Plant Protection Quarterly Vol.12(2) 1997 57 static at around 200–300 ha (Agriculture Protection Board of Western Australia St. John’s wort ( perforatum) in Western 1988). Australia In recent years there has been little spread of the weed. Only one new infesta-

A B tion was reported between 1988 and 1992. Jonathan Dodd and John K. Scott The lack of spread may partly be because A Weed Science, Agriculture Western Australia, Baron-Hay Court, South Perth, there is little agricultural activity in in- Western Australia 6151, Australia. fested areas that could lead to dispersal of B CRC Weed Management Systems, CSIRO Division of Entomology, Private Bag, rhizomes by cultivation etc., or the spread PO, Wembley, Western Australia 6014, Australia. of seeds in contaminated hay or via stock movement.

Abstract neglected small-holdings located within Early control activities St. John’s wort () forested country. The first weed control trials were in 1936. is a minor localized weed in south-west The initial extent of St. John’s wort in- Salt was used extensively in the early trials Australia where its distribution and festation in 1935 was 70 acres (28 ha) be- (Hamilton 1946). After a decade of control abundance have been stable in recent tween Margaret River and Augusta (cen- attempts, Hamilton (1946) concluded that years. Factors significantly contributing tred on Karridale, Figure 1), with smaller eradication would be difficult. In 1950 St. to the weed’s status may include: infestations later reported over the range John’s wort was declared a noxious weed i. noxious weed legislation that ensured currently occupied (Figure 1) (Hamilton under the new Agriculture and Related regular inspection and chemical con- 1946). In recent years the average area re- Resources Protection Act. It was noted trol activities, ported each year to be infested by St. that maintaining a dense pasture, espe- ii. the successfully established biological John’s wort in Western Australia is about cially one containing subterranean clover control agent, quadrigemina 44 ha on 11 properties (Table 1). While and kikuyu grass, helped suppress the and these data reflect the resources and effort weed (Meadly 1956). Also, the new, cheap, iii.climate. made in inspecting the weed, they do rep- easily-applied herbicides (such as 2,4-D These factors should form the basis for resent the minimum levels of infestation. ester) appeared in the early 1950s and future research activities in Western Aus- The actual area is estimated to be relatively their application resulted in a substantial tralia before the introduction of further biological control agents is considered.

Introduction St. John’s wort was first collected in West- ern Australia by C.A. Gardner in 1934 (Hamilton 1946) and was identified as H. perforatum var. angustifolium (Clusiaceae). The weed is believed to have been intro- duced late last century, or early this cen- tury, with fodder imported for horses and oxen involved with the timber industry. Hamilton (1946) noted that ‘prolific growth’ was not reported until 1935. St. John’s wort has not achieved the prominence in Western Australia that it has attained elsewhere in Mediterranean- type climates of the world, for example in California. Nevertheless, the plant has been subjected to control activities since its discovery in the State. In this note we summarize what is known about St. John’s wort and its control in Western Australia.

Distribution St. John’s wort is restricted to high rainfall areas of south-western Western Australia. It occurs mainly as scattered small populations, with the greatest densities being found from Dwellingup, Bodding- ton, Balingup, Greenbushes, Nannup, Manjimup, Karridale to Pemberton (Fig- ure 1). Recorded populations near Perth, Northam and Narrogin (Figure 1) are not thought to be extant. Only a few popula- tions occur in pastures. Most major infes- Figure 1. Distribution of St. John’s wort (Hypericum perforatum) in Western tations are found along roadsides and Australia. Dots indicate localities where plants have been recorded either as on the sites of former timber camps, herbarium specimens or in the records of Agriculture Western Australia. former milling towns (e.g. Holyoake) and Isohytes indicate average annual rainfall in mm. 58 Plant Protection Quarterly Vol.12(2) 1997 reduction in the extent of St. John’s wort agents in eastern Australia (Campbell et al. recent removal of St. John’s wort from (Meadly 1956). 1995), but have not been released in West- noxious weed lists leads to a change in its ern Australia. This is presumably due to abundance and distribution. In South Af- Biological control the minor nature of the weed problem in rica, St. John’s wort occurs in a similar cli- Four colonies of the biological control Western Australia. mate to south-west Australia and is now agents, Chrysolina quadrigemina (Sufrian) regarded as under satisfactory biological and Chrysolina hyperici (Forster) (these Current status and control measures control mainly due to C. quadrigemina. were originally identified as Chrysomela Between 1986 and 1992, the Agricultural Gordon and Kluge (1991) conclude that gemellata (Rossi) and Chrysomela hyperici Protection Board of Western Australia other factors such as climate, unsuitable (Forster)) were sent from Bright, Victoria treated an average of 56 ha of St. John’s habitats and lack of dispersal agents may to Western Australia by the Council for wort infestations by herbicide each year have contributed to containing the weed Scientific and Industrial Research and re- (Table 1). In 1989 the declaration of St. and these factors may also be relevant to leased in November 1948 at Holyoake, John’s wort under the Agriculture and Re- the situation in south-west Australia. Mornington Mills and Karridale in West- lated Resources Act was cancelled in As is unfortunately often the case, the ern Australia (Figure 1). It was estimated south-west areas (Agriculture Protection biological control attempts in Western that about 100 000 (presumably Board of Western Australia 1989). In other Australia have not been subjected to criti- adults) were released (Western Australia areas of the State, where the plant is pres- cal assessment that measures their impact Department of Agriculture 1948). A fur- ently absent, it remains a declared plant in on St. John’s wort. Huffaker’s (1967) infer- ther consignment was received from Vic- the P1 (plants which must not be intro- ence that biological control had been suc- toria in November 1949 and released in duced to the State) and P2 (plants which cessful in Western Australia may be valid the Karridale district (Western Australia must be eradicated) categories. because the ability of the weed to recover Department of Agriculture 1950). Estab- Current recommended chemical con- from C. quadrigemina attack is known to be lishment of the beetles and destruction of trol measures are to use 2,4-D amine or increased by summer rainfall and variabil- one small outbreak of the weed was re- ester for spot applications and boom- ity in winter rainfall in south-eastern Aus- ported the following year. spraying. Diuron can also be used for tralia (Huffaker 1967, Julien 1992), but In 1952 G.R.W. Meadly wrote ‘Some spot-treatment of small infestations and these factors are both negligible in south- encouraging results have been obtained glyphosate is used for non-selective con- west Australia. The chrysomelid–St. John’s [by the chrysomelids], but this method trol (Peirce and Smith 1994). wort interaction may be the best example cannot be regarded as highly reliable’. It is of successful biological control in Western likely that the beneficial aspects of biologi- Discussion Australia, so successful that the weed is cal control were eclipsed by the arrival of Five Clusiaceae species occur in Western now comparatively insignificant and its the modern herbicides that were easy to Australian flora; three native species, biological control agent almost forgotten. apply, cheap and gave readily observable Calophyllum sil Lauterb., Hypericum Clarification of the weed’s identification results. By 1956 the familiar pattern of bee- gramineum G. Forster, H. japonicum Thunb. and an assessment of the potential spread tles controlling plants in pasture, but not in and the introduced H. androsaemum L. and and impact of existing biological control shaded areas, had been observed (Meadly H. perforatum L. (Green 1985, Keighery agents would provide a basis for further 1956). 1995). However, there has been no critical biological control releases. Some of the Huffaker, a leading biological control reassessment of the of these available agents (e.g. the aphid Aphis entomologist from California, visited Aus- species. Detailed identification of the Hy- chloris) would be relatively simple to re- tralia in 1963 to examine St. John’s wort. pericum populations in Western Australia lease in Western Australia. However, it He reported that the control of St. John’s may be important to ensure correct appears that the weed is under successful wort in Western Australia was potentially matching of biological control agents (e.g. biological control by the chrysomelids in similar to the situation in California where the mite, Aculus hyperici Liro) to their host open or pasture areas and future control high levels of persistent control occurred plant. attempts should be directed at the plant in (Huffaker 1967). He stated that ‘very ef- Gordon and Kluge (1991) comment on shaded or forested areas. fective, well synchronized attack [by the the lack of invasion by St. John’s wort into chrysomelid] that resulted in a high degree fynbos (native heath) vegetation in south- Conclusions of death of the defoliated plants observed west South Africa. However, St. John’s The availability of additional agents opens was obvious in West Australia’. wort is regarded as a major environmen- the opportunity for further releases in Chrysolina quadrigemina adults were tal weed in eastern Australia (this work- Western Australia. The minor importance present on St. John’s wort at Holyoake in shop). There appears to be little invasion of the weed make it unlikely that funding November 1992, and at Karridale in De- of native vegetation in Western Australia would be attracted to implement a large cember 1994. The fate of C. hyperici is un- at present, even though the plant has scale research program. However, the bi- known. Four other species have been present in forested areas for a con- ology and ecology of the weed and the been established as biological control siderable time. It remains to be seen if the control agents represents an excellent

Table 1. Records of the number of properties and area infested and treated for St. John’s wort in Western Australia. Data were taken from the Annual reports of the Agricultural Protection Board, Western Australia (1986–1992). Year New properties New area No. properties Area No. properties Area infested infested (ha) infested (ha) treated treated (ha) 1986/87 –* – – – 59 200 1987/88 – – 10 60 31 71 1988/89 1 1 18 80 10 10 1989/90 0 0 11 34 10 23 1990/91 0 0 5 25 2 20 1991/92 0 0 11 20 5 11 * not recorded. Plant Protection Quarterly Vol.12(2) 1997 59 opportunity for basic studies, perhaps at the post graduate student level. Population dynamics of St. John’s wort in south- Acknowledgments eastern Australia We thank Sandy Lloyd for helping search the files on St. John’s wort. We also thank Kathryn Robinson and Paul Yeoh for their D.T. Briese, CSIRO Entomology and Co-operative Research Centre for Weed comments on the manuscript. Management Systems, GPO Box 1700, Canberra, ACT 2601, Australia.

References Agriculture Protection Board of Western Summary key element in the development of man- Australia (1988). Annual report for the A thorough understanding of the ecology agement strategies for the weed. year ending June 30 1988, 79 pp. of St. John’s wort and the processes that N. Clark’s (1953) studies indicated that Agriculture Protection Board of Western drive the population dynamics of infesta- St. John’s wort can have quite different Australia (1989). Annual report for the tions is a key to the management of this forms and population structures, depend- year ending June 30 1989, 96 pp. weed. However, while some aspects of ing on the habitat in which it grows (Table Campbell, M.H., Briese, D.T., and the weed’s biology have been studied in 1). In Type B infestations, on poorer shal- Delfosse, E.S. (1984). The biology of the seventy years of research on its con- low soils, the weed is smaller and tends to Australian weeds 13. Hypericum trol in Australia, there has been only one sucker more frequently. Stresses such as perforatum L. In ‘The Biology of Austral- study devoted to its ecology; in the Ovens low nutrient levels, shading, defoliation ian Weeds’, eds. R.H. Groves, R.C.H. Valley, Victoria, over a two year period. and fire (N. Clark 1953, Briese 1996) tend Shepherd and R.G. Richardson, pp. To complement this detailed work and to promote such vegetative reproduction 149-68. (R.G. and F.J. Richardson, Mel- provide a picture of longer term fluctua- and in fact render such infestations rela- bourne). tions, four St. John’s wort populations tively resistant to control. Where soils are Gordon, A.J. and Kluge, R.L. (1991). Bio- were monitored over a period of seven better, larger deep-rooted plants may give logical control of St. John’s wort, Hy- years, from 1981–87. This paper summa- the impression of a more vigorous infesta- pericum perforatum (Clusiaceae), in South rizes the type of data collected, and uses tion, but in fact, plants of these Type A in- Africa. Agriculture, Ecosystems and Envi- it, together with that from the Ovens Val- festations are more susceptible to stresses ronment 37, 777-90. ley study, to define those properties of St. such as defoliation (N. Clark 1953). More- Green, J.W. (1985). ‘Census of the vascular John’s wort that have contributed to its over, once the weed is reduced the richer plants of Western Australia’, 312 pp. success as a weed and which may render it soils can more readily support competing (Western Australian Herbarium, Perth). vulnerable to particular methods of con- vegetation and so St. John’s wort is more Hamilton, A.L. (1946). The eradication of trol. readily controlled. This partly explains the St. John’s wort in Western Australia. original success of biological control Western Australian Journal of Agricul- Introduction agents in areas where good agricultural ture, 2nd Series 23, 70-4. Since its introduction into Australia in the land was infested. The problem remaining Huffaker, C.B. (1967). A comparison of the late nineteenth century, St. John’s wort is that the majority of land currently in- status of biological control of St. Johns (Hypericum perforatum L.) has spread fested by St. John’s wort falls into the Type wort in California and Australia. Mushi widely and invaded a number of different B category, with over 80% of infestations 39, 51-73. habitat types, ranging from prime agricul- occurring under native forest (Shepherd Julien, M.H. (1992). ‘Biological control of tural land to native Eucalyptus forest 1983) or in poorer quality pastoral land. weeds: a world catalogue of agents and (Campbell et al. 1995). Aspects of the gen- In addition to aspects of the physical their target weeds’, 3rd edition, 186 pp. eral biology of the weed that favour its habitat, factors such as rainfall pattern can (CAB International, Wallingford). success in colonizing and persisting in new affect stem and seed production (by as Keighery, G.J. (1995). An annotated list of areas are now reasonably well understood much as 7.5 to 26-fold, respectively the naturalized vascular plants of West- in the light of numerous efforts to control (Campbell et al. 1995)), while defo- ern Australia. In ‘Invasive Weeds and the weed (see Campbell et al. 1995), but, liation can cause fluctuations in crown den- Regenerating Ecosystems in Western apart from a two year study in the Ovens sity and stem production (Clark and Clark Australia’, ed. G. Burke, Appendix 2, pp. Valley, Victoria, by N. Clark (1953), there 1952). To better understand the longer 71-101. (Institute for Science and Tech- has been no attempt to specifically investi- term changes in infestations of St. John’s nology Policy, Murdoch University). gate the population dynamics of infesta- wort in such habitats, a study was under- Meadly, G.R.W. (1956). St. John’s wort tions of the weed over a longer period. taken by CSIRO Entomology between (Hypericum perforatum L. var angusti- Given that St. John’s wort is a relatively 1981 and 1987 at four sites in south-eastern folium DC.). Western Australia Journal of long-lived perennial, such information is a Australia, infested in varying degrees by Agriculture 3rd Series 5, 661-4. Peirce, J.R. and Smith, F. (1994). ‘Declared Table 1. Characteristics of the two types of St. John’s wort infestation Plant Control Handbook’, 4th edition. recognised by N. Clark (1953). The infestations have been designated Type A (Agriculture Protection Board of West- and Type B for this paper. ern Australia, Perth). Western Australia Department of Agricul- Parameter Type A Type B ture (1948). Annual report of the De- partment of Agriculture for the year Location deeper soils shallow or stony soils 2 ended 30th June, 1947, 46 pp. (Govern- Densities (plants per m ) 12–37 12–124 ment Printer, Perth). Plant height taller shorter Western Australia Department of Agricul- Number of stems multistemmed fewer stems ture (1950). Annual report of the De- Root system deep tap-root shallow lateral roots partment of Agriculture for the year Vegetative growth rare common ended 30th June, 1950, 46 pp. (Govern- Maximum age of plants usually more than 3 years usually less than 3 years ment Printer, Perth). Response to defoliation more susceptible less susceptible