More on Flowerhead Weevils in Weed Biocontrol

In a previous issue [BNI 18(4)], we reported on the debate surrounding non-target effects of the introduced weevil, Rhinocyllus conicus, on native thistles in the USA. Amongst the items we presented, Dr Paul Boldt of the US Department of Agriculture described the history of this introduction against alien thistles and raised some questions regarding the interpretation of non-target effects. Here, scientists leading research on these effects give their views.

Insights from Data on the Nontarget Effects of the Flowerhead Weevil

In a recent comment¹, Boldt criticized our research documenting negative ecological effects by the introduced biocontrol agent, Rhinocyllus conicus². Here we respond to the criticisms and argue that our results can help improve the safety and efficacy of biological control.

None of the criticisms are warranted. First, our sample of flowering Platte thistles was randomized. Our transect-based, stratified sampling regime included all rosettes on transects that met the a priori criterion that they would flower within the season2. This method is not biased toward large plants as claimed, but representative of the flowering plant cohort.

Second, our seed mortality estimates were conservative. The average number of viable seeds produced by weevil-infested heads was 14.1% of that produced by those without weevils². Further evaluation of the effect of the weevil relative to native insects yields the same outcome (ANOVA, F_3,208 = 30.89, P < 0.001). Flowerheads without insects averaged 79.5 viable seeds in 1996. Flowerheads infested with only native tephritid flies averaged 27.3 (S.E. 6.20, N = 39), whilst flowerheads with only the introduced weevil averaged 4.9 (S.E. 1.85, N = 82), a five-fold reduction. Furthermore, when weevils fed on flowerheads with flies, number of viable seeds fell from 27.3 to 2.8 (S.E. 1.13, N = 83), a further 89.7% reduction. Also 36% more heads were attacked by weevils (N = 165) than by flies (N = 121), suggesting a higher frequency of attack by the weevil plus a higher impact per head. Since recruitment, fitness, and density of Platte thistle are proportional to number of viable seeds³, we conclude that a five-fold reduction in this relatively sparse native species is likely.

Third, our samples were adequate to evaluate weevil influence on seed production, both biologically and statistically. This is clear from the original paper², the analyses above, and multiple refereed publications (see ^{3, 4}). Site relocation can be important for fugitive species. However, here the spatial dynamics are seed-dependent, small-scale, and relatively slow. Median seedling establishment distance is under one metre³. The questioned duration of the study is actually a strength. Seed production and inflorescence insects have been quantified at Arapaho Prairie for over 20 years: 1976-1978 (see ⁴) and 1984-1997^{2, 3}. We also have demographic data for 1990-1998 (S. M. Louda, unpublished data). Thus, population swings in flowering, seed reproductive effort, flowering success and plant dynamics are exceptionally well documented. These data are the main reason the population influence of this weevil on this species can be confidently evaluated.

Fourth, Platte and other native prairie thistles are not serious weeds5. Native thistles have a long history of economic and environmental use. Globe artichoke is a delicacy. Thistle honey is a lucrative commodity. At least ten thistles have medicinal value. Many birds, including the goldfinch (Carduelis tristis), have intimate association with thistles, relying on seed heads for food and nesting materials. Three native finches have the Latin word for thistle, carduelis, as part of their name. And many US insects depend on thistle nectar and pollen, including those named after thistle hosts, such as the swamp metalmark (Calephelis muticum) named for Cirsium muticum and the painted lady butterfly (Vanessa cardui) named for thistles in general (Carduinae). The ecological costs of reducing Platte thistle densities likely extend further than we suggested.

Fifth, our prediction of impact on federally threatened Pitcher's thistle (Cirsium pitcheri) in the Great Lakes dunes rests on striking, documented parallels between Platte thistle, the putative progenitor species, and Pitcher's thistle. Not only are these sister species, they are similar in habitat requirements, life histories, and interactions with native insects^{6, 7, 8}. Most flowerheads of both species, for example, are initiated early, thus vulnerable to oviposition by R. conicus. Data on insect herbivore load on Pitcher's thistle (e.g. ^{6, 8}) suggest the effect on seed production of adding another inflorescence-feeder would resemble that for Platte thistle. Also, wind during the growing season, which could influence insects, is similar in velocity and persistence in the prairie dunes to those of the Great Lakes. The most parsimonious hypothesis is that Pitcher's thistle, already threatened by habitat destruction⁷ and limitation by native insects⁶ (also Louda & McEachern, unpublished data), will decline further if R. conicus joined the inflorescence-feeding guild. Smaller populations would increase the chance of extinction for this species.

What does this case suggest for future biocontrol projects? Biocontrol rests on several critical assumptions, including: (i) the target species poses major economic or environmental problems, (ii) no less risky alternatives exist, (iii) control by introduced natural enemies is predicted, (iv) significant harm to native species is unlikely, and (v) release involves known risks acceptable to the public. Although these assumptions are fundamental to good biocontrol, examination of Rhinocyllus and the musk thistle control programme suggests the validity of each of these must be better quantified.

Need for accurate quantification seems obvious in hindsight. For musk thistle, neither economic nor environmental data are currently adequate to weigh benefits and costs. Economic costs were not well quantified before Rhinocyllus was introduced. The best study⁹ was done after release and relied on county-level anecdotal information. Presence or absence on a scale of hundreds of square kilometres per county likely overestimated density of a patchy weed. And scoring a county as having a serious economic threat if "one or more pesticide applications had been used or would have been used if funds were available", or a potential economic threat if "the weed occurred but was not considered a problem" overestimated economic impact. Better estimates should be used today. Further, no environmental damage was likely. A weed of disturbed and overgrazed places, musk thistle loses in competition with grasses through normal plant succession and good pasture management^{10, 11}. More precise evaluations of pest problems are needed now.

The Rhinocyllus case also suggests a hierarchical pest management approach, in which the least risky options are implemented first. Several options, including mechanical control, localized spraying, and augmentation of indigenous natural enemies, exist as effective, less risky possibilities. All introductions carry some risk, so classical biocontrol could be reserved for the worst problems with lowest probabilities of nontarget effects.

Another insight from this case is that better evidence on the probability of control and on the potential for environmental side effects is needed. Host preference and range are necessary but insufficient evidence of safety. Given the existence of risk and potential interference among agents, the best biocontrol strategy would be to introduce the fewest and most effective agents with the lowest probability of nontarget effects. To do this, we need better prediction of both control and nontarget effects. For Rhinocyllus, evidence for control was equivocal¹² and subsequent studies have confirmed the importance of grass competition for thistle control (e.g. ^{10, 11}). Also, the magnitude of direct nontarget effects on related natives would have been better predicted if ecological criteria had been used to select the native species tested before release. Studying natives in which flowering was phenologically synchronized with Rhinocyllus oviposition would have warned about potential nontarget effects on Platte thistle13. Such studies suggest more field experiments in the indigenous region would improve the assessment of effectiveness and potential for host range expansion.

Thus, some lessons are clear. However, the direct effects, the lag in their development, and the indirect effects of R. conicus on native insects also suggest that anticipating the outcome of new interactions and the long-term consequences of introductions is difficult. It requires more information on the dynamics of interactions, activity of already introduced agents, and consequences of invasions. Incorporation of such information, plus increasingly rigorous economic and environmental assessments, should increase the safety and sustainability of biocontrol.

References

¹ Boldt, P. (1997) Response of a Rhinocyllus researcher. Biocontrol News and Information 18, 100N.

² Louda, S. M.; Kendall, D.; Connor, J.; Simberloff, D. (1997) Ecological effects of an insect introduced for the biological control of weeds. Science 277, 1088-1090.

³ Louda, S. M.; Potvin, M. A. (1995) Effect of inflorescence-feeding insects in the demography and lifetime fitness of a native plant. Ecology 76, 229-245.

⁴ Lamp, W. O.; McCarty, M. K. (1982) Predispersal seed predation of a native thistle, Cirsium canescens. Environmental Entomology 11, 847-851.

⁵ McCarty, M. K.; Scifres, C. J.; Robison, L. R. (1967) A descriptive guide for major Nebraska thistles. University of Nebraska Agricultural Experiment Station Bulletin 493, 1-24.

⁶ Keddy, C. J.; Keddy, P. A. (1984) Reproductive biology and habitat of Cirsium pitcheri. Michigan Botanist 23, 57-67.

7 Pavlovik, N. B.; Bowles, M.; Crispin, S. R.; Gibson, T. C.; Herman, K. D.; Kavetsky, R. T.; McEachern, A. K.; Penskar, M. R. (1992) Pitcher's thistle, Cirsium pitcheri, Recovery Plan. U. S. Fish and Wildlife Service, Region 3, Minneapolis MN.

⁸ Stanforth, L. M.; Louda, S. M.; Bevill, R. L. (1997) Insect herbivory on juveniles of a threatened plant, Cirsium pitcheri, in relation to plant size, density and distribution. EcoScience 4, 57-66.

⁹ Dunn, P. H. (1976) Distribution of Carduus nutans, C. acanthoides, C. pycnocephalus, and C. crispus, in the United States. Weed Science 24, 518-524.

¹⁰ Austin, M. P.; Groves, R. H.; Fresco, L. M. F.; Kaye, P. E. (1985) Relative growth of six thistle species along a nutrient gradient with multispecies competition. Journal of Ecology 73, 667-684.

¹¹ Popay, A. I.; Medd, R. W. (1990) The biology of Australian weeds 21. Carduus nutans L. Plant Protection Quarterly 5, 3-13.

¹² Zwoelfer, H.; Harris, P. (1984) Biology and host specificity of Rhinocyllus conicus (Froel.) (Col., Curculionidae), a successful agent for biocontrol of the thistle, Carduus nutans L. Zeitschrift für Angewandte Entomologie 97, 36-62.

¹³ Louda, S. M. (1998) Population growth of Rhinocyllus conicus (Coleoptera: Curculionidae) on two species of native thistles in prairie. Environmental Entomology 27(4), in press.

By: Svata M. Louda, School of Biological Sciences, University of Nebraska, Lincoln, NE 68588, USA.
Daniel Simberloff, Department of Ecology and Evolutionary Biology, University of Tennessee, Knoxville, TN 37996, USA.
George Boettner, Entomology Department, University of Massachusetts, Amherst, MS 01003, USA.
Jeff Connor, Rocky Mountain National Park, Estes Park, CO 80512, USA.
Deborah Kendall, Biology Department, Ft. Lewis College, Durango, CO 81301, USA.
Amy Arnett, School of Biological Sciences, University of Nebraska, Lincoln, NE 68588, USA.
Contact: Svata M. Louda
Email:
Fax: + 1 402 472 2083

 

Another Classic for Eucalyptus Psyllid

Eucalyptus psyllid (Ctenarytaina eucalypti) is a widespread, if not always serious, pest of eucalyptus, found in Africa, North and South America and Europe (particularly the Mediterranean basin), as well as in its native Australia/New Zealand. It has been the subject of a number of successful classical biological control programmes in the past, for example in the USA (California) and the UK.

It was first found in southeast France in 1994, and since 1995 has been recognized as a serious pest of eucalyptus plantations in Alpes-Maritime and Var. Species of eucalyptus have been grown for ornamental foliage since the beginning of the 1950s, and represent an important economic resource, in this area. Current production of some 700 tonnes of stems annually is worth some Fr10 million (UK£1 million). Chemical control proved to be impractical in the characteristically hilly terrain, and there were also concerns about the environmental impact of the insecticides.

In 1996, the INRA (Institut National de la Recherche Agronomique) laboratory at Antibes began to investigate the possibility of classical biological control. They considered natural enemies of the psyllid in its region of origin, and, in collaboration with the University of California at Berkeley, matched climatic and agronomic data for the putative area of introduction and the area of origin of prospective biocontrol agents. The operation identified an encyrtid parasitoid, Psyllaephagus pilosus from New Zealand, as the most promising candidate. This species had previously been introduced to California, which has a similar Mediterranean climate, where the psyllid is now under good control. Shipments of the parasitoid were imported into quarantine facilities at the INRA laboratory at Valbonne where they were mass-reared for release.

First releases were made in late April 1997, in a plot heavily infested with psyllids, in the heart of the eucalyptus growing area in the Tanneron Massif. Almost immediately it was realized that the dispersal and population growth of the introduced parasitoid were spectacular: one month after release, the whole of the 500 m² site was colonized. In less than three months, the parasitoid colonized the entire Massif, and was beginning to exert an excellent control of psyllid populations on plantations first colonized, with a rate of parasitism for older nymphs reaching 100% in some instances. By the end of the season, the whole of Alpes-Maritime and eastern Var had been colonized. In the original release sites, there was no sign of either psyllid or parasitoid at the end of July, and there was no sign of psyllid recolonization by the end of August. These results had an early impact on growers: many decided to abandon chemical control measures having seen the impact of the parasitoids.

The first year's results were extremely promising. The parasitoid had proved itself adapted to the local environment, capable of rapid population build-up and dispersal, and of having a rapid and dramatic impact on psyllid populations in the target area. However, the project aimed to provide a long-term sustainable solution to the psyllid problem, so two questions still remained to be answered: would the parasitoid survive the winter, and, if so, would it be able to multiply and disperse sufficiently quickly in the spring to prevent psyllid recolonization and damage. It was recognized that the parasitoid would need to react with a minimum of delay to control the psyllid below a threshold damage level. Although its great multiplication and dispersal rates made the outlook hopeful, the Valbonne laboratory continued to work on a rearing system that would allow them to make early season augmentative releases each spring, should these prove necessary.

However, in the spring of 1998, the parasitoid was recovered even further away than the previous autumn, more than 85 km west of the first site of release, on the Ile de Porquerolles in the Hyères group. It is now anticipated that the parasitoid will be found similarly far east, and will soon reach the Italian region of Liguria where eucalyptus are widespread. So far, therefore, results are very positive and expectations are high that this may represent another classical biological control solution for eucalyptus psyllid. The spectacular and swift success has been ascribed, at least in part, to the care that was taken in matching the potential biological control agents to the conditions and climate of southeastern France.

Sources: Malausa, J.-C.; Girardet, N. (1997) Lutte biologique contre le psylle de l'eucalyptus. Phytoma - La Défense des Végétaux, No. 498, pp. 49-51.
Malausa, J.-C. (1998) Des insectes au secours des eucalyptus. Biofutur 176, 34-37.

 

A New Mealybug Threat in the Caribbean

Paracoccus marginatus, a pest mealybug new to the Caribbean islands, is causing serious damage similar to that caused by the hibiscus mealybug (Maconellicoccus hirsutus) [see BNI 18(3)], although it does not occur on many islands yet. Paracoccus marginatus is a polyphagous species native to the Central American mainland that can damage numerous host plants including cassava; Carica papaya (pawpaw or papaya) is a favoured host.

The present distribution of P. marginatus in Central America includes Mexico, Guatemala, Belize and Costa Rica. The pest was first reported in the West Indies from St Martin (French Antilles) in May 1996¹. Since then, material from the British Virgin Islands has been identified and there are unconfirmed reports of its presence in several other islands.

It is evident from this news that authorities in the Caribbean will need to be vigilant for any new mealybug problem in addition to M. hirsutus. Although authoritative identification of P. marginatus is difficult (it may be confused with species of Pseudococcus), it is quite distinct from M. hirsutus, in having blackish body contents when squashed alive, and (in slide-mounted specimens) 17 pairs of cerarii and eight-segmented antennae. Maconellicoccus hirsutus has pink to grey body contents when squashed alive, and (in slide-mounted specimens) about five pairs of cerarii (on the abdomen only) and nine-segmented antennae.

¹ Matile-Ferrero, D; Etienne, J. (1996) Presence of the hibiscus mealybug, Maconellicoccus hirsutus in Saint Martin (Hemiptera, Pseudococcidae). Revue Francaise d'Entomologie 18(1), p 38.

By: Dr Gillian W. Watson, Taxonomist, Sternorrhyncha, CABI Bioscience, c/o Entomology Department, The Natural History Museum, Cromwell Road, London SW7 5BD, UK.
Email:
Fax: + 44 171 938 8937

Contact: Dr M. T. K. Kairo, CABI Bioscience Caribbean Regional Centre, Gordon Street, Curepe, Trinidad & Tobago, West Indies.
Email:
Fax: + 1 868 663 2859

 

Biological Suppression of Water Hyacinth in India

The free-floating weed water hyacinth, Eichhornia crassipes, was described first from southeast Brazil, although it was already widespread in occurrence in Central and South America. There is no definite report of the time of its entry into India, but 1888 or 1889 has been suggested, and it became established in Bengal around 1896. However, it was not recorded in a 1911 flora of the upper Gangetic plains and adjacent Shiwalik Hills, and apparently the weed spread in northern India much later. Sporadic occurrences were reported in Orissa in 1924. Today, water hyacinth occurs throughout India and is believed to occupy over 200,000 ha of water bodies in India.

This weed has a very serious impact on the use and management of water resources. The plants in dense infestations interfere with navigation and hydroelectric power generation. The flow of water is reduced by 40-95% in irrigation channels. Water hyacinth also interferes with seed germination and seedling establishment in paddy rice. In West Bengal alone the losses have been estimated at 110 million rupees (about US$3 million). It has also been reported from a study in West Bengal that about 45 million kilograms of fish were lost owing to water hyacinth. In another study, fish production was reported to be zero in a water body with 100% water hyacinth cover, and was reduced by about 75 per cent in a partially covered water body.

In India, the control programme is mainly conducted by the Project Directorate of Biological Control, Bangalore with funds provided by the Indian Council of Agricultural Research (ICAR), New Delhi, and in collaboration with the then Commonwealth Institute of Biological Control, UK for introduction of weed insects from different sources. The weevils Neochetina eichhorniae and N. bruchi, and the mite Orthogalumna terebrantis were introduced from Argentina in 1982. After elaborate host specificity tests and successful breeding in glasshouses, the weevils and mite were field released in 1983. These weed insects and mite became established at all release sites in India within a year.

Both N. eichhorniae and N. bruchi established readily under field conditions all over India. Successful biological control ranging from 90-98% was achieved in six different tanks with a total surface area of 1000 ha within four years after releasing about 25,000 adults in and around Bangalore (Karnataka State). Similarly in the Loktak Lake (286 km²) in Manipur, 75% weed cover was cleared within three years after the release of about 18,500 adults in 1987. The weevils have been distributed in 15 different states of the country, and success has been achieved with these weevils in Assam. Similarly Orthogalumna terebrantis has established and promising results are being obtained from several tanks covered with water hyacinth in Kerala and Karnataka States.

By: Dr S. P. Singh, Project Directorate of Biological Control (ICAR), P. B. No. 2491, H. A. Farm Post, Bellary Road, Bangalore - 560 024, India.
Email:
Fax: + 91 080 3411961