Final report: Project CSE 1 Identification of species and the origin of found in areas of dried vine fruit production

Prepared for The Australian Dried Fruits Research and Development Council

Prepared by Dr J. K. Scott and Ms. S. M. Morrison CSIRO Division of Entomology Western Australia 31 August 1994

a DEPARTMENT OF AGRICULTURE C S I RO WESTERN AUSTRALIA AUSTRALIA

CSE 1 IDENTIFYING SPECIES AND ORIGIN OF TRIBULUS FOUND IN AREAS OF DRIED VINE FRUIT PRODUCTION

Organisation: CSIRO Division of Entomology Location: South Perth Department of Agriculture, W.A.

Supervisor: Dr J.K. Scott Time Span: July 1991 to June 1994

Objective: To identify, and determine the origin of Tribulus species (caltrop) that occur in areas of dried vine fruit production as a pre-requisite for identifying suitable control measures.

Progress: 1. Burrs of s./. were obtained from 54 collections throughout the world distribution of this weed.

2. Morphological analysis of the burrs indicated that the Queensland and Northern Territory collections form a separate group from collections in southern Australia, and that a third group exists in northern Western Australia. The combination of height and length of burrs was most useful to separate the groups and is a suitable quick technique for the preliminary identification of major taxonomic groups. Morphology however, did not reflect all of the variation detected by cytology and isozyme analysis.

3. Chromosome counts of 2n = 24, 36 and 48 were detected in root tips of germinated seed. This polyploid series appears to have an autopolyploid origin. The cytogenetic studies showed that the Queensland and Northern Territory collections are different from all other Australian collections, except possibly two southern collections. The southern Australian and northern Western Australian collections are similar to those from the USA, southern Africa, Mediterranean and the Middle East, indicating that they are introduced. A separate, unrelated group occurs in north west and parts of the Middle East.

4. Isozyme analysis of seedlings indicated that the Queensland and Northern Territory collections are very similar to each other, are native to Australia and are possibly a separate species from T. terrestris. Populations in southern Australia which are a problem to agriculture, are genetically similar to populations from the USA, southern Africa, the Mediterranean and the ¥iddle East. It is likely that at least two separate introductions into Australia have occurred and that these populations originated from the Mediterranean, Middle East or southern Africa, but not from north west India.

5. Biological control methods have been effective against T. terrestris, but the few known agents are likely to attack native Australian Tribulus species. Most native Tribulus are found in the central and north western Australia and an option for the control of introduced T. terrestris would be to use biological control agents that are restricted to Mediterranean climates. However parts of the world that have this climate and that are similar to dried vine fruit production areas, have not been surveyed for biological control agents.

6. In summary, the techniques of morphological, cytogenetic and isozyme analysis are useful tools to determine the identity and origins of a poorly-described species. The results of this study show that the form of Tribulus terrestris which is a problem to the Australian Dried Fruit Industry is introduced into Australia and has originated from the Mediterranean, Middle East or southern Africa. The option of biological control is still open, but requires the identification of suitable agents restricted to areas of similar climate to Australian dried vine fruit production.

Final report: Project CSE 1 Identification of species and the origin of Tribulus found in areas of dried vine fruit production

Contents Page

1. Background to the project. 1

2. Variation in Australian and world-wide populations of Tribulus 2 terrestris L. 1. Burr morphology.

3. Variation in Australian and world-wide populations of Tribulus 23 terrestris L. 2. Chromosome numbers.

4. Variation in Australian and world-wide populations of Tribulus 33 terrestris L. 3. Isozyme analysis.

5. Prospects for biological control of introduced Tribulus terrestris L. 52 in Australia.

Appendices

1. The origin of caltrop (Tribulus terrestris L.) in Australia and the likelihood of biological control. (1994). Proceedings of the Fourth Biennial Proclaimed and State Conference.

2. Assessment of the origins of Tribulus terrestris in Australia. (1993). Proceedings ofthe lOth Australian and 14th Asian-Pacific Weed Conference.

3. Feedback articles; number 70, March 1992; number 72, July 1992.

4. Weed Update articles; number 10, November 1991; number 11, March 1992.

5. Collecting instructions for caltrop.

6. List of caltrop collectors.

This report contains unpublished material and is not to be quoted or referred to without permission of the authors 1

Background to the project

At a workshop held in Mildura, Victoria, 13-14 August 1990, four spiny weeds were discussed, three cornered jack and lesser jack, Emex spp., caltrop, Tribulus terrestris and spiny burr grass, Cenchrus longispinus. Seeds of these are important contaminants of dried vine fruits in Australia and the workshop considered short and long term strategies for their control (Johnstone 1990). The reports of the working groups from the meeting (Anon. 1990) recognised that an issue critical to the development of long term strategies for the control of caltrop, such as biological control, is the question of identification and origins of caltrop found in Australia.

A co-operative project between the CSIRO Division ofEntomology and the Western Australian Department of Agriculture was established to investigate the identification and origins ofT. terrestris. A grant was obtained from the Australian Dried Fruits Research and Development Council that enabled Ms S. Morrison to join the project. The research was based at the Department of Agriculture Western Australia, South Perth and involved a large number of volunteer collectors from Australian and world­ wide locations, who supplied the burrs used in the study. The project was carried out on a half-time basis, extended over three years, to allow sufficient time to obtain the material. Permission was received from the Australian quarantine and inspection service for the importation of the seeds, which have been stored under quarantine. At the end of the project, voucher samples were lodged in herbaria. All extra-Australian samples were either sterilized by gamma radiation in the case of voucher samples or incinerated so as to comply with Australian quarantine regulations.

The project encompassed three approaches to identifying T. terrestris in Australia, using morphological, cytogenetic and isozyme techniques. Morphological characteristics of the spiny burrs were measured from 31 world-wide populations of T. terrestris. Seeds from burrs were germinated, 48 collections examined for chromosome counts and 54 collections were used in isozyme analysis.

The report that follows comprises the manuscripts resulting from the study. In the appendices are copies of publications and other material, including in the final appendix, a list of collectors to whom we are most grateful for their time and interest. We also thank the Department of Agriculture Western Australian, Weed Science Branch and especially Drs F. D. Panetta, R. J. Martin and R. Cousins for their continued support. We also thank the Australian Dried Fruits Research and Development Council for making this investigation possible.

References

Anon. ( 1990). Reports of working groups. Plant Protection Quarterly 5, 129-131 . Johnstone, R. B. (1990). Opening address: the problems caused by weed seeds to the dried vine fruits industry. Plant Protection Quarterly 5, 84. Morphology 2

Variation in Australian and World-Wide Populations ofT. terrestris L. 1. Burr Morphology

J K. Scott and S. M Morrison

CSIRO Division of Entomology, Private Bag, P.O. Wembley, W.A. 6014, Australia.

Abstract Measurements were made of the morphology of Tribulus terrestris s.l. burrs from 31 Australian and world-wide locations. The variables included four size variables, four spine angles and the number of seeds in each burr. Cluster analysis using all variables identified four groups ofburrs, but was influenced by a single site that had spines angled differently from other sites. On some plants the basal spine was absent, so its measurements were excluded from further analyses, Four clusters were indicated on re-analysis, likewise with the removal of other variables, until only the length and height ofburrs and abaxial spine length remained. Southern Australian sites were grouped with southern African, Indian and Israeli sites. Three Western Australian sites formed one group, as did Northern Territory and Queensland sites. A fourth group included one Australian, Iran, Israel and USA site. The relationship between length and height distinguished the northern Australian collections that either have taller burrs or more elongate burrs, from the introduced T. terrestris in southern Australia. Analysis of the morphology showed that it is possible to separate northern Australian burrs from other collections and indicated that southern Australian populations are likely to be introduced. However the morphology only approximately detected groupings within the southern Australian and overseas collections that were identified by isozyme and cytogenetic studies.

Introduction Tribulus terrestris s.l. L. () is widespread in Mediterranean, subtropical and desert climates world-wide. It is listed as a weed in 37 countries and affects at least 21 crops (Parsons and Cuthbertson 1992). Weedy populations ofT. terrestris are widespread in Australia (Squires.I979, Parsons and Cuthbertson 1992), and the plant is thought to have been accidentally introduced before 1895 (Squires 1979, Bourke 1987b). Tribulus terrestris is a major problem to the Australian dried vine fruit industry because the spiny burrs contaminate harvested dried fruit (Johnstone 1990, Parsons and Cuthbertson 1992). The plant is most notable for being toxic to livestock, especially sheep (Bourke 1984, 1987a, 1987b, Jacob and Peet 1987, Bourke eta/. 1992). It also invades some agricultural crops (Young 1988, Charles 1991, Parsons and Cuthbertson 1992). Tribulus terrestris is a prostrate, summer growing annual. Stems grow up to three metres long, radiating out from a central woody taproot. The flower consists of five yellow petals 5-6 mm long. The seeds germinate rapidly after late spring and summer rains in warm conditions (24 - 27°C). Seeds can be produced as soon as five weeks after germination (Parsons and Cuthbertson 1992). The fruit consists of a woody burr in five wedge-shaped schizocarps (cocci), each with one or two pairs of large spines (Nabil el Hadidi 1978). There are two to four seeds per schizocarp (Squires 1979). Under optimum conditions a single plant can produce up to 400 burrs per square metre (Squires 1969) and several germinations can occur in one season. The seeds can remain viable for many years if buried in the soil, thus after successive generations a Morphology 3

large reservoir of seeds can accumulate (Squires 1979, Parsons and Cuthbertson 1992). The plant evidently has the potential to spread rapidly and persist for years. Control by chemical or mechanical means is lengthy and expensive because of the persistent nature of the seeds and multiple germinations. Biological control has been proposed as an alternative control method, especially as it has been successful in the USA, using seed and stem feeding weevils, Microlarinus spp. (Maddox 1976, Huffaker eta/. 1983). It is a recognised part of a biological control program that the of the target weed and its origins are understood to ensure the correct matching ofbiocontrol agent with target species. It is therefore essential to clarify the taxonomy ofT. terrestris in Australia before biological control measures can be implemented (Bruzzese 1990, Scott 1990, Shepherd 1990). The taxonomic status ofT. terrestris is unclear (Bourke 1987b, Scott 1990, Parsons and Cuthbertson 1992, Wilson 1992), largely because the morphology of the T. terrestris complex is extremely variable (Schweickerdt 1939, Malik 1966, Squires 1969, Kumar 1978, Bhansali 1980, Hilu 1981, Husain 1986, Bourke 1987b). In Australia it has been suggested that there are at least two forms of T. terrestris, a native and an introduced weedy form (Meadly 1954, Squires 1969, Bourke 1987b, Wilson 1992). Morphology is commonly used to define identifying characteristics of plant species and it forms the basis of much of the world's plant taxonomy. Cryptic species or those poorly defined in descriptive taxonomy can often be identified following detailed study of morphological characteristics, particularly if the measurements are combined with other means of species identification such as genetic variability and chromosome counts. Isozyme and cytogenetic studies are two methods being used to clarify the identity and to indicate the origins ofT. terrestris introduced into Australia (Morrison and Scott 1994a, b). Part of this study includes grouping collections ofT. terrestris burrs according to the similarity of their morphology measurements. The groups formed were compared with those from the isozyme and cytogenetic analyses for the same collections, with the intention of identifying native and introduced Australian T. terrestris and providing evidence of their origins.

Methods Collections Burrs ofT. terrestris sensu lato were obtained by writing to potential collectors in Australian and world-wide locations. Material was received from North America, the Mediterranean, the Middle East, India, southern Africa and mainland Australia. The supplier's identification of the burrs as belonging to the taxon, T. terrestris, was accepted to enable the study to encompass as broad a range of material as possible. However, burrs ofT. micrococcus Domin, T. occidenta/is R. Br. and T. cistoides L. were also received from Australian locations and these were excluded. The latter three species could be distinguished from T. terrestris with the descriptions in Harden (1992), Wilson (1992) and Rye et al. (1992) respectively. Vouchers specimens are held at the Division ofEntomology in Perth and will be lodged at either the National Herbarium ofNew South Wales, Sydney (NSW) or the Western Australian Herbarium (PERTH). Burrs were chosen for the morphology study because they are persistent, easily collected, transported and stored. The morphology of burrs is one of the most important characters of systematic value in the Tribulus (Schweickerdt 1939, Nabil el Hadidi 1978). Thirty burrs were randomly selected from each of31 Morphology 4

collections ofT. terrestris (Table 1). Two or three sites were chosen from each part of the world where the plant is introduced (USA) and where the plant is possibly native (Australia, southern Africa and West Asia).

Measurements The burrs were dried at room temperature and stored in paper bags for at least a year before measurement. Seven measurements were taken on the right side of each coccus as illustrated in Fig. 1. Length and height measurements (excluding the spines) were taken with callipers (Helios). The burrs have one or two pairs of spines, one on the abaxial surface of the burr spreading at right angles to the burr (the ''abaxial" spine) and one pair near the base of the burr, spreading downwards towards the stem (the "basal" spine) (Wilson 1992). Inside length measurements were taken on basal and abaxial spines. Angles were measured by positioning the base (surface of attachment in the earlier schizocarp stage) at zero degrees and marking the angle of the abaxial and basal spines on Polar graph paper (Gormack Graph Papers, 10 em radius). The angle between each of the abaxial and each of the basal spines was measured by positioning one spine at zero and measuring the angle to the other spine.

Number ofseeds Burrs were attached on their right side by double-sided tape, to plastic sheets placed on a cassette containing Kodak X-ray film and exposed to X-rays at 35 Kilovolts for 16 milliamp seconds. The number of seeds in each coccus was counted by examination of the developed X-ray film and the number confirmed by dissection of 20 burrs. The arrangement of the seeds inside the burr is illustrated in Fig. 1.

Statistical analysis The mean of each measurement was calculated for each collection and then used in cluster analyses to determine the number and distance between groups (SAS 1988, James and McCulloch 1990). The data were analysed at two levels. First, all variables were included, standardised to a mean of 0 and variance of 1. Secondly, only measurements on a similar scale (abaxial spine length, length ofbase and height) were incorporated into the cluster analysis without standardisation. Average Linkage was chosen as the method of analysis, and other cluster methods (SAS 1988) gave essentially similar results. The number of clusters was determined from the cubic clustering criterion and the Pseudo F (SAS 1988). A canonical descriminant analysis was determined using the number of clusters previously identified.

Results The means for each collection are given in Table 2 and the frequency distribution of each variable are given in Fig. 2. As would be expected, there is a high degree of correlation between many of the measurements of size (Table 3). The average linkage cluster analysis using all measurements, but excluding four sites with missing values (Carnarvon, Nita Downs Station, Broome and S. Pemiscot) and Prosser where only 7 burrs had all measurements, indicated that there were four clusters. Fig. 3 shows clusters in the form of a dendrogram. The A group included samples from southern Australia, Namibia and . Group B consisted of all the Northern Territory, Queensland, Indian and Kuwait samples and one sample from Namibia. Group C comprised one site in New South Wales, two in Iran and one in Israel. The D group comprised one site, Shefayim in Israel. Morphology 5

The distance between the groups is shown in Fig. 4. The first canonical variate represents size (Table 4) and shows that group B sites have smaller basal spines and more seeds for their size. The second canonical variate separates burrs that are more elongate (length ofbase by height) (group C) from the others. The third canonical variate shows the influence of the single outlying Israel population (Shefayim) which has abaxial spines with a comparatively acute angle to the base and parallel basal spines (Table 2). The basal spine was absent in all burrs from Carnarvon, Nita Downs Station, Broome and S. Pemiscot and in many burrs from Mildura I, Pulwama, Prosser and Ashqelon (Table 2). Consequently the angle of the basal spine and the angle between the basal spines were also absent. The frequency distribution for the length and angle of the basal spine appeared multi nodal due to collections from Darwin, Mil dura I, Walpeup, Taldra, Goomalling, Moorine Rock, Shefayim, Gobabeb 2 and Rondebosch, having some burrs with spines pointing above the base line at about 40° and some with spines pointing at or below zero degrees (Table 2, Fig. 2). To investigate this further, an additional sample of burrs was taken from five individual plants within the highly variable collection ofT. terrestris at the Sunraysia Research Station, Mildura (S.E. = 0.24 mm, Table 2). Isozyme and chromosome analysis (Morrison and Scott I994a, b) indicated that the Mildura collection is from a single population. These burrs had a range of size of basal spine indicating that basal spine length can be very variable within a population (Fig. 5). Because of this variation, basal spine measurements were excluded from further analyses. The cluster analysis using standardised data was recalculated successively by removing the basal spine measurements, all angle measurements and the seed counts. Similar dendrograms were produced to that incorporating only the remaining three variables, length and height of the burr and the abaxial spine length. The dendrogram is shown in Fig. 6. The sites group into four clusters. Group A came from Western Australia. Other southern Australian sites were placed in group B with sites from India, Israel, southern Africa and USA. Group C comprised Northern Territory and Queensland sites and two sites from Kuwait. Group D were from widespread localities. The distance between groups is shown in Fig. 7. The first canonical variate represents size, where group A sites have small burrs and group D sites are larger than other sites. The second canonical variate represents a separation based on the length ofbase and the height. Group C sites are more elongate than the others. The third canonical variate was not significant (Table 5) and does not aid the separation of groups as shown by the plot ofvariable 3 against variable 2 in Fig. 7. The relationship between length and height is shown in Fig. 8 for one Australian site from each of the four groups identified in the dendrogram (Fig. 6). The range in shape of the burrs is from those with a short base length but high burrs of Carnarvon to the long base length and short burrs ofDarwin. Burrs from Attunga are larger than the others and the introduced form represented by the Mildura I collection has an intermediate burr form.

Discussion The cluster analysis shows that the collections of burrs have heterogeneous morphology and that it is unlikely that all collections belong to the same species. Four groups were identified based on the measurements of base length, height and abaxial spine length. Three Western Australian sites form a separate group that was not detected in the isozyme and cytogenetic studies (Morrison and Scott I994a, b). A few Morphology 6 plants of this group were grown and the flowers were much larger than weedy populations ofT. terrestris in southern Australia. It is likely that this group will prove to be the Australian endemic, Tribulus ranunculoides F. Muell., once a full range of morphological characters is examined. The second group represents weedy T. terrestris and shows the similarity of morphology between Australian weedy populations and overseas populations from southern Africa, India and Israel, thus providing evidence of the introduced origin of T. terrestris. The Northern Territory and Queensland sites were morphologically different from other Australian sites, forming a separate group with burrs that are more elongate when compared with most southern Australian collections ofT. terrestris. These sites also were clearly separated in the isozyme and cytogenetic studies, having distinctive isozyme bands and a chromosome count of2n = 24, and probably represent a different species from T. terrestris. Samples from Kuwait were included in this group, however these burrs have chromosome counts of2n = 36 and 48, and are separate in the isozyme study. Thus group C in the morphology study probably represents two species rather than indicating an overseas origin of the species found at Northern Territory and Queensland sites. The fourth group includes Attunga in New South Wales which possibly represents a separate introduction into Australia as it classifies with sites in Iran, USA and Israel. This group was not recognised in the isozyme study and all have a chromosome count of2n = 36 (Israel was not counted). However, the burrs used in this study come from a wide range of environments. The large burrs from Attunga, for example, may represent variation due to climate. The variation found throughout the collections of burrs was examined further with studies of chromosome numbers and isozymes (Morrison and Scott a, b).

Acknowledgements We thank Kristy Hollis for technical assistance and the many collectors of burrs. Dr R. Davidson of Queen Elisabeth the Second Hospital kindly helped with the X-rays. We also thank Bert De Boer for help with statistical analyses. This work was partly funded by the Australian Dried Fruit Research and Development Council. J. R . Hosking and P. B. Yeoh commented on early.drafts ofthe manuscript.

References Bhansali, A. K. (1980). Cytological study ofthe morphological variants of Tribulus terrestris. Indian Forester 106, 734-737. Bourke, C. A. (1984). Staggers in sheep associated with the ingestion of Tribulus terrestris. Australian Veterinary Journal61, 360-363. Bourke, C. A. (1987a). A novel nigrostriatal dopaminergic disorder in sheep affected by Tribulus terrestris staggers. Research in Veterinary Science 43, 347-350. Bourke, C. A. (1987b). Some taxonomic, agronomic and animal health aspects of Tribulus. Proceedings of the 8th Australian Weeds Conference, Sydney. pp. 182- 185. Bourke, C. A., Stevens, G. R. and Carrigan, M. J. (1992). Locomotor effects in sheep of alkaloids identified in Australian Tribulus terrestris. Australian Veterinary Journal69, 163-165. Bruzzese, E . (1990). Protocols for biological control of weeds and current Victorian priorities. Plant Protection Quarterly 5, 98-99. Morphology 7

Charles, G.W. (1991). A grower survey ofweeds and herbicide use in the New South Wales cotton industry. Australian Journal ofExperimental Agriculture 31, 387- 392. Gardner, C. A. (1948). Caltrop (Tribulus terrestris Linn.). Western Australian Department ofAgriculture Journal ofAgriculture 25, 120-124. Harden, G. J. (1992). Zygophyllaceae. In 'Flora ofNew South Wales, Volume 3.' (Ed G. J. Harden.) p. 13. (NSW Press: Sydney.) Hilu, K. W. (1981 ). Cytotaxonomical studies in Tribulus terrestris and T. a latus (Zygophyllaceae). Nordic Journal ofBotany l, 531-534. Huffaker, C. B., Hamai, J. and Nowierski, R. M. (1983). Biological control of puncturevine, Tribulus terrestris in California after twenty years of activity of introduced weevils. Entomophaga 28, 387-400. Husain, S. A. (1986). Cytotaxonomic studies in Tribulus from . Kromosoma II 42, 1316-1329. Jacob, R. H. and Peet, R. L. (1987). Poisoning of sheep and goats by Tribulus terrestris (caltrop). Australian VeterinaryJournal64, 288-289. James, F. C. and McCulloch, C. E. (1990). Multivariate analysis in ecology and systematics: panacea or Pandora's box. Annual Review ofEcology and Systematics 21, 129-166. Johnstone, R. B. (1990). Control of Emex, Tribulus, and Cenchrus in vineyards. Plant Protection Quarterly 5, 84. Kumar, A. (1978). Intraspecific polyploidy in Tribulus terrestris L. Science and Culture 44, 427-428. Maddox, D. M. (1976). History ofweevils on puncturevine in and near the United States. Weed Science 24, 414-419. Malik, C. P. (1966). Corrected basic chromosome number and intraspecific polyploidy in Tribulus terrestris Linn. Chromosome Information Service 7, 7-8. Meadly, G. R. W. (1954). Weeds of Western Australia. Caltrop (Tribulus terrestris L.) Journal ofAgriculture of Western Australia 3, 673-675. Morrison S. M . and Scott J. K. (1994a). Variation in Australian and world-wide populations of Tribulus terrestris L. 2 Chromosome numbers. (in preparation). Morrison S. M. and Scott J. K. (1994b). Variation in Australian and world-wide populations of Tribu/us terrestris L. 3 Isozyme analysis. (in preparation). Nabil el Hadidi, M . (1978). An introduction to the classification of Tribulus L. Taeckholmia 9, 59-66. Parsons, W. T. and Cuthbertson, E. G. (1992). 'Noxious weeds of Australia.' (Inkata Press: Melbourne.) Rye, J. R., Koch, B. L. and Wilson, A. J. G. (1992). Zygophyllaceae. In 'Flora ofthe Kimberley Region.' (Ed. J. R. Wheeler) p. 678 (Western Australian Herbarium: Como.) SAS (1988). ' SAS/STAT user's guide, Release 6.03 edition.' (SAS Institute: Cary, North Carolina.) Schweickerdt, H. G. (1939). An account of the South African species of Tribulus Tourn. ex Linn. Bothalia 3, 157-178. Scott, J. K. (1990). Tribulus terrestris L. (Zygophyllaceae) in Southern Africa: An outline ofbiology and potential biological control agents for Australia. Plant Protection Quarterly 5, 103-106. Shepherd, R. C. H. (1990). Past Victorian work on Emex australis Stenheil and Tribulus terrestris L. Plant Protection Quarterly 5, 100-102. Morphology 8

Squires, V. R. (1969). Distribution and polymorphism of Tribulus terrestris sens. lat. in Australia. Victorian Naturalist 86, 328-334. Squires, V. R. (1979). The biology of Australian weeds. 1 Tribulus terrestris L. Journal ofthe Australian Institute ofAgricultural Science 45, 75-82. Wilson, K. L. (1992). A new species and a neotypification in Australian Tribulus (Zygophyllaceae). Telopea 5, 21-29. Young, K. R. (1988). Post emergent control of caltrop (Tribulus terrestris L.) in peanuts. Australian Weeds Research Newsletter 37, 36-39. Morphology 9

Table 1. Collection sites of Tribulus te"estris s.L burrs

Location Latitude; Longitude AUSTRALIA New South Wales Attunga 31 o 54'S; 150° 50'E Coonamble 30° 57'S; 148° 24'E Northern Territory Darwin, Casuarina Beach 12° 21'S; 130° 52'E Katherine 1 14° 32'S; 132° 24'E Queensland Cranbrook 19° 16'S; 146° 49'E Home Hill 19° 52'S; 147° 15'E South Australia Taldra 34° 21'S; 140° 51'E Wunkar 34° 29'S; 140° 18'E Victoria Mildura1 34° ll'S; 142° IO'E Rutherglen 36° 01'S; 146° 25'E Walpeup 35° 08'S; 142° 02'E Western Australia, North Broome 17° 58'S; 122° 14'E Carnarvon 24° 53'S; 113° 40'E Nita Downs Station l9° 05'S; l21° 41'E Western Australia, South Busselton 33° 39'S; 115° 20'E Goomailing 31° 18'S; 116° 50'E Moorine Rock 31° 19'S; 119° 04'E WORLD-WIDE India, Ladakh, Nubra River 34° 40'N; 77° 40'E* India, Kashmir, Pulwama 33° 55'N; 74° 55'E* Iran, Gaoh Sar 1 36° IO'N; 51° 25'E* Iran, Karaj 2 35° 48'N; 50° 58'E Israel, Ashqelon 31 o 40'N; 34° 33'E Israel, Shefayim 32° 12'N; 34° 50'E Kuwait, AI Qurain 29° 02'N; 48° 08'E* Kuwait, city 29° 20'N; 48° OO'E* N arnibia, Gobabeb 1 23° 34'S; 15° 03'E Namibia, Gobabeb 2 23 ° 34'S; 15° 03'E Namibia, Tiroll 26° IS'S; 16° 45'E* South Afiica, Rondebosch 33° 56'S; 18° 28'E USA, Prosser 46° 13'N; 119° 46'W USA, S. Pemiscot 36° 05'N; 89° 50'W* * approximate location only Morphology 10

Table 2. Mean± SE and sample size of measurements made on burrs ofT. te"estris s.L

Site Basal spine Basal spine Abaxial Abaxial Base length Height Angle of Angle of No. of length (mm) angle to spine length spine angle (mm) (mm) basal spines abaxial seeds base {0}. {mm} to base {0} {0} SEines {0} NSW, Attunga 4.1 ±0.14 25.0± 1.22 5.7±0.19 78.7 ± 0.91 7.1±0.10 5.1 ± 0.12 47.0 ± 1.65 90.6 ± 1.91 3.0 ± 0.06 30 30 30 30 30 30 30 30 30 NSW, Coonamble 3.4±0.11 26.9 ± 1.01 4.9 ± 0.11 83.4 ± 0.90 5.2 ± 0.10 5.0 ± 0.09 55.2±2.39 107.6± 1.69 3.0±0.10 30 30 30 30 30 30 30 30 30 NT, Darwin 2.1±0.16 26.2±3.67 4.7±0.18 90.6 ± 1.00 6.3 ± 0.16 4.0 ± 0.09 54.1±2.19 102.4±2.15 4.4±0.13 29 29 30 30 30 30 29 30 30 NT, Katherine 1 2.5 ± 0.11 37.7 ± 0.88 4.6 ± 0.10 96.2 ± 0.95 6.9 ± 0.08 3.7 ± 0.05 53 .5 ± 1.77 94.2 ± 2.09 4.6 ± 0.09 30 30 30 30 30 30 30 30 30 QLD, Cranbrook 1.4±0.11 38.6 ± 1.06 3.6 ± 0.12 91.9 ± 1.50 6.0 ± 0.11 3.8±0.07 60.3±1.35 103.1±1.62 3.7±0.11 29 29 30 30 30 30 29 30 30 QLD, Home Hill 2.2 ± 0.11 39.5 ± 0.85 4.1 ± 0.11 90.3 ± 0.98 5.4 ± 0.08 3.7 ± 0.07 61.3 ± 1.74 94.8 ± 2.65 2.5 ± 0.09 30 30 30 30 30 30 30 30 30 SA, Taldra 2.6±0.15 30.3 ± 1.99 5.0 ± 0.12 89.3 ± 1.24 5.8 ± 0.12 4.7±0.12 61.2±3.25 100.3±1.97 2.7±0.09 30 30 30 30 30 30 30 30 30 SA, Wunkar 2.7±0.17 27.0±1.42 5.5±0.18 83 .9 ± 1.85 5.7±0.19 4.6±0.10 40.9±2.45 94.1 ± 2.62 2.6 ± 0.1 2 29 29 30 30 30 30 29 30 30 VIC, Mildura1 2.1 ± 0.24 10.9 ± 3.67 5.0 ± 0.15 81.3±1.95 4.7±0.14 4.2± 0.11 63 .6 ± 4.59 107.9 ± 3.22 2.1 ± 0.10 14 14 30 30 30 30 14 30 30 VIC, Rutherglen 2.7±0.10 27.3±1.30 4.8±0.10 80.5 ± 0.98 5.6 ± 0.10 4.9±0.10 42.0±2.29 110.6±2.62 2.7±0.10 30 30 30 30 30 30 30 30 30 VIC, Walpeup 2.7±0.15 8.3 ± 3.44 4.4 ± 0.20 89.8 ± 2.06 5.8 ± 0.14 5.1±0.15 44.9 ± 2.03 96.6 ± 2.36 2. 8 ± 0.12 30 30 29 29 30 30 30 29 30 W A, Busselton 3.5±0.10 33.6 ± 1.21 4.7 ± 0.11 91.2 ± 1.07 4.4 ± 0.07 4.6 ± 0.10 63 .0 ± 2.27 102.2 ± 2.01 2.2 ± 0.07 30 30 30 30 30 30 30 30 30 WA, Broome 4.6 ± 0.09 86.9 ± 1.29 5.1 ± 0.07 4.9 ± 0.06 117.0 ± 2.22 2.6 ± 0.09 0 0 30 30 30 30 0 30 30 Morphology 11

Table 2. Continued. Site Basal spine Basal spine Abaxial Abaxial Base length Height Angle of Angle of No. of length (mm) angle to spine length spine angle (mm) (mm) basal spines abaxial seeds base {0} {mm} to base {0} {0} seines {0} W A, Carnarvon 4.1±0.18 87.5 ± 1.97 4.5 ± 0.11 4.4±0.16 110.6 ± 3.56 2.4 ± 0.10 0 0 30 30 30 30 0 30 30 W A, Goomalling 3.3±0.12 19.3±3.79 3.8±0.17 91.8±1.48 5.2±0.10 5.2±0.11 56.1 ± 2.55 107.1 ± 3.84 2.6 ± 0.09 30 30 23 23 30 30 30 23 30 WA, Moorine Rock 2.2±0.14 -2.1 ±4.72 3.7±0.18 83 .3 ± 1.98 4.2 ± 0.10 4.1±0.13 42.5 ± 2.57 100.2 ± 3.44 2.5 ± 0.1 0 30 30 30 30 30 30 30 30 30 W A, Nita Downs Stn 3.0 ± 0.10 93.1 ± 1.36 4.5 ± 0.08 3.9 ± 0.08 108.1 ± 1.90 2.3 ± 0.08 0 0 30 30 30 30 0 30 30 India, Nubra R. 1.4±0.10 34.6 ± 1.12 4.2 ± 0.15 83 .7 ± 1.07 5.5 ± 0.07 4.4 ± 0.08 50.4 ± 2.15 93.3 ± 1.50 3.7 ± 0.09 30 30 30 30 30 30 30 30 30 India, Pulwama 1.1 ±0.05 34.9 ± 1.07 4.3 ± 0.13 85.4 ± 0.80 5.7 ± 0.10 4.7 ± 0.08 55.7 ± 1.74 93.4 ± 0.90 3.7 ± 0.09 23 23 30 30 30 30 23 30 30 Iran, Gaoh Sar 1 3.6 ± 0.12 36.8 ± 1.30 6.2 ± 0.13 92.9 ± 1.38 7.2 ± 0.12 5.5 ±0.11 62.5 ± 1.79 100.0 ± 1.85 3.2 ± 0.08 30 30 30 30 30 30 30 30 30 Iran, Karaj 2 3.4±0.13 37.0 ± 1.90 6.0 ± 0.15 91.7 ± 1.31 7.1±0.11 5.7±0.11 56.8 ± 2.50 98.3 ± 1.50 3.1±0.10 30 30 30 30 30 30 30 30 30 Israel, Ashqelon 3.5±0.16 37.7±1.11 6.7±0.19 85.1 ± 1.54 6.3 ± 0.14 5.2±0.11 60.4 ± 1.88 88.2 ± 1.65 2.9± 0.11 19 19 30 30 30 30 19 30 30 Israel, Shefayim 3.5 ±0.11 11.7 ± 2.86 5.8 ± 0.13 79.8 ± 1.21 5.2 ± 0.07 4.7 ± 0.07 39.6 ± 1.54 67.9 ± 1.43 2.7 ± 0.08 30 30 30 30 30 30 30 30 30 Kuwait, city 1.4 ± 0.07 32.0 ± 0.71 4.2±0.17 82.5 ± 1.04 6. 7 ± 0.11 4.7 ± 0.06 55.2 ± 1.56 94.1 ± 1.92 3.6± 0.13 30 30 30 30 30 30 30 30 30 Kuwait, AI Qurain 1.2 ± 0.05 29.2± 1.04 3.7±0.14 82.4± 1.07 6.1 ±0.11 4.3 ± 0.07 48.9 ± 1. 73 89.7 ± 0.87 3.4 ± 0.09 30 30 30 30 30 30 30 30 30 Morphology 12

Table 2. Continued. Site Basal spine Basal spine Abaxial Abaxial Base length Height Angle of Angle of No. of length (rnm) angle to spine length spine angle (rnm) (rnm) basal spines abaxial seeds base {0} {mm} to base {0} {0} SEines {0} Namibia, Gobabeb 1 2.8 ± 0.07 34.5 ± 0.95 4.5 ± 0.12 82.3 ± 1.20 5.9 ± 0.08 4.5 ± 0.05 65.1 ± 1.54 79.4 ± 1.55 3.4±0.10 30 30 30 30 30 30 30 30 30 Namibia, Gobabeb 2 3.5 ± 0.20 28.3 ± 1.94 5.3 ± 0.25 84.6 ± 1.82 5.2 ± 0.13 4.8 ± 0.12 54.0 ± 2.25 93 .7 ± 2.02 2.9 ± 0.10 30 30 30 30 30 30 30 30 30 Namibia, Tiroll 3.0 ± 0.09 38.8 ± 1.03 4.5± 0.13 87.2 ± 1.02 4.8 ± 0.08 5.1±0.05 63 .6±2.27 94.9±1.76 2.9 ± 0.07 30 30 30 30 30 30 30 30 30 S. Africa, Rondebosch 3.1 ± 0.11 21.9 ± 2.93 4.4 ± 0.16 80.8 ± 1.45 5.1±0.13 5.1±0.12 50.8 ± 2.24 105.3 ± 3.64 3.2 ± 0.10 30 30 30 30 30 30 30 30 30 USA, Prosser 2.3 ± 0.08 36.0 ± 0.84 6.1 ± 0.13 85.5 ± 0.95 6.0 ± 0.09 5.8 ± 0.09 57.9 ± 1.05 93.9 ± 2.39 3. 1 ±0.08 7 7 30 30 30 30 7 30 30 USA, S. Pemiscot 4.4 ± 0.13 89.5 ± 1.24 5.3 ± 0.07 5.1 ± 0.08 104.8 ± 1.55 2.9 ± 0.09 0 0 30 30 30 30 0 30 30 Morphology 13

Table 3. Pearson correlation coefficients between measurements on T. te"estris s.L burrs The correlation coefficent, significance test and the sample size (in parentheses) are given for each combination of variables. Basal Abaxial Abaxial Base Height Angle of Angle of No. of spine spine length spine length (nun) basal abaxial seeds 0 angle to (mm) angle to (nun) spines CO) spines ( ) 0 0 base ( ) base ( ) Basal spine -0.03 0.59 0.05 0.10 0.44 0.01 -0.07 -0.20 length n.s. **** n.s. ** **** n.s. n.s. **** (mm) (750) (742) (742) (750) (750) (750) (742) (750) Basal spine 0.09 0.35 0.28 0.04 0.36 -0.05 0.19 angle to ** **** **** n.s. **** n.s. **** 0 base ( ) (742) (742) (750) (750) (750) (742) (750) Abaxial spine -0.05 0.38 0.47 -0.02 -0.19 0.07 length n.s. **** **** n.s. **** * (mm) (922) (922) (922) (742) (922) (922) Abaxial spine 0.08 -0.07 0.23 0.07 0.04 angle to * * **** * n.s. 0 base ( ) (922) (922) (742) (922) (922) Base 0.37 0.03 -0.20 . 0.50 length **** n.s. **** **** (mm) (930) (750) (922) (930) Height 0.05 0.02 0.06 (mm) n.s. n.s. n.s. (750) (922) (930) Angle of 0.21 0.07 basal **** n.s. 0 (750) spines ( ) (742) Angle of -0.12 abaxial *** ~n ~ n.s., not significant;* P < 0.05; ** P < 0.01; *** P < 0.001 ; **** P < 0.0001. Morphology 14

Table 4. Canonical discriminant analysis of Tribulus terrestris s.l. burr morphology, incorporating all variables

Canonical variate Variable I 2 3 Basal spine length (mm) 0.81 0.25 -0.05 Basal spine angle to base e) -0.40 0.46 0.37 Abaxial spine length (mm) 0.62 0.67 -0.11 Abaxial spine angle to base e) -0.17 0.11 0.33 Base length (mm) -0.20 0.89 0.23 Height (mm) 0.71 0.33 0.19 0 Angle of basal spines ( ) -0.18 0.15 0.48 Angle of abaxial spines e> 0.16 -0.45 0.83 No. ofseeds -0.70 0.34 0.04 Statistics Eigenvalue 18.39 2.83 1.49 Likelihood ratio 0.005 0.105 0.401 Approx. F 7.65 3.92 3.41 Numerator df 27 16 7 Denominator df 41.52 30 16 p **** *** * * p < 0.05; *** p < 0.001; **** p < 0.0001.

Table 5. Canonical discriminant analysis of Tribulus te"estris s.L burr morphology, incorporating the variables, abaxial spine length, base length and height

Canonical variate Variable 1 2 3 Abaxial spine length (mm) 0.91 -0.03 -0.41 Base length (mm) 0.62 . 0.77 0.15 Height (mm) 0.84 -0.41 0.36 Statistics Eigenvalue 6.68 1.95 0.001 Likelihood ratio 0.044 0.338 0.999 Approx. F 17.67 9.35 0.03 Numerator df 9 4 1 Denominator df 61 52 27 p **** **** n.s. n.s., not significant; *** P < 0.001; **** P < 0.0001. Morphology 15

Angle between abaxial spines

Angle between -E basal -E spines .c+- ~~~.,...... _.. CJ) Q) I

Base length (mm)

Abaxial spine Angle of length abaxial spines

Angle of basal spines

Seed Basal spine length

Fig. 1. Measurements-taken on burrs of Tribulus terrestris s.l. The illustration is based on Gardner ( 1948) Morphology 16

300 200 {)' g 200 {)' 150 c:r:::J c: ~ Q.) u. 100 :::Jc:r 100 ~ u. 50 -60 -40 -20 0 20 40 60 80 1 00 0 Basal spine angle to base (degrees) 0 1 2 3 4 5 6 7 200 Basal spine length (mm) 250 {)' 150 c: Q.) 200 :::J 100 {)' c:r ~ g 150 u. :::J 50 ~ 100 u. 0 1 2 3 4 5 6 7 8 9 10 11 Abaxial spine length (mm) 0 20 40 60 80 100 120 140 300 Angle between basal spines (degrees) 200 {)'c: 200 Q.) {)' 150 c:r:::J c: ~ 100 Q) u. :::Jc:r 100 ~ u. 50 40 60 80 100 120 140 Abaxial spine angle (degrees) 3 4 5 6 7 8 9 10 300 Base length (mm)

{)'c: 200 300 Q.) :::J {)' c:r g 200 u.~ 100 c:r:::J ~ u. 100 0 20 40 60 80 100120140160180 Angle between abaxial spines (degrees) 2 3 4 5 6 7 8 600 Height (mm)

{)'c: 400 Q.) :::J

u.~ 200

0 2 3 4 5 6 Number of seeds

Fig. 2. Frequency distribution of measurements of Tribulus terrestris s.l. burrs Morphology 17

Average distance between clusters 1.3 1. 0 0.5 0.0

I WA, Moorine Rock VIC, Mlldura 1

I WA, Goomalling VIC, Walpeup S. Africa, Rondebosch I"" I I NSW, Coonamble Namibia, Gobabeb 2 VIC, Rutherglen SA, Wunkar Ci;J SA, Taldra rl Namibia, Tiroll WA, Busselton

I NT, Darwin l NT, Katherine 1 QLD, Horne Hill ® I I QLD, Cranbrook

...._ Kuwait, city India, Pulwama ....__ -I India, Nubra River Kuwait, AI Qurain Namibia, Gobabeb 1 NSW, Attunga Iran, Karaj 2 J ©- L Iran, Gaoh Sar 1 Israel, Ashqe lon D Israel, Shefayim

Fig. 3. Dendrogram of Tribulus terrestris s.l. collections using all variables Morphology 18

4 c 3 ccc N Q) 2 ..() B B B co 1 D ·c A co B B A > 0 B B B AA -1 B . ~ A AM c A 0 -2 k c rn -3 (.) -4 A -5 -10 -8 -6 -4 -2 0 2 4 6 Canonical variable 1 3 2 (") B A/tn CC Q) 1 B A AA :0 rn EB A ~ ·c 0 rn E9 ~ ~Ac > -1 B A ~ -2 c 0 -3 c co -4 (.) -5 D -6 -10 -8 -6 -4 -2 0 2 4 6 Canonical variable 1 3 2 (") c A B c Q) 1 A t A B ..() A co 0 A B B c ·c Bs co B B c > -1 ~ A B A gco -2 c 0 -3 c co -4 (.) -5 D -6 -5 -4 -3 -2 -1 0 1 2 3 4 Canonical variable 2

Fig. 4. Canonical discriminant analysis of Tribulus terrestris s.l. collections using all variables. N = 26 sites. Morphology 19

0.0 .5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 Length of basal spine (mm)

Fig. 5. Frequency distribution ofbasal spine length for 50 Tribulus terrestris s.l. burrs from each of five plants collected at Mil dura, Victoria Morphology 20

Average di~ar-ce between clusters 1.5 1.0 0.5 0.0 I I I I I I I I I I WA. Moorine Rock

WA, Carr~aNon WA, Nita Downs Stn WA, Goomalling VIC, Walpeup SA. Taldra VIC, Rutherglen India, Pulwama Namibia, Gobabeb 1 India, Nubra River S. Africa, Rondebosch USA, S. Pemiscot WA, Broome Namibia, liroll NSW, Coonamble Namibia, Gobabeb 2 VIC, Mildura 1 WA. Busselton SA. Wunkar - Israel, Shefayim NT, Darwin NT, Katherine 1 Kuwait, city QLD, Home Hill QLD, Cranbrook Kuwait, AI Qurain NSW, Attunga Iran, Karaj 2 Iran, Gaoh Sar 1 USA, Prosser Israel, Ashqelon

Fig. 6. Dendrogram of Tribulus terrestris s.l. collections using length of base and abaxial spine and height Morphology 21

5 c N 4 Q.) 3 ..0 c c ro c ·c D ro 2 c > c D 1 B D ~ B B B B c: 0 § D 0 A c: -1 B B ro PaBs8 () A A B D -2 B B -3 -6 -4 -2 0 2 4 6 Canonical variable 1 3

(") 2 B Q.) ..0 cs ro 1 c ·c B ers D ro A c > 0 B sB ~ D D D ro A B .2 A c: -1 c ec B 0 B c: D ro B B () -2 -3 -6 -4 -2 0 2 4 6 Canonical variable 1 3

(") 2 B Q.) ::0 B c ro c ·c 1 B B D ro B B ~ D c > 0 ~ B B D ~ ~ B c: B B c c c 0 -1 c: B ro m D () -2

-3 -3 -2 -1 0 1 2 3 4 5 Canonical variable 2

Fig. 7. Canonical discriminant analysis of Tribulus terrestris s.l. collections incorporating length of base and abaxial spine and height N = 31 sites. Morphology 22

Fig. 8. Plots ofT. te"estris s.l. burr base length by height showing regression lines for Australian sites representing the four groups detected in Figures 6 and 7 Chromosome numbers 23

Variation in Australian and world-wide populations of Tribulus te"estris L. 2. Chromosome numbers

S. M Morrison and J. K. Scott

CSIRO Division ofEntomology, Private Bag, P.O. Wembley, W.A. 6014, Australia.

Running head: Chromosome numbers of Tribulus terrestris

Abstract Chromosome numbers in 24 Australian and 24 overseas collections of Tribulus terrestris s.l. revealed three ploidy levels, tetraploid 2n = 24, hexaploid 2n = 36 and octoploid 2n = 48. An Australian native Tribulus species (previously thought to be T. terrestris) occurring in the Northern Territory and ·Queensland, had counts of 2n = circa 24. The majority of weedy collections ofT. terrestris in southern and north west regions of Australia had counts of2n =circa 36, but two collections had counts of2n =circa 24. Collections from South Africa and Namibia also had a mixture of tetraploid and hexaploid forms. The USA, Mediterranean and Middle East samples (except for three Kuwait samples) all were hexaploid. All Indian and three Kuwait samples were the only octoploid collections observed. These ploidy levels agree with those previously reported for T. terrestris. It was concluded that the polyploid complex has an autopolyploid origin. The predominantly hexaploid weedy collections in Australia indicate an overseas origin, and the presence of two tetraploid collections point to two or more separate introductions of T. terrestris into Australia. Counts for T. micrococcus and T. occidentalis were 2n :::; circa 48 which provide further evidence for a base number ofx = 6 or 12 for the genus.

Introduction An introduction to Tribulus terrestris s.l. L. is given in Scott and Morrison (1994). Cytological studies from outside Australia have indicated that there is a great variation in ploidy levels in the T. terrestris complex (Table 1). Ploidy levels range from diploid to octoploid with a basic chromosome number ofx = 6 (Baquar et al. 1965, Malik 1966, Porter 1968). The work reported here consists of cytogenetic analysis ofT. terrestris from within Australia and from world-wide locations. The aim was to determine chromosome numbers and to provide evidence that would indicate the origins of the weed in Australia.

Methods Burrs ofT. terrestris sensu Jato were obtained from a wide range of geographical locations (see Scott and Morrison (1994) for further details). Some of the material received included T. micrococcus Domin and T. occidentalis R. Br. Seedling root tips were used for chromosome preparations and were obtained in the following manner. Seeds were excised from the woody burr, surface sterilized with 1.25% sodium hypochlorite for 5 minutes, then rinsed in distilled water. The seeds were germinated on moist filter paper in a sealed petri dish on a 16 hour 20°C, and 8 hour 35°C cycle. Germination success was between 50% and 75%. Root tips were harvested from 3 to 4 day old seedlings. Approximately 10 mm of root tip was excised and immersed in 0.01% colchicine for under 2 hours. The root tips were then fixed in a cold mixture of acetic acid and 100% ethanol in the ratio 1:3 and stored below -20°C until needed. Chromosome numbers 24

Root tips were prepared for chromosome squashes as follows. The extreme end of the root tip (I mm) was excised, hydrolysed in a drop of 1.0 M hydrochloric acid on a glass slide and gently heated to approximately 60°C. After 5 minutes of acid treatment, the material was placed on a clean slide and finely macerated with a scalpel. Once the cells had almost dried, they were stained with 4 drops of aceta-carmine stain (50% acetic acid) and left for 5 minutes. The cells were squashed under a coverslip and the preparation sealed with nail varnish. Between two and five counts were made for each collection. Chromosome preparations were made from each of two collections ofT. micrococcus and T. occidentalis. Vouchers specimens are held at the Division of Entomology in Perth and will be lodged at either the National Herbarium of New South Wales, Sydney (NSW) or the Western Australian Herbarium (PERTH).

Results Several preparations of T. terrestris chromosomes were obtained in which the counts were clearly euploid. These included Moorine Rock (Western Australia) 2n = 24 (Fig. 1), Katherine (Northern Territory) 2n = 24, Carnarvon (northern Western Australia) 2n = 36 (Fig. 1), Tamworth (New South Wales) 2n = 36, Kuwait (Khaldiya) 2n = 48 and India (Pampur) 2n = 48. The majority of the counts for other locations however, frequently ranged from 1 to 4 below these three ploidy levels. It was assumed that the chromosome numbers for this species follow a euploid series with a basic chromosome number of x = 6, as do other species in this genus (Porter 1968, Husain 1986). The chromosome numbers are therefore given the prefix "circa". Chromosome counts were obtained for 24 world-wide and 24 Australian locations (Table 2). The Queensland and Northern Territory collections had counts of2n = circa 24. The majority of the collections in southern Australia and northern Western Australia had counts of2n = circa 36 and two had counts of2n =circa 24. Of the world-wide samples, those from the USA, Middle East and the Mediterranean all had counts of2n = circa 36, whereas samples from South Africa and Namibia had counts ofboth 2n = circa 24 and 2n = circa 36. All the Indian and three of the Kuwait collections had a higher chromosome count of2n = circa 48 (Table 2). Counts of2n = circa 48 were obtained forT. occidentalis from northern Western Australia and 2n = circa 48 forT. micrococcus from central New South Wales.

Discussion Ploidy levels The three ploidy levels (tetraploid, hexaploid and octoploid) observed in our sample ofT. terrestris fall within the range reported for populations world-wide (Table 1) , but no diploid individuals were found. Determination ofwhether polyploids have arisen by autopolyploidy, allopolyploidy or segmental allopolyploidy is a complex matter and evidence is needed from several sources (Soltis and Rieseberg 1986). Although it was not possible to examine meiotic chromosome behaviour in this study, it has been observed in T. terrestris by several researchers. The majority of observations at diakinesis and metaphase I of meiosis, report the presence of normal bivalents only (Fotedar and Roy 1969, Ramachadran and Kuriachan 1970, Kumar 1978, Bhansali 1980), and a few report a prevalence of normal bivalents with a small number of multivalents (Rao and Tandon 1971, Hilu 1981, Husain 1986). This has been observed in tetraploids (Bhansali 1980), hexaploids (Ramachadran and Kuriachan 1970, Rao and Tandon 1971, Kumar 1978, Hilu 1981, Husain 1986) and in octoploids (Fotedar and Roy 1969, Rao and Tandon 1971, Husain Chromosome numbers 25

1986). Although multivalents are uncommon in T. terrestris, the predominance of bivalents does not preclude an autopolyploid origin. Soltis and Rieseberg (1986) observed a complete lack of multivalents in Tolmiea menziesii (Pursh) Torrey & A. Gray (Saxifragaceae) which they showed to be an autopolyploid, and commented that normal bivalents alone have been found in other naturally occurring autopolyploids. Additionally, Elliot (1958) stated that the meiotic chromosomes of many induced autopolyploids are also almost always associated as bivalents. Allopolyploidy in T. terrestris has been suggested by Ramachadran and Kuriachan (1970) and by Kumar (1978) largely on the basis of a lack of multivalents in meiosis, which has since been shown to be insufficient evidence of either ploidy type. A further complication is the presence of dominant genes in some populations which prevent homeologous gene pairing in allopolyploids (Jackson and Casey 1982). As a consequence, pairing only occurs between homologous chromosomes resulting in predominance ofbivalents. Allozyme analysis however, provides evidence -in favour of autopolyploidy in T. terrestris (Morrison and Scott 1994). There was unbalanced staining in heterozygotes at five loci (6PGD-1, 6PGD-2, MDH-1, IDH-1 and IDH-2), evidence of enzyme multiplicity in two enzymes ( 6PGD and PGI) and fixed heterozygosity was only observed at one locus (6PGD-2) out of eleven loci examined. These factors indicate polysomic inheritance which is a characteristic of autopolyploidy. Overall, the evidence indicates that polyploidy in T. terrestris is more likely to have arisen by autopolyploidy than allopolyploidy. Stebbins (1950) concluded that autopolyploidy is unlikely to occur beyond the tetraploid level. If this is the case, then the hexaploid and octoploid forms could be autoallopolyploid. Further evidence from diploid populations and controlled crosses are needed to clarify the situation.

Basic chromosome number Originally the basic chromosome number for the Tribulus genus was thought to be x = 12 (Darlington and Wylie 1955). Subsequent chromosome counts on T. terrestris (Baquar eta/. 1965, Malik 1966, Ramachadran and Kuriachan 1970, Rashid 1974) however, indicated that the basic chromosome number was x = 6. Porter (1968) suggested that the Tribulus genus has a euploid series of chromosome numbers with a basic number ofx = 6. In 1976, Bhandari and Sharma described a new species, Tribulus rajasthanensis, which was previously included in the T. terrestris complex. This new species had a chromosome count ofn = 6. Additionally, it was collected from the same region, N.W. Rajasthan desert, that Baquar eta/. (1965) and Malik ( 1966) obtained plants with chromosome counts of n = 6 and 2n = 12 respectively. After morphological studies, the voucher specimens lodged by Baquar et al. (1965) were found to be T. rajasthanensis (Husain 1986). Husain (1986) also suggested that the specimens described by Malik (1966) belonged to the new species. No other counts indicating a basic chromosome number of x = 6 have been reported for T. terrestris. It appears likely that although a diploid form ofT. terrestris with 2n = 12 has not been recorded, the basic chromosome number for the Tribulus genus is x = 6. The previously unreported counts of2n = circa 48 for both T. micrococcus and T. occidentalis support the base chromosome number for the Tribulus genus being x = 6 or 12.

Comparison with isozyme results A study of isozyme variation in T. terrestris was assessed for the same collections ofburrs and is reported in the following paper (Morrison and Scott 1994). The weedy Chromosome numbers 26 form ofT. terrestris in southern Australia and northern Western Australia as distinguished by isozyme data, mainly consists of hexaploid collections, but two collections (in Victoria and southern Western Australia) are tetraploid. Overseas collections that are genetically most similar to the weedy form in Australia, are also comprised of tetraploid (1 from South Africa, 1 from Namibia) and hexaploid (1 from South Africa, 3 from Namibia, USA, Mediterranean and the Middle East) forms. One exception in this group is a Kuwait sample that is octoploid. Northern Territory and Queensland samples were, however, consistently tetraploid and had isozymes bands sufficiently different to indicate that the samples were from a separate species to T. terrestris. A small genetically separate group consisting of two Namibian collections was also tetraploid. This could represent a native form of Tribulus in southern Africa. All Indian and two Kuwait collections, which have the least genetic similarity to other groups, were octoploid. Evidently chromosome numbers alone cannot be used to predict isozyme patterns. However, there is a reduced number of ploidy levels (predominantly hexaploid with some tetraploid), in the regions where weedy forms of T. terrestris occur. The two ploidy levels in the Australian weedy form of T. terrestris indicates that there were either at least two separate introductions of the plant into Australia, or that a second ploidy level evolved in Australia subsequent to the introduction of one or more tetraploid form(s). An increase in ploidy level with greater genetic and biochemical diversity, could result in an increased capacity to adapt to a new environment compared with the diploid progenitor (Soltis and Rieseberg 1986, Crawford 1990). Selection would therefore favour the development of a hexaploid form from a tetraploid form, resulting in the presence oftwo ploidy levels. However, there is no clear association between polyploidy and colonizing ability (Barrett and Shore 1990). Alternatively, the development of a hexaploid form would be facilitated in regions where diploid populations are sympatric with tetraploids, as in India. If this were the case, the separate introduction of two ploidy forms would be favoured, as no diploid forms have been reported from Australia.

Biogeographical patterns and centres ofspeciation India and Pakistan display the greatest range of ploidy levels in their T. terrestris populations including diploid, tetraploid, hexaploid and octoploid levels (Tables 1 and 2). Diploid and polyploid populations only overlap in West Pakistan and India (Hilu 1981). However, diploid populations ofT. terrestris from Rajasthan, India (Baquar et a/. 1965, Malik 1966) are likely to be a different species, T. rajasthanensis (Husain 1986). In other areas where T. terrestris is native including the Mediterranean and the Middle East, tetraploid, hexaploid and octoploid forms occur. In regions where T. terrestris has been introduced, the USA (Huffaker eta/. 1983), possibly South Africa (Schweickerdt 1939) and Australia, only tetraploid and hexaploid populations exist. It is most likely that the areas with diploid populations and the greatest diversity in ploidy levels (India and Pakistan) represent the centre of diversification among plants identified as T. terrestris, while regions where it has been introduced are less diverse.

Acknowledgments This study was supported by a grant from the Australian Dried Fruits Research and Development Council. We thank the collectors of caltrop seed, the Department of Agriculture Western Australia, Weed Science Branch, for the use of laboratory facilities and Dr David Coates of the Western Australian Herbarium for advice and the Chromosome numbers 27

use of his microscope. J. R. Hosking and P. B. Yeoh commented on early drafts ofthe manuscript.

References Baquar, S. R., Aktar, S. and Husain, A. (1965). Meiotic chromosome numbers in some vascular plants oflndus Delta I. Botaniska Notiser 118, 289-298. Barrett, S.C. H. and Shore, J. S. (1990). Isozyme variation in colonizing plants. In 'Isozymes in Plant Biology.' (Eds D. E. Soltis and P.S. Soltis.) pp. 106-126 (Chapman and Hall: London.) Bhandari, M . M. and Sharma, V. S. (1976). A new Tribulus (Zygophyllaceae) from India. Botaniska Notiser 129, 367-369. Bhansali, A. K. (1980). Cytological study of the morphological variants of Tribulus terrestris Linn Indian Forester 106, 734-737. Bir, S. S. and Sidhu, M. (1980). Cyto-palynological studies on weed flora of cultivable lands of Patiala district (Punjab). Journal ofPalynology 16, 85-105. Crawford, D. J. (1990). Enzyme Electrophoresis and Plant Systematics. In 'lsozymes in Plant Biology.' (Eds D. E . Soltis and P.S. Soltis.) pp. 146-164 (Chapman and Hall: London.) Darlington, C. D. and Wylie, A. P. (1955). 'Chromosome atlas offlowering plants.' 2nd edn. (Allen and Unwin: London.) Elliot, F. E. (1958). Polyploidy and Plant Breeding. In 'Plant Breeding and Cytogenetics.' (McGraw-Hill Inc.: USA.) Fotedar, J. L. and Roy, S. K. (1969). Tetraploid Tribulus Linn. Science and Culture 35, 25-26. Heiser, C. B. and Whitaker, T. W. (1948). Chromosome number, polyploidy and growth habit in California weeds. American Journal ofBotany 35, 179-186. Hilu, K. W . (1979). Report. Taxon 28, 395. Hilu, K. W . ( 1981 ). Cytotaxonomical studies in Tribulus terrestris and T. a latus (Zygophyllaceae). Nordic Journal ofBotany 1, 531-534. Huffaker, C. B., Hamai, J. and Nowierski, R. M. (1983). Biological control of puncturevine, Tribulus terrestris in California after twenty years of activity of introduced weevils. Entomophaga 28, 387-400. Husain, S. A. (1986). Cytotaxonomic studies in Tribulus from Pakistan. La Kromosomo II 42, 1316-1329. Jackson, R. C. and Casey, J. (1982). Cytogenetic analysis of autopolyploids: Models and methods for triploids to octoploids. American Journal ofBotany 69, 487-501. Kumar, A. (1978). Intraspecific polyploidy in Tribulus terrestris L. Science and Culture 44, 427-428. Malik, C. P. (1966). Corrected basic chromosome number and intraspecific polyploidy in Tribulus terrestris Linn. Chromosome Information Service 7, 7-8. Morrison, S. M. and Scott, J. K. (1994). Variation in Australian and World-Wide populations of Tribulus terrestris L. 3. Isozyme analysis. (Unpublished manuscript.) Negodi, G. (1939). Cariologia delle Rutaceae e delle Zygophyllaceae. Scientia Genetica 1, 180-185. Pastor, J., Fernandez, I. and Diez, M. J. (1988). Numeros cromosomocos para Ia flora Espanola 528-543. Lagascalia 15, 124-129. Porter, D. M . (1968). The basic chromosome number in Tribulus (Zygophyllaceae). Wasmann Journal ofBiology 26, 5-6. Chromosome numbers 28

Ramachadran, K. and Kuriachan, P. I. (1970). A hexaploid Tribulus terrestris. Linn. Science and Culture 36, 421-422. Rao, G. R. and Khan, R. (1972). Genetic breakdown of chromosome behaviour of Tribulus terrestris. Experientia 28, 217. Rao, G. R. and Tandon, S. L. (1971). Mechanics of evolution in higher chromosomal forms in puncture vine weed. Indian Journal ofHorticulture 28, 63-67. Rashid, A. ( 1974). Chromosome numbers of some Mediterranean plants. Bangladesh Journal ofBotany 3, 75-82. Rostovtseva, T. S. (1977). Chromosome numbers of some plant species from the south of Siberia. II. Botanicheskii Zhumal SSR 62, 1034-1042. Sanjappa, M. (1979). Reports. Taxon 28, 274. Sarkar, A. K., Malik, R., Dutta, N. and Chatterjee, U. (1977). Reports. Taxon 26, 449. Schnack, B. and Covas, G. (1947). Kariological studies in the flowering plants. Haumania 1, 32-41. (Not seen, taken from Biological Abstracts 25, 9, (1951), abstract no. 90). Schweickerdt, H. G. (1939). An account ofthe South African species of Tribulus Toum. ex Linn. Bothalia 3, 157-178. Scott, J. K. and Morrison, S.M. (1994). Variation in Australian and World-Wide populations of Tribulus terrestris L. 1. Burr morphology. (Unpublished manuscript.) Shetty, B. V. (1961). Chromosome numbers in some South Indian plants. Journal of Scientific and Industrial Research 20C, 28-30. Singh, P. (1984). Reports. Taxon 33, 759. Soltis, D. E. and Rieseberg, L. H. (1986). Autopolyploidy in Tolmiea menziesii (Saxifragaceae): Genetic insights from enzyme electrophoresis. American Journal ofBotany 73, 310-318. Stebbins, G. L. (1950). 'Variation and evolution in plants.' (Columbia University Press: New York.) (Not seen, taken from Kumar, 1978). Sugiura, T. (1940a). A list of chromosome numbers in angiospermous plants. Proceedings of the Imperial Academy of Tokyo 16, 15-16. Sugiura, T. (1940b). Studies on the chromosome numbers in higher plants V. Cytologia 10, 363-370. Zacchareva, 0 . I. and Astanova, S. B. (1968). Chromosome numbers of some wild species of flowering plants of middle Asia. Doklady-Academiia Nauk Tadzhikskoi SSR 11, 72-75. (in Russian.) Chromosome numbers 29

Table 1. Summary of chromosome counts of Tribulus terrestris s.L Chromosome Location Source count n=6 Indus delta Baquar et a/. 1965 2n= 12 India, Udaipur, Rajasthan Malik 1966 n= 12 India, Jodhpur, Coimbatore Bhansali 1980 N. America?, ? Negodi 1939 Europe, botanical gardens Sugiura 1940a, b 2n=24 India, Udaipur, Rajasthan Malik 1966 California Heiser and Whitaker 1948 , Seville Pastor et a/. 1988 n= 16 India, Maharastra, Pal Forest Sarkar et a/. 1977 n = 18 India, Cape Comorin, Ramachadran and Kuriachan 1970 Manavalakurichi India, Aligarh Rao and Khan 1972 India, Agra Kumar 1978 India, Coimbatore? Shetty 1961 India, Delhi Rao and Tandon 1971 Pakistan Husain 1986 , Montpellier Rashid 1974 2n= 36 India, Udaipur, Rajasthan Malik 1966 Iraq, Baghdad Hilu 1979, 1981 Iraq, Babylon, Karbala, Omarah Hilu 1981 Russia Zacchareva and Astanova 1968 Southern Siberia Rostovtseva 1977 n=24 India, V aranasi Fotedar and Roy 1969 India, Karnataka, Bangalore Sanjappa 1979 India, Rajasthan, Jodhpur Singh 1984 India, Agra Kumar 1978 India, Delhi Rao and Tandon 1971 Pakistan Husain 1986 2n=48 India, Udaipur, Rajasthan Malik 1966 India, Punjab Bir and Sidhu 1980 Argentina (Squires 1979) Schnack and Covas 1947 Old world tropics (Hilu 1981) Chromosome numbers 30

Table 2. Chromosome numbers in Australian and overseas populations of Tribulus te"estris s. L Chromosome Country State Location Latitude; Longitude count 2n=24 Australia Northern Territory Darwin 12° 21'S; 130° 52'E (tetraploid) Katherine 1 14° 32'S; 132° 24'E Katherine 2 14° 28'S; 132° 16'E Queensland Cranbrook 19° 16'S; 146° 49'E Home Hill 19° 52'S; 147° 15'E Mingela Range 19° 53'S; 146° 38'E Sellheim 20° OO'S; 146° 25'E Victoria Walpeup 35° OS'S; 142° 02'E Western Australia Moorine Rock 31° 19'S; ll9° 04'E South Africa Rondebosch 33° 56'S; 18° 28'E* Namibia Rosh Pinah 28° OS'S; 16° SO'E* TiroII 26° IS'S; 16° 45'E* Thomasberg 23° lO'S; 15° 31'E 2n= 36 Australia New South Wales Attunga 30° 56'S; 150° SO'E (hexaploid) Coonamble 30° 57'S; 148° 24'E Tamworth 31 o 09'S; 150° 59'E South Australia Naracoorte 36° 58'S; 140° 45'E Taldra 34° 21'S; I40° Sl'E Wunkar 34° 29'S; I40° 18'E Victoria Mildura 2 34° II'S; I42° lO'E Western Australia Broome 17° 58'S; I22° 14'E Busselton 33° 39'S; II 5° 20'E Carnarvon 24° 53'S; 113° 40'E Dongara 32° 57'S; 123° 25'E Goomailing 3I0 IS'S; 1I6° SO'E Lake Biddy 33° OO'S; 118° 56'E Nita Downs Stn. I9° OS'S; 12I 0 4I'E Wickepin 32° 47'S; I17° 30'E Crete 35° 20'N; 25° OO'E Cyprus Larnaca 34° 54'N; 33° 39'E Iran Gaoh Sari 36° IO'N; 51 o 25'E* Gaoh Sar 2 36° IO'N; 51° 25'E* Karaj 1 35° 48'N; 50° 58'E* Karaj 2 35° 48'N; 50° 58'E* Kuwait Kuwait city 29° 20'N; 48° OO'E Namibia Namib Naukluft Pk Gobabeb Station I 23° 34'S; I 5° 03'E Gobabeb Station 2 23° 34'S; I 5° 03'E South Africa Prince Albert 33° 13'S; 22° 03'E USA Kansas Manhattan Res Stn 39° II 'N; 96° 35'W Pemiscot Co. Southern region 36° OS'N; 89° SO'W* Washington State Prosser Res Stn 46° 13'N; Il9° 48'W 2n=48 Kuwait Al Qurain 29° 02'N; 48° 08'E ( octoploid) Kuwait city, Ag 29° 20'N; 48° OO'E Stn Kuwait, Khaldiya 29° 20'N; 48° OO'E Chromosome numbers 31

India Ladakh Kargil 34° 31 'N; 76° 13'E Nubra 34° 40'N; 77° 40'E* Kashmir Pam pur 34° 05'N; 74° 55'E* Pulwama 33° 55'N; 74° 55'E* * Approximate location only Chromosome numbers 32

a) ' . "-

~' - ft, . I ; '", ...... I ,· r ·, ·• t - '

I' b) I - ; ~..,.,., ')~ ·'-I ,...... ,.'· t·

.';J ~ &

Fig. I. Examples of Tribulus terrestris s.l. chromosome spreads with interpretation. a) Moorine Rock, Western Australia (2n = 24), b) Carnarvon, Western Australia (2n = 3~ .

Isozymes 33

Variation in Australian and World-wide Populations of Tribulus te"estris L. 3. Isozyme analysis

S. M Morrison and J K. Scott

CSIRO Division ofEntomology, Private Bag, P.O. Wembley, W.A. 6014, Australia.

Running head: Isozyme variation in Tribulus terrestris L.

Abstract Isozyme variation was examined in seedlings obtained from 30 Australian and 24 world-wide collections of Tribulus terrestris s.l. burrs. Polymorphism was detected in eight ofthe eleven putative loci scored. Queensland and Northern Territory collections were identified as distinct from other Australian and world-wide collections indicating that the former are native to Australia and represent a separate species. All other Australian collections had a high genetic similarity to burrs obtained from the Mediterranean, the Middle East, South Africa, Namibia and the USA, indicating possible origins of the Australian material. Two Namibian collections formed a separate group that could represent a native species in this country. All Indian and two Kuwait collections were grouped together and had little similarity with any other group. The introduced Australian collections are most likely to have originated in the Mediterranean I Middle East region or southern Africa.

Introduction An introduction to Tribulus terrestris s.l. L. is given in Scott and Morrison (1994). Isozyme analysis has been used for clarifying the taxonomy of several weed species including Xanthium occidentale Bertol. (Moran and Marshall 1978), Emex spinosa (L.) Campd. (Marshall and Weiss 1982), Emex australis Steinh. (Panetta and Carstairs 1989, Panetta 1990) and Prosopis spp. (Panetta and Carstairs 1989). Here we report on an analysis using isozyme techniques to aid the identification of the variable species, T. terrestris in Australia, and to determine its origin.

Methods Tribulus terrestris sensu lata accessions were obtained by written requests to potential collectors (Scott and Morrison 1994). Collections ofburrs were received from world-wide locations, although some sites were only represented by small samples. Thirty collections were analysed from mainland Australia (Fig. 1), and 24 from world-wide sites (Fig. 2). Thirty seedlings were assayed for a minimum ofthree collections from each location (excluding South Africa which had two collections, and Crete and Cyprus which had only one collection each). Altogether, 30 seedlings were assayed from each of32 sites, and from 4 to 27 seedlings from each of a further 22 sites (Table 1). The range of samples analysed should provide a 95% probability of detecting all alleles at a locus in diploid individuals (Marshall and Brown 1975). Dry burrs were stored at room temperature until needed. Seeds were excised from the woody burr and surface sterilised with 1.25% sodium hypochlorite for 5 minutes and then rinsed in distilled water. The seeds were germinated on moist filter paper in a sealed petri dish on a 16 hour 20°C, and 8 hour 35°C cycle. Three to four day old seedlings were used for isozyme work. Cotyledon and root tip tissue from each seedling were homogenised in 7 ~I of0.16M phosphate buffer pH 7.0 (containing Isozymes 34

2.5% sucrose, 0.12% bromophenol blue and 0.75 mg mt-1 dithiothreitol), in a plastic microwell tray on an ice bath (Panetta and Carstairs 1989). Three 2 x 4 mm filter paper wicks were placed in each sample. Horizontal starch gels were made from 9.6% hydrolysed starch (Connaught) with three different buffer systems (Table 2). The three buffer systems used were histidine (Brewer and Singh 1970), tris-citrate (Richardson eta/. 1986) and lithium borate (Panetta and Carstairs 1989). Electrophoresis was carried out at 4°C in a refrigerator, with an ice bath placed over each gel and with a maximum current of 50 rna. Gels were run until the bromophenol blue front had moved approximately I 0 em. Each gel was sliced horizontaiiy into three sections and each slice stained for an individual enzyme. The assay methods of Shaw and Prasad (1970) were used for EST, IDH and PGM, except that an agar overlay was used for PGM. The methods of Richardson eta/. (1986) were used for PGI, AC and MDH assays, and the method of Brown eta/. (1978) for 6PGD assays. Stained gels were fixed with 7% acetic acid, then rinsed in water, wrapped in plastic and photocopied for a permanent record. Isozyme banding patterns were scored immediately after staining. Cluster analysis was performed using the UPGMA method (Sneath and Sokal 1973) with Nei's unbiased genetic identity (Nei 1972) on the Biosys-1 program (Swofford and Selander 1981, Swofford 1989). Voucher specimens are held at the Division ofEntomology in Perth and will be lodged at either the National Herbarium of New South Wales, Sydney (NSW) or the Western Australian Herbarium (PERTH).

Results Six enzymes with a total of eleven putative loci were used in the analysis. PGI loci were clearly polymorphic within and among collections, but were excluded from the analysis because of multiple overlapping bands that could not be scored. The six enzymes used in the analysis demonstrated a diploid level of expression for the polymorphic loci examined and thus could be scored. The number of loci detected in each system was generaily consistent with those found in many other plant taxa, as reported by Gottlieb (1982), Kephart (1990) and Weeden and Wendel (1990). The observed electrophoretic banding patterns are depicted in Fig. 3. Variation was detected in eight ofthe eleven loci scored. The PGM-1, MDH-2 and EST-1 loci were monomorphic for all collections. Allozyme frequencies in the collections are shown in Table 3. At the 6-PGD-1 locus, all samples were homozygous slow except for all the Indian, three Namibian and two Kuwait samples, which were heterozygous. At the 6-PGD-2 locus, all samples were heterozygous. In several samples (Indian, three Namibian and two Kuwait samples), the heterozygotes had extra bands not normally expressed by a heterozygous diploid dimer. This was possibly due to heterodimers formed between the two loci. Heterozygotes at both loci frequently displayed unbalanced staining. The PGM system had simple banding patterns, but they could be interpreted in several different ways. The distance between the two bands in the most common PGM pattern indicated that there were two overlapping loci. At the PGM-1 locus all the samples were homozygous. The PGM-2 locus displayed mostly homozygous slow allozymes with fast alleles found in India, Kuwait 1,2,3 and Namibia 2. A rare fourth PGM-2 allele (D) was observed only in a Namibian collection from Gobabeb, at a low frequency of0.07. Isozymes 35

In the MDH system the most frequent three-banded pattern was interpreted as two homodimers with one interlocus heterodimer. At MDH-1 all samples were homozygous slow, except for Tiroll (Namibia) and Katherine 1 and 2 (Northern Territory) which were heterozygous. Unbalanced staining was observed in some of these heterozygotes. At the MDH-2 locus all samples were homozygous fast. In the IDH system the two loci were next to each other with the slow band ofiDH- 1 overlapping with the fast band ofiDH-2. This system displayed the most polymorphism with heterozygous and homozygous slow bands at locus one, and heterozygous and homozygous fast bands at locus two. Some asymmetry of staining was observed in the heterozygotes at both loci. Faint staining was seen in the fast allele at IDH-1 and the slow allele at IDH-2. The overlapping of the IDH-2 fast and IDH-1 slow alleles could have resulted in an intense central band, making the others appear faint in comparison. Alternatively, there may be weak expression of the faint alleles. At the IDH-1 locus many heterozygotes and homozygotes were observed, with 32 collections completely homozygous slow. The IDH-2 locus also displayed many heterozygotes and homozygotes, with 30 collections completely homozygous fast. The single AC locus displayed many heterozygotes and homozygotes, with only seven collections completely homozygous slow. The EST system was difficult to interpret, despite only having a small number of bands. It was assumed, as in the PGM system, that there was overlap between two loci. At the EST-1 locus all samples were homozygous. Most samples were homozygous for the slow allele at the EST-2locus, except India 1, 2, and 4, Kuwait 1 and 2, NT, and QLD which were homozygous for the fast allele. The dendrogram resulting from cluster analysis revealed four or possibly five major groups within the collections of burrs, with identity values ranging from 0. 94 to 0. 72 (Fig. 4). The largest and most closely related group (A) (similarity of0.94) was comprised of all the southern Australian, northern Western Australian, USA, Mediterranean, Iranian and two South African sites, plus three of the Namibian and two of the Kuwait sites. The second group (B) consisted of two Namibian (Thomasberg and Tiroll) collections (similarity of0.90). The Queensland and Northern Territory sites made up a distinct third group (C) (similarity of0.86). All four Indian and two of the Kuwait sites formed the fourth group (D) with the least similarity to all other collections (similarity of0.72). One Indian site (Kargil) possibly represents a fifth group. The percentage standard deviation of the best fitting dendrogram was 4.3%. Other methods of cluster analysis produced essentially the same results. The proportion of polymorphic loci (with a frequency of0.99 or less) per collection ranged from 0.091 to 0.455. The mean number of alleles per locus (including monomorphic and polymorphic alleles) was between 1.1 and 1. 5. The mean unbiased heterozygosity per locus was 0.091 to 0.364, and the expected heterozygosity per locus was 0.046 to 0.210 (Table 4). The mean heterozygosity between each group of sites as identified by isozyme analysis were compared using a one way ANOVA test (F = 5.9, df3,50, P< 0.01). Using a LSD (t) pairwise comparison of means, the group consisting of the Indian and two Kuwait collections were significantly different from all other groups, having a greater mean heterozygosity. Isozymes 36

Discussion Interpretation ofzymograms The putative loci identified for the PGM, MDH, IDH, AC and EST systems were possible to interpret as diploid phenotypes. The electrophoretic banding patterns observed in Indian and two Kuwait collections in the 6-PGD system were difficult to interpret because of extra bands compared with other sites, which overlapped between the two loci. These extra bands were thought to be heterodimers formed between the two loci. The patterns were consistently different from those observed in all other collections, which provides evidence that the Indian and two Kuwait collections form a genetically separate group. The extra bands in these samples indicate that they have a higher ploidy level than the rest ofthe collections examined. Qualitatively the PGI zymograms also support this observation. Further elucidation of the loci is not possible without data from known diploids and controlled crosses.

Indications oforigins The bulk of the Australian collections form an almost identical group with overseas collections from the Middle East, Mediterranean, South Africa, Namibia and the USA (group A similarity= 0.95). This indicates that the T. terrestris found in southern Australia and northern Western Australia is introduced. It is likely that group A represents one taxon, because identities for plant collections in the same taxon are often above 0.9 (Crawford 1990). The Queensland and Northern Territory collections (group C), however, have less genetic similarity with the others examined in this survey (similarity= 0.85), either within Australia or world-wide. This indicates that they are highly likely to be a native Australian species in the genus Tribulus. The Indian and two of the Kuwait sites (group D) also have little genetic similarity with the introduced Australian collections (similarity= 0.73). These regions therefore, are unlikely to be the origin of the weedy form ofT. terrestris in Australia. Similarly, the locations of Tiroll and Thomasberg (Namibia) (group B) are unlikely origins (similarity= 0.90). Initially it is not apparent which of the overseas locations within group A is the origin of the Australian weedy form ofT. terrestris. The plant is known to have been introduced into the USA from Europe before 1903 (Huffaker eta/. 1983). In South Africa T. terrestris is considered to be native; but it has been suggested that it was introduced from southern Europe before 1794 (Schweickerdt 1939). There are possibly both native and introduced forms ofT. terrestris in South Africa, a similar situation to Australia. Since the southern African, USA, Mediterranean, and many of the Middle East samples have a high genetic identity with the introduced T. terrestris in Australia, it appears that the weedy form ofT. terrestris originally came from the Mediterranean/Middle East region or southern Africa. The introduction into Australia could have been directly from the Mediterranean/Middle East or southern Africa, or as a secondary introduction via the USA or South Africa. Evidently there could have been multiple introductions from all or several of these locations. Although the introduced form in Australia is distributed over a wide geographical range it is possible that it could have been dispersed from one source. This is because the burrs are readily spread by in their feet or coats and by humans on rubber tyres and shoes. The burrs become embedded in the fleece of sheep and this is a likely means by which it has become spread around the world (Parsons and Cuthbertson 1992). Data from morphology (Scott and Morrison 1994) and chromosome numbers (Morrison and Scott 1994) also indicates similar origins for introduced T. terrestris in Australia, with possibly two separate introductions corresponding to chromosome numbers 24 and 36. Isozymes 37

Ploidy levels Many ploidy levels have been recorded in world-wide collections of Tribulus terrestris with diploid, tetraploid, hexaploid and octoploid forms being described (Malik 1966, Husain 1986) with the basic chromosome number of x = 6 (Porter 1968). It is therefore possible that some or all collections examined in this study are polyploid. The multiple allozyme banding patterns observed in most of the PGI zymograms and in the Indian and two Kuwait 6-PGD zymograms, indicate that there could be duplication or polyploidy occurring in these systems. Additionally, it is thought that polyploids can revert to a diploid level of expression (Ferris eta/. 1979). This reversion could have occurred in some of the 6PGD samples and in all PGM, MDH, IDH, ACO and EST samples. Conversely, multiple forms of an enzyme can co-migrate to the same position on a gel, thus a polyploid would have the same electrophoretic phenotype as a diploid. It has been reported that higher levels ofheterozygosity exist in both autopolyploids and in allopolyploids compared with their diploid progenitors (Soltis and Rieseberg 1986, Crawford 1990). In allopolyploids this is due to fixed heterozygosity and in autopolyploids, polysomic inheritance. Higher frequencies of heterozygotes than expected by the Hardy-Weinberg equilibrium present in all collections, also support the existence of polyploidy in T. terrestris (Table 4). The occurrence of enzyme multiplicity in the 6-PGD and PGI enzymes and unbalanced staining in 6-PGD, MDH and IDH heterozygotes are evidence of polysomic inheritance. This therefore indicates that these samples are more likely to be of autopolyploid rather than allopolyploid origin. Chromosome counts confirm that all the collections included in this analysis are polyploid (Morrison and Scott 1994). In addition, all Indian and three ofthe Kuwait samples are octoploid, which is the highest ploidy level recorded among this collection. This would explain the extra bands present in the 6-PGD zymograms for these locations.

Variability and breeding systems Moderate levels of variation in the number of alleles per locus and the proportion of polymorphic loci were observed, compared with values reported for Maybra spp. (Elisens and Crawford 1988) and for a range of selfing and outcrossing species (Gottlieb 1981). The mean heterozygosity for all collections however, was very high, ranging from 0.091 to 0.364, compared with that reported for a range ofselfers (mean heterozygosity 0.001) and outcrossers (mean heterozygosity 0.086) by Gottlieb (1981). This increased level of heterozygosity is likely to be due to polyploidy (Soltis and Riesesberg 1986, Crawford 1990, Kephart 1990). This effect is likely to have masked any reduction in heterozygosity that occurs in selfing collections (Gottlieb 1981 ). It is difficult therefore, to determine the mode of fertilisation from these elevated heterozygosity levels. India being part of the centre of origin of the Tribulus genus (Scott 1994), is likely to have collections with the maximum genetic diversity and hence higher levels of heterozygosity. The southern Australian collections ofT. terrestris included in group A, may have had multiple introductions which would explain the heterozygosity levels which are similar to the suspected native northern Australian group.

Implications for weed management in Australia Isozymes 38

In Australia, dried vine fruit is produced around Mildura in northern Victoria and southern New South Wales. It is evident from this study that the introduced weedy form ofT. terrestris is present in this region. Biological control using exotic agents can be considered, however the possible interactions with native Tribulus species in the area will need to be considered. Tribulus terrestris is well known for being toxic to livestock in Australia (Bourke 1984, Glastonbury eta/. 1984, Jacob and Peet 1987) and South Africa (Watt and Breyer-Brandwijk 1962, Kellerman and Coetzer 1984). It is of interest to note that three collections belonging to the introduced group (Walpeup in Victoria and Taldra and Wunkar in South Australia), were associated at the time of collection with sheep poisoning due to T. terrestris. This does not however, exclude the possibility that some of the native forms of Tribulus are also toxic to livestock. In central New South Wales sheep poisoning has been recorded in regions where both T. terrestris and T. micrococcus are abundant. Bourke (1987) observed however, that less severe and reversible symptoms of Tribulus staggers developed in sheep that had grazed on pastures dominated by T. micrococcus compared with those which had eaten T. terrestris.

Acknowledgments This study was supported by a grant from the Australian Dried Fruits Research and Development Council and made possible by the availability of facilities at the Department of Agriculture Western Australia, Weed Science Branch. We are especially grateful to collectors ofT. terrestris seed (Appendix 6). Technical help was provided by Kristy Hollis and Rod Randall. Dr Dane Panetta helped in establishing the project. J. R. Hosking and P. B. Yeoh commented on early drafts of the manuscript

References Bourke, C. A. (1984). Staggers in sheep associated with the ingestion of Tribu/us terrestris. Australian Veterinary Journal 61, 360-363. Bourke, C. A. (1987). Some taxonomic, agronomic and animal health aspects of Tribulus. Proceedings of the 8th Australian Weeds Conference, Sydney. pp. 182- 185. Brewer, G. J. and Singh, C. F. (1970). 'An introduction to isozyme techniques.' (Academic Press: New York.) Brown, A. H. D., Nevo, E., Zohary, N. E. and Dagan, 0. (1978). Genetic variation in natural populations ofwild barley (Hordeum spontaneum). Genetica 49, 97-108. Crawford, D. J. (1990). Enzyme electrophoresis and Plant Systematics. In ' Isozymes in Plant Biology.' (Eds D. E. Soltis and P. S. Soltis.) pp. 146-164. (Chapman and Hall: London.) Elisens, W. J. and Crawford, D. J. (1988). Genetic variation and differentiation in the genus Mabrya (Scrophulariaceae-Antirrhineae): Systematic and evolutionary inferences. American Journal ofBotany 75, 85-96. Ferris, S. D., Portnoy, S. L. and Whitt, G. S. (1979). The roles of speciation and divergence time in the loss of duplicate gene expression. Theoretical Population Biology 15, 114-139. Glastonbury, J. R. W., Doughty, F. R., Whitaker, S. J. and Sergeant, E. (1984). A syndrome of hepatogeous photosensitisation, resembling geeldikkop, in sheep grazing Tribulus terrestris. Australian Veterinary Journa/61, 314-316. Isozymes 39

Gottlieb, L. D. (1981). Electrophoretic evidence and plant populations. Progress in Phytochemistry 1, 1-46. Gottlieb, L. D. (1982). Conservation and duplication ofisozymes in plants. Science 216, 373-380. Huffaker, C. B., Hamai, J. and Nowierski, R. M. (1983). Biological control of puncturevine, Tribulus terrestris in California after twenty years of activity of introduced weevils. Entomophaga 28, 387-400. Husain, S. A. (1986). Cytotaxonomic studies in Tribulus from Pakistan. Kromosoma II 42, 1316-1329. Jacob, R. H. and Peet, R. L. (1987). Poisoning of sheep and goats by Tribulus terrestris (caltrop). Australian VeterinaryJournal64, 288-289. Kellerman, T. S. and Coetzer, J. A. W. (1984). Hepatogenous photosensitivity diseases in South Africa. Technical communication, Department of Agriculture, Republic of South Africa. (Government Printer: Pretoria.) Kephart, S. R. (1990). Starch gel electrophoresis of plant isozymes: a comparative analysis of techniques. American Journal ofBotany 77, 693-712. Malik, C. P. (1966). Corrected basic chromosome number and intraspecific polyploidy in Tribulus terrestris Linn. Chromosome Information Service 7, 7-8. Marshall, D. R. and Brown, A. H. D. (1975). Optimum sampling strategies in general conservation. In 'Genetic resources for today and tomorrow.' (Eds 0. H. Frankel and J. G. Hawkes.) pp. 53-80. (Cambridge University Press: Cambridge.) Marshall, D. R. and Weiss, P. W. (1982). Isozyme variation within and among Australian populatins of Emex spinosa (L.) Campd. Australian Journal of Biological Science 35, 327-332. Moran, G. F. and Marshall, D. R. (1978). Allozyme uniformity within and variation between races of the colonizing species Xanthium strumarium L. (Noogoora burr). Australian Journal ofBiological Science 31, 283-291. Morrison, S.M. and Scott, J. K. (1994). Variation in Australian and world-wide populations of Tribulus terrestris L. 2. Chromosome numbers. Nei, M. (1972). Genetic distance between populations. American Naturalist 106, 283- 293. Panetta, F. D. (1990). Isozyme variation in Australian and South African populations of Emex australis Steinh. Australian Journal ofBotany 38, 161-167. Panetta, F. D. and Carstairs, S. A. (1989). Isozymic discrimination oftropical Australian populations of mesquite (Prosopis spp.): implications for biological control. Weed Research 29, 157-165. Parsons, W. T. and Cuthbertson, E. G. (1992). 'Noxious weeds of Australia.' (Inkata Press: Melbourne.) Porter, D. M. (1968). The basic chromosome number in Tribulus (Zygophyllaceae). Wasmann Journal ofBiology 26, 5-6. Richardson, B. J., Haverstock, P. R. and Adams, M. (1986). 'Allozyme electrophoresis. A handbook for animal systematics and population studies.' (Academic Press: Sydney.) Schweickerdt, H. G. (1939). An account of the South African species of Tribulus Tourn. ex Linn. Bothalia 3, 157-178. Scott, J. K. (I 994). Pests and diseases of Tribulus terrestris and selection of biological control agents suitable for Australia. (Unpublished manuscript). Isozymes 40

Scott, J. K. and Morrison, S. M. (1994). Variation in Australian and world-wide populations of Tribulus terrestris L. 1. Burr morphology. (Unpublished manuscript). Shaw, C. R. and Prasad, R. (1970). Starch gel electrophoresis of enzymes- a compilation of recipes. Biochemical Genetics 4, 297-320. Sneath, P. H. A. and Sokal, R. R. (1973). 'Numerical Taxonomy.' (Freeman: San Francisco.) Soltis, D. E. and Rieseberg, L. H. (1986). Autopolyploidy in Tolmiea menziesii (Saxifragaceae): Genetic insights from enzyme electrophoresis. American Journal ofBotany 73, 310-318. Swofford, D. L. (1989). 'Biosys- 1, release 1.7.' (Illinois Natural History Survey: Champaign.) Swofford, D. L. and Selander, R. B. (1981). BIOSYS-1: a FORTRAN program for the comprehensive analysis of electrophoretic data in population genetics and systematics. Journal ofHeredity 72, 281-283. Watt, J.M. and Breyer-Brandwijk, M.G. (1962). Zygophyllaceae. In 'The Medicinal and Poisonous Plants of Southern and Eastern Africa.' pp. 1063-1076. (E. and S. Livingstone Ltd.: Edinburgh and London.) Weeden, N. F. and Wendel, J. F. (1990). Genetics ofplant isozymes. In 'Isozymes in Plant Biology.' (Eds D. E. Soltis and P. S. Soltis.) pp. 46-72. (Chapman and Hall: London.) Isozymes 41

Table 1. Collection sites of Tribulus te"estris s.L burrs

Location Site numbers Latitude; Longitude AUSTRALIA New South Wales Attunga 1 31 o 54'S; 150° 50'E Coonamble 2 30° 57'S; 148° 24'E Mudgee 3 32° 36'S; 149° 37'E Tamworth Agric. Res. Centre 4 31 o 09'S; 150° 59'E Northern Territory Darwin, Casuarina Beach 5 12° 21'S; 130° 52'E Katherine 1 6 14° 32'S; 132° 24'E Katherine 2 7 14° 28'S; 132° 16'E Queensland Cranbrook 8 19° 16'S; 146° 49'E Home Hill 9 19° 52'S; 147° 15'E Mingela Range 10 19° 53'S; 146° 38'E Sellheim 11 20° OO'S; 146° 25'E South Australia Ceduna 12 32° OS'S; 133° 4l'E Naracoorte 13 36° 58'S; 140° 45'E Taldra 14 34° 21'S; 140° 5I'E Wunkar 15 34° 29'S; 140° 18'E Victoria Merbein 16 34° 1O'S ; 142° 04'E Mildural 17 34° II'S; 142° IO'E Mildura 2 18 34° II'S; 142° lO'E Rutherglen 19 36° Ol'S; 146° 25'E Walpeup 20 35° OS'S; 142° 02'E Western Australia, North Broome 21 • 17° 58'S; 122° 14'E Carnarvon 22 24° 53'S; 113° 40'E Dongara 23 32° 57'S; 123° 25'E Moora 24 30° 39'S; 116° OO'E Nita Downs Station 25 l9° 05'S; 121 ° 41'E Western Australia, South Busselton 26 33° 39'S; 115° 20'E Goomailing 27 31° IS'S; 11 6° 50'E Lake Biddy 28 33° OO'S; 118° 56'E Moorine Rock 29 31° 19'S; 119° 04'E Wickepin 30 32° 47'S; 117° 30'E WORLD-WIDE Crete 31 35° 20'N; 25° OO'E Cyprus, Larnaca 32 34° 54'N; 33° 39'E India, Ladakh, Kargil 33 34° 31 'N; 76° 13'E India, Ladakh, Nubra River 34 34° 40'N; 77° 40'E* India, Kashmir, Pampur 35 34° 05'N; 74° 55'E* India, Kashmir, Pulwama 36 33° 55'N; 74° 55'E* Isozymes 42

Iran, Gaoh Sar 1 37 36° lO'N; 51 o 25'E* Iran, Gaoh Sar 2 38 36° 1O'N; 51 o 25'E* Iran, Karaj 1 39 35° 48'N; 50° 58'E Iran, Karaj 2 40 35° 48'N; 50° 58'E Kuwait, Agric. Exp. Station 41 29° 20'N; 48° OO'E* Kuwait, Al Qurain 42 29° 02'N; 48° 08'E* Kuwait, City 43 29° 20'N; 48° OO'E* Kuwait, Khaldiya 44 29° 20'N; 48° OO'E* Namibia, Gobabeb 1 45 23° 34'S; 15° 03'E Namibia, Gobabeb 2 46 23° 34'S; 15° 03'E Namibia, Rosh Pinah 47 28° OS'S; 16° 50'E* Namibia, Thomasberg 48 23° 10'S; 15° 31 'E Namibia, Tiroll 49 26° IS'S; 16° 45'E* S. Africa, Prince Albert 50 33° 13'S; 22° 03'E S. Africa,Rondebosch 51 33° 56'S; 18° 28'E USA, Kansas, Manhattan 52 39° 11 'N; 96° 35'W* USA, Prosser 53 46° 13'N; 119° 46'W USA, South Pemiscot 54 36° 05'N; 89° 50'W* * approximate location only Isozymes 43

Table 2. Enzymes resolved in T. terrestris s.l. H, histidine; L. lithium borate; T, tris-citrate buffer system.

Enzyme Abbreviation EC number Estimated Buffer No. of loci Aconitase AC 4.2.1.3 1 T Esterase EST 3.1.1.1 2 L Isocitrate dehydrogenase IDH 1.1.1.42 2 T Malate dehydrogenase MDH 1.1.1.37 2 T Phosphoglucoisomerase PGI 5.3.1.9 2 H Phosphoglucomutase PGM 2.7.5.1 2 H 6-phosphogluconate 6PGD 1.1.1.44 2 H dehydrogenase .Isozymes 44

Table 3. Allele frequencies for populations ofT. te"estris s.l. which exhibited isozyme variation Alleles are labelled (a,b,c,d*) in order of decreasing electorophoretic mobility. N =sample size. *except for allele din PGM-2 which is the fastest allele.

Population 6PGD-1 6PGD-2 PGM-2 MDH-1 N A B N A B c N A B c D N A B AUSTRALIA NSW, Attunga 29 0.00 1.00 29 0.50 0.50 0.00 29 0.00 1.00 0.00 0.00 29 0.00 1.00 NSW, Coonamble 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 NSW, Mudgee 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 NSW, Tamworth 28 0.00 1.00 28 0.50 0.50 0.00 28 0.00 1.00 0.00 0.00 28 0.00 1.00 NT, Darwin 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 NT,Katherine 1 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.50 0.50 NT, Katherine 2 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.17 0.83 QLD, Cranbrook 29 0.00 1.00 29 0.50 0.50 0.00 29 0.00 1.00 0.00 0.00 29 0.00 1.00 QLD, Home Hill 22 0.00 1.00 22 0.50 0.50 0.00 21 0.00 1.00 0.00 0.00 22 0.00 1.00 QLD, Mingela Range 18 0.00 1.00 18 0.50 0.50 0.00 18 0.00 1.00 0.00 0.00 18 0.00 1.00 QLD, Sellheim 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 SA, Ceduna 7 0.00 1.00 7 0.50 0.50 0.00 7 0.00 1.00 0.00 0.00 7 0.00 1.00 SA, Naracoorte 5 0.00 1.00 5 0.50 0.50 0.00 5 0.00 1.00 0.00 0.00 5 0.00 1.00 SA, Taldra 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 SA, Wunkar 19 0.00 1.00 19 0.50 0.50 0.00 19 0.00 1.00 0.00 0.00 19 0.00 1.00 VIC, Merbine 16 0.00 1.00 16 0.50 0.50 0.00 16 0.00 1.00 0.00 0.00 16 0.00 1.00 VIC, Mildura 1 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 VIC, Mildura2 20 0.00 1.00 20 0.50 0.50 0.00 21 0.00 1.00 0.00 0.00 21 0.00 1.00 VIC, Rutherglen 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 VIC, Walpeup 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 W A(N), Broome 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 W A(N), Carnarvon 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 WA(N), Dongara 4 0.00 1.00 4 0.50 0.50 0.00 4 0.00 1.00 0.00 0.00 4 0.00 1.00 .Isozymes 45

WA(N), Moora 4 0.00 1.00 4 0.50 0.50 0.00 4 0.00 1.00 0.00 0.00 4 0.00 1.00 WA(N), Nita Downs Stn. 24 0.00 1.00 24 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 WA(S), Busselton 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 WA(S), Goomalling 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 W A(S), Lake Biddy 14 0.00 1.00 14 0.50 0.50 0.00 14 0.00 1.00 0.00 0.00 14 0.00 1.00 WA(S), Moorine Rock 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 WA(S), Wickepin 24 0.00 1.00 24 0.50 0.50 0.00 24 0.00 1.00 0.00 0.00 24 0.00 1.00 WORLD-WIDE Crete 6 0.00 1.00 6 0.50 0.50 0.00 6 0.00 1.00 0.00 0.00 3 0.00 1.00 Cyprus 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 India, Kargil 14 0.50 0.50 14 0.50 0.00 0.50 14 0.00 0.00 1.00 0.00 14 0.00 1.00 India, Nubra 30 0.50 0.50 30 0.50 0.00 0.50 30 0.00 0.00 1.00 0.00 30 0.00 1.00 India, Pampur 6 0.50 0.50 6 0.50 0.00 0.50 6 0.00 0.00 1.00 0.00 6 0.00 1.00 India, Pulwama 30 0.50 0.50 30 0.50 0.00 0.50 30 0.00 0.00 1.00 0.00 30 0.00 1.00 Iran, Gaoh Sar 1 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 Iran, Gaoh Sar2 9 0.00 1.00 9 0.50 0.50 0.00 9 0.00 1.00 0.00 0.00 9 0.00 1.00 Iran, Karaj 1 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 Iran, Karaj 2 10 0.00 1.00 10 0.50 0.50 0.00 10 0.00 1.00 0.00 0.00 10 0.00 1.00 Kuwait, Agric. Exp. Stn 7 0.00 1.00 7 0.50 0.50 0.00 11 0.18 0.64 0.18 0.00 12 0.00 1.00 Kuwait, AI Qurain 25 0.50 0.50 25 0.50 0.00 0.50 28 0.00 0.00 1.00 0.00 30 0.00 1.00 Kuwait, City 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 Kuwait, Khaldiya 26 0.50 0.50 26 0.50 0.00 0.50 29 0.00 0.00 1.00 0.00 30 0.00 1.00 Namibia, Gobabeb 1 30 0.50 0.50 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 Namibia, Gobabeb 2 28 0.32 0.68 28 0.50 0.39 0.11 30 0.00 0.93 0.00 0.07 30 0.00 1.00 Namibia, Rosh Pinah 21 0.00 1.00 21 0.50 0.50 0.00 21 0.00 1.00 0.00 0.00 18 0.00 1.00 Namibia, Thomasberg 5 0.60 0.40 5 0.80 0.20 0.00 5 0.40 0.40 0.20 0.00 5 0.00 1.00 Namibia, Tiroll 30 0.00 1.00 30 0.50 0.50 0.00 30 0.90 0.10 0.00 0.00 30 0.03 0.97 S.Africa, Prince Albert 15 0.00 1.00 15 0.50 0.00 0.50 15 0.00 1.00 0.00 0.00 12 0.00 1.00 S.Africa, Rondebosch 30 0.00 1.00 30 0.50 0.00 0.50 30 0.00 1.00 0.00 0.00 30 0.00 1.00 USA, Kansas 25 0.00 1.00 25 0.50 0.50 0.00 25 0.00 1.00 0.00 0.00 25 0.00 1.00 USA, Prosser 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 USA S.Perniscot 30 0.00 1.00 30 0.50 0.50 0.00 30 0.00 1.00 0.00 0.00 30 0.00 1.00 .Isozymes 46

Table 3. Continued.

Population IDH-1 IDH-2 AC-1 EST-2 N A B N A B N A B c N A B AUSTRALIA NSW, Attunga 29 0.47 0.53 29 0.50 0.50 29 0.50 0.50 0.00 29 0.00 1.00 NSW, Coonamble 30 0.00 1.00 30 1.00 0.00 30 0.07 0.93 0.00 27 0.00 1.00 NSW, Mudgee 30 0.00 1.00 30 1.00 0.00 30 0.03 0.97 0.00 30 0.00 1.00 NSW, Tamworth 28 0.50 0.50 28 0.50 0.50 28 0.50 0.50 0.00 28 0.00 1.00 NT, Darwin 30 0.00 1.00 30 1.00 0.00 30 0.00 0.50 0.50 30 1.00 0.00 NT, Katherine 1 30 0.00 1.00 30 0.93 0.07 30 0.00 0.60 0.40 30 1.00 0.00 NT, Katherine 2 30 0.00 1.00 30 1.00 0.00 30 0.00 1.00 0.00 30 1.00 0.00 QLD, Cranbrook 29 0.00 1.00 29 1.00 0.00 29 0.00 0.74 0.26 29 1.00 0.00 QLD,HomeHill 22 0.00 1.00 22 1.00 0.00 20 0.00 0.43 0.57 19 1.00 0.00 QLD, Mingela Range 18 0.00 1.00 18 1.00 0.00 12 0.00 0.58 0.42 6 1.00 0.00 QLD, Sellheim 30 0.00 1.00 30. 1.00 0.00 30 0.00 0.75 0.25 30 1.00 0.00 SA, Ceduna 7 0.14 0.86 7 0.86 0.14 7 0.14 0.86 0.00 7 0.00 1.00 SA, Naracoorte 5 0.50 0.50 5 0.50 0.50 5 0.00 1.00 0.00 5 0.00 1.00 SA, Taldra 30 0.50 0.50 30 0.63 0.37 29 0.40 0.60 0.00 27 0.00 1.00 SA, Wunkar 16 0.03 0.97 16 0.97 0.03 15 0.13 0.87 0.00 13 0.00 1.00 VIC, Merbine 16 0.44 0.56 16 1.00 0.00 4 0.50 0.50 0.00 16 0.00 1.00 VIC, Rutherglen 30 0.00 1.00 30 1.00 0.00 28 0.21 0.79 0.00 30 0.00 1.00 VIC, Mildura1 30 0.05 0.95 30 1.00 0.00 30 0.33 0.67 0.00 30 0.00 1.00 VIC, Mildura 2 20 0.18 0.82 20 1.00 0.00 5 0.50 0.50 0.00 21 0.00 1.00 VIC, Walpeup 30 0.28 0.72 30 0.83 0.17 28 0.30 0.70 0.00 22 0.00 1.00 W A(N), Broome 30 0.00 1.00 30 1.00 0.00 30 0.17 0.83 0.00 25 0.00 1.00 W A(N), Carnarvon 30 0.00 1.00 30 1.00 0.00 30 0.02 0.98 0.00 30 0.00 1.00 W A(N), Dongara 4 0.00 1.00 4 1.00 0.00 4 0.00 1.00 0.00 4 0.00 1.00 W A(N), Moora 4 0.00 1.00 4 1.00 0.00 4 0.00 1.00 0.00 4 0.00 1.00 W A(N), Nita Downs Stn. 30 0.02 0.98 30 1.00 0.00 28 0.07 0.93 0.00 29 0.00 1.00 .Isozymes 47

W A(S), Busselton 29 0.50 0.50 29 0.50 0.50 20 0.50 0.50 0.00 21 0.00 1.00 WA(S), Goomalling 30 0.50 0.50 30 0.52 0.48 30 0.50 0.50 0.00 30 0.00 1.00 W A(S), Lake Biddy 13 0.50 0.50 13 0.50 0.50 14 0.50 0.50 0.00 14 0.00 1.00 W A(S), Moorine Rock 30 0.13 0.87 30 0.87 0.13 30 0.17 0.83 0.00 30 0.00 1.00 W A(S), Wickepin 24 0.19 0.81 24 0.75 0.25 20 0.18 0.82 0.00 24 0.00 1.00 WORLD-WIDE Crete 2 0.00 1.00 2 0.50 0.50 2 0.50 0.50 0.00 6 0.00 1.00 Cyprus 30 0.50 0.50 30 0.55 0.45 30 0.08 0.92 0.00 30 0.00 1.00 India, Kargil 14 0.00 1.00 14 0.50 0.50 5 0.40 0.60 0.00 6 0.00 1.00 India, Nubra 30 0.00 1.00 30 0.50 0.50 30 0.10 0.55 0.35 30 1.00 0.00 India, Pampur 6 0.00 1.00 6 0.50 0.50 6 0.42 0.58 0.00 4 1.00 0.00 India, Pulwama 30 0.00 1.00 30 0.53 0.47 24 0.42 0.58 0.00 22 1.00 0.00 Iran, Gaoh Sar 1 30 0.50 0.50 30 1.00 0.00 30 0.50 0.50 0.00 30 0.00 1.00 Iran, Gaoh Sar2 9 0.50 0.50 9 1.00 0.00 9 0.50 0.50 0.00 9 0.00 1.00 Iran, Karaj 1 30 0.50 0.50 30 1.00 0.00 30 0.50 0.50 0.00 30 0.00 1.00 Iran, Karaj 2 10 0.50 0.50 10 1.00 0.00 10 0.50 0.50 0.00 10 0.00 1.00 Kuwait, Agric. Exp. Stn 12 0.00 1.00 12. 0.50 0.50 12 0.38 0.62 0.00 7 0.00 1.00 Kuwait, AI Qurain 30 0.00 1.00 30 0.50 0.50 30 0.08 0.62 0.30 25 1.00 0.00 Kuwait, City 30 0.00 1.00 30 0.50 0.50 30 0.33 0.67 0.00 30 0.00 1.00 Kuwait, Khaldiya 29 0.00 1.00 29 0.50 0.50 30 0.07 0.76 0.17 26 1.00 0.00 Namibia, Gobabeb 1 30 0.00 1.00 30 1.00 0.00 30 0.05 0.95 0.00 30 0.00 1.00 Namibia, Gobabeb 2 30 0.00 1.00 30 1.00 0.00 30 0.02 0.85 0.13 30 0.00 1.00 Namibia, Rosh Pinah 21 0.00 1.00 21 1.00 0.00 21 0.02 0.98 0.00 18 0.00 1.00 Namibia, Thomasberg 5 0.20 0.80 5 1.00 0.00 5 0.20 0.80 0.00 5 0.00 1.00 Namibia, Tiroll 30 0.00 1.00 30 1.00 0.00 30 0.05 0.95 0.00 30 0.00 1.00 S.Africa, Prince Albert 15 0.00 1.00 15 1.00 0.00 14 0.00 1.00 0.00 12 0.00 1.00 S. Afiica, Rondebosch 30 0.00 1.00 30 1.00 0.00 30 0.00 1.00 0.00 30 0.00 1.00 USA, Kansas 25 0.00 1.00 25 1.00 0.00 23 0.00 1.00 0.00 23 0.00 1.00 USA, Prosser 30 0.00 1.00 30 1.00 0.00 30 0.23 0.77 0.00 30 0.00 1.00 USA, S.Perniscot 30 0.00 1.00 30 1.00 0.00 30 0.02 0.98 0.00 30 0.00 1.00 Isozymes 48

Table 4. Genetic variability of Tribulus terrrestris s.L at llloci Standard errors in ~arentheses. Mean sample Mean %of Mean unbiased Mean Collection size per number of polymorphic heterozygosity heterozygosity locus alleles per loci per locus Hardy-Weinberg locus ex~ected AUSTRALIA NSW, Attunga 29.0 (0.0) 1.4 (0.2) 36.4 0.357 (0.150) 0.185 (0.077) NSW, Coonamble 29.5 (0.4) 1.2 (0.1) 18.2 0.103 (0.091) 0.058 (0.047) NSW, Mudgee 30.0 (0.0) 1.2 (0.1) 18.2 0.097 (0.091) 0.052 (0.046) NSW, Tamworth 28.0 (0.0) 1.4 (0.2) 36.4 0.364 (0.152) 0.185 (0.077) NT, Darwin 30.0 (0.0) 1.2(0.1) 18.2 0.182 (0.122) 0.092 (0.062) NT, Katherine 1 30.0 (0.0) 1.4 (0.2) 36.4 0.267 (0.131) 0.148 (0.069) NT, Katherine 2 30.0 (0.0) 1.2 (0.1) 18.2 0.121 (0.093) 0.072 (0.051) QLD, Cranbrook 29.0 (0.0) 1.2 (0.1) 18.2 0.138 (0.098) 0.082 (0.055) QLD, Home Hill 21.1(0.4) 1.3 (0.1) 27.3 0.168 (0.113) 0.130 (0.068) QLD, Mingela 15.3 (1.5) 1.2 (0.1) 18.2 0.167 (0.112) 0.093 (0.062) QLD, Sellheim 30.0 (0.0) 1.2 (0.1) 18.2 0.136 (0.097) 0.081 (0.055) SA, Ceduna 7.0 (0.0) 1.4 (0.2) 36.4 0.169 (0.092) 0.121 (0.055) SA, Naracoorte 5.0 (0.0) 1.3 (0.1) 27.3 0.273 (0.141) 0.152 (0.078) SA, Taldra 29.4 (0.4) 1.4 (0.2) 36.4 0.321 (0.136) 0.180 (0.075) SA, Wunkar 17.0 (0.8) 1.4 (0.2) 36.4 0.127 (0.091) 0.080 (0.048) VIC, Merbine 14.9 (1.1) 1.3 (0.1) 27.3 0.261 (0.135) 0.145 (0.075) VIC, Mildura 1 30.0 (0.0) 1.3 (0.1) 27.3 0.161 (0.103) 0.096 (0.058) VIC, Mildura 2 19.2 (1.4) 1.3 (0.1) 27.3 0.214 (0.121) 0.124 (0.067) VIC, Rutherglen 29.8 (0.2) 1.2 (0.1) 18.2 0.130 (0.095) 0.077 (0.053) VIC, Walpeup 28.4 (1.0) 1.5 (0.2) 36.4 0.228 (0.106) 0.149 (0.064) WA, Broome 29.1 (0.6) 1.2 (0.1) 18.2 0.121 (0.093) 0.072 (0.051) W A, Carnarvon 30.0 (0.0) 1.1 (0.1) 9.1 0.091 (0.091) 0.046 (0.046) WA, Dongara 4.0 (0.0) 1.1 (0.1) 9.1 0.091 (0.091) 0.052 (0.052) WA, Moora 4.0 (0.0) 1.1 (0.1) 9.1 0.091 (0.091) 0.052 (0.052) W A, Nita Downs 28.5 (0.7) 1.3(0.1) 27.3 0.107 (0.090) 0.062 (0.047) W A, Busselton 27.3 (1.3) 1.4 (0.2) 36.4 0.364 (0.152) 0.185 (0.078) W A, Goomalling 30.0 (0.0) 1.4 (0.2) 36.4 0.361 (0.151) 0.185 (0.077) W A, Lake Biddy 13.8 (0.1) 1.4 (0.2) 36.4 0.364 (0.152) 0.189 (0.079) W A, Moorine Rk 30.0 (0.0) 1.4 (0.2) 36.4 0.170 (0.092) 0.115 (0.053) W A, Wickepin 23.6 (0.4) 1.4 (0.2) 36.4 0.202 (0.098) 0.136 (0.059) WORLD-WIDE Crete 4.4 (0.6) 1.3(0.1) 27.3 0.273 (0.141) 0.171 (0.089) Cyprus, Lamaca 30.0 (0.0) 1.4 (0.2) 36.4 0.279 (0.134) 0.152 (0.070) India, Kargil 11.7 ( 1.2) 1.4 (0.2) 36.4 0.345 (0.145) 0.190 (0.079) India, Nubra 30 (0.0) 1.5 (0.2) 36.4 0.355 (0.149) 0.191 (0.080) India, Pampur 5.7 (0.2) 1.4 (0.2) 36.4 0.348 (0.146) 0.197 (0.082) India, Pulwama 28.0 (1.0) 1.4 (0.2) 36.4 0.342 (0.144) 0.184 (0.077) Iran, Gaoh Sar 1 30.0 (0.0) 1.3(0.1) 27.3 0.273 (0.141) 0.139 (0.072) Iran, Gaoh Sar 2 29.0 (0.0) 1.3(0.1) 27.3 0.273 (0.141) 0.144 (0.075) Iran, Karaj 1 10.0 (0.0) 1.3(0.1) 27.3 0.273 (0.141) 0.144 (0.074) Iran, Karaj 2 22.1 (4.1) 1.3(0.1) 27.3 0.273 (0.141) 0.183 (0.102) Kuwait, Ag Stn 10.0 (0.7) 1.5 (0.2) 36.4 0.250 (0.131) 0.191 (0.080) Isozymes 49

Kuwait, AI Qurain 27.8 (0.7) 1.5 (0.2) 36.4 0.342 (0.145) . 0.187 (0.078) Kuwait, City 30.0 (0.0) 1.3(0.1) 27.3 0.242 (0.128) 0.134 (0.069) Kuwait, Khaldiya 28.2 (0.5) 1.5 (0.2) 36.4 0.315 (0.139) 0.174 (0.074) Namibia, Gob 1 30.0 (0.0) 1.3 (0.1) 27.3 0.191 (0.121) 0.101 (0.061) Namibia, Gob 2 29.6 (0.2) 1.5 (0.2) 36.4 0.177 (0.103) 0.130 (0.064) Namibia, R.Pinah 19.9 (0.5) 1.2 (0.1) 18.2 0.095 (0.091) 0.051 (0.046) Namibia, T'berg 5.0 (0.0) 1.5 (0.2) 45.5 0.182 (0.083) 0.210 (0.079) Namibia, Tiroll 30.0 (0.0) 1.4 (0.2) 36.4 0.106 (0.090) 0.078 (0.047) S. Africa, P.Albert 13 .8 (0.4) 1.1 (0.1) 9.1 0.091 (0.091) 0.047 (0.047) S. Africa, R'bosch 30.0 (0.0) 1.1 (0.1) 9.1 0.091 (0.091) 0.046 (0.046) USA, Kansas 24.5 (0.3) 1.1 (0.1) 9.1 0.091 (0.091) 0.046 (0.046) USA, Prosser 30.0 (0.0) 1.2 (0.1) 18.2 0.133 (0.096) 0.079 (0.054) USA, S. Pemiscot 30.0 (0.0) 1.2 (0.1) 18.2 0.094 (0.091) 0.049 (0.046) 6-7

25

Vl 0- ~ 3 0> Vl

3 2 1 24 4 27 3 ~~028

Vl \] 0

Fig. 1. Distribution of Australian populations of Tribulus terrestris s. I. surveyed for isozvme v~ri::ttinn . .() ~ Q, . ~.o ~0~

D

7383940 535254,,, ;x. '

(I) 0- ~ 3 (D (I) ••o••

<>

o• .. 0~

0 ;.~. 0 48 \J""' ~o'•• •• 45

.. o 46 1 49 \--50 • Ll-7 51 • [} c;::} D -Vl Distribution of Tribulus terrestris

Fig. 2. Distribution of world populations of Tribulus terrestris s.l. surveyed for ;c,..,...,..,...... o • , ...... : ..... :-- Isozymes 52

PGM 6PGD PGI

Ll----- L1 { L2- {----= Ll =----- BBBBBBBB- {----- L2 ------{~==~ BB CCAA 00 - -- L2 =---- BB BB AB BB AB --- AB AAAB ACAC ---- NOT SCORED

MDH IDH AC EST u{ = ll{ == u{- u-- -- l2 ------== l2- L2{---- - .... BB AB BC AAAA BBAB BB AB AB BB BB AA AAAA AAABAAM

Fig. 3. Representative illustrations ofbanding patterns for seven enzymes in Tribulus terrestris s. J. Isozymes 53

Similarity Chromosome 0.6 0.8 1.0 number

SA. CedLna WA. Moora * NSW. MJdgee * VIC. Ml

Fig. 4. Dendrogram showing relationships among Tribulus terrestris s.l. populations based on isozyme similarity. Biological control 54

Pests and diseases of Tribulus terrestris and selection of biological control agents suitable for Australia

J. K. Scott

CSIRO Division ofEntomology, Private Bag, P.O. Wembley, W.A. 6014, Australia.

Summary A review of the world literature found 67 records of and fungi that attack T. te"estris. The available information indicates that many of the potential biological control organisms have host ranges extending to species of Tribulus other than T. terrestris. Australia has at least a dozen native Tribulus species mostly found in the dry center and north west and few species are found in Mediterranean climates where control ofT. te"estris is desired. The conservation of native Tribulus species will need to be considered if biological means are used for control. The options include 1. finding agents with host ranges restricted to T. te"estris. Examples of organisms that could be studied include the weevil Microlarinus humeralis and the fungus Peronospora tribuli; 2. finding agents with biologies that restrict their distribution to Mediterranean climates and accepting the risk to populations of the few native Tribulus species, all of which are also found in dryer climates. New biological control agents with these characteristics may be found in north African and western Mediterranean areas that have not been previouly surveyed; 3. accepting the likelihood of damage to native species irrespective of where they are found. The known successful agents, Microlarinus lareynii and M. lypriformis, could be released in Australia with the acceptance (currently unlikely) of the risk to native species; or 4. not attempting biological control. Further research to narrow these options include surveys for biological control agents and quantification of the benefits and costs of introduced and native Tribulus.

Introduction Tribulus terrestris L. is a weed that has been successfully controlled by biological means. The first introductions were of the weevils, Microlarinus lareynii (Jacq. de Val) and M lypriformis (YIollaston) from Italy which were released in 1961 in the USA. From there they were distributed to Hawaii in 1962 and Canada in 1986 (Julien 1993). The weevils were also released against Tribulus cistoides L. on the islands of St. Kitts, Hawaii, Papua New Guinea and Nevis between 1962 and 1968 (Julien 1993). Twenty years after the first releases, the weevils were shown to significantly reduce seed production and growth ofT. terrestris (Huffaker eta/. 1983). Surveys for potential biological control agents have been undertaken in South Africa and Namibia (Kluge 1975), in Bangalore and nearby areas in South India (Sankaran and Ramaseshiah 1981 ), in areas of the Rajasthan desert and near Bangalore and New Delhi, India and coastal and interior areas of southern France and Italy (Andres and Angalet 1963), and in Israel (Gerling and Kugler 1973a, b). Scott (1990) and Shepherd (1990) reviewed the potential ofbiological control against T. terrestris in Australia and concluded that the question of origins should be resolved first. It now Biological control 55

seems clear that T. terrestris in southern Australia is of exotic origin and that the source area is likely to be the Mediterranean/west Asia region, although a southern Mrican origin is not excluded (Morrison and Scott 1994). This allows a re-examination of where searches for biological control agents should be made and whether there are potential biological control agents known which could be introduced into Australia. This report adopts three approaches to selecting suitable biological control agents: 1. The literature was examined to records of organisms associated with T. terrestris; 2. Climates ofT. terrestris infested areas in Australia were matched with climates elsewhere and compared with the regions that have been searched for biological control agents and; 3. the world distribution of high Tribulus species density was collated to identify regions likely to have a correspondingly high density of Tribulus associated organisms and to examine the extent of native Australian Tribulus species overlap with the distribution ofT. terrestris.

Methods Organisms associated with T. terrestris Abstracts for the scientific literature were surveyed for arthropods and pathogens found attacking T. terrestris. References were accessed on-line from CAB abstracts, Biosis, Agricola and Life Sciences Collection. References earlier than the on-line sources (generally pre 1972) were sought in Review of Applied Entomology A Vol1- 59 (1913-1971), Biological Abstracts Vol. 1 -49 (1927-1968) and Entomology Abstracts Vol. 1 - 9 (1969- 1978). References to plant pathogens were sought in Biological Abstracts, Review of Applied Mycology/Review ofPlant Pathology Vol. 1 - 51 (1922- 1971), Index ofFungi Vol1-6(4) (1940- 1992). References where T. terrestris was fed artificially to were excluded. Information on the host range was obtained either from the cited reference or from texts on pest species.

CLIMEX selection ofsurvey areas Climate stations in dried vine fruit production areas of Australia were matched with world climate stations contained in the CLIMEX data base (Sutherst and Maywald 1985).

World-wide Tribulus species density National or regional floras were examined to determine the number of Tribulus species found per country throughout the world. From the numbers of species per country or region, species isohytes were estimated for the world. The survey covered Africa and Eurasia, but not the Americas where Tribulus is absent as a native species. The survey included 26 flora treatments (references available from author). The Australian Tribulus species density was estimated from herbarium records (Hosking and Scott unpublished) and Hnatiuk (1990).

Results and Discussion Organisms associated with Tribulus species The literature reports 55 and 12 pathogen species associated with the weed (Table 1). Twenty six ofthe arthropods are known polyphagous species and can be excluded from further consideration. The remainder have unknown host ranges or have only been reported from Tribulus species including T. terrestris. Only Microlarinus lareynii and Microlarinus lypriformis have been extensively tested for host specificity. Biological control 56

Other Microlarinus species have been collected from Tribulus, but little is known of their biology although it seems that the genus is restricted to the Zygophyllaceae. Kluge (1975) carried out preliminary tests on two species. He reared the , Prodotis stolida from South Africa, on T. terrestris, but flax was an unsuitable host even though the moth has been reported as a minor pest of this crop in India and Russia. He also studied the bug, Deroplax sp. which has a host range that extends to at least the genus Tribulus (Kluge 1975). Other insects or mites that appear to have host ranges restricted to the genus Tribulus include the , subditalis and Ephysteris subdiminutella and the mite Eriophyes tribu/i. A mite causing identical damage to that caused by E. tribuli on T. terrestris has now been found in Australia where it is widespread and attacks Tribulus species including T. terrestris (Hosking and Scott, unpublished). Previously, Scott (1990) considered the mite to be the most promising candidate for biological control in Australia. Little is known ofthe host range of most of the pathogen species listed in Table 1. Strains ofColletotrichium sp. have been developed as mycoherbicides (Templeton and Greaves 1984) and other of the pathogens, for example, downy mildew, Peronospora tribu/i, Phyllostricta and Xanthomonas species, may have potential to be used as a classical biological control agents or have mycoherbicide potential.

Selection ofsurvey areas suitable for Australian conditions With the wide distribution ofT. terrestris, one approach is to identify areas of the world that closely resemble the region in Australia where control is desired. One area is the Sunraysia and Mallee regions ofVictoria/N.S.W. and South Australia where caltrop affects the dried fruit industry and causes sheep mortality. Five climate stations, which are included in the CLIMEX database (Sutherst and Maywald 1985), were chosen to represent the region (Table 2). These locations are very similar in climate to each other as any value greater than 0.8 represents a very high level of climatic similarity. These sites were compared with all climate stations included in the world data base in CLIMEX (Table 3). Dryer areas of the south western and southern Cape Province of South Africa are represented by five sites. Kluge (1975) collected insects from Stellenbosch which is between 93 and 310 km from these sites, but wetter and in a different biome (Rutherford and Westfall 1986) from that ofthe match sites in Table 3. However, the survey covered a number of regions and it is likely that the same species found by Kluge (1975) would be found in a survey of the area described by the sites in Table 3. A short survey over one season would be necessary to confirm whether this is the case. The second group, which consists of four sites, is in Spain. This area has not been examined for biological control agents for T. terrestris. The third area includes Algeria and Tunisa, neither of which has been examined for biological control agents forT. terrestris. None of the sites identified in Table 3 are found near to the centres of diversity of the genus Tribulus which is in eastern Mediterranean/West Asia region (Figure 1).

World-wide Tribulus species density Two regions had high Tribulus species density, north east Africa with south west Asia and secondly, central and north west Australia. Southern Africa also has a high number of species. A senario to explain the present distribution is that the genus Tribulus originally evolved in Gondwanaland. Following the break-up of Gondwanaland, species remained and continued to evolve in Australia, southern Africa and India, which drifted north to found the diversity now in north west Africa and south west Biological control 57

Asia. The organisms associated with Tribulus are shared between two of the three regions, possibly indicating a later evolution of the association with Tribulus. If this were the case, then the three centers of diversity have been separated for a long time and have had separate evolution of associated organisms, some of which may be sufficiently host specific to introduce into a new region like Australia.

Conclusions The level of host specificity is likely to be a problem since Australia is a centre of species diversity in the genus Tribulus (Figure 1) and many native species may be subject to damage by biological control agents. The native species are mostly restricted to the central and north west of the country and are largely absent from regions of Mediterranean climate. This suggests that biological control could be implemented by selecting only those agents restricted to Mediterranean climates, even though their host range might include species of Tribulus other than T. terrestris. An example could be Microlarinus humeralis, which has been reported from Israel where it attacks both stems and seeds (Gerling and Kugler 1973a, b). In conclusion, the search for agents should concentrate on regions ofMediterranean climate. Alternatively, known successful agents could be released in Australia only with the acceptance of the likelihood of damage to native species. Not attempting biological control is also an option. Further information on the costs and benefits of native and introduced Tribulus species will help authorities in Australia decide between the options for control.

References Allen, T. C., McMorran, J.P. and Locatelli, E. A. (1983). Isolation oftomato spotted wilt virus from hydrangea and four weed species. Plant Disease 67, 429-431. Not seen. Taken from Review ofPlant Pathology 62, 318. 1983, abstract no. 3466. Andres, L. A. and Angalet, G. W. (1963). Notes on the ecology and host specificity of Microlarinus lareynii and M lypriformis (Coleoptera: Curculionidae) and the biological control of puncture vine, Tribulus terrestris. Journal ofEconomic Entomology 56, 333-340. Aziz, S. A. and Rizvi, S. K. A. (1967). New record of damage to certain important vegetable and medicinal plants by Oxya velox Fabr. (Orthoptera: Acrididae) in Aligarh, U.P. Labdev. J. Sci. Tech. 5, 341. Not seen. Taken from Review ofApplied Entomology 58, abstract 2980. Bhatia, D. R. (1940). Observations on the biology of the desert locust (Schistocera gregaria Forsk.) in Sind-Rajputana desert area. I. The prefered food plants ofthe locust. Indian Journal ofEntomology 2, 187-192. Not seen, Taken from Review of Applied Entomology A. 29, 519. 1941 abstract no. xx. Bondarzeva-Montverde (1945). Not. Syst. Sect. Crypt. Inst. Bot. Acad. Sci. U.S.S.R. 5, p 160. Not seen. Taken from Index ofFungi 2, 422. Bradbury, J. F. (1986). 'Guide to plant pathogenic bacteria'. (CAB International, Slough). Brisley, H. R. (1924). Blister beetle- serious pest in northern Arizona. Journal of Economic Entomology 17, 343. Common, I. F. B. (1964). 'Australian butterflies'. (Jacaranda Press, Brisbane). Common, I. F. B. (1990). 'Moths of Australia'. (Melbourne University Press, Melbourne). Eastop, V. F. and Raccah, B. (1988). Aphid and host plant species in the Arava valley oflsrael: epidemiological aspects. Phytoparasitica 16, 23 -32. Biological control 58

Farr, D. F., Bills, G. F., Chamuris, G. P. and Rossman, A Y. (1989). 'Fungi on plants and plant products in the United States'. (APS Press, St Paul). Gerling, D. and Kugler, J. (1973a). An examination ofthe possibilities for biological control of some weeds in Israel. Phytoparasitica 1, 80. Gerling, D. and Kugler, J. (1973b). Final Technical Report, evaluation of enemies of noxious plants in Israel as potential agents for the biological control of weeds. Tel Aviv University, 241 pp. Gorter, G. J. M. A (1981). Index of pathogens, II, and the diseases they cause in wild growing plants in South Africa. Science Bulletin 398, 84pp. Hnatiuk, R. J. (1990). Census of Australian vascular plants. Australian Flora and Fauna Series 11, 1-650. Horie, Y and Udagawa, S. I. (1990). New or interesting Chaetomium spp. from herbal drugs. Transactions of the Mycological Society ofJapan 31, 249-258. Not seen. Taken from Biosis 90107945. Horie (1978). Transactions of the Mycological Society ofJapan 19, 313. Not seen. Taken from Index ofFungi 4, 588. Huffaker, C. B., Hamai, J. and Nowierski, R. M. (1983). Biological control of puncturevine, Tribulus terrestris in California after twenty years of activity of introduced weevils. Entomophaga 28, 387-400. Jooste W. J. (1975). A new species of Drechslera on Tribulus terrestris. Bothalia 11, 511-513. Julien, M. H. (1993). 'Biological control ofweeds a world catalogue of agents and their target weeds'. 3rd Edition. (CAB International, Wallingford). Kluge, R. L. (1975). Observations on insects associated with Tribulus terrestris L. in southern Africa. Unpublished MSc Thesis, University ofPretoria, 128 pp. Lodos, N. (1971). Microlarinus lareynii andM lypriformis (Coleoptera Curculionidae): two useful species of insect for the biological control of weeds and occurring in our country on Tribulus terrestris L. (iron thorn). Ege Universitesi Ziraat Fakultesi Dergisi 8, 55-74. Not seen. Taken from Review ofApplied Entomology A 61, 188. 1973, abstract no. 726. Morrison, S. M. and Scott, J. K. (1994). Variation in Australian and World-wide populations of Tribulus terrestris L. 3. Isozyme analysis. (unpublished manuscript). Narayanan, E. S., Subba Rao, B. R. and Sangwan, H. S. (1960). Host and parasite specificity in certain insects. Indian Journal ofEntomology 20, 306-307. Oudemans, C. A J. A (1921). 'Enumeratio systematica fungorum'. Vol 3. Pathak, P. S. (1968). Biological control of Tribulus terrestris by an insect of Hemiptera. Proceedings of the Symposium on Recent Advances in Tropical Ecology pp. 697-701 Pemberton, R. W. and Hoover, E. M. (1980). Insects associated with wild plants in Europe and the Middle East. Biological control ofweeds surveys. U.S. Department ofAgricultur e Miscellaneous Publication 1382, 1-33. Ramaseshiah, G. (1976). Survey for natural enemies of Tribulus terrestris. Trinidad, Commonwealth Institute ofBiological Control. Report ofwork carried out during 1975. p 46. Not seen. Taken from Weed Abstracts 27, xx. 19xx, abstract no. 2356. Riaz, M. (1974). A new host (Tribulus terrestris L.) for Colletotrichum dematium in Pakistan. Pakistan Journal of Science and Industrial Research 17, 89-90. Roffey, J. and Popov, G. (1968). Environmental and behavioural processes in a desert locust outbreak. Nature 219, 446-450. Biological control 59

Rutherford, M. C. and Westfall. R. H. (1986). Biomes of southern Mrica- an objective categorization. Memoirs of the Botanical Survey of South Africa 54, 1-98. Saccardo, P. A. (1888). 'Sylloge Fungorum'. Vol. 7. Sankaran, T. and Ramaseshiah, G. (1981). Studies on some natural enemies of puncturevine Tribulus terrestris occurring in Kamataka State, India. Proceedings of the V Interational Symposium on Biological Control ofWeeds. 153-160. Scott, J. K. (1990). Tribulus terrestris L. (Zygophyllaceae) in Southern Afiica: An outline ofbiology and potential biological control agents for Australia. Plant Protection Quarterly 5, 103-106. Shepherd, R. C. H. (1990). Prospects for the biological control of Tribulus terrestris L., and the possible use of the weevils, Microlarinus spp., as biological control agents. Victorian Department of Conservation and Environment Research Report 7, 1-21. Singh Sandhu, G. (1975). Seasonal movement ofthe painted bug, Bagrada cruciferarum Kirkaldy from cruciferous plants to graminaceous plants and its occurrence as a serious pest of maize, sorghum and pearl millet during spring in the Punjab. Indian Journal ofEntomology 37, 215-217. Not seen. Taken from Review ofApplied Entomology A 66, 538, 1978, abstract no. 4339. Squires, V. R. (1965). A note on Aristotelia sp. (Lep., Gelechiidae) attacking Tribulus terrestris L. Journal of the Entomological Society ofAustralia (N.S. W.) 2, 43-44. Srinivasan, M. C. and Patel, M. K. (1956). Three undescribed species of Xanthomonas. Current Science 25, 366-367. Sutherst, R. W. and Maywald, G. F. (1985). A computerised system for matching climates in ecology. Agriculture, Ecosystems and Environment 13, 281-299. Templeton, G. E. and Greaves, M. P. (1984). Biological control ofweeds with fungal pathogens. Tropical Pest Management 30, 333-338. Thomas, W. D. (1949). Studies on the host range of the Colorado red-node virus. J. Colo. - Wyo. Acad Sci. 4, 40. Not seen, Review ofApplied Mycology 29, p. 488. Varma, B. K. and Sharma, T. R. (1967). Studies on factors affecting population density and phase of desert locust in India. Pl. Prot. Bull. New Delhi 18, 9-24. Not seen. Taken from Review ofApplied Entomology B 59, abstract 63 . Verma, S. K. (1983). Host plants of Amsacta.moorei Butler in the Rajasthan Desert. Bulletin ofEntmology 24, 49-50. Not seen. Taken from Review ofApplied Entomology A 71,861. 1983 abstract no. 7415. Woodruff, R. E. (1972). New United States record- a sugarcane weevil (Nicentrus saccharinus) - Florida. Cooperative Economic Insect Report 22, 431 . Not seen. Taken from Review ofApplied Entomology A 61 , 509. 1973, abstract no. 1906. Biological control 60

Table 1. Arthropods and pathogens recorded from Tribulus terrestris. Species with unknown host ranges are indicated by a question mark, except for those species which have a host range limited to "Probably Tribulus sp." based on the biology of related species.

Family Species Country (Reference Host range Mites Eriophyidae Eriophyes cf tribuli Keifer Australia (Hosking and Scott Unpubl.) Tribulus spp. Eriophyidae Eriophyes tribu/i Keifer India (Sankaran and Ramaseshiah 1981) Probably Tribulus sp. Tenuipalpidae Tenuipalpus sp. India (Sankaran and Ramaseshiah 1981) ? Insects Curculionidae Baris sp. India (Sankaran and Ramaseshiah 1981) ? Curculionidae Microlarinus angustulus Mshl. India (Sankaran and Ramaseshiah 1981) Probably Tribulus sp. Curculionidae Microlarinus humeralis? Toum. Israel (Gerling and Kugler (1973b) Probably Tribulus sp. Curculionidae Microlarinus lareynii (Jacq. du Val) India (Narayanan et al. 1960, Pemberton and Hoover 1980), Tribulus spp. Italy (Andres and Angalet 1963), Turkey (Lodos 1971) Curculionidae Microlarinus lypriformis (Wollaston) India (Narayanan et al. 1960, Andres and Angalet 1963), Italy Tribulus spp. (Pemberton and Hoover 1980), South Africa (Kluge 1975), Turkey (Lodos 1971) Curculionidae Microlarinus pilosus Gyll. South Africa, (Kluge 1975) Probably Tribulus sp . Curculionidae Microlarinus rhinocylloides Hch India (Sankaran and Ramaseshiah 1981) Uzbekistan ( ref) Probably Tribulus sp. Curculionidae Microlarinus spp. (two) Namibia (Kluge 1975) Probably Tribulus sp. Curculionidae Nicentrus saccharinus Mshl. USA, Florida (Woodruff 1972) Polyphagous Curculionidae Protostrophus amplicollis F. South Africa (Kluge 1975) Polyphagous Meloidae (2 unidentified species) South Africa (Kluge 1975) ? Meloidae Epicauta corvina Lee. U.S.A.,Arizona (Brisley 1924) Polyphagous Melyridae Astylus atromaculatus Blanch South Africa (Kluge 1975) Polyphagous Orthoperidae Arthrolips sp. Italy (Pemberton and Hoover 1980) ? Aphidae Aphis craccivora Koch Namibia, South Africa (Kluge 1975), Israel (Eastop and Polyphagous Raccah 1988) Biological control 61

Aphididae Myzus persicae (Sulz.) Israel (Eastep and Raccah 1988) Polyphagous Coreidae Cletus binotulatus Stal South Africa (Kluge 1975) ? Lygaeidae Lygus sp. South Africa (Kluge 1975) ? Lygaeidae Naphius apicalis (Dalla) Namibia, South Africa (Kluge 1975) Polyphagous Lygaeidae Oxycarenus hyalinipennis (Costa) South Africa (Kluge 1975) Polyphagous Lygaeidae Paramius gracilis Ramb. South Africa (Kluge 1975) ? Lygaeidae Spilostethus pandurus var. elegans (Wolff) South Africa (Kluge 1975) ? Pentatomidae Bagrada cruciferarum Kirkaldy India (Singh Sandhu 1975) Polyphagous Pentatomidae Eupododus sp. Namibia, South Africa (Kluge 1975) ? Pentatomidae Nezara viridula var. smaragdula F. South Africa (Kluge 1975) Polyphagous Pentatomidae Piezodurus purus Stal Namibia, South Africa (Kluge 1975) Polyphagous Pentatomidae Poecilocoris sp. India (Pathak 1968) ? Pseudococcidae Ferrisia virgata (Ckll.) India ( Sankaran and Ramaseshiah 1981) Polyphagous Scutelleridae Deroplax sp. Namibia, South Mrica (Kluge 1975) Tribulus spp. Scutelleridae Xerobia sculpturata Stal Namibia, South Africa (Kluge 1975) ? Arctiidae Amsacta moorei Butler India (Verma 1983) Polyphagous Gelechiidae Dichomeris sp. India (Sankaran and Ramaseshiah 1981) ? Gelechiidae Ephysteris subdiminutella ferritincta Australia (Squires 1965, Common 1990) Probably Tribulus sp. (Tum.) Gelechiidae Ephysteris subdiminutella Stn. India (Sankaran and Ramaseshiah 1981) ? Lycaenidae Zizeeria knysia karsandra Moore Australia (Common 1964) T. terrestris Lycaenidae Zizeeria knysna Trim. Nambia (Kluge 1975) Polyphagous Lycaenidae Zizeeria maha Kollar India (Sankaran and Ramaseshiah 1981) Polyphagous Noctuidae Achaea catella Guen. South Africa (Kluge 1975) Polyphagous Noctuidae Agrotis segetum (Denis & Schiffennuller) South Africa (Kluge 1975) Polyphagous Noctuidae Agrotis spinifera Hb. India (Sankaran and Ramaseshiah 1981) Polyphagous Noctuidae Heliothis armigera var. armigera (Hubn.) South Africa (Kluge 1975) Polyphagous Noctuidae Heliothis sp. South Africa (Kluge 1975) ? Noctuidae Prodotis stolida (F) Namibia, Italy, South Africa (Kluge 1975) Tribulus spp. Biological control 62

Noctuidae Spodoptera exigua (Hubn.) South Afiica (Kluge 1975) Polyphagous Pterophoridae Trichoptilius congrualis Walk. India (Sankaran and Ramaseshiah 1981) ? Pyralidae Tegastoma sp. Italy (Ref?) ? Pyralidae Hb. India (Sankaran and Ramaseshiah 1981) Probably Tribulus sp. Pyralidae Tegostoma subdita/is Zell. Namibia (Kluge 1975) Tribulus spp. Pyraustidae Species unidentified Turkey (Kluge 1975) ? Acrididae Oxya velox Fabr. India (Aziz and Rizvi 1967) Polyphagous Acrididae Schistocera gregaria F orsk. Mali (Roffey and Popov 1968), India (Bhatia 1940, Varma Polyphagous and Sharma 1967) Thripidae Frankliniella schultzei (Trybom) India (Ramaseshiah 1976, Sankaran and Ramaseshiah 1981) Polyphagous Thripidae Haplothrips nigricornis India (Ramaseshiah 1976) Polyphagous Pathogens, fungi and viruses Bacteria Xanthomonas campestris pathovar tribuli India (Srinivasan and Patel 1956), Bradbury 1986 ? (Srinivasan & Patel) Dye Deuteromycetes Ascochyta tribuli Bondarzeva-Monteverde U.S.S.R. (Bondarzeva-Montverde 1945) ? Ascomycetes Aspergillus foveo/atus Rorie, Em'erice/la India (Rorie 1978) ? foveo/ata Rorie (perfect stage) Deuteromycetes cf Alternaria zinniae Hbl. South Afiica (Kluge 1975) ? Deuteromycetes Colletotrichum dematium Pakistan (Riaz 1974) ? Drechslera multiformis Jooste South Africa (Jooste 1975) ? Ascomycetes Chaetomium megalocarpum Rainier Japan (Rorie and Udagawa 1990) ? Coelomycetes Phyllosticta sp. USA (Farr eta/. 1989) ? Deuteromycetes Rhizoctonia so/ani USA (Farr eta/. 1989) Polyphagous Peronosporales Peronospora tribu/ina Pass. Italy (Saccardo 1888) (Oudemans 1921) South Afiica (Gorter Probably Tribulus sp. 1981) Virus Colorado red node virus USA (Thomas 1949) ? Virus Tomato spotted wilt virus USA (Allen eta/. 1983) Polyphagous Biological control 63

Table 2. Similarity of climate within the Sunraysia and Mallee regions.

Mildura Ouyen Swan Walpeup Hill Ouyen 0.90 Swan Hill 0.88 0.92 Walpeup 0.89 0.94 0.94 Loxton 0.89 0.88 0.83 0.85

Table 3. Similarity ofworld climate stations to climate stations in the Sunraysia and Mallee regions. Matching climate stations in Australia are excluded. Match Index ~ 0.7.

World climate station Climate stations in the Sunra:ysia and Mallee regions Mil dura Ouyen Swan Hill WalEeuE Loxton Montagu (South Africa) 0.79 0.80 0.80 0.79 0.75 Robertson (South Africa) 0.77 0.74 0.75 0.75 0.75 Murcia (Spain) 0.75 Tebessa (Algeria) 0.73 0.75 0.76 0.75 Oudtshoom (South Africa) 0.73 0.71 Riversdale (South Africa) 0.72 0.72 0.73 0.74 Zaragosa (Spain) 0.72 0.71 0.72 0.71 Kairouan (Tunisia) 0.71 Alicante (Spain) 0.71 0.71 0.70 Mecheria (Algeria) 0.70 Badajoz (Spain) 0.71 Touwsriver {South Africa} 0.70 <:::~. ~~ o"' ~o .

at:1:'· ~c:; · ~ •oo 0 •• 0 a= "

0~

0 ...... '\f· .

.o ; / . 8+ Tribulus spp.· 0 r})b

6+ Tribulus spp. 0\ ~ • 1 + Tribulus spp. Figure 1. Isohytes of Tribulus species density excluding introductions. Distributions wP.rP. t~il"P.n from ~ winP. nmae of taxonomic text~ on the Zvaonhvll~~e~e

Published in 'Proceedings ofthe South Australian Animal and Plant Conference,' 1994.

51 THE ORIGIN OF CALTROP IN AUSTRALIA AND THE LIKELffiOOD OF BIOLOGICAL CONTROL

John K . Scott & S.M. Morrison CSIRO Division of Entomology, Private Bag, P.O. WEMBLEY WA 6014

Abstract

Caltrop (Tribulus terrestris) is a summer growing aruma! weed, widespread in subtropical and warm temperate regions of the world. Clarification of the origins and identity of caltrop's introduced and native populations thought to exist in Australia is being undertaken as a precursor to biological control. Isozyme data indicated that southern Australian populations were very similar to populations in South Africa, Namibia, USA and the Mediterranean region. These populations had chromosome numbers of 2N = circa 24 and 36. A single South African and a single Namibian population formed a separate group, and had a chromosome number of 2N = circa 24. The isozyme banding patterns of Queensland and Northern Territory populations differed from all others and had chromosome numbers of 2N = circa 24. A fourth group of T. terrestris was detected in populations from India and Kuwait. These had chromosome numbers of 2N = circa 48. The morphology of burrs from northern Australian (Queensland and Northern Territory) collections differed from those of southern Australia and overseas collections, the latter being similar. The southern Australian form is most likely to have originated in the Middle East I Mediterranean region, which is the centre for Tribulus species diversity, and from where it has spread to South Africa and USA. Detailed surveys for biological control agents have not been undertaken in the area of origin and biological control has not been attempted in Australia. However, Microlarinus weevils from the Mediterranean region have successfully controlled caltrop in California and Hawaii. Any attempt to biologically control caltrop in Australia will need to consider the possibility of attack on the many species of closely related native Tribulus.

Introduction

Caltrop, Tribulus terrestris L. (Zygophyllaceae), is widespread in Mediterranean, subtropical and desert climates world-wide. It is listed as a weed in 37. countries and affects at least 21 crops (Parsons & Cuthbertson, 1992). It is thought to have been accidentally introduced into Australia in New South Wales before 1895 (Bourke, 1987). Weedy populations are now found in many locations in Australia (Parsons & Cuthbertson 1992; Squires, 1979). In Australia the plant can be toxic to livestock (Bourke et al. , 1992), infests some agricultural crops (Parsons & Cuthbertson, 1992) and the spiny burrs can contaminate dried fruit (Johnstone, 1990).

The taxonomic status of Tribulus terrestris is unclear (Bourke, 1987; Parsons & Cuthbertson, 1992; Scott, 1990; Wilson, 1992) largely because the morphology of the T. terrestris complex is extremely variable. In Australia it is thought that there are at least two forms of T. terrestris, a native and an introduced weedy form (Squires, 1979; Bourke, 1987; Wilson, 1992). Additionally, Tribulus terrestris has at times been confused with eight other Tribulus species in Australia (Bourke, 1987).

One option for the control of T. terrestris in Australia is to use biological control, especially since this method has been successful against the weed in California and Hawaii (Huffaker et al., 1983). However, any biological control program would need to clarify. the taxonomy ofT. terrestris in Australia to ensure the correct matching of biological control agent with target species (Scott, 1990; Shepherd, 1990). A combination of isozyme analysis, cytogenetic studies and burr morphology measurements were used to identify the weedy form of T. terrestris in Australia and to determine its overseas origin. In this paper we provide a summary of our results and conclusions.

1lUs paper w;u prepared for u><: ;u a working docwneru at the SA Aninul and Plant Control Conference. 1994. None of this ma u:rial may be abstracu:d or ciu:d ;u a reference without specific pennission or the authors. 52

Methods

Seeds of T. terrestris were obtained by writing to potential collectors in Australia and overseas. Material was received from North America, Europe, the Middle East, India, South Africa and all mainland states of Australia. A core sample of 31 populations, representing 8 countries including Australia, was used to carry out isozyme, cytogenetic and morphological studies. An extra 21 populations were analysed in the isozyme studies and for cytogenetics.

Isozyme analysis involved starch gel electrophoresis using well-established buffer and staining systems. Allozyme frequencies were examined using cluster analysis with the UPGMA method with Nei's unbiased genetic identity. For cytogenetic work, seedling root tips were treated with colchicine before fixing in acetic acid and 100% ethanol 1:3 and stored at -20· C. Pollen mother cells were fixed by the same method, but not colchicine treated. Chromosome preparations were made by hydrolysing the tissue in warm (60. C) 1.0M HCl for 5 minutes, finely macerating the material, then staining in 1% aceto-orcein for 4 minutes before squashing with a coverslip. Measurements were made of burr size and spine length and angle. Thirty burrs per site were measured. The number of seeds in each burr was determined from X-ray exposures.

Results

Six enzymes (PGM, 6PGD, MDH, IDH, AC and EST) with a total of 11 putative loci were used in the analysis. These enzymes demonstrated a diploid level of expression for the polymorphic loci examined and thus could be scored. Despite suspected polyploidy, the number of loci detected in each system was generally consistent with those found in many other plant taxa, as reported by Gottlieb (1982).

A dendrogram resulting from cluster analysis revealed 4 major groups within the samples, with identity values ranging from 0.94 to 0 .72. All four Indian and two of the Kuwait samples formed one group with the greatest separation from all other populations (similarity of 0. 72). The Queensland and Northern Territory samples made up a distinct second group (similarity of0.86). The third group consisted of one south African (Tiroll) and one Namibian (Thomasberg) sample (similarity of 0.90). The fourth and most closely related group (similarity of 0. 94) included all the southern Australian, northern W A, USA, Mediterranean and Iranian samples plus two each of the South African, Namibian and Kuwait samples.

So far, chromosome counts have been obtained for 22 overseas and 19 Australian populations. The Queensland, Northern Territory, some Victorian and Western Australian and one Namibian populations had 2N = circa 24. Other populations from southern Australia and northern W A, Victoria, South Australia, New South Wales, Mediterranean region, the Middle East, USA, Namibia and South Africa had counts of 2N = circa 36. Two of the Kuwait and the Indian populations had chromosome counts of 2N = circa 48.

Preliminary analysis of the morphology results show that despite great variation in size, length and position of spines and overall shape, the Queensland and Northern Territory burrs could be statistically distinguished from all others due to their more elongate shape.

Discussion

Isozyme, chromosome and morphology results indicate that the Queensland and Northern Territory populations are not closely related to the rest of the populations sampled, either within Australia or overseas. This indicates that they are highly likely to be a native Australian form of Tribulus terrestris. The rest of the Australian populations however, form a large closely related group with the overseas populations, excluding India and 2 Kuwait populations. This supports the hypothesis that the forms of Tribulus terrestris found in southern Australia and northern W A are introduced. The Indian and 2 of the Kuwait samples have the least genetic similarity with the rest of the populations. This suggests that these 53

regions are unlikely to be the origin of the introduced fonn of Tribulus in Australia. Similarly, the locations ofTiroll (S. Africa) and Thomasberg (Namibia) are unlikely origins. Initially, it is not apparent which of the remaining overseas locations is the origin of the Australian weedy fonn of Tribulus.

Tribulus is known to have been introduced into the USA from Europe before 1902 (Huffaker et al. , 1983). In South Africa, Tribulus terrestris is considered to be native, but it has been suggested that it was introduced from southern Europe before 1794 (Schweickerdt, 1939). The dendrogram groupings indicate that there are possibly both native (Tiroll) and introduced (Rosh Pinah, Rondebosch and Prince Albert) populations in South Africa, a similar situation to Australia.

A survey of the literature shows 2 regions, the Middle EastiN. E. Africa and N. W. to central Australia which contain the greatest number of species in the Tribulus genus (8 to 9). Southern Africa has fewer species (6 to 7). Radiating out from these 3 centres the number of species declines rapidly.

The greatest range in chromosome numbers reported from T. terrestris is in the Mediterranean/ West Asia to India region. The only diploid forms of T. terrestris (2N = 12) have been reponed in Pakistan (Baquar et at. 1965). None were found in the samples in our survey. The highest chromosome numbers (2N = 48) are from India (Bir & Sidhu, 1980) and South America (Schnack & Covas, 1947).

It appears therefore, that the original population of the weedy fonn of Tribulus came from the Mediterranean/ West Asia region. The introduction into Australia could have been directly from the Mediterranean/West Asia or as a secondary introduction via the USA or South Africa. Evidently there could have been multiple introductions from all or several of these locations. Although the introduced fonn in Australia is distributed over a wide geographical range it is possible that it could have been dispersed from one source. This is because the burrs are readily spread by animals in their feet or coats and by humans by way of rubber tyres and shoes. The burrs become embedded in the fleece of sheep and this is a likely means by which it has become spread around the world (Parsons & Cuthbertson, 1992).

Potential for biological control

The most comprehensive survey for biological control agents on T. terrestris was undenaken by Kluge (1975) in South Africa and Namibia. Other surveys have been made in Bangalore and nearby areas in South India (Sankaran & Ramaseshiah, 1981) an~ in Israel (Gerling & Kugler, 1974). Scott (1990) concluded that the leaf-feeding mite, Eriophyes tribuli, and the pathogen, Peronospora tribulina, would be of high priority to be studied in a biological control program. One product of the study of variability in caltrop was the discovery in Australia of an eriophyid mite causing damage corresponding to that of E. tribuli. This mite is widespread and found in the northern half of Australia on at least four species of Tribulus (Hosking & Scott, unpublished). Too little is known of other potential agents such as the pathogen, P. tribulina, the moths, Prodotis stolida (: Noctuidae), Tegostoma subditalis (Lepidoptera: Family name), Ephysteris subdiminutellaferritincta (Tum.) (Lepidoptera: Gelechiidae) and Deroplax sp. (Hemiptera: Pentatomidae) to judge whether they could be used in biological control, although it is likely that their host ranges will include at least other species in the genus Tribulus, and not just T. terrestris.

~uccessful biological control of caltrop has involved using the weevils Microlarinus species (Coleoptera: Curculionidae) imponed into the USA from Italy (Huffaker et al., 1983). Microlarinus lypriformis feeds in the stem. It is now established in wanner states in USA, Hawaii, St Kitts, Nevis and Papua New Guinea (Julien, 1992). The second species, Microlarinus lareynii feeds inside the burrs and is established ., in mainland USA and Hawaii (Julien, 1992). These insects may not show the required level of host plant specificity for introduction into Australia. Firstly, they are likely to feed on native Australian Tribulus species. Secondly, the adults will feed on a range of plant species (Andres & Angalet, 1963). 54

With the wide distribution of T. terrestris, one approach is to identify areas of the world that closely resemble the region in Australia where control is desired. One such area is the Sunraysia and Mallee regions of Victoria/ N.S.W. and South Australia where caltrop affects the dried fruit industry and causes sheep mortality. · Five climate stations, which are included in the CLIMEX database (Sutherst & Maywald, 1985), were chosen to represent the region. The sites were Mildura, Swan Hill, Walpeup, Loxton and Ouyen. These sites were compared with all climate stations included in the world data base in CLIMEX. Drier areas of the south west and southern Cape Province of South Africa were represented by five matching sites. Kluge (1975) collected insects from Stellenbosch which is nearby. A short survey over one season would confirm whether the same species are present. The second group, which consists of four sites, is in Spain. This area has never been examined for biological control agents for T. terrestris. The third area includes Algeria and Tunisia, both of which have not been examined for biological control agents for T. terrestris. None of the sites identified in the CLIMEX analysis are found near the centre of diversity of the genus Tribulus probably because the Australian sites used in the climate matching process are near to the edge of the natural distribution of T. terrestris .

The distribution of native species of Tribulus does not overlap with that of introduced T. terrestns m Mediterranean climatic areas in Australia. This suggests that since native Tribulus species should be spared attack by introduced biological control agents, a strategy would be to use biological control agents restricted to Mediterranean climatic areas. This implies that the successful agents, Microlarinus lypriformis and M. lareynii may have to be excluded, especially as M. lypriformis has already shown that it can colonise a wide climatic area, for example it is established in New Guinea.

In any future biological control program it is recommended that the Mediterranean region be examined and that agents with distributions restricted to Mediterranean regions or with very restricted host range within the genus Tribulus be used.

Acknowledgments

This study was supported by a grant from the Australian Dried Fruits Research and Development Council. We thank the many collectors of caltrop seed and the Department of Agriculture, Western Australia, Weed Science Branch for use of facilities.

References

Andres, L. A. & Angalet, G. W. (1963). Notes on the ecology and host specificity of Microlarinus lareynii and M. lypriformis (Coleoptera: Curculionidae) and the biological control of puncture vine, Tribulus terrestris. f. Econ. Ent. 56:333-340.

Baquar, A. K., Aktar, S., & Husain, A. (1965). Meiotic chromosome numbers in some vascular plants of Indus Delta. Bot. Not. 118: 289-298.

Bir, S. S. & Sidhu, M. (1980). Cyto-palynological studies on weed flora of cultivable lands of Patiala district (Punjab). 1. Palynology 16: 85-105.

Bourke, C. A. (1987). Some· taxonomic, agronomic and animal health aspects of Tribulus. Pro c. 8th Aust. Weeds Conf. 182-1 85.

Bourke , C. A., Stevens, G. R. & Carrigan, M. J. (1992). LocomotOr effects in sheep of alkaloids identified in Australian Tribulus terrestris. Aust. Vet. 1. 69:163-165 .

Gerling, D. & Kugler, J. (1974). Evaluation of enemies of noxious plants in Israel as potential agents for the biological control of weeds. Final technical report, Tel Aviv University, 241 pp. 55

Gottlieb, L D. (1982). Conservation and duplication of isozymes in plants. Science (216, 373-380.

Hilu, K. W. (1981). Cytotaxonomical studies in Tribulus terrestris and T. alatus (Zygophyllaceae). Nord. 1. Bot. 1:531-534.

Huffaker, C. B., Hamai, J., & Nowierski, R. M. (1983). Biological control ofpuncturevine, Tribulus terrestris in California after twenty years of activity of introduced weevils. Entomophaga 28:387-400.

Julien, M. H. (1992). Biological Control of Weeds. A World Catalogue of Agents and their Target Weeds. CAB International, 186 pp.

Johnstone, R. B. (1990). Control of Emex, Tribulus, and Cenchrus in vineyards. Plant Protection Quanerly 5(3):84.

Kluge, R. L (1975). Observations on insects associated with Tribulus terrestris L. in southern Africa. Up..tii

Parsons, W. T. & Cuthbertson, E. G. (1992). Noxious weeds of Australia. (lnkata Press: Melbourne).

Sankaran, T. & Ramaseshiah, G. (1981). Studies on some natural enemies of puncturevine Tribulus terrestris occurring in Karnataka state, India. Proc. V Int. Symp. Bioi. Control Weeds. 153-160

Schnack, B. & Covas, G. (1947). Kariological studies in the flowering plants. Haumania 1: 32-41 (not seen, taken from Hilu 1981).

Schweickderdt, H. G. (1939). An account of the South African species of Tribulus Tourn. ex Linn. Bothalia 3:157-178.

Scott, J. K. (1990). Tribulus cerrescris L (Zygophyllaceae) in Southern Africa: An outline of biology and potential biological control agents for Australia. Plant Prot. Quan. 5: 103-106.

Shepherd, R. C. H. (1990). Past Victorian work on Emex australis Stenheil and Tribulus cerrestris L Plant Protection Quanerly 5:100-102.

Squires, V. R. (1979). The biology of Australian weeds . 1 Tribulus terrestris L. 1. Aust. Instit. Ag. Sci. 45: 75-82.

Sutherst, R. W. & Maywald, G. F. (1985). A computerised system for matching climates in ecology. Agric. Ecosyst. Environ. 13: 281 -299.

Wilson, K. L. 1992. A new species and a neotypification in Australian Tribulus (Zygophyllaceae). Telopea 5: 21-29. Published in 'Proceedings ofthe lOth Australian and 14th Asian-Pacific Weed Conference,' 1993.

Weed morphology and distribution

ASSESSMENT OF THE ORIGINS OF TRIBULUS TERRESTRIS IN AUSTRALIA

S.M. Morrison and J.K. Scott CSIRO Division of Entomology, Private Bag PO, Wembley WA 6014, Australia

Summary. Tribulus terrestris is an introduced weed of unknown origin in Australia It is a problem to the Australian drie

INTRODUCTION

Tribu/us terrestris (Zygophyllaceae) (caltrop) occurs in semi-arid and Mediterranean-type climates world-wide. It is a prostrate, summer growing annual. The seeds germinate after late spring and summer rains in warm conditions (24 - 21·q. The fruit consists of a woody burr comprised of five wedge-shaped cocci each with one or two pairs of large spines (8). There are 2 to 4 seeds per coccus (16). Seeds can remain viable for many years if buried in the soil, thus after successive generations a large reservoir of seed can accumulate (16).

Weedy populations of T. terrestris are found in many locations in Australia (10, 16). It is thought to have been accidentally introduced to New South Wales before 1895 (16). Tribulus terrestris is a problem to the Australian dried fruit industry because the spiny burrs may contaminate harvested dried fruit (7). Additionally the plant can be toxic to livestock, especially sheep. (2, 3). The taxonomic status of T. terrestris is unclear (2, 10, 12 19) and forms of the plant are variously considered to be native or introduced. Before considering biological control of this weed, it is essential to identifY. the introduced weedy form of T. terrestris and to determine its overseas origin (12). Here we report on a preliminary analysis, using isozyme techniques, to aid the identification of forms of T. terrestris as an indication of their geographic ongm.

MATERIALS AND METHODS

Tribulus terrestris accessions were obtained by written requests to potential collectors. Over 80 samples were received from mainland Australia, and 13 from overseas locations with Mediterranean-type climates. The analysis presented here used seven populations from Australia and six from overseas. Dry seeds were stored at room temperature until needed. Seeds were excised from the woody burr and dipped in 1.25% sodium hypochlorite for 5 minutes, then washed in distilled water. The seeds were germinated on moist filter paper in a sealed petri dish incubated at a 16 hour 2o·c, 8 hr 35"C cycle. Seedlings were used for isozyme work when they were 2 to 4 days old.

Cotyledon and root tip tissue of each seedling were homogenised together in 7 pL of 0.16M phosphate buffer (pH 7.0}, containing 2.5% sucrose, 0.12% bromophenol blue and 0.75 mg mL·'

388 Weed morphology and distribution dithiothreitol, in a plastic microwell tray on an ice bath (9). Three 2 x 4 mm filter paper wicks were placed in each sample. Five to ten seedlings from each of three plants was assayed for each population, except for Crete where seeds from only one plant was available.

Horirontal starch gels were made from 9.6% hydrolysed starch (Connaught) in each of three different buffer systems. Three enzymes were assayed on each of three buffer systems as follows: histidine gels at pH 8.0 (4), phosphoglucomutase (PGM) EC 2.7.5.1, 6- phosphogluconate dehydrogenase (6PGD) EC 1.1.1.44, phosphoglucoisomerase (PGI) EC 5.3.1.9; tris-citrate gels at pH 7.0 (II), aconitase (AC) EC 4.2.1.3, isocitrate dehydrogenase (IDH) EC 1.1.1.42, malate dehydrogenase (MDH) EC 1.1.1.37; lithium borate gels at pH 8.0 (9), esterase (ES1) EC 3.1.1.1, leucine aminopeptidase (LAP) EC 3.4.11 or 13, glutamate-oxaloacetate tnrnsaminase (G01) EC 2.6.1. I.

A wick from each sample was inserted into a slot cut in the cathodal side of each gel. To prevent the gels overheating, electrophoresis was carried out at 4·c in a refrigerator with an ice bath placed over each gel and a maximum current of 50 rnA. Electrophoresis was tenninated when the bromophenol blue front had moved about 10 em. Each gel was sliced horirontally into 3 sections and each slice stained for an individual enzyme.

The assay methods used for EST, IDH and PGM and LAP were those described by Shaw and Prasad (13), except that an agar overlay was used for PGM. PGI, AC and MDH assays followed the methods of Richardson eta/. (II), and 6PGD and GOT the method of Brown eta/. (5). Stained gels were fixed with acetic acid, then rinsed in water, wrapped in plastic and photocopied for a permanent record. Isozyme banding patterns were scored irrunediately after staining.

Cluster analysis was performed using the UPGMA method (14) with modified Rogers distance (20) on the Biosys-1 computer program ( 17).

RESULTS AND DISCUSSION

Six of the enzymes with a total of eleven polymorphic loci were used in the analysis. In GOT and LAP systems the resolution proved too poor to score confidently, so they were excluded. PGI loci were clearly polymorphic, but were also excluded because they expressed complex zymogmms with multiple overlapping bands. This indicates that there are more than the expected two loci for PGI (18), possibly with heteropolymers between the loci. It is therefore likely that the populations examined were polyploid. Additionally, fixed heterozygosity at the 6PGD - 2 locus in all populations examined, plus preliminary chromosome counts (4n = 24) support this view. Only those enzymes that demonstrated a diploid-like expression for the polymorphic loci examined were used in the analysis.

The dendrogram (Fig. I) shows that some of the populations are very similar to each other. The southern Australian, Broome (northern Western Australia), South African and U.S.A. populations formed a close group. Goomalling (southern Western Australia), Middle East and Crete populations formed a second group. The Indian population was separate from any other groups and the Queensland and Northern Territory populations are closely related to each other, but were the greatest distance from the other populations. The percentage standard deviation of the best fining dendrogmm was 15.3% (6). Other methods used to derive genetic distances (17) produced essentially the same overall dendrogram.

389 Wl't'd morphology and di.l·frihution

ACKNOWLEDGMENTS

This study was supported by a grant from the Australian Dried Fruits Research and Development Council. We thank J. Dodd, S. Lloyd, J. Matthiesscn, R. Randall and I Tommerup for c.:o mrncnts on the manuscript.

REFERENCES

I. Bourke, C.A. 1987. Res. Vet. Sci. 43, 347-350. 2. Bourke. C. A. 19!!7. Proc. 8th Aust. Weeds Con f., Sydney. pp 182-185. J. Bourke, C.A., Stevens, G.R. and Carrigan, M.J. 1992. Aust. Vet. J. 69, 163-165. 4. Brewer. G.J. and Sing, C.F. 1970. An introduction to isozyme techniques. Academic Press, New York, I R6 pp. 5. Brown. A.H.D., Zohary, N.E. and Dagan, 0. 1978. Genetica 49, 97-108. o. Fitc.:h. W.M. and Margoliash, E. 1967. Science 155, 279-284. 7. Johnstone, R.B. 1990. Plant Protection Quarterly 5, 84. X. Nabil el Uadidi, M. 197R. Taeckholmia 9, 59-66. 9. Panetta. F.D. and Carstairs, S.A. 1989. Weed Res. 29, 157-165. 10. Parsons, W.T. and Cuthbertson, E.G. 1992. Noxious weeds of Australia. Inkata Press Melbourne, 692 pp. II. Richardson, B.J., Bavcrstock, P.R. and Adams, M. 1986. Allozyme electrophoresis. A handbook for animal systematics and population studies. Academic Press. Sydney, 410 pp. 12. Scott. J.K. 1990. Plant Protection Quarterly 5, 103-106. 13. Shaw, C.R. and Prasad, R. 1970. Biochem. Genet. 4, 297-320. 14. Sneath. P.H.A. and Sokal, R.R. 1973. Numerical Taxonomy. Freeman, San Francisco. 15. Squires. V.R. _1969. Viet. Nal 86, 328-334. lo. Squires, V.R. 1979. J. Aust. Instil. Ag. Sci. 45, 75-82. 17. Swofford. D.L. 19!!9. Biosys- I, release 1.7. Ulinois Natural History Survey, Champaign, 43 pp. I X. Weeden. N.F. and Wendel. J.F. 1990. In: Isozymes in Plant Biology. (Eds. D.E. Soltis and P.S. Soltis) (Chapman and Hall, London). pp 46 -72. 19. Wilson, K.L. 1992. Telopea 5, 21 -29. 20. Wright. S. 197!!. Evolution and the Genetics of Populations. Vol. 4. University of Chicago Press. Chicago. 5!!0 pp.

391 Weed morphology and distribution

This preliminary analysis supports the view that there are two forms of T. terrestris in Australia, an introduced form and a na.tive form. The form found mainly in southern Australia is very similar to the plants from some of the overseas collections. The Northern Territory and Queensland forms of T. terrestris had the greatest genetic distance from overseas populations indicating that they are native to Australia. Possible origins of the introduced form of T. terrestris are the Mediterranean, Middle East and South Africa. The inclusion of further samples and cytogenetic data in the analysis will help to clarify the likely origins of the introduced T. terrestris .

.40

.33 I

.27 I I w (_) z: I- - -;=:::1..-.-, ~ .20 (/) 0 .13 I

.07 I i !

.00 n ~ 1 2 3 4 5 6 7 8 9 10 11 12 13

Figure 1. Dendrogram showing relationships among Tribulus terrestris populations, based on isozyme dissimilarity. Populations are as follows: (1) Darwin, Northern Territory, (2) Charters Towers, Queensland, (3) Kashmir, India, (4) Rondebosch, South Africa, (5) Loxton, South Australia, (6) Mudgee, New South Wales, (7) Mildura, Victoria, (8) Prosser, WA, USA, (9) Broome, Western Australia, (10) Crete (11) Kuwait City, Kuwait, (12) Tehran, Iran, and (13) Goomalling, Western Australia.

390 Published in 'Weed Update' number 10, November 1991. Department of Agriculture, Western Australia. · .

Geneticist to work on caltrop

Sue Morrison has recently joined the Weed Science Branch as a geneticist. She will be working on a joint CSIRO/WADA project examining the amount of genetic variability present in caltrop populations throughout Australia. Sue has a background in zoology, with further studies in marine biology. She worked for-the last few years as a research assistant in the Department of Anatomy and Human Biology at UW A, on projects involving the genetics of small native mammals, glaucoma research and salinity adaptation in estuarine fish.

Drs Dane Panetta and John Scott received funding from the Australian Dried Fruits Industry to run the 3 year project on the weed, caltrop (Tribulus terrestris).

Caltrop is a major problem for the dried fruit industry because its spiny burrs readily contaminate harvested dried fruit. It is also a problem with Hvestock and can cause staggers in sheep, besides causing injury to the feet of sheep and other stock.

Caltrop is extremely variable in Australia. Some of the forms are considered to be native, others are possibly introduced. Caltrop has been successfully controlled in the USA by biological control agents. However, ·this method of control cannot be considered without knowing exactly what forms of caltrop occur in Australia.

The aim of the project is to characterize all coll¢ctions of caltrop. This will be done using isoz.yme and cytogenetic techniques to examine variation within and between populations, both Australian and overseas. It is planned to use the information to establish how many forms of caltrop there are in Australia as well as likely regions of origin. This will permit an evaluation of the possibility of biological control or alternative control techniques.

For this study, seeds are required from as many populations as possible. Samples of caltrop from anywhere in Australia would be greatly appreciated, Details for methods of collection will be given in a later Weed Update. Published in 'Weed Update' number 11, March 1992. Department of Agriculture, Western Australia.

Genetic variation in caltrop Recommendations for the control of caltrop are Caltrop has been successfully controlled in U.S.A. as follows: by biological control agents. However this method of control cannot be considered without knowing • Amitrole (320g/L), atrazine (320 g/L) exactly what forms of caltrop we have in proprietary mix + 2,4-D amine (500g/L). Apply Australia. Correct identification will be of use for at 2L + 2L/ha or for knapsack, 20ml + 20ml per other control techniques and in understanding the 10L water/100m2. Wetting agent dilution 0.25- causes of tribulus staggers. We are seeking your 5%. Apply before flowering. Gives short term assistance in obtaining fruits of caltrop from as residual control. many locations as possible. We will germinate the seeds and study the genetic variation in the • Glean® at 20g/ha or for knapsack, 0.2g per 10L seedlings. It is particularly important that seeds water/100m2. Wetting agent dilution 0.25-5%. from different plants are not mixed. Apply to small plants while soil is moist. Addition of 1-2L/ha 2,4-D amine (500g/L) Method of collection gives better control of adult plants.

1. Find five large plants with mature fruits. • Glyphosate 360 or 450g/L at 2-3 L/ha or for knapsack, 50-70ml per 10L water/100m2. 2. Place at least 100 fruits from each plant into a Wetting agent dilution 0.25-5%. Apply before separate paper bag. Heavy gloves are useful to seed is formed. Use higher rates for the less remove the fruits. concentrated product

3. If most of the fruits drop off the plant use the • 2,4-D ester 800g/L at l.OL/ha or for knapsack, collection method below. 10ml per 10L water /100m2. Wetting agent dilution 0.25%. Apply before seed is formed. 4. Write on a large paper bag your name, the More effective on seedlings than on adult collection site (plus nearest town), the date, the plants. land use (e.g. irrigated grapes, sheep pasture, peanut crop) and any other potentially useful • 2,4-D amine 500g/L at 1.4 L/ha or for knapsack, observations i.e. in relation to stock poisoning. 20ml per 10L water /100m2. Wetting agent dilution 0.25-5%. Apply before seed is formed. 5. Place the paper bags containing the seeds into the larger paper bag and seal.

6. Allow the fruits in the paper bags to dry.

Mailing instructions

Send the paper bag containing the seeds to Dr Dane Panetta or Mrs Sue Morrison, Weed Science Branch. Control of caltrop

Summer rains between December and February this year have provided favourable conditions for a number of summer weeds including caltrop (Tribulus terrestris) both in agricultural and metropolitan areas.

Caltrop is a prostrate, branching plant with a woody taproot. the yellow flowers are followed' by a spiny fruit which can injure feet and puncture tyres. Where caltrop occurs in high densities, it can cause hindquarter staggers in sheep and nitrate poisoning.

Caltrop germinates quickly after summer rains and commences forming fruit within three weeks. Published in 'Feedback' number 70, March 1992. Department of Agriculture, Western Australia. ·

COLLECTING INSTRUCTIONS- TRIBULUS TERRESTRIS (CALTROP) PREFERRED METHOD OF Please include your address so Dr F. D. Panetta or COLLECTION that we can acknowledge your Ms S. Morrison, 1. Find five large plants with assistance. at the above address, mature fruits (seeds). A word of warning: DO NOT Tel. (09) 368 3381 2. Placeatleast100fruitsfrom USE PLASTIC BAGS, as these Fax. (09) 474 2658, or each plant into a separate cause the seeds to rot. Dr John K. Scott. paper bag. Fold the top so ENQlllRIES CSIRO Division of Entomology, that loose fruits do not es­ Private Bag, cape. Heavyglovesareuse­ Should you wish to know more about the project please contact: P.O., Wembley 6014 ful to remove the fruits. Tel. (09) 387 0644 3. If most of the fruits drop off Fax. (09) 387 8991 the plant use the collection method below. 4. Write on a large paper bag your name, the collection site (plus nearest town), the date, the land use (e.g. grapes, sheep pasture) and any other potentially use­ ful observations i.e. in rela­ tion to stock poisoning. 5. Place the paper bags con­ taining the seeds into the larger paper bag and seal. 6. Allow the fruits in the pa­ per bags to dry. ALTERNATIVE COLLECTION METHOD Where fruits (seeds) only are available or plants are small. 1. Write on a paper bag the details as in (4) above. 2. Take a rubber soled shoe (a thong is ideal), or similar. ~:--~.~ ·.· · · 3. Press onto the ground where caltrop fruits (seeds) are present. 4. Using a knifeorsimilartool, scrape the fruits from the shoe into a paper bag. Con­ tinue until you have at least 100 fruits. 5. Seal the paper bag so that ~~ ··_.. · ~~'\\.~ the fruits do not escape. . ., ~ : ~:::E~;;;;: 6. Allow the bag to dry if the ~:- · ~ ··~· ~;Ao~- ·'4· ~~ fruits are moist ~ D .·''j : F MAILING TNSTRUCTIONS .

Send the paper bag by mail to: , ... ~

Ms. Sue Morrison, Fig. 2. Habit of the plant and details of fruits. flower s and seeds (after Gardner and Weed Science Branch, Bennetts 1956). a, habit of mature plant; b. detail of compound leaf; c, detail of W.A. Dept of Agriculture, flower; d, schizocarp showing the five component mericarps; e, side view of 3 Baron-Hay Court, schizocarp; f, s ide view of a mericarp showing the two lateral and two divergent South Perth, W.A. 6151. spines; g, section of a mericarp showing three immature seeds. Published in 'Feedback' number 72, July 1992. Department of Agriculture, Western Australia.

PRICKLY PROBLEM PURSUED Read91'S will recall a reqiHJst in the the vicinity of quarter a million in a starch gel by putting an elec­ March edition of FEEDBACK asking dollars per year through poison­ tric charge through the gel. The for samples ofcaltrop to be sent in to ing and the physical damage banding patterns of the enzymes the Agriculture 09partm6nt. H8re is caused to sheep. Wehaven'tgot in the gel will enable us to deter­ the story behind the request. that problem here in W A yet. but mine the relationship (if any) '"For a weed that germinates so it could come, particularly with between the different samples." readily in the field, I have one summer rains such as we've had "Then by using caltrop samples devil of a job getting it to germi­ this year," she said. from its horne ranges in South­ nate in the laboratory!" Sue "There is also the social aspect of em Africa, India, the Middle East Morrison was explaining one of the weed. It is difficult for shires and the Mediterranean area, and the difficulties she is having to control, and punctured bicy­ comparing their genetic make­ studying the weed caltrop in her cle tyres, injures children's feet, up with our Australian samples I laboratory at the Department of and temporarily-crippled pets hope to be able to identify where Agriculture in South Perth. are common in areas where ours originated. '1 have to cut out each seed indi­ caltrop grows prolifically," Ms '7hen we can start searching vidually from its woody, Morrison added. there to find a biological control spiny carpel with a scalpel ~----.....!----· agent suitable to our condi­ so I can get them to germi­ . tions. We'll then be well on nate. I need to be able to · thewaytosuccess,although speed up the process which, · it will be a long, drawn-out apart from being extremely process," she said. hard on my fingers, is tak­ 'There appears to be some ingup too much of my time.n variation in the plants I've MsMorrison is trying to find tested so far from W A, Ms if there is any genetic Morrison said. (isoenzyme) variation in "We had an exciting find at caltrop, particularly be­ Mount Magnet recently, tween the native and im­ ported varieties, with the when we found an Eriophy­ id mite attacking caltrop. All eventual aim of introducing biological control to the in­ plants there were being at­ troduced species of the tacked, so it could be a pos­ weed. She is working half sible control agent. Even if it doesn' t turn out that way, time on the three year pro­ {· at least we know that some­ gram, which is a co-opera­ tive venture between Dr . thing eats it," she added.

Dane Panetta of the W A r , >, > "Caltrop has been the target Dept of Agriculture and Dr i :-"' of a successful biological John Scott at CSIRO, and control program in Califor­ funded by the Australian nia. There they have intro­ Dried Fruit Industry. d uced two weevils from "Caltrop is a big problem in Italy which exert some con­ the east, and the dried fruit trol over the plant. One at­ industry is very keen to get tacks the seed of caltrop while the other attacks the controloftheweed. Caltrop • is nasty at the best of times, stem and crown of the plant. but you can imagine what it ~>iUfl.!llll~. ...il--..:...._-,-_ _ Let's hope that our research would be like if it contaminated ABOVE: SUE MORR&>N SHOWS A PAIR successfully identifies the varie­ dried fruit. The last thing you OF TI!ONGSSTUDDED wrT1i CALntOP ties of caltrop we have here in would want is to bite into one of The process she uses to identify Australia, so we can go ahead these seeds. the different varieties of caltrop with our overseas search for a In addition to this, one estimate is called electrophoresis. In this bio-rontrol agent," Ms Morrison last year put the cost of caltrop in procedure, the enzymes in each concluded. the MalleeregionofVictorian in plant sample are separated out Ron Diver Collecting instructions for caltrop sent out to potential collectors.

COLLECTING INSTRUCTIONS - TRIBULUS TERRESTRIS (CALTROP, PUNCTURE VINE) PREFERRED METHOD OF COLLECTION 1. Find five large plants with mature fruits (seeds). 2. Place at least 100 fruits from each plant into a separate paper bag and fold the top of the bag so that loose_fruits do not escape. Heavy gloves are useful to remove the fruits. 3. If most of the fruits drop off the plant use the collection method below. 4. Write on a large paper bag your name, the collection site (plus nearest town), the date, the land use (e.g. irrigated grapes, sheep pasture, peanut crop) and any other potentially useful observations i.e. in relation to stock poisoning. 5. Place the paper bags containing the seeds into a larger paper bag and seal. 6. Allow the fruits in the paper bags to dry. ALTERNATIVE COLLECTION METHOD Where fruits (seeds) only are available or plants are small. 1. Write on a paper bag your name, the collection site (plus nearest town), the date, the land use (e.g. irrigated grapes, sheep pasture, peanut crop) and any other potentially useful observations i.e. known to cause stock poisoning 2. Take a rubber soled shoe (a thong is ideal), or similar object. 3. Press onto the ground where caltrop fruits (seeds) are present. 4. Using a knife or similar tool, scrape the fruits from the shoe into a paper bag. Continue until you have at least 100 fruits. 5. Seal the paper bag so that the fruits do not escape. 6. Allow the bag to dry if the fruits are moist MAILING INSTRUCTIONS Send the paper bag containing the seeds via the mail to Ms.Sue Morrison, Weed Science Branch, W.A. Department of Agriculture, 3 Baron-Hay Court, South Perth, W.A. 6151. Please include your address so that we can acknowledge your assistance. ENQUIRIES Should you wish to know more about the project please contact · · : . Ms S. Morrison, at the above address Tel. (09) 368 3381 Fax. (09) 474 2658

or Dr John K. Scott. CSIRO Division of Entomology, Private Bag, P.O., Wembley W.A. 6014 Tel. (09) 387 0644 Fax. (09) 387 8991

Fig. 4. Fruits from plants identified as T. terrestrls from a number of localities th roughout Australia and overseas. (France); (b) Kansas (U.S.A.); (c) Alice Springs (Central Australia); (d) Koonamore (South Australia); (e) Darwin Australia); (f) Nanango (Queensland); (g)Trangie (New South Wales); (h) Fitzroy Crossing (Western Australia). . . . Examples of Tribulus terrestris seeds (from Squires, 1979, J.Aust.InstAgric.Sci. 45,p78) 1

Appendix 6. List of caltrop seed collectors Name Address Dr Rick Boydston Plant Physiology, Agricultural Research Services, U.S. Department of Agriculture, Irrigated Agricultural Research, Prosser, Washington, 99350, USA. Dr D. Johnston Department of Agronomy, University of Arkansas, Arkansas Agricultural Experimental Station, Fayetteville, AK 72703, USA. Dr Loren Moshier Kansas State University, Throckmorton Hall, Manhattan Kansas 66506, USA. Dr R. Talbert Department of Agronomy, University of Arkansas, Arkansas Agricultural Experimental Station, Fayetteville, AK 72703, USA. Dr I.A. Nawhoo Department ofBotany, University ofKashmir, Hazratbal, Srinagar 190006, India. Dr M.Sanei Chariat- Herbarium, Horticulture Department, Faculty of Agriculture, Karaj Pahini Agricultural College, University ofTehran, Karaj, Iran. Dr L. Boulos Herbarium, Botany and Microbiology Department, Kuwait University, PO Box 5969, 13060 (Kuwait) Safat, Kuwait. Dr George Cyprus Herbarium, Department of Agriculture, Ministry of Markoullis Agriculture and Natural Resources, Nicosia, Cyprus. Dr Richard Groves CSIRO Department ofPlant Industry, GPO BOx 1600, Canberra, ACT 2601. Dr Werner Greuter Herbarium, Botanischer Garten und Botanisches Museum Berlin­ Dahlem, Konigin-Luise-Strasse 6-8, D-1000 Berlin 33, Germany. Dr J.J. van Biljon Ciba-Geigy Pty Ltd., PO Box 92, Isando 1600, Transvaal, South Africa. Drs W.R.J. and S. Tierberg Karoo Research Centre, PO Box 47, Prince Albert 6930, Dean South Africa. Dr Penny Edwards CSIRO Biological Control Unit, Zoology Department, University of Cape Town, Rondebosch 7700, South Africa. Dr Christine Hanel Gobabeb, PO Box 1592, 9000 Swakopmund, Namibia. Mark Ashley c/o Graham Schultz Piers Barrow Berrimah Agricultural Research Station, Strath Road, PO Box 79, Berrimah, NT 0828. M. Crothers c/o Graham Schultz Steve Newbold c/o Colin Wilson Dr Graham Schultz Department ofPrimary Industry and Fisheries, PO Box 990, Darwin, NT 0801. Ron Smith c/o Colin Wilson Dr Colin Wilson Department ofPrimary Industry and Fisheries, PO Box 79, Berrimah, NT 0828. Dr Richard Carter Animal and Plant Control Commision, 242 Marion Road, Netley, PO Box 1671, Adelaide, SA 5001. Dr Richard Klingberg Department of Agriculture, PO Box 411, Loxton, SA 5333. Dr Marie Ahlin Tropical Weeds Research Centre, Charters Towers, QLD 4820. Dr Mark Hallam Australian Nature Conservation Agency, PO Box 636, ACT 2601. Peter Mesch DuPont (Australia) Ltd., 1 Warralong Street, The Gap, PO Box 149, Ashgrove, QLD 4061 . 2

Dr Dane Panetta Land Protection Branch, Lands Department, PO Box 168, North Quay, Brisbane, QLD 4002. Dr Chris Bourke Specialist Veterinary Research Officer, Agricultural Research and Veterinary Centre, Forest Road, Orange, NSW 2800. Dr John Hosking NSW Agriculture and Fisheries, Agricultural Research Centre, RMB 944, Tamworth, NSW 2340. Bob Lewis CSIRO, Division ofEntomology, Trangie Station, NSW. Robyn Turner Mogen, RMB 67, Burgooney, NSW 2672. Allison Chambers Rutherglen Research Institute, Chiltern Valley Road, Rutherglen VIC 3685. K. Clarke c/o Doug Harris Dr Bill Easton Mallee Research Station, Walpeup, VIC 3507. Dr Doug Harris District Veterinary Officer, 1-3 McCallum Street, Department of Agriculture, PO Box 501, Swan Hill VIC 3585. A. Heslop c/o Doug Harris J.Hulland c/o Doug Harris J. Linklator c/o Doug Harris M. McLean c/o Doug Harris B. Singleton c/o Doug Harris J. Smith c/o Doug Harris Norm Stone Schering Pty Ltd., 141 Burnley Street, PO Box 411 , Richmond VIC 3121. C. Threadgold c/o Doug Harris Martin Atwell APB, PO Box 34, Wickepin, WA 6370. Mrs Bartlett Stoneville, Perth Ken Battison APB, PO Kondinin, WA 6367. Steve Brough APB, PO Box 26, Perenjori, W A 6620. Julie Cooper Weed Science, Department of Agriculture Western Australia, South Perth, WA 6151. A.M. Davies 25 The Grove, Wembley, WA. Ray Eakins APB, PO Mullewa, WA 6630. Lorna Edwards 117 Epsom Avenue, Redcliffe, WA 6104. Geoff Gordon APB, Narrogin, WA6312. Bill Gerrie APB, Greenough, WA 6530. Ray Gwynne APB, Esperance, WA 6450. Christine Johnstone APB, PO Box 91 , Fitzroy Crossing, WA 6765. Rob Klemm Weed Science, Department of Agriculture Western Australia, South Perth, WA 6151 . Harley Lacy Polelle Station, Meekatharra, c/o Department of Agriculture, PO Box 108, Meekatharra, W A 6642. Steve Lawn Engineering Department, City ofPerth, PO Box C120, GPO Perth, WA6000. Sandy Lloyd Weed Science, Department of Agriculture Western Australia, South Perth, WA 6151 . A. Longbottom APB, PO Box 147 Onslow, WA 6701. Paul Manera APB, 16 Clayton Street, Hyden, WA 6359. Elisa Marold c/o Weed Science, WA Department of Agriculture, South Perth, WA 6151. D. Mather c/o Andrew Mitchell 3

Tony McNamara APB, Sandalwood Yards, 179 Avon Terrace, York, WA 63020 S 0Merewether APB, Bunbury, WA 62300 Lyle Metcalfe APB, PO Box 66, Wyalkatchem, WA 64850 Andrew Mitchell Department of Agriculture Western Australia, Broome, PO Box 350, Broome,WA 67250 Moved back to Perth? Peter Nielsen APB, PO Boddington, WA 63900 Peter O'Brien APB, Moora, WA 65100 Elisa Pemich APB, Shire of Corrigin, Corrigin W A 63 7 50 John Pierce Weed Science, Department of Agriculture Western Australia, South Perth, WA 61510 So Puzey APB, PO Box 254, Dalwallinu, WA 66090 Rod Randall Weed Science, Department of Agriculture Western Australia, South Perth, W A 61510 Ross Reading Robert Richardson Engineering Department, City ofPerth, PO Box C120, GPO Perth,WA 60000 Peter Robson APB, Lake Grace, WA 6353 0 Kathy Rogers APB, Northam, WA 6401. R. Salkilld APB, 4 Russ Street, Port Denison, W A 6525 0 Nadeem Samnakay APB, PO Box 254, Dalwallinu, WA 66090 Ron Sewell Juri en Greg Shea APB, c/o Department of Agriculture, Merredin, WA 64150 Dr Roger Shivas Plant Pathology, Department of Agriculture Western Australia, South Perth, WA 61510 Logan Stace APB, 4A Parade Street, Pingelly, WA 63080 Eddie Staporak APB, WAWA South West Highway, Waroona, WA62150 Ric Stellard Perth Lisa Stewart APB, PO Box 207, Goomalling, WA 65400 Dean Wainwright APB, PO Box 2, Boyup Brook, W A 6244 0 Tim Willing c/o Andrew Mitchell Noel Wilson APB, Pingrup, WA 63430