Published September, 1996
TURFGRASS SCIENCE
Bermudagrass Carbohydrate Levels as Influenced
Fertilization and Cuitivar
GradyL. Miller* and Ray Dickens
ABSTRACT
Storednonstructural carbohydratesarea primary sourceof reserve et al. (1991) indicated that starch or sucrose levels
energyin vegetative organsof warm-season
grasses.Research is needed stolons did not reduce lethal low temperatures for ’Flo-
concerning the influenceof potassiumoncarbohydrate levels in turf- ratam’ St. Augustinegrass [Stenotaphrum secundatum
grasses.FourKrates rangingfrom24 to 390kg ha-~per growing
month wereappliedtwicemonthly from1992to 1994to ’Tifdwarf’
and’Tifway’ bermudagrasses[Cynodondactylon(L.) Pers.× C. trans-
vaalensisBurttDavy]field grown onUcheeloamy sand(loamy,sili-
ceous,thermicArenicHapluduit) andsand-peat (9:1 byvolume)
dium.Root÷ rhizomeandshoot+ stolon sampleswereevaluated
monthly for total non-structuralcarbohydrate (TNC) concentrations (1965)
by means of nearinfraredreflectancespectroscopy. TNCconcentra- September
tions variedamong cultivars,plantparts,andtimeof year.Tifdwarf and rhizomes. Goatley et al. (1994) indicated no differ-
produced 30%greaterroot + rhizomeTNC concentrationand55%
greatershoot+ stolon concentration thanTifway.HigherKrates
eitherdidnotinfluenceor reduced root+ rhizome TNC concentrations
compared to lowKrates. Correlation analysisindicatedthatroot+
rhizomeyields andTNC concentrationswerenegativelyinfluenced
by temperature. Carbohydrate androot + rhizomeyield responses
suggestthereis nobenefitto applying Kat ratesbeyond thosethat
providesufficientKlevels for normal growth.
B ER~UOAGRASS is one of the more widely used grasses
on athletic fields and golf courses in the warmhumid
and warmsemiarid regions of the world (Beard, 1973).
The application of high rates of K has been suggested
to turf growers to increase bermudagrasswinter hardiness
(Beard, 1990; Gilbert and Davis, 1971). Studies have
indicated that total nonstructural carbohydrates (TNC)
mayserve an important role in the freezing tolerance of
manyplants (Levitt, 1980). Although the mechanism
which carbohydrates affect freezing stress tolerance is
not well understood (Alden and Hermann, 1971), many
plant species exhibit poor freezing tolerance whencarbo-
hydrate reserves are low.
Research on nonstructural carbohydrates of bermu-
dagrass in the southeastern USAhas been limited. In
Missouri, Dunn and Nelson (1974) indicated that TNC
levels in bermudagrass increased during fall as harden-
ing-off of turf occurred and then decreased during winter.
Rogers et al. (1975) noted that in the morecold tolerant
’Meyer’ zoysiagrass (Zoysia japonica Steud), starch lev-
els remained high during the autumn and winter in Mis-
souri. Goatley et al. (1994) found no differences in TNC
levels of Tifgreen bermudagrass (Cynodon dactylon
G.L.Miller,Environmental
Horticulture
Dep.,Univ.of Florida,Gaines-
ville, FL32611-0670;
andR. Dickens,Agronomy
andSoils Dep.,Auburn
Univ.,AuburnUniversity,AL36849-5412.
Contribution
as JournalSeries
Paperno. R-04957of the FloridaAgric. Exp.Stn. Received5 April
1995.*Corresponding
author(glmi@gnv.ifas.ufl.edu).
Publishedin CropSci. 36:1283-1289
(1996).
1283
by Potassium
C. transvaalensis) measured in Decemberor March. Fry
(Walt.) Kuntze]. Few studies have evaluated bermu-
dagrass TNClevels in response to temperature on a
12-mo basis.
Limited information has been reported on the influence
of high K rates on TNClevels. Thompson and Ward
reported that applications of K in March and
did not influence TNCin bermudagrass roots
ences in TNClevels of Tifgreen bermudagrass following
late-season applications of either 0, 41, or 82 kg K ha-~.
The objectives of this study were to: (i) quantify TNC
in Tifdwarf and Tifway bermudagrasses throughout the
year; (ii) determinethe influence of a high K fertilization
rate on carbohydrate concentration in these two cultivars;
and (iii) determine whethersignificant correlations exist
between TNCconcentrations and air temperature.
MATERIALS AND METHODS
A2-yr study wasinitiated at the AuburnUniversityTurfgrass
ResearchCenter, Auburn,AL,in the spring of 1992. Tifdwarf
and Tifwaybermudagrasseswere established on Ucheeloamy
sand (loamy,siliceous, thermic Arenic Hapludult)and a sand-
peat (9:1 by volume) medium.These experiments utilized
completely randomizeddesigns, with four replications of
4.2-m2 plots. Treatmentsconsisted of two cultivars and two
rates of K fertilization. Thehigh Krate consistedofa combina=
tion of 195 and 390 kg K ha-~ growingmonth-~ rates on the
sand peat mediumor 98 and 195 kg K ha-~ per growingmonth
on the loamysand. The low K rate consisted of 24 and 49 kg
K ha-~ per growingmonthrates on the sand peat mediumor
12 and 24 kg K ha-~ per growing monthon the loamysand.
TwoK rates were combinedto form one to ensure adequate
plant material was available for carbohydratetesting, since
early in the study it was determinedthat plot size was not
adequateto allow sufficient samplingfroma single K rate on
a monthlybasis. Therefore tissue samples for each K rate
were combinedon an equal volumebasis to ensure adequate
samplefor analysis. Potassiumas KCIand urea (25 kg N ha-~
on the loamysand and 49 kg N ha- ~ on the sand-peat medium)
were applied twice monthlythroughout the growingseason.
Data from the two plant growing media were analyzed sepa-
rately since N andKrates differed. Initial extractableKconcen-
trations were 36 and 15 kg ha-~ for the loamy sand and
sand-peat medium,respectively. Based on AuburnUniversity
Soil testing reports, these K concentrationswere lowfor the
loamysand and very low the for the sand-peat medium.Phos-
phorous was applied to the sand-peat mediumat 49 kg ha-~
in April each year accordingto AuburnUniversitySoil testing
Abbreviations:
TNC,total nonstructuralcarbohydrate.
1284 CROPSCIENCE, VOL 36,
5O
Maximum
~.0
o,,_,
30
o~ 20
E Performance coefficient of determination
._.. lO
o
-lO
ASONDJ FMAMJ d ASONDJ FMAMJ d A 0 dard error of performance for the samples reserved from
1993 1994
1992
Month
Fig. 1. Maximum and minimumweekly-air temperatures during the
1992-1994 sampling periods at Auburn, AL.
recommendations.Neither loamy sand nor sand-peat medium
required amendment to maintaina soil pHof 6.0 (___0.2).
Irrigation during the summermonthswas applied as needed
to maintainhigh quality turf. Weedcontrol was obtained with
yearly applicationof 0.56 kg ha-~ dithiopyr [3,5-pyridinedicar-
bothioic acid, 2-(difluoromethyl)-4-(2-methylpropyl-6-(tri-
fluoromethyl)-S,S-dimethyl ester] in mid-March, an application
of 2.24 kg ha-t MSMA [monosodiumacid methanearsonate]
in August1992and 0.56 kg ha-~ bromoxynil[3,5-dibromo-4-
hydroxybenzonitrile] in mid-Marchin 1994. Tifdwarf was
clipped four to five times per weekat a 5-mmcutting height
while Tifway was mowedfour to five times per week at a
12-mmcutting height. Plant clippings were removedwith each
mowing, Maximumand minimumdaily and mean maximum
and minimum weeklyair temperatures at 1 m above the soil
surface during the study were recorded (Fig. 1).
Tissue samplesfor TNCanalysis were taken at 1330h (+30
min) on the 17th day (+4) of each month by removing
6-cmdiameter by 15-cm-lengthplug from a randomlyselected
location within each plot. Plugs were washedfree of soil and
all above-groundplant material was excised from the top of
the plug with scissors at the thatch-soil interface. Stolons
were included with the above-groundportion of the plant and
rhizomeswith the below-ground portion. After separation the
tissue samples were weighed, immediately dried in an oven
for 24 h at 100°C, reweighed for yield determination, and
placed in a freezer at -20°C until TNCanalysis. Prior to
TNCdetermination, samples were ground in a cyclone-type
samplemill to pass through a l-ram screen.
Total nonstructural carbohydrate content was determined
using a Model6250 NIRmonochromator(Pacific Scientific,
Inc., Silver Spring, MD~) in conjunctionwith a personal com-
puter with software as described by Shenk(1985). The near
infrared reflectance spectroscopy(NIRS)was calibrated with
samplesthat wereanalyzed for carbohydrateconcentration by
the starch acid-hydrolysis technique described by Whistler
and Wolfrom(1962). Stepwise regression calibrations were
~Trade names are included for the benefit of the reader and do not
imply endorsementby the Univ. of Florida, AuburnUniv., or that other
equipmentis not suitable.
SEPTEMBER-OCTOBER
1996
Table 1. Calibration and performancestatistics for near infrared
reflectance spectroscopic (NIRS) determination of carbohydrate
concentration in Tifdwarf and Tifway bermudagrasses.
"R2~ r2:[: tl§ Mean SE Slope¶
g kg -I --
Calibration 0.86 - 275 126.2 22.7 -
Performance - 0.81 54 125.1 21.8 1.09
Calibration coefficient of determinationachieved in forwardstepwise
multiple regressionof spectraon data.
achievedin linear regression of
NIRS-predicteddata.
Numberof samples.
Slope calculated in regression of NIRS-predicted
data.
performedfor TNCaccording to the equation generation and
selection protocols described by Windham et al. (1989). The
performanceof calibration wastested by calculating the stan-
original populations before calibration. A summaryof NIRS
regression statistics are presented in Table 1. The optimum
equation ’used a first derivative function calculated over a
10-nmsegment, and it involved 5- and 1-nmsmoothings. The
r 2 and standardizedH (Mahalanobis) statistics werecalculated
over a range of wavelengths.Wavelengths selected in calibra-
tion were 2294 and 2274 nm.
Leaf tissues were sampledbetweenthe 1st and 5th day of
each sampling month (Maythrough October) from each plot
for elemental analysis. Elementalanalysis was conductedby
dry ashing samples at 450°Cfor 4 h, digesting with 10 mL
1 MHNO3and 10 mL1 MHC1, and bringing the volume to
100 mLwith water (Hue and Evans, 1986). At the same time
soil samplesweretaken from the 0- to 15-cmzone and prepared
according to proceduresused by AuburnUniversity Soil Test-
ing Laboratory(Hueand Evans,1986). Soil and plant extracts
were analyzed using an Inductively Coupled.Argon Plasma
spectrophotometer (ICAP9000 Jarrel-Ash, Franklin, MA).
Cultivar and K rates were treated as fixed variable and
analyses of variance (ANOVA) were conducted using the
Statistical Analysis Systemsprocedure(SASInstitute, Inc.,
1987). Appropriate error terms were used dependingon which
interactions weresignificant. Partial correlation coefficients
were calculated by meansof the multivariate ANOVA function
of the SASgeneral linear modelprocedure, to determine the
relationship betweenroot + rhizomeyield and TNCconcentra-
tion and meanmaximum and minimumtemperatures occurring
in the preceding growthperiod.
RESULTS AND DISCUSSION
Meanair temperatures at 1 m above soil surface at
the Auburn Turfgrass Research Center are presented in
Fig. 1. Temperaturesduring the winter and spring months
of 1992-1993 can be described as mild except for two
periods of below freezing temperatures during winter
to spring transition. Notes taken during the winter of
1992-1993indicated that the turf never went completely
dormant. The temperatures during the winter months of
1993-1994 were much colder than those of the previous
winter. There were 44 d with minimumtemperatures
below 0°C. Visual evaluations indicated the cooler tem-
peratures induced complete dormancy. The summer of
1993 had 29 d of 35°C and above temperatures whereas,
the 1994 summer had no days above 34°C.
Differences in TNCconcentrations were attributable
 MILLER & DICKENS: POTASSIUM AND BERMUDAGRASS CARBOHYDRATE LEVELS 1285

240 T Rhizome + Root

240
Rhizome + Root

210 210

180 180
7
~o~ 150

120 120

9O 90

6O 60
S 0 N D J F M A M J J A S 0 S °0 N D J F M A M J J A S
Month (1992 - 1993) M~nth (1992 1993)
AN OVA ANOVA
Cultivor Cultivor NS NS" NS * NS" NS NS NS NS NS NS NS
K-rote NS NS NS * * NS NS NS NS NS NS NS K-rote NS * NS NS NS NS NS * NS NS NS *
CXK NS NS NS NS NS NS NS NS NS NS NS NS CXK NS NS NS NS NS NS NS * NS NS NS NS

¯ Tifdworf - High K ¯ Tifwoy - High K ¯ Tifdworf - High K ¯ Tifwoy - High K

V Tifdwerf - Low K 13 Tifwoy - Low K
Fig. 2. Total nonstruetural carbohydrate concentrations and signifi-
cance of treatment effects in root and rhizometissues in Tifdwarf
and Tifway bermndagrassesgrownin a loamy sand at two K levels
in 1992-1993. Bars represent standard error. Points without bars
indicate a negligible standard error.
~7 Tifdworf - Low K 13 Tifwoy - Low K
Fig. 4. Total nonstructural carbohydrate concentrations and signifi-
cance of treatment effects in root + rhizome tissues in Tifdwarf
and Tifway bermudagrasses grown in a loamy sand at two K levels
in 1993-1994. Bars represent standard error. Points without bars
indicate a negligible standard error.

to Krate, cultivar, their interaction, andmonth(Fig. organs(Hull, 1992). GreaterTNCconcentrationin the

2-9). Therewerecultivar x Krate interactions several
months (Fig. 4, 5, 6, and8). Theseinteractionsoccurred
in the springor early summer whenthere wereincreases
in TNCconcentrationfromthe previousmonthin Tif-
dwarfbut not necessarily increases in TifwayTNCcon-
centrations.
Root + rhizometissues had greater TNCconcentra-
tions thanshoot+ stolontissues (Fig. 2-9). Starchrarely
accumulates in leaves but concentratesmostlyin storage
below-ground plant parts waslikely a result of greater
TNC concentrationin rhizomes.DunnandNelson(1974)
reportedthat starch contentof bermudagrass cultivars
tested washigherin rhizomesthanstolons. Likewise,a
significant portion of the TNCvalues in the shoot +
stolon tissues wasprobablya reflection of stolon TNC
concentration. Stolons serve as a majorcarbohydrate
storage site andas majorenergysource for spring re-
growthof bermudagrass (Dunnand Nelson, 1974).

210 210
Shoot + Stolon Shoot + Stolon
180 T 180

150 150

12o 12o
9O 9O

60 60

3O 3O

0 0
S 0 N D J F M A M O J A S S 0 N D d F M A M d d A S 0
Month (1992-199,3) Month (1992-1993)
ANOVA AN OVA
Cultivar ** ** ** ** ** ** * ** NS * ** **
K-rote NS NS NS NS NS NS NS NS NS NS NS NS K-rote NS NS NS NS NS NS NS NS NS NS NS NS
CX K NS NS NS NS NS NS NS NS NS NS NS NS CX K NS NS NS NS NS ** NS NS NS NS NS NS

¯ Tifdworf - High K ¯ Tifwoy - High K ¯ Tifdworf - High K ¯ Tifwoy - High K

~ Tifdworf - Low K O Tifwoy - Low K ~ Tifdwarf - Low K [] Tifway - Low K
Fig. 3. Total nonstructural carbohydrate concentrations and signifi- Fig. 5. Total nonstructural carbohydrate concentrations and signifi-
cance of treatment effects in shoot + stolon tissues in Tifdwarf cance of treatment effects in shoot + stolon tissues in Tifdwarf
and Tifway bermudagrasses grown in a loamy sand at two K levels and Tifway bermudagrasses grown in a loamy sand at two K levels
in 1992-1993. Bars represent standard error. Points without bars in 1993-1994. Bars represent standard error. Points without bars
indicate a negligible standard error. indicate a negligible standard error.
 1286 CROP SCIENCE, VOL. 36, SEPTEMBER-OCTOBER1996

Rhizome + Root Rhizome + Root

240 240

210 210
T ~ T
180 180

~ 15o 15o
120 120

90 90

60 60
S 0 N D J F M A M J J A S O N D d F M A M J J A S
Month (1992 1993) Month (1992 1993)
ANOVA ANOVA
Cultiver Cultivar
K-rote NS NS NS NS NS NS NS NS NS NS NS NS K-rote * NS- NS NA NS NS NS NS NS NS NS NS
CXK NS NS NS NS NS ** NS NS NS NS NS NS CXK NS NS NS NS NS NS * NS * NS NS NS

¯ Tifdwarf - High K ¯ Tifway - High K ¯ Tifdworf - High K ¯ Tifwoy - High K

V Tifdworf - Low K 13 Tifway - Low K V Tifdworf - Low K [3 Tifwoy - Low K
Fig. 6. Total nonstructural carbohydrate concentrations and signifi- Fig. 8. Total nonstructural carbohydrate concentrations and signifi-
cance of treatment effects in root + rhizome tissues in Tifdwarf cance of treatment effects in root + rhizome tissues in Tifdwarf
and Tifway bermudagrasses grown in a sand-peat medium at two and Tifway bermudagrasses grown in a sand-peat medium at two
K levels in 1992-1993. Bars represent standard error. Points without K levels in 1993-1994. Bars represent standard error. Points without
bars indicate a negligible standard error. bars indicate a negligible standard error.

Cultivar had the greatest effect on TNCconcentration
among the variables examined (Fig. 2-9). Tifdwarf
grown on loamy sand had 21% greater root + rhizome
TNCconcentration and 56% greater shoot + stolon
TNCconcentration than Tifway. Concentrations fol-
lowed similar patterns with cultivars grownin the sand-
peat medium, with 38 and 54% greater concentrations
in Tifdwarf root + rhizome and shoot + stolons, respec-
tively as compared to Tifway. These differences may
partially be attributed to cutting height and dormancy
characteristics of the cultivars. Thompson(1966) indi-
cated that Tifgreen plots clipped at 6.4 mmhad higher
TNCconcentrations in the winter months than those
clipped at 12.7 mm.Lowerclipped plots generally main-
tain chlorophyll and continue to produce sugars longer
than plots with a higher cut because of soil heat transfer.
Heat from the soil dissipates in the taller grass canopy
subjecting the grass to cooler temperatures. Therefore,
taller turfs are more often damagedby light frosts and
suffer reduced photosynthesis and carbohydrate produc-

210 210
Shoot + Stolon Shoot + Stolon
180 180

150 150

120 120

90 gO

60 60

30 30

0 0
S O N 13 d F M A M J J A S O S F M AM d d A S O

Month (1992-1993) Month (1992-1993)

ANOVA ANOVA
Cultivar ** ** ** ** ** * NS ** NS * ** NS
K--rote NS NS NS NS NS NS NS NS NS NS NS NS K-rate NS NS NS NS NS NS NS NS NS NS NS NS
CX K NS NS NS NS NS NS NS NS NS NS NS NS C X K NS NS NS NS NS NS NS NS NS NS NS NS

¯ Tifdwarf - High K ¯ Tifway - High K ¯ Tifdwarf - High K ¯ Tifwoy -- High K

W Tifdwarf - Low K [] Tifwoy - Low K ~ Tifdwarf - Low K [3 Tifway - Low K

Fig. 7. Total nonstructural carbohydrate concentrations and signifi- Fig. 9. Total nonstructural carbohydrate concentrations and signifi-
cance of treatment effects in shoot ÷ stolon tissues in Tifdwarf cance of treatment effects in shoot -I- stolon tissues in Tifdwarf
and Tifway bermudagrasses grown in a sand-peat medium at two and Tifway bermudagrasses grown in a sand-peat medium at two
K levels in 1992-1993. Bars represent standard error. Points without K levels in 1993-1994. Bars represent standard error. Points without
bars indicate a negligible standard error. bars indicate a negligible standard error.
 MILLER & DICKENS: POTASSIUM AND BERMUDAGRASS
CARBOHYDRATE
LEVELS 1287

Table 2. Pearson correlation coefficients for the relationship of total nonstructural carbohydrate
(TNC)concentrations of root + rhizomes,
shoot + stolons, root + rhizome yield, and mean maximumtemperature for 15 d, mean minimum temperature for 15 d, mean
maximumtemperature for 3 d, and mean minimum temperature for 3 d occurring before sample collection for Tifdwarf (values
above diagonal) and Tifway (values below diagonal) bermudagrasses grown on Ucbee loamy sand.
Root + Shoot + Root + Max. Min. Max. Min.
rhizome stolon rhizome temp. temp. temp. temp
TNC TNC Yield 15-d 15-d 3-d 3-d
Root + rhizome TNC 0.36** - 0.58** - 0.21"* - 0.19"* - 0.11 - 0.18"
Shoot + stolon TNC 0.44** - 0.30** - 0.09 - 0.07 - 0.03 - 0.05
Root + rhizome Yield - 0.28** - 0.10 0.23** 0.20** 0.16" 0.21"*
Max. temp. 15-d - 0.38** - 0.11 0.27** - - -
Min. temp. 15-d - 0.42** - 0.08 0.30** - - -
Max. temp. 3-d - 0.38** - 0.06 0.26** - - -
Min.temp. 3-d - 0.37** - 0.08 0.30** - - -
*, ** Significant at 0.05 and 0.01 level of probability, respectively.

tion. Carbohydrateconcentrationshould be moresimilar wasnot influencedby Kfertilization eventhoughtissue

during the peak growingmonthsif differences in TNC
concentrations were caused by mowingheight. However,
Tifdwarfconsistently had higher carbohydrateconcentra-
tions than Tifwayfrom June to September(Fig. 2-9).
Since cultivar responseto K rate wasgenerally similar,
the factor mostlikely responsiblefor differences in TNC
is genetic variation betweencultivars.
Potassiumis generally considered to be an important
elementin carbohydratesynthesis (Thomaset al., 1959)
and accumulationin plants (Adamsand Twersky,1960;
Matcheset al., 1963). Averageextractable Kconcentra-
tions across this study were 32 and 30 kg ha-~ in the
low K rate and 143 and 145 kg ha-~ in the high K rate
in the loamysand and sand-peat medium,respectively.
Plant tissue K concentrations averaged12%greater in
plants that received the high K rate comparedto those
that received the lower K rate (data not shown).
Analysisof variance indicated that Krate influenced
root + rhizome TNCconcentrations but not shoot +
stolon TNCconcentrations.Theproximityof K fertilizer
in relation to the root and rhizomesmayhave influenced
the below-groundtissues. The high K rate resulted in
lower root + rhizome TNCconcentrations several
monthsduring the study (Fig. 2, 4, and 8). Tifdwarf
grown on the sand-peat mediumand Tifway grown on
loamysand that had received the high K rate had 11%
lower root + rhizome TNCconcentration than those
receiving the low K rate. In contrast, Cookand Duff
(1976) reported that accumulationof carbohydrates
’Kentucky31’ tall fescue (Festuca arundinaceaSchreb.)
concentrationsrangedfrom 7 to 30 g Kkg-t dry matter.
Thompsonand Ward(1965) and Goatley et al. (1994)
reported that K applications did not influence TNCcon-
centrations in bermudagrassroots and rhizomes. The
application rates used by Thompsonand Ward(1965)
and Goatleyet al. (1994) were muchlower than the high
rate evaluatedin this study. It is not knownwhythe high
K rate reduced TNCconcentration or whyit influenced
concentration during somemonthsand not in others.
Cooil and Slattery (1948) found that carbohydratecon-
centration in guayule (PartheniumargentatumL.) in-
creasedwith increasingKapplicationsto a certain Krate,
then carbohydratesbeganto decrease with increasing K
applications. AdequateK is neededfor condensationof
reducingsugarsto polysaccharides(starch), but at higher
rates the starch-sugar balance maybe adverselyaffected
(Cooil and Slattery, 1948). Dunnand Nelson (1974)
found that during winter, sucrose increased slightly in
bermudagrassrhizomes and stolons while starch de-
creased. Fry et al. (1991) reported similar results
’Floratam’St. Augustinegrass.
There were increases in TNCconcentration of grasses
grownin the loamy sand and sand-peat mediumduring
the fall of 1992and a general decrease beginning in
winter andcontinuinginto spring (Fig. 1, 2, 6, and 7).
An increase in carbohydrates during the fall, which
frequently coincides with cold hardeningin plants, has
been reported (Dunnand Nelson, 1974; Rogers et al.,
1975; Svec and Hodges, 1971). Throughoutthe growing
season until autumnTNCconcentrations remainedrela-

Table 3. Pearson correlation coefficients for the relationship of total nonstructural carbohydrate(TNC)concentrations of root + rhizomes,
shoot + stolons, root + rhizome yield, and mean maximum temperature for 15 d, mean minimum temperature for 15 d, mean
maximumtemperature for 3 d, and mean minimum temperature for 3 d occurring before sample collection for Tifdwarf (values
above diagonal) and Tifway (values below diagonal) bermudagrass grown on on sand-peat medium.
Root + Shoot + Root + Max. Min. Max. Min.
rhizome stolon rhizome temp. temp. temp. temp
TNC TNC Yield 15-d 15-d 3-d 3-d
Root + rhizome TNC 0.30** - 0.56** - 0.20** - 0.19"* - 0.15" - 0.22**
Shoot + stolon TNC 0.22** - 0.14 - 0.12 - 0.04 - 0.08 - 0.15"
Root + rhizome Yield - 0.36** - 0.15" 0.30** 0.29** 0.28** 0.32**
Max. temp. 15-d - 0.50** - 0.02 0.20** - - -
Min. temp. 15-d - 0.47** 0.02 0.22** - - -
Max. temp. 3-d - 0.42** 0.06 0.22** - - -
Min. temp. 3-d - 0.50** - 0.08 0.23** - - -
*, ** Significant at 0.05 and 0.01 level of probability, respectively.
1288 CROP SCIENCE, VOL. 36, SEPTEMBER-OCTOBER 1996
lively constant in this study. There were increases in
Tifdwarf shoot + stolon TNC concentrations beginning
in June 1993 leveling off by September (Fig. 3 and 7).
The TNC concentrations of the shoot + stolons for both
cultivars grown in both soil types remained much more
stable than the root + rhizome concentrations. Despite
monthly differences in TNC concentrations, it was
difficult to associate concentration responses with season.
Goatley et al. (1994) reported that bermudagrass TNC
concentrations in March 1990 and 1991 declined com-
pared with rates for the previous December, but concen-
trations in samples collected during December 1991 and
March 1992 were virtually identical.
Root + rhizome TNC concentration was negatively
correlated with the mean maximum and minimum tem-
perature for the 15 d prior to sample collection (Tables
2 and 3). McKell and Youngner (1967) evaluated bermu-
dagrass (Cynodon dactylori) carbohydrate levels over
climatic extremes and found that the greatest accumula-
tion of carbohydrate occurred at the coolest temperature.
Analyses indicated that there were stronger correlations
between TNC concentrations and 15 d mean temperatures
than 3 d mean temperatures. This is most likely due to
time required for low temperature acclimation. Johnston
and Dickens (1977) reported that centipedegrass [Ere-
mochloa ophiuroides (Munro) Hack.] reached a maxi-
mum level of freezing tolerance after 10 d of acclimation
at 16/4°C day/night temperatures. Soluble protein frac-
tions in bermudagrass rhizomes were altered after the
grass was subjected to a hardening environment for 15
or more days (Davis and Gilbert, 1967). Correlation
coefficients obtained between temperature and TNC con-
centration are low since the independent variable, tem-
perature, could not be controlled (Tables 2 and 3). The
confounding effects of temperature and photoperiod are
factors in the relationship unaccounted for in this experi-
mental design.
There was no correlation between shoot + stolon TNC
concentrations and mean 15-d temperatures (Tables 2
and 3). There were positive correlations between shoot +
stolon concentration and root + rhizome concentrations.
Seasonal shoot + stolon fluctuations typically followed
root + rhizome levels (Fig. 2-9). The generally more
significant correlation of temperature with Tifway root
+ rhizome TNC concentrations than with Tifdwarf im-
plies a greater influence of temperature on root + rhi-
zome carbohydrates in Tifway.
Root + rhizome yields were not influenced by K rate
or cultivar (Fig. 10). In contrast, Belesky and Wilkinson
(1983) indicated that root and rhizome weights of the
forage type bermudagrass Tifton 44' were increased by
application of 70 or 140 kg K ha"1. Keisling et al. (1979)
reported that an application of 112 kg K ha"1 increased
rhizomes 800 kg ha"1. Higher root + rhizome weights
in 1993-1994 than 1992-1993 probably reflect differences
in turf maturation. Since these plots were established in
the spring of 1992, the grasses had not developed a
mature root and rhizome system during much of the first
year. Positive correlations were obtained between root
+ rhizome yield and temperature (Tables 2 and 3). Even
6000
5500
• Loamy Sand
5000
V Sand-Peat
4500
4000
3500
3000
2500
2000
1500
S O N D J F M A M J J A S O N D J F M A M J J ASD
1992 1993 1994
Month
Fig. 10. Seasonal fluctuations of root + rhizome yield of Tifdwarf
and Tifway bermudagrasses grown in a loamy sand or sand-peat
medium. Means are averaged across cultivar and K-rate. Bars
represent standard error. Points without bars indicate a negligible
standard error.
though root + rhizome weights remained high during
the winter months, generally high temperatures were
associated with higher root + rhizome weights. DiPaola
et al. (1982) indicated that root growth continued after
winter shoot dormancy occurred, and at soil temperatures
below 10°C. In this study, no attempt was made to
distinguish between live actively growing roots and rhi-
zomes and those not capable of functioning. Separating
out the nonviable roots might have resulted in greater
seasonal variation.
Root + rhizome TNC concentrations of Tifdwarf and
Tifway were responsive to temperature; however, dis-
tinct changes in concentration due to season, especially
winter to spring transition, were not always evident.
These results suggest that use of stored carbohydrates
was low, or that carbohydrate accumulation/utilization
ratio was remaining relatively constant. The high K rate
provided no benefit in terms of TNC concentration.
There was an unexplained negative response from high
K rate on root + rhizome concentration in both cultivars.
More detailed work is needed to evaluate the influence
of K on carbohydrate concentration in roots and rhizomes
of these cultivars.
MILLER & DICKENS: POTASSIUM AND BERMUDAGRASS CARBOHYDRATE LEVELS 1289