III. Biological Control of Weeds with Plant Pathogens

118



D. 0. TEBEEST



The endemic fungal pathogen Phytophthoru pulmivoru (Feichtenberger er ul.,

1984) (= Phytophthora citrophthora ( R . E. Sm. and E. H. Sm.) (Burnett e t a l . ,

1973, 1974) was first used commercially as DeVine in 1981 to control stranglervine (=Milkweed vine), Morreniu odoruta, in citrus groves (Kenney, 1986).

Morrenia odorutu was introduced from South America and has become a serious

weed problem in citrus, competing for water, sunlight, and nutrients, girdling

tree limbs, and interfering with harvesting, pesticide applications, and irrigation.

The weed infests approximately 120,000hectares in Florida. Phytophthoru palmivoru was initially isolated in 1972 from diseased and dead plants found in

Orange County, Florida. In small-scale field tests, 96% of the vines were killed

within 10 weeks after infestation of soil with the fungus (Burnett et at., 1973;

Ridings et al., 1976, 1977).

Colletotrichum gloeosporioides fsp. aeschynomene was developed in the

United States and marketed as a microbial pesticide (mycoherbicide) in 1982 as

COLLEGO for the control of northern jointvetch (NJV), A . virginica, in rice and

soybeans in several states in the lower Mississippi River delta (Bowers, 1986).

Northern jointvetch reduces the quality of milled rice and at densities of 1 to I 1

plants/m2, also reduces grain yields from 4 to 19%, respectively. In 1980, 11%

of the rice crop was discounted due to the presence of weed seeds (Smith et al.,

1973), and the discount loss in Arkansas was estimated to be $7.6 million.

Viable spores of the fungus were formulated as a wettable powder to be used,

handled, and applied much like any postemergence chemical herbicide.

Collerotrichum gloeosporioides f. sp. aeschynomene causes an anthracnose on

NJV seedlings, infecting stems, petioles, and leaflets (Daniel et ul., 1973; TeBeest, 1982, 1988, 1990; TeBeest et al., 1978a,b). Enlargement and coalescence

of stem lesions result in the girdling and death of the plant above the lesions. The

fungus sporulates profusely on the lesion surfaces, and rainfall contributes to

dispersal of the fungus spores on the plant, increasing the severity of infection.

The fungus also is dispersed by infected seed (TeBeest and Brumley, 1978) and

by rain-splash (Yang and TeBeest, 1992a). Small-scale field tests demonstrated

that 100% of the seedlings inoculated with the fungus were controlled within a

few weeks after treatment (Templeton et ul., 1981). In the hands of growers, the

commercial formulation of the fungus provides greater than 90% control of NJV

when used according to label directions (Bowers, 1986). COLLEGO has not

been marketed by the registrant since 1992.

A third mycoherbicide, BIOMAL, is composed of spores of C. gloeosporioides (Penz.) Sacc. f.sp. malvue and was registered in 1992 in Canada for the

control of round-leaved mallow (Mulvupusilla Sm.) in wheat (Triticum aestivum

L.) (Grant et al., 1988, 1990; Mortenson, 1988, 1991). Round-leaved mallow

plants inoculated with spore suspensions were killed within 17 to 20 days after

inoculation. The fungus infects leaves, petioles, stems, and crowns of this weed

and kills the plant within a few weeks after application. Control is 90 to 100%

effective in the field and the infected plants do not reappear the following year.



BIOLOGICAL CONTROL OF WEEDS



119



The fungus infects several Malvu species, velvetleaf (Abutilon theophrasti Medic.), and hollyhock (Althea rosea (L.) Cav.), but the disease is severe only on M .

pusilla. Though registered, BIOMAL has not been available commercially.

The following examples of other fungal pathogens under investigation illustrate the variety of targeted weeds and organisms that are being evaluated as

biological control agents.

Colletotrichum orbiculare (Berk. et Mont.) v. Arx is being reevaluated as a

biological control agent for Bathurst bur (Xanthium spinosum L.) in Australia

(Auld et al., 1990; Auld and Tisdell, 1985) after tests conducted earlier were

occasionally promising (Butler, I95 I). When applied as a mycoherbicide, the

fungus controlled 50 to 100% of the seedlings in field tests conducted in 1987

and 1988. The highest levels of control, 98 to loo%, were achieved in a dryland

grazing site (Auld et a / . , 1990). The efficacy of the fungus increases with increasing periods of high humidity, and the presence of the extracellular conidial

matrix hastened the onset of disease symptoms and increased disease levels on

X . spinosum (McRae and Stevens, 1990).

In Japan, two fungi, Drechslera rnonoceras and Epicoccosorus nematosporus

are being investigated for control of two of the four major weeds in rice fields in

Japan. Drechslera monoceras has been reported to give excellent control of

barnyardgrass, Echinochloa species, in greenhouse and field tests (Gohbara and

Yamaguchi, 1993). Combined use of this fungus and the herbicide pyrszosulfuronethyl controlled most of the weeds growing in paddy fields. Similarly, E . nematosporus has been repeatedly effective in controlling water chestnut (Elocharis

kuroguwai) in greenhouse and field tests (Gohbara and Yamaguchi, 1993; Suzuki,

1991). Spore suspensions sprayed onto plants in June and July decreased plant

populations and the number of tubers from which plants emerged into the next

year.

Despite the success of DeVine, much of the work in microbial pesticides has

focused on microbial control of weeds by postemergence application of plant

pathogens to foliage. However, Boyette er al. (1984), Weidemann (1988), and

Weidemann and Templeton (1988) have demonstrated that the soilborne fungus

Fusarium solani (Mart.) Appel & Wr. f.sp. cucurbirae effectively controls Texas

gourd (Cucurbita texana A. Gray) following preemergence application of inoculum to soil. Inoculum consisting of either spores or alginate granules containing

spores and mycelium controlled up to 95% of emerging seedlings.

The soilborne fungus Sclerotinia sclerotiorum (Lib.) de Bary also has been

investigated for control of Canada thistle (Cirsium arvense (L.) Scop.), spotted

knapweed (Centuuria maculosa Lam.), and dandelion (Turuxucumoficinale Weber) (Brosten and Sands, 1986; Miller et a l . , 1989; Riddle et al., 1991). Populations of dandelions in turfgrass were reduced 80 to 85% following repeated

applications of heat-killed perennial ryegrass seed infested with S. sclerotiorum

(Riddle et a/., 1991). Kentucky bluegrass (Pou pratensis L.), creeping bentgrass

(Agrostis pulustris L.), annual bluegrass (Poa annua L, ), and quackgrass



120



D. 0. TEBEEST



(Agropyron repens (L.) Beauv.) were not infected. The potential of S. sclerotiorum as a biological control agent for thistle was demonstrated in the United

States in Montana (Brosten and Sands, 1986). In field trials, 20 to 80% of

Canada thistle shoots were killed after treatment with the fungus applied as

sclerotia or as infested wheat kernels, and plant populations were also reduced

the following year. The fungus also has been reported to be a pathogen of

Centaurea difusa (Watson et a!. , 1974).

Sclerotinia sclerotiorum has a very wide host range and infects many plant

species of economic value. A genetic approach has been utilized to make the

fungus environmentally safe by reducing the effective host range of this pathogen. Miller et al. (1989) reported that an auxotrophic strain was avirulent to four

of seven susceptible hosts unless an exogenous source of cytosine was applied at

the inoculation site. The apparent virulence of the auxotrophic strain was dependent on the inoculum used in these tests. Inoculum consisting of infested millet

seed with no amendments added prior to inoculation resulted in infection of

lettuce (Lactuca sativa L.), clover (Trifolium hybridum L.), and sunflower (Helianthus annuus L.), whereas inoculation with PDA cultures without additives

resulted in infection of lettuce. The addition of yeast extract or cytosine increased

the number of plants of the seven test species infected. Such a genetic approach

to limiting host range may permit the use of pathogens with broad host ranges,

such as S. sclerotiorum, as bioherbicides (Miller et a l . , 1989).

In the United States an endemic rust has also been evaluated for control of a

weed utilizing the augmentative approach (Bruckart and Hasan, 1991) rather

than a truly classical approach. Pucrinia canaliculatu (Schw). Lagerh. has been

evaluated for control of nutsedges, Cyperus rotundus L. and C . esculentus L., in

the United States (Callaway et al., 1985; Phatak et a l . , 1983). When released

early in the spring, the rust inhibits flowering and tuber formation (Calaway et

a!. , 1985). Phatak er al. (1983) had earlier reported that a simple release of the

rust into plots resulted in reductions in root growth and fresh and dry weight of

shoots. This fungus is capable of rapid dispersal and infection. Within 60 days,

approximately 78% of the leaf area was rusted in these tests. In one test, rust

pustules were found in a previously healthy stand of yellow nutsedge within 12

days after a single pot of rusted seedlings were placed in similar yellow nutsedge

stand 7 km away. An epiphytotic reportedly developed over the entire area. The

rust appears to be adapted to a wide range of environmental conditions since

epiphytotics developed over several hectares following releases made thoughout

the growing season under a variety of conditions.



B. CONTROL

OF WEEDS

WITH



THE



CLASSICAL

STRAGEGY



Pathogens used in the classical approach are expected to reduce weed populations to economically insignificant levels as a result of the natural epidemics they

would cause.



BIOLOGICAL CONTROL OF WEEDS



121



The introduction of the rust fungus Puccinia chondrillina into Australia in

1971 from the Mediterranean region for the control of rush skeletonweed

(Chondrilla juncea) appears to constitute the first deliberate introduction of a

pathogen for weed control in any country in what has become known as the

classical approach to biological control of weeds with plant pathogens (Cullen et

a ( . , 1972). Factors that contributed to excellent control of the weed in Australia

included rapid dispersal of spores, a high density of susceptible plants (C.

juncea), virulence of the pathogen to the dominant biotype of the weed, and

favorable weather conditions (Hasan, 1972; Hasan and Jenkins, 1972; Hasan and

Wapshere, 1973). Two strains of P . chondrillina from Eboli, Italy, were introduced into the United States in 1975 (Adams and Line, 1984; Emge e t a l . , 1981).

Within 2 years, the fungus caused severe infections of plants throughout populations of skeletonweed in California, Oregon, Idaho, and Washington (Lee, 1986;

Supkoff et al., 1988).

In 1975, Enryloma agerutinae sp. nov. (Barreto and Evans, 1988) was introduced into Hawaii from Jamaica to control hamakua pamakani (Ageratina riparia (Regel) K.&R.) (Trujillo, 1976, 1985; Trujillo er al., 1988). Ageratina

riparia was determined to be the most serious weed pest of the Hawaiian range

from 800 to 6500 ft in elevation. Releases of the fungus from November of 1975

through May of 1976 resulted in an epidemic and devastation of hamakua pamakani. Weed populations were reduced from 80 to <5% of the original population within 1 year. Control of the weed was most effective between 1500 and

6000 ft elevation where the temperature and rainfall were conducive of disease

development. More than 50,000 hectares of pastureland have been rehabilitated

by removal of the weed by C. ageratinae.

Coiletotrichum gloeosporioides f . sp. clidemiae was isolated from diseased

leaves of Clidemia hirta (L.) D. Don collected in Panama by Trujillo et al.

(1986) and Trujillo (1992). Clidemia hirta is a major noxious weed of tropical

American origin and was introduced accidentally into Hawaii in 1941 where it

now infests more than 40,000 ha of rain forest areas on Oahu. It has also spread

to several other Hawaiian islands. Trujillo er al. (1986) have found that C.

gloeosporioides f.sp. clidemiae is an aggressive pathogen and specific to C. hirta

and concluded that this pathogen can be an effective biological control agent for

this weed in Hawaii if introduced.

A study of Septoria passijlorae in Hawaii by Trujillo er al. (1994) confirmed

that this pathogen is specific to Passzjlora rripurtira (Juss.) Poir. var. tripartita

Holm-Nie. Jorg. & Laws. This plant was first introduced into Hawaii in 1921 as

an ornamental but has become an aggressive, weedy species of high-elevation

areas on Kauai and Hawaii. Cultures of S.passiflorue were obtained from infected plants of PassiJora planted near Aldana, Narino, Colombia. Septoriu passiflorue appears to be specific to P . tripartita and P . foetida after host range tests

were conducted in Hawaii. Both are introduced weedy species in Hawaii. The

fungus does not appear to be pathogenic to economic plants. Since the Pas-



122



D.0. TEBEEST



sifloraceae are not native to Hawaii, S. pussiflorue is presumed to be an environmentally safe biological control agent for these weeds in Hawaii following introduction (Trujillo et uf., 1994). Introduction has been recommended because the

probability for success appears to be excellent and host range is very restricted

(Trujillo et af., 1994).

Currently three imported rust fungi are being evaluated in the United States for

biological control of leafy spurge (Euphorbiu esulu L.), musk thistle (Curduus

nutuns L. ssp. leiphyflus (Petrovic) Stoj. & Stef.), and yellow starthistle (Centuureu sofstitiafisL.) in the United States (Bruckart and Dowler, 1986; Melching

et ul., 1983).

A study of I8 isolates of Mefumpsoru euphorbiue (Schub.) Cast. collected in

Austria, Hungary, Romania, Switzerland and Yugoslavia resulted in only one

compatible combination on two collections of cypress spurge (Euphorbiu cyparissius L.) (Bruckart and Dowler, 1986). Several other isolates produced infections that were difficult to maintain on both leafy and cypress spurge (Turner et

a / . , 1983).

An isolate of Puccinia curduorum Jacky collected in lbrkey was found to be

very aggressive on 23 of 27 collections of musk thistle from Canada, France, and

the United States (Bruckart and Dowler, 1986). The rust was pathogenic to 8 of

17 Cirsium species tested but was much less aggressive on species other than C.

nutans L. The susceptibility of musk thistle and related composites has also been

determined (Politis et a l . , 1984). This rust was released for field studies in the

United States in 1987 after it was determined that it was not aggressive on globe

artichoke (Cynaru scolymus L.) (Politis et al., 1984). Baudoin ef ul. (1993) have

suggested that P . carduorum, particularly in combination with insects, can contribute to the reduction of seed production and control of musk thistle.

Pucciniu jaceae Otth. was evaluated for control of yellow starthistle with rust

collections from Tbrkey. Isolates of the rust were very aggressive on yellow

starthistle but also infected safflower (Carthumus tinctorius L.) (Bruckart and

Dowler, 1986).

The effects of the rust Pucciniu lagenophorue Cooke on groundsel (Senecio

vulgaris L.) have been studied under summer and winter conditions in the United

Kingdom (Paul and Ayers, 1986a,b, 1987a,b) after introduction into Great Britain in 1961 (Paul and Ayers, 1987b). Field-grown groundsel, infected as a result

of inoculations made in the autumn, showed 70% mortality when measured in

the spring whereas mortality was only 40% for plants inoculated during the

spring and summer. The higher mortality of plants inoculated in the autumn was

attributed to infection of the hypocotyls which usually killed the host plants

within 1 to 2 weeks and compromised the ability of surviving seedlings to

withstand water stress. Infection of hypocotyls in the summer, though as severe,

did not result in significant mortality although infection had very substantial

effects on growth and reproduction of populations during the summer. Infection



BIOLOGICAL CONTROL OF W E D S



123



resulted in a 50% reduction of vegetative growth, a 15% reduction in plant

density, and a 24% reduction in floret production. The fungus does not naturally

survive at sufficiently high levels to cause significant mortality. In this case,

manipulation of the host-pathogen population relationship over time has a large

effect on host survival. In glasshouse and field experiments, rust infections of

groundsel decreased the competitive ability of groundsel with lettuce, although

lettuce yields were not significantly reduced by rust-infected plants until weed

densities exceeded 4000 plants per square meter (Paul and Ayers, 1987a). Rust

infections reduced the impact of groundsel on lettuce yield without causing any

significant increase in groundsel mortality.



C. BIOLOGICAL

CONTROL

OF AQUATIC

WEEDS

Several plant pathogens have been or are currently under investigation for

biological control of aquatic weeds such waterhyacinth, water milfoil, duckweeds, alligator weed, waterlettuce, or blue-green algae, in a variety of aquatic

environments (Joye, 1990). Pathogens of aquatic weeds that have been tested as

microbial pesticides include species of Fusarium and Macrophomina, on hydrilla

(Hydrilla verticillatu (L.F) Royle), and species of Acremonium, Colletotrichum,

Fusarium, Pythium, and Phytophzhoru for control of eurasian watermilfoil (Myriophylfum spicatum L.), but no promising control agents have been found among

these isolates (Joye, 1990).

Experiments were conducted with potential commercial formulations of Cercospora rodmanii Conway for control of waterhyacinth (Eichhornia crassipes

(Mart) Solmes) (Charudattan et a/., 1985; Conway, 1976a,b). This fungus has

been released in South Africa for control of waterhyacinth in the Crocodile river

using a classical approach (Morris and Cilliers, 1992).

An isolate of Colletotrichum gloeosporioides was also tested as a potential

biological control agent for eurasian watermilfoil (Myriophyllum spicatum L.)

(Smith et a l . , 1989). Under realistic conditions, the effect of this fungus on

watermilfoil was too small to warrant further consideration as a possible biological control agent. Recently, Microleptodiscus rerrestris (Gerdemann) Ostazeski

was reported to have considerable impact on the populations of milfoil in Florida

tests (Joye, 1990). In recent work, Verma and Charudattan (1993) showed that

this fungus was pathogenic to 3 (Hydrilla verticilata, Myriophyllum aquaticum,

and Ceratophyllum demersum L.) of 16 aquatic plant species tested. Only on

Hydrillu did the fungus cause plant mortality comparable to levels achieved by

infection of watermilfoil. The remaining 13 species were not infected. Of 17

terrestrial species tested, seed germination was significantly affected by the fungus although postemergent symptoms of disease developed on seedlings of only

10 species. On 4 of these, Medicago sativu L., Lotus corniculatus L., Trifolium



I24



D. 0.TEBEEST



hybridum L., and T . repens L., disease developed to affect from 26 to 50% of the

host plant tissues. Shearer (1994) reported that application of a formulation of M.

terrestris was ineffective in reducing aboveground biomass of eurasian watermilfoil under natural conditions in the field, but poor field performance was

attributed more to fungus/formulation problems than to biological, chemical, or

physical factors encountered in this test.



D. BIOLOGICAL

CONTROL

OF WEEDS

WITH BACTERIAL

PLANTPATHOGENS

Although bacterial diseases of weeds have been known and described for

many years (Rosen, 1924), until recently, few have been investigated for potential biological control of weed species.

Recently, Caesar ( 1994) suggested that strains of Agrobacterium tumefaciens

(E. F. Smith & Town) isolated from important rangeland weeds may be effective

as biological control agents for their respective hosts. Host ranges of isolates of

A . tumefaciens from Russian knapweed and leafy spurge and strains representing

biovars I and 2 of A . tumefaciens and A . viris varied greatly with six strains being

pathogenic to no more than one species in addition to the original host. Strains

from New Mexico were highly pathogenic to diffuse knapweed (Centaurea diffusa L.), causing girdling, stunting, and death of this host.

Zhou and Neal (1995) recently compared strains of Xanrhomonas campestris

(L. R . Jones et a / . ) pv. poannua as biocontrol agents for annual and perennial

subspecies of annual bluegrass ( P . annua L.). Results of controlled growth

chamber and field tests showed that two strains of this bacterium were similarly

virulent in both tests. In growth chamber tests, annual and perennial subspecies

of P . annuu were controlled 82 and 92%.respectively, while in field tests control

reached only I 1 and 7%, respectively, following repeated weekly applications.

Control of annual bluegrass was only 40% following 4 weeks of three applications per week. The weeds also recovered 2 to 5 weeks after weekly inoculations

were stopped. Johnson (1994) reported that three applications of two strains of X .

campestris pv. poannnua controlled between 52 and 82% of the annual bluegrass

in dormant bermudagrass (Cyanodon transvaafensis Burtt-Davy X C . ducryfon

(L) Pers.) field plots.

Begonia et al. (1990) have demonstrated in culture tube assemblies that isolates of Pseudomonas and Erwinia herbicola caused velvetleaf (Abutilon theophrasti) seedlings to become chlorotic and develop abnormal root systems

compared to noninoculated controls.



125



BIOLOGICAL CONTROL OF WEEDS



IV. BIOLOGICAL CONTROL OF WEEDS BY MICROBIAL

MANAGEMENT OF SEED BANKS

As seen above and as widely reported in previous reviews (Templeton et a l .,

1979; TeBeest, 1991), considerable attention has been given to developing pathogens as microbial pesticides or as classical agents for control of weeds. Weed

seed banks are considered to be the major source of weed infestations in arable

lands (Cavers and Benoit, 1989). Recent work by Kremer (1993) and Kremer et

al. (1990) has suggested that deleterious rhizobacteria may be useful in reducing

weed seed banks in yet another approach to reducing weed infestations in crop

and range lands.

Kennedy et al. ( 1991 ) screened more than 1000 isolates of pseudomonads, and

81 inhibited downy brome (Bromus rectorum) but not wheat. Six isolates consistently inhibited downy brome growth but not wheat in controlled environments.

In some field tests, isolates reduced downy brome populations up to 30% and

shoot dry weights by 42%. Winter wheat yields were increased in two of three

field locations because of suppression of downy brome.



V. SYNERGISMS THAT MAY AFFECT THE

EFFECTIVENESS OF MICROBIAL, AGENTS

The term synergism is used loosely here to mean a combined use of insects,

chemicals, or pathogens that enable pathogens to control weeds when the individual activities of the interactive participants are less effective. Synergism as

used here should also not be confused with the integrated use of various components (i.e., biological and chemical pesticides) that may or may not be inhibitory

of each other’s activities but that, nevertheless, have been employed in effective

control schemes employing both components (Klerk et al., 1985).



A. SYNERGISM

OF PATHOGENS

WITH *HER



PATHOGENS



Several examples have been reported in which pathogens incapable of causing

significant levels of disease when infecting alone were more severe in combination with other pathogens. For example, Dimock and Baker (1951) showed that

Fusariiim roseum Lk. emend. Snyd. & Hans. infected snapdragon (Antirrhinum

majus L . ) through lesions caused by the rust fungus Puccinia antirrhini D.& H.

Apparently, F . roseum infected healthy tissue beyond the rust lesion and caused



126



D. 0. TEBEEST



death of leaves and shoots or even entire plants whereas infection by the rust

alone seldom caused death. Rust-free plants, even when severely wounded, were

not infected by F. roseum. Thus it appears that a facultative parasite, incapable of

infecting a plant alone, contributed to increased disease severity by invading

lesions produced by another pathogen. This phenomenon has been extended to

biological control of weeds.

Alternaria macrospora Zimm. has been investigated as a potential mycoherbicide to control spurred anoda (Anoda crisrata (L.) Schlecht.) in the United

States (Crawley et al., 1985; Walker, 1981; Walker and Sciumbato, 1979). The

susceptibility of spurred anoda to A . rnacrospora is highly correlated with plant

age. Approximately 100% of seedlings inoculated at the cotyledonary stage were

killed by infection, but less than 10% were killed at the three- to four-leaf stage

(Walker and Sciumbato, 1979). However, 100% of plants inoculated at the fourto five-leaf stage were killed by the interaction of A . rnacrospora and Fusarium

lateritium Nees ex Fr. Fusarium lateritium is a weak pathogen of spurred anoda

and causes less than 20% mortality when inoculated to seedlings alone, although

F. laterilium usually killed wound-inoculated spurred anoda seedlings (Crawley

et al., 1985). The highest mortality occurred when Afternaria was inoculated 5

days before Fusarium. When Fusarium was inoculated 5 days before Alternaria,

only 1 1 % of the seedlings were killed. One suggested explanation for the Afternarial Fusarium interaction was that F . lateritium penetrated and infected

through the lesions caused by Alternaria. These results indicate that sequential

applications of both fungi were more effective than either fungus used alone for

control of spurred anoda.

Mortality of groundsel infected by P . lagenophorae has been attributed to

invasion of rust lesions by Botntis cinerea Pers. (Hallett et al., 1990a,b). Inoculation of healthy groundsel with B . cinerea caused only 10% mortality and only

40% mortality of abiotically wounded plants: however, all plants previously

infected by P . fagenophorae died after inoculation with Botrytis. Death of plants

was attributed to growth of Botryris into stems. The time necessary to kill plants

was dependent upon several factors, including the inoculum concentration of

Botrytis and initial pustule numbers of the rust (Hallett, 1990a.b).



B. SYNERGISM

OF PATHOGENS

WITH INSECTS

Insects and plant diseases may have played a role in the control of prickly pear

in Australia and other countries (Wilson, 1969). However, the relative importance of pathogens in biological control of prickly pear is not clear since most of

the intensive research focused on the role of insects and Cactoblastis cacrorum

Berg rather than on potential fungal and bacterial pathogens (Wilson, 1969).

Similarly, introduction of the insect Proceidorchures utilis Stone into Australia