across periods might result. 51 Richardson and Kroeber indicated that fashion does not change at a regular rate, but ranges from periods of high stability and low variability to periods of insta- bility and high variability.3 They attribute the differ- ences in variability to the fact that women's dress styles tend toward an ideal. Once the ideal is achieved it is followed by equilibrium, relative stability, and low vari- ability. Then saturation occurs and style strives for a new ideal, causing instability, and high variability until the new ideal is reached. It may be noted that the costume periods in this investigation with the highest standard deviations or variability coincide with the beginning of Young's backfullness and tubular silhouette cycles. The 1870's has a standard deviation of 3.38 and the 1900's, the beginning of the second tubular cycle,4 has a standard deviation of 3.32. The means for these two periods and that of the 1930's, perhaps the beginning of another cycle, are slightly higher than the means surrounding them. This seems to indicate that periods of instability or high variation coincide with the beginning of a new silhouette cycle and perhaps more generally with the period following the introduction of a new fashion. 3 p. 149. Richardson and Kroeber, "Quantitative Analysis," 4Young, Cycles of Fashion, p. 115. 52 Relationship of Complexity of Structural Design of'Daytime Dresses to Educational‘andIi Occupational Levels of WOmen The occupational level of women was defined for this investigation as the percentage of women of working age who are actually employed outside the home during any given period. The educational level of women was defined as the percentage of females of school age actually attending school. The educational and occupational levels of women in the United States from 1860 to 1940 are presented in Table 3. TABLE 3.--Mean scores for complexity of structural design of daytime dresses and educational and occupational levels of women in the United States by costume period. Mean of Costume Complexity of Educational Occupational Period Structural Level Level Design 1860 4.15 45.6 12.4 1870 6.96 43.8 17.9 1880 6.44 52.8 14.7 1890 5.61 51.1 17.4 1900 9.59 51.6 18.8 1910 6.19 60.5 23.4 1920 6.37 65.5 21.1 1930 8.24 69.7 28.5 53 Hypotheses 4 and 5 proposed relationships between the complexity of structural design of daytime dresses and the educational and occupational levels of women from 1860 to 1940. Pearson product-moment correlation coefficients were computed to test these relationships. Table 4 con- tains the results of the computations. TABLE 4.--Correlations relating complexity of structural design of women's daytime dresses and educational and occupational levels of women from 1860 to 1940. Educational Occupational Level Level Complexity of a a structural design .30 .52 aNot significant at .05 level. No significant relationship was found to exist be- tween the complexity of structural design of women's day- time dresses and the educational and occupational levels of women from 1860 to 1940. However, the possibility of the existence of such a relationship cannot be ruled out on the basis of this investigation since the results of this study may be due to the coarseness of the indexes used. On the other hand, the century long trend toward more education and higher employment for women may be more powerful than other factors such as cultural ideal which might simultaneously affect the educational and occupational levels as well as the dress of women. CHAPTER VI SUMMARY AND CONCLUSIONS The major concern of this study was to develop and to test the reliability and validity of a quantitative index of the complexity of structural design of clothing. The index was used to compare the structural design of women's daytime dresses to changes in fashion from one costume period to another from 1860 to 1940, and to deter- mine whether or not a relationship exists between the com- plexity of structural design of women's daytime dresses and the educational and occupational levels of women be- tween 1860 and 1940. A review of the literature revealed that attempts have been made by Richardson and Kroeber to quantify fashion silhouettes but no research attempt to quantify structural design has been made. Examination of year-to- year changes in women's dresses indicated that these changes manifest themselves in changes in structural de- sign. Therefore the structural design of garments worn during a given period should be similar to that of other fashionable garments worn during that period. The 54 55 literature also indicated that changes in the design of women's clothing are related to the changing status of women. Hypotheses were proposed suggesting differences in the complexity of structural design of women's daytime dresses worn during each costume period between 1860 and 1940 and relationships between complexity of structural design of women's daytime dresses and the educational and occupational levels of women. The sample was composed of daytime dresses, randomly selected from the collections of the College of Human Ecology and the Michigan State University Museum. A total of eighty dresses, ten from each of eight costume periods were included. The variables included in the index of complexity of structural design were the area of garment fabric, the area of supportive fabric, the area of trim, and the total length of stitching in the garment. Trim was defined as decoration which stands away from garment fabric or is inserted into garment fabric and constitutes a separate structural area of the garment. The index of complexity of structural design was the number of meters of stitching per square meter of fabric in a garment. The index of the educational level of women used was the percentage of women in the United States of school age who were actually attending school. The occupational 56 level of women was defined as the percentage of the women in the United States of working age who were actually em- ployed outside the home. F-tests were used for significance of difference be- tween the means of the complexity of structural design scores for each costume period. Post hoc comparisons were used to identify the exact differences between means. Correlations were computed to establish whether or not a relationship exists between the complexity of structural design of women's daytime dresses and the educational and occupational levels of women across the eight costume periods included in the study. The prOposed hypotheses and the results are sum- marized below: Hypothesis 1 A quantitative measure assigning a numerical score to the complexity of structural design of women's daytime dresses worn in a given period results in similar scores for all such dresses worn during that period. Hypothesis 2 The mean score of the complexity of the structural design of women's daytime dresses worn during each of the coStume periods between 1860 and 1940 will differ from the mean score for typical dresses worn during every other costume period from 1860 to 1940. Hypotheses 1 and 2 were not confirmed by the results of the study as differences in the variation within periods and between periods and in means for each combination of 57 costume periods were not significant. Only three of the twenty-seven differences between means were significant at the .05 level. Hypothesis 3 Scores of complexity of structural design of gar- ments worn during a given costume period will differ more from nonadjacent periods than from adjacent ones. The means of nonadjacent periods were not signifi- cantly different from the means of adjacent periods for any of the eight costume periods. Therefore, this hypothe- sis was not confirmed. Hypothesis 4 The complexity of structural design of women's day- time dresses will be inversely related to the edu- cational levels of women in the United States for the decades between 1860 and 1940. The correlation between the educational levels of women in the United States with the complexity of struc- tural design of women's daytime dresses worn for the decades from 1860 to 1940 was not significant. Hypothesis 4 was rejected. Hypothesis 5 The complexity of structural design of women's day- time dresses will be inversely related to the occu- pational levels of United States' women for the decades between 1860 and 1940. No significant relationship was found to exist be- tween the complexity of structural design of women's day- time dresses and the occupational levels of women. Hypothesis 5 was rejected. 58 The results of this study are inconclusive in that significant results were not obtained except in the case of three of twenty-seven mean differences but definite trends can be noted which indicate that relationships may exist but were not significant because of confounding variables present in the investigation. CHAPTER VII RECOMMENDATIONS AND IMPLICATIONS Although the results of the present investigation were not found to be significant, the investigation ful- fills a purpose in providing guidelines for further study. This chapter including recommendations for improvement of the index of complexity of structural design and impli- cations for further research, identifies some of these guidelines. Recommendations for Improyement of the Index of Complexity of Structural Design The reliability and validity of the index of com- plexity of structural design was tested and found to be satisfactory for this investigation. Further testing of reliability and validity, however, would be advantageous. Testing done in the development of the index involved gar- ments from different costume periods. Present day gar- ments, children's and men's garments, as well as clothing from other cultures might be used to establish more firmly the reliability and validity of the index. 59 60 Refinement of the index through further study and breakdown of variables would be desirable. The variables included and all of their component parts should be re- evaluated individually and in combination with all other variables. weighting some scores such as garment fabric and hand stitching more than others should be tested and analyzed. Finer distinctions as to exact items which should or should not be included in the index would possi- bly increase the reliability of the index. A correction factor to eliminate variability due to the body size of the wearer should be developed as it could reduce some of the error variance in the index. Perhaps the greatest potential of the index of com- plexity of structural design is that of providing an ob- jective means of analyzing an aspect of costume design. To date, all methods of analyzing design in clothing are subjective in nature. The development of an equivalent index of decorative design and its use in conjunction with the index of complexity of structural design would provide an objective, quantitative method for analyzing complexity of design in clothing. Implications for Further Research Refinement of the index of complexity of structural design could provide a quantitative measure which could facilitate quantitative comparisons between design in clothing and cultural phenomena as well as providing an 61 objective method of documenting aspects of fashion change and variability in fashion. This portion is devoted to recommendations for further research which are not directly related to the index but which evolved from the present investigation. Definite trends were noted in the data for this study even though significant results were not found. The present study should be repeated to determine whether the findings were due to true lack of significance or to inter- ference of confounding variables. By controlling the intervening variables, more valid and perhaps more sig- nificant results could be obtained. Modification of the procedure of the investigation regarding the time span in costume periods, the definition of a daytime dress, and garments combining characteristics of two or more costume periods should be made. Costume periods, as defined for this investigation, were periods of approximately ten years. Garments worn during a costume period were described in terms of general characteristics. A study with costume periods spanning shorter periods of time and more specific descriptions of garments worn during a period might result in significant findings. Uncontrolled variability within the period would be reduced by these changes which would be likely to reveal any significant differences in scores for com- plexity of structural design of daytime dresses across costume periods. 62 The limited number of dresses available for this study necessitated a broad definition of a daytime dress including both walking dresses and tea or afternoon dresses. Although the scores for the two types were not significantly different across the eight costume periods, sample sizes were too small to detect differences within a costume period. Only ten dresses for each period were used and therefore comparisons within costume periods were based on a total of only ten observations. Significant differ- ences were not found to exist within a costume period but data indicated that a larger sample size might have pro- duced significant differences. A study including only more tailored or walking dresses, only dressy afternoon dresses, or a sample sufficiently large to allow inclusion of type of daytime dress as an independent variable, would eliminate the possibility of type of daytime dress as an intervening variable. Some of the garments used in the investigation had characteristics of two adjacent costume periods. These garments might have been the newest fashion during a costume period, slightly outdated in a later costume period, or actually transitional garments combining some new characteristics with the characteristics of the cos- tume period which was coming to a close. An investigation eliminating garments combining characteristics of two adjacent costume periods would increase independence 63 between groups, reduce within group variance, and result in increased precision. Garments worn during the first costume period in each of Young's silhouette cycles exhibited higher vari- ability and higher mean scores than garments of other costume periods. Perhaps the higher the complexity scores and higher variability during these periods are related to the process of fashion change. Further, more detailed study is needed to test this relationship. By quanti- tatively measuring the variability within a costume period, the index of complexity of structural design might provide a means of pin-pointing and measuring fashion change. The index of complexity of structural design might be used in a study comparing changes in the design of specific parts of garments such as bodices, sleeves, and skirts during a given costume period or across periods. Fashion changes from time to time tend to concentrate on a given area of a garment. Perhaps the area of emphasis of a garment coincides with the area of the garment with the greatest complexity of structural design. Garments worn within a given period may yield similar ratios of complexity of structural design of bodices to complexity of structural design of skirts. The index would provide an objective method of analyzing the changing area of emphasis of a garment. 64 Another research topic worthy of investigation is the variation in the shape of bodices pieces. The amount and type of fitting in the bodices of garments were found to vary in the years from 1860 to 1940. Changes in the shape of garment pieces, particularly the bodice back piece, seemed to parallel changes in fitting. Variations in the shape of bodice pieces are more subtle than vari- ation in other parts of garments and appear to remain more stable through a costume period. Shape of bodice pieces should be explored further as a possible method of dating historic costumes. The literature reveals the commonly accepted notion that women's clothing became simpler in the years from 1870 to 1930. The scores of garments on the index of com- plexity of structural design in this study indicated that complexity of structural design did not change appreciably during those years. The silhouette appears to have become more simple but the complexity of structural design does not. Garments of the 1920's were characterized by simple silhouettes and straight lines but the line and space divisions within the garment were comparable in complexity to those of garments of the 1870's and the 1880's. Another assumption often made without much foundation is that the invention of the sewing machine was responsi- ble for the lavish amounts of trim on dresses worn during the 1870's. Based upon the close examination of garments 65 used in this investigation, this assumption was found to be invalid. Although garments of the 1870's were highly trimmed, the trim was made and sewn to the garment with hand stitching. The machine stitching found in the gar- ments of the 1870's was found only in the basic seamlines. The largest amount of stitching, that to finish and attach the trim, was done by hand. Therefore, the existence of a relationship between the invention of the sewing machine and the trim of women's dresses during the 1870's needs to be tested and evaluated through research. Assumptions such as those above appear to be widely accepted despite the fact that they may be based entirely upon speculation rather than validated research. The possible invalidity of these and other such speculative statements can be detected only through research. However, develOpment of reliable and valid measures of costume must precede such research. The development and testing of the index of complexity of structural design was intended as a small step toward that goal. BIBLIOGRAPHY BIBLIOGRAPHY Books and Articles Bradley, Carolyn G. Western World Costume. New York: Appleton-Century-Crofts, Inc., 1954. Chambers, Bernice G. Color and Design. New York: Prentice-Hall, Inc., 1954. Cunnington, C. Willett, and Cunnington, Phyllis. Handbook 9f English Costume in the Nineteenth Century. London: Faber andFaBer, Ltd., 1966. Goldstein, Harriet, and Goldstein, Vetta. Art in Everyday Life. New York: The Macmillan Company, 1954. Hansen, Harald. Costume and Styles. New York: E. P. Strutton and Co., 1956. Kroeber, A. L. "On the Principle of Order in Civilization as Exemplified by Changes of Fashion." American Anthropologist, XXI (July, 1919), 235-63. Laver, James. WOmen's Dress in the Jazz Age. London: Hamish Hamilton, Ltd., 1964. Mchmsey, Harriet. Art in Clothing Selection. New York: Harper & Row, Publishers, 1963. Morton, Grace Margaret. The Arts of Costume andggersonal A earance. n.p.: University of Nebraska Foundation, 1955. Payne, Blanche. History of Costume. New York: Harper & Row, 1965. Richardson, Jane, and Kroeber, A. L. "Three Centuries of WOmen's Dress Fashions: A Quantitative Analysis." Anthropological Records, V, No. 2 (1940), 111-53. Tate, Mildred, and Glisson, Oris. Family Clothing. New York: John Wiley & Sons, Inc., 1961. 66 67 Wilcox, R. Turner. The Mode in Costume. New York: Charles ScribnerTs Sons, 1958. Young, Agatha Brooks. Recurring Cycles of Fashion. New York: Cooper Square Publishers, Inc., 1966. Other Sources Allan, Barbara. "The Flapper: A Study of the Evolution of an American Fashion Stereotype from 1910 to 1929." Unpublished Master's thesis, Michigan State Uni- versity, 1960. Alton, Leota R. "A Study of WOmen's Fashions in the United States Between 1860 and 1900." Unpublished Master's thesis, State University of Iowa, 1961. Basralian, Karen Marie. "The American WOmen's Day Dress, 1890 Through 1900, as Reflected in the American Fashion Magazines." Unpublished Master's thesis, University of Maryland, 1969. COOper, Muriel Sieving. "The Growth of the Commercial Pattern Industry and Its Contribution to Contemporary Dress." Unpublished Master's thesis, State Uni- versity of Iowa, 1961. DeLong, Marilyn Ruth Revell. "Analysis of Costume Organi- zation: Development of a Model for Considering Style." Unpublished Ph.D. dissertation, Ohio State University, 1967. Hager, Margaret Ruth Wood. "The Costume of Four Selected Cultures, Analyzed on the Basis of Line Factors." Unpublished Master's Thesis, Pennsylvania State University, 1963. Joseph, Marjory Lockwood. "Changes in WOmen's Daytime ' Dress as Related to Other Selected Cultural Factors During the First and Third Decades of the Twentieth Century." Unpublished Ph.D. dissertation, Pennsyl- vania State University, 1962. Lumbra, Elaine. "Everyday WOmen's Dresses Worn in the United States During the Period 1870-1900." Unpub- lished Master's thesis, University of Indiana, 1966. Nugent, Elinor Roth. "The Relationship of Fashion in Women's Dress to Selected Aspects of Social Change from 1850 to 1950." Unpublished Ph.D. dissertation, Louisiana State University, 1962. 68 Strand, Sandra Lou. "The Psychology of Dress and Its Effect Interpreted in Nineteenth Century Costume of Women." Unpublished Master's thesis, University of Washington, 1964. U.S. Department of Commerce. Bureau of the Census. Fifteenth Census of the United States, 1930: Population. Vol. I, II, and III. . Bureau of the Census. Fourteenth Census of the United States, 1920: Population. Vols. I and IV. . Bureau of the Census. Thirteenth Census of the United States, 1910: Population. Vols. I and IV. U.S. Department of Interior. Bureau of the Census. Eighth Census of_the United States, 1860: Popu- lation. Vol. III. . Bureau of the Census. Eleventh Census of the United States, 1890: Population. Vol. I. . Bureau of the Census. Ninth Census of the United States, 1870: Pqpulati5n. Vol. I. . Bureau of the Census. Tenth Census_of the United States, 1880: Population. Vol. I. . Bureauof the Census. Twelfth Census of the United States, 1900: Population. Vol. II. APPENDICES APPENDIX A SKETCHES OF DRESSES IN SAMPLE [Front View] A): A». ' AV VI III 2 3 \ M AA A ~ - 9 APPENDIX B IDENTIFYING NUMBERS AND SOURCES OF GARMENTS IN SAMPLE APPENDIX B IDENTIFYING NUMBERS AND SOURCES OF GARMENTS IN SAMPLE Garment Identifying Costume Number Number Collection 1 6-26-50#5 Human Ecology 2 6-26-50#6 Human Ecology 3 6-26-50#7 Human Ecology 4 6-26-50#10 Human Ecology 5 6-26-50#3 Human Ecology 6 631.3 Museum 7 2569.1 Museum 8 9523cw Museum 9 11593 Museum 10 6-28-50#8 Human Ecology 11 6-28-50#17 Human Ecology 12 9222cw Museum 13 6-27-56#1 Human Ecology 14 6-27-5043 Human Ecology 15 850.91 Museum 16 466.7 Museum 17 850.85 Museum 18 2544.185 Museum 19 1827.9 Museum 20 9086cw Museum 21 7-27-50#3 Human Ecology 22 6-28-50#7 Human Ecology 23 6-28-56#l Human Ecology 24 6-28-50#12 Human Ecology 25 6-28-50#15 Human Ecology 26 6-28-50#ll Human Ecology 27 2580.1 Museum 28 11756 Museum 29 None Museum 30 401.1 Museum 31 6-29-50#39 Human Ecology 79 80 Garment Identifying Costume Number Number Collection 32 6-29-50#37 Human Ecology 33 6-29-50#33 Human Ecology 34 6-29-50#32 Human Ecology 35 6-29-50#30 Human Ecology 36 6-29-50#41 Human Ecology 37 6-29-50#29 Human Ecology 38 6-29-50#43 Human Ecology 39 850.167 Museum 40 3376.2 Museum 41 6-30-50#28 Human Ecology 42 6-30-50#34 Human Ecology 43 6-30-50#20 Human Ecology 44 6-30-50#26 Human Ecology 45 6-30-50#30 Human Ecology 46 6-30-50#43 Human Ecology 47 8544.169 Museum 48 1310.584 Museum 49 2499.4 Museum 50 850.106 Museum 51 6-31-50#26 Human Ecology 52 6-31-50#25 Human Ecology 53 6-31-50#31 Human Ecology 54 850.166 Museum 55 2575.35 Museum 56 1432.14 Museum 57 None Museum 58 1737.8 Museum 59 1737.2 Museum 60 850.61 Museum 61 6-32-50#52 Human Ecology 62 6-32-50#28 Human Ecology 63 6-32-50#43 Human Ecology 64 6-32-50#32 Human Ecology 65 6-32-50#61 Human Ecology 66 6-32-50#36 Human Ecology 67 2239.4 Museum 68 None Museum 69 850.170 Museum 70 2750.62 Museum 71 6-33-50#l6 Human Ecology 72 6-33-50#21 Human Ecology 73 6-33-50#11 Human Ecology 74 6-33-50#l9 Human Ecology 75 6-33-50#33 Human Ecology 76 6-33-50#12 Human Ecology 81 Garment Identifying Costume Number Number Collection 77 1729HM Museum 78 None Museum 79 None Museum 80 850.1 Museum APPENDIX C ILLUSTRATION OF MEASURING TECHNIQUE The figure above illustrates the technique used in dividing garment pieces into geometric shapes to measure the area of the fabric (see pages 38 and 39). 82 T U V E NI . LIB | IIIIHIH III HIIHIHII 163 1 22 nT 1}" Q) (14. :3 MICHIGAN 5 "WWW"? 3129 IT