Session 1 Pre-Release Testing of Weed Biological Control Agents 3

Pre-release Studies and Release of the Cornops aquaticum in South Africa – a New Biological Control Agent for Water Hyacinth,

A. Bownes1, A. King2 and A. Nongogo3

Agricultural Research Council – Plant Protection Research Institute, Private Bag X6006, Hilton, 3245, South Africa [email protected]; [email protected]; [email protected]

Abstract

The grasshopper, Cornops aquaticum Brüner (: ) has recently been released in South Africa as a biocontrol agent for water hyacinth, Eichhornia crassipes (Martius) Solms-Laubach (), the country’s worst invasive aquatic weed. The release follows 15 years of pre-release studies to assess C. aquaticum’s safety and potential value as a new agent. Cornops aquaticum was first introduced into quarantine in South Africa from Manaus, Brazil in 1995. Host specificity testing was completed by 2001 but release of the grasshopper was delayed, initially, due to difficulties in obtaining release permits for weed biocontrol agents. A permit was finally granted in 2007 by which time pre-release efficacy studies had been initiated and new concerns over compatibility of C. aquaticum with the Neochetina (Coleoptera: Curculionidae) weevils, the most damaging agents in the field in both South Africa and other parts of Africa, had arisen. The efficacy and agent interaction studies were first concluded to guide the decision on whether C. aquaticum’s introduction into the country was justifiable. Pre-release impact studies indicated that C. aquaticum damage is directly associated with density and that herbivory at relatively low grasshopper densities can disrupt water hyacinth growth and productivity when growing under optimal nutrient conditions. Interaction studies with C. aquaticum and Neochetina eichhorniae Warner suggested a synergism whereby pairing of these agents, under laboratory conditions, had the greatest negative impact on biomass accumulation compared to the agents alone or other combinations of agents tested. In August 2010, the South African biocontrol community supported a decision to release C. aquaticum and field releases began early in 2011. Four initial release sites have been selected to encompass different nutrient and climatic conditions and are being monitored to assess establishment, impact and population dynamics of C. aquaticum.

Introduction 2011). A biological control programme for water hyacinth was initiated in 1974 with the release of the Water hyacinth, Eichhornia crassipes (Martius) petiole-mining water hyacinth weevil, Neochetina Solms-Laubach (Pontederiaceae) is a free-floating eichhorniae Warner (Coleoptera: Curculionidae). perennial herb, native to South America that was Following this, an additional four introduced into South Africa in the early 1900’s biocontrol agents were released between 1989 and via the ornamental plant trade. By the 1970’s it 1996: a leaf-mining mite, Orthogalumna terebrantis had reached pest proportions in many systems Wallwork (Acarina: Galumnidae) in 1989; another around the country and to date remains South petiole-mining weevil, Neochetina bruchi Hustache Africa’s worst invasive aquatic weed (Coetzee et al., and, a petiole-mining moth, Niphograpta albigutallis

XIII International Symposium on Biological Control of Weeds - 2011 4 Session 1 Pre-Release Testing of Weed Biological Control Agents

Warren (Lepidoptera: Pyralidae) in 1990; and a sap- support for the grasshopper’s release but new sucking mirid, Eccritotarsus catarinensis Carvalho concerns over compatibility of C. aquaticum with (Hemiptera: Miridae) in 1996 (Hill and Cilliers, 1999). the Neochetina weevils had arisen. The weevils are While South Africa’s biocontrol programme had the most important water hyacinth biocontrol agents a few successes, most water hyacinth sites around the in South Africa as well as other parts of Africa so country have been difficult to control biologically. any disruption to their efficacy or populations would By the end of the 1990’s, after almost 30 years of an have diminished prospects for the grasshopper’s active biological control programme, it was clear release. that success was variable and levels of control were This paper presents a subset of results from deemed unsatisfactory. Hill and Olckers (2001) the efficacy and agent interaction studies that outlined several factors that were speculated to primarily motivated the decision to proceed disrupt or reduce the efficacy of the biocontrol with the release of C. aquaticum into the South agents already released, the most important of African biocontrol programme for water which were eutrophication, where plant growth hyacinth. It also summarizes details of the first rates outpace the damage caused by the biocontrol releases of the grasshopper in South Africa. agents, and incompatibility of the agents with the temperate climate of many regions of the country. A Methods and Materials potential solution was to consider new agents in the hope that a better agent could be found or that the correct, complimentary suite of agents had not yet 1. Agent efficacy studies been released. 1.1 Effect of water nutrient levels on the im- The neotropical grasshopper, Cornops aquaticum pact of Cornops aquaticum herbivory Brüner (Orthoptera: Acrididae) was considered the most promising candidate based on reports on its Water hyacinth plants were grown in plastic tubs damage potential from the native range (Perkins, (43 x 31 x 19 cm) in a quarantine glasshouse for a 1974), and its wide distribution in South America, period of four weeks prior to the introduction of extending to climatically similar regions to South the . Each tub contained 15 L of water Africa (Adis et al. 2007). The first collections of and two water hyacinth plants and was enclosed the grasshopper took place in Manaus, Brazil in with a net canopy. Nutrient levels in the water were 1995 and subsequent collections were made in manipulated to represent levels of nitrates and Trinidad and Venezuela in 1997 and Mexico in 1997. phosphates present in South African water bodies. Oberholzer and Hill (2001) studied the host range Nitrogen and phosphorus were added as potassium of C. aquaticum by testing 64 plant species in 32 nitrate (KNO3) and potassium dihydrogen families and concluded that it is oligophagous within orthophosphate (KH2PO4) respectively. Commercial the family Pontederiaceae, with a strong preference chelated iron was also added at a rate of 1.3g/15L for water hyacinth. of water. The nutrient levels were classified as Although C. aquaticum was considered safe eutrophic (high), mesotrophic/eutrophic (medium) for release in South Africa based on its host and oligotrophic (low) (Table 1) according to the specificity, its release was initially delayed due to South African Water Quality Guidelines (Holmes, difficulties in obtaining release permits. Regulatory 1996). Water in the tubs was changed once a week to authorities delayed granting permits as they lacked maintain the required nutrient supply to the plants. in-house expertise needed to critically evaluate After the four-week growth period, all daughter release applications. A permit was finally granted plants, dead leaves and stems were removed, and in 2007, by which time agent efficacy studies had the plants weighed to determine wet weight. Adult already been initiated. It was decided to complete C. aquaticum grasshoppers were introduced into the this research in order to determine whether the experimental tubs at a density of one per plant and grasshopper’s introduction into South Africa was one male/female pair per tub. The treatments were justified based on its potential to be an effective replicated six times and the trial was run for a period biocontrol agent. The efficacy results showed strong of ten weeks. Plants were weighed at termination

XIII International Symposium on Biological Control of Weeds - 2011 Session 1 Pre-Release Testing of Weed Biological Control Agents 5

Table 1. Nutrient concentrations used to represent the range of levels found in South African river systems and impoundments. High Medium Low (eutrophic) (eutrophic/mesotrophic) (oligotrophic) Nitrates (mgL-1) 7.6 2.52 0.034 Phosphates(mgL-1) 1.37 0.316 0.024

of the trial to determine end wet weight. The effect 1.3 Density-damage relationships between of nutrient treatment, herbivory by C. aquaticum Cornops aquaticum and water hyacinth and their combined effect on the difference in wet weight from the start to the end of the trial were The experimental design followed the same analyzed using a two-way ANOVA. Tukey’s HSD protocol as trial 1.1 with the exception that all plants test was used as a post-hoc comparison of the means. were grown at the high nutrient level, when plant growth and productivity would be optimal. Male and 1.2 Effect of plant nutrient levels on Cornops female C. aquaticum grasshoppers were introduced aquaticum (a) survival and (b) fecundity into the tubs at a density of 2, 3 and 4 per plant (= four, six and eight grasshoppers per tub). The sexes were (a) Twenty-eight newly emerged C. aquaticum separated so that each tub had either only males or nymphs were reared on water hyacinth plants only females. Each treatment was replicated six times grown at the high, medium and low nutrient and two tubs per replicate were used as controls. levels (Table 1) for a period of three months. Water hyacinth plants were weighed at the start and Water and plant nutrient levels in water the end of the trial to determine wet weight and the hyacinth are highly correlated (Gossett and trial was run for a period of eight weeks. The plant Norris, 1971) so there was a corresponding biomass data was subjected to a regression analysis to increase in plant tissue nutrient levels with determine the relationship between biomass an increase in nutrient supply to the plants (as the independent variable) and plant biomass (as (Bownes, 2009). The total number of nymphs the dependent variable). Insect biomass was used as to survive to adulthood and the proportions a surrogate for insect density since densities of male of males and females were recorded. and female grasshoppers were the same. For this, a random sample of male and female grasshoppers (b) Eight pairs of adult grasshoppers reared in were weighed (males n =47; females n = 50) to trial (a) were confined in tubs with water hya- obtain a mean wet weight (g) for each sex (Bownes cinth plants grown at the same nutrient levels et al. 2010a). The biomass and insect data were fitted on which they were reared. The number of egg to a damage curve, similar to that suggested by packets oviposited by females and the number McClay and Balciunas (2005), and which is used to of nymphs to emerge from each egg packet identify agents that are not sufficiently damaging to were recorded. The number of egg packets their host plant to justify release. The damage curve per female and the number of nymphs per egg relates a critical aspect of weed performance such as packet were compared by one-way ANOVA to growth rate or final biomass to increasing densities test for the effect of nutrient treatment on fe- of the biocontrol agent. cundity of C. aquaticum. Tukey’s HSD test was used as a post-hoc comparison of the means.

XIII International Symposium on Biological Control of Weeds - 2011 6 Session 1 Pre-Release Testing of Weed Biological Control Agents

Table 2. Treatments testing interactions between Cornops aquaticum and two biocontrol agents already released in south Africa, Neochetina eichhorniae and Eccritotarsus catarinensis. Treatment Species combination (density/plant) Control No 1 CA (1 female) + NE (2pairs) 2 CA (1 female) + EC (10 adults) 3 NE (2 pairs) + EC (10 adults) 4 NE (2 pairs) 5 EC (10 adults) Table 3. Effect of nutrient levels on survival of Cornops aquaticum from first instar to adult, and sex ratio of survivors. Nutrient level Survival to adult Sex ratio Female:Male High 82% 65:35 Medium 71% 55:45 Low 64% 39:61

0.40

0.35 b

0.30 bd

0.25

0.20

0.15 a

0.10

ace 0.05 c

0.00 Mean difference inweight wet (kg) difference Mean High -0.05 Medium

ce -0.10 Low Experiment Control -0.15 Nutrient/herbivory treatment Figure 1. Mean change in wet weight of water hyacinth plants in response to herbivory by Cornops aquaticum. Plants grown at high, medium and low nutrient levels for ten weeks. Means with the same letter are not significantly different (Tukey’s HSD, P<0.05). Error bars represent the standard error of the mean.

XIII International Symposium on Biological Control of Weeds - 2011 Session 1 Pre-Release Testing of Weed Biological Control Agents 7

a A 5

4 b

3

c

2

Mean no. ofMean egg packetsper female 1

0 High Medium Low Mean Mean±SE Nutrient treatment

B

22 a

20

ab 18

16 b 14

12

10

8

6

Mean no.Mean of nymphs per egg packet 4

2

0 High Medium Low Mean Mean±SE Nutrient treatment

Figure 2. Mean fecundity of Cornops aquaticum females measured as (A) the no. of egg packets per female and (B) the number of nymphs to emerge from each egg packet. Means with the same letter are not significantly different (Tukey’s HSD test, P <0.05). Error bars represent the standard error of the mean.

XIII International Symposium on Biological Control of Weeds - 2011 8 Session 1 Pre-Release Testing of Weed Biological Control Agents

2. Agent interaction studies measured: number of leaves per plant, number of Effect of Cornops aquaticum on populations ramets per plant, proportion of petioles mined by and feeding damage of the Neochetina weevils Neochetina larvae and the total number of larvae and the impact of combinations of these agents recovered. The means of each parameter were on water hyacinth growth and productivity compared over time by one-way ANOVA and Tukey’s HSD test used as a post-hoc comparison.

(a) Manipulative, small-scale experiment One water hyacinth plant per treatment was Results grown in an 8 L bucket in a temperature controlled quarantine glasshouse. Nitrates and phosphates were 1. Agent efficacy studies added to the water at a rate of 2 mg L-1 and 0.29 mg 1.1 Effect of water nutrient levels on the L-1 respectively. Flowers and ramets were removed impact of Cornops aquaticum herbivory from the plants two weeks prior to the start of the trial to allow for the plants to recover and stabilise. Nutrient treatment (F2;29 = 48.53; P < 0.0001) Nutrients and water were replaced after the two- and herbivory (F1;29 = 81.80; P < 0.0001) had week recovery period and 25 days subsequent to this. a significant effect on the change in wet weight Table 2 shows the combinations and densities of the of water hyacinth plants (Fig. 1) from the start insect species that were tested. Plants were weighed at to the end of the ten week trial. The interaction the start and end of the trial to determine wet weight between nutrient supply and herbivory was also and were sampled weekly to record insect activity significant (F2;29 = 5.56; P = 0.009). Plant tolerance such as the number of N. eichhorniae feeding scars. to herbivory by C. aquaticum increased with an Each treatment was replicated 10 times and the trial increase in nutrient supply to the plants however was run for a period of 50 days. To analyze the effect feeding by the grasshoppers significantly reduced of the different insect treatments on the change in biomass accumulation at all three nutrient levels. wet weight of water hyacinth plants, the biomass data were compared by one-way ANOVA. The mean 1.2 Effect of plant nutrient levels on Cornops number of weevil feeding scars when in combination aquaticum (a) survival and (b) fecundity with C. aquaticum and E. catarinensis and alone were also compared by one-way ANOVA. Tukey’s HSD (a) Survival of C. aquaticum nymphs to test was used as a post hoc comparison of the means. adulthood was influenced by plant nutrient levels which also had an effect on the proportions of males (b) Pond experiment with an already estab- and females to be reared through to adulthood lished weevil population (Table 3). Higher levels of nitrogen in the plant tissue (Bownes, 2009) elicited higher rates of The trial was conducted in a 1300 L portable survival and greater numbers of females survived to pool (215 x 45 cm) housed in a semi-quarantine adulthood in the high nutrient treatment (Table 3). glasshouse and enclosed with a net canopy to confine the insects to the plants. The pool was 100% (b) Nutrient treatment had a significant effect covered with water hyacinth and had a combined (F2;20 = 26.06; P < 0.0001) on fecundity of female density of 2.6 (± 0.87) adult N. eichhorniae and grasshoppers that were reared and maintained, after N. bruchi per plant that were resident for a period pairing at adulthood, on plants grown at the high, of 3 months prior to the introduction of the medium and low nutrient levels, with fewer egg grasshoppers. At the start of the trial, 97 adult C. packets being produced at the low nutrient level. aquaticum (46 females: 51 males) were released Nutrient treatment had a significant effect (F2;18 onto the plants which equated to 0.3 grasshoppers = 7.58; P = 0.0041) on the number of nymphs to per plant. A random sample of five water hyacinth hatch from egg packets of females (Fig. 2). The mean plants were destructively sampled fortnightly and number of nymphs per egg packet increased with an the following plant and insect parameters were increase in nutrient supply to the plants although

XIII International Symposium on Biological Control of Weeds - 2011 Session 1 Pre-Release Testing of Weed Biological Control Agents 9 only the high and low treatments were statistically and feeding damage of the Neochetina weevils significantly different from one another. and the impact of combinations of these agents on water hyacinth growth and productivity 1.3 Density-damage relationships between Cornops aquaticum and water hyacinth (a) Manipulative, small-scale experiment

The relationship between plant biomass at A combination of C. aquaticum and N. the end of the trial as a function of increasing C. eichhorniae was the only treatment to significantly aquaticum biomass was curvilinear (Fig. 3). Biomass reduce biomass accumulation of water hyacinth of water hyacinth plants decreased with an increase plants compared to control plants (F5;54 = 3.62; P = in feeding intensity by the grasshoppers. Exponential 0.0068). Although the other combinations of insects regression best described the relationship between or treatments with N. eichhorniae and E. catarinensis plant yield and insect biomass and was highly alone hampered biomass accumulation relative to significant (F6;43 = 73.20; P < 0.0001) accounting insect free plants, none of these differences were for 75% of the variance (Bownes et al. 2010a). statistically significant (Fig. 4). The presence of C. aquaticum had no effect on N. eichhorniae feeding 2. Agent interaction studies intensity compared to when the weevils were alone Effect of Cornops aquaticum on populations or in combination with the mirid E. catarinensis.

0.7

r2 = 0.747 0.6 y = 0.074 + 0.420(0.089x)

0.5

0.4

0.3 E. crassipesE. (kg/plant) plants

0.2

0.1 End weights of

0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Insect biomass (g/tub)

Figure 3. Regression of Cornops aquaticum biomass (g) and final weight (kg) of water hyacinth plants after eight weeks. Insect biomass is represented by a mean weight of male or female grasshoppers multiplied by the respective density.

XIII International Symposium on Biological Control of Weeds - 2011 10 Session 1 Pre-Release Testing of Weed Biological Control Agents

0.36 0.34 b 0.32 0.30 ab 0.28 ab 0.26 0.24 ab ab 0.22 0.20 0.18 0.16 a 0.14 0.12 Biomassaccumulation (kg) 0.10 0.08 0.06 0.04 Control CA + EC CA + NE EC + NE EC NE Treatment Figure 4. Mean change in wet weight of water hyacinth plants exposed to different combinations of insect biocontrol agents (Cornops aquaticum, Neochetina eichhorniae and Eccritotarsus catarinensis). Means with the same letter are not significantly different (Tukey’s HSD, P <0.05). Error bars represent the standard error of the mean.

40

a 35 a

30

25

20

15

Total number of feeding scars of Totalnumber 10 a

5

0 With CA With EC Alone Treatment

Figure 5. Mean numbers of Neochetina eichhorniae feeding scars when tested alone and in combination with Cornops aquaticum and Eccritotarsus catarinensis. Means with the same letter are not signifficantly different (Tukey’s HSD, P <0.05). Error bars represent the standard error of the mean.

XIII International Symposium on Biological Control of Weeds - 2011 Session 1 Pre-Release Testing of Weed Biological Control Agents 11

1.4 Ramets 11 Leaves a 1.2 10

ab 1.0 9 ab ab

0.8 b 8

0.6 7

0.4 6 Numberof leaves/plant Numberof ramets/plant c 0.2 5

0.0 4 08 Dec 26 Dec 08 Jan 21 Jan 05 Feb 19 Feb Date

Figure 6. Mean numbers of ramets (daughter plants) and leaves produced by water hyacinth plants over time in response to feeding by the Neochetina weevils and Cornops aquaticum. Means with the same letter are not significantly different (Tukey’s HSD, P <0.05).

0.9 14 Petioles mined c D Larvae

0.8 12 bc 0.7 10

0.6 abc C 8 ab ab 0.5 BC 6 0.4

a AB 4 0.3 Total number of larvae/plant AB Proportionof petioles mined/plant A 0.2 2

0.1 0 08 Dec 26 Dec 08 Jan 21 Jan 05 Feb 19 Feb Date

Figure 7. Proportions of petioles mined by Neochetina larvae and numbers of Neochetina larvae per water hyacinth plant when in combination with Cornops aquaticum. Means with the same letter are not significantly different (Tukey’s HSD,P <0.05). Error bars represent the standard error of the mean.

XIII International Symposium on Biological Control of Weeds - 2011 12 Session 1 Pre-Release Testing of Weed Biological Control Agents

Interestingly, the presence of E. catarinensis invested in developing this agent, a decision was caused a reduction in feeding of N. eichhorniae; made in August 2010 to proceed with the release of however the differences in the number feeding the grasshopper. scars were not statistically significant (Fig. 5). For the initial releases and monitoring of C. aquaticum, four sites were selected to encompass (b) Pond experiment with an already estab- a range of both nutrient and climatic conditions lished weevil population such as high or low water nutrient conditions and temperate to sub-tropical climates. All four sites Plant productivity, as measured by the were monitored for at least nine months prior to number of leaves and ramets per water hyacinth release in order to evaluate site-specific conditions plant decreased over time, although this was only such as microclimate, water nutrient conditions statistically significant for the number of leaves per and the status of the plants and insect biocontrol plant (F5;24 = 12.39; P < 0.0001). Although there agents already present. The first release took place in were no significant differences, ramet production January 2011 followed by a second release in March of the water hyacinth plants ceased after 5-6 2011. Three hundred adult and late instar nymphs weeks (Fig. 6), and by the 10th week, most of the were released at each site. Sites were monitored plants in the pool had died back. Neochetina larval three months post-release, but to date evidence activity increased over time after the introduction of establishment of the grasshoppers has not been of C. aquaticum. There was a significant increase found. in the proportion of petioles mined (F5;24 = 6.80; With assistance with mass rearing from the P = 0.0004) and in the total number of larvae (F5;24 South African Sugar Research Institute (SASRI) = 24.38; P < 0.0001) recovered from the plants from which has a specialized rearing facility for insect the start of the trial to the last sampling event (Fig. 7). biocontrol agents, repeated releases will take place during the summer of 2011/2012. All sites will be monitored on a quarterly basis to determine Discussion establishment and efficacy of C. aquaticum as well population dynamics of the insect biocontrol agents The results from these and other studies (Bownes, already present on water hyacinth in South Africa. 2009; Bownes et al. 2010b) strongly suggested that C. aquaticum has the potential to be a valuable Acknowledgements biocontrol agent for water hyacinth in South Africa

and the following conclusions were made: (1) C. The Working for Water (WfW) Programme of the aquaticum has the potential to reduce populations of Department of Environmental Affairs is gratefully water hyacinth under eutrophic nutrient conditions acknowledged for funding research on this agent. and that these nutrient conditions should have a Prof. Martin Hill and Prof. Marcus Byrne are positive effect on their population dynamics; (2) thanked for their guidance on certain aspects of the the damage caused by C. aquaticum is density- pre-release research which contributed to a PhD dependent in that increasing densities will lead to a degree. The authors also thank WfW implementation corresponding reduction in water hyacinth growth officers, Daleen Strydom and Ryan Brudvig for their and productivity, supporting the conclusion that assistance with locating suitable field release sites it would be sufficiently effective to warrant release and with field work. (McClay and Balciunas, 2005); (3) C. aquaticum does not appear to have a negative effect on feeding and populations of the Neochetina weevils; and (4) an References apparent synergism between C. aquaticum and the Neochetina weevils could potentially lead to better Adis, J., Bustorf, E., Lhano, M.G., Amedegnato, C. levels of control of water hyacinth in South Africa. & Nunes, A. (2007) Distribution of Cornops On the basis of these conclusions and on the fact that grasshoppers (Leptysminae: Acrididae: a substantial amount of time and resources had been Orthoptera) in Latin America and the Caribbean

XIII International Symposium on Biological Control of Weeds - 2011 Session 1 Pre-Release Testing of Weed Biological Control Agents 13

Islands. Studies on Neotropical Fauna and Hydrobiologia 38, 15–28. Environment 42, 11–24. Hill, M.P. & Cilliers, C.J. (1999) A review of the Bownes, A. (2009) Evaluation of a plant-herbivore arthropod natural enemies, and factors that system in determining potential efficacy of a influence their efficacy, in the biological control candidate biological control agent, Cornops of water hyacinth, Eichhornia crassipes (Mart.) aquaticum for water hyacinth, Eichhornia Solms-Laubach (Pontederiaceae), in South crassipes. PhD dissertation, Rhodes University, Africa. African Entomology Memoir 1, 103–112. Grahamstown, South Africa. Hill, M.P. & Olckers, T. (2001) Biological control Bownes, A., Hill, M.P. & Byrne, M.J. (2010a) initiatives against water hyacinth in South Africa: Assessing density-damage relationships between constraining factors, successes and new courses water hyacinth and its grasshopper herbivore. of action. In Proceedings of the Second Meeting Entomologia Experimentalis et Applicata 137, of the Global Working Group for the Biological 246–254. and Integrated Control of Water Hyacinth (eds Bownes, A., Hill, M.P. & Byrne, M.J. (2010b) Julien, M.H., Hill, M.P., Center, T.D. & Jianqing, Evaluating the impact of herbivory by a D.), pp. 33–38. ACIAR, Canberra, Australia. grasshopper, Cornops aquaticum (Orthoptera: Holmes, S. (1996) South African Water Quality Acrididae) on the competitive performance Guidelines. World Wide Web. http://www.dwaf. and biomass accumulation of water hyacinth, gov.za/IWQS/wq_guide/index/html Eichhornia crassipes (Pontederiaceae). Biological McClay, A.S. & Balciunas, J.K. (2005) The role Control 53, 297–303. of pre-release efficacy assessment in selecting Coetzee, J.A., Hill, M.P., Byrne, M.J. & Bownes, A. (2011) A review on the biological control classical biological control agents for weeds – the programmes on Eichhornia crassipes (C.Mart.) Anna Karenina principle. Biological Control 35, Solms (Pontederiaceae), 197 – 207. D.S. Mitch (Salviniaceae), Pistia stratiotes L. Oberholzer, I. G. & Hill, M.P. (2001) How safe is the (Araceae), Myriophyllum aquaticum (Vell.) grasshopper for release on water hyacinth in South Verdc. (Haloragaceae) and Azolla filiculoides Africa? In Proceedings of the Second Meeting of Lam. (Azollaceae) in South Africa. African the Global Working Group for the Biological and Entomology 19, 451–468. Integrated Control of Water Hyacinth (eds Julien, Gossett, D.R. & Norris, W.E. (1971) Relationship M.H., Hill, M.P., Center, T.D. & Jianqing, D.), pp. between nutrient availability and content of 82–88. ACIAR, Canberra, Australia. nitrogen and phosphorus in tissues of the aquatic Perkins, B.D. (1974) that stress water hyacinth. macrophyte, Eichhornia crassipes (Mart.) Solms. Pest Articles and News Summaries 20, 304–314.

XIII International Symposium on Biological Control of Weeds - 2011