Bulk Density and Fragipan Identification in Loess Soils
of the Lower Mississippi River Valley
D. L. Lindbo,* F. E. Rhoton, J. M. Bigham, W. H. Hudnall, F. S. Jones, N. E. Smeck, and D. D. Tyler
High bulk density (pb) compared with overlying soil horizons has long been considered one of the principal characteristics of
fragipans, yet recent investigations have noted inconsistent relationships between fragipans and p b. This study was conducted to
evaluate the fragipan pb relationship in several major loessial soils of the lower Mississippi River Valley. Average p b (field-moist,
saran-coated clods) obtained from five Memphis-Loring-Grenada catenas extending from northwest Tennessee to southeast Louisiana
show that pb in the fragipans of the Loring (fine-silty, mixed, thermic Typic Fragiudalf) and Grenada (fine-silty, mixed, thermic Glossic
Fragiudalf) pedons range from 1.45 to 1.78 Mg m (x = 1.56). A wider range of 1.38 to 1.86 Mg m (x = 1.55) occurred at similar
depths in the Memphis (fine-silty, mixed, thermic Typic Hapludalf) pedons. Within most Loring and Grenada pedons, the p b of the
fragipan horizon is not statistically unique (based on Duncan’s new multiple-range test) and does not always represent the maximum
for a given pedon. There is also a weak positive correlation (r = 0.44) between total sand content and p b, suggesting that higher pb in
these soils are more closely related to lithologic factors than to pedogenic densification. Specifically, pedons with a signi ficant
influence of Coastal Plain sediment have higher p b than those developed entirely in loess, suggesting that the high p b are partially
inherited. These results indicate that p b should not be used as a principal test for differentiating between fragipan and nonfragipan
horizons in these soils. It appears that root and water restriction typically associated with fragipans in the lower Mississippi River Valley
are related to physical or chemical characteristics of the fragipans other than p b.
NUMEROUS FIELD CLUES are listed by the Soil Survey Staff (1975, 1992) to assist in the recognition and delineation of fragipan
horizons. Although there are no established chemical or laboratory procedures for fragipan identification, it is widely acknowledged that
fragipans have a high pb, which decreases permeability and root penetration and profoundly impacts properties and land use.
Consequently, high pb relative to overlying horizons has been suggested as a possible quantitative means for fragipan identification
(Witty and Knox, 1989; Quandt and Glocker, 1993).
The proposal presents several problems. For example, some researchers have noted that: (i) a dense horizon does not always
indicate the presence of a fragipan (Buntley et al., 1977; Parfitt et al., 1984; McIntosh and Kemp, 1991); (ii) the fragipan may not
always have the highest Pb in a profile (Anderson and White, 1958; Grossman et al., 1959; Norton and Franzmeier, 1978; Habeck er et
al., 1990); and (iii) pb may be related to degree of development of a fragipan (Smith and Daniels, 1989). Research on a variety of
parent materials has also indicated that high Pb may be an inherited feature rather than a pedogenic one (Miller et al., 1971 ; Wang et
al., 1974; Smith and Callahan, 1987; Franzmeier et al., 1989; Smeck et al., 1989; Habecker et al., 1990; Lindbo and Veneman, 1993)
or that the densification process may not be entirely pedogenic (Calhoun, 1980). For example, Bryant (1989) noted that high d ensities
could occur soon after deposition of sediments due to physical ripening (desiccation) or self-weight collapse (Hand, 1973). Ray (1967,
p. D221-D227) suggested that slow deposition of loess was often accompanied by greater weathering and compaction as fine silts
settled into packing voids. As a result, internal drainage decreased and pore spaces became clogged by illuvial clay and other
weathering products, thereby increasing pb and decreasing permeability. Such occurrences fall in the gray area between geogenic and
Data from several Grenada pedons, located in Arkansas, Kentucky, Louisiana, and Mississippi, showed that the fragipans were not
consistently denser than overlying horizons (Romkens et al., 1986). Additionally, some Pennsylvania soils developed in either eolian or
lacustrine materials did not exhibit significant differences in pb between fragipan and nonfragipan soils (Petersen et al., 1970). In one
New Zealand study, field morphology (vertical gray veins, prismatic structure, brittleness, and root restriction) was the dis tinguishing
factor, as pb was the same in both fragipan and nonfragipan pedons (McIntosh and Kemp, 1991).
Soils of the Memphis catena extend throughout the lower Mississippi River Valley. Several members of the catena exhibit classic
fragipan characteristics (brittleness, vertical gray tongues, prismatic structure, and root restriction), whereas the Memphis soil does
not. The objectives of this investigation were to: (i) compare the p b of fragipan and nonfragipan soils within the Memphis catena; and
(ii) evaluate the effectiveness of pb in delineating the fragipan horizons in silty soils of the lower Mississippi River Valley.
MATERIALS AND METHODS
The sites for this investigation are within the southern Mississippi Valley Silty Uplands designated as Major Land Resource Area 134
by the Soil Conservation Service (1981). Five sites were selected, approximately 150 km apart, from northwestern Tennessee to
southeastern Louisiana (Fig. 1). Three of the sites (Obion County, Tennessee and Tate and Yazoo counties, Mississippi) are located in
deep loess (Zones B and C; loess thickness >3 m, Fig. 1), whereas the other two (Fayette County, Tennessee, and West Felician a
Parish, Louisiana) are located in shallow loess underlain by Coastal Plain sediments (Zone E; loess thickness - 2 m, Fig. 1). At each
site, three pedons (Memphis, Loring, and Grenada) were described and
D.L. Lindbo and F.E. Rhoton, USDA-ARS, National Sedimentation Lab., P.O. Box 1157, Oxford, MS 38655; J.M. Bigham, F.S. Jones, and N.E. Smeck,
Dep. of Agronomy, Ohio State Univ., Columbus, OH 43210; W.H. Hudnall, Agronomy Dep., Louisiana Agric. Exp. Stn., Louisiana State Univ.
Agricultural Center, Baton Rouge, LA 70803; and D.D. Tyler, Dep. of Plant and Soil Science, Univ. of Tennessee, Jackson, TN 38301. Contribution
from the USDA-ARS National Sedimentation Lab. Received 7 Apr. 1993. *Corresponding author.
Published in Soil Sci. Soc. Am. J. 58:884-891 (1994).
sampled via soil pits to a depth of approximately 3 m. A fourth pedon representing a transitional soil between the Memphis and Loring
was also investigated at the West Feliciana Parish site. Sites were selected so that pedons representing soils of the catena were
spaced between 20 and 500 m apart. An exception occurred in West Feliciana Parish, where the Memphis pedon is separated from
the others by approximately 2 km. The general topographic position of each pedon was the same at all sites, with Memphis at t he
summit, Loring at the backslope, and Grenada at the toeslope position. At the time of sampling, land use at all sites was pasture or
hay. The fragipan horizons were identified by morphology (vertical gray seams or tongues, coarse prismatic structure, and rel ative
absence of roots in the prisms) and brittleness (Soil Survey Staff, 1975, 1992); transitional horizons (i.e., E/Btx, Bt/Btx) were those with
<60% brittleness in the matrix material. Horizons designated with a prefix of 2 at the Tate and Fayette sites are developed i n Roxana
Silt; those with a prefix of 3 at the Fayette site or 2 at the West Feliciana site are developed in Coastal Plain material.
Particle-size distribution was determined for each horizon using the pipette method (Soil Survey Staff, 1984, Method 3Al). Bulk density
(field moist) was measured using the average of three saran-coated clods for each horizon (Soil Survey Staff, 1984, Method 4AI). The
Pb values were compared between fragipan and nonfragipan horizons at each site, each loess depth, and for the entire study (Table 1).
Those horizons within the control section and designated with an x were considered fragipan horizons in the analysis; non-fragipan
horizons in Memphis pedons (identified by italics in Table 2) were those horizons in the control section that span the approximate
depth range of fragipan horizons in the Loring and Grenada at a given site. The deep fragipan horizons present in the West Feliciana
Memphis soil were not included as fragipans in the data analysis, as they occurred below the control section. Although these horizons
were not included in the data analysis, they appear identical to other fragipan horizons observed in this study and are probably related
to the overall hydrology of the site. Comparisons were also made among all horizons within a given pedon (Table 2). Regression
equations of pb vs. total sand, silt, and clay contents were developed using SAS (SAS Institute, 1985) (Table 2).
RESULTS AND DISCUSSION
Relationships between Pedons
The pb of fragipans in the Loring and Grenada pedons range from 1.45 to 1.78 Mg m (x = 1.56) (Tables I and 2) and are
similar to those reported in the literature (Romkens et al., 1986; Rhoton and Tyler, 1990; Switzer and Pettry, 1992). A somewhat
wider range of 1.38 to 1.86 Mg m (x = 1.55) is observed at comparable depths in the Memphis pedons. Fragipan p b within the
Loring and Grenada pedons in Yazoo County and West Feliciana Parish are significantly higher than p b at similar depths in the
associated Memphis pedons (Table 1). Bulk densities among pedons are not different in the Tate County samples. In Obion
County only the Grenada fragipan p b is greater than the pb at similar depths in the Memphis pedon. In Fayette County the p b of the
Memphis pedon is actually greater than the p bs of fragipan horizons from both the Loring and Grenada pedons.
The data (fragipan horizons) for all Grenada pedons were combined first by loess depth (deep and shallow) and then for all
Grenada pedons and compared with simidarly combined data from the Loring (fragipan horizons) and Memphis (nonfragipan
horizons) pedons. Combined data from the three deep loess sites (Obion, Tate, and Yazoo Counties) indicate Pb is greatest (by
only 0.02 Mg m ) in the Grenada pedon and statistically similar in both the Memphis and Loring pedons. The combined data for
the shallow loess sites (Fayette County and West Feliciana Parish) indicate no statistical differences between pedons. There is no
difference among Grenada, Loring, and Memphis Pb when the data from all five sites are combined by pedon (Table 1).
Such similarity in pb between fragipan and nonfragipan soils was also observed by Buntley et al. (1977) in West Tennessee.
They noted that some Memphis pedons contained a dense, brittle zone, but this zone either lacked the typical morphology and
root-restricting properties of a fragipan or occurred beneath the control section and did not influence classification. More recently,
McIntosh and Kemp (1991) observed that the fragipan horizons in some loess-derived soils of New Zealand were only slightly
denser than horizons at similar depths in nonfragipan soils occurring within the same landscape. The differences are similar to
those in this investigation and suggest that p b may differ slightly across a loess landscape, but is unrelated to the occurrence of
Relationships within Pedons
If pb is to be used quantitatively to delineate fragipans, it should be consistently and significantly different from overlying and
underlying horizons. The statistically highest pb occurs in the fragipan in only one of the 10 Loring and Grenada pedons studied
(Grenada pedon, Tate Co.) (Table 2 and Fig. 2a). Even in that case, an underlying horizon has a statistically similar but sli ghtly lower
pb. Otherwise, the pb is higher than overlying horizons throughout the fragipan but is similar to lower B and C horizons (Loring and
Grenada pedons, Yazoo Co., and Loring pedon, Tate Co.), increases with depth (Loring pedon, Fayette Co.), or is similar to ho rizons
above and below (Grenada pedon, Obion Co.). The remaining four pedons have Pb that are highest in horizons other than the
fragipan. Three of these pedons (Loring and Grenada, West Feliciana. Pa., and Grenada, Fayette Co.) exhibit an increase in p b with
depth, reaching a maximum below the fragipan. The maximum p b occurs both above and below the fragipan in the Loring pedon from
Obion Co., but not in the fragipan.
Our data show that pb is not consistently greatest in the fragipan, a phenomenon that has also been observed by others in this
region (Romkens et al., 1986) and in other loessial soils (Anderson and White, 1958; Grossman et al., 1959). The instances where p b
increases with depth are restricted to the shallow loess pedons, where underlying Coastal Plain sediments have a high p b (Fig. 2b).
Thus, as others have indicated, delineation of the lower boundary of the fragipan is rarely possible based on p b, particularly if a high pb
is inherited from the parent material (Miller et al., 1971; Wang et al., 1974; Smith and Callahan, 1987; Habecker et al., 199 0). The
similarity of pb between the fragipan and those of overlying and underlying horizons in the majority of the deep loess pedons furffier
illustrates the impossibility of using p b to delineate fragipans in this region. The loessial parent materials (C horizons) investigated have
a pb ranging from 1.40 to 1.59 Mg m (x = 1.51). The average fragipan pb is only slightly greater, suggesting that there has been little
pedogenic densification of these soils. Despite little evidence for densification, these soils have undergone significant pedogenesis
involving eluviation-illuviation, rubification, and redox processes (Lindbo et al., 1994, unpublished data).
Since high pb is not unique to the fragipan horizons, it follows that some other property or properties are responsible for the
observed root restriction, impedance of water movement, and brittle consistence of the fragipans. Both Guthrie et al. (1983, p. 66-68)
and Jones (1983) have suggested p b values exceeding those derived from the following equation: pb = 1.95 - 0.0186x, where x is the
clay percentage, would restrict root growth (Hudnall and Williams, 1989, p. 43-68). Application of this formula to our data indicates that
pb for both fragipan and nonfragipan horizons exceed the derived value (Fig. 3); however, only the fragipan horizons appear to restrict
root growth in the field.
The highest pb observed in this study are in the fragipans developed in shallow loess. These pedons typically have an increasing
total sand content due to basal mixing between the loess and the underlying Coastal Plain sediments (Fig. 2b). A regression analysis
of total sand content vs. pb for the shallow loess pedons yields an r = 0.56 (the r for individual pedons ranges from 0.61 to 0.95,
Table 2). The deep loess pedons have an r = 0.001 (the r for individual pedons ranges from 0.00 to 0.28, Table 2). While sand
content cannot account for all the observed increase in p b (overall r = 0.44), it appears to have some relationship. Similar
observations have been made in other regions (Norton and Franzmeier, 1978; Habecker et al., 1990; Lindbo and Veneman, 1993).
Total silt content shows similar correlations, but typically the r values are lower and less significant (Table 2). Such similarity is
expected as silt contents are inversely related to sand. Clay contents are poorly correlated to p b (Table 2).
Bulk densities in the fragipans range from 1.45 to 1.78 Mg m (x = 1.56). At similar depths in the Memphis pedons, p b ranges from
1.38 to 1.86 Mg m (x = 1.55). Since the differences between fragipan and nonfragipan horizons are so small (<0.1 Mg m ), pb must
not be used as a primary test for differentiating fragipans in these soils. Furthermore, the fragipan horizon pb are neither consistently
unique (dissimilar to overlying or underlying horizons) nor greatest within a given pedon, making positive identification impossible
based solely on this criterion. In some of these loess-derived pedons, higher pb are attributed to increases in sand content inherited
from underlying Coastal Plain materials. In lieu of pb, morphology (vertical gray seams and polygonal expression) must be used as the
primary criteria for delineating fragipan horizons in the loess soils of the lower Mississippi River Valley.
We thank Vince Campbell, John Cox, and Dan McChesney for their field, laboratory, and computer assistance. We also thank Drs.
Robert Cullum, Chris Evans, and Peter Veneman for their reviews and constructive comments. This manuscript has been approved for
publication by the Director of the Louisiana Agricultural Experiment Station as Manuscript 9309-7259. All programs and services of the
U.S. Department of Agriculture are offered on a nondiscriminatory basis without regard to race, color, national origin, religion, sex,
marital status, or handicap.
Anderson, J.U., and J.L. White. 1958. A study of fragipans in some southern Indiana soils. Soil Sci. Soc. Am. Proc. 22:450-454.
Bryant, R.B. 1989. Physical processes of fragipan formation. p. 141150. In N.E. Smeck and E.J. Ciolkosz (ed.) Fragipans: Their
occurrence, classification, and genesis. SSSA Spec. Publ. 24. SSSA, Madison, WL
Buntley, G.J., R.B. Daniels, E.E. Gamble, and W.T. Brown. 1977. Fragipan horizons in soils of the Memphis-Loring-Grenada se-
quence in west Tennessee. Soil Sci. Soc. Am. J. 41:400-407.
Calhoun, T.E. 1980. Recommended reclassification or disposition of Northeast Region series now classified as having fragipans . In
E.J. Ciolkosz (ed.) Proc. Northeast Coop. Soil Surv. Conf., University Park, PA. 23-27 June 1980. Agron. Series no. 65.
Pennsylvania State Univ., University Park.
Franzmeier, D.P., L.D. Norton, and G.C. Steinhardt. 1989. Fragipan formation in loess of the midwestern United States. p. 69-98. In
N.E. Smeck and E.J. Ciolkosz (ed.) Fragipans: Their occurrence, classification, and genesis. SSSA Spec. Publ. 24. SSSA, Madison,
Grossman, R.B., J.B. Fehrenbacher, and A.H. Beavers. 1959. Fragipan soils of Illinois: 1. General characteristics and field
relationships of the Hosmer silt loam. Soil Sci Soc. Am. Proc. 23:65-70.
Guthrie, R., J. Baker, R. Haberman, W.H. Hudnall, G. Huntington, B. Rouke, N. Smeck, and P. Veneman. 1983. Rep. of Soil Taxon-
omy Committee. Tech. Comm. 1. In Proc. Nad. Work Plan Conf. Coop. Soil Survey, USDA-SCS, Washington, DC. 28 Mar. -1 Apr.
1983. U.S. Gov. Print. Office, Washington, DC.
Habecker, M.A., K. McSweeney, and F.W. Madison. 1990. Identification and genesis of fragipans in Ochrepts of north central Wiscon-
sin. Soil Sci. Soc. Am. J. 54:139-146.
Handy, R.L. 1973. Collapsible loess in Iowa. Soil Sci. Soc. Am. Proc. 37:281-284.
Hudnall, W.H., and D. Williams. 1989. Fragipan distribution in the south central states. In N.E. Smeck and E.J. Ciolkosz (ed.)
Fragipans: Their occurrence, classification, and genesis. SSSA Spec. Publ. 24. SSSA, Madison, WI.
Jones, C.A. 1983. Effect of soil texture on critical bulk densities for root growth. Soil Sci. Soc. Am. J. 47:1208-1211.
Lindbo, D.L., and P.L.M. Veneman. 1993. Morphological and physical properties of selected fragipan soils in Massachusetts. Soil Sci.
Soc. Am. J. 57:429-436.
McIntosh, P.D., and R.A. Kemp. 1991. Consolidated horizons in soils formed in aeolian deposits in Southland, New Zealand. Geo -
Miller, F.P., N. Holowaychuk, and L.P. Wilding. 1971. Canfield silt loam, a Fragiudalf. L Macromorphological, physical, and che mical
properties. Soil Sci. Soc. Am. Proc. 35:319-324.
Norton, L.D., and D.P. Franzmeier. 1978. Toposequence of loessderived soils in southwestern Indiana. Soil Sci. Soc. Am. J. 42:
Parfitt, R.L., M. Saigusa, and D.N. Eden. 1984. Soil development processes in an Aqualf-Ochrept sequence from loess with admix-
tures of tephra, New Zealand. J. Soil Sci. 35:625-640.
Petersen, G.W., R.W. Ranney, R.L. Cunningham, and R.P. Matelski. 1970. Fragipans in Pennsylvania soils: A statistical study of
laboratory data. Soil Sci. Soc. Am. Proc. 34:719-722.
Quandt, L.A., and C.L. Glocker. 1993. Fragipans - New criteria. p. 302. In Agronomy abstracts. ASA, Madison, WI.
Ray, L.L. 1967. An interpretation of profiles of weathering of the Peorian Loess of western Kentucky. U.S. Geol. Sury. Prof. Pap.
R6mkens, M.J.M., H.M. Selim, H.D. Scott, R.E. Philips, and F.D. Whisler. 1986. Physical characteristics of soils in the southern region:
Captina, Gigger, Grenada, Loring, Olivier, and Sharkey series. South. Coop. Ser. Bull. 264. Mississippi Agric. For. Exp. Sm.
Rhoton, F.E., and D.D. Tyler. 1990. Erosion-induced changes in the properties of a fragipan soil. Soil Sci. Soc. Am. J. 54:223228.
SAS Institute. 1985. SAS/STAT guide for personal computers. SAS Inst., Cary, NC.
Smeck, N.E., M.L. Thompson, L.D. Norton, and M.J. Shipitalo. 1989. Weathering discontinuities: A key to fragipan formation. p.
99-112. In N.E. Smeck and E.J. Ciolkosz (ed.) Fragipans: Their occurrence, classification, and genesis. SSSA Spec. Publ. 24.
SSSA, Madison, WI.
Smith, B.R., and L.L. Callahan. 1987. Soils with Bx horizons in the Upper Coastal Plains of South Carolina. Soil Sci. Soc. Am. J. 51:
Smith, B. R., and R. B. Daniels. 1989. Occurrence and characteristics of fragipans on the Coastal Plains of the southeastern United
States. p. 33-42. In N.E. Smeck and E.J. Ciolkosz (ed.) Fragipans: Their occurrence, classification and genesis. SSSA Spec. Publ.
24. SSSA, Madison, WI.
Soil Conservation Service. 1981. Land resource regions and major land resource areas of the United States. USDA-SCS Agric.
Handb. 296. U. S. Gov. Print. Office, Washington, DC.
Soil Survey Staff. 1975. Soil taxonomy: A basic system of soil classification for making and interpreting soil surveys. USDA-SCS Agric.
Handb. 436. U.S. Gov. Print. Office, Washington, DC.
Soil Survey Staff. 1984. Procedures for collecting soil samples and methods of analysis for soil survey. USDA-SCS Soil Surv. Invest.
Rep. 1. U.S. Gov. Print. Office, Washington, DC.
Soil Survey Staff. 1992. Keys to soil taxonomy. 4th ed. Soil Manage. Support Serv. Tech. Monogr. 19. Virginia Polytechnic Inst. and
State Univ., Blacksburg.
Switzer, R.E., and D.E. Pettry. 1992. Phosphorous distribution in loessial Fragiudalfs as affected by relief. Soil Sci. Soc. Am. J.
Wang, C., J.L. Nowland, and H. Kodama. 1974. Properties of two fragipan soils in Nova Scotia including scanning electron
micrographs. Can. J. Soil Sci. 54:159-170.
Wascher, H.L., R.P. Humbert, and J.G. Cady. 1948. Loess in the southern Mississippi Valley: Identification and distribution of the loess
sheets. Soil Sci. Soc. Am. Proc. 12:389-399.
Witty, J.E., and E.G. Knox. 1989. Identification, role of soil taxonomy, and worldwide distribution of fragipans. p. 1-10. In N.E. Smeck
and E.J. Ciolkosz (ed.) Fragipans: Their occurrence, classification, and genesis. SSSA Spec. Publ. 24. SSSA, Madison, WI.