<<

Protection Quarterly Vol.9(3) 1994 101 whereas flower and fruit production in triacanthos L. (Caesalpiniaceae), another females is highly variable from year to year (Schnabel 1988, Schnabel and thorny, exotic fodder gone wild Hamrick 1990b). Flowers emerge in spring, either after or at the same time as S.M. CsurhesA and D. KriticosB the . Female individuals produce loose racemes containing 10–40 small, yellow-green, perfect flowers with abor- A Queensland Department of Lands, Land Protection Branch, GPO Box 1401, tive stamens. Most male individuals pro- Brisbane, Qld 4001, Australia. duce dense racemes that often contain B Queensland Department of Lands, Applied Research Unit, Locked Bag 40, more than 100 greenish staminate flowers Coorparoo DC, Qld 4151, Australia. which lack pistils (Schnabel and Hamrick 1990b). Schnabel and Hamrick (1990b) have observed that the number of Summary G. triacanthos is reported to reach its produced by hermaphrodite varies Attention is drawn to the proliferation greatest size in the valleys of small from year to year, but is always much less of Gleditsia triacanthos L. (honey locust streams in southern Indiana and Illinois than the number produced by female in- tree) at several locations in south east (Sargent 1965), where it can form dense dividuals. Yearly variation in pollen and Queensland and New South Wales. Al- thickets. Throughout the mid-western production is also common for though promoted and planted as an at- United States G. triacanthos invades pas- purely male and female individuals, re- tractive garden ornamental and as a high tures and old fields (Schnabel et al. 1991, spectively. protein fodder tree, this thorny, exotic Schnabel and Hamrick 1990a) and often In the United States, G. triacanthos is tree has the potential to smother pas- co-occurs in woodlands and early succes- pollinated by a variety of insects, includ- tures and replace native riparian vegeta- sional habitats with Maclura pomifera ing bees, moths and butterflies (Schnabel tion. G. triacanthos appears well suited (osage orange), another deciduous tree 1988). Schnabel (1988) has shown that to sub-tropical and temperate climates species. 15–50% of pollination in three G. tria- and could become a serious pest on allu- Although capable of tolerating most canthos sites was by pollen originating vial soils throughout the south-eastern soil types, including soils which are mod- outside the sites. These values are similar states of Australia. In Queensland, erately saline, G. triacanthos grows most to pollen migration rates recorded for G. triacanthos was declared noxious in prolifically on fertile, alluvial soils with a wind-pollinated species (Friedman and March 1993 and an eradication program pH of 6.0–8.0 (NAS 1980). The plant de- Adams 1985, Nagasaka and Szmidt 1985, has been implemented. Where feasible, velops an extensive lateral root system Harju and Muona 1989). early control of this species in other with a weakly developed tap root (Bedker Like many weeds, G. triacanthos is a States is recommended. and Blanchette 1983). Good growth oc- prolific seeder. It produces long (20–30 curs in areas receiving from 500 mm to cm), flattened pods which may contain Introduction more than 1500 mm annual precipitation up to 30 seeds separated by a sweet pulp A range of exotic shrub and tree species (NAS 1980), although the plant has been (Gordon 1966, Mathwig 1971). Pods are have been imported into Australia to im- reported to survive in areas which receive first produced when the plant is 3–5 years prove the quality and availability of fod- an annual rainfall as low as 355 mm old. The number of pods produced can der for cattle and sheep. Unfortunately, (Mitchell 1978). increase dramatically as the tree matures. some species have spread to become de- G. triacanthos is subdioecious—a small over 12 years of age may produce in structive weeds, forming dense thickets number of hermaphroditic individuals excess of 500 kg of pods per tree (Hart, and outcompeting pasture species over may occur within a population of undated). Seed dispersal can be highly lo- large areas. This paper addresses the dioecious individuals (Schnabel and calized since most of the pods fall directly long-term pest potential of Gleditsia Hamrick 1990b). Male trees flower yearly, beneath the parent tree. Secondary triacanthos (the honey locust tree), an in- troduced fodder tree which is believed to be in its early stages of spread in Aus- tralia.

Description and biology Gleditsia triacanthos L. is a deciduous, le- guminous tree often reaching 25 m in height. The trunk and branches are pro- tected by an armoury of long, strong spines some of which may be 15 cm in length (Figure 1). G. triacanthos is native to central and eastern North America, ranging from On- tario south to Texas (Little 1971). The spe- cies occupies a variety of habitats throughout its natural range, including woodlands, rocky upland slopes and abandoned pastures (Schnabel and Hamrick 1990a). It is particularly suited to the Central Plains region of the United States and grows vigorously on fertile al- luvial soils associated with the Ohio and Mississippi Rivers (Skerman 1977). Figure 1. Thorns of G. triacanthos. 102 Plant Protection Quarterly Vol.9(3) 1994 dispersers include cattle, horses, deer, form and foliage, is tolerant of poor soils, Hamrick 1990a). Sargent (1965) com- rabbits and a variety of small mammals is easily propagated, and is readily trans- ments that G. triacanthos in the United (Bugbee and Riegel 1945, Dice 1945, planted (Dirr 1983, Bastian and Hart States occasionally covers considerable Peattie 1953, Fowells 1965) all of which 1990). A large number of and areas on moist fertile soils, excluding may facilitate spread. Field observations variants have been developed by horticul- most other species. of seed dispersal in south east Queens- turists. So called “thornless” variants of land suggest the species is spread by this species, included under the group Current distribution and pest cattle (which ingest the pods and pass the name G. triacanthos var. inermis, include potential in Australia seed in their dung) and by floodwater “Moraine”, “Shademaster”, “Rubylace” Due to its desirability as a garden orna- (the pods float). and “Sunburst” (varieties described by mental and fodder tree, G. triacanthos has The seed of G. triacanthos requires dam- Bradley 1978, Rowell 1991). Seeds pro- been planted throughout Australia. Dis- age before it will germinate and is be- duced by thornless honey locust varieties tribution data collected from State Her- lieved to be effectively scarified by pass- bear true for that characteristic for up to baria indicate, however, that the species ing through the digestive systems of 80% of any one seed lot (Hart, undated) has become naturalized only in Queens- cattle. Scarified seed has 10 times the whereas the remaining 20% may produce land and New South Wales. Major infes- emergence of unscarified seed (Burton thorny trees. Ornamental varieties can be tations occur in south east Queensland and Bazzaz 1991). Laboratory experi- grafted onto thorny rootstock, which pro- along the Condamine River and its tribu- ments showed 100% of scarified seed ger- duce thorny suckers if damaged (Hart, taries (Darling Downs region) and in the minated within 60 days. Unscarified seed undated). Brisbane Valley (Esk and Toogoolawah field sown in 1985 emerged in 1987 and Whereas G. triacanthos grows pest-free area). The latter infestation is believed to even into 1988 (Burton and Bazzaz 1991) in Australia, it is subject to attack by a have originated from a small number of suggesting a dormancy period of up to range of insects and fungal diseases in ornamental specimens planted by a Mr. three years under field conditions. North America. North American pests in- McConnell early this Century (hence the The ability of G. triacanthos to resprout clude the canker causing pathogens local name of “McConnell’s curse”). The multiple stems after injury to the primary Nectria cinnabarina (Bedker and Blanch- species was first reported as a pest in stem suggests that many clumps may be ette 1983) and Thyronectria austro- Queensland in 1955 at a location adjacent single individuals (Fowells 1965, Smith americana (Crowe et al. 1982) and the to the original planting by McConnell and Perino 1981). honey locust plant bug (Diphnocoris (Lands Department records). In Queens- G. triacanthos has been promoted and chlorionsis) (Herms et al. 1987). The most land, G. triacanthos has formed dense planted as a fodder tree in North damaging pest of ornamental varieties is thickets amongst native vegetation along America, Europe, Australia, New Zea- the mimosa webworm moth (Homadaula creek and river banks and has invaded land and South America (NAS 1980). The anisocentra Meyrick) which can cause re- improved pastures on surrounding allu- leaves have a crude protein content rang- peated defoliation (Bastian and Hart 1990, vial clay soil flats (Figure 2). Similarly, ing from 14.3–17.3% (Skerman 1977) and North and Hart 1983). G. triacanthos has formed dense infesta- the sweet pods are relished by grazing tions at several locations in New South animals. Although young plants provide Pest status overseas Wales in areas surrounding the original a valuable source of protein-rich cattle G. triacanthos has been recorded as a pest planting sites. Major infestations exist feed, the leaves of mature trees are either in South Africa, Chile and the United along the Nepean River and its tributaries out of the reach of grazing stock or pro- States (Holm et al. 1979, Wells et al. 1986). and at Schofields. The infestations in New tected by the thorns. It has become naturalized in parts of Eu- South Wales are believed to have origi- In addition to its fodder value, rope (NAS 1980). Even within its natural nated from original plantings by William G. triacanthos is also a popular garden or- range, G. triacanthos can invade pastures Mac-arthur at “Camden Park” in the mid namental in North America and Aus- and old-fields (Burton and Bazzaz 1991, 19th Century. tralia. It grows rapidly, has attractive Schnabel et al. 1991, Schnabel and Sizeable infestations in south east Queensland and New South Wales, which have originated from small num- bers of specimens planted over 50 years ago, provide a valuable insight into the long-term propensity of this thorny tree to become a destructive pest. As sug- gested for other serious weed species such as Mimosa pigra (Lonsdale, in press), there appears to have been a significant “lag phase” between initial introduction and the development of significant infesta- tions of G. triacanthos. In the long-term, therefore, the continued naturalization of G. triacanthos from cultivated specimens in other areas of Australia is likely.

CLIMEX modelling The CLIMEX modelling package devel- oped by the CSIRO Division of Entomol- ogy (Maywald and Sutherst 1989, 1991, Sutherst and Maywald 1985, 1991) was used to predict the potential distribution of G. triacanthos in Australia. The ‘com- Figure 2. Infestation of Gleditsia triacanthos invading improved pasture at pare climates’ module in CLIMEX was Esk, south east Queensland. used to generate a prediction, using a Plant Protection Quarterly Vol.9(3) 1994 103 0.7 threshold for closeness of match. It has Louisville Toowoomba become accepted that a climate match 0 threshold of 0.7 (from a possible range of 0–1) generally provides an indication of a 50 high degree of climatic similarity (Maywald and Sutherst 1991, Scott 1991). 100 The climatic parameters were derived from locations detailed in published 150 sources for the distribution of G. tria- canthos in the United States. The output 200 from the model was interpreted in terms Rainfall (mm) of known distributions of the species in Australia and the United States. 250 Using a threshold of 0.7 and prime lo- cations in the United States produced 300 very few climate matches. Close examina- 40 tion of the model’s output indicated that the range of climatic conditions in Aus- 30 tralia is considerably smaller than in the United States. For example, a comparison 20 between the climate of Louisville (where G. triacanthos is recorded to grow prolifi- cally) and Toowoomba (in south east 10 Queensland) produced a climate match

temperatures index of only 0.54 (Figure 3). 0 The fact that one of Queensland’s worst infestations of G. triacanthos is located Maximum and minimum near Toowoomba (at Clifton on the Dar- -10 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec ling Downs), suggests that a climate match index of 0.5 would be a more ap- Match index = 0.54 propriate threshold for climate suitability assessment in this case. Using a 0.5 Figure 3. Climate comparison between Louisville (USA) and Toowoomba threshold for closeness of match, a second (Queensland, Australia). The dashed line for maximum and minimum prediction was generated (Figure 4). This temperature is for Toowoomba. prediction suggests that G. triacanthos may have the potential to colonize allu- vial soils over large areas of south-eastern Australia. The ‘match index’ process used in CLIMEX to identify sites with a similar climate employs a least-squares technique as an indication of the combined close- ness of match of maximum temperature, minimum temperature, rainfall and evaporation (Maywald and Sutherst 1989, 1991). This index is insensitive to the di- rection of the difference between the cli- matic parameters being compared. A cli- mate which is ‘milder’ in terms of the range of any of the four climate param- eters may show up as having a poor match compared with the climate from which the species has originated. Despite a match index of 0.54, the ‘milder’ climates at locations such as Toowoomba may be suitable for the growth of G. triacanthos, due perhaps to the lack of climatic extremes in combina- tion with the reduced effects of limiting 0 km 1000 factors such as natural enemies or interspecific competition. The insensitivity of the climate match index to the direction of the difference for each climate parameter calls into question its usefulness as a sifting mechanism for Figure 4. Similarity of climates between selected weather stations in the indicating locations which are climati- USA (Dallas, Louisville, St. Louis and Wichita Falls), and climate stations cally suitable for a particular plant spe- in Australia with a climate match index threshold of 0.5. The size of the cies. An index which is sensitive to the shaded circles indicates the relative size of the climate match index. direction of the difference between each 104 Plant Protection Quarterly Vol.9(3) 1994 of the climatic parameters of the target complicate and reduce the effectiveness of temperature and moisture gradients and origin site would be preferable for future biological control programs in a and in patches of old-field vegetation. predicting the climatic suitability of a site. similar manner to that which has been American Journal of Botany 78, 131-49. The problem with the climate match in- experienced with other pest plants in Crowe, F., Starkey, D. and Lengkeek, V. dex underlines the fact that models serve Australia such as Lantana camara (Willson (1982). Honey locust canker in Kansas only as a guide and are no replacement 1993) and Prosopis species. caused by Thyronectria austro-americana. for a clear understanding of the real proc- Plant Disease 66, 155-8. esses being modelled. Summary Dice, L.R. (1945). Some winter foods of The observation that G. triacanthos has In view of the plant’s climatic require- the cottontail in southern Michigan. formed dense infestations in south east ments, broad natural range and prolific Journal of Mammalogy 26, 87-8. Queensland and on the central coast of growth on alluvial soils, Gleditsia tria- Dirr, M.A. (1983). ‘Manual of woody New South Wales tends to support the canthos is considered a threat to native landscape plants: their identification, validity of the prediction generated by riparian vegetation and pastures along ornamental characteristics, culture, CLIMEX. In addition, as predicted by the watercourses and on flood plains propagation and uses’. 2nd edition. model, G. triacanthos has failed to natural- throughout much of south-eastern Aus- (Stipes, Champaign, Illinois). ize in Western Australia, the Northern tralia. Pastures established on alluvial Friedman, S.T. and Adams, W.T. (1985). Territory or South Australia (P. Pheloung, soils associated with the Murray-Darling Estimation of gene flow into two seed personal communication, State Herbaria river system may be particularly prone to orchards of loblolly pine (Pinus taeda L.). records). The plant has not, however, invasion since pods floating down the Theoretical Applied Genetics 69, 609-15. naturalized in Victoria or Tasmania, as Condamine River (from Queensland into Fowells, H.A. (1965). Silvics of Forest could be expected from the prediction. A New South Wales) will find ideal condi- Trees of the United States. Agricultural more accurate assessment of the validity tions for establishment. Handbook Number 271. US Depart- of the prediction can only be made over G. triacanthos possesses similar at- ment of Agriculture, Washington DC. the next 50–100 years, when G. triacanthos tributes to Acacia nilotica and the Prosopis Gordon, D. (1966). A revision of the ge- has had sufficient time to become more complex, some of Australia’s most de- nus Gleditsia (Leguminosae). Ph.D The- firmly established in Australia. structive introduced fodder trees (i.e. sis, Indiana University, Champaign- rapid growth, thorniness, prolific seed Urbana. Risks posed by cultivated varieties production, animal and water dispersal, Harju, A. and Muona, O. (1989). Back- of G. triacanthos and other species freedom from pests and propensity to ground pollination in Pinus sylvestris of Gleditsia form impenetrable thickets along water- seed orchards. Scandinavian Journal of There are 14 species of Gleditsia with a courses and out onto pastures). Unlike Forest Research 4, 513-20. broad range of natural distribution from the situations for the latter species, how- Hart, A.J. (undated). Honey locust Asia to North and South America ever, infestations of G. triacanthos in Aus- (Gleditsia triacanthos). Information (Gordon 1966). Other aggressive, thorny, tralia are still relatively localized. There is Sheet Number 48, Western Australian rapidly growing species of Gleditsia may an opportunity to control small popu- Department of Conservation and Land have significant pest potential in Aus- lations of G. triacanthos before they be- Management. tralia. Two species of particular concern come firmly established. An eradication Herms, D.A., Nielsen, D.G. and Davis are G. sinensis which is reported to be a program for G. triacanthos was initiated in Syndor, T. (1987). Impact of honey- “rapidly growing tree, thriving along Queensland in March 1993. Restrictions locust plant bug (Heteroptera: ) stream courses, upland glens and moun- on the continued sale of G. triacanthos and on ornamental honeylocust and associ- tain slopes of the temperate zone” (Paclt the importation of closely related Gleditsia ated adult buprestids. Environmental 1982) and G. caspica which can form thick species are also recommended. Entomology 16, 996-1000. impenetrable groves, as exemplified in Holm, L.G., Pancho, J.V., Herberger, J.P. forests in the region of Astara in Talys Acknowledgments and Plucknett, D.L. (1979). ‘A Geo- (Talish, USSR) (Safarov 1960, 1965, cited We thank Paul Pheloung for his assist- graphical Atlas of World Weeds’. in Paclt 1982). Additional Gleditsia species ance with climate modelling. We also (Krieger Publishing Co., Florida). should not be imported or sold until a thank Dane Panetta and Rachel Mc- Little, E.L. (1971). ‘Atlas of US Trees’. Vol- thorough review of the pest potential of Fadyen for their comments on the manu- ume 1. United States Department of the entire genus is complete. script. Agriculture Forest Service, Miscellane- The potential for G. triacanthos to hy- ous Publication Number 1146. (United bridize with other species of Gleditsia is References States Government Printing Office, also of concern. Artificial hybridization Bastian, R.A. and Hart, E.R. (1990). Washington DC). between male G. triacanthos (‘Moraine’ Honeylocust clonal effects on develop- Lonsdale, W.M. (1994). Rates of spread of and ‘Skyline’ varieties) and G. melana- mental biology of Mimosa webworm an invading species: Mimosa pigra in the cantha Tang and Wang (Syn. G. caspica) (Lepidoptera: Plutellidae). Journal of Northern Territory. Journal of Ecology has been documented by Santamour Economic Entomology 83, 533-8. (in press). (1976). The hybrid seed germinated and Bedker, P.J. and Blanchette, R.A. (1983). Mathwig, J.E. (1971). Relationships be- grew normally. In addition, G. texana is Development of cankers caused by Nec- tween bruchid beetles (Amblycerus believed to be a naturally occurring hy- tria cinnabarina on honey locusts after robinae) and honey locust trees brid between G. aquatica Marsh. (water root pruning. Plant Disease 67, 1010-13. (Gleditsia triacanthos). Ph.D Thesis, Uni- locust, native to the southern USA) and Bradley, R.J. (1978). ‘Encyclopedia of versity of Kansas, Lawrence. G. triacanthos (Schnabel and Hamrick Australian Gardening’. (Bay Books, Maywald, G.F. and Sutherst, R.W. (1989). 1990b). Hybridization between Gleditsia Sydney). CLIMEX: a computer program for com- species and between varieties of G. Bugbee, R.E. and Riegel, A. (1945). Sea- paring climates in ecology: recent de- triacanthos may lead to increased morpho- sonal food choices of the in velopments. Proceedings of the SSAI- logical variability and an ever-increasing western Kansas. Kansas Academy of Sci- MACS 1989 Biennial Conference on variation of genotypic combinations ence Transcripts 48, 199-203. Modelling and Simulation, pp. 134-40. in the hybrid off-spring. Resultant mor- Burton, P.J. and Bazzaz, F.A. (1991). Tree Maywald, G.F. and Sutherst, R.W. (1991). phological and genetic variability may seedling emergence on interactive User’s Guide to CLIMEX, a computer Plant Protection Quarterly Vol.9(3) 1994 105 program for comparing climates in within and among populations of ecology. 2nd Edition. CSIRO Australian Gleditsia triacanthos (Leguminosae). Division of Entomology , Report No. American Journal of Botany 77, 1060-69. 48, pp. 1-35. Schnabel, A. and Hamrick, J.L. (1990b). Mitchell, D. (1978). Notes compiled by the Non-random associations between sex Nelson branch of the New Zealand Tree and 6-phosphogluconate dehydrogen- Crops Association, New Zealand Journal ase isozyme genotypes in Gleditsia of Agriculture November 1978, 21-2. triacanthos L. The Journal of Heredity 81, Nagasaka, K. and Szmidt, A.E. (1985). 230-3. Multilocus analysis of external pollen Schnabel, A., Laushman, R.H. and contamination of a Scots pine (Pinus Hamrick, J.L. (1991). Comparative ge- sylvestris L.) seed orchard. In ‘Popula- netic structure of two co-occurring tree tion Genetics in Forestry’, ed. H.R. species, Maclura pomifera (Moraceae) Gregorius, pp. 134-8. (Springer-Verlag, and Gleditsia triacanthos (Legumin- Berlin). oseae). Heredity 67, 357-64. NAS (1980). ‘Firewood Crops: Shrub and Scott, J. (1991). Acacia karroo Hayne Tree Species for Energy Production’. (Mimosaceae), a potentially serious Volume 1, pp. 36-7. (National Academy weed in Australia. Plant Protection of Sciences, Washington, DC.). Quarterly 6, 16-8. North, R.C. and Hart, E.R. (1983). Skerman, P.J. (1977). ‘Tropical Forage Oviposition preference of the mimosa ’, p. 507. (FAO United Na- webworm, tions, Rome). (Lepidoptera: Plutellidae). Environmen- Smith, J.L. and Perino, J.V. (1981). Osage tal Entomology 12, 546-51. orange (Maclura pomifera): History and Paclt, J. (1982). On the repeatedly con- economic uses. Economic Botany 35, 24- fused nomenclature of Chinese species 41. of Gleditsia (Caesalpiniaceae). Taxon 31, Sutherst, R.W. and Maywald, G.F. (1985). 551-3. A computerized system for matching Peattie, D.C. (1953). ‘A Natural History of climates in ecology. Agriculture, Ecosys- Western Trees’. (Houghton Mifflin, tems and Environment 13, 281-99. Boston, Massachusetts). Sutherst, R.W. and Maywald, G.F. (1991). Rowell, R.J. (1991). ‘Ornamental Flower- Climate-matching for quarantine, using ing Trees in Australia’, pp. 155-6 (NSW CLIMEX. Plant Protection Quarterly 6, University Press). 3-7. Santamour, F.S. (1976). Metaxenia in Wells, M.J., Balsinhas, A.A., Joffe, H., interspecific honeylocust crosses. The Engelbrecht, V.M., Harding, G. and Journal of Heredity 67, 185-6. Stirton, C.H. (1986). A catalogue of Sargent, C.S. (1965). ‘Manual of the Trees problem plants in southern Africa. of North America (Exclusive of Memoirs of the Botanical Survey of South Mexico)’. (2 Volumes), pp. 608-9, (Do- Africa No. 53. ver Publications Inc, New York). Willson, B.W. (1993). A renewed attempt Schnabel, A. (1988). Genetic structure and at the biological control of Lantana gene flow in Gleditsia triacanthos L. camara. Proceedings of the 10th Austral- Ph.D Thesis. University of Kansas, ian Weeds Conference and 14th Asian Lawrence. Pacific Weed Science Society Confer- Schnabel, A. and Hamrick, J.L. (1990a). ence, Brisbane, Australia, Weed Society Organization of genetic diversity of Queensland, Brisbane.