United States
Environmental Protection
Agency
Environmental Research
Laboratory
Corvallis OR 97333
Research and Development
EPA-600/S3-83-004  June 1983
Project  Summary
Acidity,  Nutrients,  and
Minerals  in Atmospheric
Precipitation  Over  Florida:
Deposition  Patterns,
Mechanisms and
Ecological  Effects
Patrick L Brezonik, Charles D. Hendry, Jr., Eric S. Edgerton, Randy L Schulze,
and Thomas L. Crisman
  A monitoring network of 21 bulk and
 4 wet/dry collectors located through-
 out Florida was operated to determine
 spatial and temporal trends in atmos-
 pheric deposition of acidity, nutrients,
 and minerals. During an intensive study
 year.  May 4,  1978 to April,  1979,
 nitrogen (N) and phosphorus (P) depo-
 sition via bulk precipitation averaged
 0.77 and 0.050 g/mz-yr, respectively.
 Highest deposition  rates occurred  in
 agricultural areas and lowest deposition
 rates occurred in coastal and forested
 areas.  Nutrient concentrations were
 higher during summer (convective) rains
 than in winter (frontal) events. Wet-only
 input accounted for most of the deposi-
 tion of inorganic N, but dryfall was
 more  important for organic N and
 especially for total  phosphorus (TP).
 Inorganic forms accounted for most of
 the N  and  P in  rainfall.  Statewide
 deposition rates of N and P were below
 the loading rates associated with eutro-
 phication, but the average N loading
 from bulk precipitation approached the
 mesotrophic criterion of Vollenweider,
 and N and P loadings exceeded  meso-
 trophic loading criteria at a few agri-
 cultural sites.
  The acidity of  Florida rainfall has
 increased markedly in the past 25 years,
 and concentrations of nitrate and sul-
 fate have risen correspondingly. Annual
 average pH values of less than 4.7 occur
 over the northern two-thirds of the
 state. Summer rain averaged 0.2-0.3
 pH units lower than winter rain, and
 excess sulfate levels were higher at
 most sites during summer. Sulf uric acid
 accounted for about 70% of the ob-
 served acidity, and nitric acid accounted
 for the remainder. Local (within-state)
 emissions of SO2 (and NOx) seem to
 control the acidity of Florida rainfall.
 The annual deposition  of H+ is about
 250-500 eq./ha over interior northern
 Florida, or about one-third to one-half
 the deposition rate for H* over the
 northeastern United States.
  Levels of pH in some softwater lakes
 of north Florida have declined by up to
 0.5 units over the past 20 years; no
 changes were observed in similar lakes
 of south-central Florida. Chlorophyll a
 and TP decreased with pH in a survey of
 20 softwater  lakes. Aluminum levels
 increased with decreasing pH, but ob-
 served maximum levels (100-150/ug/L)
 probably are not high enough to cause
 fish toxicity. The major change in phyto-
 plankton populations was a replace-
 ment of blue-green algae by green algae
 in acidic lakes. Species diversity and
 abundance also declined with decreas-
 ing pH, but the data exhibited much
 scatter. Some trends were noted in
 zooplankton and benthic invertebrates

-------
along a pH gradient, but both composi-
tion and abundance changes were rela-
tively subtle. Results indicate that acidic
conditions (as low as pH 4.6-4.7) do not
have major impacts on community
structure in Florida lakes.
  This Project Summary was developed
by EPA's Environmental Research Lab-
oratory, Corvallis.OR. to announce key
findings of the research project that is
fully documented in a separate report of
the  same  title  (see Project  Report
ordering information at back).

Introduction
  Although the chemistry of  rainfall has
been studied for  over a century,  many
questions remain about its importance in
biogeochemical cycles and in  transport of
pollutants from the atmosphere to terres-
trial and aquatic  ecosystems. Because
the nitrogen cycle involves several vola-
tile or  gaseous compounds,  the impor-
tance of atmospheric reactions and trans-
fers in  this cycle has been recognized for
a long time. Much less information is
available on levels of "rock-bound" nutri-
ents (e.g., phosphorus) in rainfall. More-
over, information in the literature exhibits
a wide variation in  concentrations both
spatially  and temporally  in rainfall.
Causes for this variability have not been
explained.
  The  role  of  rainfall  as a transport
mechanism for various pollutants such as
heavy metals and acidity has  been recog-
nized in recent years, and a considerable
volume of data has been assembled on
the pH of rainfall in  Scandinavia and the
northeastern United States.  The delete-
rious effects of acid rainfall on aquatic
systems in temperate climates also has
been documented.  Previous studies on
acid rainfall  in the United  States have
been skewed geographically to the North-
east, where  the problem apparently is
most severe. Little  information  is avail-
able on the extent of the problem in other
areas of the United States, the Southeast,
for example. Edaphic conditions in that
region, especially in  Florida, suggest  a
high susceptibility for deleterious ecolog-
ical effects;  and  demographic  patterns
suggest that the Southeast  may experi-
ence increasingly acidic precipitation in
the future.
  Most of the project results are based on
two large-scale field studies. The first, a
statewide sampling network  for bulk and
wet-only precipitation, was used to eval-
uate the importance of rainfall and dryfall
as  sources  of  nutrients, minerals, and
acidity to Florida ecosystems. The net-
work was established to allow analysis of
the influence of surrounding land-use-
patterns on deposition rates  of these
substances. Samplers were located in
urban, agricultural, forested, coastal, and
in pristine areas; and transects were
established to evaluate north-south and
east-west  (coastal-inland) gradients in
deposition  patterns (Figure 1). The net-
work  provided valuable information on
nutrient and mineral deposition patterns,
and yielded the first comprehensive anal-
ysis of the acid  rainfall problem in the
state of Florida.
  The second field effort  involved  a
sampling program on 20 softwater lakes
in north-central  and south-central  Flor-
ida. Routine limnological measurements
and complete chemical and biotic analy-
ses were done on each lake to evaluate
the effects of acidification. Phytoplankton,
zooplankton  and  benthic  invertebrate
communities were analyzed for species
diversity and abundance. The results of
the project are summarized according to
the three major phases of the project: (1)
atmospheric deposition of nutrients; (2)
the spatial and temporal distribution of
rainfall acidity in Florida, and (3) the
effects of acidification on chemical anc
biological conditions in soft-water lakes
of Florida.

Atmospheric Loadings of
Nutrients
  Bulk  precipitation is an  important
source of nitrogen for both terrestrial anc
aquatic  systems. Average annual  load-
ings of total nitrogen (TN) for the 21 bulk
sampling  sites  ranged from 0.3 to 1.3
g/m"2-yr during thetwo-year study period
with a statewide mean of 0.72 g/m2-yi
(Figure 2). About 70% of the nitrogen was
inorganic (ammonium and nitrate) anc
thus  was readily  available  to  plants
Deposition rates were  highest in rura
agricultural areas  (0.88  g/m2-yr) anc
lowest along the coast (0.58 g/m2-yr). Ir
comparison with critical loading rates foi
lake eutrophication, the annual depositior
of TN at all 24  collection sites (Figure 2
was below the values associated with
eutrophic conditions (assuming nitroger
was the limiting nutrient). The statewide
average deposition rate was about 75% ol
the "permissible"  loading for shallow
lakes suggested by Vollenweider (1968)
      KEYS TO STATIONS

      AB  Apopka (wet/dry)
      BA  Bradenton
      BG  Belle Glade (wet/dry)
      BH  Bahia Honda
      BN  Bronson
      CH  Chipley
      CK  Cedar Key (wet/dry)
      CL  Clewiston
      CW  Corkscrew Swamp
      FM  Fort Meyers
      GV  Gainesville (wet/dry)
      HA  Hastings
      JA  Jacksonville
      JS  Jasper
      JY  Jay
      LA  Lake Alfred
      LP  Lake Placid
       LI  Lisbon
      MC  MacArthur Farms
      Ml  Miami
      ML  Marine/and
      ST  Stuart
      TA  Tallahassee
      WD  Waldo
                                                                    BH
Figure 1.    Location of sampling stations in the Florida Atmospheric Deposition Network (FADNJ.

-------
deposition at several agricultural  loca-
tions slightly exceeded the permissible
criterion but none approached the eutro-
phic loading criterion.
  Bulk precipitation has significant levels
of total phosphorus (TP); the statewide
two-year average was 33 jug/L, and the
mean deposition rate was 48 mg P/m2-
yr. Land use had an important effect on
atmospheric  deposition  rates for "TP.
Rural (non-agricultural) and coastal sites
had the lowest rates (27 and 31 mg P/m2-
yr,  respectively), and  agricultural  sites
had the highest permissible loading rate
for the lakes that are most vulnerable to
eutrophication,  i.e.  shallow lakes  with
long hydraulic residence times. Figure 2
compares annual bulk deposition rates of
TP  for the rainfall  network sites to the
phosphorus loading criteria proposed by
Vollenweider (1975), according to whom
the critical (i.e., eutrophication-causing)
rogen
- 2.0
5*15
*E
\
^
^1
' ^
•3 0.5






fl—
k
.u


rr~T"
ti
AP | flG I fi/V

-
r
—
b
-


CHlCW
hi

I
GV
-n-

i



j/»
-
-


jy
--


Phosphorus
-




-
-

1
/.p
	 > 	 n.
f| [In f]|| p
llllill
wcT/w/ frxi T
400 -
300 «; -
E
^N
Q»
200 E -
'S
100 §-
E(5)
Ed)



P(S)

""
             BH  BA   CK   CL  FM  HA  JS  LA   LI  ML  ST  WD
                                       Site


Figure 2.    Loadings of TN (open bars) and TP (closed bars) by bulk precipitation at each FADN
            site. Statewide average loadings shown by dashed lines.  Permissible loading
            criteria  (P) and excessive loading criteria (E;  i.e.,  inducing eutrophy) for TP
            are for two values ofareal water loading (q, - 1 and 5 m/yr), as given byVollenweider
            (1975). Loading criteria for TN are from Vollenweider (1968) for lakes with mean
            depth < 5 m.
loading rate is  a function of the area!
water loading rate, q*.  For most Florida
lakes q. is in the range of 1 -5 m/yr. Based
on Vollenweider's loading criteria appro-
priate to this range of qs, bulk precipita-
tion in Florida supplies only 12-16% of
the loading required to  induce eutrophic
conditions.
   Average concentrations of TP in sum-
mer rainfall were about 1.5 times as high
as those in winter rainfall. Whereas most
of the atmospheric deposition of inorganic
N and TN occurs via rainfall rather than
dryfall, the opposite is true for phospho-
rus. At the four sites with wet-only/dry-
fall collectors, wet deposition accounted
for an average  of only 20% of the TP.
Thus, most of the phosphorus  in bulk
precipitation  is dryfall, presumably of
large wind-blown particles of dust and
soil  that are not transported large dis-
tances.
   Average concentrations of total sulfate
in rainfall ranged from  about 0.2 to 1.2
mg/L (as S); about two-thirds of the total
deposition  of sulfate  was by  rainfall.
Deposition rates (Figure 3) ranged from 3
to 23 kg S/ha-yr (for all sites over the
two-year study),  indicating that the  at-
mosphere is a significant source of sulfur
to soils in Florida. Except at sites very
close (a few km) to the coast, most of the
sulfate in Florida rain is excess SO2" i.e.,
sulfate derived from SOz. Sea salt aero-
sols contributed only about 0.5 kg/ha-yr
of sulfate-S to bulk deposition at inland
                                                 ••"3.2
                                                                                                               2.03
Figure 3.    Annual deposition (kg/ha) of (A) excess sulfate-S and (B) sea-salt sulfate-S across Florida for period May 1978- April 1979.

                                                                                   3

-------
sites. Because formation of excess sulfate
from 862 in the atmosphere concomitant-
ly results in stoichiometric production of
hydrogen ion (SOZ + YiQz + H20 — 2H* +
SOD, the occurrence of excess sulfate in
Florida rainfall also has important implica-
tions regarding the acidity of the rain.
Acidity of Precipitation
  Rainfall throughout  Florida  is acidic
(Figure 4), with average values for all but
a few stations in south Florida being less
than geochemical neutrality (pH~5.7).
Single  events as low  as pH 3.9 were
measured at Gainesville, and the lowest
pH  of a bulk precipitation  sample (col-
lected weekly or biweekly) was 3.73 (at
Jay in the western panhandle during
August  1978). A definite geographic
pattern exists for acid deposition in bulk
precipitation around the state; mean
annual pH values (volume-weighted) for
1978-1979 were around 4.6-4.7 through-
out the panhandle and the northern two-
thirds of the peninsula.  Mean annual
values  south  of Lake Okeechobee were
around 5.0  or above.
  Neutralization of bulk precipitation was
found for  coastal sites, but wet-only
precipitation collected near both the Gulf
and Atlantic coasts was approximately as
acidic as inland stations of comparable
latitude. Partial neutralization of acidity in
coastal rain apparently results from dry
deposition of alkaline particles containing
calcium carbonate  of local  (terrestrial)
origin. Analysis of the ionic composition
of coastal bulk precipitation indicates that
sea spray is not the agent of neutraliza-
tion. Sea-salt sulfate levels were only
modestly  elevated compared to inland
bulk precipitation,  and the calculated
amount of sea-salt calcium carbonate is
too low to account for the neutralization.
  A seasonal pattern was found in precip-
itation acidity throughout the state with
summertime pH averaging 0.2-0.3 units
lower than wintertime values. Possible
reasons for the  difference  include  (1)
increased summertime emissions of S02
and  NOx (caused in  part by seasonal
demands for air conditioning); (2) greater
thunderstorm activity in summer, result-
ing in greater fixation of N0« by lightning;
(3) enhanced  scavenging efficiency of
summer convective showers compared to
winter frontal storms; and (4) differences
in the frequency and size of individual
rain events between summer and winter.
Further  studies are needed  to evaluate
the importance of these factors.
                                                         (4.7)
(4.7)-
                                                                          (5.0)
                                                          •-••5.45

Figure 4.    Volume-weighted mean pH of precipitation throughout Florida,  May 1978-April
            1979.
  In spite of the north-south gradient of
decreasing rainfall acidity, long-range
(interstate) transport of acid precursors is
not a wholly satisfactory explanation for
acid precipitation in Florida. A substantial
portion of the H2S04 and HNOs must be
derived from in-state emissions of  SO:
and NOx, which are widespread  and
substantial. These conclusions are sup-
ported by several lines of evidence, includ-
ing the fact that summer rainfall through-
out the entire state is more acidic than
winter rainfall.  From a  meteorological
viewpoint, peninsular Florida is isolated
from the  rest of the United States during
summer. Large-scale weather patterns
for the peninsula come from the south-
east (Caribbean) or southwest  (Gulf of
Mexico) during summer, and cold fronts
from the  north rarely penetrate the state
during this period.
  Granat-type analysis  indicates  that
about 70% of the rainfall acidity in Florida
is derived from sulfuric acid and 30%
from nitric acid. A multiple regression
equation with  [H+] as the dependent
variable  and  [S042lxs  and [Ca2+]*s as
independent variables explained about
75% of  the  variance in  hydrogen ion
concentration over  the  statewide  net-
work: [Hi =  6.1  + 0.54[SO42~]«s - 0.35
[Ca2+]xs, where xs refers to the fraction ol
the ions  of non-marine origin. Thus the
pH of rainfall in Florida primarily reflects
the degree  to which sulfuric acid has
been  neutralized by terrestrial calcium
carbonate.


   Bulk precipitation throughout northerr
 and central  Florida deposited  250-500
 equiv HVha-yr during 1978-1979, which
 is about one-third to  one-half  of the
 annual  deposition of H+  in the heavily
 impacted northeastern  United States
 Comparable values for excess sulfate are
 7-11  kg/ha-vr  in northern Florida anc
 ~13 kg/ha-yr in the northeastern  U.S
 (the 10-year average for Hubbard Brook,
 N.H.). Thus Florida  ecosystems receive
 50-90% of the  excess sulfate from the
 atmosphere as their northern  counter-
 parts. Although historical data are lacking
 on the pH of  Florida rainfall, calculatec
 values for rainfall pH during  the  mid-
 1950's  indicate wet-only precipitatior
 was  not acidic at that  time  (Table 1)
 Moreover, present values of sulfate depo-
 sition in northern Florida are up to foui
 times higher than values for the earl\
 1950's.  Scattered information for the pl-
 of rainfall at Gainesville from 1973 tc
 the present, however, do  not show an^
 long-term trends.

-------
 Table 1.    Comparison of Parameters Related to the Acidity of Florida Rainfall in 1955-56 and
           1978-79

                                  Weighted-Mean Concentration (ueq/L)
Location
Mobile
Tallahassee
Jacksonville
Tampa/Bradenton
W. Palm Bch. /Stuart
Mean
1978-9/1956

1956
<2.5
<2.5
<2.5
<2.5
<2.5
<2.5

rf
1978-9^
24.0
17.4
18.3
20.1
6.9
17.3
>8.4
Excess-SOl
1956*
16.0
18.8
27.9
28.8
13.5
21.0

1978-91
34.7
33.0
4^.5
36-4
20.1
33.5
7.6ft
7956
2.6
2.9
2.9
2.7
4.1
3.1

A/Oa
* 7978-9f
13.9
13.9
16.2
14.3
12.1
14.1
4.511
  * Data from Junge and Werby 11958) and Junge (1958); proton concentration inferred via
   anion/cation balance.                                    ,


  t This study; May 1978-April 1979.

tt Present data are for bulk precipitation; 1956 data are for rainfall-only. Adjacent wet-only (W)
   and bulk (B) collectors at Gainesville in this study yielded the following volume-weighted
   average concentrations fin fjieq/IJ: excess sulfate, 35.11B) and26.6 (W), B/W=1.3; nitrate, 16.9
   (B) and 13.6 (W). B/W1.24. Differences in collector type thus do not wholly explain the increases
   in concentrations.
 Effects of Acid
 Precipitation on
 Florida Lakes
  A large  number of soft-water  lakes
occur in the sand-hill highlands region of
peninsular Florida. Based on comparison
of historical and current data for 13 such
lakes in northern Florida and 7 soft-water
lakes in  south-central  Florida, pH has
decreased  by up to about 0.5 units in
some of  the  northern soft-water lakes,
whereas no  temporal trends could be
discerned for the southern group. Corre-
sponding decreases  in alkalinity  and
increases in sulfate concentration  were
observed in  the  northern  lakes.  The
northern (Trail Ridge) lakes lie about 40-
50  km east of Gainesville  in a region
receiving rainfall with a (volume-weight-
ed)  average annual pH of 4.5-4.7. The
southern (Highlands  Ridge)  lakes  lie
northwest of Lake Okeechobee, near the
current southern terminus of pronounced
rainfall acidity. The 20 lakes surveyed had
annual average pH levels ranging  from
4.72 to 6.80, but otherwise had generally
similar characteristics (soft water, oligo-
trophic to mesotrophic nutritional condi-
tions). The  group thus served as a  good
data base to evaluate the effects of acid
precipitation on vulnerable aquatic eco-
systems in Florida.
   Ageneraltrend of increasing aluminum
with decreasing pH was found in the 20
Florida lakes (Figure 5). However, maxi-
mum values (100-150 ug/L) were below
the  levels associated with fish toxicity;
this may explain the occurrence of large-
mouth bass and several other common
game fish species in lakes with pH values
below 5.0. Aluminum concentrations are
generally  low in Florida's sandy soils, but
further studies are needed to evaluate the
impact of acid precipitation on leaching of
aluminum from these soils to Florida's
soft-water lakes.
   A general trend of increasing chloro-
phyll a concentration with increasing pH
was found. However, total  phosphate
concentration also tended to rise with pH
(Figure 6). The trend of greater oligo-
trophic conditions in  more acidic lakes
may be caused  by lower rates of nutrient
cycling  at lower pH, or it may reflect
watershed nutrient loading factors that
just  happened to correlate with lake pH.
Further studies are needed on this point.
  The number of phytoplankton species
and their abundance in a lake decreased
with increasing acidity, but much scatter
occurred for both parameters. Although
the data are limited, a trend of increasing
phytoplankton abundance with increas-
 ing pH was found for a series of lakes with
 similar levels of  phosphate, the lake
 survey also indicated that species compo-
 sition varied along a pH  gradient, with
 green algae replacing blue-greens at low
 pH. (Figure 7). In lakes with pH values of
 4.5-5.0, 60% of the algae were green
 (Chlorophyta), and 25% were blue-green
 (Cyanophyta). Corresponding values for
 lakes in the pH range 6.5-7.0 were 31 %
 green algae, 63% blue-green algae.
   Similar trends were found in the zoo-
 plankton; i.e.,  slight decreases  in the
 numbers of species and individuals with
 pH, but the trends exhibited considerable
 scatter. In general, the number of zoo-
 plankton species found at a given pH was
 greater than the number found in temper-
 ate lakes of comparable pH. Six species of
 zooplankton were dominant at  all  pH
 levels, and five other species were always
 present but never dominant. Two types of
 multivariate analysis (principal compo-
 nent  and cluster analysis) showed that
 the zooplankton populations  could  be
 grouped along  pH gradients,  but the
 population differences with pH are rela-
 tively subtle. Rare species showed greater
 differences  with pH than did common
 species. Some of the observed changes in
 zooplankton community structure may
 not have been directly related to changes
 in pH but  rather to  changes in overall
 trophic conditions that varied somewhat
 along the pH gradient. Acidic lakes tended
 to be less productive, were more nutrient-
 depauperate, and had low standing crops
 of phytoplankton. Whether these  trends
 were  caused by or merely correlated with
 pH cannot be determined from the survey
 data,  and further  experimental  studies
 are needed to evaluate this matter.
   No clear trends were seen in either the
 diversity of the abundance of benthic
 invertebrates with pH, and the differences
 that were found among the lakes may
 reflect differences in trophic conditions
 and substrate type more than  direct
 effects of pH.
   Overall, pH appears to have relatively
 small effects (in the range 4.7-6.8)  on
 community structure in soft-water Florida
 lakes  that otherwise have similar chemi-
 cal composition. More dramatic effects
 may occur, of course, under more acidic
 conditions  (pH <  4.7), and significant
 changes may occur in community metab-
 olism, productivity, and nutrient cycling
 processes within the pH  range of the
 lakes  included in  the survey. Further
 studies on the biological effects of acidi-
fication on Florida lakes should be direct-
ed at these processes.

-------
     750 |-
     125 -
     100  -
 I
a
75 -
      50 -
      25 -

-




-




McCloud t—





1
1
1
Acidic Lake Group \ Nonacidic Lake Group
i


••

Cowpen





flaliloo I—
"*


Brooklyn » [ 	


1 , 1 i! lilt
| O -j^L^t^etj
                          5.00
                                                  5.50
                                                                    6.00
6.50
Figure 5.    A verage and range of labile aluminum in the 20 survey lakes vs. pH.
References
Junge,  C.E.  1958.  The distribution  of
  ammonia and nitrate in rain water over
  the United States. Trans. Amer. Geo-
  phys. Union 39:241-248.
Junge, C. E. and R. T. Werby. 1958. The
  concentration of chloride, sodium, po-
  tassium, calcium, and sulfate in rain-
  water over the United States. J. Mete-
  orol. 15:417-425.
Vollenweider, R. A.  1968. Scientific fun-
  damentals of the eutrophication  of
  lakes and  flowing waters, with partic-
  ular reference to nitrogen and phos-
  phorus as factors in  eutrophication.
  Org. Econ. Cooperation and  Develop-
  ment, Paris. Rept. DAS/CSL/68.27.
Vollenweider, R. A. 1975. Input-output
  models with special reference to the
  phosphorus loading concept in limnol-
  ogy. Schweiz. Z. Hydrol. 37:53-84.

-------
Figure 6.    Three-dimensional plot of chlorophyll a as function of pH and TP in the 20 survey
             lakes.
3
I
3
•Q

-------
     Patrick L Brezonik, Charles D. Hendry. Jr., EricS. Edgerton, Randy L Schulze, and
       Thomas L. Crisman are with the University of Florida, Gainesville, FL 32611.
     Charles F. Powers is the EPA Project Officer (see below).
     The complete report, entitled "Acidity, Nutrients, and Minerals in Atmospheric
       Precipitation Over Florida: Deposition Patterns, Mechanisms and Ecological
       Effects," (Order No. PB 83-165 837; Cost: $16.00, subject to change) will be
       available only from:
             National Technical Information Service
             5285 Port Royal Road
             Springfield, VA 22161
             Telephone: 703-487-4650
     The EPA Project Officer can be contacted at:
             Environmental Research Laboratory
             U.S. Environmental Protection Agency
             200 SW 35th Street
             Corvallis, OR 97333
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
                  SSS
                    HICAGO  IL  60601

-------