Biology, ecology and management of cogongrass (Imperata cylindrica (L.) Beauv.)

BIOLOGY, ECOLOGY AND MANAGEMENT OF COGONGRASS
[Imperata cylindrica (L.)Beauv.]

By

TOMMY RAY WILLARD

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1988

; ;

iii .;. i
ii
~iiil;;;..il li',iilEiiiiiiiiii ;il

To my loving wife Donna and my daughter Stacey I

humbly dedicate this work. No words can adequately describe

how you have enhanced the quality of my life.

ACKNOWLEDGEMENTS

For their support during my tenure at the University of

Florida I wish to express my sincere appreciation to my

committee members: Dr. William Haller, Dr. Donn Shilling,

Dr. David Hall, Dr. Barry Brecke, and Dr. Paul Mislevy. I

also wish to thank Dr. Wayne Currey for his guidance during

my Masters program and the majority of my Ph.D. studies.

I especially acknowledge Dr. Donn Shilling, co-chairman of

my committee; without his unending patience and willingness

to teach I would not be at this point today.

I thank Mr. David Studstill and Mr. Tim Pederson,

without whose technical assistance it would have been

extremely difficult, if not impossible, to conduct the

research presented.

During my tenure at the University of Florida I was

fortunate to have been associated with graduate students of

the highest caliber. The friends I made here will always be

remembered.

I thank my parents, Mr. and Mrs. Morris Willard for

their love and understanding. The values they instilled in

me during my youth are responsible for all I have attained.
k.

iii

TABLE OF CONTENTS

PAGE

ACKNOWLEDGEMENTS.................................... iii

LIST OF TABLES................................. ....... vii

LIST OF FIGURES............. .................... ..... ix

ABSTRACT.............................................. x

CHAPTERS

I INTRODUCTION................................... 1

II COGONGRASS [Imperata cylindrica (L.)Beauv.]
DISTRIBUTION ON FLORIDA HIGHWAY RIGHTS-OF-WAY 3

Introduction.....*** .......................*... 3
Materials and Methods.......................... 7
Results and Discussion......................... 9

III THE INFLUENCE OF HERBICIDE COMBINATIONS AND
APPLICATION TECHNOLOGY ON COGONGRASS [Imperata
cylindrica (L.)Beauv.] CONTROL............... 18

Introduction..........................,........ 18
Materials and Methods.......................... 21
Cogongrass rhizome biomass sampling.......... 21
'Sequential herbicide application programs.... 22
Tank-mix combinations of imazapyr and
glyphosate... ...... ...................... 24
Comparison of low and conventional volume
applications of imazapyr and glyphosate.... 25
Ropewick applications of imazapyr and
glyphosate................................. 27
Imazapyr longevity bioassay.................. 28
Results and Discussion........*........,...... 32
Cogongrass rhizome biomass sampling......... 32
Sequential herbicide application programs.... 34
Tank-mix combinations of imazapyr and
glyphosate.. ...... ............. o.......... 40
Comparison of low and conventional volume
applications of imazapyr and glyphosate.... 42

Ropewick applications of imazapyr and
glyphosate................................. 44
Imazapyr longevity bioassay.................. 46

IV THE INFLUENCE OF MECHANICAL AND CHEMICAL INPUTS
ON COGONGRASS [Imperata cylindrica (L.)Beauv.]
CONTROL....... ......... ...................... 49

Introduction............................... 49
Materials and Methods...................... 51
Integration of mowing and disking with
herbicide treatments................... 51
The influence of photoperiod on shoot
initiation in cogongrass rhizomes...... 55
The influence of cogongrass stage of
development and defoliation on
glyphosate efficacy ................... 56
Results and Discussion..................... 60
Mowing-herbicide studies................. 60
Disking-herbicide studies................ 65
The influence of photoperiod on shoot
initiation in cogongrass rhizomes...... 71
The influence of cogongrass stage of
development and defoliation on
glyphosate efficacy.................... 73

V THE INFLUENCE OF STAGE OF DEVELOPMENT AND MOWING
ON BAHIAGRASS [Paspalum notatum var. saurae
Parodi 'Pensacola'] AND COGONGRASS [Imperata
cylindrica (L.)Beauv.] INTERFERENCE.......... 79

Introduction............................... 79
Materials and Methods...................... 81
Preliminary Studies...................... 81
Seedling bahiagrass and emerging cogon-
grass interference studies............. 83
The influence of mowing on established
bahiagrass and emerging cogongrass
interference........................... 84
Experimental design and analysis......... 86
Results and Discussion..................... 86
Seedling bahiagrass and emerging cogon-
grass interference studies............. 86
The influence of mowing on established
bahiagrass and emerging cogongrass
interference..............** ........... 90

VI SUMMARY AND CONCLUSIONS......................... 95

APPENDICES

A OCCURENCE AND SEVERITY OF Andropogon sp. ON
FLORIDA HIGHWAY RIGHTS-OF-WAY................. 98

B OCCURRENCE AND SEVERITY OF Phragmites australis
ON FLORIDA HIGHWAY RIGHTS-OF-WAY ............. 99

C OCCURRENCE AND SEVERITY OF Panicum maximum ON
FLORIDA HIGHWAY RIGHTS-OF-WAY................. 100

D OCCURRENCE AND SEVERITY OF Sorghum halepense ON
FLORIDA HIGHWAY RIGHTS-OF-WAY................ 101

E OCCURRENCE AND SEVERITY OF Pennisetum purpureum
ON FLORIDA HIGHWAY RIGHTS-OF-WAY.............. 102

F OCCURRENCE AND SEVERITY OF Brachiaria mutica ON
FLORIDA HIGHWAY RIGHTS-OF-WAY................. 103

G OCCURRENCE AND SEVERITY OF Sporobolus indicus ON
FLORIDA HIGHWAY RIGHTS-OF-WAY................. 104

H OCCURRENCE AND SEVERITY OF Paspalum urvillei ON
FLORIDA HIGHWAY RIGHTS-OF-WAY................. 105

LITERATURE CITED................ ..................... 106

BIOGRAPHICAL SKETCH................................... 113

LIST OF TABLES

TABLE PAGE

2.1 Perennial grass weeds frequently occurring
on Florida highway rights-of-way....6....... 5

2.2 Right-of-way distance surveyed by district
in 1984-85 for perennial grass weeds by the
Florida Department of Transportation........ 10

3.1 Cogongrass rhizome biomass collected in
Chiefland, FL from 1986 to 1988............. 33

3.2 The effect of sequential herbicide treatments
on cogongrass foliage and rhizome dry weight
(Experiment 1 1985 through 1987 Area A). 35

3.3 The effect of sequential herbicide treatments
on cogongrass foliage and rhizome dry weight
(Experiment 2 1986 through 1988 Area B). 36

3.4 The effect of imazapyr and glyphosate tank-
mixes on cogongrass foliage and rhizome dry
weight...................................... 41

3.5 The effect of carrier volume and imazapyr and
glyphosate rate on cogongrass foliage dry
weight...................................... 43

3.6 The effect of ropewick applications of
imazapyr and glyphosate on cogongrass
foliage and rhizome dry weight.............. 45

3.7 The influence of imazapyr dissipation on the
inhibition of dry weight (IDW) response of
four grass species.......................... 47

4.1 Timing of mechanical and chemical treatments
and harvest performed for evaluation of
integrated cogongrass control studies....... 54

4.2 The effect of mowing and herbicides on cogon-
grass foliage and rhizome dry weight (Experi-
ment 1 1985 through 1987)................. 61

vii

4.3 The effect of mowing and herbicides on cogon-
grass foliage and rhizome dry weight (Experi-
ment 2 1986 through 1988)................. 64

4.4 The effect of disking and herbicides on cogon-
grass foliage and rhizome dry weight (Experi-
ment 1 1985 through 1987)................. 66

4.5 The effect of disking and herbicides on cogon-
grass foliage and rhizome dry weight (Experi-
ment 2 1986 through 1988)................. 69

4.6 The influence of photoperiod on shoot
initiation from cogongrass rhizomes at one
week after planting.......*................. 72

4.7 The influence of time of defoliation on
greenhouse-grown cogongrass growth.......... 74

4.8 The influence of defoliation on glyphosate
efficacy in cogongrass...................... 76

5.1 The interaction of seedling bahiagrass (B)
and emerging cogongrass (C) as measured by
relative yield (RY), relative yield total
(RYT) and relative crowding coefficient (RCC)
of leaf dry weight and height and total pot
dry weight at eight weeks after planting.... 87

5.2 The influence of mowing on the interaction of
established bahiagrass (B) and emerging
cogongrass (C) as measured by relative yield
(RY) and relative crowding coefficient (RCC)
using leaf dry weight produced by each
species..................................... 91

5.3 The influence of mowing and bahiagrass
competition on cogongrass height, shoot
number and rhizome dry weight............... 92

viii

LIST OF FIGURES

FIGURE PAGE

2.1 Occurrence and severity of Imperata
cylindrica on Florida highway rights-of-way. 12

3.1 Weekly precipitation in Gainesville, FL,
1986 through 1987........................... 38

ix

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

BIOLOGY, ECOLOGY AND MANAGEMENT OF COGONGRASS
[Imperata cylindrica (L.) Beauv.]

By

Tommy Ray Willard

December, 1988

Chairman: Dr. W. T. Haller
Major Department: Agronomy

Field and greenhouse studies were conducted to deter-

mine the distribution and management of cogongrass and the

effects of this weed and control strategies on desirable

vegetation. A survey of 22% of the Florida highway rights-

of-way indicated that cogongrass is distributed along major

thoroughfares from Lake Okeechobee to the Panhandle region.

The greatest infestation level was found in north-central

Florida.

Field studies conducted on herbicide and application

technology for cogongrass control indicated that single

herbicide applications would not provide long-term control,

but control could be acheived with sequential treatments of

imazapyr and glyphosate. Ropewick applications of imazapyr

were as effective as conventional spray applications.

Imazapyr or glyphosate applied in 46 L'ha-1 water did not

increase cogongrass control over 234 L'ha- applications.

The half-life of imazapyr in sandy soils was determined to

be eight months and residual imazapyr affected seedling

growth of four grass species in the following order (most to

least affected): ryegrass > browntop millet > bermudagrass >

bahiagrass.

Field and greenhouse studies, which were conducted to

determine the influence of mowing and disking alone and when

combined with herbicide treatments, indicated that mowing

alone has little effect on cogongrass. Disking cogongrass

one time caused an increase in stand density, however moder-

ate control was acheived with two diskings. The combination

of imazapyr or glyphosate with two diskings provided nearly

100% cogongrass control. Other studies indicated that

glyphosate should be applied during the fall to maximize

translocation to, and control of, cogongrass rhizomes.

Replacement studies indicated that cogongrass emerging

from rhizomes competed effectively with, and could displace,

seedling bahiagrass. However, cogongrass growth from rhi-

zomes was severely inhibited when competing interspecifically

with established bahiagrass. Mowing negatively impacted

cogongrass rhizome dry weight when competing intraspec-

ifically, but shoot number increased. This did not occur

under interspecific competition due to the overwhelming

xi

impact of bahiagrass. Mowing had less impact on bahiagrass

than cogongrass. Differences in leaf morphology (cogongrass

being erect with little leaf lamina below 10 cm and bahia-

grass being more prostrate with significant leaf area below

10 cm) appear to favor bahiagrass under mowed conditions.

xii

CHAPTER 1
INTRODUCTION.

Cogongrass [Imperata cylindrica (L.)Beauv.] is an

exotic, aggressive, perennial grass. During the last fifty

years it has become a serious pest in pastures, pines, and

other uncultivated areas in Florida. It went virtually

unnoticed as a pest in this state until approximately 1972,

even though it already covered four thousand hectares in

south Alabama. By 1982, botanists and weed specialists in

Florida were being routinely questioned as to the identifica-

tion and control of this grass that had "suddenly appeared".

The need for soil stabilization materials and low-input

forages in Florida led to the introduction and intentional,

as well as unauthorized, dissemination of cogongrass. How-

ever, researchers realized too late that the cogongrass

spread beyond soil stabilization areas and was not a

quality, productive forage (Tabor, 1949). During this same

period (1940-50s) authorities in Alabama and Mississippi

were destroying plantings of cogongrass on their experiment

stations.

The Florida Department of Transportation in 1984 funded

cogongrass research to be conducted by the Agronomy Depart-

ment of the University of Flotida. The goals of this

research were to characterize the growth and development of

cogongrass in Florida and to use this information to develop

environmentally-sound and effective control strategies. The

objectives were to: 1) define the distribution of cogongrass

on Florida highway rights-of way, 2) evaluate available

herbicide and application technology, 3) integrate mechan-

ical and chemical control strategies, and 4) to evaluate the

competitive interaction of cogongrass and bahiagrass as it

is influenced by vegetation management practices. Cogongrass

does not discriminate between roadsides and other favorable

habitats; therefore, these results should be used not only

to guide cogongrass control programs on Florida highway

rights-of-way, but could be applied to other infested areas.

It may not be possible to eradicate cogongrass from Florida;

however, a diligent effort to manage as many infestations as

possible could reduce the threat of this species.

CHAPTER 2
COGONGRASS [Imperata cylindrica (L.)Beauv.] DISTRIBUTION
ON FLORIDA HIGHWAY RIGHTS-OF-WAY.

Introduction

The Florida Department of Transportation (FDOT)

maintains 45,326 hectares of roadside vegetation along

18,500 km of highway Maintenance practices, which include

mowing, fertilization, tree and brush removal, wildflower

management, and selective herbicide treatments are necessary

to create a safe, aesthetically pleasing thoroughfare for

the state's large resident and tourist populations (Evink,

et al., 1983). Therefore, the FDOT allocates approximately

twenty million dollars annually for these activities.

The same environmental conditions that are conducive to

the establishment of turfgrass species [predominantly bahia-

grass, Paspalum notatum var. saurae Parodi 'Pensacola' #2

PASNS, with sporadic populations of common bermudagrass,

Cynodon dactylon (L.)Pers. # CYNDA, and centipedegrass

Eremochloa ophiuroides (Munro)Hack. # ERLOP] on these

Personal communication, 1986, J.A. Lewis, Florida
Department of Transporation, Tallahassee, FL.

Letters following this symbol are a WSSA-approved computer
code from the Composite List of Weeds, Weed Sci. 32,Suppl.
2. Available from WSSA, 309 West Clark Street, Champaign,
IL 61820.

rights-of-way also favor the establishment of undesirable

perennial grasses. Rank growth of these weedy species,

which rapidly exceeds the height of desirable vegetation,

causes several problems. Motorist safety is jeopardized by.

limited sight distance caused by tall vegetation during the

growing season and also by smoke from fires fueled by this

vegetation during the dry, winter months when the foliage is

dessicated. Excessive vegetation increases maintenance

costs by reducing mowing efficiency (km per man-hour) and by

increasing equipment repair and subsequent downtime. Mowing

has provided insignificant control of these species, and has

been shown to stimulate seed production and dispersal in

some of them (McCaleb and Hodges, 1971). The use of herbi-

cides to control weeds and reduce mowing costs has increased

as infestations have expanded (Lewis, 1985). Although less

tangible, motorist reaction to the poor aesthetics caused by

weedy vegetation and the use of herbicides has been negative.

In 1981, the FDOT Bureau of Environment cited nine

grassy weed species that required additional efforts beyond

those needed to maintain turf species (Evink, et al., 1983).

Table 2.1 lists the common and scientific names of these

grasses.

Selective control of these weeds is very difficult, as

the weeds and the turf are both warm-season perennial grasses.

Specialized equipment, such as ropewick applicators, can be

used in some situations to achieve selective control. For

ropewick treatments to provide optimal control without turf

Table 2.1. Perennial grass weeds frequently occurring on
Florida highway rights-of-way.

Common Name

Bluestems or
Broomgrasses

Cogongrass

Giant Reed

Guineagrass

Johnsongrass

Napiergrass

Paragrass

Smutgrass

Vaseygrass

Scientific Name

Andropogon virginicus L. I ANOVI

Andropogon virginicus L. # ANOVI
Andropogon glomeratus (Walt.)BSP. # ANOGL

Imperata cylindrica (L.)Beauv. # IMPCY

Phragmites australis (Cav.)Trin. ex Steud
# PHRCO

Panicum maximum Jacq. # PANMA

Sorghum halepense (L.)Pers. # SORHA

Pennisetum purpureum Schumach. # PESPU

Brachiaria mutica (Forsk.)Stapf # PANPU

Sporobolus indicus (L.)R. Br. # SPZIN

Paspalum urvillei Steud. # PASUR

injury, a significant height differential between the weed

and the turf is required. Herbicide applications made with

a ropewick are slow compared to broadcast applications, more

expensive, and may conflict with the mowing protocols of the

FDOT management districts (Lewis, 1985). Nonselective

control, usually attained with handgun applications, is

undesirable from an aesthetic standpoint due to the dessica-

tion of large areas of vegetation and could also lead to

erosion. These selective and nonselective approaches are

intensive in terms of chemical and manpower expenditures.

The applications are usually employed only when infestations

are large enough to warrant expenditure of manpower and

equipment. This approach to weed control has resulted in

reduced herbicide efficacy on large infestations (making

repeat applications necessary to achieve long-term control)

and the continued expansion of smaller untreated infesta-

tions.

Glyphosate, isopropylamine salt of N-(phosphonomethyl)-

glycine, dalapon, magnesium and/or sodium salts of 2,2-

dichloropropionic acid, and hexazinone, 3-cyclohexyl-6-(di-

methylamino)-l-methyl-1,3,5-triazine-2,4(1H,3H)-dione are

the primary chemicals used for control of perennial grass

weeds in Florida (Lewis, 1985). However, several applica-

tions may be necessary for optimal control, and in the case

of cogongrass, improved application and/or herbicide

technology is needed.

The FDOT surveys its rights-of-way to assess needs and

develop programs for roadway and roadside maintenance.

Previous surveys have not focused on the identification,

distribution, and intensity of perennial grass weed infesta-

..tions. To obtain this information FDOT district maintenance

engineers surveyed major rights-of-way for the nine weed

species listed in Table 2.1.

MATERIALS AND METHODS

A survey of Florida highway rights-of-way was conducted

in 1984-85 for the occurrence and severity of nine perennial

grass weeds. The purpose of the survey was, first, to give

district personnel the opportunity to identify the problem

species on a county or individual roadway basis and second,

to assess the impact of maintenance activities (i.e., mowing,

chemical control, or roadside construction) on these species.

District maintenance engineers, who were involved in

roadway selection and survey format, conducted the survey,

and university investigators verified and compiled the data.

The criteria for roadway selection were: 1) all limited

access roadways including interstate highways and the

Florida Turnpike; 2) where possible, at least one north-

south and one east-west roadway per county in addition to

those selected under criteria 1; and 3) urban sections of

roadway were excluded. The survey format included location

data (county name, section number, roadway number, and

mileage marks) and a zero to three rating scale with a zero

rating indicating that the particular species was not present

on that section of roadway, a rating of one indicating that

the species was present in only sparse infestations and that

present maintenance/control practices provided adequate

..control, a rating of two indicating the species was more

frequently encountered and present maintenance/control

practices provided only partial control, and a rating of

three indicating numerous large infestations and/or little

control was realized from existing programs. Additional

space was provided on the survey form to note any unusual

conditions or to more precisely locate major cogongrass and

napiergrass infestations.

Using this rating scale, cogongrass and napiergrass

would receive a rating of three, in most situations, due to

the lack of effective control programs. These two species

pose an extreme problem for roadway maintenance personnel

and are a threat to adjoining agricultural, silvicultural,

and residential areas. To gain a perspective of the relative

intensity of these two species, those counties with one to

five infestations of 100 m2 or greater were assigned a

rating of one (low), counties with six to fifteen infesta-

tions were assigned a rating of two (moderate), and counties

with greater than fifteen infestations were assigned a rating

of three (severe).

Training seminars were conducted to familiarize

personnel with the survey format and identification of each

species. Trial surveys of selected roadways were also

conducted. Surveyors were given eighteen months to complete

their surveys of assigned roadways, which afforded them the

opportunity to evaluate mowed or reconstructed areas after

sufficient weed regrowth had occurred.

Districts IV and VI did not use the survey form as the

recording medium. A K-5000 Nu-Metrics Distance Measuring

Instrument (DMI)3 and printer were programmed with event

numbers corresponding to each species. This instrument

continuously monitored miles traveled. As infestations were

identified, the operator entered the appropriate event

number which was printed with the mileage mark. From these

printouts the principle investigator assigned severity

ratings for the counties based on the number and size of

infestations in a manner similar to that used for cogongrass

and napiergrass.

It should be emphasized that the data presented is

subjective in nature and is limited to those roadways

surveyed. Therefore, a species that is represented as not

occurring in a particular county may, in fact, occur on

other unsurveyed roadways or in off right-of-way areas.

Results and Discussion

Table 2.2 presents the geographical scope of the survey

by district. The 8234 km surveyed was 92.1% of that

designated for survey. Though varying with locale and

surveyor, the districts using'the survey sheet recorded an

3Nu-metrics Instrumentation, Vanderbilt, Pa, 15486-0471.

Table 2.2. Right-of-way distance surveyed by district in
1984-85 for perennial grass weeds by the Florida Department
of Transportation.

Distance
Geographic No. of Surveyed % of
District Region Counties km Centerline

I SW Gulf Coastal 14 2182 28.1
through S Central

II N Central through 18 1709 21.4
N Atlantic Coastal

III West Florida 16 1377 18.5
(Panhandle)

IV SE Atlantic Coastal 7 1078 21.0

V Central through 10 1415 21.3
Central Atlantic
Coastal

VI Everglades 2 473 22.5
(Monroe and Dade
counties)

Statewide Total 67 8234 22.2

1Percentage of centerline distance computed by dividing the
kilometers surveyed by two to account for surveying each
shoulder.

average of one observation for every 10.1 centerline km. In

Districts IV and VI where the DMI was utilized, recorded

observations averaged one for every 0.5 centerline km. This

points out the greater accuracy that can be achieved by

using an automatic recording device such as the DMI.

The distribution and severity of cogongrass has been

reported for Alabama, Mississippi, and Louisiana (Dickens,

1974; Patterson, et al., 1983). The major emphasis of this

survey was to determine the range and severity of cogongrass

on Florida highway rights-of-way. The distribution maps for

the other eight species surveyed are presented in Appencies

A through H but will not be elaborated upon further. The

distribution and severity of cogongrass on Florida highway

rights-of-way is presented in Figure 2.1. Rights-of-way in

the north central region (Districts II and V) and the south

central region (Districts I and IV) were severely infested

with cogongrass. Historical records indicate that these

locations correspond to intentional propagation of this

species for forage and soil stabilization purposes during

the 1940s and 1950s (Dickens and Buchanon, 1971; Tabor,

1949). Points of introduction in Florida include Gainesville

(University of Florida Experiment Station), Brooksville

(USDA Plant Introduction Station), and Withlacoochee (Soil

Conservation Service reclamation area).

Once introduced, dispersal of cogongrass can occur by

two mechanisms, seeds and rhizomes. Cogongrass produces

plummed wind-blown seeds beginning in January (in response

DISICT
IS DIST

I

O NONEREPORTED

MLow

MODERATE

ICT

DISTRICT
I

SEVERE

DISTRICT
L v

DISTRICT
L IV

Figure 2.1. Occurrence and severity of Imperata cylindrica
on Florida highway rights-of-way.

DISTRICT

to the freezing temperatures, Holm, et al., 1977) and

continuing until May (personal observation). Dickens and

Moore (1974) found that cogongrass seed collected from Grand

Bay, Alabama, was viable soon after anthesis and germinated

readily under laboratory conditions. They also reported a

significant temperature response, with the optimum germina-

tion (60%) occurring at 30 C, but less than 10% germination

at 20 C. The viability of the seed declined steadily under

storage in sealed vials. The significance of cogongrass seed

production in Florida is undetermined at this time. Several

factors appear to limit the impact that seeds could have on

the overall spread of cogongrass on Florida highway rights-

of-way. First, the plant begins seed production during the

coldest and driest months of the year, conditions which do

not favor immediate germination. Secondly, Eussen and

Soerjani (1975) reported that flowers produced in response

to environmental stress, such as freezing temperatures,

seldom contained seed. If seed is produced, research has

shown that they do not accumulate in the soil seed bank

(Hopkins and Graham, 1984). Thirdly, the orientation of the

majority of infestations on highway rights-of-way is parallel

to the roadway. Winds funneling through these tree-lined

corridors and turbulence caused by vehicles would tend to

disperse the seeds within the length of the infestation, as

the distance cogongrass seeds travel has been estimated to

average only fifteen meters from the seedhead (Holm, et al.,

1977). Fourthly, if viable seeds are dispersed beyond the

bounds of the infestation area, they are unable to penetrate

into adjoining forested areas due to the hairs which catch

on vegetation. If they are deposited onto the right-of-way,

research has shown that their survival is dependent on con-

--tacting moist, bare soil that is free of competing vegetation

(Kushwaha, et al., 1983). Lastly, cogongrass seedlings have

been reported to be frail and unlikely to survive (Kushwaha,

et al., 1983) except under ideal conditions (Dickens, 1976).

However, environmental and soil conditions in the phosphate

mining area of the state near Bartow (Polk county, District

I) could be favorable cogongrass seedlings on off right-of-

way areas. This ability to spread by seeds into open areas

could also pose a major threat to citrus production on the

central Florida ridge.

The second method of cogongrass dispersal is by

rhizomes. From established infestations it has been reported

that cogongrass will expand laterally 25-40 cm per year

(Dickens, 1973). Infestations on Florida highway rights-of-

way tend to be isolated and occurring at irregular intervals.

Roadway construction requires that large quantities of soil

be transported to the site to create front and back drainage

slopes. This bulk movement of potentially rhizome-contami-

nated soil from maintenance unit barrow pits may be a major

avenue of dispersal in Florida. This may have occurred on

interstate highway 75 (1-75) and U.S. 441/301. These two

roadways which traverse Alachua (District II) Marion, Sumter

(District V), Polk, Highlands (District I), and Okeechobee

(District IV) counties are heavily infested with cogongrass.

Alachua and Marion counties are the most heavily infested in

the state, on rights-of-way as well as in pastures and

pulpwood production areas. These heavy infestations

- occurred due to the aforementioned intentional planting of

cogongrass followed by intense roadway construction in these

counties during the mid to late 1960s (Lewis, 1986).

District III reported several isolated infestations.

These infestations occurred on highways U.S. 90 and

interstate 10 (1-10). During the 1940s cogongrass was

removed from university experiment stations without

authorization. It was estimated that four hundred hectares

were planted in west Florida during this period (Tabor,

1952). There is little doubt that transport of cogongrass

occurred on U.S. 90 which was the major east-west corridor

during this time, prior to the completion of I-10 in the

1950s. Movement of cogongrass from Florida to Alabama was

reported by Dickens (1974). West to east movement of

cogongrass may have also occurred via U.S. 90 and 1-10

(which parallels U.S. 90) as these two roadways pass through

Grand Bay and Mobile, Alabama, which have been infested with

cogongrass since the 1940s (Tabor, 1949). However, observa-

tions made in 1988 of cogongrass along 1-10 from Grand Bay

to Pensacola, Florida, showed that Alabama infestations

ended 15-25 km west of the state line and Florida infesta-

tions began approximately 15 km east of the state line.

Tallahassee, the state capital (Leon county, District

III), is a major transportation hub because of the conver-

gence of I-10, U.S. 90, and U.S. 19/27. The interchanges

surrounding Tallahassee are heavily infested with cogongrass.

.This reinforces the hypothesis that roadway construction and

maintenance has played a major role in the distribution of

cogongrass in Florida.

The dispersion of cogongrass in Florida does not appear

to be as rapid as that experienced in Alabama. However, its

spread will continue to increase unless control programs are

enacted immediately. In the author's opinion, there should

be a coordinated effort headed by the Florida Department of

Agricultural and Consumer Services that will at least

partially subsidize the cost of cogongrass control in all

areas of the state. This program should be conducted in a

manner similar to that implemented for witchweed [Striga

asiatica (L.)Ktze. # STRLU] in North Carolina (USDA, 1970).

Efforts should also be made to more strictly regulate the

introduction of plant species into the United States under

the Federal Noxious Weed Law. There are numerous examples

of species that have been introduced in the past for forage

or soil stabilization/improvement purposes that have, with

time, proven to be unsuitable for the intended purpose or

have exhibited characteristics that enable them to success-

fully compete with and eventually displace desirable vegeta-

tion [e.g., kudzu, Pueraria lbbata (Willd.)Ohwi # PUELO,

melaleuca, Melaleuca quinquenervia (Cav.)Blake # MLAQU, and

torpedograss, Panicum repens L. # PANRE] (Williams, 1980).

Efforts by the Animal and Plant Health Inspection Service

(APHIS) have prevented the introduction of potentially

hazardous insects and plant pathogens. The effort placed on

Identifying, locating, and eventually controlling noxious

plant species should receive the same level of diligence.

The sedentary nature of weeds and the relatively slow

encroachment that they make have allowed many introduced

species to become established to the point that they cannot

be eradicated. It has been forty years since R. L.

Pendleton wrote:

Certainly its [cogongrass'] hazard as a poten-
tial weed for upland crops in the tropical and
subtropical portions of the western hemisphere
is a very much more serious threat to agriculture
than the small amount of benefit it can possibly
be as a forage. The writer feels very strongly
that steps should be taken at once to completely
eradicate this noxious weed from the western
hemisphere. (Pendleton, 1948, p. 1048).

If control measures are not implemented in the near future

in all states with cogongrass infestations, many hectares of

valuable pasture and silvicultural land could become

unproductive.

CHAPTER 3
THE INFLUENCE OF HERBICIDE COMBINATIONS AND APPLICATION
TECHNOLOGY ON COGONGRASS [Imperata cylindrica (L.)Beauv.]
CONTROL.

Introduction

Cogongrass [Imperata cylindrica (L.)Beauv. # IMPCY] is

a serious pest throughout the sub-tropical and tropical

regions of the world. It ranks as the seventh most trouble-

some weed worldwide (Holm, et al., 1977). Banana (Musca x

paradisiaca L.), citrus (Citrus species), coconut (Cocus

nucifera L.), oil palm (Elaeis guineensis Jacq.), pastures,

pineapple [Ananas comosus (L.)Merrill], pine (Pinus species),

rubber [Hevea brasiliensis (Willd. ex A. Juss.)Mull. Arg.],

and tea [Camellia sinensis (L.)0. Kuntze] are major crops

that have been reported to be adversely affected by the

presence of cogongrass. With little utility except for

thatch and short-term forage production and soil stabiliza-

tion, cogongrass research has been geared toward control.

However, control of this species has proven to be extremely

difficult. Slash and burn and shifting agriculture have

resulted in transient control, usually allowing only a year

1Letters following this symbol are a WSSA-approved computer
code from the Composite List of Weeds, Weed Sci. 32,Suppl.
2. Available from WSSA, 309 West Clark Street, Champaign,
IL 61820.

or two of crop production before reinfestation takes its toll.

By eliminating natural vegetation that competes effectively

with cogongrass and concomitantly distributing seeds and

rhizomes, these control strategies have increased the area

of cogongrass infestation (Prommool, 1984).

With the advent of chemical weed control in the 1950s

efforts were made to apply this technology to cogongrass

management. Since that time, at least thirty compounds and

hundreds of combinations have been evaluated and reported

for cogongrass control (Dickens and Buchanon, 1975; SEAWIC,

1988). Of these herbicides, glyphosate, N-(phosphonomethyl)-

glycine, dalapon, 2,2-dichloropropanoic acid, and imazapyr,

()-2-[4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-lH-

imadazol-2-yl]-3-pyridinecarboxylic acid, have shown the

greatest activity on cogongrass with the fewest adverse

effects (i.e., bioaccumulation of heavy metals, extended

periods of soil sterilization, non-target species injury, or

applicator injury). In most situations, long-term control

was not acheived from a single application of any of these

compounds. Repeat applications have proven to be necessary

to directly kill or deplete the rhizome biomass, which is

the basis of cogongrass' ability to survive and spread.

Various innovations in application technology have also

been tested to improve the activity of these herbicides in

cogongrass. Low and ultra-low volume (ULV) applications

(usually in the range 20 to 100 L'ha-1 carrier) have been

reported to enhance the activity of glyphosate in common

bermudagrass [Cynodon dactylon (L.)Pers. # CYNDA] (Erickson

and England, 1983; Jordan, 1981), quackgrass (Agropyron

repens (L.)Beauv. I AGRRE) (Sandberg, et al., 1978), and

torpedograss [Panicum repens L. # PANRE] (Baird, et al.,

1983). The use of this low-volume technology with glyphosate

(Arif, et al., 1986) and imazapyr (Bacon, 1986) on cogongrass

has been reported. However, results have been variable and

not definitive as to enhancement of activity or the long-term

control realized.

Ropewick and roller-type applicators have been used for

the past ten years for selective control of perennials in

susceptible crops such as cotton (Gossypium hirsutum L.),

soybean [Glycine max (L.)Merr.], and pastures (Dale, 1981;

Wiese and Lavake, 1980). The type of applicator, speed,

herbicide concentration, and number of applications have all

been reported to influence the efficacy of glyphosate when

applied in this manner (Derting, 1981).

The concept of control has been frequently defined as

the initial effects of the herbicide treatments. In a few

cases, excellent control has been reported in as little as

two weeks (Al-Juboory and Sarmaly, 1984) and usually within

the same growing season. The research that provides the

most useful information (i.e., long-term efficacy) about

cogongrass control is that which allows sufficient time to

elapse following treatment before evaluation, such as those

studies conducted by Dickens (1973). The most accurate

measure of perennial weed control is the inhibition of

regrowth from perenniating organs. This can be accomplished

by harvests of foliage regrowth and the amount of productive

perenniating tissue remaining after an extended period of

time (i.e., one growing season following application).

Therefore, the objective of this research was to study

the influence of herbicide combinations, both in-tank

mixtures and sequential herbicide programs, and application

methodology (i.e., low-volume and ropewick applications) for

long-term cogongrass control.

Materials and Methods

Cogongrass Rhizome Biomass Sampling.

To determine the influence of time and to characterize

the infestation area that was to be used for the research

henceforth described, cogongrass rhizomes were harvested

from a site in Chiefland, Florida. Sampling of two

designated areas was conducted from October, 1986, through

April, 1988. Historical information provided by the land-

owner indicated that cogongrass was introduced in the 1950s

following intentional planting for fire break stabilization

within a pulpwood production area. Cogongrass was not a

problem in this area until approximately 1975. A lightning

strike caused the site to burn following harvest of mature

trees, land preparation, and planting of seedling pines.

With the competing vegetation removed, cogongrass dominated

this eight hectare area within a year. This infestation

then spread to adjoining cleared areas and throughout mature

pines to cover an area of approximately 20 ha. Other species

present were dogfennel [Eupatorium capillifolium (Lam.)Small

# EUPCP], partridgepea (Cassia fasciculata Michx. # CASFA),

saw palmetto [Serenoa repens (Bartr.)Small # SERRE], and

maidencane (Panicum hemitomon Schultes # PANHE).

The infestation was divided into two equal size areas.

Area A was known to be the older, more established section

from which Area B originated. With the exception of January,

April, and December, 1987, monthly samples were taken from

each of these sections. For each sampling time, six soil

cores from each area were taken using a bucket auger 182 cm2

in surface area to a depth of 23 cm. Rhizomes which were

not decayed were removed and dried at 100 C for 24 hrs.

From these samples rhizome biomass on a per hectare basis

was computed.

Analysis of variance (Helwig and Council, 1982) was

used to test for main factor (area and sampling time)

effects and the interaction. Least significant difference

(LSD) procedure was used to compare rhizome biomass between

areas within a sampling time.

Sequential Herbicide Application Programs.

To determine the influence of combinations of initial

and sequential herbicides on cogongrass control, two experi-

ments were conducted over a three-year period at Chiefland,

Florida. The area was located in a flatwoods, pulpwood

production site. The soil type was a Sparr fine sand

(loamy, siliceous, hyperthermic, Grossarenic Paleudults).

The herbicides and rates utilized for these studies were

selected from a herbicide screening field study conducted in

1983-1984 (data not shown). The herbicides and rates applied

were: 1) imazapyr at 0.8 kg (active ingredient) ai'ha-1, 2)

glyphosate at 3.4 kg ai'ha-1, 3) dalapon at 16.8 kg ai'ha-

4) sulfometuron, 2-[[[[(4,6-dimethyl-2-pyrimidinyl)amino]-

carbonyl]amino]sulfonyl]benzoic acid, at 1.1 kg ai'ha- and

5) untreated control. Initial herbicide applications were

made to cogongrass foliage that was 60-90 cm tall, and

sequential treatments were made to regrowth that was 60-120

cm tall. Treatments were applied using a C02-pressurized
-1
backpack boom-sprayer calibrated to deliver 280 L'ha- at

210 kPa. A three by five factorial design (initial treat-

ments of imazapyr, glyphosate, and sulfometuron and sequen-

tial treatments of imazapyr, glyphosate, dalapon, sulfomet-

uron, or untreated) was utilized with three replications on

1.8 m x 4.6 m plots. The first experiment was located in

Area A of the infestation and the second experiment in Area

B. Initial and sequential applications for the first experi-

ment were made on July 9, 1985, and September 19, 1986,

respectively. In the second experiment, initial and sequen-

tial applications were made on September 19, 1986, and

October 8, 1987, respectively. In January, 1986, (first

experiment) and January, 1987, (second experiment) plots were

mowed to ground level. This allowed the sequential applica-

tions to be made to foliage regrowth without interference

from persistent dead foliage. In January, 1987, (first

experiment) and January, 1988, (second experiment) a 1.8 m

swath was mowed through the plots. From this area foliage

regrowth within a 0.25 m2 area was harvested. Rhizome

biomass was harvested from within these quadrants by taking

.a single soil core in the same manner as that described in

the rhizome biomass sampling study. Untreated control plots

were also harvested and used to calculate inhibition of

foliage and rhizome dry weight. These harvests were made on

September 17, 1987, (first experiment) and June 23, 1988

(second experiment). All samples were dried at 60 C for 72

hrs. This substantial regrowth period was necessary in

order to provide an accurate assessment of the long-term

effect of these herbicide programs on cogongrass.

Analysis of variance was used to test for main factor

(initial and sequential herbicide treatments and experiment)

effects and the interactions. Least significant difference

(LSD) procedure was used to separate treatment means.

Tank-mix combinations of imazapyr and glyphosate.

To determine the influence of tank-mixes of imazapyr

and glyphosate on cogongrass control, two experiments were

conducted during a three-year period. Single applications

were made to cogongrass foliage on July 9, 1985, (first

experiment Area A) and on September 16, 1986, (second

experiment Area B). Applications were made using the

CO2-pressurized boom system previously described. The

absolute amount of each herbicide added to the tank-mix was

computed by designating 1.1 and 3.4 kg ai'ha-1 of imazapyr

and glyphosate, respectively, as the 100% dosage. Using

these rates imazapyr + glyphosate mixtures of 0 + 100, 25 +

75, 50 + 50, 75 + 25, and 100 + 0% were formulated. The

experimental design utilized was a randomized complete block

with three replications on 1.8 m x 4.6 m plots. Plots were

mowed in January, 1986, and January, 1987, for the first and

second study, respectively. Both foliage regrowth in a

0.25m2 area and a rhizome-soil core from within this area

were harvested on September 17, 1987, for the first experi-

ment. For the second experiment, only foliage regrowth was

harvested on June 23, 1988. Untreated control plots were

harvested for both studies and this data was used to compute

inhibition of foliage and rhizome dry weight.

Analysis of variance was used to test for treatment and

experimental effects and for the interaction. Least signif-

icant difference (LSD) procedure was used to separate

treatment means.

Comparison of low and conventional volume applications of

imazapyr and glyphosate.

To determine the influence of carrier volume on the

efficacy of imazapyr and glyphosate, two experiments were

conducted during a three-year period. Imazapyr and

glyphosate were applied in 46 and 234 L'ha-1. Glyphosate

was applied at 1.7 and 3.4 kg ai'ha-1 and imazapyr was

applied at 0.4 and 0.8 kg ai*ha-1. In the first experiment,

applications were made using a tractor-mounted boom sprayer

travelling at 6.4 km'hr-1. To apply 46 L'ha-I the boom was

equipped with 11001LP2 flat fan nozzles calibrated at 124 kPa

pressure. To apply 234 L'ha-1 the boom was equipped with

110052 flat fan nozzles calibrated at 276 kPa. In the second

experiment a C02-pressurized backpack boom sprayer was

.utilized. Using this system, 46 L'ha-1 was applied by using

TX-62 hollow cone nozzle calibrated at 207 kPa and travelling

at 8 km'hr-1 while, 234 L'ha-1 was applied by using 110032

flat fan nozzles calibrated at 221 kPa and travelling at 4.8

km'hr-1.

Treatments were applied on July 9, 1985, (first experi-

ment) and September 16, 1986, (second experiment). The

experimental design utilized was a 23 factorial with three

replications. Plots were 3.0 x 6.1 m and 1.8 x 4.6 m in the

first and second experiment, respectively. Both studies

were conducted in Area B of the infestation and cogongrass

was 60-90 cm tall at the time of application. In January,

1987, and January, 1988, plots were mowed. Regrowth

foliage and soil-rhizome cores were harvested, as previously

described, on September 17, 1987, and June 23, 1988, for the

first and second experiment, respectively. Untreated

control plots were harvested for use in computing inhibition

values.

Data were subjected to analysis of variance to test for

significance of herbicide, herbicide rate, carrier volume,

and experimental effects and the interactions. Since there

2Spraying Systems Co., North Avenue, Wheaton, IL 60188.

was no experimental interaction, data for the two experiments

experiments were pooled. Treatment means were separated

using least significant difference (LSD) procedure after

appropriate sorting.

Ropewick applications of imazapyr and glyphosate.

To determine the influence of ropewick-applied imazapyr

and glyphosate on cogongrass control, two experiments were

conducted during a three-year period. The effects of

concentrations of imazapyr and glyphosate (33 and 50% v/v)

applied once or twice (in opposite directions) within a plot

were evaluated.

The ropewick apparatus utilized was 2.1 m long with a

reservior capacity of 17.3 L. Two rows of half-overlapping

Pistachio3 rope (nine sections per row each 20 cm long) were

attached using rubber bushings within a screw-cap compression

fitting, the body of which was glued to the PVC reservior

creating 1.8 m of wicking surface. Two sets of ropes were

constructed, one for each herbicide. Applications were made

by two people carrying the ropewick through the plot at 4.8

km*h-1 with the wicking surface held approximately 150 below

perpendicular to the leaf surface.

The experimental design utilized was a 23 factorial

with three replications in the first experiment and four

replications in the second experiment. The plot size in

both experiments was 1.8 x 4.6 m. Both experiments were

Gulf Rope and Cordage Inc., P.O. Box 5516, Mobile, AL
36605.

conducted in Area A of the infestation. Applications of

both herbicides were made on July 9, 1985, for the first

experiment. In the second experiment, glyphosate was

applied October 2, 1986, and imazapyr was applied October 3,

1986. In both experiments, cogongrass was 60-90 cm tall at

the time of application. In January, 1987, and January,

1988, plots were mowed. Regrowth foliage was harvested, as

previously described, on September 17, 1987, and June 23,

1988, for the first and second experiment, respectively.

Untreated control plots were harvested for use in computing

inhibition values.

Data were subjected to analysis of variance to test for

significance of herbicide, herbicide concentration, number

of passes, and experimental effects and the interactions.

Since there was no experimental interaction, data for the

two experiments were pooled. Treatment means were separated

using least significant difference (LSD) procedure after

appropriate sorting.

Imazapyr longevity bioassay.

To determine the persistence of imazapyr under Florida

conditions and the effects of residual imazapyr on four

seedling grass species frequently used for roadside cover,

an imazapyr rate titration study and a field soil bioassay

study were conducted. Of the four species used, two

(Italian ryegrass, Lolium multiflorum Lam., and browntop

millet, Brachiaria ramosa (L.)Stapf.) are annuals used to

establish rapid cover in reestablishment areas and two

(Pensacola bahiagrass, Paspalum notatum var. saurae Parodi,

and common bermudagrass) are perenniating species established

from seed. The mixture of the four species is dependent on

the time of year and the location.

Greenhouse experiments were conducted to develop stan-

dard curves for inhibition of dry weight of each species.

Field soil (containing cogongrass leaf and rhizome matter)

from the Chiefland site was placed in 1.0 L cups. Imazapyr

was surface-applied at 0, 0.001, 0.01, 0.1, 1.0, and 10.0 kg

ai'ha-1 using a CO2-pressurized micro-applicator4 calibrated

to deliver 376 L'ha-1 at 276 kPa. A 0.5 cm layer of silica

sand was put on the soil surface and 30 seeds of each

species (1 species per pot) were planted and covered with

and additional 1.0 cm of sand. Plants were watered daily

and maintained in a greenhouse with the following environ-

mental conditions; day temperature 35 5 C, night tempera-

ture 25 5 C, light intensity at noon 1000 pE'm"s- ,

and a 16-hr light/8-hr dark photoperiod. At thirty days

after treatment, the plants were cut at the soil surface and

dried at 100 C for 24 hours. Each treatment was replicated

four times and the experiment was repeated.

Regression analysis was used to determine the best-
2
fitting relationship (based on F-values, r values, and

significance of the equation-parameter estimates) between

imazapyr rate and inhibition of dry weight (IDW) for each

R & D Sprayers, Opelousas, LA 70570.

species. Both log-transformed and nontransformed herbicide

rates and plant responses were used to determine the best

model. The following models were chosen based on this

criteria:

Equation [1]:

Annual ryegrass (AR):

Logl0JIDWI = 1.74 + 0.22 (Log10Rate) r2 = 0.96

Pensacola bahiagrass (PB):

IDW = -36.1 54.4 (Logl0Rate) 11.1 (Logl0Rate)2

r2 = 0.99

Common bermudagrass (CB):

IDW = -69.6 28.2 (Logl0Rate) r2 = 0.94

Browntop millet (BTM):

IDW = -5.8 71.1 (Rate) + 6.2 (Rate) r = 0.99

These data were utilized to: 1) determine that browntop

millet (based upon the model significance, its intermediate

response to imazapyr, and its consistently high germination)

would be used in the field soil bioassay, and 2) develop

relational models between the response of browntop millet

with the other species. These relationships are:

Equation [2]:

IDWAR = 3.6 + 0.98 (IDWBTM) r2 = 0.85

IDWpB = 34.0 + 1.25 (IbWBTM) r2 = 0.97

IDWCB = -5.9 + 0.9 (IDWBTM) r2 = 0.74

To determine the dissipation rate of imazapyr under

Florida conditions following treatment of cogongrass, a 0.1

ha block was treated with 1.1 kg ai'ha- imazapyr on

September 16, 1986. Application was made using a CO2

pressurized backpack boom-sprayer calibrated to deliver 235

L'ha- at 221 kPa. From this area four PVC columns (16 cm

in surface area and 23 cm long) were extracted at 2, 4, 5,

6, 8, 9, 10, 11, 12 and 14 months after treatment, sealed

and stored at 0 C until all samples had been retrieved.

Samples were allowed to thaw for 24 hrs at which time

the top 15 cm was removed intact and placed vertically in

1.0 L cups. The fact that imazapyr is more or less immobile

in soil would indicate that the active ingredient would be

in the top portion of these soil columns (American

Cyanamide, 1983). Four untreated pots were included to

calculate IDW. Thirty browntop millet seeds were planted

and covered with 1.0 cm sand. These pots were maintained in

a greenhouse with environmental conditions similar to those

described for the titration study. Thirty days after

planting above-ground biomass was harvested and dried.

The response (IDW) data for browntop millet bioassay

were used to develop a regression model for predicting IDW

of browntop millet at a given time after application. The

following equation describes that relationship:

Equation [3]:

IDWBTM = -91.2 + 7.7 (Months after treatment)

r2 = 0.80

From equations [2] a desired response (IDW) is set (in this

case, the 0, 25, and 50% IDW) for each of the three species

and the corresponding repsonse (IDW) for browntop millet is

determined. This response is used in equation [3] to

determine the TO (time required to allow imazapyr to

dissipate before the IDW of a given species equals 0), T25,

and T50 values for each of the species.

Results and Discussion

Cogongrass Rhizome Biomass Sampling.

Table 3.1 presents the rhizome biomass harvested from

the two areas of the infestation used for the control

studies. There was a significant interaction between area

and time of sampling. Overall, Area A, the older section of

the infestation, contained a greater amount of rhizome

biomass than did Area B. Both areas tended to be cyclic in

terms of rhizome production. The greatest quantity, in both

areas, was harvested during the late fall in both 1986 and

1987. The least amount of rhizome was present during the

late spring and early summer. This phenomenon has been

documented for other perennial grasses (Johnson and

Buchholtz, 1962; Horowitz, 1972). It has also been shown

that during regrowth periods (i.e., in the spring or

following mowing) there is a quantitative shift from carbo-

hydrate storage to remobilization for leaf production

(Potter, et al., 1986; Arny, 1932). There is also a shift

from rhizome biomass production during periods of rapid leaf

Table 3.1. Cogongrass rhizome biomass collected in
Chiefland, FL from 1986 to 1988.

Month Year AREA A AREA B LSD

October

November

December

February

March

May

June

July

August

September

October

November

January

February

March

April

1986

1986

1986

1987

1987

1987

1987

1987

1987

1987

1987

1987

1988

1988

1988

1988

- - kg'ha-

7355 (387)2

6898 (521)

6734 (318)

7465 (707)

7355 (631)

4623 (537)

7218 (396)

7300 (66)

7163 (275)

8031 (544)

8598 (495)

8159 (258)

7076 (428)

7044 (420)

7181 (333)

6012 (343)

1
Significance levels for rhizome biomass differences between
areas within a sampling time using LSD are 0.01 (***),
0.05 (**), and nonsignificant (NS).
2
Mean rhizome dry weight of 6 replicates/area computed from
4170 cubic cm soil cores followed by the standard error of
the mean.

NS

***

5592

7328

4797

3984

4806

2960

3682

4806

4185

4020

6789

4879

4061

4651

5217

4048

(360)

(443)

(398)

(305)

(212)

(181)

(214)

(272)

(398)

(500)

(504)

(197)

(97)

(198)

(303)

(135)

***

area expansion which then reverses after sufficient leaf

area has been produced.

The efficacy of translocated herbicides (e.g.,

glyphosate) has been shown to be affected by the stage of

development of perennial species (Atkinson, 1985). The best

long-term control is attained when perennial grasses are: 1)

utilizing rhizome for expansion purposes, thereby not

harbouring large numbers of metabolically quiescent nodes

and 2) transporting photosynthate more to the perenniating

organs than to expanding leaf tissue, which would occur

during the latter part of the growing season. Relating

these factors and the characteristics of the infestation

will be used to explain year-to-year interactions that

occurred in the following control studies.

Sequential Herbicide Application Programs.

There was a significant three-way interaction between

initial herbicide treatment, sequential herbicide treatment,

and experiment. Therefore, experiments were analyzed

separately. Tables 3.2 and 3.3 present the cogongrasss leaf

and rhizome inhibition data for the first and second experi-

ment, respectively.

Overall, control was superior in the second experiment.

Two factors may have contributed to this. First, the two

areas within the infestation were shown to be significantly

different in terms of rhizome biomass, with Area A (first

experiment) containing a greater quantity (Table 3.1).

There was also a greater amount of leaf material present in

Table 3.2. The effect of sequential herbicide treatments on
cogongrass foliage and rhizome dry weight (Experiment 1 -
1985 through 1987 Area A).

Initial Application

(kg ai'ha1)

Sequential Imazapyr Glyphosate Sulfometuron
Application (0.8) (3.4) (1.1)

(kg ai'ha -1 FDW1 RDW2 FDW RDW FDW RDW
----% Inhibition -- -

Imazapyr
(0.8)

Glyphosate
(3.4)

Dalapon
(16.8)

Sulfometuron
(1.1)

100

82

64

55

100 -37

88 -42

-17

-25 -48

28 -71

Untreated

LSD(0 1)Initial 20 30

LSD(0.1)Sequential 25

Cogongrass foliage dry weight (regrowth) harvested from
0.25 square m quadrants at 12 months after sequential
herbicide application.
2
Cogongrass rhizome dry weight harvested from 4170 cubic cm
soil cores at 12 months after sequential herbicide
application.

Inhibition values computed using untreated plots which
contained 53.5 and 3.9 g of foliage and rhizome,
respectively.

Table 3.3. The effect of sequential herbicide treatments on
cogongrass foliage and rhizome dry weight (Experiment 2 -
1986 through 1988 Area B).

Initial Application

(kg ai'ha1)

Sequential Imazapyr Glyphosate Sulfometuron
Application (0.8) (3.4) (1.1)

(kg ai'ha-1) FDW1 RDW2 FDW RDW FDW RDW
- % Inhibition3

Imazapyr 98 68 100 78 89 68
(0.8)

Glyphosate 94 59 98 29 100 15
(3.4)

Dalapon 84 55 100 60 43 -49
(16.8)

Sulfometuron 68 78 91 64 12 -29
(1.1)

Untreated 96 66 83 69 48 28
----- ---- ---- -- -- -- --m - ---- -- -- --
LSD(0.1)Initial 13 13

LSD(0.1)Sequential 16 16

1Cogongrass foliage dry weight (regrowth) harvested from
0.25 square m quadrants at 10 months after sequential
herbicide application.

2Cogongrass rhizome dry weight harvested from 4170 cubic cm
soil cores at 10 months after sequential herbicide
application.

3Inhibition values computed using untreated plots which
contained 10.8 and 3.3 g of foliage and rhizome,
respectively.

the untreated controls of Area A (first experiment). There-

fore, a denser, more established stand would be more diffi-

cult to control. Second, the weather patterns occurring

during the two experiments were different (Figure 3.1).

During the first experiment, a dry summer 1985 was followed

by normal precipation through 1986. In the second

experiment, there was adequate rainfall through the summer

and fall 1986, but was followed by a below normal rainfall

through the summer and fall 1987 except during the month

prior to the sequential applications. Weather patterns

could have adversely affected herbicide efficacy in the

first experiment. First, the effect of the initial

applications may have been reduced by drought-stressed

cogongrass which was demonstrated by the relatively poor

control obtained from single applications. Second, adequate

rainfall following the sequential application favored

cogongrass regrowth (which is evident from the 53.5 g of

foliage regrowth in the untreated control plots). In the

second experiment, initial applications to vigorously

growing cogongrass were quite efficacious. This is

demonstrated by the excellent control obtained from single

applications of imazapyr and glyphosate. Though the

sequential applications were made during an overall dry

period, the rainfall during September and early October was

adequate to maintain cogongrass growth. After this time,

rainfall was low which resulted in suppressed cogongrass

growth. This evident from the 10.8 g of foliage regrowth in

Gainesville, Florida Weekly Precipitation
CM 1986
12.71

1

U,

7.6 -

5.1 -o. --
2.5 ,

Jan Feb Mar Apr May Jun Jul Aug Sept Oct leyv Dec

Total Precipitation 132.9 cm

Gainesville, Florida Weekly Precipitation
cm 1987

Jan Feb Mar Apr May Jun Sl Aug SCp OC Nov Dec
Total Precipitation 115.5 cm

Figure 3.1. Weekly precipitation in Gainesville, FL, in
1986 and 1987 (Adapted from McCloud and Hill, 1987 and
McCloud and Hill, 1988).

~rr r

the untreated control plots. The combination of a less

dense stand (Area B) and the efficacious initial herbicide

treatments followed by less than favorable conditions for

regrowth following the sequential applications resulted in

better overall control in this experiment.

In both experiments two compounds, imazapyr and

glyphosate, were most effective. However, single applica-

tions provided adequate control only in the second experi-

ment. The response of cogongrass has been shown to be

impacted by the overall soil-water status. Arif (1979)

found that cogongrass control acheived with glyphosate was

greater during the dry season (unless there was serious

water stress) than during the rainy months. This may

explain the improved glyphosate activity in the second

experiment. In the first experiment, single applications of

these herbicides had no effect on rhizome growth and

consequently poor control of foliage regrowth was obtained.

The use of dalapon as a sequential treatment provided some

additional control in the first experiment when glyphosate

or sulfometuron were used as an initial treatment, but none

when imazapyr was used. In the second experiment, dalapon

provided no additional control when used as a sequential

treatment. Sulfometuron apparently had no adverse effect on

cogongrass. In fact, there was significant stimulation of

rhizome growth in both experiments. This may have occurred

not because of a direct chemical effect on cogongrass, but

from the control of broadleaf species that may have been

competing with cogongrass.

These experiments indicated that near 100% control of

cogongrass on a long-term basis may be consistently

achieved, regardless of extraneous field and climatic

factors, only with an initial and sequential herbicide

.program based on imazapyr and glyphosate.

Tank-mix combinations of imazapyr and glyphosate.

Table 3.4 presents cogongrass control data obtained

from tank-mixes of imazapyr and glyphosate. There was a

significant interaction between treatment and experiment,

therefore experiments were analyzed separately. Overall,

control of foliage regrowth was superior in the second

experiment. This occurred possibly for the same reasons as

those proposed for the sequential herbicide programs study.

In the first experiment no clear trend is apparent for the

level of control achieved and the herbicide components.

This indicates that in-tank antagonistic interactions did

not occur. In the second experiment control of foliage

regrowth was greater from tank-mixes with 50% or more

imazapyr. The results of the two experiments may indicate

that these two herbicides were additive in their effect on

cogongrass, with imazapyr being more active. These data

also indicate that the use of imazapyr + glyphosate ratio

tank-mixes did not alleviate the need for multiple

applications to consistently achieve long-term control.

Table 3.4. The effect of imazapyr and glyphosate tank-mixes
on cogongrass foliage and rhizome dry weight.

Tank-mix Treatments
Imazapyr + Glyphosate

Experiment 1
1985-87
(Area A)

FDW1

RDW2

Experiment 2
1986-88
(Area B)

FDW

- kg ai'ha-1
- kg a i'ha

1.1

0.8

0.6

0.3

0.8

1.7

2.5

3.4

3nhibition
- % Inhibition

60 36 82
- -

LSD(0.1) 33 40 9

Cogongrass foliage dry weight (regrowth) harvested from
0.25 square m quadrants at 26 and 21 months after treatment
for the first and second experiment, respectively.
2
Cogongrass rhizome dry weight harvested from 4170 cubic cm
soil cores at 26 and 21 months after treatment for the
first and second experiment, respectively.

Inhibition values computed using untreated plots which
contained 81.0 and 13.5 g of foliate and rhizome,
respectively in experiment 1 and 17.3 g of foliage in
experiment 2.

- -

Comparison of low and conventional volume applications of

imazapyr and glyphosate.

There was not a significant interaction between experi-

ment and the other factors (volume, rate and herbicide).

However, experiments were significantly different (P < 0.05).

The differential rainfall discussed previously could have

caused the overall increased control acheived in the second

experiment, but the efficacy of the treatments was not

differentially affected. There was a significant two-way

interaction between herbicides (imazapyr or glyphosate) and

carrier volume (46 or 234 L'ha1 ) for the inhibition of

cogongrass foliage regrowth. These data are presented in

Table 3.5.

Imazapyr provided significantly greater control than
-1
glyphosate at both rates when applied in 234 L'ha1 water.

For both herbicides, rate did not influence control.

However, in the case of glyphosate, the half rate was not

significantly different from the untreated control on an

absolute foliage dry weight basis, but the higher rate did

cause a significant reduction in foliage dry weight as

compared to the untreated control plots. In either case,

control obtained using glyphosate was poor. Imazapyr

provided a greater level of control when applied in 234

L'ha-1 volume application than when applied in 46 L'ha-.

This may indicate that in dense stands of cogongrass the

greater foliage coverage attained with this volume is more

important than the enhanced activity that has been reported

43

Table 3.5. The effect of carrier volume and imazapyr and
glyphosate rate on cogongrass foliage dry weight.

Imazapyr

Carrier
Volume
(L'ha-1)
(L'ha )

Glyphosate

-i
kg ai'ha1

0.4 0.8 1.7 3.4
1
-% Inhibition

333

732,3

234

483

712,3

122

Inhibition values computed using untreated plots which
contained 24.8 g of foliage regrowth/0.25 square m quadrant.

2Means within a given volume and herbicide rate indicates
that herbicides are significantly different using LSD(0.1).

Means with a given herbicide and herbicide rate indicates
that volumes are significantly different using LSD(0.1).

-102

to occur with ultra-low volume applications. This study

again points out the necessity of multiple applications to

achieve long-term cogongrass control.

Ropewick applications of imazapyr and glyphosate.

There was no significant interaction between

experiments and the tested factors (herbicide, concentration,

and number of passes). However, experiments were signifi-

cantly different (P < 0.05). There was a significant three-

way interaction between herbicides (imazapyr or glyphosate),

herbicide concentration (33 or 50% v/v), and number of passes

for the inhibition of both cogongrass foliage regrowth and

rhizome dry weight. These data are presented in Table 3.6.

Control of cogongrass rhizomes with glyphosate was

unaffected by the concentration or the number of passes

made. This may have occurred due to the solutions being too

concentrated. Boerboom and Wyse (1988) speculated that the

reason for poor control of Canada thistle [Cirsium arvense

(L.)Scopoli] achieved using ropewick-applied glyphosate was

that the concentrations being applied were greater than the

optimum level needed to maximize translocation to the roots.

In essence, by overdosing the leaf tissue, glyphosate

provides more contact activity to the detriment of systemic

activity. Therefore, the use of glyphosate at concentra-

tions greater than 33% would be economically unsound. In

fact, the use of concentrations lower than 33% may provide

as much, if not more, control of cogongrass.

Table 3.6. The effect of ropewick applications of imazapyr
and glyphosate on cogongrass foliage and rhizome dry
weight.

Imazapyr Glyphosate

.% Solution # Passes FDW RDW2 FDW RDW

- % Inhibition3

33 1 67 -175 55 17

33 2 68 115 65 31

50 1 764 245,6 384 28

50 2 784 455,6 604 27

Cogongrass foliage dry weight (regrowth) harvested from
0.25 square m quadrants.

Cogongrass rhizome dry weight harvested from 4170 cubic cm
soil cores.

Inhibition values computed using untreated plots which
contained 43.6 and 5.8 g of foliage and rhizome,
respectively.

Means within a given % solution and number of passes
indicates that herbicides are significantly different using
LSD(0.1)

5Means within a given herbicide and number of passes
indicates that % solutions are significantly different
using LSD(0.1).

6Means within a given herbicide and % solution indicates
that number of passes are significantly different using
LSD(0.1)

Imazapyr at the 50% concentration provided better

control of cogongrass foliage regrowth than did glyphosate

at this concentration. Unlike glyphosate, imazapyr efficacy

on rhizome growth was significantly affected by concentration

.and number of passes. The highest level of control was

achieved at 50% v/v applied twice. This would seem to

indicate that imazapyr adsorption and/or translocation is

not as sensitive to high concentrations as glyphosate. The

use of two passes at the 50% concentration provided the

highest level of control of any of the treatments.

Although 100% control was not acheived from either

herbicide, multiple ropewick applications of glyphosate or

imazapyr may be a viable alternative for cogongrass control

in situations where broadcast applications are not

desirable.

Imazapyr longevity bioassay.

Data from the imazapyr bioassay are presented in Table

3.7. The most tolerant of the four species tested was

bahiagrass with an 150 rate of 1.40 kg ai'ha-1. Common

bermudagrass and browntop millet were intermediate in their

response and annual ryegrass was the most sensitive. The

half-life of imazapyr in sandy soils of Florida was found to

be approximately eight months. Research has shown that under

temperate conditions imazapyr persists from three months to

one year, while under tropical conditions biological activity

persists for three to five months (American Cyanamide, 1983).

The fact that the applications were made late in the growing

Table 3.7. The influence of imazapyr dissipation on the
inhibition of dry weight (IDW) response of four grass
species.

1
50
Species kg ai'ha

2
T T T
0 T25 T50

Months After Treatment

Annual ryegrass (AR) 0.65 11.4 8.0 4.7

Pensacola bahiagrass (PB) 1.40 8.3 5.7 3.1

Common bermudagrass (CB) 0.70 12.7 9.1 5.5

Browntop millet (BTM) 0.70 11.8 8.6 5.4

1Imazapyr rate required to cause 50% IDW.
2
Dissipation time (months) required to cause 0, 25, and 50%
IDW of each species.

season (September 16) should be taken into consideration as

degradation by soil microbes would have been reduced during

the cold, dry winter months. However, research has shown

that the major route of imazapyr degradation is via photo-

lysis (American Cyanamide, 1983). Photolysis of imazapyr in

the soil would probably be reduced by the dense and

persistent cogongrass foliage during the winter months.

Therefore, the conditions under which this study was

conducted would have tended to increase the persistence of

imazapyr.

Therefore, if imazapyr were to be used for cogongrass

control on highway rights-of-way, reestablishment of this

area with one of the four grass species tested would have to

be delayed until the late spring following the fall applica-

tion. The amount of imazapyr remaining at this time would

possibly be reduced by the removal of cogongrass foliage

after the first frost, thereby increasing imazapyr photolytic

breakdown. However, this would also reduce the amount of

imazapyr available for control of cogongrass regrowth. The

use of multiple applications (which may be necessary to

achieve 100% control) should also be considered. The use of

imazapyr as an initial treatment, followed by the use of

glyphosate or dalapon as a sequential spot treatment to

cogongrass regrowth following turf establishment may provide

the highest level of control with the shortest reestablish-

ment period.

CHAPTER 4
THE INFLUENCE OF MECHANICAL AND CHEMICAL INPUTS ON
COGONGRASS [Imperata cylindrica (L.)Beauv.] CONTROL.

Introduction

In the tropical South Pacific and Southeast Asia

cogongrass (known there as alang-alang or lalang-alang)

covers several million hectares (Holm, et al., 1977). These

developing countries, whose economies are based largely on

agriculture, do not have the capital, expertise, or tech-

nology necessary to effectively control cogongrass on a

large-scale basis (Holm, 1969). Therefore, localized

efforts relying on slash-and-burn, grazing, and tillage are

the most widely used control methods. Controlled studies on

the effectiveness of these techniques have indicated that;

1) shifting agriculture may provide short-term control (long

enough to produce one or two.crops-prior to reinfestation),

but in the long-term, cogongrass populations are increased,

2) repeated burning followed by grazing will marginally

support animal production, but provides little in the way of

control, and 3)- intensive tillage has repeatedly shown to be

an effective method off bogongrass management, however the
C I **<;
availability of ilaplmefts, soil type climatic conditions,

Said. tera.n limit it use. in southeast Asia (Falvey, et al.,

':Ip~l, fl8xiea j 2,833; Prorasool, 1184). ,*

.. ... ..
.. .. .. ... W.FF;
TE: .~:

:, ." .....
J .: ....=

The recent introduction of several new (and more .

efficacious) herbicides in these areas have demonstrated:.the

utility of chemical control (Arif, 1979; Bacon, 1986).

However, application technology and expertise in the proper,

and most effective use of these compounds is limited. Much

of the chemical control conducted was (and still is) with

dalapon, 2,2-dichloropropanoic acid, paraquat, 1,1'-dimethyl-

4,4'-bipyridinium ion, and glyphosate, N-(phosphonomethyl)-

glycine applied with knapsack sprayers or antiquated high-

volume spray systems (Hartley; 1949, Keepings and Matheson;

1949; Sandanam and Jayasinghe, 1977; SEAWIC, 1987; Seth,

1970). Multiple applications are expensive, but a require-

ment to acheive any significant control and then long-tend

control is seldom realized.

The integration df til age :and herbicide treatments has

been used in the United States :for the control of several

perennial weeds. Rhizome johnsongrass [Sorghum halepense

(L.)Pers. # SOP A] control programs based on dishing after

corn (Zea mays L.) harvest, fellow applications of

glyphosate to re growth, and tri lra li F 2,6-dinitto-,N--
": *" .. 4
A.*
dipropyl-4-(tri-fluoromethyl) benzenaWi nr treat ent'at twice
.. .. : .. :
the normal use rate :: io to soybean. [Gi a;. (L.)Mer;.]
4 4 ....F'" ". ," : i,":""..a. ;. :.::. i .. ..:
planting provides excellent a ontol (7LA1O~ 1.87; lonanto,

1988)., Programs ustni il Ie e either _91r following

. ..' ,* .. .. ..,...:..
.. .. .. .
Letters follow ing thseb a vd
code from the oapo.dite, 4Psdst t N ..
2. Availabib frin SSA, 3C9WMw J. 0.201&
LX 61820. ." ""
S ... .. .... ."
A;, 2pr. :t':.j .J..
.."t rr-'.. i'. ... 2. "" "" :" """ ..~ a.'n.: .': :,"i

glyphosate applications have been utilized for control of

other perennial weed species (Baird and Begeman, 1972;

Sandberg and Meggitt, 1977).

Mowing has been shown to have only short-term effects

Son perennial grass growth (Beard, 1973). However mowing is

utilized extensively in Florida for right-of-way and pasture

management. If chemical control is to be effectively

interfaced with mowing then the combined effects of mowing

and herbicides needs to be investigated.

The objectives of these studies were to determine the

influence of mowing and disking on the efficacy of selected

herbicides. Studies were also conducted to determine the

influence of light and photoperiod on shoot initation in

cogongrass rhizome segments and to determine the influence

of cogongrass stage of development and defoliation on the

efficacy of glyphosate.

Materials and Methods

Integration of mowing and disking with herbicide treatments.

To determine the interaction of mowing or disking with

herbicide treatments on long-term cogongrass control, two

studies were conducted from 1985 to 1988 at Chiefland, Levy

County1. Florida. Studies were located on an eight hectare

cogongrass infestation that had became established during

the: mid-l970. The soil type was a Sparr fine sand (loamy,

*siliceo per thermic, Grssaarenic Palenulduts),.
.' :. :* 1:. *c ; .:..
;;.. ::1:ib ......... .. :.ii: ." ". .... *. *
,A '.. 1 ...
iii.1 ,
:.4 :"

The experimental design was a split-plot with the

mechanical treatment (mowing or disking) being the main plot

and herbicide treatment as the subplot. Three levels of

mowing or disking were used: 1) no mechanical treatment, 2)

mowing or disking prior to herbicide treatment, and 3)

mowing or disking prior to and subsequent to herbicide

treatment. Five herbicide treatments were applied: 1)

untreated control, 2) imazapyr, () -2-[4,5-dihydro-4-methyl-

4-(1-methylethyl)-5-oxo-lH-imidazole-2-yl]-3-pyridinecar-

boxylic acid, applied at 0.8 kg ai'ha-1, 3) glyphosate

applied at 3.4 kg ai'ha-1, 4) dalapon applied at 16.8 kg

ai'ha-1, and 5) sulfometuron, 2-[[[[(4,6-dimethyl-2-pyrimi-

dyl)amino]carbonyl]amino]sulfonyl]-2-thiophenecarboxylic
-1
acid, at 1.1 kg ai'ha1. In 1985 (first experiments) each

mechanical-chemical treatment was replicated three times on

1.8 m x 3.7 m plots. In 1986 (second experiments) each

treatment was replicated four times on 1.8 m x 4.6 m plots.

Mowing treatments were made using a tractor-mounted

horizontal impact-mower2 which was set to remove approxi-

mately the top half of the foliage then set to cut ground

level. The disking treatments were made by first mowing the

blocks followed by repeated disking until all vegetation was

destroyed and the soil was tilled to a depth of 10-15 cm.

Due to the dense rhizome layer, eight to ten passes per

block were required. Herbicide applications were made using

2Bush Hog Manufacturing, Inc., Selma, AL.

a CO2-pressurized backpack sprayer calibrated to deliver 280
-1
L'ha1 at 207 kPa.

On July 5, 1985, initial mowing and disking treatments

for the first experiments were implemented. Herbicide

treatments were applied on November 15, 1985, to cogongrass

that was 30 cm tall in the mechanically treated blocks and

100 cm tall in the undisturbed blocks. Following the first

freeze in January, 1986, all blocks were mowed to ground

level. On July 29, 1986, the sequential mowing and disking

treatments were implemented. For the second experiments in

each study, initial mowing and disking treatments were

imposed on September 16, 1986, followed by herbicide

applications on November 14, 1986. In January, 1987, all

blocks were mowed to ground level. Sequential mechanical

treatments were completed on June 25, 1987 (Table 4.1).
2
Cogongrass control was determined by harvest of 0.25 m

quadrants from the center of each subplot and a soil-rhizome

core was extracted from this quadrant using a bucket auger

182 cm2 in surface area to a depth of 23 cm. Rhizomes which

were not decayed were removed from these samples. These

harvests were made on September 17, 1987, (first experiments)

and on June 23, 1988 (second experiments). Foliage and

rhizome samples were dried at 60 C for 72 hrs. Foliage and

rhizome dry weights were converted to percent inhibition

based on the untreated controls. Data were subjected to

analysis of variance (Helwig and Council, 1982) to test for

single factor effects and interactions. In both the mowing

-.. "., .'-:. ....

:: ,. 5 ':: '';:"
4.

Table 4.1. Timing of mechanical and chemical treatments and
harvests performed for evaluation of integrated cogongrass
control experiments.

Operation
SPerformed

Experiment 1
1985-1987

Initial Mowing
and Disking

Herbicide
Applications

Mowing for Removal
of Dead Foliage

Sequential Mowing
and Disking

Foliage Regrowth
and Rhizome Harvest

7/5/1985

11/15/1985

1/15/1986

7/29/1986.

9/17/1987

Experiment 2
1986-1988

9/16/1986

11/14/1986

1/15/1987

6/25/1987

6/23/1988

I .

F- *

1*.

'* *'* **

:... ....
*' '> ..* :. .,." .
I ". .,

T- : ..
; ;,

,: .: ... 5.
** : : .... : .* : :- ..; : .:- y
. .. .: .. ". .* 1 J ". :.
.. .i ":... :. ; "," "::" : ... .
"',i .. .:,. ; ,. ... .

; .

.:."*k .
"ii

:
..s -,

I
E.

..

iPt
:.." ,E
i.*; ; ::;^ :

.*. ** :-. -t .

*" "Ii
.- ; *.. ; *'*.: ,

S:... .. ... .:. :. .
S:. .. .. ;...: ..... :. ". .. '" :.,. .'.." ; "

i :: ".. .*' .: ." H. : ..

... "i :
:.... '." :. '. :. ** .. *" ". .. .:.' :.. *** : *:.* : ..'
..." : : :... ..*' ..^ ^ A S .,. .. .

.;
.. ~ ~ ~ ~ .. ':E .. ." :. ..:. .; :!., .. . .* *
: .; 'I :: : : ", / ::; :

:.. ; :: .'" : : .. : :iii .' .; : E ..

--

:*
1
rr;

.1.

;;-

and disking studies there was a significant interaction (P <

0.05) between experiment and treatment, therefore experiments

were analyzed separately. Treatment means were separated

using least significant difference (LSD) procedure.

Influence of photoperiod on shoot production in cogongrass

rhizomes.

Growth chamber experiments were conducted to determine

the effect of photoperiod on shoot initiation in cogongrass

rhizomes. Rhizomes, harvested from an infestation at the

University of Florida Agronomy Farm in 1987, were washed,

de-scaled, and cut into six-node sections which showed no

evidence of injury (i.e., pathogen or insect damage). No

attempt was made to differentiate sections from various

positions on the rhizome. Five or six rhizomes sections

were placed on filter paper3 in petri dishes, 2 ml of

deionized water was added, and petri dishes were covered.

Dark controls were prepared first and immediately wrapped in

aluminum foil. This procedure took approximately ten

minutes during which time the dark controls were exposed to

light. Preliminary studies indicated that this amount of

exposure was insufficient to initiate shoots (data not

shown). Water was added to petri dishes during the light

portion of the photoperiod on the fourth day after planting.

Dark controls were watered at night in a darkened room. The

conditions in the growth chambers were 30 2 C and

3Whatman Paper Ltd., Maidstone, England.

fluorescent lights were used to obtain a light intensity of

200 pE'm-2s-. Two growth chambers were utilized for each

study. The first study evaluated continuous light and a

16-hr light/8-hr dark photoperiod. The second study

.evaluated a 9-hr light/15-hr dark photoperiod and a 8 + 1-hr

light(8 hr with a 1 hr light period during the dark

period)/15-hr dark photoperiod. The third study evaluated a

16-hr light/8-hr dark photoperiod and a 9-hr light/15-hr

dark photoperiod. Light treatments were replicated (5 petri

dishes within the growth chamber). Dark controls were

evaluated in each study and studies were repeated. One week

after photoperiod treatments were imposed the following data

were collected: number of productive segments, shoot number

per segment, and total node number.

Based on the dark controls only, all three studies

(each conducted twice) were not significantly different (P >

0.05), thus all the data were combined and analyzed.

Analysis of variance procedure was used to test for the

significance of photoperiod on shoot initiation. Duncan's

Multiple Range test was used to separate treatment means.

The influence of cogongrass stage of development and defoli-

ation on glyphosate efficacy.

Greenhouse studies on cogongrass growth and development

were conducted in 1987 and 1988 to characterize the effects

of defoliation at various times after planting. Cogongrass

was propagated in a growth chamber from rhizome segments

that were harvested from an infestation in Gainesville,

Florida in 1987. Rhizomes were cleaned, de-scaled, and cut

into six-node segments that were free of insect or pathogen

injury. Rhizomes were planted horizontally in trays filled

with 2 cm of vermiculite and covered with an additional 2 cm

of vermiculite. Rhizomes were watered daily. Growth chamber

conditions were 30 2 C with a 16-hr light/8-hr dark
-2. -1
photoperiod. Fluorescent lights provided 2001uE'm- s-

Two weeks after planting, nodes that had produced a shoot 5

cm long were cut from the six-node segments. These shoots

were planted into 2.8 L pots filled with 2.4 L of soil mix

containing 50% fumigated4 Arredondo fine sand (loamy,

siliceous, hyperthermic, Grossarenic Paledults) and 50%

commercial potting mix5 on a volume basis. Three shoots

were planted in each pot and fertilized with 100 ml of 5X

Hoagland's solution one week after planting (Hoagland and

Aronon, 1950). Plants were watered three times per week.

Greenhouse conditions were as follows: day temperature 27 -

5 C, night temperature 15 5 C, and light intensity at noon

of 1000 uE'm 2s-1 Four pots were harvested every three

weeks after planting. At these harvest periods four other

pots were defoliated to ground level. Defoliated pots were

harvested three weeks later. Harvests were carried out by;

clipping leaf tissue at ground level and measuring leaf

Brom-o-gas (methyl bromide + chloropicrin), Great Lakes
Chemical Corp., West Lafayette, IN 47906.

Metro-mix 200, Gracewood Horticultural Products, Cambridge,
MA 02140.

area6, composite root-rhizome biomass was washed and dried

at 100 C for 24 hrs and weighed, as was leaf tissue. Leaf

and root-rhizome dry weight were converted to ratios based

on total plant dry weight. This study was conducted twice.

To determine the influence of stage of development and

defoliation on glyphosate efficacy in cogongrass, greenhouse

studies were conducted during 1988. Cogongrass propagation

was identical to that used in the growth analysis study.

Three stages of cogongrass were used in these experiments.

Three-week-old plants (Stage A), six-week-old plants (Stage

B), and plants that were defoliated at three weeks and

allowed to regrow for three weeks (Stage C). To allow

glyphosate applications to be made at the same time to all

stages, cogongrass for stages B and C was planted first,

then three weeks later cogongrass for stage A was planted

(also at this time stage C plants were defoliated). Glypho-

sate was applied three weeks after the second planting.

Glyphosate solutions were prepared using technical

glyphosate (96.5% acid7). The surfactant MON-08187 (poly-

oxyethylene tallow amine) was added at 0.5% (v/v).
8
Solutions were applied using a single nozzle (TJ-60 8002E )

CO2-pressurized micro-applicator calibrated to deliver 860

Licor Model 3100 Area Meter, Licor, Inc., Lincoln, NE

7Monsanto Agri. Prod. Co., St. Louis, MO 63617.

8Spraying Systems Co., North Avenue, Wheaton, IL 60188.

R & D Sprayers, Inc. Opelousas, LA 70570.

L'ha- at 276 kPa. Glyphosate was applied at 0.1, 0.3, 0.6,

and 1.1 kg (acid equivalent) ae'ha- (first experiment) and

0.01, 0.06, 0.1, 0.3, and 0.6 kg ae'ha-1 (second experiment).

Untreated plants were also included in both experiments.

Treatments were replicated four times within each cogongrass

stage. At the time of application four pots from each stage

were harvested. The data collected from these harvests were

leaf area and dry weight and composite rhizome-root biomass.

Samples were dried at 100 C for 24 hrs. At three weeks

after application leaf tissue was cut off at ground level.

Plants were allowed to regrow for three weeks, at which time

leaf and root-rhizome tissue was harvested and dried.

Data was converted to percent inhibition of dry weight

change over the treatment period. Data were subjected to

analysis of variance procedure to test for single factor

effects and interactions. There were no significant inter-

actions (P > 0.05) between experiments and treatments or

experiments and cogongrass growth stages, therefore data

from the two experiments were combined. There was a

significant interaction (P < 0.05) between treatment and

cogongrass stage. Therefore, regression models (% inhibition

versus transformed glyphosate rate) were developed for each

growth stage separately. The best models were chosen based

on the significance level (P < 0.05) of the model, the

equation components and the r2 value.

Results and Discussion

Mowing-herbicide studies.

Data on the effects of mowing and herbicides on cogon-

.grass are presented in Tables 4.2 and 4.3. The experiments

are presented separately due to a significant (P < 0.05)

experiment by treatment interaction. This interaction may

have been a result of locational and/or environmental

factors enhancing the effect of effiacious treatments while

having little, if any, effect on treatments that provided

little control. The area in which the first experiment was

conducted contained five times more foliar biomass (95.8 g

vs. 15.7 g'0.25 m-2) and two times more rhizome biomass
-3
(15.0 g vs. 5.2 g'4170 cm-3) than the area in which the

second experiment was conducted. On a yearly basis, rain-

fall for 1986 was normal with no extended dry periods

(McCloud and Hill, 1987 and McCloud and Hill, 1988). Rainfall

in 1987 was 28 cm below normal with three extended periods

of near-zero rainfall during the growing season.

In the first experiment, there was a significant (P <

0.05) interaction between mowing and herbicide treatments on

inhibition of foliage dry weight (Table 4.2). Mowing had a

significant impact when no herbicides were applied or when

dalapon or sulfometuron were applied. With these two

herbicides, as the number of mowings increased control (%

inhibition) of foliage dry weight (FDW) also increased. The

use of mowing prior to sulfometuron treatment apparently

Table 4.2. The effect of mowing and herbicides on cogon-
grass foliage and rhizome dry weight (Experiment 1 1985
through 1987).

Mowings1

0 1 2
Rate
Rate-1 2 3
Herbicide kg ai'ha FDW RDW FDW RDW FDW RDW

-- % Inhibition4 - -

Imazapyr 0.8 94 44 89 61 95 42

Glyphosate 3.4 86 33 85 42 97 49

Dalapon 16.8 65 12 74 -3 90 24

Sulfometuron 1.1 11 -2 61 -2 79 52

Untreated 0 0 36 13 65 26

LSD(0.1)Mowings 7 18

LSD (0.1)Herbicides 9 24
1
0 mowings indicates herbicides applied to undisturbed
cogongrass, 1 mowing indicates mowing prior to herbicide
application, and 2 mowings indicates mowing prior and
subsequent to herbicide application.

2Cogongrass regrowth foliage dry weight harvested from 0.25
square m quadrants at 14 months after second mowing
operation.

Cogongrass rhizome dry weight harvested from 4170 cubic cm
soil cores at 14 months after second mowing operation.

Inhibition values computed using unmowed and untreated
plots which contained 95.8 and 15.0 g of foliage and
rhizome, respectively.

increased herbicide efficacy. This may have been due to the

fact that sulfometuron has both foliar and soil activity

(DuPont, 1982). By removing a large portion of the foliage

prior to application, more herbicide may have directly

.contacted the soil and was affecting emerging cogongrass

shoots at that site. Though sulfometuron does exhibit post-

emergence foliar activity in many species, it did not appear

to affect cogongrass in this manner. Subsequent mowing had

less of an affect than did prior mowing in conjunction with

sulfometuron. Mowing in conjunction with glyphosate

increased cogongrass control only when done both prior and

subsequent to herbicide treatment. Mowing did not have a

significant effect on the control of foliage regrowth by

imazapyr. Control with imazapyr was excellent regardless of

mowing treatment. Imazapyr, like sulfometuron, has both

foliar and soil activity (American Cyanamide, 1983). How-

ever, unlike sulfometuron, imazapyr exhibited significant

foliar activity on cogongrass.

Mowing had less impact on cogongrass rhizomes than

foliage. Two mowings were required to significantly impact

rhizome dry weight and then only when no herbicides were

applied or dalapon or sulfometuron were applied. The use of

mowing prior to and subsequent to imazapyr treatment signif-

icantly reduced control as compared to mowing prior to

application only. This may have occurred due to the fact

that the major cause of imazapyr degradation is photolysis

(American Cyanamide, 1983). By removing the cogongrass

foliage, more light reached the soil increasing the

degradation of imazapyr.

In the second experiment (Table 4.3), herbicides and

mowings significantly interacted to affect cogongrass

foliage regrowth. When used without chemical treatments,

mowing only inhibited foliage regrowth when done twice.

This may have been caused by a depletion of carbohydrates

that were needed to re-establish above-ground biomass.

Inhibition of foliage dry weight with sulfometuron increased

from -24% (growth stimulation) with no mowings to 77% with

two mowings. The stimulation of foliage regrowth found when

sulfometuron was used with no prior mowing may have been

caused not by direct chemical stimulation of cogongrass, but

by decreased competition from broadleaf species which sulfo-

meturon controlled. This hypothesis is supported by the fact

that sulfometuron with one or two mowings was not signifi-

cantly different from mowing treatments alone. The use of

mowing with dalapon gave greater control than dalapon alone

or two mowings alone. When cogongrass was mowed prior and

subsequent to dalapon treatments control was 88%, however

this level of control probably would not be adequate consid-

ering cogongrass' ability to survive and spread (Dickens,

1973). As opposed to the first experiment, the use of mowing

with glyphosate and imazapyr treatments did improve cogon-

grass control over the use of these herbicides alone, with

glyphosate efficacy being improved the most. Control of

foliage regrowth with glyphosate was increased from 27% with

Table 4.3. The effect of mowing and herbicides on cogon-
grass foliage and rhizome dry weight (Experiment 2 1986
through 1988).

Mowings1

0 1 2
Rate
-ah1 2 3
Herbicide kg ai'ha FDW RDW FDW RDW FDW RDW

--- % Inhibition

Imazapyr 0.8 64 53 61 52 94 82

Glyphosate 3.4 27 33 61 73 95 73

Dalapon 16.8 -8 3 55 50 88 76

Sulfometuron 1.1 -24 -11 16 49 77 70

Untreated 0 0 5 31 65 51

LSD(0.1)Mowings 16 18

LSD(0.1)Herbicides 21 23

0 mowings indicates herbicides applied to undisturbed
cogongrass, 1 mowing indicates mowing prior to herbicide
application, and 2 mowings indicates mowing prior and
subsequent to herbicide application.

2Cogongrass regrowth foliage dry weight harvested from 0.25
square m quadrants at 12 months after second mowing
operation.

3Cogongrass rhizome dry weight harvested from 4170 cubic cm
soil cores at 12 months after second mowing operation.

Inhibition values computed using unmowed and untreated
plots which contained 15.7 and 5.2 g of foliage and
rhizome, respectively.

no mowing to 61 and 95% when cogongrass was mowed once

or twice, respectively. Imazapyr efficacy was increased

over imazapyr alone only when cogongrass was mowed both

prior and subsequent to herbicide treatment, but was overall

.the most efficacious herbicide when used without mowing.

The inhibition of rhizome growth in the second

experiment followed a pattern similar to that found for

foliage regrowth except that the effect of mowing alone

increased with each mowing (0 to 31 to 51% inhibition with

0, 1, and 2 mowings, respectively). This indicated that

although mowing cannot provide long-term control, it can be

utitlized to reduce the potential spread of cogongrass.

Overall, from these experiments it was apparent that imazapyr

and glyphosate were the most active herbicides tested.

Disking-herbicide studies.

There was an interaction between experiments and treat-

ments in the disking-herbicide studies. Differences in cogon-

grass density and environmental conditions may have differ-

entially impacted the mechanical and herbicide treatments

between experiments. Tables 4.4 and 4.5 present the data

collected for the 1985-1987 and the 1986-1988 experiments,

respectively.

In the first experiment (Table 4.4), the use of a

single disking operation, either alone or prior to herbicide

treatment, did not control cogongrass regrowth more than the

herbicides applied to undisturbed cogongrass. Detachment

and decapitation (removal of the apex) of rhizomes caused by

Table 4.4. The effect of disking and herbicides on cogon-
grass foliage and rhizome dry weight (Experiment 1 1985
through 1987).

Diskings1

0 1 2
Rate
Rate 1 2 3
Herbicide kg ai'ha FDW RDW FDW RDW FDW RDW
--% Inhibition - -

Imazapyr 0.8 91 68 85 51 94 85

Glyphosate 3.4 69 74 76 60 85 85

Dalapon 16.8 42 40 56 34 94 85

Sulfometuron 1.1 19 15 4 33 64 51

Untreated 0 0 4 29 57 59

LSD(0.1)Diskings 14 12

LSD(01)Herbicides 17 16
(o.1)

0 diskings indicates herbicides applied to undisturbed
cogongrass, 1 disking indicates disking prior to herbicide
application, and 2 diskings indicates disking prior and
subsequent to herbicide application.

Cogongrass regrowth foliage dry weight harvested from 0.25
square m quadrants at 14 months following second disking
operation.

Cogongrass rhizome dry weight harvested from 4170 cubic cm
soil cores at 14 months following second disking operation.

4Inhibition values computed using undisked and untreated
plots which contained 66.9 and 17.6 g of foliage and
rhizome, respectively.

disking may have released dormant nodes from the inhibiting

influence of endogenous growth regulator or nutrient

gradients (Chancellor, 1974; Leaky, et al., 1975; McIntyre,

1965; and Nyahoza, et al., 1971). These activated nodes may

.have produced new shoots in a quantity equivalent to or

greater than the amount of shoot tissue destroyed by disking,

herbicide treatment, or a combination of the two. However,

a second disking reduced foliage regrowth 57% indicating

that depletion of reserves and subsequent regrowth potential

had been reduced. When dalapon or glyphosate were applied

between two diskings the control was 94 and 85%, respectively.

Control of cogongrass regrowth by imazapyr was unaffected by

disking treatments.

Disking alone significantly reduced cogongrass rhizome

biomass. Rhizome dry weight was inhibited by 29 and 59% by

one and two diskings, respectively. The effects of disking

varied with the herbicide evaluated. With sulfometuron,

disking caused the greatest enhancement of herbicide activity.

When used with dalapon, disking increased control only when

implemented prior and subsequent to herbicide treatment.

When imazapyr or glyphosate were used with disking, control

was significantly reduced by a single pretreatment disking.

In the case of imazapyr this apparent reduction in activity

may have been caused by increased photolysis which would

have occurred in a more open canopy caused previously by

disking. This reduction in control may have also been caused

by leaf tissue production at the expense of rhizome growth.

By changing the source-to-sink pattern from leaf-to-rhizome

to rhizome-to-leaf less imazapyr and glyphosate may have been

translocated to the rhizomes. This apparent effect of

source-sink pattern on glyphosate and other translocatable

herbicides has been proposed for several perennial grasses

(Atkinson, 1985; Kivlin and Doll, 1988). When disking was

implemented prior and subsequent to imazapyr treatment

rhizome biomass was reduced 85% which was greater than

imazapyr alone. However, the use of glyphosate with two

diskings did not inhibit rhizome growth more than glyphosate

alone. This may indicate that rhizomes being produced had

dormant nodes that were activated by the subsequent disking

treatment.

In the second experiment (Table 4.5), disking alone had

a greater effect on both foliage regrowth and rhizome prod-

uction than in the first experiment. Two diskings provided

89 and 72% control of foliage and rhizomes, respectively.

This may have occurred due to increased rhizome mortality

caused by the dry periods in 1987 following the sequential

disking. As opposed to the first experiment, the efficacy

of herbicide treatments were increased with each disking

performed. This may have been due to the the density of the

cogongrass stand in the area in which the second experiment

was conducted. This area contained one-fourth the foliage

biomass and one-fifth the rhizome biomass of the first

experiment area.

Table 4.5. The effect of disking and herbicides on cogon-
grass foliage and rhizome dry weight (Experiment 2 1986
through 1988).

Diskings1

0 1 2
Rate
-1 2 3
Herbicide kg ai'ha FDW RDW FDW RDW FDW RDW

---- % Inhibition4 - -

Imazapyr 0.8 63 40 71 77 95 95

Glyphosate 3.4 72 30 75 76 99 93

Dalapon 16.8 20 -20 74 79 98 85

Sulfometuron 1.1 -65 -41 61 58 96 81

Untreated 0 0 -9 25 89 72

LSD(0.1)Diskings 20 15

LSD(0.1Herbicides 26 19

0 diskings indicates herbicides applied to undisturbed
cogongrass, 1 disking indicates disking prior to herbicide
application, and 2 diskings indicates disking prior and
subsequent to herbicide application.

Cogongrass regrowth foliage dry weight harvested from 0.25
square m quadrants at 12 months following second disking
operation.

Cogongrass rhizome dry weight harvested from 4170 cubic cm
soil cores at 12 months following second disking operation.
4
Inhibition values computed using undisked and untreated
plots which contained 17.4 and 3.3 g of foliage and
rhizome, respectively.

Cogongrass control with the less efficacious herbicides

sulfometuron and dalapon were the most affected by the

addition of disking. A single disking prior to treatment

with sulfometuron increased foliage inhibition from 65%

.stimulation (-65% inhibition) to 61% inhibition. Control

attained with dalapon increased 270% with the addition of

prior disking and 390% with prior and subsequent disking as

compared to dalapon alone. The use of prior disking in

conjunction with imazapyr or glyphosate did not improve

control of foliage regrowth over that attained with either

herbicide alone. However, rhizome inhibition did increase

with each disking. This indicated that shoot production

caused by a single disking was sufficient to overcome

rhizome mortality. When imazapyr or glyphosate were used

with prior and subsequent disking, control of foliage

regrowth and rhizomes was near 100%. Although disking may

not be feasible in some situations (i.e., steep terrain or

heavily-wooded areas), in pasture renovation and silvicul-

tural site preparation tillage is an integral part, there-

fore the integration of herbicides (preferably imazapyr or

glyphosate) with disking could provide long-term cogongrass

control.

The influence of photoperiod on shoot initiation in

cogongrass rhizomes.

Three measurements of shoot initiation potential (total

shoot number per productive rhizome segment, percent shoot

initiation on productive segments, and percent shoot

initation from total node number tested) were utilized to

determine the effect of photoperiod. These data are

presented in Table 4.6. Continuous dark or continuous light

inhibited all three measures of shoot initiation in

comparison to the three photoperiods. Leaky, et al., (1978)

reported that nodes of rhizome sections of quackgrass

[Agropyron repens (L.)Beauv. #AGRRE] were inhibited by

light. The cogongrass rhizomes exposed to continuous light

were green at the time of harvest which indicated that

chlorophyll was present which may have at least in part,

negated the need for shoot intiation. This is supported by

the fact that rhizomes subjected to the various photoperiods

did not visibly produce chlorphyll, but did initiate shoots

more consistently (higher initation percentage) and in

greater quantity (higher shoot percentage). There was no

significant difference in shoot number or percent initiation

between long-day, short-day, and interrupted photoperiods.

The short-day photoperiod resulted in a higher shoot

percentage than did the long-day photoperiod, but not when

compared to the interrupted photoperiod. These results,

taken in toto, seem to discount a classic photoperiodic and

phytochrome response. The implications of this data to

shoot initiation following disking seems to point to a

minimum amount of daylength required to initiate shoot

morphogenesis, but no absolute dark requirement. Rhizomes

that become buried during the'disking operation would seem

to be less likely to produce shoots than those closer to the

Table 4.6. The influence of photoperiod on shoot initiation
from cogongrass rhizomes at one week after planting.

Light Dark Shoot #1 Initiation2 Shoots3

- hr - -

Continuous

Continuous

15

8

8 +1

1.7 B4

1.0 B

3.0 A

2.3 AB

2.2 AB

25 B

26 B

64 A

55 AB

67 A

13 C

9 C

45 A

25 BC

35 AB

Shoot number per productive rhizome
least one shoot).

(segments producing at

2Number of segments with at least one shoot/total number of
segments tested.

Number of shoots/total number of nodes.

Means within a column followed by the same letter are not
significantly different according to Duncan's Multiple
Range Test (o = 0.05).

One hour of light during the dark cycle.

-----$----

surface. This is even more likely to occur in cogongrass

which has specialized rhizome anatomy that reduces the

effects of dessication (Holm, 1977).

The influence of cogongrass stage of development and

defoliation on glyphosate efficacy.

Table 4.7 presents data collected from greenhouse

studies on the influence of time of defoliation on

cogongrass growth. Plants were monitored for 14 weeks,

however, after 9 weeks rhizomes were growing out of the pot

and leaf area was difficult to determine as lower leaves

senesced. Therefore, these data are not presented.

Defoliation and subsequent regrowth of three-week-old

cogongrass produced plants with a leaf area similar to

three-week-old plants. In response to this defoliation,

root-rhizome growth temporarily ceased as the plants

produced leaf tissue. This is a common phenomenon in

perennial grasses (Sturkie, 1930; Youngner and Nudge, 1976).

The leaf and root-rhizome weight ratios for three- and

six-week-old plants were similar, whereas the plants

defoliated at three weeks had a higher root-rhizome weight

ratio and a correspondingly lower leaf weight ratio. Plants

older than six weeks, whether defoliated or not, had leaf

areas that were judged to be too great to attain adequate

spray coverage in the subsequent glyphosate studies. There-

fore, three growth stages, three-week-old (Stage A), six-

week-old (Stage B), and plants defoliated at three weeks

after planting and allowed to regrow for three weeks (Stage

Table 4.7. The influence of time of defoliation on cogon-
grass growth.

Weeks after
Planting +
Weeks Regrowth

3+0

3+3

6+0

6+3

9+0

Leaf Area
2
cm

15.8 (5.4)3

13.2 (5.3)

40.7 (19.4)

189.3 (21.0)

391.0 (91.6)

Leaf Weight

Ratio1

0.69 (0.06)

0.21 (0.04)

0.58 (0.05)

0.70 (0.04)

0.55 (0.04)

Root-Rhizome

Weight Ratio2

0.31 (0.04)

0.79 (0.04)

0.40 (0.03)

0.30 (0.03)

0.45 (0.03)

1Leaf dry weight (g)/total plant dry weight (g).
2Root-rhizome dry weight (g)/total plant dry weight (g).

3Means followed by 1 standard deviation.

C), were selected for use in studies to determine the influ-

ence of relative source-to-sink pattern and relative leaf to

below-ground biomass ratio on glyphosate efficacy in

cogongrass.

The growth stages selected significantly influenced

glyphosate efficacy (Table 4.8). Stage A cogongrass was

extremely sensitive. A predictive model could not be

developed for inhibition of regrowth leaf dry weight (IRLDW)

because of a lack of differential response. The predictive

model for inhibition of root-rhizome growth (IRRG) indicated

that 0.02 kg ae'ha-1 would inhibit growth 50% (I50). Stages

B and C were more tolerant of glyphosate than Stage A.

There was little difference between the 150 for IRLDW between

Stage B and C (0.14 and 0.15, respectively). However, the

150 for IRRG was 0.03 for Stage B and 0.09 for Stage C.

This would seem to indicate that Stage B was in a leaf-to-

root/rhizome source-sink pattern while Stage C was in a

root/rhizome-to-leaf pattern. Glyphosate has been shown to

be translocated to rhizomes more in plants that are actively

producing rhizomes (therefore a metabolic sink) (Lolas and

Coble, 1980). The Stage B cogongrass was known to be

actively producing rhizome (Table 4.7) whereas the Stage C

plants were producing leaf tissue to the detriment of

root-rhizome growth. The fact that the 150 for IRLDW was

the same for both these stages does not detract from this

hypothesis because the majority of glyphosate recovered by

Lolas and Coble was on or in the treated leaf while 2 to 8%

Table 4.8. The influence of defoliation on glyphosate
efficacy in cogongrass.

Regrowth 2 2
Parameter Regression Equation R I50

- - - - Stage A3 -- - -- --

Leaf Dry
Weight (RLDW) NS -

Rhizome-Root
Dry Weight
(RRG)

%I = -130 45*(Logl0Rate)

- - - - Stage B3 - - -

Leaf Dry
Weight (RLDW)

Rhizome-Root
Dry Weight
(RRG)

%I = -212 189*(Log10Rate)

0.90

0.14

%I = -92 34*(Log10Rate) + 35*(Log10Rate)2

0.99

0.03

- - - - Stage C3 --------

Leaf Dry
Weight (RLDW)

Rhizome-Root
Dry Weight
(RRG)

%I = -228 217*(Log10Rate)

%I = -114 60*(Logl0Rate)

Glyphosate efficacy based on percent inhibition (%I) of
leaf regrowth (RLDW) and composite rhizome-root growth
(RRG) over the treatment period.

2Glyphosate rate in kg ae'ha-1 required to cause a 50%
reduction in the particular regrowth parameter.

Stage A cogongrass was treated 3 weeks after planting,
Stage B cogongrass was treated 6 weeks after planting, and
Stage C cogongrass was treated 3 weeks after being
defoliated at 3 weeks after planting.

0.86

0.02

0.92

0.88

0.15

0.09

was translocated to the rhizomes of johnsongrass. Similar
14
results have been reported using 14C-labelled glyphosate in

other species (Wyrill III and Burnside, 1976; Gougler and

Geiger, 1981).

Stages B and C cogongrass utilized in this study are

analagous to established cogongrass infestations in late

summer to fall and spring, respectively. Willard (1988)

found that cogongrass produced more rhizomes during the

later part of the growing season than during the spring.

Therefore, it could be assumed that glyphosate applications

would be more efficacious when applied during the fall than

during the spring. Increased efficacy of fall treatments

of glyphosate has been reported for johnsongrass (Derting,

et al., 1973) and bermudagrass [Cynodon dactylon (L.)Pers.

#CYNDA] (Andrews, et al., 1974; Whitwell and Santelmann,

1978).

The results of this research could also be applied to

the proper timing of glyphosate following mowing. It would

appear that glyphosate applied before sufficient leaf area

has regrown to continue production of rhizomes would result

in reduced efficacy. Baird, et al., (1973) indicated that

mowing within a week of glyphosate treatment reduced control

of quackgrass. Mowing outside this time range had no effect

indicating that sufficient leaf area had been produced.

From a practical standpoint these data indicated that

disking or mowing in the spring to remove dead foliage and

stimulate new shoot growth then waiting until early fall

78

before applying glyphosate would provide the greatest level

of cogongrass control.

CHAPTER 5
THE INFLUENCE OF STAGE OF DEVELOPMENT AND MOWING ON
BAHIAGRASS [Paspalum notatum var. saurae Parodi 'Pensacola']
AND COGONGRASS [Imperata cylindrica (L.)Beauv.] INTERFERENCE.

INTRODUCTION

Cogongrass [Imperata cylindrica (L.)Beauv. # IMPCY]

has been reported to negatively impact many annual and

perennial crop species (Eussen, 1979; Eussen, et al., 1976;

Eussen and Wirjahardja, 1973; Ivens, 1975; Jagoe; 1938).

Competition for light, water, and nutrients, physical plant

injury (caused by the rhizome apices penetrating crop

roots), and allelopathy have been reported to be mechanisms

of cogongrass interference (Boonitee and Ritdhit, 1984;

Eussen and Soerjani, 1975; Eussen and Soerjani, 1976;

Soerjani, 1970; Sunarwidi and Batugal, 1982; Zaenuddin, et

al., 1986). The combination of these factors and man's

inability to effectively control cogongrass has resulted in

infestations of 200 million hectares in southeast Asia,

several thousand hectares in the southern United States and

over 500 million hectares worldwide (Falvey, 1981; Dickens,

1974). The impact of cogongrass on crop production (and

Letters following this symbol are a WSSA-approved computer
code from Composite List of Weeds, Weed Sci. 32,Suppl. 2.
Available from WSSA, 309 West Clark Street, Champaign, IL
61820.

subsequently the quality of human life) on a worldwide basis

has earned cogongrass the distinction of being one of the

ten worst weeds (Holm, et al., 1977).

Bahiagrass [Paspalum notatum var. saurae Parodi

'Pensacola' # PASNS] is the predominant grass used on

rights-of-way and in pastures in Florida due to its ease of

establishment from seed, drought tolerance, and lack of

insect and disease problems, while responding to additional

inputs of fertilizer, etc. (Beard, 1973; Heath, et al.,

1985;

Turgeon, 1980). Establishment and mowing are the two major

cultural practices used for bahiagrass management on highway

rights-of-way (Lewis, 1986). In Florida, cogongrass has

become a serious pest on these rights-of-way. It has been

hypothesized that on highway rights-of-way cogongrass has

become a problem in these areas due to use of rhizome-

contaminated soil during bahiagrass establishment following

roadway construction and the relatively short-term effect

that mowing has on cogongrass growth and development

(Willard, 1988). However, the spread of cogongrass into

undisturbed, established bahiagrass appeared to be

inhibited. Therefore, research was conducted to determine

the influence of bahiagrass stage of development and mowing

on its interaction with cogongrass. This information will

help predict the most efficient route for cogongrass

invasion into bahiagrass as well as lead to an understanding

the influence that mowing has on the spread of cogongrass

into bahiagrass.

Materials and Methods

Preliminary studies.

Prior to conducting replacement studies to determine

the influence of bahiagrass stage of development and mowing

on its interaction with cogongrass, two greenhouse exper-

iments were necessary. The carrying capacity and the

response to fertility of each species was determined for the

system utilized for these studies.

A density study was conducted for each species to

determine the carrying capacity of the experimental system

utilized in the replacement studies. Cogongrass was

propagated from cleaned, de-scaled rhizome segments

collected from an infestation in Gainesville, Florida in

1988. Multi-node segments were planted horizontally in

trays covered with vermiculite and placed in a growth

chamber. The environmental conditions in the growth chamber

were; 16-hr light/8-hr dark photoperiod with an intensity of

200)E'm 2"s' and 30 C 2 C. Two weeks after planting

single nodes that had produced a two-leaf shoot were exised.

Shoots were planted into 2.8 L pots filled with 2.4 L of soil

mix containing 50% fumigated2 Arredondo fine sand (loamy,

siliceous, hyperthermic, Grossarenic Paleudults) and 50%

vermiculite on a volume basis. Densities of 2, 4, 8, 16,

and 32 shoots per pot were planted. Commercially obtained

2Bromo-O-Gas (methyl bromide + chloropicrin), Great Lakes
Chemical Corp., P.O. 2200, West Lafayette, IN 47906.

Pensacola bahiagrass seed3 were planted 0.5 cm deep at 50,

100, 200, 400 per pot. These seeding rates produced plant

densities of 29, 54, 104, and 222 plants. Pots were

maintained in a greenhouse with the following environmental

conditions; 16-hr light/8-hr dark photoperiod, day tempera-
+ +
ture of 30 5 C, night temperature of 25 +5 C and a mean

light intensity at noon of 900 pE'm2s -1. Pots were

watered daily and fertilized weekly with nutirent solution

(Hoagland and Aronon, 1950). Cogongrass and bahiagrass leaf

dry weights were determined eight weeks after planting.

Leaf dry weight per plant (LDWPP) was determined by dividing

the leaf dry weight by the corresponding plant density (PD).

The relationship between LDWPP and PD is described in the

following equations:

Bahiagrass LDWPP = 0.1 0.0003*(PD); r2 = 0.97

Cogongrass LDWPP = 12.2 0.9*(PD) + 0.02*(PD)2;

r2 = 0.94

The LDWPP declined as PD for bahiagrass and cogongrass

increased. From these studies a PD of 100 for bahiagrass

(to be acheived by planting 200 seed) and sixteen for cogon-

grass was utilized for the subsequent replacement studies.

These densities insured that the replacement studies were

conducted at a density that was independent of yield and

would therefore maximize interspecific competition.

B and G Seed Co., Williston, FL.

The effect of added fertility was determined for each

species by establishing a PD of 200 and 16 for bahiagrass

and cogongrass, respectively, and on a weekly basis applying

0, 40, 200, 400, and 800 ml of nutrient solution per pot

(Hoagland and Aronon, 1960). Bahiagrass and cogongrass

propagation methodology and greenhouse environmental

conditions were identical to those in the density studies.

Leaf dry weight (LDW) was determined eight weeks after

planting. Leaf dry weights were converted to relative

yields (RY) based on the the LDW yield of the highest

fertility input. Regression analysis (RY versus ml nutrient

solution) was performed on these data.

Bahiagrass RY = 0.007 + 0.004*(ml) 0.00003*(ml)2;

r2 = 0.96

Cogongrass RY = 0.18 + 0.002*(ml) 0.000001*(ml)2

r2 = 0.97

From these models replacement-series studies were fertilized

weekly with 500 ml nutrient solution. This fertility level

was utilized to alleviate interspecific competition for

nutrients.

Seedling bahiagrass and emerging cogongrass interference

studies.

The interaction of seedling bahiagrass and emerging

cogongrass was studied under greenhouse conditions using a

replacement-series experimental design (Wit, 1960). Bahia-

grass and cogongrass were planted in 2.8 L pots containing

the same soil mix used in the preliminary studies. Five

treatments consisting of varying ratios of bahiagrass PD to

cogongrass PD were utilized; 100:0, 75:4, 50:8, 25:12, and

0:16. All pots were watered daily and fertilized weekly

.with nutrient solution. The first experiment was conducted

in a greenhouse with the same environmental conditions as

those described previously. The second experiment was

conducted in a greenhouse with the same environmental condi-

tions except that the light intensity at noon averaged 1300

pE'm-2.s-1 (averaged 25% higher intensity at any given time
during the day)4. Eight weeks after planting, leaf dry

weight (LDW) and height of each species and shoot number and

rhizome dry weight of cogongrass were determined. Relative

yield (RY) of LDW for each species was obtained by dividing

the absolute yield at each ratio by the yield of each

species in monoculture. Relative yield totals (RYT) were

determined by adding the RY of each species within each

ratio mixture. Relative crowding coefficients (RCC) were

determined for each species as described by Harper (1977).

The influence of mowing on established bahiagrass and

emerging cogongrass interference.

To determine the influence of cogongrass on established

bahiagrass as affected by mowing frequency, bahiagrass

densities of 25, 50, 75, and 100 plants per pot (under

environmental and soil conditions identical to those used in

4Li-Cor Model LI-188B Integrating Quantum/Radiometer/Photo-
meter, Li-Cor, Inc., Lincoln, NE.

the seedling study) were established. Eight weeks after

planting bahiagrass foliage above 10 cm was removed using

hand-held, horizontally-reciprocating shears5. Into these

pots were planted single-node cogongrass rhizome segments

that had produced a two-leaf shoot. A replacement-series

design was utilized in which the following bahiagrass PD to

cogongrass PD were established; 100:0, 75:4, 50:8, 25:12,

and 0:16. These pots were watered daily and fertilized

weekly with nutrient solution. Four weeks after estab-

lishing these mixtures two-thirds of the pots were clipped

at 10 cm above the soil surface. During this procedure the

leaf weight (LDW) of each species in each mixture was

determined. Four weeks later half of the previously clipped

pots were clipped at 10 cm and species dry weights recorded.

This produced three clipping regimes (0, 1, and 2) for each

species ratio. Four weeks following the second clipping all

pots were harvested. At this time, LDW and height of each

species and shoot number and rhizome dry weight of

cogongrass was determined. Total leaf dry weight (TLDW) of

each species was determined by adding the LDW harvested at

each clipping interval to the final LDW. Relative yield

(RY) of each species was computed by dividing the TLDW of

each species in competition by the TLDW produced by each

species in monoculture. The relative yield total (RYT) was

determined by adding the RY of each species within each

Diston Model EGS-HD4, Danville, VA 24543

ratio mixture. The relative crowding coefficients (RCC)

were also determined.

Experimental design and analysis.

All experiments were conducted twice using a randomized

complete block design with four replications. Analysis of

variance was utilized to determine if the influence of

single factors (species, ratio, mowing, and experiment) and

interactions between these factors significantly affected

yield components (Helwig and Council, 1982). When experiment

significantly interacted with species-ratio treatments (P <

0.05) the individual experiments are presented separately.

When there was no experiment by species-ratio treatment

interaction (P > 0.05) data were pooled.

Results and Discussion

Seedling bahiagrass and emerging cogongrass interference

studies.

The two experiments conducted to determine the

interaction of seedling bahiagrass and emerging cogongrass

differentially impacted the competitiveness of the two

species. Table 5.1 presents the separate results of these

two experiments. The first experiment was conducted in a

greenhouse that transmitted, on average 25%, less light than

the greenhouse in which the second experiment was conducted.

The influence of reduced light intensity on cogongrass

growth and development has been reported by Soerjani (1970)

and Patterson (1980). Reduced light caused a greater

Table 5.1. The interaction of seedling bahiagrass (B) and
emerging cogongrass (C) as measured by relative yield (RY),
relative yeild total (RYT) and relative crowding coefficient
(RCC) of leaf dry weight and height and total pot dry weight
at eight weeks after planting.

Experiment 11

Experiment 2

--------------------- Leaf Dry Weight --------------------

0.58 (0.04)3

0.41 (0.02)

0.99

1.57

0.64

0.20 (0.03)

0.54 (0.05)

0.74

0.48

2.08

-------------------------- Height -------------------------
- cm- - -

HeightB (Monoculture)

35.6 (0.9)

Heights (50:50% Mixture) 36.6 (3.6)

HeightC (Monoculture)

59.0 (1.5)

HeightC (50:50% Mixture) 51.8 (2.4)

41.5 (2.3)

28.8 (1.9)

71.5 (4.3)

67.5 (2.2)

------------------- Total Pot Dry Weight -------------------
- -- -g-- ----

Monoculture B

50:50% Mixture

Monoculture C

33.3 (2.5)

31.6 (1.6)

33.9 (2.6)

29.4 (3.8)

28.1 (3.4)

45.6 (6.5)

Experiment 1 conducted in a greenhouse with 25% less light
intensity as compared to Experiment 2.
2
RY of total leaf dry weight of each species and RYT
computed at the 50:50% ratio mixture.

Where appropriate mean values are followed by standard
errors in parenthesis.

RCC values computed using mean dry weight per plant of each
species at the 50:50% ratio mixture and in monoculture.

2
RY
B
2
RY

RYT2

RCCB4
4
RCCC
C

reduction in rhizomes than leaves. However, as rhizomes are

a product of carbon fixation in the leaf, effects on leaf

biomass must be considered paramount.

Under the reduced light conditions in the first experi-

-ment seedling bahiagrass was more competitive than emerging

cogongrass, with RY's of 0.58 and 0.41, respectively. Using

RCC, bahiagrass was 2.4 times more competitive than cogon-

grass (RCC of 1.57 and 0.64, respectively). The RYT was 0.99

indicating that the two species in mixture were competing

interspecifically. Since water and nutrient status was

maintained at an optimal level, this indicated that light

was the limiting factor. Though cogongrass was taller on an

absolute basis, its erect leaf morphology may have allowed

sufficient light penetration to maintain bahiagrass. Bahia-

grass height was not affected by cogongrass competition,

(as compared to bahiagrass in monoculture), but cogongrass

height was reduced 12.2% when competing interspecifically

(59.0 versus 51.8). Total biomass produced was equivalent

for each species in monoculture and when the two species

were competing interspecifically.

Under the higher light intensity in the second experi-

ment, cogongrass was more competitive than bahiagrass, with

RCC of 2.08 and 0.48, respectively. This indicates that

cogongrass became 4.3 times more competitive under the high

light conditions. The RY values of 0.54 and 0.20 for cogon-

grass and bahiagrass, respectively, indicated that cogongrass

was contributing over half of the total yield in the system.

	Title Page
	Dedication
	Acknowledgement
	Table of Contents
	List of Tables
	List of Figures
	Abstract
	Introduction
	Cogongrass [imperata cylindrica...
	The influence of herbicide combinations...
	The influence of mechanical and...
	The influence of stage of development...
	Summary and conclusions
	Appendices
	Literature cited
	Biographical sketch

MILLISECOND	CLASS.METHOD	MESSAGE
0	sobekcm_page_globals.constructor
0	sobekcm_page_globals.constructor	Application State validated or built
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.constructor	Navigation Object created from URI query string
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.display_item	Retrieving item or group information
0	sobekcm_page_globals.get_entire_collection_hierarchy	Retrieving hierarchy information
0	sobekcm_assistant.get_entire_collection_hierarchy
0	cached_data_manager.retrieve_item_aggregation
0	cached_data_manager.retrieve_item_aggregation	Found item aggregation on local cache
0	item_aggregation_builder.get_item_aggregation	Found 'all' item aggregation in cache
0	system.web.ui.page.page_load (ufdc.page_load)
0	sobekcm_page_globals.constructor.on_page_load
0	html_echo_mainwriter.add_style_references	Adding style references to HTML
0	html_echo_mainwriter.add_text_to_page	Reading the text from the file and echoing back to the output stream
40	html_echo_mainwriter.add_text_to_page	Finished reading and writing the file