Neem and Papaya Fruit Extracts and Ferric Phosphate for Control of Golden Apple Snail in Wetland Taro: Efficacy Testing

Final Report for SW03-010

Project Type: Research and Education
Funds awarded in 2003: $31,831.00
Projected End Date: 12/31/2004
Matching Non-Federal Funds: $6,750.00
Region: Western
State: Hawaii
Principal Investigator:
Lance Santo
Hawaii Agriculture Research Center
Mel Jackson
Hawaii Agriculture Research Center
Expand All

Project Information


Water-based extracts of neem and papaya fruits were tested for their effectiveness in the control of the Golden Apple snail (Pomacea canaliculata) in wetland taro cultivation systems. Water extracts were tested at two rates and in combination. Neem extracts were found to be most effective with in excess of 90% snail mortality observed over the test period. No phytotoxicity was observed, however, younger taro plants were found to be much more susceptible to attack than older ones. In addition, a commercially available ferric phosphate formulation was tested. A significant effect on mortality, egg laying and protection of plants was not seen.

Project Objectives:

1. To determine the efficacy of various formulation mixes and rates of application of neem (Azadirachta indica ) and papaya (Carica papaya L.) extracts on apple snail populations in commercial lo'i.

2. To determine the effects of the various formulations selected from Objective 1 on non-target species residing within the lo'i.

3. To disseminate the results of the project to taro growers through a field demonstration, publications and web sites and through grower association seminars.

4. To initiate the EPA registration process of a botanical for use in taro lo'i

5. To determine the efficacy of NEU1165M Slug and Snail Bait™ (1% ferric phosphate) on apple snail populations in commercial lo'i at the currently registered label rate and at a 5X exaggerated rate.

6. To determine the effects of NEU1165M Slug and Snail Bait™ (1% ferric phosphate) on non-target species residing within the lo'i.

7. To initiate the EPA process (through IR-4 and the manufacturer) of amending a current label so that that this material is registered for use in wetland taro lo'i


Hawaii taro production in 2003 was estimated at 5.0 million pounds, an 18% decrease from 2002 (Hawaii Agricultural Statistics Service, Feb 9, 2004). This is the lowest ever recorded production and was in large part due to the effects of the Golden Apple snail (Pomacea canaliculata). The value of sales in 2003 was about $2.7 million, compared with $3.294 million the previous year. Golden apple snail was introduced into Hawaii, Japan and many other countries in Southeast Asia from South America as a source of food in the early 1980s. However, after its commercial markets had failed, discarded and escaped snails invaded taro and rice ecosystems and have been causing significant economic damage. In Hawaii, these snails were also purposely introduced into taro paddies (lo`i), the reasoning being that they could be harvested for food (Tanji, 1990). However, the consequences of this action were not fully understood at the time. The snails are voracious, fast growing and have a huge reproductive potential. A single female can produce as many as 15,000 offspring per year, and can thrive in water at a density of 1,000 snails per square meter (Anderson, 1993). They mature within 60 – 85 days and spawn at weekly intervals and have been described as the most damaging pest ever to hit neotropical areas (Halwart, 1994). The snails very quickly spread throughout taro lo`i, via the extensive and interconnected irrigation system (Tanji 1990, Ashizawa, 1992). In 1996 the Hawaii State Department of Agriculture statistical services recorded that in 1992 approximately 60,000 lb. of fresh taro was marketed from the islands of Oahu, Molokai and Maui, yet in 1996, only approximately 10,000 lb. was marketed; an 84% decline, largely due to apple snails. This picture of rapid and overwhelming infestation is reflected elsewhere in the world. For example, $1 million has been spent annually to control the snails in rice paddies in Taiwan since 1982 (Cheng, 1989). In addition, an estimated 20% of farm income was spent on apple snail control in the Philippines in 1993 (Anderson, 1993). Because taro grows in water where there are aquatic animals sharing the same ecosystems, chemicals are not allowed for pest control.

Biological control agents such as predatory animals and parasites are not effective for control of this snail. Similarly, conventional pesticides have been found to be non-specific and can pose significant ecological problems. In 1994 a formulation of copper sulfate was registered for use in taro lo'i (EPA registration number 1109-21), but was soon discontinued due to its non-specific and highly toxic nature. Extracts from natural products have been screened for potential molluscicide activity throughout Asia. For instance, in the Philippines, Starflower (Calotropis gigantea) was tested and showed anti-molluscicide activity (Lobo, et al., 1991), however, the effective rate of application was very high (200Kg/ha.). Morallo-Rejesus and Punzalan, (1997) examined a total of 138 extracts from Philippine plants as potential molluscicides, including extracts from neem leaf. While some were found to be effective, none were subsequently followed up. In addition, mechanical control has also been studied (Awadhal and Quick, 1994), however, this is an extremely labor intensive practice.

In preliminary research undertaken by this laboratory in 2001 and 2002, neem (Azadirachta indica) and papaya (Carica papaya) extracts were shown to be toxic to the Apple snail. The active ingredient in papaya was shown to be the proteolytic enzyme papain. The active ingredients in the neem extract were thought to be the azadirachtins, but this was not demonstrated due to a lack of suitable purified materials.

The original proposal to study the effects of papaya and neem extracts was completed within the original time frame of the grant. The results were discussed with the EPA's Biopesticide Division and the consensus in the EPA was that even though the extracts are botanical in nature, the neem extract would require a full registration procedure. This would include toxicological studies on non-target organisms and a full residue study on the crop, before the extract could be registered for use on wetland taro. A study of this nature usually costs millions of dollars and takes many years to complete. Unfortunately, there is no current neem extractor willing to invest in a botanical extract with such a limited market in the U.S. However, in the meeting with the EPA, the chief of the Biopesticide Division suggested trying a relatively new inorganic biochemical molluscicide with ferric phosphate as the active ingredient. The product is currently registered by the EPA (Registration # 67702-3) for terrestrial snail and slug control. Ferric phosphate is found to be ubiquitous in the environment. The EPA waived a number of ecological effects toxicity data requirements because of the known lack of toxicity of iron phosphate to birds, fish and non-target insects, its low solubility in water and the conversion of the ferric form to the even lower soluble ferrous form in soil. An acute oral toxicity study of the Bobwhite quail indicated that ferric phosphate was practically non-toxic to avian species. Based on the above factors, the data requirements for toxicity studies in Mallard duck, rainbow trout, freshwater invertebrates and non-target insect/honeybees were waived (EPA Technical Document 034903, October, 1998).

A sample of this material was received from the manufacturer (W. Neudorff GmbH KG, Germany) and preliminary laboratory data suggested that the current formulation (extruded pellet) containing 1% active ingredient is stable in water for more than 7 days at ambient temperature. The material does not float and so is ideal for targeting aquatic snails. Stability in a taro lo'i should therefore not be a problem. This project, in addition to testing botanical extracts, also sought to test the efficacy of the ferric phosphate formulation on Golden Apple snails.


Click linked name(s) to expand/collapse or show everyone's info
  • Adam Asquith
  • Rodney Haraguchi
  • Roy Yamakawa


Materials and methods:

Production of Water extracts of Papaya and Neem:
Ripe neem fruits were harvested at HARC's Kunia experiment station from a stand of neem trees planted approximately 15 years previously. Water extracts of whole neem fruits were made by homogenizing ripe fruits in the presence of water at a ratio of 1 Kg of fresh fruit to 1 liter of water. The resulting mixture was filtered and stored refrigerated prior to use in the field. The neem extract appearance is shown in a photograph in Appendix B. Papaya extracts were made from green (unripe) papayas at HARC's Kunia experiment station. Previous studies had shown that the latex from green papaya had a significant effect on snail mortality, with 100% mortality within 48 hours after application in an aquarium system in the laboratory. Latex was obtained by scoring green papaya fruits and collecting the milky latex that exuded. The latex was dried at room temperature and then pulverized to form a white powder. The dried papaya latex was stored in an airtight container prior to dilution and use in the field. Neither the neem nor papaya fruits had been exposed to pest control chemicals and therefore the potential of confounding any results due to chemical contamination was avoided.

Determination of Efficacy of water extracts and Liquid formulations (Objective 1)
Wetland taro lo'i that are infested with large snail populations were selected for the field testing of the extracted materials. Two cooperators on the island of Kauai (Mr. R. Haraguchi and Dr. Adam Asquith) agreed to cooperate on this project, allowed use of lo`i at their farms for this test and provided day-to-day maintenance. (Dr. Asquith owns a taro farm and is also an ag. extension agent for the University of Hawaii.) Prior to the initiation of the test, the lo'i (taro at approximately six months after planting) was drained such that only about a one inch depth of water remained. This is a typical practice, prior to fertilizer application. Flow into and out of the lo'i was stopped for three days after the initiation of the test. Circular areas (four feet in diameter) were physically separated from the rest of the lo'i by placing four-foot-high wire mesh around the perimeter and lining the interior of the perimeter with heavy gauge plastic. The structure did not need any further stabilization, providing a sturdy barrier to snails leaving or entering the test area. The mesh served to trap the snails within the perimeter, and the plastic sheet prevented the escape of snails by burrowing and minimized drift of test substances from the test area. The base of the mesh and plastic sheeting was buried in mud bottom of the lo'i. Each circular test plot contained four taro plants, based upon existing commercial spacing in each of the test locations. So that results were not confounded all snails and eggs within the test areas were removed. Each test area was then stocked with 20 female snails, recognized by the concave shape of the operculum, (in contrast to the males, which have a convex lip on the outer edge of the operculum) and 5 males. Only snails with a shell size greater than ½ inch were placed in the test plots. Seven circular test plots were set up for each test. Test material applications were made to each of the plots. Each plot was subjected to one of the following applications: Papaya extract only (1 x rate), Neem extract only (1 x rate), Papaya extract only (5x rate), Neem extract only (5 x rate), Papaya/Neem (50/50 mix at the 1 x rate), Papaya/Neem (50/50 mix at the 5 x rate) and an untreated control. For the water extract of neem fruits, the 1 x rate was 50 mls of extract per test plot, the 5 x rate was an application of 250 mls of extract per test plot. For the papaya extract, a 1 x rate was 10 grams of latex dissolved in 250 mls of water, and the 5 x rate was 50 grams of latex dissolved in 250 mls of water. Applications were made using a watering can with a perforated nozzle, giving a fin shower of test material into the test plots at application. Duplicate test plots were arranged in a randomized complete block design within each lo`i.

There were two locations with identical treatments, and the overall test design for each location is given in figures 1a and 1b in Appendix A. In addition, photographs of each of the test locations are given in Appendix B.

After the initial application, water to the lo'i remained shut off and the lo`i kept partially drained for three days. After this time, the lo'i were re-flooded but the test circles remained intact, except for the inner plastic sheeting, which was removed to allow for free flow of water. The mesh remained to prevent snails from leaving the test plots. Snail and egg counts were made 0, 1, 3, 9, 35 and 66 days after the application of test substance. In counting the snails, the number of dead snails vs. live was recorded. The total cumulative number of egg clusters (each cluster containing between 50 and 100 eggs, and each cluster representing a single egg-laying event) were counted. In addition, the number of plants remaining in each of the test plots was counted at each observation time. Any obvious phytotoxicity was looked for throughout the test period.

Determination of Efficacy of NEU1165M Slug and Snail Bait™ (Objective 5)

The ferric phosphate-based snail bait was tested in the same manner as the liquid extracts from papaya and neem. The same circular test plot designs were used as for the botanical extracts. Mr. Haraguchi and Dr. Asquith both consented to let lo'i on their taro farms be used for the tests, to provide day-to-day maintenance and to aid in counting snails, plants and eggs.

After clearing each plot of snails and eggs, 10 female and 5 male adult snails were placed within each plot. Eight circular areas were set up for each test, with four duplicate test applications at each location. The lo'i used for the test were different from the ones used previously when testing the botanical extracts. Each test plot was subjected to one of the following treatments: test material at the label rate (1 lb. per 1,000 sq. feet), exaggerated 5X label rate (5lb. Per 1,000 sq. feet), one-half the label rate (0.5 lb. per 1,000 sq. feet) and an untreated control. Duplicate treatments were arranged in a randomized complete block design were set up in each lo`i. Figures 2a and 2b in Appendix A show the overall layout of the test plots at each location. Photographs are shown in Appendix B.

After the initial application, water to the lo'i remained shut off and the lo`i kept partially drained for three days. After this time, the lo'i were re-flooded but the test circles remained intact, except for the inner plastic sheeting, which was removed to allow for free flow of water. The screen remained to prevent snails and non-target species from leaving the test plots.

Observations were made 3, 7 and 21 days after the test substance application. Snail counts were made of live vs. dead snails, egg cluster counts were made and the number of plants remaining in each test plot were counted. In addition, any phytotoxicity was noted.

Research results and discussion:

Effect of Water Extracts of Papaya and Neem

The results of the observations made over the 66 day period following the test application are shown in Table 1, Appendix A.

At the first location (Dr. Asquith's farm), the application was made in rainy weather, however, as the formulations were in liquid form, this had no consequence. No difficulties were encountered during the applications, with the exception of the observation that due to the uneven nature of the lo'i, the area where the test plots had been placed had an average depth of 3 inches, not the prescribed 1 inch set out in the protocol. The plants at this location had only been in the lo'i for 2-3 weeks and consequently were only about 1 1/2 feet high and consisted of a single stem splitting into two, with 2 leaves. The test at this location was terminated after 35 days, because at that time it was observed that the majority of the plants had been consumed by the snails.

The results shown in Table 1, Appendix A clearly demonstrated that the 1x and 5 x applications of the neem extract had a significant impact on snail mortality. The data show that after 35 days post-application, in the duplicate test plots where 5X neem/papaya combination applications were made (plots 6A and 6B), 17 of 25 and 11 of 25 snails, respectively, were dead. This gives a combined % mortality of 56%. This was the highest combined % mortality seen at this location, followed by the 5 x neem application (54%), the 1 x neem application (44%), the 5 x papaya application (32%), the 1 x papaya application (24%) and finally the control with 12% mortality. When the mortality data for this location were subjected to analysis of variance (ANOVA) the difference in mortality between the treatments was highly significant (p=0.0215). From this data it seems that the neem alone was almost as effective as the combination of neem and papaya. It is interesting to note that papaya extracts were extremely effective in the laboratory aquaria tests undertaken prior to this study, with 100% mortality within 3 days. This was not the case in this particular field test. A number of factors could have contributed to this difference, including the fact that in the field there is layer of muddy, floculant soil under the water. This soil has a high organic content, which may have adsorbed the non-specific protease papain, in high concentration in the papaya extract, attenuating its effect on the snails.

No significant difference could be seen in the number of egg clusters laid between the different treatments during the 35-day observation period. Even though snails died in the neem treatment, the remaining snails continued to lay eggs.

The reason for terminating the trial in this location by day 35 was that the number of plants remaining was too low to garner any further meaningful results. Out of a total of 56 plants that started in this trial, only 5.5 plants remained; less than 10%. Even though snail mortality was significant, the remaining snails and the relatively fragile nature of the young plants meant that most were consumed. This indicates that this treatment might be more effective on older and hardier plants. In addition, it had been noted that the depth of the water in this location was 3 inches. There may have been a dilution effect bringing the concentration of active ingredients under the threshold for an effect that may have killed more snails and so avoided the crop loss. Dr. Asquith therefore agreed to re-plant the area. The test was repeated with only the neem treatments and negative controls. The depth of the water in each of the test areas was measured and the amount of material added was changed to compensate for depth differences. Three test plots had neem treatments at the depth compensated 5 x rate and two test plots had no applications made and were the controls. Observations were made 3 days and 7 days after the application and are shown in Table 2 of Appendix A. It can be seen that when the depth of water is compensated for, the results are very striking. Three days after the application, one of the three 5 x rate application tests showed 22 of 25 snails dead, while the other two showed 8 dead and 5 dead, respectively, while in the controls there were 0 dead and 3 dead, respectively. Seven days after the application, the three test plots exposed to the 5 x application showed 23, 21 and 20 dead snails, respectively, giving an average mortality rate of 85.3%. In contrast, the two control plots there were 0 dead and 5 dead respectively, giving an average mortality rate of 6.6%. It therefore seems that if the neem extract were used commercially, particular attention would have to be paid to the drained depth of the lo'i in order maximize efficacy. The increased snail mortality in the test plots where the neem extract was applied showed a total of 8 egg clusters laid in the 7-day period. These were all laid within the first three days. This gives an average egg laying rate of 2.6 clusters per plot, and compares with 18.5 clusters per plot in the controls.

At the second location (Mr. Haraguchi's farm), taro is grown on a much larger scale than in the the first location, which is a more typical single family series of plots totaling less than an acre. In contrast Mr. Haraguchi's farm is the largest in the state, and employs more technology in the cultivation practices. For this reason, the lo'i are all absolutely level. Consequently Mr. Haraguchi was able to lower the water level to exactly 1 inch for the field test. The lo'i that was chosen had plants in it that were approximately 4 months old and had an average height of about 3 feet, with multiple stems and leaves. Test plots were set up as previously described, with 4 ft. diameter circles encompassing 4 plants. Applications of test substances were made as previously described and on the same day as those made in the first location. Observations were made 0, 1, 3, 9, 35 and 66 days after the applications were made. The results are shown in Table 3, Appendix A and show that the neem extracts were the most effective in terms of snail mortality. The data show that after 66 days post-application, in the duplicate test plots where 5X neem/papaya combination applications were made (plots 6A and 6B), 25 of 25 and 23 of 25 snails, respectively, were dead. This gives a combined % mortality of 96%. This was the highest combined % mortality seen at this location, followed by the 5 x neem application (90%), the 1 x neem application (58%), the 5 x papaya application (42%), the 1 x papaya application (40%) and finally the control with 16% mortality. When the mortality data for this location were subjected to ANOVA the difference in mortality between the treatments at 66 days was highly significant (p=0.0029). The difference in effect of the neem treatment as little as 3 days after the application was striking, with 78% mortality in the 5X neem/papaya combination applications, compared with 66% in the 5 x neem applications, 22% in the 5 x papaya applications, 12% in the 1 x neem applications and 2% in the controls. ANOVA showed that the differences in snail death at 3 days were also highly significant (p= 0.0085). Similarly, egg laying was severely curtailed in the neem application plots. Over the 66 day period a total of 2 egg clusters were observed in the 5 x neem/papaya combination application, compared with 26 in the 5 x neem application, 31 in the 1 x neem application, 96 in the 5 x papaya application, 112 in the 1 x papaya application and 108 in the control. This is a logical extension of the mortality data, given that in the higher rate neem containing treatments, the majority of the snails were dead within 3 days. The differences observed in egg laying were highly significant after 66 days (p=0.0002). No phytotoxicity was observed in the plants throughout the test period. The number of remaining plants was also a good indicator of the effectiveness of each of the treatments. In the 5 x neem/papaya combination and 5 x neem treatments none of the plants were eaten. In contrast in the 5 x papaya and 1 x neem applications 25% of the plants were eaten, 50% in the 1 x neem application and in the control. Analysis of variance shows these differences to be significant (p=0.0394). It can be seen that the percentage of plants eaten across all treatments in this location was lower than those in the first location (Dr. Asquith's farm). This is probably due to the age and resulting hardiness of the plants. It lends further evidence to support the notion that if neem applications are adopted, plant age might be a factor in the application protocol.

Collecting enough non-target species in the lo'i for any meaningful study of extract effects proved extremely difficult to do and consequently this aspect of the study had to be abandoned. This will require a separate effort from this project. Other than Golden Apple snails, no other species were found within the test plots.

Effect of NEU1165M Slug and Snail Bait™

The observation results from the test plots are shown in Table 4 of Appendix A. It can be seen that the bait had little effect, when compared with the neem extracts. General observations were made one day after the bait had been applied and it appeared that the bait had been consumed as none remained. It was shown in the lab in preliminary work that the bait was stable for at least 1 week in water, so it does not seem that the bait would have disintegrated in the intervening period since application. Snail counts over the 21 day test period showed that there was a dose effect on snail mortality, with the % mortality across both locations for the ½ lb/1,000sq.ft. rate at 6.25%, for the 1.0 lb/1,000sq.ft. rate at 18.75%, and for the 5.0 lb/1,000sq. ft. rate at 40.6%. These figures compare with the % mortality in the control (no bait) of 25%. The plantings in the test lo'i were all between 3 and 6 weeks old, which in the tests with the botanical extracts had been shown to be vulnerable to destruction by the snails. This was unavoidable because these were the only plots available at the time of the test. Consequently, after 7 days only 37.5% of the plants remained and after 21 days only 9.4% remained. For this reason, the test was terminated after 21 days. The number of egg clusters was not affected by the treatments.

There are a number of possible reasons why the bait did not exhibit the efficacy of the neem extracts. It is not absolutely certain that the snails ate the bait. It is possible that other organisms consumed it, but given that the plastic sheet remained around the test rings for three days, not allowing other animals into the test plots, it is difficult to imagine what else may have consumed it, unless they burrowed under the perimeter and then burrowed back out again within 24 hours. Secondly, it may be that the Golden apple snail metabolizes ferric phosphate by a different mechanism than its terrestrial counterparts, rendering the active ingredient inert, or expelling it. However, there was a slight dose effect. It therefore may be that the application rate needs to be much higher in an aquatic system. Further studies are needed to determine the true efficacy of ferric phosphate in this system. Further funding is being sought to expand this study.

Research conclusions:

The results showed that water-based extracts of neem fruits work well in controlling Golden Apple snails in wetland taro cultivation systems. There are now a number of regulatory hurdles to overcome in order to register the water extracts for use. If this can be achieved, possibly through Inter-region IV (IR-4), then taro farmers in Hawaii will have the first means to effectively control this devastating pest.

Participation Summary

Research Outcomes

No research outcomes

Education and Outreach

Participation Summary:

Education and outreach methods and analyses:

At Dr. Asquith's farm, where one of the test plots was located for these studies, local school children were taken to the plots weekly and the experiment explained to them, including an explanation of why the Apple snails were a threat to wetland taro cultivation. They were also encouraged to take part in the data acquisition, which included counting dead and live snails, eggs and plants.

In addition, the data were presented as a one-hour seminar at the Annual Hawaii Taro Growers' Festival and Conference in October 2004. The seminar was videotaped and included in a presentation to be made to the Taro Association affiliates at other locations throughout the state. A handout was produced for the taro growers attending this meeting and is enclosed in Appendix B. The principal investigator will prepare a manuscript for publication in a peer-reviewed journal, dealing with the comparative effect of the two water-based botanical extracts from papaya and neem.

Education and Outreach Outcomes

Recommendations for education and outreach:

Areas needing additional study

A number of areas require more in-depth studies. Neem oil contains azadirachtin and the other neem derived limnoids as the active ingredients. The water extract needs to be analyzed to determine if azadirachtin is the active ingredient therein. Azadirachtin is not very water soluble, however, the preparation of the water extract is likely to have produced an emulsion, which probably contains the compound. If azadirachtin is shown to be the active ingredient, it is already known that it is not very stable in the environment and particularly in water. This may aid in trying to develop a more integrated approach to its delivery, minimizing the potential for non-target species interactions. In addition, further attention needs to be directed at which non-target species remain in the lo'i after draining and what the effect of the water extract is on them. During the course of this study, draining the lo'i apparently drove out many of the non-target species, including fish and frogs. This resulted in there not being enough of any species to monitor in the lo'i and so from this study, those effects are not known. The combination of draining the lo'i, isolating it from the rest of the water flow may be enough to minimize any non-target effects. IR-4 will be contacted for advice on how to proceed. The economics of using neem extracts for Golden Apple snail control have yet to be fully studied. This will need to be a priority in future studies.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.