Title Page
 Description of region
 Scope of present study, field observations...
 Preliminary study of chemical properties...
 Intensive study of chemical properties...
 Literature cited

Group Title: Bulletin University of Florida. Agricultural Experiment Station
Title: Effect of frequent fires on chemical composition of forest soils in the longleaf pine region
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00026834/00001
 Material Information
Title: Effect of frequent fires on chemical composition of forest soils in the longleaf pine region
Series Title: Bulletin University of Florida. Agricultural Experiment Station
Physical Description: 39 p. : ill., map ; 23 cm.
Language: English
Creator: Heyward, Frank
Barnette, R. M
Publisher: University of Florida Agricultural Experiment Station
Place of Publication: Gainesville Fla
Publication Date: 1934
Copyright Date: 1934
Subject: Longleaf pine -- Soils   ( lcsh )
Forest soils -- Composition   ( lcsh )
Burning of land   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Includes bibliographical references (p. 39).
Statement of Responsibility: by Frank Heyward and R.M. Barnette.
General Note: Cover title.
 Record Information
Bibliographic ID: UF00026834
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltuf - AEN4732
oclc - 18206661
alephbibnum - 000924126

Table of Contents
    Title Page
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
    Description of region
        Page 7
        Page 8
    Scope of present study, field observations on vegetation and physical properties of soils
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
    Preliminary study of chemical properties of soils from burned and unburned areas
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
    Intensive study of chemical properties of soils from burned and unburned areas
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
    Literature cited
        Page 39
Full Text

Bulletin 265 March, 1934

Wilmon Newell, Director


Frank Heyward, Junior Forester,
Southern Forest Experiment Station, U. S. Forest Service
R. M. Barnette, Chemist,
Florida Agricultural Experiment Station


Bulletins will be sent free upon application to the

John J. Tigert, M.A., LL.D., President of the Geo. H. Baldwin, Chairman, Jacksonville
University A. H. Blanding, Bartow
Wilmon Newell, D.Sc., Director A. H. Wagg, West Palm Beach
H. Harold Hume, M.S., Asst. Dir., Research Oliver J. Semmes, Pensacola
Harold Mowry. B.S.A., Asst. Dir., Adm. Harry C. Duncan, Tavares
J. Francis Cooper, M.S.A., Editor J. T. Diamond, Secretary, Tallahassee
R. M. Fulghum, B.S.A., Assistant Editor
Ida Keeling Cresap, Librarian
Ruby Newhall, Administrative Manager BRANCH STATIONS
K. H. Graham, Business Manager
Rachel McQuarrie, Accountant NORTH FLORIDA STATION, QUINCY
L. O. Gratz, Ph.D., Plant Pathologist in Charge
R. R. Kincaid, M.S., Asst. Plant Pathologist
MAIN STATION, GAINESVILLE J. D. Warner, M.S., Associate Agronomist
R. M. Crown, B.S.A., Assistant Agronomist
AGRONOMY Jesse Reeves, Farm Superintendent
W. E. Stokes, M.S., Agronomist** CITRUS STATION, LAKE ALFRED
W. A. Leukel, Ph.D., Agronomist
G. E. Ritchey. M.S.A., Associate* John H. Jefferies, Superintendent
Fred H. Hull, M.S., Associate Geo. D. Ruehle, Ph.D., Associate Plant Pathol-
W. A. Carver, Ph.D., Associate ogist
John P. Camp, M.S., Assistant W. A. Kuntz, A.M., Associate Plant Pathologist
B. R. Fudge, Ph.D., Associate Chemist
ANIMAL HUSBANDRY W. L. Thompson, B.S., Assistant Entomologist
A. L. Shealy, D.V.M., Animal Husbandman** EVERGLADES STATION, BELLE GLADE
R. B. Becker, Ph.D., Specialist in Dairy Hus- ane, Ph.D., Agronomist in Charge
bandry A. Daane, Ph.D., Agronomist in Charge
bandry A N Lobdell, M.5., Entomologist
W. M. Neal, Ph.D., Associate in Animal Nutri- R. odell, .S., Entomologronomist
tion F. D. Stevens, B.S., Sugarcane Agronomist
D. A. Sanders, D.V.M., Veterinarian G. R. Townsend, Ph.D., Asst. Plant Pathologist
M. W. Emmel, D.V.M., Assistant Veterinarian A. Bourne, Ph.D., Sugarcane Physiologist
W. W. Henley, B.S.A., Assistant Animal Hus- R. idder, PB.S., ima Husbandmn
bandman R. W. Kidder, B.S., Asst. Animal Husbandman
P. T. Di Arnold, B.S.A., Assistant in Dairy In- Ross E. Robertson, B.S., Assistant Chemist

CHEMISTRY AND SOILS H. S. Wolfe, Ph.D., Horticulturist in Charge
R. W. Ruprecht, Ph.D., Chemist** W. M. Fifield, M.S., Assistant Horticulturist
R. M. Barnette, Ph.D., Chemist Stacy O. Hawkins, M.A., Assistant Plant
C. E. Bell, Ph.D., Associate Pathologist
H. W. Winsor, B.S.A., Assistant

ECONOMICS, AGRICULTURAL W. F. Ward, Asst. Animal Husbandman in
C. V. Noble, Ph.D., Agricultural Economist**
Bruce McKinley, A.B., B.S.A., Associate
M. A. Brooker, Ph.D., Associate
Zach Savage, M.S.A., Assistant FIELD STATIONS

Ouida Davis Abbott, Ph.D., Specialist** Leesburg
L. W. Gaddum. Ph.D., Biochemist M. N. Walker, Ph.D., Plant Pathologist in
C. F. Ahmann, Ph.D., Physiologist Charge
W. B. Shippy, Ph.D., Asso. Plant Pathologist
K. W. Loucks, M.S., Asst. Plant Pathologist
ENTOMOLOGY J. W. Wilson, Ph.D., Associate Entomologist
J. R. Watson, A.M., Entomologist** C. C. Goff, M.S., Assistant Entomologist
A. N. Tissot, Ph.D., Associate
H. E. Bratley, M.S.A, Assistant Plant City
A. N. Brooks, Ph.D., Plant Pathologist
HORTICULTURE R. E. Nolen, M.S.A., Asst. Plant Pathologist
A. F. Camp, Ph.D., Horticulturist** Co
M. R. Ensign, M.S., Associate Ccoa
A. L. Stahl, Ph.D., Associate A. S. Rhoads, Ph.D., Plant Pathologist
G. H. Blackmon, M.S.A., Pecan Culturist
Roy J. Wilmot, M.S.A., Specialist, Fumigation Hastings
Research A. H. Eddins, Ph.D., Asso. Plant Pathologist
R. D. Dickey, Assistant Horticulturist
PLANT PATHOLOGY -- Assistant Entomologist
W. B. Tisdale. Ph.D., Plant Pathologist** Bradenton
George F. Weber, Ph.D., Plant Pathologist B eon
R. K. Voorhees, M.S., Assistant David G. Kelbert, Asst. Plant Pathologist
Erdman West, M.S., Mycologist Sa
*In cooperation with U.S.D.A. E. R. Purvis, Ph.D., Assistant Chemist, Celery
"**Head of Department. Investigations



Frank Heyward and R. M. Barnette'

Introduction ...................................... ................... ......... 3
Description of Region .............. .. ........................... ................ 7
Scope of Present Study ........................................................... 9
Field Observations on Vegetation and Physical Properties of Soils ....................... 9
Preliminary Study of Chemical Properties of Soils from Burned and Unburned Areas.... 18
Intensive Study of Chemical Properties of Soils from Burned and Unburned Areas....... 24
Discussion ............................................ .................. 34
Summary .............. ........................................ 38
Literature Cited ....................................... .. .. ................... 39


Throughout the longleaf pine region, one of the most important
sources of lumber and other forest products in the United States,
the practice of setting fire to the forest each year has prevailed
for generations. This burning is attributable in general to two
groups: cattlemen who believe that annual burning of the woods
improves the natural range by stimulating the growth of forage
plants, particularly grasses, and naval stores operators who in-
tentionally burn the woods as a method of preventing later un-
controlled fires. It is the general custom of the latter before
burning to rake all combustible material, such as chips, pine
needles and grass, away from the bases of trees that are being
worked for resin. Trees thus protected by individual firebreaks
apparently are not damaged to any great extent by the intention-
al fires. Burning under such conditions largely obviates the like-
lihood of accidental fires during the succeeding year, after which
the raking and burning are repeated.
It has been suggested that removal of organic materials from
the surface of the soil by frequent fires reduces the nitrogen

'The authors express their appreciation of the advice of V. L. Harper and
J. G. Osborne in connection with the statistical handling of data, and of
laboratory assistance by H. W. Jones.

4 Florida Agricultural Experiment Station

and organic matter content of the soil. Hilgard(13)2, as long ago
as 1860, decried the burning of the natural range in the South and
recommended the use of longleaf pine litter as a fertilizer. His
analysis showed that a ton of longleaf pine needles collected in
Mississippi contained 6.5 pounds of calcium, 2.5 pounds of potas-
sium, 2.4 pounds of magnesium, and 0.5 pounds of phosphoric
acid. Advocates of fire protection point out that pine litter has a
nitrogen content of approximately 1 percent, and that the annual
removal of this material by fires before it has begun to decompose
must ultimately result in a deficiency of soil nitrogen and organic
matter. In support of the policy of annual burning, on the other
hand, it has been argued that on burned areas the fertility of the
soil is improved by the stimulated development of the native
grasses, which add organic matter to the soil through the decay of
their roots; also, that through the addition of ash to the soil
plant nutrients are kept in constant circulation instead of being
retained in undecomposed organic matter.
In Europe the opinion has long prevailed that fires increase the
calcium content of the soil and decrease its acidity. Ehren-
berg(6) mentions that the custom of burning over agricultural
soils to increase their fertility was practiced by the Romans. This
investigator in his studies of agricultural soils found that burning
broke down the absorptive complex, particularly in clay soils, and
increased the percentage of water-soluble salts. He states that
light burning may be expected to result in an increased yield of
crops, but, that severe or too frequent burning may be detrimental
to crop production. All Ehrenberg's work dealt with agricultural
and not forest soils, and his studies dealing with the availability
of plant nutrients were based in part on soils heated in an oven.
The applicability of his results to forest soils in the South is
doubtful. Hesselman(12) attributed the successful regeneration
of spruce in Sweden to increased nitrification resulting from recent
fires. Spruce reproduction on burned areas made more rapid
growth than spruce on areas long protected from fire. In pure
coniferous forests the heavy raw humus layer renders electro-
lytes immobile and creates high acidity in the upper soil horizons.
These conditions are unfavorable for bacterial activity and cause
decreased nitrification. The release of these substances from the
ash in the form of soluble salts following fire greatly accelerates
nitrification. Glomme(9), in a series of pot experiments on nitri-
2Figures in parentheses (italic) refer to "Literature Cited" in the back
of this bulletin.

Bulletin 265, Effect of Frequent Fires 5

fiction, found that the addition of varying quantities of birch
ashes to different types of forest humus favored nitrate forma-
tion, although the degree and rate of the response varied widely
according to the type of humus. Hess(ll), working in Switzer-
land, reported greater percentages of calcium carbonate and less
acid conditions in the soils for five distinct forest areas subjected
to fires, as compared with adjacent unburned areas. Hess pointed
out that as early as 1853 a beneficial effect of fire on marly and
peaty soils was referred to by Hartstein(10), who reported that
agricultural lands were often burned in order to increase their
fertility. Salisbury(15), in England, concluded that increased
carbonates and nitrates and decreased acidity were important fac-
tors determining better establishment and growth of reproduc-
tion on old charcoal heaths compared with adjacent unburned
In this country, little research has been done on the effect of
fire on soils. Snyder(16) reported a loss to the soil of 2,500
pounds of nitrogen per acre following the Hinckley, Minnesota,
fire in 1893. This estimate was based on only two individual soil
samples from the burned area and two from the unburned. More
recently, Alway and Rost(1), examining forest soil in Minnesota
that had been subjected to one severe fire, found no changes in
nitrogen and moisture equivalent but a slight decrease in acidity
in the top 3 inches of soil. These investigators point out that the
only marked change in the soil by fire was the consumption of the
layer of organic matter on the soil surface. They state, "Thus
there is no evidence that fire caused any significant change in
chemical composition or physical properties of the soil below the
surface layer of organic residues, even where this was entirely
burned off. The lime, phosphoric acid and potash of the leafmold
would suffer no loss by burning and much would be left in a more
immediately available form." Barnette and Hester(2), working
in Florida, found greater quantities of nitrogen and organic mat-
ter in a forest soil on an island off the coast, that had been protect-
ed from fire for many years, than in a mainland forest soil that
had been subjected to annual burning. The mainland soil show-
ed a greater percentage of replaceable calcium and higher pH
The present study was undertaken in an endeavor to obtain
additional information on the effect of fire on forest soils, with
particular reference to the longleaf pine region, where fires are
of widespread and frequent occurrence.



S. s r r

Fig. 1.-Soil Divisions of the Coastal Plain (after Bennett), and location of study areas.

Bulletin 265, Effect of Frequent Fires 7

The longleaf pine region of commercial importance forms a belt,
rarely more than 200 miles wide, extending within the Atlantic
and Gulf Coastal Plain from South Carolina to eastern Texas. The
coastal plain consists of a series of indistinctly separated terraces,
each representing a different stage in the recession of the ocean
in past ages. Certain soil differences are apparent between each
terrace and the next, and Bennett(3) has recognized several soil
divisions within the Atlantic and Gulf coastal plain (Fig. 1). Of
these divisions, the four within which most of the field work of
the present study was done are the Atlantic and Gulf flatwoods,
sandy lands, middle coastal plain, and upper coastal plain. The
soils of the flatwoods consist largely of poorly drained fine sands
characterized by a low content of nitrogen and of organic matter.
Those of the sandy lands are composed of deep, excessively drain-
ed, fine to medium sands of even lower fertility. Those of the
middle and upper plains are somewhat heavier in texture, partic-
ularly the sub-soils, with fine sandy loams as the predominating
type. These soils, also, have low percentages of nitrogen and or-
ganic matter, but are in general more productive than the lighter
types. Their drainage varies widely according to topography;
but, in general, they may be regarded as intermediate between
those of the flatwoods and those of the sandy lands.
Within this region, longleaf pine (Pinus palustris Mill.) is the
principal tree species, occurring in pure open stands on most of the
well-drained sites. The old-growth timber has practically all
been cut and replaced by extensive forests of second growth up
to 35 or 40 years of age (Figs. 2 and 3). Stands of longleaf
pine characteristically contain little or no underbrush, although
a dense ground cover consisting chiefly of several species of wire-
grass (Aristida spp.), or of broomsedge (Andropogon virginicus
L.) is usually present. The scarcity of underbrush is attributable
to the action of frequent fires.
In this region forest fires are more common than in any other
part of the United States, but are rarely of the devastating crown-
fire type. Because the pine stands lack underbrush and are open,
most of the fires occurring in them are confined entirely to the
forest floor. Surface fires, however, when burning before a brisk
wind or occurring during dry periods, sometimes burn fiercely
and cause great damage to stands of young timber,

8 Florida Agricultural Experiment Station

Fig. 2.-Typical stand of longleaf pine second growth. This stand is about
12 years old and has been protected from fire for five years.

Fig. 3.-This stand is a portion of the one shown in Fig. 2 but has been sub-
jected to annual fires. The two areas are separated by the fire line at the

Bulletin 265, Effect of Frequent Fires 9

The purpose of this study was to compare the chemical proper-
ties of soils from areas frequently burned over with those of soils,
similar in all other characteristics, that had not been burned
over for at least 10 years. It was necessary that the soils com-
pared be from pairs of adjoining or closely adjacent areas, and
that they show no marked differences as to soil series, type, or
With these conditions in mind, an extensive survey was made of
the territory from southern South Carolina to Louisiana. More
than 30 long-unburned stands of pine were located, but of this
number only 18 were found suitable for study. The locations of
the 18 study areas are shown in Fig. 1; and their location, for-
est and soil types, and fire history are shown in Table I.
At the beginning of the study, no means were at hand for es-
timating the number of individual soil samples needed to give
reliable results. Therefore the study areas were sampled super-
ficially, in an attempt to detect any large differences caused by
Following chemical analysis of these samples, 8 areas were
selected for detailed study, and sufficient samples were collected
from each to permit statistical treatment of the data.

On the burned areas, a ground cover of either wiregrass or
broomsedge, together with a wide variety of herbaceous plants,
was present (Fig. 4). Various genera of the family Leguminosae
were usually abundant, including Dolicholus, Meibomia, and Les-
pedeza; several species of Chrysopsis, also, were common. This
ground cover on annually burned areas was rarely dense enough
to cover the surface of the soil completely. On many spots the
soil was exposed directly to the elements, as illustrated by Fig. 5.
On the unburned areas, ground-cover conditions varied mark-
edly according to whether the forest stand was open-grown or
dense. In openings in the stand, 'a dense ground cover of wire-
grass was typical; usually few other herbaceous plants were pres-
ent. Individual plants of wiregrass sometimes produced leaves

30ne of the 18 study areas, the Chiefland area, was not sampled in this
manner, because at the time when this collection was completed it had not
been located.


Fire History
Area Fre Soil Designation and Year
Number Location Forest Type Drainage Unburned Area. Burned Area. Sampled
No fire for given Subjected to fire:
number of years
1 Summerville, S. C... Longleaf pine....... Norfolk fine sandy loam 14 Annually 1931
Drainage poor

"2 Ridgeland, S. C.I.... Loblolly pine....... Norfolk loamy fine sand One fire in preceding Annually 1931
Drainage good 12 years

3 Waycross, Ga.'..... Slash pine .......... Leon fine sand 15 to 20 Frequently 1931
Drainage poor

4 Racepond, Ga....... Longleaf-slash pine. Leon sand 38 Frequently 1931
Drainage poor

5 Lake City, Fla.t..... Longleaf-loblolly pine Blanton fine sand 12 Frequently 1931 Q
Drainage good

6 Starke, Fla ......... Slash pine............ Portsmouth fine sand 12 to 15 Frequently 1931
Drainage very poor

7 Raiford, Fla......... Slash-loblolly pine... Blanton fine sand 11 Annually 1931
Drainage good 1933
8 Trenton, Fla.1....... Longleaf pine...... Blanton fine sand 15 to 20 Frequently 1931
Drainage excessive 1932

9 Chiefland, Fla.1..... Longleaf pine....... Unclassified fine sand Possibly two fires in Frequently 1933
S __Drainage good preceding 50 years
'Fire histories for these areas are approximate only.


Fire History
Area Soil Designation and Year
Number Location Forest Type Drainage Unburned Area. Sampled
No fire for given Burned Area.
number of years Subjected to fire:
10 Bartow, Fla........ Longleaf pine....... Fort Meade fine sand 44 Frequently 1932
Drainage excessive
11 Valparaiso, Fla...... Longleaf pine-turkey Norfolk sand, deep phase 23 Annually 1931
oak.............. Drainage excessive

12 Adrian, Ga... ...... Longleaf pine....... Ruston sand 15 Biennially 1931
Drainage good 1932
13 Chipley, Fla.'........ Longleaf pine....... Norfolk fine sand 15 Frequently 1931
Drainage good
14 Stapleton, Ala....... Longleaf pine....... Norfolk fine sandy loam 11 Annually 1931
Drainage good
15 State Line, Miss.'.... Slash pine.......... Plummer sandy loam 14 Frequently 1931
Drainage poor
16 Wiggins, Miss.1...... Longleaf pine....... Norfolk fine sand 12 Frequently 1931
Drainage good
17 McNeill, Miss...... Longleaf pine....... Orangeburg fine sandy
loam 10 Annually 1932
Drainage good 1933
18 Urania, La.......... Longleaf pine....... Montrose silt loam 10 Annually 1931
Drainage good _1933
1Fire histories for these areas are approximate only.

12 Florida Agricultural Experiment Station

Fig. 4.-Typical annually burned area, McNeill, Miss. Burned Dec. 20,
1932, and photographed June 4, 1933.

Fig. 5.-Detail of ground cover at a point selected at random on the area
shown in Fig. 4. The soil is exposed because of continued removal of litter
by fire.

Bulletin 265, Effect of Frequent Fires 13

30 inches in length. These leaves, tangled and interwoven, form-
ed a dense "rough" which completely covered the soil surface to
a depth of 12 inches or more (Fig. 6). On areas having a cover-
ing of such "rough," as the pines increase in age and their crowns
tend to close in and form a continuous canopy over the forest
floor, the wiregrass may be smothered out by the accumulation of
pine litter (Fig. 7). Complete elimination of wiregrass, even
under dense timber, appears to require a number of years. In
openings in the stand, this grass may persist indefinitely.

Fig. 6.-Dense "rough" of wiregrass in opening in pine stand on area un-
burned for at least 10 years.
Under dense stands of pine protected from fire, the forest floor
was characterized by a well-defined Ao horizon consisting of 1 to
2 inches of pine Jitter underlain by a layer of duff from 1/3 inch to
1 inch deep (Fig. 8). In places, the entire Ao horizon was as much
as 31/2 inches deep. Depths in excess of this are rare in the long-
leaf pine region.
The surface horizon of the soil from burned and unburned areas
was found to differ considerably in physical condition. On the un-
burned areas this horizon was loose, mellow, and very permeable.
The soil could be scooped up with the fingers to a depth of 4 to 5
inches. The activity of insects and earthworms was much in evi-
dence, and the soil contained many of their holes, sometimes to
depths of 10 or 12 inches. Earthworm casts were very abundant
immediately beneath the Ao horizon. Frequently, numerous holes

14 Florida Agricultural Experiment Station

Fig. 7.-In this young stand of longleaf pine, unburned for six years, a
heavy ground cover was present originally but it has been completely cov-
ered and smothered by pine litter.

Fig. 8.-Upper portion of soil profile on area unburned for 11 years. The
depth of the Ao horizon was approximately three inches at this particular
location, averaging somewhat less for the entire area.

Bulletin 265, Effect of Frequent Fires 15

from 1 inch to I1 inches in diameter, made by small burrowing
animals, were found ramifying through the surface horizon (Fig.

Fig. 9.-Holes of burrowing animals commonly found under pine litter such
as that shown in Fig. 8.

The surface horizon of soils on burned areas was impermeable
and extremely compact, as a result of direct exposure to the sun
and to driving rains. With the exception of occasional ant hills,
evidence of an active macrofauna was rare. The surface of these
soils was divided by a network of dams approximately 1/2 inch in
height, extending from one stool of wiregrass to another, behind
which rain water collected. These dams were composed for the
most part of fragments of charcoal, together with bits of bark,
decayed wood, and other organic material, and apparently had
been formed by surface movement of rain water during heavy
On the unburned areas a great many tiny feeding roots of the
pines were found just beneath the Ao horizon. These roots fre-
quently had worked their way well into the duff layer. No such
concentration of feeding roots was apparent in the surface layer of
soil on burned areas, where no protecting layer of litter was pres-
ent except that which had accumulated during the current year.
With the possible exception of several species of scrub oaks,
longleaf pine is much more resistant to fire than any associated

16 Florida Agricultural Experiment Station

tree species. For this reason, pure stands of pine with little un-
derbrush were found on burned areas, whereas on unburned areas
a large variety of hardwood trees and smaller woody plants were
always associated with the pine (Fig. 10). Tree species common-
ly associated with longleaf pine on unburned areas in the flat-
woods portion of the longleaf region, in addition to slash pine
(Pinus caribaea More), were laurel oak (Quercus laurifolia
Michx.), water oak (Q. nigra L.), sweet bay (Magnolia virginiana
L.), swamp bay (Persea pubescens Sarg.), red maple (Acer
rubrum L.), tree huckleberry (Vaccinium arboreum Marsh.), and

Fig. 10.-Left: Area burned annually. Note heavy ground cover of grass
and absence of hardwood species. Right: Area protected from fire since
1916. Note layer of pine litter and presence of hardwood undergrowth.
Both areas, Urania, La.
wax-myrtle (Myrica cerifera L.). In this part of the longleaf
region gallberry (Ilex glabra (L.) A. Gray) and saw palmetto
(Serenoa serrulata (Michx.) Hook.) are common, because of
their ability to sprout after fire. On areas protected from fire,
these underbrush species attain much greater size than on fre-
quently burned areas.
On the soils of the middle and upper coastal plain, southern red
oak (Q. rubra L.), red gum (Liquidambar styraciflua L.), dog-
wood (Cornus florida L.), and wax-myrtle are commonly found

Bulletin 265, Effect of Frequent Fires 17

associated with longleaf pine. Because of frequent fires, these
trees are rarely prominent in the stand. Even on frequently
burned areas, however, they sometimes persist. When protected
from fire, they form a dense undergrowth (Fig. 10, right) and
sometimes develop into co-dominant or even dominant trees in
what was originally a pure stand of pine. The ultimate develop-
ment sometimes attained by hardwoods in an oak-longleaf associ-
ation on an area protected from fire is illustrated in Fig. 11.
Laurel oak and live oak (Q. virginiana Mill.) have here developed
with the pine into dominant and co-dominant trees. An adjoin-
ing area characterized by the same site factors, but subjected to
frequent fires, is shown in Fig. 12. These areas are only 20
feet distant from each other, being separated only by a country
road. The view of the burned area is typical of the moderately

Fig. 11.-Area near Chiefland, Fla., burned possibly twice during the past
50 years.

dry sandy soils of the peninsula of Florida. Invasion by oaks and
their development have been largely prevented by fires. Sur-
rounding the numerous poorly-drained "sink-holes" in the same
locality both live oak and laurel oak are abundant, as fires do not
burn over such sites except infrequently during years of unusual

18 Florida Agricultural Experiment Station

Fig. 12.-Frequently burned area separated from that shown in Fig. 11
only by a woods road. Note small size of liveoak in background at left.
Underbrush of chinquapin (Castanea nana Muhl.), which sprouts prolifically
following fire.
In the preliminary sampling, 4 individual samples were col-
lected from each of the burned and unburned plots. Four holes,
approximately 1 foot square and varying in depth with soil type,
were dug on each plot, at locations selected at random but con-
fined to an area not greater than 1/8 acre. If the 8 profiles thus
exposed were found to be closely alike, particularly with respect
to the depths of horizons, three 1,000-gram samples were collect-
ed, as follows: (1) one from the surface of the soil usually to a
depth of 2 to 3 inches (Fig. 13); (2) one from that portion of the
soil profile just beneath the dark organic discoloration, approxi-
mately the portion between the depths of 4 and 8 inches; and,
lastly, (3) one from an arbitrary depth that varied widely with
soil type but that was usually between 18 and 24 inches beneath
the surface. Because only matched profiles were sampled, it was
possible to collect the samples from each respective horizon at the
same depth on each pair of burned and unburned areas4. If decay-
4Reference to definite horizons has been purposely omitted because of the
difficulty of distinguishing the A and B horizons in the sandy soils of the

Bulletin 265, Effect of Frequent Fires 19

Fig. 13.-Method of collecting soil sample from first depth.

ed roots or large deposits of charcoal5 were encountered, the pro-
file was discarded and another hole was dug. The four samples
for each of the respective horizons were thoroughly mixed and
1,000 grams of the composite taken for analysis.
In the laboratory, the composite samples were air dried and
passed through a 2-mm. round-hole sieve. Suitable portions were
then analyzed, according to standard methods, for total nitrogen,
loss on ignitionG, and replaceable calcium. Hydrogen-ion concen-
trations of distilled water suspensions of the dried and sieved soils
were determined with a Youden quinhydrone apparatus.

"5Fragments of charcoal were found in every soil examined, although at
least two of the areas were definitely known to have escaped fire for the
preceding 50 years. Charcoal appears to be present in practically all the
longleaf pine soils of the region, frequently to depths of 30 inches or more.
The presence of this material at such depths is difficult to explain unless
it is attributed to roots that have been completely burned and then dispersed
throughout the soil. In places the charcoal is so uniformly distributed, in
small fragments, that it appears to have been deposited simultaneously
with the mineral portion of the soil.
6Loss on ignition was used in this study as an estimate of organic matter.
In the region studied, because of the general absence of carbonates and the
small percentage of combined water in the common soils, loss on ignition
affords a fairly reliable means of estimating organic matter. However, such
a method is subject to an error, due to the almost universal presence of char-
coal in the soils. Such an error is partly compensating, as the soils from
long-unburned areas contain noticeable quantities of charcoal. No satisfac-
tory means of eliminating this source of error have been worked out.

20 Florida Agricultural Experiment Station


Area Location of Depth Total Loss Replace
Number area sampled Plot of Nitroge on a pH
sample Ignition Calcium

Inches Percent Percent Percent

1 Summerville, S. C. Unburned 0-3 .137 8.358 .021 4.23
6-10 .037 3.116 .012 4.57
12-16 .021 3.161 .014 4.49

Burned... 0- 3 .101 6.317 .026 4.66
6-10 .024 2.463 .014 4.74
12-16 .020 2.430 .014 4.66

2 Ridgeland, S. C.... Unburned 0- 3 .169 9.449 .093 4.70
5- 9 .034 1.950 .029 5.17
12-18 .013 .886 .016 5.00

Burned... 0-3 .128 7.263 .123 5.58
5- 9 .027 1.946 .016 5.67
12-18 .008 .748 .013 5.67

3 Waycross, Ga..... Unburned 0- 3 .109 7.502 .025 3.72
6-10 .008 .233 .016 5.33
10-14 .038 2.237 .016 4.95

Burned... 0- 3 .093 5.447 .031 3.90
6-10 .014 .450 .016 5.33
10-14 .065 4.842 .016 4.53

4 Racepond, Ga..... Unburned 0- 3 .033 2.525 .022 4.15
6- 9 .007 .369 .016 5.17
11-14 .055 2.983 .018 5.00

Burned... 0- 3 .057 3.515 .031 4.40
6- 9 .015 .497 .019 5.04
11-14 .058 4.695 .021 4.99

5 Lake City, Fla.... Unburned 0- 2 .035 1.682 .015 5.37
5- 8 .016 .766 .008 5.92
18-20 .010 .470 .005 6.00

Burned... 0- 2 .035 2.115 .015 5.46
5- 8 .016 1.042 .005 5.62
18-20 .011 .710 .004 5.62

6 Starke, Fla........ Unburned 0- 3 .118 6.419
5- 9 .016 .495 .
12-16 .009 .196

Burned... 0- 3 .079 3.541
5- 9 .017 .440
12-16 .010 .170 .

7 Raiford, Fla....... Unburned 0- 2 .034 1.854 .013 5.37
4- 8 .017 .995 .... 5.29
18-20 .012 .634 .003 5.33

Burned... 0-2 .040 2.056 .014 5.33
4- 8 .018 .853 .005 5.46
18-20 .010 .613 .004 5.46

Bulletin 265, Effect of Frequent Fires 21

Area Location f Depth Total Loss Replace-
Number ara amnled Plot of Nitrogen on able pH
"area stamp sample Ignition Calcium

Inches Percent Percent Percent

8 Trenton, Fla...... Unburned 0-2 .032 2.249 .024 5.29
6- 9 .018 1.152 .011 5.54
16-18 .011 .815 .009 5.58

Burned... 0- 2 .040 2.925 .047 5.97
6- 9 .020 1.346 .010 5.67
16-18 .012 .990 .008 5.75

9 Chiefland, Fla.. . Not sample ed at this time.
10 Bartow, Fla...... Unburned 0- 3 .061 4.218 .064 5.17
10-12 .022 2.132 .033 5.67
24-28 .018 1.748 .024 5.71

Burned... 0- 3 .046 3.600 .059 5.67
10-12 .018 2.244 .035 5.71
24-28 .013 1.521 .028 5.75

11 Valparaiso, Fla.... Unburned 0- 2 .042 2.518 .020 5.00
6-10 .020 1.110 .016 5.25
28-30 .011 .681 .016 5.45

Burned... 0- 2 .031 1.905 .020 5.12
6-10 .016 .934 .016 5.20
28-30 .011 .626 .016 5.25

12 Adrian, Ga........ Unburned 0- 3 .033 2.606 .042 5.25.
6-10 .011 .914 .021 5.58
14-20 .005 .570 .016 6.01

Burned... 0- 3 .034 2.171 .027 5.75
6-10 .010 .806 .019 5.79
14-20 .003 .651 .017 5.71

13 Chipley, Fla....... Unburned 0- 2 .034 2.583 .021 5.54
7-10 .022 1.238 .010 5.71
14-17 .012 .871 .009 5.84

Burned... 0-2 .032 2.153 .021 5.92
7-10 .021 1.263 .011 5.84
14-17 .012 1.055 .009 5.97

14 Stapleton, Ala.... Unburned 0- 2 .064 3.874 .034 5.62
6-8 .042 3.388 .032 5.54
10-12 .030 4.110 .028 5.33

Burned... 0- 2 .070 5.122 .057 5.62
6- 8 .035 2.536 .025 5.37
10-12 .021 2.216 .025 5.37

15 State Line, Miss... Unburned 0- 2 .066 2.932 .016 4.32
5- 7 .046 2.092 .016 4.49
10-12 .034 1.855 .016 4.53

Burned... 0- 2 .092 5.423 .038 4.83
5- 7 .027 1.938 .017 4.75
10-12 .014 1.325 .018 4.87

22 Florida Agricultural Experiment Station


Area Location of Depth Total Loss Replace-
Number area sampled Plot of Nitrogen Igion able pH
sample Ignition Calcium

Inches Percent Percent Percent
16 Wiggins, Miss..... Unburned 0- 2 .060 3.166 .046 5.50
6-8 .021 1.392 .020 5.50
10-12 .013 .840 .018 5.50
Burned... 0- 2 .081 6.689 .050 5.50
6- 8 .025 1.649 .024 5.50
10-12 .022 2.435 .036 5.21
17 McNeill, Miss.... Unburned 0- 2 .061 4.147 .035 5.50
6- 9 .030 2.490 .016 5.58
10-14 .024 2.802 .016 5.21
Burned... 0- 2 062 4.516 .042 5.50
6- 9 .031 2.777 .021 5.45
10-14 .025 3.012 .021 5.37
18 Urania, La....... Unburned 0- 2 .074 4.807 .063 5.45
8-11 .034 2.283 .030 5.54
11-14 .030 3.047 .039 5.21
Burned... 0- 2 .086 5.223 .083 5.97
8-11 .036 2.476 .030 5.92
11-14 .029 2.729 .025 5.25

The purpose of the first collection, as has been mentioned, was
simply to detect any general trends in changes in soil composition
associated with fire treatment. The existence of such trends is
clearly shown by the data in Table II, which are summarized in
Table III. The heavy odds in favor of burned soils at the first
depth as to replaceable calcium and pH, 11 against 2 and 12
against 1, respectively, with results from three areas showing no
difference in each instance, are strong evidence that fire increases
the replaceable calcium content and decreases the acidity of the
soil. For the second and third depths these trends are much
weaker. The data for nitrogen and loss on ignition, even for the
first depth, reveal no indicative trend.
Although definite trends indicating an influence of fire were
found for the first soil depth; and, whereas, for replaceable cal-
cium and pH these trends persisted slightly for the second and
third depths, it is evident that the most pronounced changes in
the soil, ascribable to fire, occurred within the first few inches
beneath the surface. This was further borne out later by pH


First Depth, 0 to 3 Inches' Second Depth, 6 to 12 Inches' Third Depth, 24 to 30 Inches1
Difference No Unburned u d Dir.c
Unburned Burned De Unburned Burned No Unburned Burned Difference
Difference Difference Difference

Number Number Number Number Number Number Number Number Number -
of of of of of of of of of
Areas Areas Areas Areas Areas Areas Areas Areas Areas

Greater percentage of nitrogen.......... 7 9 1 8 8 1 8 7 2

Greater loss on ignition ................ 8 9 0 8 9 0 9 8 0

Greater percentage of replaceable calcium2 2 11 3 5 7 3 5 7 4

Higher pH .......................... 1 12 3 5 9 2 6 10 0

'These figures are only approximate; the respective horizons sampled varied considerably in depth beneath the
surface. .
2Tests were completely omitted for one area and for the second depth of one additional area.
3Tests were completely omitted for one area.

24 Florida Agricultural Experiment Station

determinations made for 12 individual samples taken at each of
three depths on the Trenton area. For the 0-2 inch depth, the pH
of the soil subjected to fire was 0.63 .0847 higher than that of
the soil protected from fire.
This difference is more than 7 times the standard error of the
difference. For the 6-8 inch and 16-18 inch soil depths, the
pH of the soil subjected to fire was only 0.10 .069 and 0.02
.054 higher, respectively, than that of the soil protected from fire.
These differences were not significant.
Although the soils studied in the preliminary phase of the work
varied widely in physical and chemical properties, no evidence
was found that these soils reacted differently to fire.

Upon completion of the preliminary work, eight study areas
were resampled in sufficient detail for the data from each area to
be treated statistically. The remaining areas sampled in the pre-
liminary work were not considered sufficiently comparable for
this purpose. The variability in chemical composition of the re-
spective soils to be sampled was not known. Accordingly, 12 in-
dividual samples were collected from each burned and unburned
portion of one area and 30 samples each from the burned and un-
burned portions of three other study areas, and on the basis of
analysis of these samples an estimate was made of the minimum
number of samples that would give reliable mean values for the
components determined chemically. The estimated numbers of
samples required for each of the four study areas are shown in
Table IV. These results apply to areas not greater in size than
1/4 acres. It was found that for these areas, with the exception of
the Adrian area, a total of 88 individual samples would probably
have been sufficient to show significant differences in nitrogen and
organic matter. For the Adrian area, the large number of
samples required was due not so much to greater variation with-
in this soil as to the small difference in the means. Table V in-
TStandard error of the difference.
8The areas sampled varied in size from 1% acre to 40 acres. Obviously
the work in sampling 40 acres conclusively would be prohibitive. There-
fore, all sample plots were restricted to % acre or less in size. Sample plots
were selected after examining the entire area, and in all cases the plots on
burned and unburned portions were practically adjacent.

Bulletin 265, Effect of Frequent Fires 25

dicates that the standard error of the difference for nitrogen and
for loss on ignition for this area compares favorably in magnitude
with those for the other areas.
On the basis of these calculations, 100 individual samples were
collected from the Trenton area. Upon analyzing these samples
and subjecting the results to statistical treatment, 160 individual
samples were then collected from each of the additional areas.
Thus, almost twice as many samples were collected, as indicated
by the estimates in Table IV, as a safeguard against a greater
variability within the areas to be sampled. The McNeill area was
again sampled because of particular interest attached to this area
with reference to additional experimental work in progress by the
Southern Forest Experiment Station. Results of the intensive
study are thus based on 30 individual samples from each burned


Individual Meanfor Mean for Difference Standard
Error Samples
Area Test Samples Unburned Burned Between of te Required
Collected Soil Soil Means Difference

Number Percent Percent Percent Percent Number
Florida.. Total
Nitrogen 12 .0403 .0464 .0061 .0052 71
Loss on
Ignition 12 2.368 2.691 .323 .3026 87
Florida.. Total
Nitrogen 30 .0480 .0528 .0048 .0026 82
Loss on
Ignition 30 3.270 3.755 .485 .1670 34
Georgia. Total
Nitrogen 30 .0360 .0370 .0010 .0019 1,037
Loss on
Ignition 30 2.358 2.252 .106 .1731 719
Miss... Total
Nitrogen 30 .0727 .0585 .0142 .0029 12
Loss on
Ignition 30 4.937 4.542 .395 .2184 83

1In these computations a difference was regarded as significant if it was
more than three times the standard error of the difference.

26 Florida Agricultural Experiment Station

and unburned portion of the Bartow and Adrian areas, 100
samples from the Trenton area, and 160 samples from each por-
tion of the remaining five areas.
To minimize analytical work, each of these sets of 160 individ-
ual samples was mixed to give 16 composite samples. The follow-
ing method of sampling was used: On each burned and each un-
burned area to be sampled a plot 80 feet square was laid out. This
was subdivided into 16 sub-plots 20 feet square. From each sub-
plot 10 individual samples were collected. These were thorough-
ly mixed, and approximately 6,000 grams9 of the composite were
sent to the laboratory for analysis. As the 80-foot sample plots
were divided into sub-plots regularly spaced in rows and columns,
it was possible to analyze the data in such a way as to eliminate
variation due to place correlation(8). Fig. 14 shows the meth-
od of laying out the sample plots; The locations of individual
samples are indicated only for sub-plot No. 1.

1. 4. S.
Z5 25 24 17 16 ? ,.
3. 7. 1A

31 26 Z3 18 1 I1 10 7 2
----- 80'
8 1'
0 27 22. 19 14 11 6

29 28 21 20 1] 12 ) 4

------------- ____ ________ ---
-so' o'
Fig. 14.-General method of laying out sample plots.

This method was used in all the later sampling except at Mc-
Neill. Because of the topographical variations and of the large
size of the McNeill area1, in order to get more representative
means than those based on one square sample plot for the entire

9Large samples were collected, so that sufficient material would be on
hand for studies in addition to those reported at this time.
IoThe approximate size of each of the areas sampled was as follows:
Adrian, 2 acres; Trenton, 40 acres; Bartow, 1 acre; Chiefland, 1 acre; Rai-
ford, 14 acre; Stapleton, 1 acre; McNeill, 10 acres; Urania, acre.


Area unsuited
CWZ WWWWWUI1JL~JW C I2JI/S becouseofob 'o,. U -
^ EaBBBBBBBB BB ^| BB1j7[[] I



Fig. 15.-Sketch map of the McNeill area. Sample plots are 20 feet square.

28 Florida Agricultural Experiment Station

area, a rectangular strip consisting of a series of 16 sub-plots
each 20 feet square was located on each side of the firebreak ex-
tending across the plots (Fig. 15). In the statistical handling of
the results the mean values of opposite sub-plots were compared
and tests of significance applied to the difference between means.
Because the preliminary results (Table III) indicated that
neither large nor consistent differences in the soils being com-
pared would be found at depths greater than 3 to 4 inches beneath
the surface, the detailed sampling was restricted to the surface
horizon. For the soils studied this is the upper portion of the
A horizon and is decidedly darker in color than the underlying
horizons. The organic discoloration usually extends from the top
of the mineral soil to depths varying from 3 to 6 inches. Each
individual sample collected represented a section of this entire
horizon (Fig. 13). Even within the small areas sampled, the
depth of this horizon varied considerably. The individual samples,
therefore, although taken from the same horizon, were collected
from the surface to varying depths.
The results of the intensive analysis are given in Table V, and
are also summarized in Fig. 16.
Seven of the eight soils from burned areas had more nitrogen
than the soils from the corresponding unburned areas; however,
the differences were greater by two standard errors for only two
study areas; viz., for Stapleton and Trenton.
The trend showing a greater loss on ignition for the burned
areas was not as strong as that of nitrogen, the score being five
cases out of eight. The Raiford area showed a significantly
greater loss on ignition on the unburned portion, whereas the
Trenton, Bartow and Stapleton areas showed significantly great-
er loss on ignition on the burned portions.
Seven of the eight areas showed a greater percentage of re-
placeable calcium on the burned portions, the differences being
significant in five of the seven cases. Only the Adrian area show-
ed a higher percentage of calcium on the unburned plot, and here
the difference was a slight one, well within the range of chance.
Consistently lower hydrogen-ion concentrations"-f. e. higher
pH values-were found for each of the eight burned areas. For

"1In the statistical analysis of the differences in acidity, actual hydrogen-
ion concentrations were used in the calculations instead of pH values. In
Table V, equivalent pH values are given for each hydrogen-ion concentra-


Individual Percent Actual Standard
Date Location of Depth samples Unburned Burned difference difference error
collected study area of collected from plot plot between between of
sample each plot means means differences
Inches Number Percent Percent Percent Percent Percent
May, 1932......... Adrian, Ga ............. 0-3 30 .0360 .0370 2.8 .0010 .0019 .
August, 1932....... Trenton, Fla............ 0-7 100 .0392 .0423 7.9 .0031* .0010
April, 1932........ Bartow, Fla............ 0-3 30 .0480 .0528 10.0 .0048 .0026 t
June, 1933......... Chiefland, Fla........... 0-4 160 .0343 .0346 0.9 .0003 .0010
June, 1933......... Raiford, Fla............ 0-6 160 .0302 .0285 5.6 + .0017 .0009
May, 1933......... Stapleton, Ala........... 0-4 160 .0626 .0712 13.7 .0086* .0015
June, 1933........ McNeill, Miss........... 0-3 160 .0692 .0721 4.2 .0029 .0034
June, 1933........ Urania, La.............. 0-4 160 .0752 .0785 4.4 .0033 .0017
May, 1932......... Adrian, Ga............ 0-3 30 2.358 2.252 4.5 + .106 .1731
August, 1932....... Trenton, Fla............ 0-7 100 2.282 2.719 19.1 .437* .0747
April, 1932........ Bartow, Fla............. 0-3 30 3.270 3.755 14.8 .485* .1670
June, 1933........ Chiefland, Fla........... 0-4 160 1.858 1.820 2.0 + .038 .0618
June,1933......... Raiford, Fla............. 0-6 160 1.669 1.511 9.5 + .158* .0572
May, 1933 ........ Stapleton, Ala........... 0-4 160 4.088 4.722 15.5 .634* .0851
June, 1933......... McNeill, Miss........... 0-3 160 4.912 5.019 2.2 .107 .2182
June, 1933.......... Urania, La............. 0-4 160 4.555 4.633 1.7 .078 .1299
May, 1932......... Adrian, Ga.............. 0-3 30 .0233 .0220 5.6 + .0013 .0027
August, 1932..... Trenton, Fla............ 0-7 100 .0291 .0380 30.6 .0089* .0028
April, 1932....... Bartow. Fla............. 0-3 30 .0310 .0473 52.6 .0163* .0035
June, 1933 ........ Chiefland, Fla........... 0-4 160 .0205 .0341 66.3 .0136* .0018
June, 1933......... Raiford, Fla............. 0-6 160 .0087 .0106 21.8 .0019* .0004
May, 1933......... Stapleton, Ala............ 0-4 160 .0164 .0329 100.6 .0165* .0019
June, 1933......... McNeill, Miss........... 0-3 160 .0403 .0436 8.2 .0033 .0034
June, 1933......... Urania, La.............. 0-4 160 .0599 .0682 13.9 .0083 .0063 M


Individual Percent Actual Standard
Date Location of Depth samples Unburned Burned difference difference error
collected study area of collected from plot plot between between of '2
sample each plot means means differences

Inches Number Percent Percent Percent Percent Percent

May, 1932......... Adrian, Ga.............. 0-3 30 .1999x10-5 .1325x10-6 .... -.0674x10-5*.+ 0170x10- .
August, 1932...... Trenton, Fla............ 0-7 100 .2866x10-5 .1421x10- .... -.1445x10I* .0235x10-5
April, 1932........ Bartow, Fla............. 0-3 30 .4145x10-5 .2032x10- .... -.2113x10-'* .0283x10-
June, 1933 ........ Chiefland, Fla.......... 0-4 160 .5258x10-6 .2906x10- .... -.2352x10-6* .0410x10-
June, 1933........ Raiford, Fla............. 0-6 160 .6948x10-6 .4689x10-5 .... -.2259x10-* .0520x10-6
May, 1933......... Stapleton, Ala .......... 0-4 160 .9346x10-5 .6544x10-5 .... -.2802x10-. .0513x10-
June, 1933.......... McNeill, Miss........... 0-3 160 .6276x10-5 .4105x10-5 .... .2171x10-j* .0419x10-S
June, 1933......... Urania, La.............. 0-4 160 .4743x10-6 .1576x10-5 .... -.3167x10-'* .0439x10-5

May, 1932 ........ Adrian, Ga.............. 0-3 30 5.70 5.88 .... -.18....
August, 1932 ...... Trenton, Fla .......... 0-7 100 5.54 5.85 .... -.31* .
April, 1932 ....... Bartow, Fla............. 0-3 30 5.38 5.69 .... -.31*
June, 1933 ........ Chiefland, Fla....... 0-4 160 5.28 5.54 .... -.26* ....
June, 1933........ Raiford, Fla............. 0-6 160 5.16 5.33 .... -.17* ...
May, 1933 ......... Stapleton, Ala........... 0-4 160 5.03 5.18 .... -.15* ...
June, 1933 ........ McNeill, Miss........... 0-3 160 5.20 5.39 .... .19* ...
June, 1933........ Urania........ ....... 0-4 160 5.32 5.80 .... .48* ....

*Significant differences, i. e., more than twice the standard error of the difference.
"**Corresponding to hydrogen-ion concentrations given above.

Bulletin 265, Effect of Frequent Fires 31

all areas the difference between means was greater than the
standard error of the difference by at least 4.
The results of the detailed sampling agree with those from the
preliminary study based on composites of four sub-samples in
showing no correlation between effect of fire and soil type. This

D* TwPtaf Loss on Reploceable -ydroqen io
n/Irogen iygni'on calc/urn cnce nvon

-^b-------+ --------------------

_+_______ _____ __ _____ ____
0o ../

-__ -/ 317 */ e


-6 .1- Q ,9 ./

Q -7
-8 ./0

-.3 .14

Fig. M.- suls of ned meo-,s / p irrn nnned an)
os in terms of on aj idiffs soaion o o o
iee with sinicancess results in r wn os-.
3,drion /o Chief/o,-x/ /* StaF /po/
0 f/Yaio'r //Bo/-So#ov /7 ,-/Vell
9 Trenton / /roaniao

Fig. 16.-Results of chemical analyses for 8 pairs of burned and unburned
plots in terms of difference of means in relation to standard error of differ-
ence with significance of results indicated.


Percentage difference in means'
Area Soil of unburned area between soils of '
burned and unburned areas

Silt Total Loss on Replaceable Total Loss on Replaceable
Sand and Clay Nitrogen Ignition Calcium pH Nitrogen Ignition Calcium pH

Percent Percent Percent Percent Percent Percent Percent Percent Percent I

Chiefland, Fla.......... 93.5 6.5 .0343 1.858 .0205 5.28 0.9 + 2.0 -66.3 -3.9
Trenton, Fla........... 93.1 6.9 .0392 2.282 .0291 5.54 7.9 -19.1 -30.6 -5.6
Bartow, Fla............. 91.9 8.1 .0480 3.270 .0310 5.38 -10.0 -14.8 -52.6 -5.5
Raiford, Fla............. 91.4 8.6 .0302 1.669 .0087 5.16 + 5.6 + 9.5 -21.8 -3.8 *
Adrian, Ga............. 89.6 10.4 .0360 2.358 .0233 5.70 2.8 + 4.5 + 5.6 -2.8

Stapleton, Ala........... 71.1 28.9 .0626 4.088 .0164 5.03 -13.7 -15.5 -100.6 -3.8
McNeill, Miss........... 68.3 31.7 .0692 4.912 .0403 5.20 4.2 2.2 8.2 -3.4
Urania, La............. 47.4 52.6 .0752 4.555 .0599 5.32 4.4 1.7 -13.9 -9.3

1Where mean is greater for burned than for unburned area, percentage difference is expressed as a negative quantity;
where mean is less for burned than for unburned area, difference is expressed as a positive quantity.

Bulletin 265, Effect of Frequent Fires 33

is apparent in Table VI, which shows widely varying differences
between the soils compared which were of similar texture and
fire history. Thus the percentage difference in total nitrogen
values for the burned and unburned soils of the first three areas
in Table VI, all of which were essentially alike in mechanical com-
position12, ranged from -6.0 to -10.0. Similarly, differences
in replaceable calcium for the McNeill and Stapleton soils, both
of which were classed as fine sandy loams, varied from -8.2 to
-100.6 percent.
A fair correlation appeared between soil type and the percent-
age of nitrogen, organic matter, and replaceable calcium, the
quantity of these constituents increasing in general with increase
in the silt and clay fraction. These findings may be summarized
by stating that no correlation was found between effect of fire
and character of soil, although the soils studied varied in both
physical and chemical composition, from deep sand of low fertility
to fine sandy and silt loams of much higher fertility.
As no correlation was found between effect of fire and physical
and chemical character of the soil, the eight areas sampled in de-
tail may be grouped and the results of each test averaged for the
burned and for the unburned portions (Table VII). In all tests


Mean Mean Difference Standard Odds2inl00
of of in error of that
unburned burned means1 the difference is
areas areas difference due to fire
Percent Percent Percent Percent

Total nitrogen........ .0493 .0521 -.0028 .0011 96
Loss on ignition....... 3.124 3.304 -.180 .1056 86
Replaceable calcium.. .0286 -0371 -.0085 .0024 99
H ions............... .520x10- .307x10- +-. 213x10-5 .0275x10-' 99

'Where burned mean is greater than unburned, difference in means is
expressed as a negative quantity; where burned mean is less than unburned,
difference is expressed as a positive quantity.
2Odds computed from Ezekiel's "Table A". (7)

12Mechanical analyses were made according to Bouyoucos' hydrometer
method (4).

34 Florida Agricultural Experiment Station

the means are higher for the burned areas. The probabilities
that the differences in the means are due to fire history and not
to chance are given in the last column of this table.

It is evident that certain changes in the chemical composition
of forest soils of the longleaf pine region are associated with fire.
No evidence was obtained indicating that fires deplete soil nitro-
gen or organic matter; on the contrary, an indication was found
that fires actually result in an increase in soil organic matter.
Likewise, strong evidence was found that an increase of total
nitrogen follows repeated fires. However, the differences both in
organic matter and in nitrogen were small. For replaceable cal-
cium the differences, on a percentage basis, were much larger
than for nitrogen and organic matter, as much as 101 percent
more being found on the burned areas. For the eight areas
sampled in detail replaceable calcium averaged 36 percent more
on the burned portions.
Although the differences found during this study are very small
in absolute units, the data indicate that in relation to the total
quantities of the individual materials present in the soil, these
differences may be relatively large. Furthermore, the importance
to plant growth of even small quantities of nutrients is well rec-
ognized by physiologists.
Since the higher replaceable calcium content of the soils of
burned areas may be attributed to the addition of ash following
fire, it is probable that the quantities of potassium, magnesium,
phosphorus and other constituents of ash are likewise somewhat
increased by fire.
The lower hydrogen-ion concentration of the burned areas,
also, may be largely attributed to the addition of ash. The avail-
ability of certain ions to the plant and the nature and activity of
the soil micropopulation are known to be influenced by hydrogen-
ion concentration. It has been observed, however, that longleaf
and slash pines can make rapid growth on soils having a pH as
low as 3.7; hence, the importance to tree growth of a pH difference
of even 0.48, as found for the Urania area, is doubtful.
Changes in the chemical composition of the soils studied that
could be attributed to fire were not detected below the 4- to 6-inch
soil depth. That the zone between that depth and the surface is
the one where such changes would exert maximum influence on

Bulletin 265, Effect of Frequent Fires 35

growth of longleaf and slash pine is suggested by the results of
studies made by the Southern Forest Experiment Station, in
which it was found that most of the feeding roots of these
species are located within this general zone.
The mellow, permeable surface soil on unburned areas con-
trasted strikingly with the compact, impermeable soil on the
burned areas. Although no volume weight determinations were
made for these two classes of soil, a difference was apparent. As
the chemical analyses reported in this study are based on soil that
has been sifted and ground, a slight error results. The point at
issue is not quantity of nutrients per unit weight of soil, but
quantity per unit volume. As the soils from the burned areas
appeared to be more compact than those from the unburned,
they probably had a higher volume weight. Therefore, a com-
parison of chemical analyses for soils from burned and unburned
areas, based on samples that had been sifted and ground, would
probably tend to minimize any differences that exist under actual
field conditions. Owing to the high variability in volume weight
determinations, as found for one area, the number of samples re-
quired to give a reliable mean was prohibitive at the time of the
present phase of the study.
The relation of inherent variations in chemical characteristics
of the soils being studied to differences in these characteristics
associated with fire history is of importance in the present prob-
lem. An illustration is afforded by the McNeill area, which was
sampled by the strip method, by which large variations in the
soil were to be expected (Fig. 15). The detailed results for each
sub-plot on this area are given in Table VIII. Had sampling been
restricted to the first 12 pairs of plots (a strip 300 feet long), no
difference would have been found between the burned and unburn-
ed plots in total nitrogen, loss on ignition, or replaceable calcium.
On the other hand, had paired plots 13 through 16 (a strip 100
feet in length) been sampled, large differences in favor of the
burned area would have been found. Lastly, had only paired
plots 9 through 12 been sampled, differences would have been
found in favor of the unburned area. Clearly, for the McNeill
area the inherent variations of total nitrogen, organic matter and
replaceable calcium in the soils being compared were greater than
differences associated with fire treatment, such differences being
based on difference in mean values of 16 composites of 10 sub-
samples. For pH, however, regardless of the grouping of sub-
plots, higher values were found for the burned area; only one

36 Florida Agricultural Experiment Station


Plot number Total Loss on Replaceable
nitrogen ignition calcium
Un- BurnedU B Un Un Burned Un- Burned Un- Burned
burned Burd burned n burned burned Bu b d Burned

1 17 .093 .072 5.994 5.125 .039 .034 5.41 "5.33
2 18 .070 .075 5.414 5.530 .042 .059 5.17 5.41
3 19 .082 .089 5.592 6.039 .059 .056 5.41 5.50
4 20 .085 .089 5.799 5.562 .058 .064 5.41 5.58
5 21 .081 .083 5.703 6.017 .055 .055 5.33 5.50
6 22 .075 .071 6.000 4.505 .050 .046 5.25 5.50
7 23 .077 .069 5.344 4.768 .047 .048 5.17 5.50
8 24 .070 .078 4.769 5.439 .041 .058 5.17 5.41
9 25 .069 .067 4.984 4.925 .038 .038 5.25 5.33
10 26 .069 .067 4.790 5.148 .031 .042 5.17 5.41
11 27 .073 .053 5.001 3.812 .065 .028 5.08 5.50
12 28 .053 .056 4.082 3.818 .021 .024 5.00 5.33
13 29 .062 .073 3.821 5.044 .024 .037 5.08 5.25
14 30 .050 .085 3.988 5.791 .025 .047 5.17 5.33
15 31 .051 .064 3.777 4.389 .025 .033 5.17 5.25
16 32 .048 .063 3.537 4.395 .025 .029 5.25 5.25

sub-plot out of 16 showed a lower pH value than the correspond-
ing unburned sub-plot.
As the fire history of many of the areas covered by this study
could not be accurately determined, it was not possible to cor-
relate degree of change in chemical composition of the soil with
number of fires. Since the increased replaceable calcium and pH
values of soils of burned areas are attributed to the addition of
ash resulting from fire, these differences must be related to dif-
ferences in quantity of vegetation that is burned. Density of
vegetation varies widely with soil type, being much greater on
heavy than on sandy types. It is somewhat surprising, there-
fore, to find no correlation between soil type and effect of fire as
indicated by Table VI. This may be accounted for in part by the

Bulletin 265, Effect of Frequent Fires 37

small number of areas representing each soil type. Had a great-
er number of areas of each type been studied, it seems likely from
theoretical reasoning, that a correlation would have been found.
However, other variables such as length of time between date of
sampling and date of last previous fire, and also occurrence of
rainfall would tend to obscure such correlation. No appreciable
change in total nitrogen or organic matter would be expected to
result from a single fire, however. The increase in nitrogen which
follows burning is attributable mainly to the addition of organic
materials to the soil through decay of roots of the grass vegeta-
tion characterizing burned areas. An appreciable change in the
quantity of nitrogen or organic matter resulting from burning
could not take place, therefore, until a considerable period after
the fire. Field observations indicate that such a change would
require at least 8 to 10 years.
The superiority of grass over forest vegetation in increasing
soil fertility has long been recognized by soil scientists. Jen-
ny(14), in a study of the nitrogen and organic matter in soils of
the central United States from Canada to the Gulf of Mexico,
found that grassland soils were uniformly higher in nitrogen and
organic matter than forest soils. Dunnewald(5) found the same
relationship in Colorado.
The Adrian area had not been subjected to so many fires as
had the other areas. The burned portion had been burned every
two or three years; at the time of the detailed soil collection, it
had not been burned for three years. The relative infrequency
of fires on this area probably accounts for the inconsistency in
the results for replaceable calcium (see Table V). Notwithstand-
ing the fact that the last fire had occurred three years previous
to the time of detailed sampling, an appreciable difference was
found in pH, as shown in Table V.
During this study, all areas sampled with the exception of the
Raiford area were located on virgin pine forest soils. This area
had been in cultivation until 25 or 30 years before the time of the
study; then it had been abandoned and had restocked naturally
to slash and loblolly pine. As F:g. 4 shows, this is the only area
out of the eight that showed a marked divergence from the gen-
eral trends in nitrogen and organic matter. The lower percent-
age of these constituents in the soil of the burned portion of this
area may possibly be attributable to the fact that in the longleaf
pine region abandoned cultivated fields do not have such dense
cover as virgin soils.

38 Florida Agricultural Experiment Station

It is desirable to call attention to several additional factors
which should be kept in mind when considering the results of the
present study. In the first place, it has not been possible, at the
present stage of the investigation, to draw definite
regarding the effect of fire on tree growth, as reflected .
in the soil due to fire. The physiological importance of the
changes in chemical composition of the soil ascribablt to fire is
entirely, unknown. Furthermore, it appears likely that certain
physical differences in the soil, such as permeability and moisture
relationships, may also be associated with fire; and it is possible
that these differences may more than offset the differences in
chemical properties reported at this time. Before final conclu-
sions can be drawn regarding the physiological effect of fire on
tree growth, through changes in the soil, many factors remain
to be measured and carefully evaluated. Until this is done, it will
be impossible to state whether fires in the longleaf region are
beneficial or detrimental to pine forest soils.

Tests for total nitrogen, loss on ignition, replaceable calcium,
and acidity were made on representative pine forest soils in the
longleaf pine region subjected to annual fires, and on comparable
adjacent soils protected from fire.
The soils subjected to frequent fires were found to be consis-
tently less acid, and to have higher percentages of replaceable
calcium and total nitrogen. An indication was found that these
burned soils were also characterized by larger quantities of or-
ganic matter as judged by loss on ignition.
Soils from burned areas showed pH values ranging from 0.15
to 0.48 units higher than those of unburned areas, whereas re-
placeable calcium totaled as much as 101 percent more on
burned soils than on the corresponding unburned soils. Differ-
ences in total nitrogen were small but significant, ranging up to
14 percent in favor of soils subjected to fire.
The observed changes in chemical composition of the soils,
ascribable to fire, were restricted to the top 4 to 6 inches.
Unburned areas studied were characterized by a layer of pine
needle litter from 2 to 3 inches deep. Except in openings in the
stands of pine, only a scant ground cover was present. On the
frequently burned areas only a small quantity of litter was pres-
ent, but a ground cover, consisting of wiregrass and a wide va-

Bulletin 265, Effect of Frequent Fires 39

riety of broadleaved herbaceous plants, including numerous mem-
bers of the Leguminosae, was typical.
Differeun1s found in nitrogen and organic matter are believed
differences in the forest floor and in ground cover
t; of burning. Changes in acidity and in replaceable
calcium can be attributed to the addition of ash following fire.


1. ALWAY, F. J., and ROST, C. 0. Effect of forest fires upon composition and
productivity of the soil. Pro. and papers 1st Int. Cong. Soil Sci. 1927.
2. BARNETTE, R. M., and HESTER, J. B. Effect of burning upon the accumu-
lation of organic matter in forest soils. Soil Sci. 29: 281-284. 1930.
.3. BENNETT, H. H. The soils and agriculture of the southern states. Mac-
millan. 1921.
4. BouYoucos, G. J. A comparison of the hydrometer and the pipette
method for making mechanical analyses of soils, with new directions.
Jour. Amer. Soc. Agron. 22: 747-751. 1930.
5. DUNNEWALD, T. J. Organic matter effects on soil. Jour. Amer. Soc.
Agron. 22: 713-718. 1930.
6. EHRENBERG, P. Die Bodenkolloide. Theodor Stein Kopff. 1915.
7. EZEKIEL, M. Methods of correlation analysis. John Wiley & Sons, Inc.
8. FISHER, R. A. Statistical methods for research workers, 4th ed. Oliver
and Boyd. 1932.
9. GWLMME, HANS. Investigations of the ammonia and nitrate production
in different types of humus and factors influencing these processes. Re-
port Norway For. Exp. Sta., 4 No. 1. 1932.
10. HARTSTEIN, -, Von englisechen und schottischen Diingerwessen. (Orig-
inal not seen.) 1853.
11. HESS, E. La sol et la foret. Mitteilungen der schweimerischen Central-
ansalt fiir das forstliche Versuchswesen, 15 No. 1. 1929.
12. HESSELMAN, HENRIK. On the effect of our regeneration measures on the
formation of saltpetre in the ground and its importance in the regen-
eration of coniferous forests. Meddelanden fran Statens Skogsforsoksan-
salt, No. 13-14. 1916-17.
13. HILGARD, E. W. Report on the geology and agriculture of the state of
Mississippi. 1860.
14. JENNY, HANS. A study on the influence of climate upon the nitrogen and
organic matter content of the soil. Mo. Agr. Exp. Sta. Bul. 152. 1930.
15. SALISBURY, E. J. The vegetation of the forest of Wyre: a preliminary
account. Jour. Ecol. 13: 314-321. 1925.
16. SNYDER, H. Effects of the rotation of crops upon the humus content and
the fertility of soils. Production of humus from manures. Minn. Agr.
Exp. Sta. Bul. 53. 1897.

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs