ABSTRACT. Metes and bounds surveys of property ownership boundaries for the period 1807-1887 have been used to characterize forest composition in three Tennessee counties. The counties (dates, numbers of surveys, numbers of corner trees) were: Blount, 1824-1826, 336, 1071; Roane, 1807-1808, 141, 519; Rhea, 1824-1889, 260, 1625. Trees most often used as corners were oaks and hickories, but pine, poplar, and chestnut were also much used. Comparison of percent composition of the surveys with modern inventories reveals that some species/species groups have remained relatively constant over time, some taxa have increased in relative importance (e.g., Pinus, Juniperus), and some have decreased (e.g., mesophytes as Fagus grandifolia). Changes in land use, modern forest "management" and disease are probably responsible for percentage shifts. Plant communities known from modern studies are no more than suggested by survey percentages or species co-occurrences. Survey records provide valuable information about Tennessee’s forest landscape near the time of settlement and a few decades thereafter.

     The vegetation cover of the whole landscape at or near the time of settlement by European-Americans is of interest to field scientists. Such information is relevant to the research of historians (Williams 1989), paleoecologists (Delcourt et al. 1993), pedologists (Jenny 1980), anthropologists (Chapman and Shea 1981), and vegetation biomass modelers (Waring and Schlesinger 1985). Vegetation ecologists (as DeSelm 1994) use past vegetation patterns to interpret the effects of historical, cultural, and environmental factors on present vegetation (Mueller-Dombois and Ellenbereg 1974). Congressional land survey records and similar rectilinear surveys (cf. DeSelm 1994) and metes and bounds surveys have been used to interpret early settlement vegetation (DeSelm 1995, 1997, 1999; DeSelm and Rose 1995).

     This paper reports vegetation results obtained from metes and bounds surveys of the period in Blount, Rhea and Roane counties (Fig. 1). I report percent forest composition as seen by surveyors and compare that with modern forest inventory composition.

     The surveyed parts of Blount County lie in the Ridge and Valley Physiographic Province and a few surveys were made in coves in the Blue Ridge Province. The surveyed part of Roane County also is in the Ridge and Valley while almost all surveys in Rhea County are in the Appalachian Plateaus (Cumberland Plateau) part of the county (Fenneman 1938). These areas (including only low elevation Blount County) have a tropical humid climatic type. Precipitation varies from 122 to 132 cm per year, rather well distributed, but floods and late-season droughts are common (Trewartha 1968, Dickson 1960). January mean maximum temperatures are in the 10-11EC range, mean minima are in the -1-0EC range. July mean maximum temperatures range from about 31-32EC, mean minima range from about 18-20EC (Dickson 1960).

     Blount County of 1820, near the beginning of the survey period, included small parts of the present Sevier and Loudon counties (Fig. 1). Parallel southwest-northeast trending ridges and valleys characterize most of the county landscape. These areas are underlain by Paleozoic sandstones, dolomites, limestones and shales (Rodgers 1953, Luther 1977, Hardeman 1966). Major rivers, the Tennessee and the Little Tennessee, border the county. Elevations range from about 300 m on the Tennessee River to 2026 m but surveying was done between about 300 and 450 m. Topography is undulating to steep. Soils are mostly well drained, shallow to deep, and loamy to clayey–some have chert in the profile or have limestone outcrops. Soils are classed as Alfisols, Inceptisols and Ultisols (Springer and Elder 1980).

51

     Of the 239 Blount County surveys located by place, 205 were in the Ridge and Valley on the Holston and Tennessee rivers or on tributaries of these and on tributaries of the Little Tennessee River; locations were: Elijoy, Crooked, Nine Mile, Nails, Six Mile creeks and the Little River. The other 34 surveys were on 14 other creeks or at other places including four at unknown places (not in the Tennessee Gazeteer, Fullerton, 1974). All of these were in the Ridge and Valley except 12 surveys in Tuckaleechee and Millers coves–calcareous valleys of the Blue Ridge (Hardeman 1966).

Roane County of 1810, at the end of the survey period, included parts of the present Loudon, a large part of Cumberland, most of Morgan, as well as most of Roane counties (Fig. 1). Surveys were in the eastern, the Ridge and Valley part of the county. The Tennessee, Clinch, and Emory rivers bisect the area. Elevations range from about 220 to 620 m on the plateau escarpment (Walden Ridge) just beyond the west edge of the survey area. The topography is undulating to steep. Soils are shallow to deep, well drained to excessively drained, clayey, cherty or stony and are classed as in Blount County (Springer and Elder 1980).

     Rhea County of 1840 (Fig. 1), near the middle of the survey period, lies both in the Ridge and Valley and on the Cumberland Plateau. The county is bordered on the east by the Tennessee River and, in the west, it extends beyond the Plateau escarpment 4-8 km. Elevations vary from about 208 to 600 m. The topography is undulating to steep. On the Plateau, bedrock is sandstone and shale, with limestone on the lower slope of the escarpment and in the bottom of deep valleys (Hardeman 1966). Soils are loamy, often stony, well drained, moderately deep, and rock outcrops may occur. They are mapped as Ultisols and Inceptisols (Springer and Elder 1980).

52

     The floras of these areas are rather well known (Wofford and Kral 1993, Chester et al. 1993, Chester et al. 1997). Known floras (native plants) range in numbers of species from 570 in Rhea to 1050 in Blount County (Univ. of Tennessee Herbarium 1998). The forest vegetation is discussed by Stephenson et al. (1993) and Hinkle et al. (1993), the barrens by DeSelm and Murdock (1993) and the sandstone outcrops (flatrocks) by Perkins (1981). Generally, the forests fall into upland oak, oak-hickory, oak-pine or oak-cedar types; lower slopes and coves exhibit mesic hardwood or hemlock-hardwood types and flooded areas exhibit wetland types. Forest vegetation class composition or average forest composition by site have been published in the Ridge and Valley and on the Plateau (Anonymous 1953, 1954, Cowan 1958, Parr and Pounds 1987, and Thor and Summers 1971). Plant community classes are mapped in these counties (Tennessee Valley Authority 1941).

     More than 10,000 years ago, people of Native American cultures began visiting the surveyed areas (Paleoindian and Archaic cultures). The major stream valleys were used by later cultures for village and field sites, uplands were hunted for game animals and wild plant food. Fire was commonly used in the forest to clear the understory (Hudson 1976, Lewis and Kneberg 1958). European-American settlement occurred following treaties with Native Americans over the period 1780-1819 (Mathews 1960, Folmsbee et al. 1969). Forests on low-slope topography were cleared and row crops cultivated. Slopes were logged for farm, or in some cases export timber, and forests were grazed and often burned using spring surface fires (Killebrew et al. 1874).

     Land Entry Books (Entry Takers Books) were available from Blount, Roane, and Rhea counties. Land transfers are described from the periods: Blount 1824-1826, Rhea 1824-1887, 1902-1929, and Roane 1807-1808. The Blount and Roan data were handwritten records from the State Archives in Nashville, typed in the 1930s (1938 Blount County, no date Roane County), as part of the Works Progress Administration Historical Records Project. They are available from Mountain Press in book form. The Rhea County Land Entry Book (Surveyors Book No. 1) is a publication of the Rhea County Historical and Geneological Society. The numbers of surveys used and (usually) trees cited were: Blount: 336-1118, Rhea: 260-1621 and Roane: 141-526. Surveys came from parts of the county areas of the period (Fig. 1). Blount surveys were all by Robert Wear. Rhea County surveys were apparently all from the Cumberland Plateau area except for two tracts of I. T. Lockes on the Tennessee River. In this county no records after 1889 were used. Roane County records include a few entries from nearby (Cumberland, Bledsoe and Overton counties–all part of the Fourth Survey District) (Crouch 1968).

     The original records were typed, or copied as they were found. Plant and topographic spellings and other nomenclature is preserved here (plant names used have been repeated in the tables). The surveys usually recorded a tree (or stump) at each survey corner, sometimes more than one tree, or topographic feature, or a stake was recorded instead. (In Blount County more stake corners were used than tree corners.) The botanical qualifications of the surveyors is not known. Surveyors used a compass and measured distances in poles or chains and links. No attempt is made to find the survey lines on the ground. No diameters or point-to-tree distances are given. Specific locations are not known.

     In a few surveys, only stakes were used. Some of these are at town lots as in Kingston. The early surveys passed through wild, little modified forest land but all started from known corners of existing property lines. Some surveys crossed roads or paths or recorded fields. Some surveys duplicated corners, apparently representing resale of lands earlier described by the same corners. This tree data has been eliminated when recognized.

53

Tree names, tree-form descriptors and landscape descriptors were similar to those used in other, mainly nineteenth century, surveys (DeSelm 1995, 1997, 1999, DeSelm and Rose 1995). Most place names could be found in modern gazetteers (Fullerton 1974, United States Geological Survey 1991).

Two unusual natural-cultural features in the county were ore banks and ironworks. The people considered the latter so important to their success there, that ore bank land was condemned for use only by the ironworks (company). Table 1 contains the percentages of "species" of survey stems; 40 taxa were named, in addition was Frenchwood which is unknown. See also the modern inventories in the table. The oak proportion in the Survey data is notable–they exceed proportions of pines and hickories by a factor of more than three. The high percentages of black and post oaks suggest that there may have been communities of these types there–in fact, types dominated by white, black and post oaks are known (Martin 1971, 1978, Martin and DeSelm 1976). The black and post oak percentages are nearly as high as those from northern Sevier County (DeSelm and Rose 1995) where

Table 1. Percent occurrences in Surveys from Blount County. Cowan (1946) data is percent board feet. TVA (1971) DEN is for stems $5 inches; TVA (1971) VOL is for stems $9 inches.

most of the data came from the "slate" knobs. The sum of percents from nine mesophytes including beech, sugar maple and poplar suggests that some lower slope and cove communities were sampled (and then cleared for agriculture). The paucity of wetland taxa, as sycamore, suggests that few bottoms were sampled. Mesic and wetland forests are well known in the area (Martin 1971).

     Comparison of Survey with inventory data (which excludes the Great Smoky Mountains National Park) reveals the existence of species which are modern increasers. These are locust (probably mostly Robinia pseudoacacia, which is an invader [Burns and Honkala 1990] and is planted), walnut (planted and saved to grow because of the value of the wood), poplar (an increaser with disturbance, Martin 1971, Smith 1968, Burns and Honkala 1990), cedar, the pines (disturbance invader and increasers), and black gum, and chestnut oak. The apparent increases in the last two may be a sampling area phenomenon. Surveys were done on deep soils, the best potentially agricultural soils of the period, but few forests available to be inventoried occur there now. Thus modern inventories are more likely placed on stony slopes where chestnut oak and black gum are typical.

     Decreasers include hickories, white oak and post oak. This may again be that the lands of these survey species are now in agriculture and modern samples come from poorer sites. Comparing percent number of stems versus volume of stem wood allows comparing tree numbers versus size. Hickory, locust (largely Robinia), pines and post oak have higher relative density than volume suggesting small tree size and overall modern management for pulp, poles, ties, and posts. Other oaks (white, chestnut and mixed oak including black and southern red) have larger relative volume than density suggesting current management for saw timber.

56

     The surveys reported plant observations among 33 taxa. Surveys were in the Ridge and Valley part of the county though some stream bottoms of the Emory River are included (which may have been in the Plateau). Features seen include a fish trap, coal, a mill seat, a saltpeter cave and a sucking spring or pond. Apparently the Fourth District (Crouch 1968) had been previously surveyed at the section level since repeated references are made to section lines, section miles and such trees as ". . . the two mile tree of the section line. . . ."

     Percentages among taxa (Table 2) are less than five percent each except for hickories, pines, poplar and three oak taxa. The sum of all oaks exceeds the sum of hickories, pines, or poplar by X5 to X8. The oak percent exceeds that of chestnut by more than X24. The relatively high percentages of white, post, and black oaks suggests that communities dominated by these taxa may have occurred in Roane County of that period. In fact, several white oak types, a black oak type, and a post oak type occur in modern forests (Martin 1971, 1978, Martin and DeSelm 1976, Stephenson et al. 1993). Several white oak types contain black and post oaks and the post oak type contains black oak (Martin 1971). Pines, poplar, hickories, and chestnut are common constituents of the white, post, and black oak types; indeed pines or poplar are dominant among other types (Martin 1971).

     Percentages of mesic taxa, including poplar, beech and sugar maple sum to 25 percent of the total indicating a good representative of moist sites and mesic forests in the area. On the other hand, percentages of wetland taxa, as sycamore, total only 4.1 although some survey lines crossed streams and some paralleled rivers.

     On Yellow Creek in Bledsoe County, surveyors Richard Hodson and John Hunt surveyed ". . . 200 poles to a stake in a gease near to four blazed black oaks . . ." (gease ? = geason, meaning producing scantily, barren or unproductive; Oxford English Dictionary, Simpson and Weiners 1989). This might have been a barren which vegetation types are well known on the Plateau (DeSelm 1992).

     Increasers between Survey and modern times are black gum (X6 – nearly X10), Spanish oak (X4), red (and scarlet) oaks (X1.8), chestnut oak (nearly X30-43.5), and pines (X5.8-11.2). These increases are again probably due to modern sample location. Survey samples came from the Ridge and Valley where few forests now remain whereas modern inventory data comes from the Plateau part of the county where forests abound. The pine increases are due to forest disturbance, where it is an invader, and due to planting. No cedar was mentioned in the Survey–it is now a successional invader (Smith 1968) and is common enough to appear in the 1973 inventory (Tennessee Valley Authority 1973).

     Decreasers include the mesophytes sugar maple, buckeye, beech, ash, poplar and hemlock–their sites now mostly in cultivation. Similarly the inventory use of the well-drained to tight soils of oak forests with white, post, and black oak, and dogwood have declined. Swamp taxa, boxelder and sycamore have declined with drainage and conversion of their sites to agriculture. The loss of chestnut and elm are due to disease (Hepting 1971).

     A table of 156 co-occurrences was prepared (not shown) where more than one tree was recorded at the same corner. Most commonly a species name was repeated on a single corner (as white oak with white oak); however, white oak also occurred at three or more corners with hickory, dogwood, and black oak. Post oak occurred most commonly with black oak and hickory. Sugar maple and beech were commonly associated. The co-occurrence of the oaks, hickories and dogwood suggest the existence of white oak, post oak, and black oak forests, or these together or with hickory as a dominant. These are reported by Martin (1971) along with a beech-sugar maple type (Martin 1971).

56

Rhea County

Surveying in Rhea County was done on the Plateau surfaces, and also on the slopes of the escarpment, at its base, and in Cranmore Cove. Surveyors were chiefly Jesse Thompson and Alfred Collins but 12 other surveyors worked for short periods. They reported a tar kiln in operation and

Table 2. Percent occurrences in surveys from Roane County. The Cowan (1946) data is percent board feet.

57

such land features as coal banks, flatrock, rockhouse, calybeate spring (also as cylebriate spring), chinquepin plain, ponds, a swamp, canebrakes, and "Gaze’s hickory flat."

The Surveys reported 38 taxa (Table 3). Percentage representation reported among taxa are all less than five except for certain oaks, chestnut, and hickory. The oak total exceeds that of chestnut or hickory by ratios of 6.7 and 4.5. The high percentages of white, black and post oaks suggest that communities dominated by these taxa occurred. Modern communities dominated by white oak, mixed oaks (high in black oak), and post oak occur on the Plateau (Hinkle 1978, 1989). These xeric

types all contain some pine and hickory, in fact, Hinkle (1978) reports both Virginia and shortleaf pine types. In the Surveys, pine on ridges and cliffs is reported repeatedly. Oak, oak-pine, oak-chestnut and oak-hickory percentages (Table 3) sum to half to two-thirds of the total. The sum of mesic taxa as poplar, beech, maple and others indicates that only a few coves were sampled. The totals of sycamore, sweet gum, water oak and hackberry indicate that few wetlands were seen although a swamp and two ponds were crossed. Fields and old fields were seen but percentages of invaders as cedar, persimmon and sassafras are low; potential invaders there as poplar and pine cannot be distinguished from individuals in natural forests.

     Of the 32 tree taxa, only about 12 names can be used with modern inventory names; others are understory plants or are included in combined species categories. Useable taxa include those which are constants (holding their percentages with time) as maple, black gum and hemlock. Decreasers include those cut heavily for lumber, post and white oaks. Chestnut was decimated by disease (Hepting 1978).

     Apparent increasers include taxa such as sweetgum, poplar, and pine (increases vary from about X2 to X18). The increase in beech may simply be that beech is often left in the forest after logging. Chestnut oak may be less accessible among the oaks and so also less frequently cut.

     A table of co-occurrences of taxa, mentioned together at the same corner was prepared (not shown); they total 888 including 260 duplicates of a single species reported more than once at the same corner. White oak occurred mainly with black oak, hickory and chestnut. Black oak was recorded most with post oak, white oak, hickory, and chestnut. Blackjack was most often with post oak. Red oak (some were doubtless scarlet oak) was most often with post oak. The pines grew mostly with chestnut, black and chestnut oaks and black gum. Dogwood grew mainly with white oaks and hickory.

     Between the three surveyed areas, 43 taxa were named, 25 to the species level, 12 to the genus level and two were unknown. The intergeneric terms gum and locust were used, in Pinus, yellow pine was cited (subgenus Pinus, Little and Critchfield 1969), and the term red oak probably included northern red and scarlet oaks. The forests were mainly hardwood–percentages totaled 86.1-94.7 and softwoods only <4.9-13.9. The proportions of post, black, and white oaks and the hickories were large. Importance trends among genera were in Blount County oak > hickory = pine > poplar > chestnut (and beech exceeds chestnut); Roane County oak > hickory > pine = poplar > chestnut (six other taxa exceed chestnut); Rhea County oak > hickory > chestnut > pine (black gum exceeds pine). The first two (easternmost) of these counties lie in the Appalachian Oak Forest (Stephenson et al. 1993) where Survey data support this classification. Rhea County lies in the Mixed Mesophytic Forest of Hinkle et al. (1993) who make it clear that uplands are chiefly oak, oak-hickory and oak-pine forests as are indicated in the surveys.

59

     Increasers in two of the three county areas are pine, black gum and chestnut oak. The first is an opportunistic invaders, spreading with disturbance (Burns and Honkala 1990), the last two now remain in areas difficult to log. Decreasers in two areas are white and post oak, hickory, and maple (chiefly red maple). The decreasers are probably due to the loss of moist soil oak sites by their conversion to agriculture. Poplar seems to be an increaser with pine and cedar in Blount County but a slight decreaser in Roane County which lost other mesophytes such as sugar maple and beech (mentioned previously) during land use conversion to agriculture.

     Metes and bounds survey records are not without deficiencies of field methods which make interpretations in terms of vegetation cover less than conclusive. Included are lack of precise location of survey starting points and survey lines, surveying through or near disturbed areas as roads, fields, and other already-in-use properties, the naming of small plants at corners, the possible non-random choice of corners, and the possible limited taxonomic expertise of surveyors as evidenced by the few species seen and their grouping of species. In spite of these deficiencies, the surveys constitute nearly the only record of vegetation present near the time of settlement. Thus, their contribution to our historical botanical geographic knowledge is great.

63