1. I roduction nt (Eichhornia crassip

1.  I roduction nt (Eichhornia  crassip

1. I

nt

roduction

Water hyacinth

(Eichhornia crassip

e

s)

is a South American plant that has spread widely and become one of the most damaging weeds in the world. South African biological control programs on the weed , which have been running for almost 20 years

(Cilliers 1991), have been fairly sllccessful. However, constraints such as climate and enriched water bodies restrict the impact of biological control agents (Hill

&

Cilliers

1999). Chemical control has also been successfully applied to confront the problem

(Findlay & Jones 1996), but gives only temporary relief against the rapidly reproducing weed . The idea of applying integrated control by using a combination of chemical and biological methods and eventually phasing out the use of chemicals has been suggested

(De Groot 1993).

In order to implement integrated control, information is needed on the susceptibility of natural enemies towards the herbicides used against water hyacinth . The arthropods that have been released as natural enemies on water hyacinth in South Africa are two congeneric weevil species

(Neochetina eichhorniae

and

N bruchi),

a mirid

(Eccritotarsus catarinensis),

a moth

(Niphograpta albiguttalis)

and a mite

(Orthogalumna terebrantis)

(Julien

&

Griffiths 1998). Some of these species have undergone tests in the USA and Australia to determine their susceptibility towards herbicides. It has been reported that

N

eichhorniae

(pellessier 1988) and

0.

terebrantis

(Roorda

et al.

1978) are susceptible toward diquat, while glyphosate is relatively non-toxic .

Herbicides that are registered for use on water hyacinth in South Africa are diquat, glyphosate , glyphosate-trimesium and terbutryn (Vermeulen

et al.

1998) . Another herbicide that is popularly used on water hyacinth is 2,4-D-arnine as it results in rapid death of the weed (Divakar & Manoharan 1979). No work has been done in South

Africa to determine the toxicity of any of these herbicides towards arthropod species, and little has been done in the rest of the world. Literature on the toxic effects of glyphosate suggests that it i s relatively non-toxic towards a variety of tested organisms , although only a few invertebrates have been tested (Sullivan 1988; Tooby

1985) . It was, however, found that a formulation of glyphosate was more toxic than the active ingredient alone, possibly as a result of added surfactants (Tooby 1985).

Another consideration for integrated control practices against water hyacinth is the management of arthropod populations subsequent to herbicide applications. Haag

(l986a) suggested that natural enemies fleeing treated mats should be provided with healthy untreated plants to serve as a reserve. This will result in higher insect survival when plants die and sink, and better re-establishment when the water hyacinth population resumes its growth. Others suggest that some arthropods might be attracted to treated plants for reasons varying from softened petioles, increased sugar content and the release of a kairomone, which could have dire consequences for the arthropod population (Delfosse & Perkins 1977; Wright & Bourne 1990) . The abilities of the different species to successfully escape sinking mats, the size of reserve mats and the distance of reserve mats necessary to implement such a system , have also been questioned (Bennett & Zwolfer 1968; Haag 1986a, b; Haag et at. 1988).

Even though we know little about the interaction between herbicides and biocontrol agents, integrated control of water hyacinth is currently pursued in South Africa. It is the aim of this study to make recommendations towards future integrated control practices of water hyacinth. The specific objectives are:

(i) To assess the relative acute toxicity of selected herbicide formulations for the mirid (Eccritotarsus catarinensis) and the weevil (Neochetina

eichhomiae), with the purpose of determining the least toxic, and therefore

the prefered herbicide to use during integrated control practices (Chapter

3).

(ii) To use feeding behaviour, i .

e. a sub-lethal indicator of relative toxicity, to compare the effect of selected herbicide formulations on the weevil

(Chapter 4) .

(iii) To establish whether normal feeding continued on treated plant material by determining the feeding behaviour of the weevil on water hyacinth leaves treated with selected herbicide formulations (Chapter 5).

(iv) To investigate the behaviour of weevil populations of water hyacinth mats sprayed with two herbicide formulations, and to determine whether migration away or toward treated plants occurs (Chapter 6).

2

2. Literature review

THE BIOLOGY OF WATER BYACINTH

2.1.1 Taxonomy

Water hyacinth

(Eichhornia crassipes

(Mart.) Solms-Laubach (family Pontederiaceae , order Liliales)) originates in the Amazon basin of South America, but has now spread to almost every tropical and subtropical region of the world (Holm

et al.

1977; Barrett

& Forno 1982). The Pontederiaceae is a monocot family that includes six genera, with one of the seven species in the genus,

Eichhornia natans,

being indigenous to Mrica and common in Ethiopia (Smith 1898 ; Divakar & Manoharan 1979; Aweke 1994).

About 30 to 35 species of the family are aquatic and indigenous to the Americas and none are considered threatened or endangered in their native habitat (Eckenwalder &

Barrett 1986; USDA 1999) .

2.1.2

Morphology

Eichhornia crassipes

is an erect , herbaceous , floating hydrophyte. The leaves are arranged in a rosette and held above the water while the roots and rhizome are submerged (Figure 2.1). Roots have visible root caps and may be purplish or black in colour if exposed regularly to light because of the presence of anthocyanines (Holm

et al.

1977; McKnight 1993) . Plants that are rooted in mud have white roots and may have a symbiotic relationship with micorhiza (Stent 1913 ; Penfound

&

Earl 1948 ;

Wright

&

Purcell 1995) .

The whorled leaves are divided into the base and the petiole, which is often bulbous , the narrow isthmus and the petiole blade, which is not a true lamina but the flattened part of the petiole (Penfound & Earl 1948; Holm

et al.

1977). At low population densities the petiole bases tend to be swollen with aerenchyma tissue , but plants occurring in thick mats have large petiole blades and slender, elongated petiole bases

(Center & Spencer 1981). Blades are heart-shaped and between 40 and 150 mm long and wide , and may act as sails

if

the plants travel downstream during dispersal (Holm

et

al. 1977)

.

3

o o

§

Figure 2.1

Morphology of

Eichhomia crassipes

,

(Drawn by R. Weber, NBI, Pretoria).

2.1.3 Flowering and reproduction

Flowering is closely correlated to temperature with night temperatures above 21 0 required for flowering (Edwards

&

Musil 1975) . The inflorescence is a spike with approximately eight flowers on a 300 mm long rachis . The flowers are white, lilac, blue, to purple and trimeric, with a dark purple and yellow blotch on the upper tepal, apparently serving as an insect nectar guide.

The presence of nectar and a nectar guide suggest that the plant is insect pollinated , but few insects have been observed visiting the flowers (Barrett 1977 ; 1980a; Wright

& Purcell 1995) . According to some studies sexual reproduction is limited within the new distribution range , possibly as a result of inefficient pollinators and environmental factors (Barrett 1980a, b) . Others suggest that the dry pollen is proof of a self-pollination reproducti v e strategy, being accomplished when the wilted spike curves downwards into the water (Holm

et at.

1977; Di v akar

&

Manoharan 1979).

-I

Tristyly, a type of genetic polymorphism, which is found in the flowers of water hyacinth, offers more proof for the sexual reproduction strategy theory (Barrett 1977 ;

Barrett

&

Forno 1982; Harley 1993). The mid-styled flower form is widespread , the long styled form occurs less often and the short-styled form is scarce, but all three occur in the Amazon area , which suggests that the plant originated from

this

region

(Barrett & Forno 1982). South African populations are of the mid-styled flower form , but long styled forms also occur in

this

country .

Each flower produces a capsule containing between 200 and 500 seeds that can remain viable for at least 15 years in the mud at the bottom of a pond (Stent 1913 ; Holm

et al.

1977 ; Divakar & Manoh a ran 1979) . Fluctuating water levels are characteristic of the natural habitat of the plant , causing exposure to sunlight which triggers germination

(Center & Spencer 1981; Barrett & Forno 1982) . Seeds may germinate on beach slimes and other periodically wet areas, and seedlings collected by the incoming tide or rising waters (McKnight 1993).

Although vegetative reproduction

IS considered the most important method of reproduction for water hyacinth , the potential of seeds to cause infestations has been underestimated. Large-scale herbicide treatment of the plant on the Nile River was followed by

high

seed germination among the dead plant material, which led to rapid re-infestation (Pettet 1964) . In Japan plants are often damaged by low winter temperatures and regenerate mainly through seeds (Ueki

&

Dki 1979) .

2.5.1

Physiology and tolerance

Water hyacinth mats have been reported to mcrease the loss of water through evapotranspiration by 1.4 to 3.7 times that of open water surfaces (Timmer & Weldon

1967; De Groot 1993 ; Singh & Gill 1996). Water hyacinth stomata are larger than those of other aquatic weeds investigated , which explains its

high

rate of evapotranspiration (penfound & Earl 1948 ; Lallana

et al.

1987) . However, many of these studies have been questioned due to large edge effects in the e x perjmp.nt~l plots

The plant grows best in a pH of 7 , but can survive at a pH ranging from 4 to 10

(Penfound & Earl 1948; Wright & Purcell 1995) . Aquatic micro environment changes studied by Ultsch (1973) , showed that water under water hyacinth mats have a lower pH , lower dissolved oxygen content , higher dissolved carbon dioxide content and lower

5

temperature than neighbouring open water surfaces. These, as well as nutrient changes are the results of the decomposition of high volumes of dead tissue (Stent 1913; Reddy

& De Burk 1991).

The plant contains 93 to 96 % water, and has a high nitrogen, phosphate and potassium content (Edwards & Musil 1975; Holm

et al.

1977; Imaoka & Teranishi 1988). It has the ability to remove and bio-accumulate nutrients and metals from polluted water, which makes it a successful inhabitant of eutrophied waters O'l"ewman & Haller 1988;

Zaranyika & Ndapwadza 1995; Mansor 1996) . Two major contributing factors toward the water hyacinth problem in South Mrica are the high nutrient content of water bodies, combined with hydrological factors such as impounded rivers which create stable water bodies (Edwards & Musil 1975).

Low temperatures limit the distribution range of the weed (Owens & Madsen 1995).

Frost may lead to the loss ofleaves, but the plant only dies when the rhizome tip freezes

(Sastroutomo

et al.

1978). Rooted plants fare better than free-floating plants when exposed to low temperatures (Owens & Madsen 1995). Forno and Bourne (1978) found that the death of taller leaves due to frost, increased light penetration and the production of shorter, more bulbous leaves .

2.5.2 Ecology

Water hyacinth behaves as a classical r-selected species throughout its native and introduced range. When conditions are suitable it germinates quickJy, grows rapidly and sets copious amounts of seed. As the seeds are very long-lived, the plant is able to survive extreme conditions of cold and drought and re-infest the water body when conditions are again suitable. Its ability to survive in eutrophied water bodies increases its competitive abilities. The paucity of indigenous free-floating aquatic macrophyte species in South Africa provides a nearly vacant niche for this fierce competitor. Water hyacinth can facilitate succession by providing a platform for other water plants to germinate and establish, but it is mostly other weed species that take advantage of this opportunity (Penfound

&

Earl 1948).

6

2.2 NEGATIVE

IMPACT

The negative impacts of water hyacinth in Mrica can be divided into its effect on the environment ; impact on socio-economy and on human health and recreation .

The weed competes directly with indigenous organisms for light and nutrients, but also changes the aquatic environment by making it unsuitable for native aquatic invertebrates, vertebrates and plants , thereby decreasing biodiversity (Wright &

Purcell 1995; Findlay & Jones 1996 ; Orach-Meza 1997). A reduction in predatory fish ; increased problematic algal blooms ; and reduced photosynthetic rates by submerged plant species have been reported as a result of water hyacinth infestations

(Penfound & Earl 1948; Nichols 1991).

In Mrica, India and Sri Lanka , water hyacinth interferes with agricultural practices by blocking irrigation and drainage systems, and increasing the loss of water through evapotranspiration (Divakar & Manoharan 1979 ; Cilliers 1991; Room & Fernando

1992; Aweke 1994; Abdelgadir 1996). Hydro-electrical power generation is obstructed and sedimentation in rivers and d a ms is increased, shortening the life of water bodies (Aweke 1994; Wright

&

Purcell 1995 ; Findlay

&

Jones 1996; Watts

1997). Movement of boats is impeded, hindering transport, communication and tourism activities (Abdelgadir 1996; Orach-Meza 1997).

Water hyacinth infested water develops bad odours , taste , colour and turbidity , resulting in reduced quality . More important , it promotes the development of water­ borne, water-based and water-related diseases e.g

. malaria, encephalitis , filariasis, amoebic dysentry and typhoid fever (Orach-Meza 1997) . Water hyacinth mats are the ideal hiding place for crocodiles and snakes that attack animal and human visitors to the water (Findlay

&

Jones 1996; Gough 1997; Orach-Meza 1997) . Mats impede human access to the water , restrict fishing activities from the shore, and prevent bathing , sWImmmg , water-skiing and other water sports (Aweke 1994; Abdelgadir

1996).

7

2.3 PEST STATUS AND SPREAD

Water hyacinth is considered the worst aquatic weed in South Africa, followed by

Salvinia molesta

D.S. Mitchell and Pistia stratiotes L., and two relatively new weed species: Myriophyllum aquaticum (Veil.) Verdc. and Azolla filliculoides Lam.

(Cilliers 1991). In the last 100 years, water hyacinth has spread to almost every tropical and sub-tropical region of the world. Its introduction to some countries is summarised in Table 2.1.

Table 2.1 The introduction of water hyacinth in various countries, with date of introduction and references (adapted from Scott et al. 1979)

Country

USA

Egypt

Australia

Java (Indonesia)

Siam (Thailand)

Bengal (India)

China

Ceylon (Sri Lanka)

Fiji Islands

Malaysia

South Africa

Burma

Philippines

Reunion and Madagascar

Borneo

Zimbabwe

Hawaii

Okinaqa (Ryukyu Islands)

Congo Brazzaville

Sudan

Senegal

Ethiopia

Date of entry or first report

1884

187911893

189011894

1894

1901

1902

1902

1905

1905

1907

190811910

1910

1912

1920

1926

1937

1946

1952

1952

1956

1964

1965

Reference

Edwards & Musil

1975

Scott

et al . 1979

Stent

1913;

Wright & Purcell

1995

Edwards & Musil

1975

Scott

et at. 1979

Divakar & Manoharan

1979

Divakar & Manoharan

1979

Edwards & Musil

1975

Scott

et at. 1979

Scott

et at. 1979

Stent

1913;

Harley

1993

Scott

et at. 1979

Scott

et al. 1979

Scott

et at. 1979

Scott

et at. 1979

Chikwenhere

1994

Scott

et at. 1979

Scott

et al. 1979

Scott

et at. 1979

Pettet

1964

Scott

et at. 1979

Aweke

1994

According to Harley (1993), the first infestations in Africa occurred in Egypt, South

Africa and the Congo River (Democratic Republic of Congo). It has recently caused serious problems in Nigeria, Cote d'Ivoire, on Lake Victoria and on Lake Chivero

(Zimbabwe) (Harley 1993). Benin, Ghana, Niger and most eastern and southern

African countries, except Botswana, Namibia and Lesotho experience problems with the weed (Harley 1993).

8

Water hyacinth was first recorded in the Western Cape Province of South Africa in

1913 , and soon spread and caused problems in Kwazulu-Natal rivers (Figure 2 .

2)

(Edwards

&

Musil 1975 ; Cilliers 1991). It has become especially troublesome in the

Kwazulu-Natal coastal areas and along the Vaal River (Edwards & Musil 1975) .

Since its introduction, it has caused problems in the Swartskops River (Eastern Cape

Province) , Hartebeespoort Dam, Crocodile and Vaal River (Edwards & Musil 1975) .

Its importance as a weed was acknowledged in 1977 , when the infestation on the

Hartebeespoort Dam increased to cover 60 % of the water surface and chemical control had to be implemented (Scott

et al.

1979) .

• Eichhornia crassipes

..

S APIA Dat abase , A RC-PP RI

I

Figure

2.2 Distribution of

Eichhornia cras

s

ipes

in South Africa

(Drawn by

L.

Henderson, Plant Protection Research Institute).

9

Human transport of the plant for aesthetic reasons was originally the main method of distribution . Further distribution within continents and countries occurred through natural water movement, as plants were swept downstream or seeds carried in mud on animal feet (Edwards & Musil 1975) . Human transport of the plant still plays a big role in its spread as plants are often used as cushions in boats or are snagged by boats and carried upstream over great distances (Holm

et al

.

1977). In Thailand, China and

Japan, the use of water hyacinth for pig fodder has contributed considerably to its dispersal through the region (Edwards

&

Musil 1975).

Eutrophied water plays an important role in creating the ideal environment for water hyacinth to grow in South Africa. The dumping of sewage and other effluent into rivers , has led to serious infestations in the past (Marshall 1993; Watts 1997). By interfering with the hydrology of rivers and providing stable or slow moving water bodies, we further contribute to its success (Edwards

&

Musil 1975). In its native range,

E. crassipes

prefers slow moving water bodies with high nutrient content. In

South America population densities seldom reach high numbers as natural enemies and competing plant species suppress the weed (Wright

&

Purcell 1995).

2.4

UTILISA

TION

Many uses for the weed have been suggested, but none have proved to be very profitable on a large scale or have led to successful control of the weed as yet. A short list of the most popular uses is provided in Table 2 .

2. For more detail on practical uses for water hyacinth, see Little (1968); Holm

et al.

(1977); Edwards and

Musil (1975); Gopal (1987) and Findlay and Jones (1996) .

To utilise water-weeds on an economically profitable basis, large industrial developments and a constant source of plants are necessary (De Groot 1993). The role of small-scale use in the spread of the weed also causes concern (De Groot 1993).

Removal and transport of the plant material contribute to spreading the seeds and plants that can survive dry periods and infest new areas.

10

Table 2.2

Some popular uses for water hyacinth

Aquarium plant

Animal food (unpalatable , must add supplements)

Detergent (use ash)

Cigar wrappers

Colour dye

Fertiliser (high nitrogen)

Fish traps

Growth medium for mushrooms

Homeopathic medicine (Indians)

Human consumption (itching may occur)

Insecticide

Paper making

Recovering polluted water

Source of chemicals , carotene, vitamin A

Weaving baskets, mats , ropes

2.5 CONTROL

2.5.1 Introduction

Traditionally the control of water hyacinth fell into one of three broad categories: mechanical; chemical and biological control. More recently the emphasis has moved to an integration of the three methods (Cilliers

et al.

1996).

2.5.2

Mechanical

control

Mechanical control or the manual removal of water hyacinth is invariably the initial control option exercised once a water hyacinth infestation has become problematic .

However, some operators consider this option impractical in infestations larger than about one hectare because of the rapid rate of increase of the weed, and its weight as a result of high water content.

Machines such as harvesters, cutters, triturators, saw-boats and crusher plants are expensive to build, while their running costs are high (Gutierrez

et al.

1996). The boats that carry these machines are usually too large to enter shallow waters, and patches of weed are left behind in these areas, contributing to regrowth. As a result, these types of harvesters are unsuccessful in African river systems, which are often meandering and shallow. Disposing of waste products after removal is problematic as dumped water hyacinth plants and seeds can cause re-infestation or spread . Drying the plants in the sun followed by burning and burying was suggested, but would be very labour intensive (Du Toit 1938). Mechanical control is labour intensive, slow and of a temporary nature, but can be applied successfully between herbicide

11

applications to contain regrowth as part of an integrated control programme (Cilliers

et al.

1996; Findlay & Jones 1996; Reinhardt 1997).

2.5.3

Chemical control

Herbicides can control infestations immediately and can be used on large infestations .

Disadvantages of herbicide use are the expense associated with continued application and the possibility of negative effects on the environment. However, putting the problem into perspective, water hyacinth control is an expensive long-term management commitment, irrespective of control method or combination thereof.

Another consideration is that the majority of herbicides currently used for water hyacinth control, whilst not ecologically benign, impact the environment considerably less than mats of water hyacinth.

Chemical control has been practised against water hyacinth since the early 1900's .

Sodium arsenate was first used in Florida in the USA but discontinued after cattle poisoning occurred in 1905 (Center 1975). A number of other herbicides have since been used . Herbicides registered for use on water hyacinth in South Africa are glyphosate, glyphosate-trimesium, diquat and terbutryn (Vermeulen

et al.

1998),

while 2,4-D is still used in many other countries. Water hyacinth is apparently very susceptible to diquat and 2,4-D amine formulations (Divakar & Manoharan 1979;

Wright

&

Purcell 1995) .

Glyphosate products have been safely and successfully used in this country, but are relatively expensive compared with 2,4-D amine, diquat and glyphosate-trimesium formulations, which are preferred in many African countries, the USA, Mexico,

Australia , Taiwan, India and Malaysia (Wang

et at.

1994; Wright

&

Purcell 1995;

Guitierrez

et al.

1996; Mansor 1996; Singh & Gill 1996) .

Herbici de s are applied from aircraft, boats and from the ground, often with the help of mounted spray equipment (Findlay & Jones 1996). Large infestations are best ' contained by aerial spraying from fixed wing aircraft, helicopters or micro-light aircraft . Follow-up application can be done from boats or with knapsack sprayers, depending on the size and location of the infestation. Especially during ' aerial

12

application, care has to be taken to avoid herbicide drift to non-target vegetation. An evaluation of application methods specifically for the control of water hyacinth

In

South Africa was done for the Department of Water Affairs (Anonymous 1990).

Successful chemical control depends on skilled operators who maintain a long-term follow-up programme, which continually controls re-infestations from scattered plants and germinating seed. Any herbicide programme against the weed requires a commitment to continued operation of unlimited duration.

It is the lack of a rigorous follow-up regime that has often led to the failure of chemical control programmes.

For further discussion of the active ingredients and formulations used in this study refer to Appendix 1.

2.5.4 Biological control

(a) Background

While mechanical and chemical control is viewed as short-term, immediate control options, biological control is perceived as a long-term or sustainable control option.

Biological control takes a long time and dedication to implement as insect populations have to be released and established, and reach levels that can have an impact on the weed population.

Natural enemies considered for release against water hyacinth include phytophagous insects, mites, pathogens, a snail, herbivorous fish and the manatee (Andres &

Bennett 1975; Divakar & Manoharan 1979; Cassani

et al.

1981). Cattle, hippopotami and even ducks have been observed feeding on the weed, but their contribution toward its control, is doubtful.

Arthropods are considered ideal biocontrol agents because of their specificity towards their host plant, which reduces the risk of the species becoming a pest on other indigenous plant species. Biocontrol agents undergo vigorous testing before release, including starvation and multiple choice tests (Bennett 1977), in addition to extensive data collection from the region of origin and the literature. Selected species must be host-specific feeders and preferably have a short life cycle, reproduce rapidly and be able to adapt to a new environment. l3

The United States Department of Agriculture in 1961 initiated the biological control programme on water hyacinth. Since that time there have been a number of surveys of water hyacinth in South America (Center 1994), with the purpose of identifying natural enemies which might be suitable for release on the weed in its introduced range as biological control agents. The first biological control agent was released on water hyacinth in 1971 and since then a further six natural enemies have been released around the world (Appendix 2) (Julien

&

Griffiths 1998).

It is not necessary for biological control agents to kill whole plants. Many natural enemies cause constant injury to the plant, thereby leading to reduced reproductive rates, decreased leaf size, discarding of leaves and changes in nutrient composition

(Center 1980; Center

&

Van 1989).

(b) Released natural enemies

The two weevil species,

Neochetina eichhorniae

Warner and

Neochetina bruchi

Hustache (Coleoptera: Curculionidae) are the most widely used agents and have been released in 29 and 26 countries respectively (Julien & Griffith 1998). The biology of both species is well documented (Center 1994), rearing and releasing techniques have been refined and a universal post-release evaluation method has been developed.

The two weevils have contributed to the control of the weed world-wide. The release of only these two species have led to successful control in Papua New Guinea (Julien

& Orapa 1999) and on Lake Kyoga in Uganda (Ogwang & Molo 1997). The recent introduction of the two weevil species to Lake Victoria is also starting to reap rewards as they have established at a number of sites, and in some areas of the lake a dramatic decline in the weed population has been reported (Murphy 2000). This success has prompted authorities to reject the use of herbicides on the lake in favour of biological control. This success contradicts the report by Pearce (1998) which suggested that biological control would be ineffective on Lake Victoria and that the large-scale use of2,4-D would be the only solution to the water hyacinth problem on the lake.

14

The moth,

Niphograpta albiguttalis

Warren

(=Sameodes albiguttalis)

(Lepidoptera:

Pyralidae) has also been introduced to a number of countries and, in combination with the two weevil species has contributed to the control of water hyacinth in Australia and the USA (Center 1994). A number of pathogens have been recorded from water hyacinth and is contributing towards control, but

Cercospora rodmanii

Conway

(Fungus: Hyphomycetes) (=

Cercospora piaropi

Tharp.) is the only one that has been intentionally introduced as a biological control agent (Julien

&

Griffiths 1998).

Although these species result in effective biological control of the weed in some areas, results in other areas have been variable. This prompted biological control practitioners to investigate additional natural enemy species. The moth

Xubida infusella,

the mite,

Orthogalumna terebrantis

and the mirid,

Eccritotarsus catarinensis

have also been released in a number of countries (Hill & Cilliers 1999) and further surveys for additional agents are being carried out.

(c) History of biological control in South Africa

The biological control programme against water hyacinth in South Africa was initiated in 1973 and the weevil,

Neochetina eichhorniae

was released in 1974 (Cilliers 1991 ;

Julien & Griffiths 1998). The programme was terminated in 1977, but restarted in 1985 when the weevil was re-released (Cilliers 1991; Hill

&

Cilliers 1999). However,

N

eichhorniae

was deemed unlikely to achieve the desired level of control throughout the weed's geographical range in South Africa and additional natural enemies were collected, tested and released on the weed.

The biological control of water hyacinth programme in South Africa currently relies on the two weevil species

(Neochetina eichhorniae

and

N bruchi)

1 , the pyralid moth

(Niphograpta albiguttalis) ,

the water hyacinth bug

(Eccritotarsus catarinensis) ,

the galumnid mite

(Orthogalumna terebrantis)

and the fungal pathogen

(C

e

rcospora rodmanii).

These species have been released and have established at a number of sites throughout South Africa .

1

Consult Appendix 2 for a list of arthropod species released worldwide on water hyacinth, and

Appendix 3 for a detailed biology of the arthropod species released in this country.

15

Biological control of water hyacinth has been very effective in some areas of the country, such as New Years Dam at Alicedale (Eastern Cape Province). The release of

Neochetina eichhorniae

in this area in 1991 resulted in an 80

% reduction of the water hyacinth mat by 1994 . In the insect "reserve " sites along the Vaal River, water hyacinth mats infected with

Neochetina eichhorniae

and

Niphograpta alb igu ttalis,

have been shown to collapse within four years if left unhindered (Hill

&

Cilliers

1999).

Four more insect species are being considered for release in this country, they are the moth

Xubida infusella

Walker

(=Acigona infusella

Walker), the grasshopper

Cornops aquaticum

Bruner, the planthopper

Megamelus scutelavis

and flies of the genus

Thrypticus

(I-Iill & Cilliers 1999).

(d) Success of biological control

In spite of the successes discussed above the outcomes of the biological control programme have been variable. Hill and Cilliers (1999) have suggested several factors that might have constrained the impact of the biological control agents .

Firstly, the inability of some of the species to proliferate and control the weed in the range of climatic conditions under which water hyacinth grows. Successful control of water hyacinth elsewhere in the world has been limited to areas with tropical or subtropical climates . In South Africa, many of the water hyacinth infestations occur in areas at high elevation, characterised by cold winters and frequent frosts. This means that the active growing period of the plant is limited, and therefore the population growth of the biological control agents is limited to approximately six months of the year.

Secondly, many of the aquatic ecosystems that are invaded by the weed in South

Africa are enriched with nitrates and phosphates that allow the weed to proliferate at such a high rate that insects are unable to suppress it.

An example of this is

Hammarsclale Dam (KwaZulu-Natal Province) where the water hyacinth mirid

(Eccritotarsus catarinensis)

and the weevil

(Neochetina eichhorniae),

have attained very high population levels . However, the natural enemies appear to have had little

16

impact on the plants , possibly as they are fertilised by enriched effluent from a nearby waste water treatment plant (M .

P.HiI1

2

Pers . comm.) .

Finally, regular removal of water hyacinth infestations with their natural enemy populations through flooding and chemical and mechanical control programmes has also had an effect on the biological control of water hyacinth. Re-infestation of cleared or treated sites by seeds and scattered plants occurs rapidly, and populations of the plant proliferate in the absence of natural enemies.

The variable success of the biological control programme on water hyacinth has prompted two courses of action. Firstly, to expand the present suite of natural enemies to include species that might be more cold tolerant and able to achieve a greater level of control over the restricted growing period of the plant. Secondly, to investigate an integrated control approach for the weed, which would incorporate aspects of mechanical, chemical, biological, hydrological and nutrient control on a site-specific basis. The management plan formulated by Cilliers

et al.

(1996) for the

Vaal River provides an example of how local water hyacinth problems should be addressed.

2.5.5 Integrated control

(a) Introduction

Integrated control is the use of a combination of methods to control the weed. The aim is to phase out short-term physical and chemical control and in the long-term rely completely on biological control to suppress the weed (De Groot 1993). The methods chosen depend on the level of control desired, the hydrological nature of the site , time of the year, intensity of the infestation, and resources such as labour and infrastructure available.

An example of integrated control is the management of Lake Chivero in Zimbabwe.

Between 1990 and 1991 a serious infestation, partly as a result of eutrophied water , could only be contained by using aerial herbicide application. Natural enemies

(Neochetina

weevils) had been released previously and were released following this

2

Hill , M .

P . PPRI, ARC, Private Bag X134, Pretoria, 0001

17

herbicide application, while a "floating fence" was used to prevent new plants from entering the lake (Chikwenhere 1994; Watts 1997). Chikwenhere (1994) highlights an important aspect of integrated management, when he emphasises the necessity of a site-specific plan.

(b) Combining herbicides and natural enemies

Charudattan (1986) documented the development of a mycoherbicide containing

Cercospora rodmanii.

He showed that 100 % control can be attained within seven months after applying the mycoherbicide with both weevil species, while used separately neither of them resulted in sufficient control. The best results were obtained when the mycoherbicide was applied and followed three weeks later, by an application of a moderate concentration of the herbicide 2, 4-D.

Caunter and Mohamed (1990) found in a laboratory trial that on its own

Neochetina eichhorniae

could not contain water hyacinth significantly when released. However, when combined with the pathogen,

Myrothecium roridum,

good results where obtained after four weeks. When

Myrothecium roridum

was applied with low rates of 2,4-D, it gave better control than when applied without herbicide (Liyanage &

Gunasekera 1989). Center

et al.

(1982) suggested that combining a plant growth retardant with the weevil may give successful control of water hyacinth.

It has subsequently been demonstrated that paclobutrazol, another growth retardant, also combined with

N eichhorniae,

can result in effective control (Van & Center 1994).

Delfosse and Perkins (1977) noted the presence of a kairomone released by damaged plants, and suggested it could be a phagostimulant or oviposition stimulant that attracts

Neochetina eichhorniae

and

Orthogalumna terebrantis

to young plants, following herbicide application. Treated plants may also exhibit an increase in sucrose, as found in plants after treatment with glyphosate. Delfosse and Perkins

(1977) suggested spraying herbicides at a low concentration and leaving plants to release the kairomone that would attract insects, to control the weed further. Wright and Bourne (1990) suggested that

Niphograpta albiguttalis

preferred feeding on leaves sprayed with 2,4-D amine as it softened the petiole.

18

When combining chemical and biological control , it is necessary to determine the direct and indirect negative effects that herbicides may have on the insects. It was suggested that reservoirs of unsprayed plants be left for insects to escape to, from sinking mats

(Bennett

&

Zwolfer 1968; Haag 1986a). This recommendation relies on the assumption that the insects have the ability to disperse and are not directly affected by the herbicides. The movement of

Neochetina

weevils from plants sprayed with 2,4-D was tested, and although there was some movement of adult weevils, a high mortality of eggs, larvae and pupae occurred (Haag 1986a, b; Haag

et al.

1988). The timing and pattern of spraying may be important. If spraying coincides with the migration period, when weevils normally develop wing muscles , better survival may be achieved

(Haag 1986b).

19

3. Acute

to

xicity of selected herbicides towards

Eccritotarsus catarinensis

and

Neochetina eichh ornia

e

3.1 INTRODUCTION

The aim of this study was to test whether the herbicides currently used in the control of water hyacinth in South Mrica are acutely toxic, and therefore incompatible with two of the natural enemies released as biological control agents.

The biological control of water hyacinth in South Africa relies on five arthropod species, only

Eccritotarslls catarinensis

(water hyacinth mirid) and

Neochetina eichhorniae

(water hyacinth weevil) were chosen for inclusion in this study . The weevil,

Neochetina eichhorniae

is presently the most widespread of the agents released in South Mrica on water hyacinth and the results on this species can be extrapolated to its congeneric

N

bruchi.

The mirid was selected because it is the most recently released natural enemy on water hyacinth worldwide, and because it is the only agent thus far released on water hyacinth where both the immature and adult stages feed externally on the plant.

Previous studies have shown that adults of another arthropod speCles used

in

the biological control program on water hyacinth, the mite

Orthogallimna terebrantis,

is very sensitive to exposure to diquat, but not to 2,4-D, glyphosate or paraquat (Roorda

et al.

1978). Other studies have shown that active ingredients combined with surfactants can be . toxic even when neither caused any mortality when applied separately to the weevil

Neochetina eichhorniae

(Grodowitz & Pellessier 1990; Wright & Skilling 1987).

The ability of natural enemles to survlve after exposure to the selected herbicide formulations and surfactants as well as combinations will determine recommendations for integrated control of the weed.

20

3.2

MATERIALS AND METHODS

3.2.1

Background

Lethal concentration tests (LCso tests) are often used to quantify the toxicity of chemicals to living organisms. Although this criterion possesses recognizable limitations , it usually provides the first and most statistically reliable measure of the toxic effect a chemical may impose on an organism (Forbes & Forbes 1993). The test is a widely used and trusted method to determine the concentration of a substance that causes 50 % of the organisms in a test population to die. Only a few concentrations need to be tested to determine lethal doses with the help of computer programs such as LST ATS PIPROBAN

(Van Ark 1983) . The response of LC so tests does not have to be mortality (Jagers op

Akkerhuis

et al.

1999), but quantifying any response other than death is less objective and therefore less accurate.

Toxicity tests can be classified according to the experiment duration relative to life span of the organism. Acute toxicity tests usually last between 24 to 48 hours for insects, rats, fish and birds (Landis & Yu 1995) . The acute toxicity experiments done during this study, continued for 120 hours and were monitored every 24 hours .

Insects from natural populations of the test organisms were collected randomly for this experiment , to give a test population representative of the genetic composition in the field. Unfortunately this precluded knowledge about the age of the insect, but the test period of approximately 5 days was short in comparison to the adult longevity of 120 days for

N.

eichhomiae

(DeLoach & Cordo 1976) and 50 days for E.

catarinensis

(Hill

et al.

1999).

The herbicide concentrations were determined according to recommendations by manufacturers as well as preliminary tests (Table3.l & 3 .

2). Manufacturers recommend that Touchdown be used with the surfactant Add-2 while Midstream is recommended with the surfactant Agral. Both were tested in combination with and without this surfactant to determine an increased toxicity as a result of the surfactant.

21

,\LQ9Lrz...?/5

\:> \ I??b'L.

UOI

The exposure scenario determines the uptake efficiency and ultimate effect on the insect.

The droplet application method was de s igned to portray a possible field scenario , while not overestimating the potential toxicity of the herbicides as a complete emergence in the solution might have . Weevils and bugs were exposed to a 20

III droplet , applied on the dorsal surface of the insect. In practice, droplets sized 300 to 400 III are specified by some herbicide labels (Sanachem 1998) . As large droplets tend to submerge and drown insects during the application process , it was decided to apply smaller droplets , as only the relative toxicity of herbicides was under investigation.

Egg , larval , nymphal and pupal mortality were not tested, but it is expected that most of the sedentary immature life stages inside the water hyacinth plant material , will be killed as a result of deteriorating plant quality and rapid plant death following herbicide application (Haag 19 86b) .

3.2.

2 Eccritotarsus catarinensis

In this acute toxicity test , the water hyacinth bug was directly exposed to different herbicides at various concentrations and combinations with surfactants (Table 3.1 & 3 .

2), and mortality was noted . The insects were collected approximately two days before treatment, from ponds forming part of the breeding program at the PPRI in Pretoria and kept in tubs with an abundance of fresh , severed leaves as food . Insects were sorted in groups of eight unsexed adults in small petri dishes (65 mm diameter), lined with moist filter paper a day before treatment to acclimatise. The sex of the bugs was not determined a s this is a difficult procedure causing further stress to the insects. Food in the form of water hyacinth leaf pieces, sized 20 x 30 mm, were provided every 24 hours and moisture was checked regularly to avoid desiccation . Temperatures varied between

22 and 28 ° C and the y were exposed to a light regime of 16 hours li g ht, 8 hours darkness .

22

Table 3.1

Information on herbicides selected for testing: formulation composition and recommended dosages

Brand name M a n ufacturer Content

2,4-D amine

Mamba 360 Se

Midstream

2

Mon 52276

Muster

2

Rode0

2

Roundup2

Roundup Ultra

.

Touchdown

2

Tumbleweed

2

2

Provided by Sanachem

Sanachem

Zeneca

Monsanto

Zeneca

Monsanto

Monsanto

Monsanto

Zeneca

Enviro Weed Control

2,4-D 480 amine

Glyphosate SL (IP A salt)

Diquat (Dibromide salt)

Glyphosate (IP A salt)

Glyphosate (Trimethyl sulfonium salt)

Glyphosate (IP A salt)

Glyphosate (IP A salt)

Glyphosate

Glyphosate (Trimethyl sulfonium salt)

Glyphosate (IP A salt)

Surlactants

Add-2

2

Agral

2

Zeneca

Zeneca

Description

Spreader adjuvant

Polysaccharide

(with Touchdown)

(hexitan ester)

Wetting and

Alkylated phenol-ethylene sticking adjuvant oxide concentrate

{wjtlll\1idstream)

600

940

Recommended

g

a.ellY-- _ Dosage

(ro

~o1!vol}

480 2.00 - 6.00

360

373.5

360

330

2.00 - 6.00

3 .

75 - 5 .

00

2.00 - 6.00

2.00 - 3.00

480

360

360

480

240

4.50 - 9.00

2.00 - 6.00

2.00 - 6.00

2.00

2.00 - 6.00

0 .

20 - 0.30

0.75

1 g a.ell

= gram acid equivalent per litre

2

Registered under Act 36 (1947)

23

Table 3.2

Concentrations of herbicide formulations and surfactants used in acute toxicity tests on two insect species

Herbicide or Active ingredient

Surfactant gil

2,4 D

Add-2

Agral

Mamba

Midstream

Mon 52276

Muster

Concentration used in toxicity tests

Eccritotarsus catarinensis

2,4-D amine 480

1 3

9

18

Adjuvant 600

0.2 0.3 0.6

940

0.5

0.75 1

1 2

Adjuvant

Glyphosate 360

3

Diquat

374

1

Glyphosate 360

3

Glyphosate330

2

4

2

4

3

12

24

3.75 4

12

12

24

24

5

1

3

2

3

2

Neochetina eichhorniae

3

4

9

12

24

4

4

4

3

5

4

(% vol/vol)

0.2 0.3

1 2.4

0.25 0.5 0.75

12 24

12 24

12 24

12

24

Rodeo trimesium

Glyphosate

480

1.5

3

2

Roundup Glyphosate 360

1

Roundup Ultra

Glyphosate

360

3

Touchdown Glyphosate480

2

4

4 trimesium

Tumbleweed Glyphosate 240

1 4

4

3

9

4

12 24

18

12 24

1.5

1

3

8 20 2

12

24 50

1

3

2

4

4

4

4

3

8

12

9

4

12 24

12

18

12 24

12 24

24

Combinations

2% TD· & 0.2% Add

2% TD & 0.3% Add

4% TD & 0.2% Add

4% TD & 0.4% Add

8% TD & 0.8% Add

Eccritotarsus catarinensis

Neochetina eichhorniae

X

X

X

X

X

X

X

0.9375% MS & 0.1875% Ag

1.875% MS & 0

.

375% Ag

3.75% MS & 0.75% Ag

X

X

• TD

=

Touchdown; Add

=

Add2; MS

=

Midstream; Ag

=

Agral

X

X

X

24

The bugs were cooled in a fridge at about 5°C for less than a minute before application of herbicide and/or surfactant , as they tend to be very active and fly or run off . Fresh concentrations of herbicides were made up less than an hour before application. A 20

III droplet of herbicide was applied dorsally on each insect. Within minutes after application the insects were checked to ensure no mortality had occurred as a result of drowning in herbicide. Insects trapped in herbicide were removed carefully with a fine, clean brush . Insect mortality was monitored every 24 hours up to 120 hours . Three replicates of eight insects were done for each of the treatments .

3.2.3 Neochetina eichhorniae

Weevils were captured from the field approximately two weeks before initiation of the experiment and kept in plastic tubs with an abundance of fresh, severed leaves as food. Six weevils (three male and three female) were placed in 500 ml plastic containers, lined with moist filter paper and aerated through small holes in the top, at least 48 hours before commencement of experiment to acclimatise. Insects were directly exposed to different herbicide concentrations (Table 3.2) by applying a 20 f...ll droplet dorsally onto the elytra of the insect. Food was provided daily in the form of a fresh water hyacinth leaf. Temperatures varied between 22 and 28°C and they were exposed to a light regime of 16 hours light, 8 hours darkness. Mortality was noted every 24 hours up to 120 hours. Three replicates of six insects each were used.

3.2.4 Statistical procedures

The Proban analysis program used to calculate LCso values in this study, determines a regression of mortality against concentration of the formulation in water (Figure 3. 1).

Concentration values are transformed to logarithms, a normalization procedure considered standard practice in toxicity testing (Van Ark 1983). Normalizing the data is a way to compensate for the asymmetrical population reaction curve, usually a result of the high tolerance exhibited by a few insects (Van Ark 1983). The log concentration value that causes 50% mortality can then be anti-logged to obtain a predicted LCso value, which represents the concentration that is expected to cause

50% mortality .

25

The L C

so

v al

ue

ind icated on F

i

g ure 3.1 gives the concentr

ation

expected to re sult in mortality

of 50%

of the p opulation at 48 hours after exposure

t

o

the

tox in

.

The lin ear correlation coefficient (r) gives the accuracy of the estimated regression line

(

-

), or

the fit of the line to the observed mortality

(D) .

-1 o

--

~

Add2

48 hrs

30

20

10

I

-0.5 0

L og concentration

Le so = r

=

0.38

0.88

%

0.5

Fi

gure 3.1

The regression line obtained for Eccritotarsus catarinensis after

48 hours exposure to the surfactant Add2

.

The estimated LC

so

value (percentage offormulation in water) and linear correlation coefficient (r) are indicated .

• =

predicted mortality

D

=

observed mortality

26

3.3 RESULTS AND DISCUSSION

3.3.1 Determining LCso values

Both Proban analysis (P) and linear regression (R) values are listed, with linear correlation coefficient values (r) where available, in Appendix 4. The Proban analysis values are considered more accurate because the program takes into account the natural mortality of control treatments. However, the program is also sensitive to the number of concentrations used and could not be used in all cases.

When comparing the more accurate Proban values with those of the regressions, they clearly show the same trends but are in most cases slightly higher. As the regression curves could be fit for all herbicides examined in this study, it will be used as a measure of comparison during further discussion, keeping in mind that it slightly overestimates the toxicity of all products tested.

3.3.2 Eccritotarsus catannensis

(a) Non-linear relationships

All herbicides increased the mortality of the bug in comparison to the control, which was treated with distilled water. For most of the herbicides, higher concentrations increased mortality, but for Rodeo, Roundup and Touchdown this was not the case (Figure 3.2h, i, k).

These three formulations showed a non-linear relationship between herbicide concentration and mortality, meaning mortality did not show a linear increase with increased concentration.

A very high LC so or great variation of the value over time is the result of such a non-linear relationship. Roundup at 1 and 2

%, for example, caused more mortality than did 12 %

(Figure 3.2i) and resulted in a high

LCso value of 61.04

% at 24 hours (Table 3.4). For

Touchdown the

LCso value dropped from 112 .

60 % at 24 hours to 0.003

% at 120 hours (Table

3.4), showing great variation over time. For these herbicides the toxicity cannot be fully explained by simply considering the

LCso values .

27

Observations made during experiments suggest that non-linear relationships may be a result of droplet surface tension that increases with increased herbicide concentration in the solution.

This c a used some droplets to roll off the insect, not coming into contact with the insect's body .

Higher concentrations were sticky , thus possibly inhibiting movement or interferin g with gaseous exchange by covering spiracles and therefore having a different " toxic" effect on the insect than lower concentrations. The reason for these non-linear relationships is not certain, but for the purpose of this investigation it will be sufficient to know what happens at recommended concentrations of these three herbicides .

(b) Relative toxicity

General

The percentage mortality of adult

Eccritotarslls catarinensis

treated with herbicides is illustrated in Figure 3.2, and means and standard deviations appear in Table 3 .

3 . Estimated

LC s o values for

E.

catarinensis

are illustrated in Figure 3.3

.

Also

see Appendix 4, 5, 6 and 7 for statistical analysis.

Bugs treated with herbicides and control insects showed increased mortality with time (Figure

3 .

2, Table 3.3). The mortality of control insects increased from 8.75 % at 24 hours to 21.25 % at 120 hours . This could simply be a result of natural mortality , or can be partly attributed to increased stress as a result of confinement.

LC s o values are illustrated in Figure 3 .

3 where herbicides are organized from most to least toxic on the x-axis at the specified time, with recommended dosages indicated for further companson . Both Agral and Midstream show a LC so value lower than the recommended dosage range (Figure 3.3a) . This means that at the concentrations recommended for the control of water hyacinth, more than 50

% mortality of

E.

catarinensis

can be expected . For

Roundup Ultra (Figure 3.3a), some of the recommended dosages cause 50

% mortality, while other lower recomm e nded dosages cause less . For Mamba , none of the recommended d o sages will cause more than 50 % mortality because the recommended dosages lie well below the LC s o value (Figure 3.3a-e).

28

Table 3.4 describes the overlap between the

Le so value and the recommended dosage for each herbicide. This description or rating is based on the results obtained during this study and is limited to the specific species. It is however, the first indication of how insects will react to the tested herbicides and therefore the preferred herbicides to use in integrated control. This relative scale indicates for example that 2,4-D amine has a toxic effect on

E.

catarinensis,

when used at recommended dosages, while Roundup is considered safe at recommended dosages.

Surfactants

The surfactant Agral, showed a

Le so val~Jower than 0.75 %, its recommended dosage, throughout the 120 hours. Its effect on the species is therefore described as toxic (Table 3.4)

The other surfactant Add-2 also exhibits a low

LCo value between 0.63 and 0.33 %, but as this never overlapped with the recommended dosage range of 0.2 to 0.3 %, it is labeled safe (Table

3.4).

Formulations

Midstream (diquat) was the most toxic herbicide throughout the 120 hours. Its LC so value of

0.38% at 24 hours fall far below the lowest recommended dosage of 3 .

75 % (Figure 3.3a).

Therefore Midstream is labelled toxic in Table 3.4. It would have been informative to test another diquat-based herbicide as this could have indicated whether the toxic nature of the herbicide could be ascribed to the active ingredient or to the additives of the formulation.

Of the glyphosate-based herbicides, Rodeo was by far the least toxic, which may be a result of the absence of additives in the formulation. Very high LC so values for Roundup suggest it is very safe, but as pointed out earlier, this is mostly a result of a non-linear relationship between concentration and mortality. Roundup caused between 29.20 % (24 hours) and 50.00 % (120 hours) mortality at both 1 and 2 % concentration (Table 3.3). Roundup is however considered to be relatively safe as it did not cause more than 50 % mortality at recommended dosages during the test period (Table 3.3). Mamba, Mon 52276 and Tumbleweed all exhibited a moderate toxic effect. Mamba was the least toxic of the three as it never reached a

Le so value that falls within the recommended dosage range (Figure 3.3a-e) . Tumbleweed was the most toxic of the three and is described as hazardous (Table 3.4).

29

Glyphosate-trimesiwn-based Muster is considered toxic, as the estimated

LC

so values fall below the recommended dosages throughout the experiment (Figure 3.3a-e). Another glyphosate­ trimesium-based herbicide, Touchdown, showed an initial high

LC

so value as a result of the non-linear relationship described earlier, appearing non-toxic. Initially Touchdown appeared to have a low toxicity, but at the recommended dosage of 2

%, it resulted in more than 50

% mortality of the population at 120 hours and is therefore labelled toxic (Figure 3.2k, Table 3.4).

2,4-D amine is considered toxic with

LC

so value of 3.15

% at 24 hours, reduced to only 0.93

% at 120 hours, with recommended dosages of2 to 6

%

(Table 3.4).

Combinations

Combinations of Midstream with the surfactant AgraI were more toxic than when applied separately (Figure 3.4a). At 120 hours, Touchdown at 2 % combined with 0.2 % Add-2 resulted in less mortality than 2 % Touchdown alone (Figure 3.4b). Only double the recommended dosage of 4 % Touchdown and 0.4 % Add-2 increased the toxicity of the combination appreciably.

30

Table 3.3

Mortality

(%

and standard deviation) of

Eccritotarsus catarinensis

after direct

Herbicide

2,4-D amine

exposure to the following herbicides

% n Mean percentage of adult mortality'

242 48 72 96 120

MEAl"! MEAl"! MEAN

MEAN

MEAl"! l.0 3

8.3 ± 15.0 16.7 ± 15.0 29 .

2 ± 28.8

37.5

±

32.5 41.7 ± 28.8

3.0 3

54.2 ± 40.0 75.0 ± 21.3 83.3 ± 18.8

83.3 ± 18.8 87 .

5 ± 12.5

9.0

3

91.7 ± 7.5 95.8 ± 7.5 95.8 ± 7.5

95.8

±

7.5 95.8 ± 7.5

18.0 3

95.8 ± 7.5 100.0 ± 0 .

0 100.0 ± 0.0

100.0

±

0.0 100.0 ± 0.0

Add-2

Agral (Ag)

Mamba

0.2 3

12.5 ± 0.0 33.3 ± 26.3 33 .

3 ± 26.3

33.3 ± 26.3 33 .

3 ± 26.3

0.3 3

41.7 ± 28.8 58 .

3 ± 18.8 58.3 ± 18 .

8 58.3 ± 18.8

66.7 ± 26.3

0.6 3

37.5 ± 25.0 45.8 ± 26 .

3 45.8 ± 26.3 50.0

±

21.3

50.0

±

21.3 l.0 3

58.3 ± 7.5 66.7 ± 18.8 66.7 ± 18.8

66 .

7 ± 18.8 66.7 ± 18.8

2.0 3

91.7 ± 7.5 91.7 ± 7.5 91.7 ± 7.5 9l.7

±

7.5

91.7

±

7.5

0.5 3

54.2 ± 15.0 58.3 ± 18.8 58.3 ± 18.8 66.7

±

31.3 66.7 ± 31.3

0 .

8 3 62.5 ± 21.3 66.7 ± 18.8 75 .

0 ± 21.3

79.2

±

26.3 79.2 ± 26.3 l.0 3 62.5 ± 12 .

5 66 .

7 ± 7.5 66.7 ± 7.5

70.8 ± 15.0 70.8 ± 12.5

3.0 3

8.3 ± 7.5 12.5 ± 12.5 16.7 ± 18.8 16.7

4.0 3

16.7 ± 18.8 20.8 ± 26.3 25 .

0 ± 25.0 25.0

±

±

18.8

25.0

16.7

25.0

±

±

18.8

25.0

12.0 3

70.8 ± 31.3 70.8 ± 31.3 83.3 ± 26.3 83.3

±

26.3 83.3 ± 26.3

24.0 3

95.8 ± 7.5 95.8 ± 7.5 95.8 ± 7.5 95.8 ± 7.5 95.8

±

7.5

Midstream (MS) l.0 3

66.7 ± 18.8 95.8 ± 7.5 100.0 ± 0 .

0 100 .

0 ± 0.0 100.0 ± 0.0

2.0

3

83.3 ± 18 .

8 91.7 ± 7 .

5 95.8 ± 7.5 100.0 ± 0.0 100.0 ± 0.0

3.8 3

83.3

±

18.8 95.8

±

7.5 100.0

±

0.0 100.0

±

0.0

100.0 ± 0.0

Mon 52276

4.0 3 100.0 ± 0.0 100.0 ± 0.0 100.0 ± 0 .

0 100 .

0 ± 0.0 100.0 ± 0.0

5.0

3 83.3 ± 15 .

0 100.0 ± 0.0 100.0 ± 0.0 100.0

±

0.0 100.0 ± 0.0

3.0 3

25.0 ± 12.5 33.3 ± 7.5 33.3 ± 7.5 37.5

±

12.5 37.5 ± 12.5

4.0 3 33.3 ± 7.5 41.7 ± 7.5 45.8 ± 7.5 50 .

0 ± 0.0 54.2 ± 7.5

12.0 3 62.5

±

32.5 79.2

±

18.8 87.5

±

21.3 87.5 ± 21.3 87.5

±

21.3

24.0 3 95.8 ± 7.5 100.0 ± 0 .

0 100 .

0 ± 0.0 100.0 ± 0.0 100.0

±

0.0

Muster

Rodeo

2.0 3

50.0 ± 12.5 58.3 ± 18.8 62.5 ± 21.3 62.5 ± 21.3

66.7 ± 15.0

3.0 3 83 .

3 ± 7 .

5 83.3 ± 7.5 83.3 ± 7.5 87.5 ± 12 .

5 87 .

5

±

12.5

12.0 3

95.8 ± 7.5 95 .

8 ± 7 .

5 95.8 ± 7.5 95.8 ± 7.5

100.0

±

0.0

24.0 3 100.0 ± 0.0 100.0 ± 0.0 100.0 ± 0 .

0

100.0 ± 0.0 100.0 ± 0.0

1.5 3

3.0 3

0.0 ± 0.0 12.5 ± 12.5 12.5 ± 12.5 12.5

4.2 ± 7.5 12.5 ± 12.5 12.5 ± 12.5 20.8

±

12.5 12 .

5 ± 12.5

±

7.5 25.0 ± 12.5

4.0 3 4.2 ± 7.5 8.3 ± 15.0 8.3 ± 15 .

0

8 .

3 ± 15.0 16.7 ± 18.8

9.0 3

4.2

±

7 .

5 4.2

±

7.5 16.7

±

18.8

16.7 ± 18.8 20 .

8 ± 26.3

18 .

0 3 25.0 ± 21.3 25.0 ± 21.3 37 .

5 ± 25.0 37.5 ± 25.0 37.5 ± 25.0

Roundup 1.0 3

29.2 ± 26.3 37.5 ± 32 .

5 41.7 ± 58.8

45.8

2.0 3 29.2

±

50.0 29.2

±

50.0 33.3

±

46.3

33.3

3.0 3 12.5 ± 21.3 29.2 ± 40 .

0 29.2 ± 40.0 29.2

±

43.8 50.0

±

45.0

±

46.3 50 .

0 ± 45.0

±

40.0 37.5 ± 32.5

4.0 3 16.7 ± 28.8 20.8 ± 36.3 20.8 ± 36.3 20 .

8 ± 36.3

29.2 ± 31.3

12.0 3 25.0 ± 21.3 25.0 ± 21.3 33.3 ± 18.8 33.3

±

18 .

8 33.3

±

18.8

24 .

0 3 50.0 ± 32.5 50 .

0 ± 32.5 54.2 ± 38.8 54.2 ± 38.8

54.2

±

38.8

Roundup-Ultra 3.0 3

54.2 ± 31.3 58.3 ± 31.3 62.5 ± 32.5

62.5 ± 32.5 66.7 ± 26.3

4.0 3 50.0 ± 25.0 62.5 ± 25.0 70.8 ± 18 .

8 75.0

±

12 .

5 75 .

0

±

12.5

12 .

0 3 95.8 ± 7.5 95 .

8 ± 7.5 95.8 ± 7.5 95.8 ± 7.5 100.0 ± 0.0

24.0 3 100.0 ± 0.0 100.0 ± 0.0 100.0 ± 0.0 100.0 ± 0.0 100.0 ± 0.0

, Eight insects per replicate

2

Hours after herbicide application

37

Table 3.3

Continued

Herbicide

Touchdown

(TD)

Tumbleweed

%

2 .

0

4.0

8 .

0

20.0

1.0

4.0

12.0

24 .

0

50.0 n

MEAl"

24

Mean Qercentage of adults d~ing

MEAN

48

MEAN

72

MEAN

96

MEAN

120

3 4.2

±

7.5

29.2 ± 15 .

0 50 .

0 ± 21.3 62.5 ± 37.5 62 .

5 ± 37.5

3 29.2 ± 50.0 50.0

± 45.0 58.3 ± 38.8 75.0 ± 25.0 75.0 ± 25.0

3 16 .

7 ± 15.0 20 .

8

± 18 .

8

41.7

± 28 .

8

66.7

± 38.8 70 .

8 ± 40.0

3

33.3

±

7.5

41.7 ± 7.5 62.5 ± 21.3 70.8 ± 26.3 70 .

8 ± 26 .

3

3 20.8 ± 26.3 29.2 ±

31.3

33.3 ±

26 .

3 33 .

3

±

26 .

3

50.0 ±

18.8

3

29.2 ± 7 .

5 37.5 ± 12.5 45.8 ± 18.8 45.8 ± 18 .

8 50 .

0 ±

25.0

3 66.7 ±

26.3 70.8 ± 18.8 75 .

0 ± 21.3 79.2 ± 26 .

3 79.2 ± 26 .

3

3

79 .

2

±

7.5 79 .

2

±

7 .

5 91.7

±

7 .

5 95.8

±

7.5 95.8

±

7.5

3 100.0

± 0 .

0 100 .

0

±

0 .

0 100.0

±

0.0 100 .

0

±

0 .

0 100.0

±

00

Combinations

2 .0

%1D&O.2%Add

. 2 .

0%1D&O.3%Add

4.0%1D&O.4%Add

3

20 .

8 ± 18.8 25.0 ± 21.3 45.8

±

28 .

8 45.8

±

28.8 45.8

±

28.8

3

33.3 ± 31.3 41.7 ± 38 .

8 50.0 ± 32.5 50.0

±

32.5 50.0

±

325

3

54.2 ± 40.0 66.7 ± 38.8

79 .

2 ±

18.8 83.3

± 15.0 87.5 ± 12 .

5

1.875%MS&O.375%Ag 3

95 .

8

± 7 .

5 100.0 ± 0.0 100.0 ±

0.0 100.0

±

0 .

0 100.0

±

0 .

0

3.75%MS&O.75%Ag

3 100.0

±

0.0 100 .

0

±

0.0 100.0

±

0.0 100.0

±

0.0 100.0

±

0 .

0

Control

H

2

O

LSD

3

(p

<

0.05)

LSD

(p

<

0.01)

3 8.3

± 14.4 8.3

±

14.4

12.5 ± 12.5 12 .

5 ± 12 .

5 20.8 ± 19.1

70.0

77.5

71.3

78.8

71.3

78 .

8

75.0

82.5

73 .

8

81.3

3

LSD

=

Least significant difference, with p-values a s specified (Statistical analyses in

Appendix 6)

38

Table 3.4 Herbicides rated according to the effect they have on Eccritotarslis

catarinensis

during acute toxicity tests

HERBICIDE

2,4-D amine

Add-2

Agral

Mamba

Midstream

Mon 52276

Muster

Rodeo

Roundup

1

Roundup Ultra

Touchdown

Tumbleweed

DOSAGES (%)

USED IN PRACTICE

2

2.0 - 6 .

0

(to 12.0)

0.2 - 0.3

0.8

2.0 - 6.0

3.8 - 5.0

2.0 - 6.0

2.0 -3 .

0

4 .

5 - 9.0

2 .

0 - 6.0

2.0 - 6.0

2 .

0

2 .

0 - 6.0

(to 12.0)

(to 12.0)

(to 12.0)

(to 12.0)

(to 12.0)

LC

so

V ALUES

3

24 hrs 120 hrs

3.15

0.93

0.33 0.63

0.33

8 .

04

0.38

6.75

0 .

95

723.43

61.04

2 .

96

112.6

6.15

0.12

6.91

3.92

0 .

26

146.99

0.95

0.00284

2 .

21

RATING

4

Toxic

Safe

Toxic

Safe

Toxic

Hazardous

Toxic

Safe

Safe

Toxic

Toxic

Hazardous

I

2

Fonnulation or s urfactant

Refer to product labels for specified dosages and application instru c tions

3

4

Values obtained by linear regression (Excel)

Safe

=

LC so never falls within recommended dosage range;

Hazardous LC so falls within or below the recommended dosage range during the 120 hours; . at some time

Toxic LC so falls within or below the recommended dosage range from 24 hours

42

3.3.3 Neochetina eichhorniae

The percentage mortality for adult

N. eichhorniae

is illustrated in Figure 3.5a-1 with means and standard deviations appearing in Table 3 .

5 . In several of the graphs the mortality of the weevils was zero and the concentration lines overlap.

Surfactants and formulations

The weevil was less susceptible to herbicide application than the bug. Only Agral

(Figure 3.5c), Midstream (Figure 3.5e), Muster (Figure 3.5g) and Roundup Ultra

(Figure 3.5j) produced appreciable mortality. Regressions could only be drawn, and

LC50 values calculated, for Agral, Muster and Roundup Ultra, because of the relatively high mortalities caused by them (Appendix 4, Table B) . Proban analysis

LCso values could only be determined for Muster and Roundup Ultra, again giving slightly lower LCso values but correlating well with the regression LCso values.

Considering the LC50 values in Table B (Appendix 4), Agral is again the most toxic substance and cause approximately 50 % mortality at the recommended dosage of

0 .

75

% concentration at 24 hours.

It is followed by Muster which should not cause appreciable mortality when applied at the recommended dosage of 2 to 3 % (Table

3 .

5) . Roundup Ultra caused slightly higher mortality than Midstream, but neither will result in significant mortality when applied at recommended concentrations. No mortality occurred after treatment with up to 24

%

Rodeo or Mon 52276 while other glyphosate-based herbicides resulted in very little mortality (Table 3.5) . 2,4-D amine showed very low mortality overall (Figure 3.5a, Table 3 .

5) .

Combinations

When combining Midstream with the surfactant Agral, significant mortality was caused to the insect population (Figure 3.5e). The lowest recommended dosage of

3 .

75

%

Midstream with 0.75

%

Agral resulted in 94.44 % mortality after 120 hours

(Figure 3 .

6a) . Even half the recommended dosage for Midstream and Agral combined, caused up to 77 .

78

% mortality (Figure 3.6a). Agral at 0 .

75% caused up to

72.22 % mortality and contributed most to the toxic effect on this species (Figure

3.6a) .

44

Touchdown and Add-2, an adjuvant, did not cause any significant mortality when combined or when used seperately (Figure 3.Sb, k; . Figure 3.6b).

It is interesting to note that the Touchdown is glyphosate-trimesium-based , as is Muster which resulted

in

much higher mortality.

45

Table

3.5 Mortality (% and standard deviation) of

Neochetina eichhorniae

after direct

Herbicide exposure to the following herbicides

% n

24 2

MEAl\! MEAN

Mean number of adults dyingl

48 72 96

MEAN MEAN

120

2 , 4-D amine

1.0 3

3 .

0 3

4.0 3

9.0

3

12.0 3

0 .

0

±

0.0 0.0

±

0.0 0 .

0

±

0 .

0 0.0

±

0.0 0.0

±

0.0

5 .

0 ± 10.0 5 .

0 ± 10 .

0 5.0 ± 10 .

0 11.7 ± 10 .

0 11.7 ± 10.0

0 .

0 ± 0 .

0 0 .

0 ± 0 .

0 0.0 ± 0 .

0 0.0 ± 0.0 0.0 ± 00

0.0

±

0 .

0 0.0

±

0.0 0.0

±

0.0 0.0

±

0 .

0 0.0

±

0.0

Add-2

18.0 3

0.2

3

0.3 3

0.0

±

0.0 0.0

±

0.0 0.0

±

0.0 0.0

±

0.0 0.0

±

0 .

0

0.0

0.0

±

±

0.0

0.0

0.0

0.0

±

±

0.0

00

0.0

0.0

±

±

0.0

0.0

0.0

0.0

±

±

0.0

0.0

0.0

0 .

0

±

±

0 .

0

0.0

1.0 3

2.4 3

0 .

0

±

0 .

0 0.0

±

0.0 0.0

±

0.0 0.0

±

0.0 0.0 ± 0.0

0.0

±

0.0 0 .

0 ± 0.0 0.0

±

0.0 0 .

0

±

0.0 5.0

±

10.0

0 .

0

±

0.0 0 .

0

±

0 .

0 0.0

±

0.0 0.0

±

0 .

0 0.0

±

0 .

0

Agral (Ag)

0 .

3

3

0.0

±

0.0 0 .

0

±

0 .

0 0 .

0 ± 0 .

0 0.0

±

0.0 0.0

±

0.0

0.5

3

16.7

±

28.3 21.7

±

38 .

3 28.3

±

48 .

3 28.3

±

48 .

3 28.3

±

48.3

Mamba

0 .

8

3

61.7 ± 20 .

0 66.7

±

16.7 71.7 ± 25.0 71.7

±

25.0 71.7

±

25.0

3.0 3

0.0

±

0.0 0.0

±

0.0 0 .

0

±

0.0 0.0

±

0.0 0.0

±

0.0

Midstream

4.0

3

12.0 3 0 .

0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0

24.0 3

11.7

±

20.0 16.7

±

16.7 16 .

7

±

16.7 21.7

±

20.0 21.7

±

20.0

(MS) 2

.

0

3

4.0

3

0.0 ± 0 .

0 0.0 ± 0.0 0 .

0 ± 0.0 0.0 ± 0.0 0 .

0 ± 0.0

0 .

0

±

0.0 0.0

±

0.0 5 .

0

±

10.0 21.7

±

38.3 21.7

±

38.3

0.0

±

0.0 5 .

0

±

10.0 11.7

±

10 .

0 11.7

±

10.0 33.3

±

43.3

5 .

0

3

12.0 3

24.0 3

0.0

±

0 .

0 0.0

±

0.0 5.0

±

10 .

0 16 .

7

±

16.7 28.3 ± 35.0

0.0

±

0 .

0 0.0

±

0.0 0.0

±

0 .

0 0.0

±

0.0 0.0

±

0.0

0.0

±

0.0 0.0

±

0.0 0.0

±

0.0 5 .

0

±

10 .

0 5 .

0

±

10.0

M an 52276

0 .

0

±

0.0 0.0

±

0 .

0 0 .

0

±

0.0 0.0

±

0.0 0.0

±

0.0

3.0 3

4.0 3

12 .

0 3

24 .

0 3

0 .

0

0.0

0.0

±

±

±

0.0

0.0

0.0

0

0

.

.

0

0

0.0

±

±

±

0

0

.

.

0

0

0.0

0.0

0.0

0.0

±

±

±

0

0

.

.

0

0

0.0

0.0

0 .

0

0.0

±

±

±

0 .

0

0.0

00

0 .

0

0.0

0.0

±

±

±

0.0

0.0

0.0

Muster

Rodeo

2.0 3

3.0 3

0 .

0

±

0.0 0 .

0

±

0 .

0 0.0

±

0.0 . 0.0

±

0 .

0 0.0

±

0 .

0

5.0

±

10.0 5 .

0

±

10.0 5 .

0

±

10.0 5.0

±

10.0 5.0

±

10.0

4.0

3

0.0

±

0.0 0.0

±

0.0 5.0

±

10.0 5 .

0

±

10.0 5.0

±

10.0

12.0 3 83.3

±

28.3 88.3

±

20.0 88.3

±

20.0 88 .

3

±

20.0 95.0

±

10.0

24.0 3 95.0

±

10.0 100.0

±

0.0 100 .

0 ± 0.0 100.0 ± 0.0 100 .

0

±

0.0

1.5 3

3.0 3

4.0 3

9 .

0 3

12.0 3

18.0 3

24.0 3

0 .

0 ± 0.0 0 .

0 ± 0.0 0.0 ± 0.0 0 .

0 ± 0.0 0.0 ± 0.0

0.0 ± 0.0 0 .

0 ± 0 .

0 0.0 ± 0 .

0 0.0 ± 0 .

0 0 .

0 ± 0 .

0

0 .

0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0

0 .

0

±

0.0 0.0

±

0.0 0 .

0

±

0.0 0.0

±

0.0 0.0

±

0 .

0

0.0

±

0.0 0 .

0

±

0.0 0.0

±

0.0 0.0

±

0.0 0.0

±

0.0

0.0

±

0 .

0 0.0

±

0.0 0.0

±

0.0 0.0

±

0.0 0.0

±

0.0

0.0

±

0 .

0 0.0

±

0.0 0.0

±

0.0 0.0

±

0.0 0.0

±

0.0

Roundup 1.0 3

2.0 3

3.0

3

4.0

3

0 .

0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0

0 .

0 ± 0 .

0 5 .

0

±

10 .

0 5.0

±

10 .

0 5 .

0

±

10.0 5 .

0

±

10.0

0.0

±

0.0 0.0

±

0.0 0 .

0

±

0.0 0.0

±

0 .

0 0.0

±

0.0

5 .

0

±

10 .

0 5 .

0

±

10 .

0 5.0

±

10 .

0 5.0

±

10.0 11.7

±

10.0

12.0 3

24.0 3

0.0

±

0.0 0.0

±

0.0 0.0

±

0.0 0.0

±

0.0 5.0

±

10 .

0

0.0 ± 0 .

0 0.0 ± 0.0 0 .

0 ± 0.0 0.0 ± 0.0 5.0 ± 10.0

Roundup-Ultra 3.0 3

5.0

±

10.0 5.0

±

10.0 5.0

±

10.0 5.0

±

10.0 5.0 ± 10.0

4.0

3

16.7

±

16.7 5.0

±

25 .

0 5.0

±

25.0 2l.7

±

25.0 2l.7

±

25.0

12.0 3

16.7

±

16.7 45.0

±

48.3 50 .

0

±

45 .

0 50.0

±

45.0 50,0

±

45 .

0

24.0 3 33 .

3

±

33.3 45,0

±

4l.7 45 .

0

±

4l.7 50.0

±

43.3 55.0

±

48.3

1

Six in se cts per replicate

2

Hours after herbicide application

52

Table 3.5

Continued

Herbicide

Touchdown

(TD)

Tumbleweed

%n

MEAl"J

24

Mean number of adults dying

48

MEAN MEAN

72

MEAN

96

MEAl"J

120

2.0 3 0.0

±

0 .

0 0.0 ± 0.0 0.0

±

0 .

0 0 .

0

±

0.0 5.0

±

10 .

0

4 .

0 3

0 .

0

±

0.0 0 .

0 ± 0.0 0.0

±

0 .

0 0 .

0

±

0.0 0.0

±

0.0

8.0 3 0 .

0 ± 0 .

0 0.0 ±

0.0 0.0

± 0 .

0 00 ± 0.0 0.0 ± 0.0

12.0 3 0 .

0 ± 0.0 0 .

0 ± 0.0 0.0 ± 0.0 0.0

±

0.0 0 .

0

±

0.0

24.0 3 0.0 ± 0 .

0 0.0 ± 0.0 0.0 ± 0 .

0 0.0 ± 0.0 0 .

0 ± 0.0

1.0 3 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0 .

0 ± 0.0

4.0 3 5.0

±

10 .

0 5.0

±

10.0 5 .

0

±

10.0 5.0

±

10.0 5.0

±

10.0

12.0 3 0.0 ± 0.0 0 .

0 ± 0.0 0 .

0 ± 0.0 0 .

0

±

0 .

0 0.0

±

0.0

24.0 3

0.0

±

0.0 5 .

0

±

10.0 5.0

±

10.0 5 .

0

±

10.0

5.0

±

10 .

0

Combinations 2.0%TD&0.2%Add 3 5 .

0 ±

10 .

0 5.0

±

10.0

5 .

0 ± 10.0 5 .

0 ± 10.0 5.0 ± 10 .

0

4.0%TD&0.2%Add

3 0.0 ± 0.0 0 .

0 ± 0.0 0.0 ± 0 .

0 0.0

±

0.0 0.0

±

0.0

4.0%TD&0.4%Add

3 0 .

0

±

0.0 0.0 ± 0.0 0.0

±

0.0 0.0

± 0.0 0.0 ± 0.0

8 .

0%TD&0.8%Add

3 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0 .

0

±

0 .

0 0.0

±

0.0

0 .

9%MS&0.2%Ag

3 0.0 ±

0.0 0.0

±

0.0 0.0

± 0.0 0 .

0 ± 0 .

0 0.0 ± 0.0

1.9%MS&O.4%Ag

3 33 .

3

±

28.3 61.7

±

25.0 66 .

7

±

28.3 66 .

7

±

28.3

28.3 ±

10.0

3.8%MS&0.8%Ag

3 50.0 ±

43.3 61.7

±

35.0 66.7

±

28 .

3 88 .

3

±

20.0 95.0

±

10 .

0

Control

3 0.0 ± 0.0 0.0 ±

0.0

0.0 ± 0 .

0 0.0 ± 0.0 0.0 ± 0.0

H

2

O

LSD

3

(p

<

0 .

05)

LSD

(p

<

0 .

01)

35.2

41.2

40.2

47.2

42 .

0

49.3

44.8

52.5

48 .

7

57.2

3

LSD

=

Least significant difference, with p-values as specified (Statistical analyses in Appendix 9)

53

3.3

CONCLUSION

The results for

Eccritotarsus catarinensis give a good indication of the relative toxicity of the formulations that were tested. All the herbicides that were rated "safe" for use with

E.

catarinensis are glyphosate-based. The least toxic substance, Rodeo, is also glyphosate-based and contains the least surractants.

Glyphosate-trimesium seemed to be acutely toxic to the mirid. However, one glyphosate-trimesium formulation had no toxic effect on the weevil, even at dosages that are much higher than recommended or when combined with a surfactant.

The diquat-based herbicide was the most toxic of all the formulations toward the bug, and when combined with a surractant it was also toxic to the weevil at recommended dosages . The toxic effects of diquat and low toxicity of glyphosate-based herbicides is echoed in the result of other studies on the mite Orthogalumna terebrantis (Roorda

et at. 1978) and Neochetina

eichhomiae

(Pellessier 1988).

Most substances resulted in high mortality when used at dosages higher than those recommended. Therefore, it is crucial that herbicides are used strictly according to manufacturers' recommendations.

Surfactants definitely contribute to the toxicity of the herbicide formulations, considering the low toxicity of Rodeo (low surfactant content) relative to other glyphosate-based herbicides consisting of varied quantities of various types of surfactants . The increased mortality resulting when diquat (Midstream) is combined with the surfactant (Agral), is further proof of the contribution of surfactants. More work needs to be done on surfactants to rate them according to their toxicity toward natural enemies of water hyacinth and make recommendations for their use. This experiment highlights the importance of choice of surfactants with regard to toxicity towards arthropods used i.ll. biological control.

55

4. Feeding

beh

aviour of

Neochetina eichhorniae

weevils tr

e at

ed with selected herbicides

4.1 INTRO,DUCTION

Neochetina ei

c

hhorniae

exhibited low susceptibility to most herbicides during acute toxicity tests (Chapter 3). The species was only sensitive to one glyphosate-trimesium­ based herbicide Muster, and the surfactant Agral when applied alone, as well as combined with a diquat based herbicide, Midstream. However, besides acute toxicity , there are several other methods to measure the toxic effects of substances on insects.

Sub-lethal effects such as changes in fecundity and behaviour, which includes feeding behaviour, are considered effective ways to monitor the health of treated subjects

(Yokoyama & Pritchard 1984; Jagers op Akkerhuis

et al.

1999) . Behaviour is a sensitive indicator of stress, fitness and disruptions at physiological level (Slobodkin 1968; Bayne

1980 In : Forbes & Forbes 1993) .

Literature on arthropod and insect toxicology is limited and usually concerned with environmental toxicology .

In these studies a single species is chosen to represent the effect of a toxin on a whole taxa.

In this way,

Daphnia

(Crustacean) and

Chironomus

(midge) species are popularly used to represent insects, arthropods or invertebrates

(Forbes & Forbes 1993; Landis & Yu 1995) . These arthropods are aquatic, short-lived , small and soft-bodied arthropods relative to the test species used in this study .

It would be impractical to do feeding tests on most insect species, including the bug

(Ec

c

ritotarsus

c

atarin

e

nsis)

,

as it is difficult to quantify feeding. However, the weevil leaves visible feeding scars that can be distinguished easily (Figure 4.1) . The aim of this study was to monitor the feeding of

Neoch

e

tina

e

ichhorniae

on water hyacinth leaves after weevils had been treated with several herbicide formulations , in an attempt to investigate the possible sub-lethal and chronic impact of the s e herbicides on the weevils.

56

4.3 RESULTS AND DISCUSSION

4.3.1 Herbicide treatments: 4 and 12 %

The mean number of feeding scars made by weevils after exposure to 4 % and 12 % concentrations of a range of herbicides are illustrated in Figure 4.2. For the 4 % treatments at 48 hours after exposure, no significant difference occurred between the control and any of the treatments (Figure 4.2a). After 96 hours the combination of

Midstream and Agral caused a significant decrease in feeding compared with the control

(p

<

0.05), while Midstream alone did not (Figure 4.2b).

Concentrations of 12 % are higher than are usually applied in practice. Muster seemed to decrease feeding at 12

% relative to the control, but this was not significant at 48 or 96 hours (Figure 4.2c, d). No significant increases or decreases in feeding could be noted , probably as a result of high natural variance in feeding and too few replicates. A more accurate method for determining normal feeding and growth is determining weight changes (Forbes

&

Forbes 1993), but as with other measurements of the insect body the increased handling contributes to further stress to the test organisms.

Herbicides that caused increased feeding were Mon 52276, Rodeo, Touchdown and

Tumbleweed at 4 %, although these increases were not significant (Figure 4 .

2a, b). Mon

52276 and Tumbleweed also resulted in increased feeding at 12

%

(Figure 4.2c, d).

Herbicides that caused decreased feeding were 2,4-D amine, Roundup and Roundup Ultra at 4 %, while 2,4-D amine, Mamba, Midstream, Muster, Roundup and Roundup Ultra led to decreased feeding at 12

%

(Figure 4.2a-d).

58

4.3.2 Effect of increased herbicide concentration

The herbicides in Figure 4.3 caused trends of decreased feeding with increases in herbicide concentration, while those not included showed no definite dose-response . Insects treated with Midstream showed no significant increase or decrease in feeding with increased concentration (Figure 4.3a). When Midstream and Agral were combined at the lowest recommended dosage (3.75 % Midstream and 0

.

75 % Agral), and even half the lowest recommended dosage , it decreased feeding significantly (p

<

0.05) (Figure 4 .

3b) . Exposure to high concentrations of Muster led to a significant decrease in feeding at 48 hours, but there was no significant difference at 96 hours (Figure 4.3c). Muster, which had previously led to mortality of the weevil (Chapter 3), only decreased feeding significantly at unusually high concentrations of the herbicide.

Of the glyphosate-based herbicides that exhibited a dose-response, Mamba caused decreased feeding at higher concentrations, but this was not significant (Figure 4.3d). Roundup led to decreased feeding at unusually high herbicide concentrations, feeding at 12 % and 24 % was significantly less (p

<

0.05), but only at 48 hours (Figure 4.3e). Roundup Ultra also caused significant decreases in feeding at unusually high concentrations of 12 % and 24% but only at 96 hours (p

<

0.05) (Figure 4.3f).

61

120r-----------------------;-------------------------------------,

i

a

Midstream

100 so

60

40

20

o

H20 2%MS 4%MS

5%MS 12%MS

24%MS

M

II

.s

~ rs

Vl

120r-------------------------------------------------------------,

b Midstream and Agral

100

SO

.~

60

"C

QJ oS! ci 40 c c ro

QJ

~

20

LSDo05(4Shr)=42.8

LSDo05(96hr)=3S .

3 o

H20 0.94%MS

&

0.19%Ag 1.SS%MS

&

0.3S%Ag 3.75%MS

&

0.75%Ag

120 r-----------r-------------------r---------------------------~

c Muster

100

LS D

OO5(

4Shr)=53.2

LSDo05(96hr)=74.0

SO

60

40

20 o

H20 2%Must

3%Must 4%Must

S%Must 12%Must 24%Must

Figure

4.3

Mean number of feeding scars produced by groups of six

Neochetina eichhorniae,

after direct exposure to selected herbicides and surfactants (n=3).

Mean number of feeding scars after 48 hours

[=:=J

Mean number of feeding scars after 96 hours

LSD

=

Least significa;:, difference; H2O

=

Control; MS

=

Midstream ; Ag

=

Agral ; Must

=

Muster

62

40

20 o

120r-------------------r-------------------------------------

--, d Mamba

100 so

60

H20

3%Mamba

4%Mamba

12%Mamba 24%Mamba

~

<"')

II

.s

~

III

U

<II

120

~----------------------------------------------------------------~

e Roundup

100

I

LSDo05(4Shr)=44 .

3

SO

.~

60

"C

CIJ

~ ci 40 t: t:

~

:2:

20 o

H20

1%RU

2%RU 3%RU 4%RU

12%RU 24%RU

120r-----------------------------------------------------~------~

f Roundup Ultra

100 I

LSDo05(96hr)=4S.2

80

60

40

20 o

H20

3%RUU 4%RUU 12%RUU 24%RUU

Figure 4.3

Continued

LSD

==

Least significant difference; H2O

=

Control; RU == Roundup ; RUU == Roundup Ultra

63

4.4 CONCLUSION

The changes in feeding could not be attributed to a specific active ingredient, as treatment with some glyphosate formulations resulted in increased feeding (including Rodeo and

Tumbleweed) and others decreased feeding (including Roundup and Roundup Ultra).

For the glyphosate-trimesium formulations , one caused increased feeding (Touchdown) while the other caused decreased feeding (Muster). The 2,4-D amine formulation resulted in decreased feeding. Diquat was the only formulation that resulted in significant changes in feeding, when combined with a surfactant. The herbicides could not be rated based on these results.

It could be that surfactants contribute to the toxic effect of the formulation more than the active ingredient, which explains why different formulations of the same active ingredient, even at the same amount of acid equivalent (Figure 3.1), lead to different effects on feeding .

Exposure to toxins may result in complex metabolic reactions that could not be explained by quantifying feeding by counting feeding scars. Increases in feeding could be the result of increased metabolism through which insects try to rid their bodies of toxins, as it has been shown by Landis and Yu (1995) that fasting results in decreased metabolic rate and toxins stay in the organism's body longer. Decreased feeding may be a strategy to avoid the intake of toxins e.g. the grass carp's

(Ctenopharyngodon idella

Val.) feeding decreased at sub-lethal concentrations of herbicides (Tooby

et at. 1980)

.

Application of a droplet of herbicide on the weevil's body does not represent the situation it will be exposed to in the field. A more likely scenario where the weevil come into contact with herbicide treated leaf material during nocturnal feeding, is investigated in the next chapter.

64

5. Fee

d i n g

be

h a

viour of

Neochetina eichhor

nia e

on wa

te

r

h y

acinth leaves treated with selected herbicides

5.1 INTR O D UCTION

The previous two chapters have shown that direct application of herbicides to the weevil

N

e

ochetina eichhorniae

did have an effect on its mortality and feeding behaviour. This effect was usually not significantly different from the control and showed the weevils to be fairly resistant or tolerant to most of the herbicides tested . In addition , the methodology in the previous two chapters was not a true reflection of what the weevils are exposed to in the field, but a study of the relative toxicity of selected formulations.

As the weevils are nocturnal and shelter at the base of the petioles during the day they are unlikely to come into direct contact with the herbicides during field application.

However , the weevils will feed on water hyacinth that has been sprayed with herbicides.

The effect of herbicide-treated plants on weevil feeding behaviour has not been studied well. Jianquin

et at.

(1999) claimed that feeding on plants treated with high concentrations of Roundup resulted in some mortality of the weevil. Water hyacinth plants exposed to 2 , 4-D amine have softer petioles and increased nitrogen content , which may favour herbivore attack (Wright

&

Bourne 1990). However , a study by Pellessier

(1988) showed that feeding by

N

e

i

c

hhornia

e

and

N bruchi

decreased on 2,4-D amine and diquat treated leaves relative to the control treatment.

This study investigates the feeding behaviour and mortality of

N

e

ochetina eichhorniae

on water hyacinth leaves tr e ated with three herbicide formulations, each containing a different active ingredient.

65

5.2 MATERIALS AND METHODS

Groups offive adult weevils were kept in plastic containers (500 ml), lined with moist filter paper and aerated through small holes in the top. A young water hyacinth leaf was removed from the plant and put into each container for a period of24 hours, after which the leaf was removed and feeding scars and weevil mortality documented (0 hours).

This initial period of 24 hours serves as a second control, to ensure that the groups of weevils showed similar feeding capacity .

At 0 hours, a new leaf painted with a freshly made-up solution of either glyphosate

(Roundup); 2,4-D amine or diquat (Midstream) made up to a 6 % concentration with distilled water, or a control leaf treated with distilled water, was provided (Table 3.1).

Weevils were allowed to feed on the treated leaves for 24 hours after which the feeding scars and mortality were recorded. Treated food was provided in the same manner, every 24 hours and feeding scars and mortality was recorded daily, at 24, 48, 72, 96,

120 and 144 hours (Figure 5 .

2) . The experiment was conducted at temperatures varying between 22 and 28°C, and a light regime of 16 hours light and 8 hours darkness. The experiment was replicated six times . The mortality data did not comply with the requirements for parametric analysis, therefore the Kruskal-Wallis ANOVA by ranks and Dunn's multiple comparison test (Zar 1984) was used (Appendix 13). The feeding data was analized using a one-way analysis of variance for parametric data (Appendix

14).

5.3 RESULTS AND DISCUSSION

5.3.1 Evaluation of methodology

The toxicity may be overestimated because weevils receive freshly treated food daily.

A more natural situation would be to feed on an intact plant that has received a single herbicide treatment at the start of the experiment. This would however make regular observation of feeding and mortality difficult.

If only one treated leaf is provided, wilting will occur quickly and the experiment could only last for approximately 48

66

hours.

This method was therefore chosen to determine whether insects detect herbicides and refuse to feed, canyon feeding or show increased feeding on treated leaves.

Herbicide-treated plants transport toxins to other parts of the plant body , i.e. glyphosate is transported to the roots and apical meristem to be stored or metabolized (Bariuan et

al.

1999). This transport is not possible in severed leaves and changes in the chemistry ofleaves may occur, that would not in an intact plant. Even though severed leaves were used in this study, the results show a similar decrease in feeding that was found by

Pellessier (1988) on whole plants.

5.3.2 Appearance of Leaves

Control leaves painted with distilled water, displayed no visible signs of deterioration or wilting. Glyphosate (Roundup) treated leaves appeared normal, sometimes disco louring slightly and becoming yellowish. Diquat (Midstream) treated leaves exhibited uneven, brown discolourations which was possibly a result of uneven coverage because of high surface tension of the formulation without added surfactant. Leaves treated with 2,4-D amine turned brown and slightly sticky on the surface and were infected with a powdery white fungus, as this fast-acting herbicide led to rapid tissue deterioration .

5.3.3

Mortality

From 120 hours insects that fed on 6 % diquat (Midstream) treated leaves showed a significant increase in mortality (56.7

% at 120 hours and 70.0

% at 144 hours) when compared with the control (Figure 5.1, Appendix 13). 2,4-D amine also resulted in some mortality but this was not significantly different from the (Figure 5.1, Appendix

13). No insects feeding on untreated leaves or leaves treated with glyphosate

(Roundup) died (Figure 5.1). Feeding on treated leaves lead to a reduction in feeding in some cases that could have led to starvation, but not in the 144 hours that the experiment lasted (Figure 5.2). Ingestion of water hyacinth leaf material treated with a

6 % concentration of diquat (Midstream) caused higher mortality in the weevils than direct exposure to the herbicide (Chapter 3). The highest mortality documented for

67

5.3.4 Feeding

There was an overaJl decrease in feeding over time, the control insects' feeding dropped from a mean of66.0 scars at 0 hours to 51.7 scars at 144 hours (Figure 5.2). The control insects' feeding decreased by 14.3 scars; glyphosate (Roundup) treatment decreased by

20 .

2 scars; 2,4-D amine by 34.7 scars and diquat (Midstream) by 46.6 scars over 144 hours (Figure 5.2).

Feeding documented at 0 hours after 24 hours offeeding on untreated leaves, showed no significant difference (Figure 5.3a, Appendix 14). After 24 hours of feeding on leaves treated with diquat and 2,4-D amine, the weevils showed a significant decrease in feeding activity in comparison to the control

(p

< 0 .

05) (Figure 5.3b, Appendix 14). This decreased feeding activity continued for the rest of the 144 hours, for both diquat and 2,4­

D amine (Figure 5.3c-g, Appendix 14).

Feeding on glyphosate (Roundup) treated leaves only reduced feeding significantly at 96 hours.

5.3.5 Literature

Wright and Bourne (1990) found that 2,4-D amine treated plants displayed softened petioles and increased nitrogen content that might induce increased feeding by herbivores .

They suggested that it might be the reason for higher activity on the plants of the moth

(Niphograpta albiguttalis)

after herbicide treatment, and suggested that weevils may show the same trend. It has also been shown that glyphosate can lead to increased feeding by mammal herbivores, when applied at low concentrations (SuJlivan 1985), possibly a result of increased sucrose.

Pellessier (1988) reported decreased weevil feeding on plants treated with diquat and 2,4­

D (Weedar64), relative to a control population. Diquat-treated plants showed significant decreases in ether extractable compounds and ash content (pellessier 1988). Both 2,4-D­ and diquat-treated plants showed significantly reduced crude protein levels and an increase

69

in nitrogen-free extracts, while calcium content in 2,4-D-treated plants increased greatly.

Pellessier (1988) also found correlations between the changes in feeding on diquat-treated plants, with changes in ether extractable compound and moisture content. In this experiment, weevils did not show increased feeding on treated, severed leaves in this study or in the work ofPellessier (1988) on whole plants.

70

CD

II

.s

.!!1

'~

QJ

::

~

I+:

80

70

60 ·

50

40 ..

30

20

10 .. a

0 Hours

a

o

I , ..

:

".

, Y," .

".

C.

:l e

OJ

....

W

C.

.... 80 co

lil

70

OJ

,~

"C al u.

60

50·

6%Glyphosate

c 48 Hours

F "" .," ,

6%2.4-0 amine

40 ..

30

20

10 o

I , .....

. . ' ,

6%Glyphosate

: .

,

6%2,4-0 amine

" .

'

"~'"

6%Oiquat

, : /'> '

6%Oiquat

."

Control

80

70

60

50

40

~ -

..

~.

,~

.

...

~ -

..

~ .

.-

.

. ...

. .

. _._.

...

.

. " _ ..

_ .

..

. -----... ..,.... .

_-

.'..

.

¥-"-'~ ~ " '.' . -.~ -

..

---.

---~

.

.,...~-~

b 24 Hours

a

30 ..

20

10

o

I

1" : " .

",.

,' . . ,

6%Glyphosate

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

6%2,4-0 amine

80

70

60

50

d 72 Hours

40

30 ..

20

10 o

~

, ,, ,

> .

"" ' .

,

Control

Treatment

, .

''' ''''.~ ''I

6%Glyphosate

, ; . .

.. : ,'.

6%2,4-0 amine

P : ' " . ,

6%Oiquat

, ' , . .. . .. ....

6%Oiquat

Control

" .

· ' d

Control

Figm'e 5.3 Mean number of feeding scars produced by a group offive

Neochelina eichhorniae

after feeding on treated leaves over a period of six days (Different alphabet letters indicate significant differences, also see Appendix 14), a, After 24 hours of feeding on untreated leaves; b, after 24 hours offeeding on treated leaves; c, after 48 hours of feeding on treated leaves; e, after 96 hours of feeding on treated leaves; g, after

144 hours offeeding on treated leaves, d, after 72 hours of feeding on treated leaves; f after 120 hours of feeding on treated leaves ;

72

8 0

1

·_·

..

··

.._·

·

-

·

96

_

· ·......

---.-

Hours

.

...

-

.

..

~-

·-· ·· .. ----· · ·.... ··· · ·· · ·-· ----· · a

60

C.

: l e

OJ

....

Q)

~ tIS o

CI) w-

50

II i

.s

40 j b

CI)

'>

(\)

(\)

~

(\)

>

.... o

30

20

10 o

.

~

1 ', · .,·,,· ·< '.!

70

6%Glyphosate

I I '

6%2,4·0 amine

9 144 Hours

b

~ I b

1'1 "~

6%Diquat

.

' 1

Control

80

T ..

· ·- .. ..-· .. • ..

..

·~

.... ....

.

.

...

-~

.• ... .. ..

- ....

...

. ...

..

- .• ..

..

. .

.

.....

..

....

- ..

.

....

... ..

.. _ ..

.

. . . .

_ ..

_ . ..

.

. ..

- . _ ..

..

.....

.. .

..

-

OJ

.5

"0 dol

Q)

I.!..

60 a

50

' ''j

I

I

I

I

40

30

20

10

0 .

1

I .. :

,"'~' ''"

'1.

6%Glyphosate

I ..

· .,

L...-..

.,------I ·.

>i''''''· · ,

6 % 2,4.. D amine

6%Diquat

Treatment

, .

'" ,-'· ..

· ·1

Control

: [ -;12 0Ho urs

60

1 i

50

J '

40 a

30

20 i

·1

': i-

6%G Iyphosate b

6%2,4-D amine

Figu r e

5 .

3 Continued

73 a b fYjW

.~ r-~

6%Diquat

"' "

.. , ...; · .· ,:/· .

: .. 1

Control

S.4

CONCLUSION

Ingestion of water hyacinth material treated with herbicides caused mortality similar to direct exposure experiments (Chapter 3). Feeding on glyphosate treated leaves resulted in no mortality; 2,4-D amine caused intermediate mortality and only diquat resulted in significant mortality of the weevils. All herbicide treatments led to decreased feeding, but only diquat significantly reduced feeding. Increased nitrogen (Wright & Bourne

1990) or sucrose in herbicide treated plant material did not lead to increased feeding .

This study has shown that weevils do not feed as well on treated plants as they did on untreated plants.

It is therefore their ability to migrate that will determine their survival in rapidly degrading plant mats, as they did not exhibit a preference for treated plants. A slow acting herbicide may give insects ample time to flee or may be so undetectable that insects continue feeding and consuming toxins that may lead to chronic effects such as reduced fecundity.

74

6. Behaviour of adult

Neochetina eichhorniae

popula

ti

o

ns

on water hyacinth mats treated with sel£cted

h

er

bi

cides

6.1 lliTRO

DU

CTION

One of the concepts suggested in the integrated control of water hyacinth is the establishment of reservoir areas of unsprayed plants for insects to move onto while sprayed plants die and sink (Haag 1986b). This suggestion relies on the assumption that insects have the ability to disperse and are not directly affected by the herbicide . It has been shown in this study that some herbicide formulations do increase mortality and decrease the feeding rate of

Neochetina eichhorniae

and

Eccritotarsus catarinensis,

and that surfactants can contribute to the toxic nature of formulations

(Chapters 3 to 5) .

This suggestion also relies on the assumption that the insects used in the biological control of water hyacinth are not attracted to the plants that have been sprayed . It has previously been shown that damaged water hyacinth plants release kairomones that stimulated feeding and oviposition by

Neochetina eichhorniae

and

Orthogalumna terebrantis

(Delfosse & Perkins 1977). The feeding of the moth

Niphograpta albiguttalis

seemed to increase on plants sprayed with 2,4-D amine (Wright & Bourne

1990).

It was however shown in the previous chapter that

Neochetina eichhorniae

exhibit decreased feeding on severed, treated water hyacinth leaves.

Haag (1986b) and Pellessier (1988) noted some movement by adult weevils from sprayed to unsprayed water hyacinth plants. Pellessier (1988) found that plants that had been sprayed with diquat degraded rapidly leading to rapid migration of weevils to unsprayed plants, but also resulting in high mortality of larvae and pupae. She suggested that low dosages or slower acting herbicides may allow more pupae . to emerge while females would lay fewer eggs as the plants lost condition, resulting in less of an impact on the insect popUlation.

75

Therefore, the aim of this study was to determine the pattern of movement and feeding behaviour of adult

Neochetina eichhorniae

on plants sprayed with formulations of two active ingredients, diquat and the slower acting glyphosate.

6.2 MATE RIA LS AND METHODS

Experiments were conducted in rectangular asbestos trays (2 100 mm long, 900 mm wide and 150 mm deep). The trays were divided in half across the tray using twine.

They were then filled with water and 25 g of the slow- release fertilizer (Osmocote) and 2.5 g of chelated iron was added (amounting to a concentration of 0.1 gil for the fertilizer and 0.01 gil for the iron chelate). Fifty young water hyacinth plants were added to each half of each tray and left for seven days to acclimatise.

Adult

Neochetina eichhorniae

were collected from the reanng facility at the

Rietondale Experimental Farm in Pretoria. The weevils were marked using the corrective fluid, Tippex, a drop of which was applied to the right elytrum. For each tray, 50 weevils were marked white and 50 were marked pink. The pink weevils were added, one per plant, to the half to be sprayed with herbicide and the white weevils to the half to be left unsprayed. The trays were left overnight for the weevils to acclimatise .

One day following the release of the weevils, one half of the tray was sprayed with a herbicide. The herbicide was applied in the morning, before

11

:00, ensuring that there was sufficient sunlight hours left to allow effective uptake of the herbicide. The 50 plants to be sprayed were lifted carefully from the tray to ensure that the leaves did not get wet, and placed in another water filled container. The group was then sprayed, left for approximately 10 minutes and carefully lifted and returned to the original tray .

This was to ensure that there was no spray drift onto the unsprayed plants, either through the air, or the shared water body. Herbicides were applied with a hand held atomiser and every effort was made to ensure an even spray of the herbicide. The

76

6.3 RESULTS AND DISCUSSION

6.3.1 Appearance of Leaves

After treatment with glyphosate the plants turned yellow within a few days and any flowers that emerged turned white . The plants treated with this low dosage did not rot within the 12 days of monitoring, except for some of the smaller plants and older leaves. Diquat treated plants turned brown within days and started to rot rapidly . By the sixth day these plants had all turned brown and it was decided to terminate the experiment at this stage .

6.3.2 Movement

of Weevils

According to the distribution of weevils at the time of collection the weevils moved from treated to untreated plants (Table 6 .

1, Figure 6.2). There was some movement of control insects, but this was probably a result of natural dispersal , 21.0 % of the surviving pink weevils had moved from the sprayed to the unsprayed section , while the remainder (79.0 %) of the pink weevils remained in the distilled water sprayed area .

At the same time , 34.4

% of the surviving white weevils had moved to the distilled­ water sprayed section with 65.6 % remaining in the unsprayed area.

The migration of surviving pink marked weevils in the diquat (Midstream) treatment was 80.2

% towards the unsprayed area, while 0.0 % of the surviving white weevils migrated towards the sprayed area (Table 6.1) . Movement of weevils after treatment with glyphosate (Roundup) was also towards the unsprayed area, 87.3 % of the surviving pink weevils had moved to the unsprayed area and 6.0

% of the surviving white weevils had moved onto plants in the sprayed area (Table 6.1) .

The weevil movement towards glyphosate treated plants (6.0 % of the sUfYlvmg insects) was less than the normal movement by the control insects (34.4 % of th e surviving insects) towards the distilled water sprayed section (Table 6.1). This does not agree with the suggestion that treated plants attract natural enemies because of an

78

increased nitrogen content, softened petioles or the release of a kairomone associated with injury.

Table 6.1 The mean percentage survival and mean number offeeding scars by

Neochetina eichhorniae

weevils, after release on trays with herbicide treated and untreated plants

Mean survival

of weevils

(%)

Mean number of feeding scars l

n

Pink White

TREAT ME NT

Glyphosate

Sprayed

Unsprayed

4

4

5.6 ±4A

2.6 ±1.0

38.6 ±11A 41.0 ±7A

9.9 ±4.5

40.2 ±5 .

0

Diquat

Sprayed

Unsprayed

Control

Sprayed

Unsprayed

4

4

2

2

10.6 ±10.2 0.0 ±O.O

43.0 ±10 .

0

41.0 ±7A

49 .

0 ±18A

22.0 ±O .

O

13 .

0 ±7 .

0

41.0 ±18A

8 .

6 ±2 .

2

36 .

6 ±3.5

35.0 ±lOA

35 .

5 ±8.8

1

On youngest open leaf

Pink == Insects released in sprayed area

White == Insects released in unsprayed area

±

== Standard deviation

Only 54.7

%

of the released population was re-captured for the whole experiment

(Table 6 .

2). As the trays were not covered, the reason for the reduction in the total number of weevils is uncertain and may be a result of mortality, or as is more likely the case, dispersal. There was no evidence of excessive mortality .

In addition, the numbers collected are likely to be an underestimate due to the cryptic behaviour of the weevils. The first trial showed lower survival overall

(44A

%), considering the overall survival of the second trial (65 .

0 %) (Table 6.2) . This could be as a result of leaving this trial for twice as long as the diquat trial , thereby allowing more time for dispersal of the insects. Treated ponds exhibited lower survival, or possibly higher dispersal, than the control pond in each trial (Table 6.2).

79

Table 6.2

Survival rate for

Neochetina eichhomiae

on water hyacinth plants sprayed with the herbicide glyphosate (Roundup) and diquat (Midstream) as a percentage of the released number of insects

Released Recovered Survival Rate

(Rec/RelxlOO)

Triall

Treated ponds (glyphosate) 400 175 43.8

Control

100 47

47.0

Total 44.4

Trial 2

Total

Treated ponds (diquat )

Control

Total

500

400

100

500

1000

222

247

78

325

547

61.8

78.0

65.0

54.7

6.3.3 Feeding

The feeding damage expressed as number of adult feeding scars on the first fully expanded leaf, was similar in both halves of the control trays (Table 6.1, Figure 6.3) .

However, in the herbicide treated trays a significantly higher number of feeding scars were recorded on the unsprayed half of the tray than on the sprayed half (one way analysis of variance, p< 0.05) (Figure 6.3, Appendix 15). This suggests that the water hyacinth plants treated with herbicides are avoided by the weevils, possibly as it becomes unpalatable. However, the number of feeding scars on the unsprayed halves of the herbicide treated trays was not significantly higher than either half of the control trays, suggesting that the weevils are dispersing away from the tray rather than to the adjacent, unsprayed water hyacinth within the same tray .

As in Chapter 5, feeding decreased significantly on diquat-treated plants, followed by movement away from the plants. Although feeding on glyphosate-treated plants did not result in mortality or seriously decreased feeding in the previous experiment, it seems that \Yeevils still prefer untreated plants and move towards them.

81

The results of the behavioural experiments using marked weevils showed that adult weevils crawl to adjacent healthy plants from sprayed, dying plants. In addition to which, they could migrate through flight to other mats of the weed, although this was not tested in this experiment. In a similar study, Haag (1986a, b) sprayed a series of water hyacinth plots with 2,4-D and monitored the movement of the two weevil species,

N eichhorniae

and

N bruchi.

The results of that experiment were similar to those reported in this study, and showed that weevils would move from sprayed to unsprayed plants up to a distance of four meters. The author suggested that it might be possible to spray mats of water hyacinth selectively and thereby herd adult weevils to unsprayed plants (Haag 1986b) and that portions of the mats should be left unsprayed and contained in distant areas where their economic impact would be small.

More research is necessary to determine the recovery rates of both the weed and the insect popUlations in order to determine the size and number of reserves of unsprayed plants needed.

83

7.

Gener

al

discussion and recommendations

7.1 A1M

The aim of this study was to determine whether herbicides applied to water hyacinth in South Africa are toxic to two of the insect species released as natural enemies.

Furthermore, feeding and movement behaviour of insects in the presence of herbicide treated water hyacinth material was investigated. Two insect species were used in the trials, a weevil

(Neochetina eichhorniae)

and the water hyacinth bug

(Eccntotarsus catarinensis).

The weevil is a well-established natural enemy of the weed in South

Africa and many other countries and the mirid is a recently released natural enemy against water hyacinth .

7.2 ACUTE

TOXICITY

TESTS

Acute toxicity tests encompassed exposure of insects to a droplet of herbicide . Ten herbicide formulations were tested of which all, except the 2,4-D amine formulation, contained one of the active ingredients registered for use on water hyacinth in South

Mrica. Although herbicide brands already contain surfactants, the efficiency of some formulations increase when more surfactant is added immediately before application .

Two such surfactants were tested on their own and in combination with a herbicide formulation, at recommended dosages. Most herbicides were tested at a concentration slightly less that it is recommended at, one or two recommended dosages and one or two higher than recommended dosages.

Increased concentrations of most herbicides resulted in increased insect mortality.

The weevil species was less susceptible to the herbicides overall, relative to the bug.

The active ingredients could be rated according to the toxic effects on the two insect species tested.

84

Direct exposure to a diquat formulation, only one diquat formulation was tested, caused the highest mortality of all the formulations, correlating with results from previous toxicity studies (Pellessier 1988; Roorda

et at.

1978). Glyphosate was the least toxic active ingredient, and the 2,4-D amine formulation was intermediately toxic. Two formulations of the active ingredient glyphosate-trimesium were tested and the toxicity varied from low to hazardous, probably the result of different surfact ants in the formulations .

The least toxic substance was one of the five glyphosate formulations tested, and the only glyphosate formulation without any surfactant Tt shetlld be noted however, that another glyphosate formulation, Roundup Ultra, was rated as toxic toward the bug. This highlights the contribution of surfactants to the toxicity of a formulation, as even the most harmless active ingredient can result in serious insect mortality when combined with certain surfactants. The toxicity of the diquat formulation also increased when applied with a surfactant. The toxicity of surfactants was not sufficiently investigated during this experiment, and needs attention during further research on the topic.

It is therefore clear that the rating of active ingredients is very much dependant on the final formulation or spray mixture. Surfactants in the formulation or added to the spray mixture must be taken into account when making a choice of herbicide in an integrated control program with insect bio-control agents.

7.3 F E EDING BEHAVIOUR AND MOVEMENT

During the first group of feeding trials, weevils were directly exposed to a range of herbicides at concentrations of 4 % or 12 %, of all ten herbicides and some of the herbicides with added surfactants . Weevil feeding scars were counted at 48 and 96 hours after exposure. Diquat combined with a surfactant at recommended dosages, resulted in a signifiCant decrease in feeding while the diquat formulation alone did not. Exposure to

85

some of the other herbicides also caused increased feeding and others decreased feeding, but these differences were not significant.

The next feeding study tested the feeding of weevils on herbicide treated leaves over a six­ day period. Freshly severed leaves treated with recommended dosages of either a diquat,

2,4-D amine or glyphosate formulation, were provided daily. Weevils feeding on diquat treated leaves showed significantly decreased feeding and significant mortality relative to the control treatment. Weevil feeding also decreased significantly on leaves treated with

2,4-D alT'jne but little mortality occurred . Leaves treated with glyphosate had no effect on feeding. The results support the suggestion that glyphosate is the least toxic active ingredient and that diquat is relatively toxic to the weevil.

The last experiment entailed the monitoring of weevil movement on a water hyacinth mat, partly treated with herbicide. Two formulations were tested at recommended dosages, a diquat formulation (Midstream) with a surfactant (Agral), and a slower reacting glyphosate formulation (Roundup) . Weevils were shown to migrate from the sprayed half of the mat to the unsprayed half, for both spray mixtures. Contradictory to previous reports, insects were not attracted by treated plants but fled the dying mats (Delfosse &

Perkins 1977; Wright

&

Bourne 1990).

7.4 CO NSID ERATIONS AND RECOMMENDATIONS

This study illustrates the important contribution of surfactants toward the toxicity of formulations or spray mixtures for insects. Keeping this in mind, the least toxic active ingredient, and therefore the one preferred for use in future integrated control of water hyacinth, is glyphosate. Formulations of both glyphosate-trimesium and 2,4-D amine can be used safely with insect bio-control agents when adhering to label specifications and concentrations, and careful consideration of formulation and surfactants. Diquat-based

86

herbicides should not be used in integrated control of water hyacinth , as insect mortality will be high.

Contrary to p r evious reports , this study has shown that weevils , unlike many mammal herbivores , are not attracted to herbicide treated leaves . Even glyphosate treated plants do not attract weevils, although it has been shown that grasses treated with glyphosate become more palatable to cattle and deer prefer feeding on glyphosate treated shrubs in the USA (Sullivan 1985) . When weevil populations are forced to continue feeding on leaves treated with diquat , high mortality follows . However , weevils have been shown to move away from sprayed to unsprayed plants, thus aiding integrated control practices.

The suggestion of using reserve mats or " refugia" for insects to flee to after herbicide applications is therefore viable, but needs further investigation . Information concerning the required size and distance of reserve mats , which may vary depending on the natural enemy used at the site, is required.

7.5

CONCLUSION

While it has been accepted that an integrated pest management approach is likely to achieve the best short- and long-term control of water hyacinth, the integrated control management plan for water hyacinth in South Mrica as currently practiced, is largely a herbicide treatment plan. Far more emphasis has to be placed on the requirements of the biological control agents . Hopefully , this study has gone some way to achieving a better understanding of what is required to ensure that herbicide control programs are compatible with biological control programs to achieve the long-term goal of reducing the impact of water hyacinth on South Mrica's aquatic ecosystems.

87

8. Su

m

mary

Water hyacinth is an aquatic plant originating from South America, and recognized as the world's most damaging aquatic weed. Large mats of water hyacinth invade water bodies, thus degrading the aquatic ecosystem and limiting all aspects of water utilization.

Although biological control of the weed has been successful it is considered too slow, while herbicide applications can render short-term relief. Therefore, the integrated use of the two methods is expected to produce both successful short-term and long-term control.

However, the success of this approach relies on the assumption that the two methods are compatible. This study was designed to determine whether herbicides registered for use on water hyacinth in South Africa are toxic to two insect species released for its biological control. The two agents chosen for this study were the weevil,

Neochetina eichhorniae,

and the mired,

Eccritotarsus catarinensis.

The study also investigated the feeding and migration of the weevil species, following direct and indirect exposure through herbicide treated plant material.

The first experiment tested the mortalities of both species as a result of direct exposure to the herbicides at a range of concentrations. The rnirid was very susceptible to herbicide exposure and high mortalities were recorded, especially from the herbicide product with diquat as active ingredient. The surfactant content of herbicides played an important role in determining the toxicity of a formulation. Even though weevils were in general less susceptible to herbicide exposure, treatment with diquat still resulted in significant mortality in the weevil population.

The feeding behaviour of the weevil was quantified after it had been treated with herbicide concentrations of 4 or 12 %. The herbicide containing diquat combined with the surfactant Agral caused a significant reduction in weevil feeding. Of the other herbicides tested, some caused an increase and some a decrease in weevil feeding, but these differences were not significant.

88

Weevils feeding on diquat- and 2,4-D amine-treated leaves exhibited a significant decrease irl feeding over a six-day period. Feeding on diquat-treated leaves also resulted ill significantly higher mortality than on 2,4-D amine- and glyphosate-treated leaves.

Movement of weevils was monitored after half a mat of water hyacinth was sprayed with either glyphosate or diquat. In both cases , the weevils mi g rated away from the sprayed plants to the unsprayed sections of the mat.

The implications for integrated control of water hyacinth in South Africa are that where possible glyphosate-based herbicides with low surfactant content and low active ingredient concentrations should be used .

It is vital that reserves or " refugia" be maintained to harbour insect populations, although the required size and distance of reserve areas warrant further investigation . Consultation between herbicide and biological control practitioners, on a site-specific basis is necessary for the successful integrated control of water hyacinth in South Africa and elsewhere where the weed constitutes a problem.

89

8. Opsom min g

Waterruasint is

'n akwatiese plant wat afkornstig is van Suid Amerika en wereldwyd beskou word as een van die nadeligste onkruide . Die plant verdring inheernse waterplante om groot matte te vorm wat die akwatiese ekosisteem drasties verander en degradeer.

Biologiese beheer van die onkruid was reeds in sommige gevalle baie suksesvol. Die implementering van biologiese beheer neem egter lank en daarom word onkruiddoders steeds gebruik om korttermyn beheer te bewerksteIlig. Die beheer van waterhiasinte is tans reeds 'n vorm van gemtegreerde beheer, al is dit onduidelik of die twee metodes mekaar aanvul. Hierdie studie het beoog om vas te stel of die onkruiddoders wat geregistreer is vir gebruik op waterhiasinte, toksies is vir twee van die insekspesies wat as natuurlike vyande in Suid-Mrika vrygelaat is . Die invloed van indirekte blootstelling aan sommige van die onkruiddoders en bymiddels word ook ondersoek, deur insekvoeding en

-migrasie te monitor in die aanwesigheid van behandelde plantmateriaal. Die twee biologiese beheeragente wat vir die studie gekies is, is die snuitkewer,

Neochetina eichhomiae,

en 'n besie wat van die familie Miridae,

Eccntotarsus catarinensis

.

Die eerste eksperirnent het die mortaliteit van beide speSles bepaal nadat hulledirek b!ootgestel is aan onkruiddoders teen 'n reeks konsentrasies . Die besie was in die algemeen meer sensitief vir onkruiddoders as die kewer, en vera! sensitief vir die onkruiddoder met dikwat as aktiewe bestandeel. Die snuitkewer het lae mortaliteit getoon in die algemeen, terwyl b!ootstelling aan die dikwat onkruiddoder duidelik die hoogste kewer mortaliteit veroorsaak het. Die bymidde!s het we! bygedra tot die toksisiteit van die onkruiddo ders.

Die voedi.!1g van snuitkewers gemonitor nadat hulle direk blootgestel is aan onkruiddoders teen konsentrasies van 4 of 12 % . Behandeling met die onkruiddoder wat dikwat bevat gekombineer met die bymiddel Agral het 'n betekenisvolle afname in voeding veroorsaak.

Ander toenames en afnames in snuitkewervoeding is gevind maar hierdie veranderinge was rue betekenisvol nie .

90

Die volgende voedingseksperiment het die reaksie van snuitkewers ondersoek, in die aanwesigheid van waterhiasintblare wat afgesny is en daarna met ondkruiddoder behandel is. 'n Dikwat, 2,4-D amien en glifosaat onkruiddoder is gekies en getoets oor 'n tydperk van ses dae. Voeding het betekenisvol afgeneem op dikwat- en 2,4-D amien-behandelde voedsel, en betekenisvolle mortaliteit is aangeteken by snuitkewers wat op dikwat­ behandelde plantmateriaal gevoed het.

Die beweging van snuitkewers is bestudeer in 'n dam waar die helfde van die waterruasintmat behandel is met glifosaat of dikwat teen 'n aanbevole konsentrasie . Die snuitkewers het in beide gevalle migreer vanaf die behandelde na die onbehandelde deel van die plantmat.

Voorstelle vir gemtegreerde beheer wat gedurende hierdie studie na yore gekom het, is dat daar verkieslik van glifosaat onkruiddoders met 'n lae bymiddelinhoud en teen lae aktiewe bestandeel konsentrasies gebruik gemaak moet word. Dit is noodsaaklik dat reserwe plante beskikbaar is tydens onkruiddodertoedienings, maar die grootte en afstand van sulke matte benodig verdere ondersoek. Konsultasie tussen diegene wat chemiese en biologiese beheer toepas is noodsaaklik vir elke infestasie, om suksesvolle gemtegreerde beheer te verseker in Suid-Mrika asook elders waar waterhiasinte probleme veroorsaak.

91

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement

Table of contents