Antiquities Section Selected Papers Volume VII Number 16

Fremont settlement pattern and architectural variation

FREMONT SETTLEMENT PATTERN AND ARCHITECTURAL VARIATION ERNEST S. LOHSE University of Utah ABSTRACT This study's explicit objective was to test Marwitt's (1970) identification of five Fremont variants. Fifty-three excavated habitation sites were utilized to generate data on settlement pattern and architecture. Architectural traits were coded and subjected to both cluster and discriminant analysis. On the basis of this study, Marwitt's variants are well supported. However, an even stronger support is generated for the iden­tification of only two variants, corresponding to the Great Basin and Colorado Plateau physiographic provinces. INTRODUCTION This paper has three primary objectives: the presen­tation of new information regarding Fremont architec­tural characteristics; an examination of both intra-site and inter-site settlement plan; and a definition of Fremont in strictly architectural terms. Within this rather general framework, previous studies defining internal Fremont variation will be challenged The analysis of architectural variation and settle­ment pattern obviously cannot produce a definitive statement of Fremont identity when considered in isola­tion from other cultural elements. Therefore, this study only aspires to present an investigation that will offer a valuable insight into the Fremont problem. BACKGROUND Other papers in this volume present a detailed de­scription of the problem of Fremont origins and past theoretical approaches. Consequently, I will assume that the reader is familiar with the basis of past research in the Fremont area, and will present only a brief preface. Much recent debate has centered on the question of just what is Fremont. Aikens (1966b) has suggested that Fremont may represent a southward migration of un­specified northern Great Plains groups, who coincident-ly adopted certain distinctive Southwestern traits. Gunnerson (1969) and Berry (1975) felt that Fremont was merely a northward expansion of Pueblo groups. The more popular view, as expressed by Wormington (1955) and Jennings (1956), was that Fremont derived from an earlier Archaic base and had merely absorbed distinctive Southwestern traits. Madsen and Lindsay (1977) have assumed a common sense stance, stating that the best answer lies in a combination of these three viewpoints. Researchers have differed on the delineation of in­ternal variation within the Fremont cultural area as well. Jennings (1956) divided the Fremont culture into two distinct variants: the Sevier in the eastern Great Basin and the Fremont on the Colorado Plateau. Am­bler (1966) identified five variants on the basis of an involved and somewhat ambiguous series of trait anal­yses: the Sevier, Provo, and Conger in the eastern Great Basin, and the Uinta and San Rafael on the Colorado Plateau. Marwitt (1970) again identified five Fremont variants, although he changed the designations and their respective boundaries: the Parowan, Sevier, and Great Salt Lake in the eastern Great Basin, and the Uinta and San Rafael on the Colorado Plateau. Madsen and Lind­say (1977), decrying the lack of a suitable definition for the Fremont, have suggested that it would be more profitable to identify two cultures: the Sevier in the east­ern Great Basin and the Fremont on the Colorado Plateau. They also reserve the possibility of a third cul­ture in the northern portion of both the eastern Great Basin and the Colorado Plateau, which exhibits a pecul­iar Great Plains character. The Fremont problem or controversy, depending 41 upon your own predilection, has seemingly gone full circle. Earlier researchers identified only gross differ­ences. Later, Ambler and Marwitt felt that finer lines could be drawn within the Fremont cultural area. And recently, Madsen and Lindsay, recognizing the rather evident lack of a definitive statement concerning the Fremont, have suggested that we again concern our­selves with only gross differences, i.e. Sevier versus Fremont. However, they prefer the term "cultures," injecting the idea that Sevier and Fremont owe their distinctiveness to different origins and influences. In their view, these two groups, previously identified only as variants of the same culture, should be viewed as two separate cultures, more different than alike. 1 accept the notion that Fremont is a recognizable cultural en­tity corresponding to the subarea, as defined by Willey and Phillips (1958): ". . . territories of geographical ex­tent intermediate between the region and the area which possess qualities and degrees of cultural unity that give them a definite usefulness in archaeological and ethno­graphical studies." Whether or not we can succinctly define Fremont does not seem to me of paramount importance; it is only necessary that we are able to recognize Fremont as different from other surrounding cultural groups. It is the usefulness of the designation in archeological contexts, and not its validity in terms of past cultural identities, which should be of concern. Fremont is recognizable; it does exist as an entity in the archeological record, distinct from the Anasazi to the south and from hunting and gathering groups to the north, east and west. 1 feel that to break Fremont into two "cultures," more different than alike with re­gard to the Southwestern cultural area, causes even more ambiguity. If it is recognizable within our archeological record, and it seems to serve as a useful pigeonhole or tool, then we should continue to make use of it. PROBLEM It was felt that a consideration of both architectural traits and settlement pattern offered a means of challeng­ing prior viewpoints. For instance, the analysis of archi­tectural traits and settlement plan should be very pro­ductive in establishing the validity of Gunnerson's (1969) and Berry's (1975) hypotheses of Fremont ori­gins. If the Fremont do indeed represent a northward migration of Pueblo groups, it is only rational to expect that this expansion should be adequately documented in the archeological record. If an actual physical move­ment of people occurred, involving representatives of a particular stage of Pueblo development, then one would expect their material culture to be virtually identical in both areas. Further, an in-depth considera­tion of available architectural and settlement data should be equally productive in dealing with the lack of a suit­able definition for Fremont. Analysis will only focus on the validity of the five variants postulated by Marwitt (1970) and the two variants suggested by Jennings (1956). 1 cite Jennings as opposed to Madsen and Lindsay (1977) because of their use of the term "cultures." The first part of this paper will involve a discriminant analysis of arch­itectural traits in an effort to quantitatively support or refute Marwitt and Jennings. The second part deals with the limited information available pertaining to settlement pattern, in a similar effort. FREMONT ARCHITECTURAL VARIATION METHOD I chose as a sample 53 excavated Fremont habita-. tion sites. (See Table V). The only criteria involved availability of information and marginally reliable ex­cavation records. Only habitation structures were con­sidered in depth, with the presence of storage structures =H?erely being noted. Presumably, dwellings would be more reliable indicators of cultural influence/identity, due to their greater elaboration and complexity. A habitation structure was defined solely on the presence of an interior hearth. A dwelling was inden-tified as subterranean if the living floor had been pur­posefully excavated below the then-occupation surface, forming a recognizable wall, and if the roof leaners had been footed outside of the structure margin. Each site was recorded on a separate form involving 22 different categories concerning 21 different discriminant variables. Categories included whether the individual structure was subterranean or surface; its postulated use (habita- 42 tion, storage); the construction technique (dry- or wet-laid masonry, coiled or coursed adobe, jacal); treatment of floor and walls (masonry, puddled clay, natural); roof construction (rectangular, peripheral, irregular); structural plan (circular, rectilinear, irregular); approx­imate angle of the subsurface walls (vertical, outward or inward sloping); the living floor area in square meters (0-19.999 being small, 20.000-39.999 being me­dium, +40.000 being large); the presence of a ventilator shaft with regard to the cardinal directions; ventilator shaft construction (floor and wall treatment, presence of masonry, adobe, jacal, earth); fireplace construction (clay-rimmed, clay-lined, paved, natural, any combina­tion); storage cist construction (bell-shaped, slab-lined, irregular); number and kind of depressions discerned on the living floor; any miscellaneous structural features (entry ramps, ash pits, etc.); postulated cultural affilia­tion; and date off the living floor, if available. See Tables VI-V1I1 for frequencies of selected architectural traits. TABLE V Fremont Habitation Sites Used . . . Colbradf^fi Alice Hunt Site (JenrungV&d/b) " ' Basket Hut Site (Jennings n.d,6) Boundary Village (Leach l96i$ Bu rn't; Ho use {Bretejrn it?. J 970a) - t a Id well Village^ Ambler 1966)!; Chas, B. Hunt Site (Jennings n;<j Crescent R idge (M adserr 1975b) 7, Cub Creek Site (Breternitz;j970af Dam Site(BrclerntU 1970a) .'-':,• Emery Site (Gunnerson 1957)-:';: -;->v Fa I len Worn a n (W i lso n a nd 5m if h WM Fetter's Hill (Shields 1967) Flat Top Butie (Shields 1967) ; Ford Site (Breternitz 1970a) h : Fremont Playhouse (Brctecnitz J970 G i Ibert Site {S hie Id s 1967) ::[::•; "•:• '• <••£$%$; Gnat Haven (Jennings n.d.b) r S Goodrich Site (Shields 1967) " U"-C'-.'.-. Innocents Ridge (Schroedf arid Hog: I vie Ridge (Wilson and Smith 1976} Macleod Site (Bretemitt 1970ar Ninas Hill (Jennings n;d,b);:: Nine Mile Canyon (GiliinJ938) North Point (Jennings n.d.b) ; .'' Old Woman (Taylor 1957) Playa Site (Jennings rul.bj • Poplar Knob (Wilson and S m U h i " ? ? ^^ Power Pole Knoll (Madsen 1975 Round Springs (unpublished surveVreDoriV * SnakeRock Village (Aikens 1967)%- Turner-Look Site {Wormington i95S Wagon Run (Breternitz 1970a) TWhiterocks Village ( S h i e l d^ Wholeplace Village (Brcternit| Windy Ridge (Madsen 197i5)C &y-z1970a) iieldsv1967) rctemitz-197ft. j . ro75i;:" ••,-%- Great Ba^g? * Back hoe Village (Madsenand Lihds^ Bradshaw Mound'(Judd J9 J 7_b) '•" UVX .. Ephraim(GilUn 1941) •'---..• ' . • • ' " ., '' Evans M ound (Berry 1972) : '/'-• Harrison (Taylor 1954) ::. 3 ' : . . George Mound (Judd .1917b5:: Grantsville (Steward 1933b; 1 njun Creek (Aikens i966b).'\\ 7£f§&p£ Kanosh (Steward 1931,1933b) Knoll Site(Dalley n.d,) - . -;' :. Levee S ite (Da 1 ley n.d.) **Marysvale.'(GilItn 194J) M ed ian V i llage (M anyit t 1970) -Nephi Mounds (Sharrock arid Marwitt 1 •:.; Paragonah (Judd 1917a, h; m9; 1926) : Pharo Village (Marwitt 1968) ''^'•'<.:•• Tooele.(Gillin 1941) •wmard(Judd 1917b, 1926) • Unfortunately, excavation techniques and subse­quent reporting have varied greatly within the Fremont area^ Site reports from over twenty years ago often Jailed to supply even the minimum information neces­sary for area-wide comparisons. Further, the dating of individual sites, let alone of individual structures has been sadly neglected. Only recently, with the excava­tions of Backhoe Village (Madsen and Lindsay 1977) and the Bull Creek sites (Jennings n.d.b), has there been any concern with the dating of individual living floors. The lack of good excavation controls and absolute dates forced me to deal with the data on architecture in a synchronic manner. Discriminant analysis was per- 43 lormed on the data as if all sites, structures, clc, were contemporary. Nevertheless, the lack of chronology is probably not much of a problem: a conservative esti­mate for the entire time span of an identifiable Frcmonl entity is perhaps only about 200-400 years; most sites are single component deposits, not enabling us to recog­nize any change through lime. Further, one may con­ceivably interpret the results of the analysis with the untreated lime factor in mind, DISCRIMINANT ANALYSIS Once the data was recorded for the 53 selected sites it was coded and transferred to computer data cards' Discriminant analysis was performed, utilizing the for­mat presented in the Statistical Package for the Social Sciences (Nie et al. 1975). In basic terms, discriminant analysis may be thought of as a method of producing hybrid variables so as to produce the best possible sepa­ration or discrimination between two or more groups It begins with the desire to statistically distinguish be­tween two or more cases. In order to execute this, the researcher selects certain discriminating variables. If as in Jennings' (1956) delineation of two Fremont vari­ants, the observations are divided into two groups to iorm the dependent variable, only a single discriminant function can be derived. The stepwise option in the S.P.S.S. subprogram DISCRIMINANT was used This method begins by selecting the single best discriminating variable, and then involving other variables in the order that they are best able to improve the discrimination criterion in combination with the first variable. At each step in the routine, variables already selected will be removed if they are found to reduce discrimination when combined with prior variables. A measure termed Wilks' lambda was used in adding the variables in the order that they made the greatest contribution to reducing the ratio of within-group to total variation Wilks1 lambda is used to determine the significance of the canonical correlation or canonical root (the canonical correlation being a method for correlation analysis when both the independent and dependent variables are com­posites of several measured variables). The reciprocal of Wilks' lambda is the usual correlation coefficient and its significance was tested using the chi-square distribution. The individual observations, e.g. rectilinear house shape, are given scores on the discriminant funcrions These scores are derived in standardized form and a mean score is computed for each group of observations This measure 1$ termed a group centroid. Analyses of variance can then be performed on these scores. The specific aim of discriminant analysis is to maximize the F ratio of the between-group to the within-group vari­ance estimates: the discriminant function is located to produce the "best" analysis of variance. Multivariate methods like discriminant analysis were developed to test hypotheses, in this case relating Site Boundary Village Burnt House Caldwell Village Cub Creek DamSite^ FcUer's-Hill Flat Top Butte Fremont Playhouse GilbertSite . Goodrich Site MacteodSite WagqnRun: Whitefocks Village Wholepiace Village Alice Hunt Site Basket Hut Site Chas.B^HuntSite Crescent Ridge Fallen Woman GnatHaven Innocents Ridge IvieRidge NinasHiii: :•• Nine-Mile Canyon NorthPoint •."': Old Woman Playa She Poplar Knob Power Pple Knoll Round Springs Snake Rock Village Turner-Look Site Windy Ridge' Backhoe Village. Evans- Mound Garrison Kanosh Marysvalc: Median.Village Paragonah Ephraim Grantsville Nephi Mounds PharoVUlage •Tooele Bradshaw Mound .Injun Creek Knoll Site Levee Site Willard Site TABLE VI House Shape and Dimensions Circular, Irrtgular- SUBTERRANEAN Rectilinear Average Floor Area in sq. m. 28.6 23.8 31.6 0.0 0.0 28,8 40.3 3.4 0.0 18.8 40.1 0.0 0.0 21.3 23.9 8.7 28.4. 0.0 22.7 28.4 33.4 22.8 11.4 32.4 28:4 22.7 28.4 0.0 0.0 10.5 17.9 0.0 0.0 17.8 16.3 21:1 13.4 8.9 20.6 19.5 11.2 20.2 33.3 29.7 0.0 0.0 0.0 26.8 38.9* 0.0 Circular, irregular SURFACE Rectilinear *At the Levee Site, during the ea square meters. During the late square meters. rlycomponent(prior to ca A.D. 1000), the average floor area was 11.1 component (ca A.D. 1090 to 1240). this average increased to 38.93 Average Floor Area in sq. m. 0.0 36.5 0.0 0.0 0.0 0,0 o!o 0.0 212 23.8 16,0 23.8 00 0.0 0.0 0.0 0.0 15.4 0.0 0.0 16.5 0.0: 0.0 12.5 27.0 20.9 0.0 23.1 0.0 0.0 16,2 25.3 15.6 0.0 0.0 27:0 22.3 0,0 .0.0 0.0 14.5 0.0 20.0 0.0 9.2 20.9 25.3 0.0 ii.r o.o measurements on nominal scales to others on collinear interval and ratio scales. Discriminant analysis is there­fore seen to have three primary uses, following John-ston( 1978): 1. Testing and generating hypotheses. The method is used to investigate whether a certain set of selected variables successfully discriminates between groups of observations on a nominal scale. 2. Evaluating a prior classification. Discriminant analysis will point out the number of misclassifications in the dependent variable. These are observations whose 44 TABLE VII Fireplace Treatment Old Woman Playa Site Popular Knob Power Pole Knoll Snake Rock Village Turner-Look WindyRidge Backhoe Village Eyans Mound Garrison Kanosh Marys vale Median Village Paragonah Ephraim Grantsville Nephi Mounds Pharo Village Tooele 0 4 5 12 2 0 1 1 o - 2 3 0 2 $ : • • 5- 0 0 2 0 1 2 4 2 0 0 I 0 0 0 0 0 9 . 0 0 0 0 0 0 0 0 0 0 0 0 o 0 0 0 • 0 " . o 0 0 0 0 0 0 0 0 0 5 I •• 3 • : : . l ' 12 . 8 • 3 ;" 9 " M 5 4 8 13 .2." 3 5 JO 3 Bradshaw Mound Injun Creek Knoll Site Levee Site Willard Site TABLE VIII Miscellaneous Structural Features Site Boundary Village BurntHouse Caldwell Village Cub Creek DamSite Fetter's Hill Flat Top Butte Fremont Playhouse Gilbert Site Goodrich Site Macleod Site Wagon Run Whiterocks Village Wholeplace Village Alice Hum Site Basket HufSke Chas. B.HuntSite Crescent Ridge Fallen Woman Gnat Haven Innocents Ridge Ivie Ridge Ninas Hill Nine-Mile Canyon North Point Old Woman , Playa Site Poplar Knob Power Po|e K no! Round Springs Snake Rock Village Turner-Look ' Windy Ridge Backhoe Village Evans Mound ' Garrison Kanosh Marysvale Median Village Paragonah Ephraim Grantsville Nephi Mounds Pharo Village Tooele Bradshaw Mound Injun Creek Knoll Site Levee Site Willard JI •2 0 0 0 9 0 0 0 4 0 0 0 0 0 12 8 "3 9 17 5 4 8 13 2 : 3 5 10 3 1 45 TABLE IX Comparison of Fremont Architectural Traits Within Marwitt's (1970) Variants Architectural Trait Uinta .San Rafael Parowan, .Sevier ; Great'..".'. SaUL-ake Total Subterranean structures Village habitation structures (versus storage) Puddled living floors.:. Rectangular roof support ,7067 .9867 .1867 .3333 .6562 .9562 .2437 .2687 .6000 .7250 .3125 .1875 .3625 .7250 .4375 .3125 .5000 .8333 .3667 J 833 .6000 .8811 .2811 .2717 Peripheral root support Circular subterranean habitation structures Rectilinear subterranean habitation structures Plastered walls .0467 .6333 .2000 ;0000 .0625 .8687 .1312 .1875- .0500 .3375 .5625 .2250 .0000 .2125: .7875 .5250 .6333 .5833 .4167 .0000 1113 .5906 .3472 .1698 Ventilators Stab deflectors Jacal deflectors Masonry deflectors .0000 .0000 .0000 .0000 .0625 .0000 .0625 .0000 .3375 .0000 .6250 .0000 .2375: .1250; 0000 .0000 .2500 .0000 -Peflectors of naturalearth Clay-rimmed firepits, Clay-rimmed, paved firepits Ciay-iined firepits .0000 .5067 .0000 .0200 .oooo .6000 .4375 .1250 .0000 ,6625 ,0500 .1875 .0000 .6125 .4625 ,2000 .0000 :1667 1340 .0189 '1132 .1)000 .0000 .0000 .0189; .552$\ .2094 .1019 Unelaborated firepits Unrimmed, Paved firepits Bell-shaped cists . Slab-lined cists Above-floor slab bins Drylaid surface structures Wetlaid surface structures /Coiled adobe surface structui Coursed adobe surface structures. Jacal surface structures Entry' ramps Ashpits'; .6333 , .0200 .1533 .0067: , -.1333 .1667 •' .0333 .0333 . .1000 " .0667 . .1333 ..2667 .3125 .1625 ; .0437 :;0187 ;';•••:. .0625. I-- 4125 .2875 oooo .0125 0000;. .1250, :• .1250 : ;- 7:5750-- - '.0000 . :025o v.; ,0000 .-'•'•; .0000 .1250 7 .0375" - .0250 7; - .5875 'V- -7":;0125 :- .0000 - ..1250 :- - ;. ;3250 '•'.. .1250 ' - .1000 7: ...0375 .0000 : .4000 ,Hv .5000 .2500 .2000 . A' :,..:.- .0250 .0000. ; -, .00.00: 7 ; 5 0 0 0 .0000 -• J667 •.oooo .. : .00005 . .0000 + '• '. 3333:.-.'! ' - - .0000 • .0000 ^V.OOOO' ;. "\ .4666;; .0736 .0755 .0321 :0755 :2S09 T774 .2509 T8877 :0245 .0755. .1321 scores on the discriminant function suggest that, accord­ing to the relevant characteristics identified, they are more like the members of another group than they are like those of the group to which they were assigned. 3. Estimating values for other observations. If dis­criminant analysis is successful in isolating differences between groups, it can then be used to predict values on the dependent variable for observations not in the origi­nal sample. OUTCOME The use of discriminant analysis upheld both the concept of five Fremont variants advocated by Marwitt (1970), and the delineation of only two Fremont vari­ants suggested by Jennings (1956). The percentage of correctly classified cases was high in both. For Marwitt's five Fremont variants, 88.6% of the cases, i.e. sites, were correctly classified: Uinta (15 sites), 93.3% predicted, 6.7% assigned to both the Paro­wan and Great Salt Lake; Parowan (8 sites), 87.5% predicted, 12.5% assigned to the Uinta; Sevier (8 sites), 87.5% predicted, 12.5% assigned to the Uinta; Great Salt Lake (6 sites), 83.3% predicted, 16.7% assigned to the Uinta. (mean frequencies) The best discriminating variables are arranged below in order of decreasing significance (chi-square values in parentheses): jacal deflectors (.000); peripheral roof supports (.000); plastered walls (.000); rectilinear habitation structures (.000); earthen deflectors (.000); clay-rimmed, paved firepits (.003); clay-lined firepits (.000); coiled adobe surface structures (.000); entry ramps (.000); bell-shaped cists (.000); ash pits (.000); unelaborated firepits (.000); clay-rimmed firepits (.000); circular habitation structures (.004). Table IX presents relative frequencies of architectural traits within the five Fremont variants. Misclassified sites included the following: the Gilbert Site, more probably Sevier than Uinta; North Point, more probably Parowan than San Rafael; the Playa Site, more probably Great Salt Lake than San Rafael; the Garrison Site, more probably Uinta than Parowan; Grantsville, more probably Uinta than Sevier; and the Willard Site, more probably Uinta than Great Salt Lake. Table X presents cases identified as misclassified, with the probabilities of their belonging to another group. For Jennings' two Fremont variants, 94.3% of the 46 TABLE X Misclassified Sites Site Gilbert Site North/'Point... Playa Site 7 Garrison Site Grantsville Willard Site Actual Group Uinta San Rafael San Rafael Sevier Sevier Great Salt Lake . ! Highest Probability* Group P(X/G) P(G/X) •Sevier;; ;':- Pa rowan Great Salt Lake Uinta ' "Uinta Uirita . .996 .973 .991 .999 .989 .998 1569 ,4887 ,931 960 .596 ,712 *P(X/G) = the P(G/X) = thc probability ol a case in group G being that far from the centroid. probability ol a case being in group G. Second Highest Group P(G/X) San Rafael San Rafael San Rafael Sevier San Rafael San Rafael .291 .369 .062 .036 .226 .283 cases were correctly classified: Fremont (31 sites), 90.6% predicted; Sevier (22 sites), 95% predicted. The best discriminating variables are arranged below in order of decreasing significance (chi-square values in parentheses): rectilinear habitation structures (.000); ventilators (.001); coursed adobe surface struc­tures (.002); peripheral roof supports (.032); slab deflectors (.023); bell-shaped cists (.027); entry ramps (.025); jacal surface structures (.054); earthen deflec­tors (.074); puddled living floors (.025). Table XI pre­sents the relative frequencies of architectural traits within the Fremont and Sevier variants. Misclassified sites included the following: Fremont Playhouse, more probably Sevier than Fremont; the Gilbert Site, more probably Sevier than Fremont; the Willard Site, more probably Fremont than Sevier. Table Xll presents cases identified as misclassified, with the probabilities of their belonging to another group. SUMMARY The delineation of five Fremont variants, i.e. Uinta, San Rafael, Parowan, Sevier, and Great Salt Lake (Marwitt 1970), appears to be statistically valid. As : TABLE XI Comparison of Fremont Architectural Traits Within the Colorado Plateau and the Eastern Great Basin Architectural Trait Colorado Plateau Subterranean structures • Village habitation structures (versus storage), Puddled living floors ; Rectangular-roof support Great Basin .6806 :9710 .2161 3000 .4864. .7545 .3727; .2318'' Peripheral roof support Circular subterranean habitation structures Rectilinear subterranean habitation structures Plastered walls .0548 .7548 .1645 .0968 .1909 .3591 .6045 ,2727 Ventilators Slab deflectors Jacal deflectors Masonry deflectors ,0323 .0000 .0323 .0000 .2773 .0455 .2273 .0000 Deflectors of naturaiearth Clay-rimmed firepits Clay-rimmed, paved firepits Clay-lined firepits .0000 .5548 .2258 .0742 .0455 .5500 .1864 .1409 Unelaborated firepits Unrimmed, paved firepits Bell-shaped cists Slab-lined cists .4677 .0935 .0968 .0129 .4636 .0455 ,0455 .0591 Above-floor slab bins Dry-laid surface sturctures Wet-laid surface structures Coiled adobe surface structures .0968 .2935 .1645 .0161 .0455 .1909 .1955 .1000 Coursed adobe surface structures Jacal surface structures Entry ramps Ash pits .0548 .0323 .1290 .1935 .3773 .0336 .0000 .0455 Total .6000 .8811 .2811 271.7 .1113 .5906 •3472 .1698 .1340 .0189 .1132 .0000 .0189 .5528 .2094 .1019 .0736 .1736 .0755 .0321 .0755 .2509 .1774 .0509 .1887 .0245 .0755 .1321 47 TABLE XII Misclassified Sites Site Fremont Playhouse Gilbert Site Willard Site Actual Group Fremont Fremont Sevier Highest Probability* Group Sevier Sevier Fremont PIX/G) 1.000 1.000 1.000 P(G/X) .656 .'; .752 .971 Second Highest Group Fremont Fremont Sevier 7 P(G/X) .344 .248 .029 *P(X/G) = the probability ol a case in group G being that far fromccntroid P(G/X) = the probability of a case being in group G may be seen in Fig. 4, the Uinta and San Rafael vari­ants are the least separated, with their respective group centroids quite close together. The Parowan and Sevier variants are distinct from either the Uinta and San Rafael, but certainly not as alike as those two variants. The Great Salt Lake variant is the most identifiable, not closely allied with any of the other four, though more similar to the Parowan and Sevier than the Uinta and San Rafael. This lends credence, but not firm support, to the view advanced by Madsen and Lindsay (1977) that a discernible Great Plains charac­ter exists in the northern portion of the Fremont area. A correctly predicted total of 88.6% is a dependable indicator for Fremont internal variation, but it can only be considered suggestive. The six misclassified sites are probably attributable to a lack of refinement in the discriminant variables and/or may be evidence of the overall cultural homogeneity of the Fremont. It most certainly would be unwise to extrapolate from architectural data to tangible cultural differences. The number of misclassified sites is greatly reduced when the cases are divided into two groups roughly corresponding to the Colorado Plateau and Great Basin physiographic regions (see Fig. 5). These two groups, termed Fremont and Sevier respectively, by Jennings (1956) and Madsen and Lindsay (1977), are viable units. Marwitt's (1970) five Fremont variants are valid, and further, fit nicely into these two macro-variants or subdivisions. However, whether these variants are reflective of actual cultural differences or are merely products of the synchronic treatment, attributable to outside cultural influence, is still a matter for debate. The use of the terms Fremont and Sevier for sub­divisions of a cultural entity already termed Fremont, seems needlessly confusing, unless like Madsen and Lindsay (1977) we are willing to identify two distinct cultures. I propose that it would be more profitable to retain Fremont as the cultural name and drop the terms Fremont and Sevier when referring to the two subdivisions, in favor of the use of Colorado Plateau and Great Basin, as these seem more indicative of the physiographic-environmental factors contributing to the over all distinctiveness and character of Fremont culture. 5 55 5 3 , 2^l 1 1 121 44 4 4 • _-4.000 -2.875 -.625 1.625 _-1.250 _-2.125 -3.000 2.250 Fig. 4. Discriminant Analysis of Five Variants: indi- Fig. 5. Discriminant Analysis of Two Variants: indi­cates a group centroid; 1 Uinta; 2 San Rafael; cates a group centroid; 1 Plateau; 2 Basin. 3 Parowan; 4 Sevier; 5 Great Salt Lake. 48 FREMONT SETTLEMENT INTRODUCTION As yet there have been no serious studies concern­ing prehistoric economy and settlement pattern within the Fremont cultural area. The meager information that has been available is not sufficient for any in-depth treatment. Further, no in-field investigation involving either subsistence or settlement pattern has even been attempted. Most of the excavated sites have been the product of salvage work with narrow research goals or the result of field schools with the primary objective centering on chronology. Sites have been considered as cultural phenomena isolated from the exterior world. Structures are identified within test trenches and completely excavated, but exploration is not continued out onto the associated occupation surface, effectively curtailing any solid statements about intra-site plan. Unfortunately, inter-site rela­tionships have not been considered with any more regard. Substantial biases exist in prior investigations of Fremont culture. First, sites have been treated as if they were developed in isolation from the surrounding cultural system. Second, a self-fulfilling prophecy has been established, wherein we know what a Fremont site is, so that is what we dig. Excavation has, with few exceptions, involved only "village sites." Other sites, i.e. open campsites, rockshelters, etc., certainly ele­ments of the Fremont cultural system, have not been touched upon. The result is our present, distorted view of Fremont. We may assume that Fremont is a recog­nizable cultural entity on an economic scale somewhere between surrounding Anasazi horticulturalists and Great Basin and Great Plains hunters and gatherers, but we don't know if dependence on horticulture equalled that of the Anasazi or whether prior hunting and gathering strategies were maintained with domes­ticated crops merely representing some sort of backup or novelty. Further, we have no real idea of Fremont settlement plan or correspondingly, social organiza­tion. A good relative chronology has been built up, although the largest stumbling block in any study focusing on settlement pattern is the general lack of radiocarbon dates off identifiable living surfaces. Therefore, any consideration of Fremont settlement pattern, like architecture, is lacking the element of time. Even more regrettably, it must be largely spec­ulative. SITE LOCATION Both Gunnerson (1969) and Berry (1974) have pointed out that Fremont farming settlements are characteristically located near perennial streams. Rela­tively little land throughout the Fremont area would have received adequate groundwater for farming, sug­gesting that prehistoric horticulturalists must either have practiced irrigation or in some way controlled natural runoff. Interestingly, no solid evidence for Fremont irriga­tion has yet been reported. There are however, a num-her of unsubstantiated European reports describing presumed prehistoric irrigation canals. Morss (1931) said that he was told of and shown a portion of an v aboriginal irrigation ditch along Pleasant Creek, south of Fruita, Utah. Reagan (1931) also described prehistoric . ditches along Brush Creek, northeast of Vernal, Utah. Gunnerson (1957) was told that early settlers frequently J - V\l discovered old irrigation ditches along Ferron Creek, west of Ferron, Utah. This author, in conversations with a landowner along Bull Creek, south of Hanksville, Utah, was told that aboriginal fields and ditches were observed when the first Europeans arrived. Whether or not irrigation was actually being prac­ticed, it does appear that Fremont farmers preferred to live in an area with available natural runoff. For example, Berry (1974) notes that every sedentary village within the Parowan Valley was built on or near the alluvial fan of a perennial stream. He argues that prox­imity to water was not the primary aboriginal concern, since surveyed areas in the vicinity often support numerous springs, yet only scattered campsites are found in direct association. He believes that village sites occur only when the conditions for floodwater farming are met. Madsen and Lindsay (1977) have stated that there has been an overemphasis on the importance of hor­ticulture with regard to Fremont subsistence. They argue that marsh resources were the important factor in the economy, basing their hypothesis on the recovery of large amounts of Typha pollen from Backhoe Vil­lage in the Sevier variant (Madsen and Lindsay 1977). They did not excavate any storage structures, and speculate that Typha was the dominant staple merely because it is plentiful and, evidently, reasonably nutri-tous. Their argument seems in contradiction to evi­dence from other village sites, where remains of domes­ticated crops were recovered. Nielson (1978) supports their position however, when he points out that Sevier sites are associated with soils developed under wetland and marsh environments. Though this is not conclusive proof of the importance of marsh exploitation, it is reasonable, and Nielson concludes that Fremont village site selection was a rational cultural decision centering on the exploitation of marsh resources. Great Basin village sites tend to be larger than their counterparts on the Colorado Plateau. Within the Parowan and Sevier variants, presumably sedentary sites are located along the western base of the Wasatch Mountains atop alluvial fans or on the valley floors adjacent to perennial streams. The largest sites, Evans Mound, Median Village, and Paragonah, effectively 49 distinguish the Sevier and Parowan variants from the Uinta and San Rafael on the Colorado Plateau. How­ever, the Great Salt Lake variant, while still within the Great Basin, does not exhibit large village sites. Small settlements like the Levee and Knoll sites (Dal-ley n.d.), are the norm. These are located on slight rises adjacent to the Bear River in areas of what were aboriginally, extensive marshlands. The difference in site size and architecture, though possibly representing cultural differences, may also be due to our sampling bias. Areas of extensive historic land development and farming lie at the northern foot of the Wasatch Front in areas which would have been quite attractive to Fre­mont farmers. If large sites once existed in these areas, similar to those to the south, the evidence is probably no longer available. This problem is not confined to the Great Salt Lake variant, it is apparent throughout the Fremont area, with Europeans having settled in pre­cisely those same optimal locations favored by prehis­toric farmers. Sites on the Colorado Plateau, are in general, much smaller than those excavated in the Great Basin. Jen­nings (1978a) refers to these as "rancherias," pointing out that they usually comprise one or two dwellings and associated storage structures. The sites are commonly situated on slight rises overlooking arable land and perennial sources of water. This pattern is repeated throughout the Uinta and San Rafael variants. This east-west, Colorado Plateau-Great Basin divi­sion seems to be primarily a physiographic one. Large settlements are more common along the western foot of the Wasatch Front at those points where alluvial fans are formed by rivers originating in the mountains, dis­charging out into the valleys below. However, large village sites are also found on the Colorado Plateau within the Uinta variant, e.g. Caldwell Village and Whiterocks Village. Most probably, Fremont peoples settled in larger numbers wherever optimal locations consisting of arable lands and perennial sources of water existed. The question of location and settlement compo­sition docs not appear to involve cultural differences as much as it involves limitations imposed by two very different environments, presented by the Colorado Pla­teau and Great Basin physiographic regions. SUMMARY The Fremont, as is the case with most known ethno­graphic groups of similar economic orientation, tended to settle in larger numbers based on the availability of resources. If horticulture was an integral part of the traditional hunting and gathering economy as has been suggested, then it is only reasonable to assume that the Fremont would have been especially sensitive to the availability of arable land and dependable water sources. If marsh resources were also a concern, so much the better; the two are not mutually exclusive, and the mat­ter still is one of available resources required to support a given population. This discussion of Fremont settlement has been, un­fortunately, somewhat vague and inconclusive. Very little may be said regarding infra structure, as the majority of excavation reports do not even treat intra-site plan, let alone inter-site relationships. Nowhere in the Fremont cultural area has a study pertaining to social organization, economy, etc., been attempted. This lack is regrettable, and should be resolved in the future if the study of Fremont is to advance beyond the con­struction of typologies, chronology, and speculations regarding origins and definitions. Fremont may be said to exist, what is needed now is an understanding of the cultural system involved. DISTINCTIVENESS OF FREMONT Recent studies by Ambler (1966) and Marwitt (1970) have attempted to define Fremont on the basis of trait analysis. The result has been the delineation of variants within the overall Fremont cultural area, with different boundaries in each study. Neither researcher was able to offer an overall definition other than to allude to the idea that the entity may only be identified by reference to total constellations of traits. Madsen and Lindsay (1977), prompted by this lack of a suitable definition, have concluded that Fremont as such must not exist. However, as has been demon­strated, Fremont may readily be distinguished on the basis of architectural traits and settlement plan. Bullard (1962) set up the following traits as diagnos­tic of the Anasazi: the standard arrangement of pit-houses and surface structures in recognizable habita­tion units; the segmentation of larger villages into sepa­rate habitation units; the ventilator shafts oriented be­tween south and east; antechambers; typical L-shaped Anasazi ventilator shafts; an interior bench and/or sup­port for roof side poles; room partitions; clay-rimmed firepits; deflectors; ash pits; sipapus; heating pits con­sisting of basin-shaped floor holes filled with burned rocks or sand; above-floor storage bins; and the use of sandstone slab in wall construction. If, as proposed by Gunnerson (1969) and Berry (1975), the Fremont are to be considered the direct result of a northward burst of Anasazi settlers, then we should expect a high frequency of the above traits in any given site. In reality, the following diagnostic Ana­sazi traits are absent from the Fremont cultural area: recognizable habitation units; consistent ventilator shaft orientations between south and east; L-shaped ventilator shafts; well-defined antechambers; interior benches supporting roof leaners; room partitions; recog­nizable sipapus; and heating pits. Clay-rimmed firepits and deflectors are relatively common throughout the Fremont area. The use of slabs 50 in wall construction is more confined, occurring only in the San Rafael variant on the Colorado Plateau. Examples are found in Nine-Mile Canyon (Gillin 1938), Snake Rock Village (Aikens 1967), Windy Ridge (Madsen 1975b), and Ninas Hill (Jennings n.d.b.). Ash pits, as such, are also largely restricted to the Colorado Plateau, although in both the Uinta and San Rafael variants: Boundary Village (Leach 1967); Dam Site (Breternitz 1970b); Macleod Site(Breternitz 1970b); Inno­cents Ridge (Schroedl and Hogan 1975); Ninas Hill (Jennings n.d.b.); and Windy Ridge (Madsen 1975b). Only one example has been recorded in the eastern Great Basin, at Bradshaw Mound (Judd 1917b). Only four above-floor storage bins have ever been reported: three on the Colorado Plateau, one each at Caldwell Village (Ambler 1966), the Macleod Site (Breternitz 1970), and Nine-Mile Canyon (Gillin 1938); and one in the Great Basin, at Bradshaw Mound (Judd 1917b). See Table XIII. Although certain Anasazi traits are present within the Fremont area, they most assuredly do not represent the high frequency expected if the Fremont merely re­present the northward migration of Anasazi settlers. With this in mind, the Fremont may be considered a distinct entity. Perhaps horticultural practices and associated Southwestern traits were introduced into the Fremont area, not by an actual movement of Pueblo peoples, but by a diffusion of ideas. Whatever the actual nature of the selective process involved in Archaic groups accepting horticultural practices, the culture to the north of the Anasazi remained distinct. Information is in large part lacking for any extensive comparison of Fremont to hunting and gathering groups to the north, east, and west. Fremont is recog­nizable, however, with reference to its mixed economy of hunting and gathering and farming, with the atten-dent use of pithouses and associated surface structures, involving typical Southwestern traits. Aikens (1966b) has presented an interesting argument for Fremont origins and subsequent disappearance. He feels that an expansion south out of the northern Great Plains occurred ca A.D. 500. These people gradually assimilated certain Southwestern traits, adopting hor­ticultural practices although still maintaining a depen­dence on hunting and gathering, and thus became the archeological entity termed "Fremont." He further pos­tulates that this culture was forced out of the Utah area and drifted eastward onto the central Great Plains under pressure from the Shoshonean expansion out of the Great Basin ca A.D. 1400. (Lamb 1958) Aikens draws evidence for his argument from the presence of Great Plains-like traits in the Fremont cultural area, including distinctive shield pictographs; small, trian­gular corner-notched and side-notched concave-based projectile points; unilaterally barbed bone harpoon points; bone fish-hooks; bone and antler wedges or chisels; bone whistles; use of the moccasin; "tipi rings"; pottery with punctation, incising, and applique; Pro­montory paddle-and-anvil pottery; and, finally, the con­struction of shallow, temporary surface dwellings. Fur­ther, he believes that the Fremont may have moved eastward to become the Dismal River culture, citing the approximate Fremont abandonment-inception of Dismal River culture as suggestive, the presence of paddle-and-anvil pottery in both areas, general simi­larities in architecture between Dismal River dwellings and a structure at Injun Creek (Aikens 1966b), and the dependence of both groups on a mixed economy of hunting and gathering. Aikens' argument in simplified form merely attempts to account for certain influences in Fremont culture not totally explained by postulated Anasazi or Southwest­ern relationships. External influences contributing to the distinctiveness of Fremont culture probably did come from the Great Plains as well as from the South­western cultural area. The limited scope of this paper does not allow for an in-depth analysis of Aikens' hypo­thesis; however, his use of architectural data is much less than suggestive either of a Great Plains origin and con­tinued influence, or of the impetus to the Dismal River culture. Rather than Fremont representing cultural ties or the movement of different peoples into and out of an area, it seems more profitable to approach the question with regard to economy. Yes, it may be that shallow surface dwellings, serrated bison metapodial fleshers and scrapers, etc., were Great Plains additions. But these could also be interpreted as elements of a culture exploiting a large, mobile resource such as the bison. Aikens himself (1966b), with regard to the so-called "Promontory Culture," states that differences in artifact assemblages between cave and Fremont habita­tion sites may simply involve differential resource ex­ploitation. It is quite possible that the reminiscent Great Plains traits are only part and parcel of a Fremont mixed economy, emphasizing hunting-gathering and horticulture. Dwellings at some sites would be temporary as opposed to those in other areas; tools would be simi­lar in both the Great Plains and the Great Basin, given a similar emphasis in subsistence. Whatever the origins and antecedents of the Fremont, the culture as a whole is clearly distinguished from sur­rounding cultures with reference to architecture and settlement plan. (See Fig. 6, for depiction of Fremont, Pueblo, and Great Plains dwellings.) Open habitation sites are most similar to those in the Southwestern cul­tural area, with the exception of certain sites excavated in the Great Salt Lake variant, which may only be in­dicative of different economic strategies. Villages com­prise pithouse architecture, and surface habitation and storage structures. Intra-site plan appears informal or random in comparison with the standardized layout of Pueblo sites to the south, although individual elements may be very similar. 51 TABLE XIII Typical Anasazi Architectural Traits Present Within the Fremont Area (after Bullard 1962). Site Boundary Village . Burnt House Caldwell Village Cub Creek DamSite Felter'sHill Flat Top Butte Ford Site Fremont Playhouse Gilbert Site Goodrich Site Maclebd Site Wagon Run Whitcrocks Village Wholeplacc Village Alice Hunt Site •Basket Hut Site Chas: B.Hunt Site Crescent Ridge Emery; Fallen Woman Gnat Haven lvie Ridge Innocents Ridge Ninas Hill Nine Mile Canyon NorthPoint Old Woman : Playa Site Poplar Knob Power Pole Knoll Round Springs :-- Snake Rock Village Turner-Look Site Windy-Ridge ^Backhoe Village Evans Mound Garrison Kanosh. Marysvale. Mediail Village Paragonah Ephraim Grantsville Nephi -Mounds.. Pharo Village Tooele Bradshaw Mound •Injun'Creek'-' Knoll Site "'. LeveeSite Willard Site Clay-rimmed Firepits - ; . • - • • . 3 0 11 0 1 '•- I . 0 0 o 1 3 • • 3 •..• - , - . v 7 3 ...••••: : . 5 • 4'. 1 ' F . . I =r • y . - ' - ! _ - . - '; 1 • 1 2 0. • ' • 3 ' 0. 4 " 0 3 j 1 10 2 " 4 9 0 .4 • 5 32 2 J 2 6 1 1 2 0 0 1 1 Ash Pits 4 0 0 0 1 „ 0 0 o 0 . ^ : o- 0 I . 0 0 0 0 0 0 0 0 - 0 • o 0 2 0 0 0 0 1 o 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 Above Floor Storage Bins 0 0 1 0 0 0 0 0 0 0. 0 0 r0. 0 1 o 0 0 o 0 0 0 0 0 • 0 • ;•;- ' - : 1 ••••• 0 0 0 0 0 0 . 0 0 0 0 0 0 0 0 . o . 0 0 0 0 0 0 1 0 0 0 0 Slab Wall Construction O O 0 0 0 O 0 0 0 0 0 O 0 0 O 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 : > • ••"•. Oi v-".•':•'•• 2? 0 0 0 0 0 0 0 0 A) 0 0 0 0 0 0 0 0 Total Houses in Sample 6 5 15 1 1 7 4 1 -'• 1 2 . 3 : . . . . • ' 4; . ''. 4 5 6 .-. •• • 1 "' i:;::k'-^: "!:l:••.'•;• / / - • " - ' • !"•': 2 1 2 1 2 4 . . I 4 3 5 1 3 I 1 12 8 3 9 17 ' 5 4 8 13 2 3 5 10 3 1 3 3 1 S 1 -.' • v'rl . i - ' •. .. • 7 52 San Rafael - ca. AD900 Parowan - ca. AD 1100 San Rafael - ca. AD900 entrryy ^ v- '' ramp Fig. 6. Fremont Architecture Alkali Ridge - ca. AD800 inferior partitions antechamber Coombs Village - ca. AD1100 Dismal River ca. AD1600 / / i ;o yo Great Salt Lake - ca. AD 1400 \ 0 ,''<S fr. \Qm proposed margin ;o Oy L~shaped venlifator o\ o: Q' a- 53 Unfortunately, the debate over the distinctiveness of Fremont ultimately becomes one of semantics. Pueblo groups and Fremont are closely related, though a sug­gestive Great Plains-like character is also clear. One may identify influence, development, etc., on the basis of certain general traits, e.g. pilhouse architecture, moc­casins and the like. However, structural details, intra-site plan and economy, are only characteristic of the Fremont. CONCLUSIONS Fremont is a recognizable and viable archeological concept, regardless of the lack of a precise and involved definition. Morss (1931) was the first to formally recog­nize this, although somewhat vaguely: ". . . Finally, in its time lag, in the modification of its Southwestern traits (for example, pottery, basketry, mats, figurines), and most strikingly, in the substitution of moccasins for sandals, the Fremont culture shows that it is not an integral part of the main stream of Southwestern devel­opment." Later researchers have become more en­meshed in the "northern periphery" puzzle, where a people are not wholly Southwestern but certainly show similar characteristics. Aikens (1967) injected new life into the study of Fremont with his hypothesis concern­ing a possible Great Plains origin and dispersal, a new approach in an area formerly considered little more than a cultural backwater of the Southwest. Debate still centers on Fremont origins and ultimate demise, internal composition, and chronology. Fremont exists, irrespective of the applied referent, and it will continue to be fruitless to attempt an ultimate definition involv­ing an origin or extinction. We already have a rough category or pigeon hole into which we can wedge Fre­mont; what is needed is a series of investigations focused on economic strategy, settlement pattern, social organi­zation, and the like. In short, we need to approach Fre­mont in terms of a cultural system, and refrain from elaborate speculation involving postulated cultural identities. This study has presented statistical support for in­ternal variation within the Fremont cultural area. Using independent data, discriminant analysis upheld Mar­witt's (1970) identification of five variants: Uinta, San Rafael, Parowan, Sevier, and Great Salt Lake. Not unexpectedly, even more precision was obtained when Fremont was broken into two macro-variants or sub­divisions, corresponding to Fremont and Sevier as pro­posed by Jennings (1956) and Madsen and Lindsay (1977). I have also suggested that Colorado Plateau and Great Basin should be adopted instead of Fremont and Sevier as these terms appear to be more insightful and des­criptive with regard to the reasons for Fremont internal variation. Parowan, Sevier, and Great Salt Lake may be grouped under the Great Basin, with Uinta and San Rafael under Colorado Plateau. This breakdown is more indicative of Fremont viewed as one cultural tradi­tion occupying two major physiographic regions, and comprising a subarea of the overall Southwestern Cul­tural area. Fremont is therefore seen as a distinct cul­tural entity, consisting of two major cultural subdivi­sions corresponding in geographical extent to the Colo­rado Plateau and the eastern Great Basin, further sub­divided into five recognizable variants. What little is known about Fremont settlement plans also supports the two-fold division. Great Basin settle­ments are generally larger village sites occupying optimal locations of arable land and available perennial water supply. The Colorado Plateau, with its more rigorous and dissected landscape, offers fewer optimal areas for horticulturalists, with correspondingly fewer large vil­lage sites. Predictably, a high correlation exists between optimal resource areas and higher population densities. The division of Colorado Plateau-Great Basin Fremont is real, though probably not evidence of a cultural boun­dary. More likely, Fremont settlement and architecture differ as a result of the exploitation of contrasting en­vironments and/or external influences from independent sources. This paper has supported the view that Fremont as a cultural area may be broken into two macro-variants or subdivisions, corresponding to the Colorado Plateau and eastern Great Basin physiographic regions. Further, these subdivisions can then be broken down into five variants. If debate on the existence of the Fremont subsides, we will be able to concentrate on the cultural system that was present. Far from being the cultural backwater of the Southwest and therefore uninteresting, studies in the Fremont area have the intriguing potential of explaining why hunting and gathering populations opted for horticulture. Researchers in the Fremont cul­tural area are in the enviable position of approaching one of the most interesting questions in anthropology. We merely need to get over the hurdle of just who were Fremont, who influenced them the most, when did they exist, and where did they go? It would be far more in­teresting and profitable to understand their relation­ships with environment and with their cultural neigh­bors, with particular emphasis on how and why they became horticulturalists. 54