Investigating the potential use of Trichogramma, a hymenopteran egg parasitoid, in the integrated management of lepidopteran pests of cabbage in Puerto Rico

Final Report for GS00-005

Project Type: Graduate Student
Funds awarded in 2000: $10,000.00
Projected End Date: 12/31/2003
Matching Non-Federal Funds: $10,000.00
Region: Southern
State: Florida
Graduate Student:
Major Professor:
Richard Pluke
University of Florida
Expand All

Project Information



From our studies in the laboratory and in the field, the egg parasitoid Trichogramma (Hymenoptera, Chalcidoidea) willingly parasitized diamondback moth (DBM) eggs on cabbage. The species of Trichogramma investigated were Trichogramma pretiosum (Riley), Trichogrammatoidea bactrae (Nagaraja) and Trichogramma minutum (Riley). All parasitized DBM eggs, even in a two-host scenario. The alternative host examined was the soybean looper (Pseudoplusia includens (Walker)), a common pest of vegetables on the south coast of Puerto Rico. Even though the soybean looper is considered a favorable host of Trichogramma, it was the DBM eggs that were preferentially parasitized in all three species. The greatest preference for DBM was shown by T. pretiosum, while T. bactrae, had the higher overall parasitism levels. Trichogramma minutum expressed the lowest levels of parasitism. The fieldwork showed that the two lepidopteran species oviposited their eggs in different locations on the cabbage plant. Their relative vertical positions on the cabbage plants varied depending on the size of the cabbage plants and on the particular experiment. Commercial T. pretiosum was used in the fieldwork and this insect conducted most of its parasitism in the lower reaches of the plants. Due to its foraging/parasitizing behavior, percent parasitism of the two host’s eggs depended on how low on the cabbage plants the bulk of the eggs were oviposited. Percent parasitism therefore was a combination of the inherent acceptability of the host eggs and the positioning of the host eggs relative to the parasitoid search. In our fieldwork there was generally higher parasitism of DBM eggs than soybean looper eggs. In those experiments where DBM eggs were oviposited close to the ground, parasitism levels were at their highest. Parasitism levels of soybean looper eggs, on the other hand were low, even when the eggs were found close to the ground. This is thought to be because of its lower acceptability as host to T. pretiosum.

A socioeconomic study was conducted in the mountainous central region of Puerto Rico; and a linear programming (LP) model was developed to study the farming system of this region. A series of interviews and other data-gathering exercises gave detailed information on farm activities and the factors that influence the farming system. The LP model combined this information and gave a fair representation of the farming system. Most of the farm activity occurs during the Christmas period, while the summer months are normally a time of low activity. The government’s incentive scheme sustains many of the activities found; and leads to plantain being a dominant crop. Poor, unstable markets seem to be the biggest constraint to this system. Agrochemical use is limited in many of the crops, especially roots and tubers. Any introduced IPM technique would have to focus, primarily, on keeping labor costs down.


Insect Pests in Cabbage and the Failure to
Control them using Insecticides

About 15 species of insect pests of cabbage in Puerto Rico have been recorded (Agricultural Experimental Station, University of Puerto Rico, 1999). These include the diamondback moth, whitefly, leafminers, loopers and the imported and gulf white cabbageworms. The diamondback moth (Plutella xylostella) (Linnaeus) is by far the most important pest of cabbage on the island. Both the cabbage looper (Trichoplusia ni Hübner) and the soybean looper (Pseudoplusia includens Walker) can also be found in the field, but the latter is more common. In the University of Puerto Rico’s Research Station’s technological package on cabbage (1999), loopers are noted as the second most important pest of cabbage.
Both the diamondback moth (DBM) and the soybean looper (SL) are known to have developed resistance to a variety of insecticidal compounds that have been used in Puerto Rico and elsewhere (Armstrong 1990, Boethel et al. 1992, Leibee and Savage 1992, Shelton et al. 1993, Gianessi et al. 2002, Zhao et al. 2002). It is for this reason that the biological control of these insect pests has been pursued. Although most of the research has been directed at the better-known larval parasitoids—in particular Cotesia spp. (Hymenoptera: Braconidae) and Diadegma spp. (Hymenoptera: Ichneumonidae)—egg parasitoids also have the potential to significantly reduce populations of these pests. DBM and the loopers are thought to have few natural egg parasitoids (Harcourt 1962, Lundgren et al 2002) and so it has been postulated that effective control may be achieved by filling this niche (Oatman et al. 1968) with the Trichogramma egg parasitoids.

Trichogramma Egg Parasitoids and Host Selection

Trichogramma’s ability to recognize and use a variety of host species is considered to be an evolutionary adaptation to its multivoltine life cycle and limited capacity to control its dispersal in the field (Sachtleben 1929). Its polyphagous nature does not signify, however, that it is indiscriminate in its choice of host eggs. Trichogramma carefully determines the number and the sex of the eggs laid into chosen hosts. It does so by an assessment of host size, age, nutritional suitability and previous parasitism of the host egg amongst other things (Quednau 1855, Klomp and Teerink 1962, Nettles et al. 1983, De Jong and Pak 1984, Pak 1986, Pak and De Jong 1987, Bin et al. 1993, Schmidt 1994, Consoli et al. 1999). In terms of egg parasitism and host quality, the small size of DBM eggs (~0.02 mg—0.44 mm long 0.26 mm wide) may reduce their perceived suitability to ovipositing Trichogramma females. It has been shown that Trichogramma wasp body size, fecundity, and longevity are dependent on larval feeding (Charnov and Skinner 1985). Below a certain host egg volume, small adults emerge, as space and availability of sufficient nutrients limit larval development (Greenberg et al. 1998). In general, most Trichogramma species are reported to prefer intermediate to large host eggs, often 0.8–1.8 mm in diameter (Schmidt 1994). In addition to possible low parasitism levels due to size, how do DBM eggs compare as potential hosts when in the presence of other host species’ eggs? As mentioned before, Trichogramma parasitizes the eggs of other members of the cabbageworm complex, attesting to their suitability as hosts. Loopers, in fact, are used as factitious hosts in laboratory colonies of Trichogramma (Hohmann et al. 1988, Kazmer and Luck 1995) and in the literature there is note that Trichogramma pretiosum is considered to have become adapted to genera of the Noctuidae (Monje et al. 1999). The global importance and ubiquitous presence of DBM makes it the chosen target in most control programs and research efforts. It is therefore important to find species or strains of Trichogramma that effectively parasitize DBM eggs, even when in the presence of other host eggs. Soybean looper (Pseudoplusia includens) was used in these experiments as the alternate host to DBM. The research was conducted in Puerto Rico where the soybean looper (SL) is an important, widespread pest in vegetable crops and likely to be found in cabbage plantings, alongside the DBM. The SL egg is significantly larger than the DBM egg (0.6 mm in diameter and 0.4 mm in height) and, as with other noctuids, is considered a good host for Trichogramma pretiosum (Monje et al 1999). How would the presence of this alternate, and larger host affect parasitism levels of DBM eggs by T. pretiosum? A series of field and laboratory experiments were conducted to examine Trichogramma pretiosum's parasitism of DBM in Puerto Rico and to see whether SL would be considered a more acceptable host. In addition to examining host preference, the series of field experiments conducted in 2001 and 2002 were designed to determine ovipositional patterns of the two host species (DBM and SL) and to look for parasitism patterns of the Trichogramma with respect to plant architecture. The laboratory experiments expanded on the basic host preference objectives and included protocols to examine the influence of prior ovipositional experience. In addition to this, two other Trichogramma species were also included in the basic host preference experiments. The two other species of Trichogramma (Trichogrammatoidea bactrae and Trichogramma minutum) were chosen, based on a literature search, which determined that both species were considered strong candidates in previous work.

Farming Systems Research

A farming system is defined by the FAO as “a population of individual farm systems that have broadly similar resource bases, enterprise patterns, household livelihoods and constraints, and for which similar development strategies and interventions would be appropriate” (F.A.O. 2003). Keating and McCown (2001) identify two key components of farming systems, namely the biophysical ‘Production System’ of crops, animals, soil and climate together with certain physical inputs and outputs and the ‘Management System’, made up of people, values, goals, knowledge, resources, monitoring opportunities and decision making. Their review of six types of farming systems analysis concludes that the challenges and opportunities lie at the interface between the ‘hard’, scientific approaches to the analysis of the biophysical system and ‘soft’ approaches to intervention in social management systems. They also conclude that the use of models in farmer decision support systems has been disappointing and a way has to be found of making models relevant to real world decision-making and management practices. This may not necessarily be achieved by only making the models more accurate or more comprehensive. Part of the scientific process is the deconstruction of problems so that individual elements can be identified, appraised, experimented on and understood. They are either left as they are, or altered in the hope of improvement. A weakness in the scientific process is that the deconstruction removes the elements from their natural position, and contextually, this can lead to misinterpretations and oversights. The place of an element in a particular system is as important as the intrinsic characteristics of the element itself. Importantly, the producers themselves see their environment as a system (Dixon et al. 2001) and evaluate new technologies by the way in which they interact with the other elements of their environment. Success is based on a perceived over-all betterment of the system. How can the totality of a system be understood? Farmer Participatory Research (FPR) is a good way of studying communities and understanding their characteristics. It relies on human interactions and observations and is built on synergistic collaborations of multi-disciplinary technical teams, community members and other stakeholders such as extension agents (Dlott et al. 1994, Biggs 1989, Chambers 1994, Cornwall and Jewkes 1995). Participatory research methods have become an integral part of Farming Systems Research (FSR), and they serve as an important means of dialogue between participants and stakeholders.
FSR is principally about technology generation, and its processes can be divided into four stages: descriptive (diagnostic), design, testing and extension (Norman 1980). Stakeholder feedback is crucial to all stages. The first stage is about understanding the livelihood system and generating research objectives based on identified problems or possibilities. The second stage determines how best the research objectives can be met by planning an effective and efficient set of research activities. The third stage is the execution of these activities. This stage is given validity by its inclusiveness, its relevance and its interactivity. The final stage disseminates results and implements new technologies. The purpose of the farming systems analysis conducted in the three municipalities of the central region of Puerto Rico (Barranquitas, Naranjito & Orocovis) was to form a better understanding of how the farms function and how IPM practices could be incorporated. Specifically, the objective was to see how cabbage might be grown on these farms using an IPM methodology for the control of diamondback moth that included the use of Trichogramma. The mountain farms in this region were never developed or supported in the same way as the larger farms of the coastal areas, and both the industrialization and agricultural intensification programs of the last century have had little positive impact. They still remain key, however, to the identity and livelihoods of the people who live in this region of Puerto Rico. It is for this reason that methods should be identified that could help characterize the farming systems present and help promote sustainable development within these areas. This research project attempted to use linear programming models as a means of characterizing and studying the farming system of the central region of Puerto Rico.

Linear Programming

This is a form of modeling using an optimization matrix program that, for the purposes of FSR, looks to examine the interface of ‘hard’ scientific approaches and ‘soft’ approaches in social management. It does so by simulating and analyzing family farm livelihood systems by determining the optimal combination of farm and non-farm activities that is feasible, given a set of fixed constraints (Cabrera 1999). LP models are not as exact in their simulation of production functions as some crop models, and are not as sophisticated as some economic models, but they do represent a robust and fairly simple means of characterizing farming systems. They also provide a way of numerating assessments of how alternative activities achieve household objectives. To construct a linear programming (LP) model, certain information is needed. Hildebrand and Araújo (1997) state that the following is needed: 1) the farm and non-farm activities and options with their respective resource requirements and any constraints on their production; 2) the fixed resources and other maximum or minimum constraints that limit farm and family production; 3) cash costs and returns of each activity; and 4) a defined objective or objectives. With this information, LP models can be made to simulate the characteristics and activities found on farms that go beyond merely identifying the most productive strategies. The use of LP models in farm planning has its origins in the late 1950s, when whole farm planning was been developed. In 1958, Heady and Candler outlined the application of LP modeling to farm planning, and by 1963, its relevance to low-income agriculture had been demonstrated (Clayton 1963). Since then, it has been widely used to examine supply changes and policy shifts in agriculture (Hazell and Norton 1986). Its impact on improving livelihoods in developing countries, however, has never been great, in part due to the laborious data collecting process and to its lack of direct applicability (Collinson 2000). How then does LP modeling fit into FSR? Its primary use is in the first stage (description/diagnostic phase) and second stage (research-planning) of FSR, but can also be used as an extension tool in stage 4. Norton et al. (1999) describe an approach to participatory IPM research that is being implemented by the USAID-supported IPM Collaborative Research Support Program (IPM CSRP). One of the early activities in the IPM CSRP approach is the participatory appraisal stage that can take from between one to two weeks. It is at this stage; with disparate sets of community data or on-farm data collected, that linear program (LP) modeling could be most useful. All the information gathered by the various participants could be distilled into a matrix representing the enterprise activities (resource requirements and production functions), the farm’s constraints and resources (land, labor, capital, costs - on both spatial and temporal levels) and household objectives. Once a model has been validated, and it accurately reflects the farming systems under question, the designing of experiments (stage 2 of the FSR approach) can proceed. Alternatively, validated models can be used to assess already existing technologies to see if they would be worth implementing into the farming system under study. Used properly, LP modeling can be a very useful tool to FSR practitioners.

Project Objectives:
  • Investigate the host preferences of Trichogramma, specifically the preferences between the soybean looper (Pseudoplusia includens) and diamondback moth (Plutella xylostella). This will be accomplished through field trials and greenhouse experiments.

    Assess field releases of Trichogramma with respect to their impact on lepidopteran pest populations.

    Examine the socioeconomic influences on cabbage growers with respect to their acceptance and use of biological control practices


Materials and methods:

Laboratory Experiments

The Trichogramma species used were shipped weekly from Beneficial Insectary (Trichogramma pretiosum) (Redding, CA) and from Rincon Vitova Insectaries (Trichogrammatoidea bactrae and Trichogramma minutum) (Ventura, CA). The diamondback moth were obtained from a culture maintained at the Mid-Florida REC (University of Florida) in Apopka, Florida. The soybean looper eggs used in the experiments came from a laboratory culture that had been established in the Fall of 2000 at the University of Puerto Rico's agricultural research station in Río Piedras. The original insects had been collected as larvae and pupae from fields of various crops in Puerto Rico (soybean, tomato and eggplant being the principal host crops) and their identity was later confirmed by Dr. Heppner (taxonomist, Department of Plant Industry, Florida Department of Agriculture and Consumer Services) as soybean looper.

FalconTM Petri dishes (10 cm in diameter and 15 mm in depth) were used in all experiments. The protocol was adapted from previous work (Hassan and Guo 1991) and the host eggs were placed in four clumps around a small, 10% honey water drop in the center of the Petri dish. The eggs were placed into the Petri dishes in arrangements dependent on the experimental treatments. There were three treatments in each experiment. The first two treatments were single host treatments with either the eggs of DBM or SL, placed in clumps of 15 eggs at the four points around the honey water droplet. The third treatment was where both DBM and SL eggs were placed in the same Petri dish (combined host treatment). The two clumps of eggs for each species were placed opposite each other, across the honey water droplet. Each Petri dish contained 60 host eggs. For each treatment there were five replicates per experiment. The control Petri dishes were assigned the same combination of DBM and SL eggs as found in the combined host treatment, with the only difference being that no Trichogramma were introduced into the Petri dishes during the experiment.

Using a dissecting microscope and a fine camel hair brush, the female Trichogramma were individually placed into Petri dishes. The Petri dishes were then placed into a plastic container and the container placed in an insect rearing room (temperature 27 ± 3oC, 65 ± 5% r.h. and 16L:8D) for 24 hours. After this period, the Petri dishes were opened and the Trichogramma removed from the Petri dishes, which were then closed, returned to the plastic container and put back into the rearing room.

After 5 days, the eggs in the Petri dishes were examined for parasitism and desiccation. After this period, those eggs not parasitized had hatched and had either been eaten by the emerging larvae or stood as empty, transparent shells. The parasitized eggs were shiny black in color and easily identified. The dry eggs were brown or yellow and were in various states of desiccation. The number of all types of eggs was recorded for later analysis. From these results it was possible to determine percent parasitism in each treatment and also to determine what percent of the dry or damaged eggs found was due to direct parasitoid-induced mortality.

Two additional sets of experiments were conducted in the laboratory; the first to examine the influence of prior ovipositional experience of T. pretiosum on subsequent host decisions and the second to examine the host preferences of two other species of Trichogramma. To examine the influence of prior ovipositional experience, the Trichogramma adults were isolated for four hours in a vial that contained 50 of one or the other host species' eggs. After this period of time the adults were removed and the experiment conducted using the same protocol as the first set of experiments. The experiments with the two additional species of Trichogramma were also conducted as outlined above.

Field Experiments

In total, 8 experiments were conducted in 2001 and 2002. All the insects used in the experiments were acquired from the same sources as those used in the laboratory experiments. The experiments were conducted at the Fortuna agricultural research station on the southern coast of Puerto Rico.

The cabbage variety Blue Vantage was used in all experiments, and the seed was bought from the Rupp Seed Company (Wauseon, OH). The cabbage was sown at a spacing of 2 feet apart along the row and 3 feet between rows. Plants were irrigated by drip irrigation and sprayed once a week with Mattch (Ecogen Inc, Langhorne PA, USA), until the experiments were ready to be conducted.

A few days before the experiment was conducted, the field cages were placed over the experimental plots of cabbage. The cabbage plants were weeded before the field cages were placed over the experimental plots. To examine the possibility of there being native Trichogramma present at the research station, some of the cages were designated control cages.

The DBM and SL host species were released under the cages late afternoon. The cages were divided into three treatments and the control. The treatments were as follows: DBM only, SL only, DBM + SL together, and control (DBM + SL together). A completely randomized design was adopted for the experimental plots, which were randomly divided amongst the treatments and replicates. Depending on the experiment, there were either four or five replicates per treatment. 24 hours after the host species release, the Trichogramma were taken to the field site and released under the field cages. The Trichogramma were released from diet cups placed on the ground. Because of high wind conditions that exist in this part of the southern coastal plain of Puerto Rico, it was thought best to release the wasps from ground level, rather than from an elevated position. There were no Trichogramma released in the control cages. After a further 24 hours, the cages were opened and sample cabbage plants removed for analysis.

Individual cabbage plants were taken apart leaf by leaf, starting from the bottom. Each plant as divided vertically into quartiles. Using a magnifying glass and hand lens, each leaf was carefully examined for eggs. Eggs were then isolated by cutting a leaf disk containing the eggs with a leaf auger and placing the disk into an egg tray. All positional information relating to the eggs on the plants was carefully recorded, as was the location of individual eggs in the egg tray. The egg trays were placed in the insect colony room (temperature 27 ± 3oC, 65 ± 5% r.h. and 16L:8D) until signs of parasitism could be discerned, which normally occurred 4–5 days later. The non-parasitized eggs normally hatched before any signs of parasitism were noticeable in the parasitized eggs. The first sign of parasitism was a graying of the egg, which was then followed by the development of silvery-black patches that eventually coalesced over the entire surface of the egg.

Farming Systems Research and the use of Linear Programming Models

This work was not meant to be a comprehensive FSR study as it lacks many of the necessary components. It is however, an attempt to evaluate how FSR could help direct agricultural research and whether LP modeling could be a useful tool in this regard.

Identifying the Study Farming System

The first task was to identify the municipalities where the work was to be done. Barranquitas, Naranjito and Orocovis were eventually chosen because, in years past, they had collectively been the center of the island’s cabbage production (Agricultural Experimental Station, University of Puerto Rico 1999). With their relative proximity to the metropolitan area of San Juan, and with the benefits of the cooler mountain climate, they had become important suppliers of cabbage to the local market. Most of the production of the cabbage was during the early months of the year, when North American production was restricted by the winter cold. Other reasons for choosing these municipalities were the homogeneity of the farming systems, the fact that family farms were the dominant farm type and the fact that the farms had a relatively low resource base as compared to the large coastal farms. They continue with more ‘traditional’ farming practices such as the use of bulls for preparing the land. Once the three municipalities had been chosen, it was necessary to choose the farmers with whom to collaborate. With the help of research personnel, the three extension agents of the region were contacted and preliminary visits were made. To explain the objectives of the study a workshop was held with the extension agents. One of the objectives of the workshop was to determine whether the farms identified could be considered part of the same ‘recommendation domain’ (RD), or not. The RD concept was first developed by the International Maize and Wheat Improvement Center (CIMMYT) in the 1970s and it defined an RD to be “a group of farmers operating the same system and for whom the same new technologies would be appropriate” (Collinson 2000). For the LP model to work, the information inputted needs to be from one RD to ensure accuracy and relevance. One of the identified farms in Orocovis was larger than the rest of the farms (~150 acres as opposed to usually less than 100 acres) and the principal crop was coffee. This set it apart from the other farms. Nevertheless, it was included because other crops were grown on the farm and, as big as it was, it was run by one family and had similar constraints and available resources. The extension agents chose six farmers from each municipality and letters were written to explain the work and to invite the farmers to a workshop. Although the farmer workshop was principally designed to introduce the study, it also incorporated presentations on the principle pests of cabbage, the natural enemies of these pests and the new insecticides available for control in cabbage. At the end of the farmer selection process there were 16 farmers who had agreed to participate in the study.

Questionnaire Development

Because of the large amount of data needed, three questionnaires were prepared and three farm visits made to each farmer. The first questionnaire covered general farm data, the second covered the main agricultural activities conducted on the farm and the third questionnaire pertained to the economics of the farm and family.


Almost all the interviews were conducted with a Spanish-speaking colleague who assisted by taking down the farmers’ responses to the questions asked by the researcher. All interviews were conducted on the farms, which gave the researcher an opportunity to walk the farms and meet the farmers’ families. On occasion, the extension agent for the municipality participated. The interviews took from between one hour to two and a half hours to complete. Often the farmer was assisted by his wife and family. Additional data was obtained through sociologists, extension agents, agronomists and economists at the agricultural research stations of the University of Puerto Rico. Also approached were officials of the Puerto Rican Department of Agriculture and USDA for information on subsidies and incentive programs. Information gathered for individual crops was compared with the technological packages produced by the University of Puerto Rico for each crop. In addition to technical data, time was spent gathering information on the economic and political development of Puerto Rico, and specifically for the central region of the island. All aspects of life on the island are affected by its association with the United States of America and by its status as the oldest colony in the world.

Linear Programming

A linear programming model was developed using the data collected from the questionnaires and other sources. The model maximized the discretionary cash produced at the end of the year after all basic family needs were satisfied. In the making of the model, several assumptions had to be made and activities and constraints defined: 1) the year was divided into quartiles (January-March, April-June, July-August and September-December) to reflect the different periods of the year. Christmas time is a time of elevated sales while the summer months are a time of little activity, and the model was designed to assimilate these differences; 2) the tropical climate is such that many different crops can be, and are, grown in these mountain farms. The crops included in the model were restricted to those crops that were of economic importance to the farms of the central region. Thirteen crops were eventually chosen (Table 2). Many of the crops could potentially be grown at any time of the year. Crop appearances in the model were restricted to the times of the year when the crops were most commonly grown; 3) farmers accepted and used the government incentives available to them; 4) the work was not divided up between the farmer and the members of his family. A common family labor pool was used to complete the tasks associated with the activities. The labor available is a limited resource and the unit of measurement was a day of work that consisted of 8 hours; 5) additional labor is available at a cost of $30 a day per laborer; 6) land is also a limiting resource; and 7) the farmer had savings available to him at the beginning of each quartile. The quantity of the money available was calculated based on size of farm and size of the workforce. The savings had to be recuperated by the end of the year. This money was carried over to the next year and made available as before. This gave the system the flexibility to chose when savings could be best used. The crops included in the LP model: bananas, beans (kidney, string), cabbage, cassava, celeriac, chayote, coffee, ginger, papaya, plantains, pumpkin, taniers, and yam. The linear model was constructed using Microsoft® Excel 2000 Professional and Microsoft® Visual Basic. The principal matrix was supported and defined by a number of input and output tables that were also Excel documents. The solver used was the Premium Solver Platform with the XPRESS solver engine from Frontline Systems (Nevada, USA).

Determining the Parameters of an IPM Control
Strategy for Diamondback Moth in Cabbage

Part of this work was to see how a novel crop protection strategy might be incorporated into an existing farming system found in the central region of Puerto Rico. To extend the entomological research conducted in this doctoral work, it was decided to use the egg parasitoid Trichogramma (Hymenoptera: Chalcidoidea) in an IPM program against the diamondback moth (DBM) (Plutella xylostella; Lepidoptera: Plutellidae). The IPM program would include the use of scouting for insecticide application decisions and the use of IPM-compatible insecticides. The experiment was conducted on the farm of Sr. Ruben Ortiz, an experienced farmer from Orocovis who had grown cabbage for many years. It was agreed that he would plant and manage a quarter acre of cabbage using his normal production methods, the only difference being that the control program for DBM would follow a pre-determined IPM methodology. Sr. Ortiz was given a notebook that was used as a diary to log all activities pertaining to the cabbage crop. This included both the costs of materials and time taken for the work completed. Two thousand cabbage seeds were planted in seedbeds close to the house on the 3rd of May 2002. Half of the seeds were the Blue Vantage (Sakata Seed America, Inc.) variety and the other half were the Río Verde (Syngenta Seeds, Inc.) variety. The planting medium was a mixture of soil from the farm and river sand, a technique that he had developed to reduce the damage to the roots at transplant. Because of the sandy nature of the soil, he watered the seedbeds once a day. Seven times during the month that the plants were in the seedbed, he included one teaspoon of the fertilizer 20-20-20 to his spray. Weekly applications of Dipel DF, (Bacillus thuringensis v kurstaki, Valent Biosciences Corporation) (1 teaspoon/gallon) was applied to the seedbeds to prevent DBM damage. While the seedlings were developing in the seedbeds, the land was prepared for transplant. Vegetation was sprayed and cut back using herbicides and machete. 750 lbs. of calcium carbonate were applied to the land to increase pH. and later the soil plowed over using a pair of oxen. Banks were plowed into the field to help with drainage, to reduce soil humidity and the threat of soil-borne diseases. The cabbage seedlings were transplanted over a three-week period in June. One half of the field was transplanted with Blue Vantage transplants and the other half was planted with Río Verde transplants. The staggered planting was his normal practice, which, apart from dividing up the work, allowed him to have cabbage ready for market for over a month. A handful of 8-8-12 (NPK) fertilizer was applied to every plant at transplant. One other application of this fertilizer was made during the crop’s growth. Unlike other farmers, he did not designate a specific time for weeding, although he did pull weeds up when he came across them during his almost daily visits to the cabbage.

The following IPM strategy was used. Imidacloprid (Admire 240, Bayer) was applied as a drench soon after transplant (1 teaspoon per 4 gallons of water). This insecticide was used to prevent root aphid damage (species unknown), a very recent and very serious pest of cabbage in Orocovis. Admire 240 was not part of the farmer’s regular crop protection strategy. The Admire 240 was also applied to prevent the build up of whitefly later on in the crop. The scouting method used was one developed by Dr. Leibee of the University of Florida. The sampling locations were determined by mentally dividing the field up into quarters and placing a marker flag at the centre of each quarter. A fifth flag was placed in the middle of the field. Each week Sr. Ortiz scouted for DBM at each of the five marker flags. At each flag, four plants were randomly chosen and examined for DBM. Only the bud/head and the next four youngest leaves for each plant were examined. As the plant got older, this area of the plant began to represent the marketable part of the plant. Two measurements were made for each plant. The first was the presence or absence of any of the life stages of the pest insect (i.e., DBM) and the second was the presence or absence of DBM damage to the bud or head. In this case, the vertical leaves and those that face inward represent the bud. Each measurement was recorded as a ‘+’ for presence and ‘-’ for absence. At the end of scouting there were 20 measurements (‘+’ or ‘-’ for each criteria), which could be used to give percentage DBM presence or DBM damage in the cabbage. Control decisions could be made using this information. The whole scouting exercise took no more than 15 minutes. Dipel DF was applied on a weekly basis as part of the IPM strategy to control the DBM populations. This was one of the main insecticides used by Sr. Ortiz and he was happy with its effectiveness. The idea was to only use this insecticide against DBM during the life of the crop, and to use it at a lower frequency than normal. Sr. Ortiz’s normal management practice was to apply Dipel DF once every 5 days. Other insecticides used by Sr. Ortiz and surrounding farmers against DBM were Ambush (permethrin, Syngenta Crop Protection, Inc), Orthene (acephate, Arvesta Corporation), Agree WG (Bacillus thuringensis v aizawai, Certis USA), Monitor (O,S-Dimethyl phosphoramidothioate, organophosphate, Bayer) and Diazinon (diazinon, organophosphate, Syngenta). The Dipel was applied at two teaspoons (10 ml) per gallon and ten gallons were applied to the crop foliage. This worked out to be nearly a pound per acre. A regular knapsack sprayer was used to apply the insecticide mix, which also contained two teaspoons of Vel, a local washing-up liquid that helped to spread the mix. Sr. Ortiz normally applied the insecticide late in the afternoon. Based on scouting data, the weekly application of Dipel DF was to be increased to once every 5 days if DBM damage was found in over 30% of the plants. The T. pretiosum used in this experiment was obtained from Beneficial Insectaries, which sent the Trichogramma egg cards on a weekly basis. The factitious host eggs used were Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) eggs. The wasps were flown overnight to Puerto Rico and were then taken the next day to Sr. Ortiz’s farm. A weekly delivery of the Trichogramma was planned. On receipt of the Trichogramma, Sr. Ortiz cut each egg card square into two and placed each piece into a diet cup that was then sealed with a lid. In this way, ten diet cups were prepared. Based on previous calculations of Trichogramma emergence, 1,500–1,600 Trichogramma would be expected to emerge from each of the diet cups. This gave a release rate of up to 64,000 wasps per acre. At the beginning of the experiment, after the methods had been explained, Sr. Ortiz took the initiative and prepared release points for the Trichogramma. With wood, a zinc panel found on the farm, plastic cups and tape he made ten release stands for the Trichogramma. On the day that the Trichogramma adults began to emerge, Sr. Ortiz took the Trichogramma to the field during late afternoon, and released them by taking off the diet cup tops and placing the cup and its top into the plastic cups attached to the release points. An advantage of the location of the experiment was that the field was sandwiched by two stands of banana plants, which reduced wind effects. The Trichogramma were normally released on a Monday or Tuesday, and Dipel DF applications were made later in the week. There was at least 2 days between Trichogramma release and application of the insecticide.

Although the agreement with the farmer was that he would sell whatever cabbage was produced, and that we were trying to produce cabbage of the highest quality, this was not the objective of the experiment. The principle objective was not to see whether this particular DBM control strategy worked or not (cabbage is not normally grown in the mountains at this time of year) but was to record the requirements of the strategy in terms of costs and labor. The work was done to collect real data for use in the development of an alternative cabbage activity that could be placed and assessed in the LP model.

Research results and discussion:

Host Preference and Parasitism Experiments (Laboratory and Fieldwork)

All field experiments and laboratory experiments were completed by 2003. In the laboratory experiments, T. pretiosum expressed a clear preference for DBM eggs (70-80% parasitism) to soybean looper eggs (<40% parasitism). This was also the case when the female Trichogramma were given prior ovipositional experience to one or the other of the host eggs before being released into the experimental Petri-dishes. This 'conditioning' did not influence Trichogramma's preference for DBM eggs. Two other Trichogramma species (T. minutum and T. bactrae) were also examined and they also showed a preference for DBM eggs. T. bactrae although preferring the DBM eggs, showed relatively high parasitism levels of soybean looper eggs (50-60% parasitism). During the fieldwork, an extensive set of data was collected for host species' egg location on cabbage. The work indicated distinct differences in positioning of the two hosts' eggs on the plant. Position and grouping of the eggs depended on host species ovipositional behavior and on cabbage plant architectural development with age. Only DBM eggs were found on the stems of the plants and they were the most common eggs oviposited on leaf petioles. This is thought to be because of the adult’s smaller size, which allows it to enter further into the plant. Soybean looper females exhibit an ovipositional behavior whereby they position themselves on the adaxial edge of the leaf and then bend their abdomens down so that a large majority of their eggs are deposited near the edge of the leaf, on the abaxial side. There was no clear pattern linking the vertical distribution of soybean looper eggs and cabbage plant growth. For DBM egg distribution it seemed that there was a vertical shift upwards with growth. It could be that the DBM females were ensuring that their future progeny emerged relatively close to the palatable apical leaves, which move further from ground level with plant growth. For whatever reason, DBM eggs tended to be closer to the ground with smaller cabbage plants and higher up in the more mature plants. Thus, for the smaller plants, the DBM eggs were found lower in the plant than soybean looper eggs. This pattern was reversed in the larger plants. Most of the parasitism by the T. pretiosum used in the field experiments occurred in the lower reaches of the cabbage plant. From observation, the Trichogramma in the experiments tended to walk to the plants and begin searching from the bottom up. Either the Trichogramma, after a certain time, made a decision to stop parasitizing and so stopped halfway up the plant or for some other reason, made the decision to leave the plant altogether before reaching the top. Whatever the reason, those host eggs found in the lower reaches of the plant were more highly parasitized. This meant that the specific parasitism levels for each host species depended on how high up the plant their eggs were found. Field parasitism of the host species’ eggs, with the numbers of Trichogramma used in these experiments was generally low. Only once did average percent parasitism exceed 50% in an experiment, and this was with DBM eggs. DBM eggs were generally parasitized more than SL eggs - 10 of the possible 12 DBM experiments (single-host and combined-host) recorded an average of more than 10% parasitism. In the 12 SL experiments only 4 of the experiments recorded average parasitism of greater than 10%. In DBM there was generally higher parasitism in the experiments with smaller plants. This was not the case with SL. These experiments showed that for cabbage of differing maturity, the vertical distribution of host eggs in the DBM/SL host system changes—more of one host’s eggs can be found lower (or higher) in the plant than the eggs of the other host. In the three experiments (Experiments 1 (2002), 5 & 6) with high levels of DBM parasitism, the DBM eggs were found lower in the plant than the SL eggs. In Experiment 3, where DBM parasitism levels were lowest, the DBM eggs were found further towards the top of the plant. In the other experiment where SL eggs were found lower in the plant (Experiment 4), SL egg parasitism was higher than DBM egg parasitism. The fieldwork also showed that parasitism levels for the individual hosts’ eggs were no different in the single-host or combined-host treatments. This would suggest that, in this case at least, parasitism of the one host was independent of the presence of the other host. It seems that the two main factors affecting parasitism levels were egg position and inherent host acceptability. Farming Systems Study and Linear Programming Over fifty interviews with farmers in the central mountain region of Puerto Rico were conducted. The information from these interviews was used to characterize the Central Region farming system of Puerto Rico. The data and other collected information were also used in the construction of a linear programming model, which was then used to examine the dynamics of the farming system. In particular it was used to see how best IPM programs could be incorporated into this farming system. An additional analysis was done to examine the feasibility of managing DBM in cabbage, using a Trichogramma-based IPM strategy. From the time spent in this region of Puerto Rico, it seems that there are many constraints to farming in this region. The mountain farms of Puerto Rico are relatively small; ~70% of the farms in central mountains of Puerto Rico are 20 acres or less in size. They struggle to compete with the large farms of the southern coasts or with the cheap imports entering the country from overseas. All crops aside from plantain, coffee and banana are grown in plots no bigger than 5-10 acres and the produce is sold to intermediaries who, once they have a sufficient quantity of the product, then sell it on to supermarkets. Thus supplies from the mountains are not substantial or without complexity. The supermarkets of the metropolitan area prefer the steadier supplies from Canada, northern USA or the south coasts of Puerto Rico. Aside from market difficulties, there are problems in obtaining sufficient labor for the mountain farms. The industrial developments on the island during the last century led to the disappearance of large numbers of agricultural workers. The introduction of the federal food stamps program in the 1970s also had a negative effect on the availability of labor for the local farms. A third influence on agriculture in Puerto Rico was the push to modernize agriculture and to develop more intensive cultivation techniques in the seventies. This led to sophisticated incentive schemes that have had an effect of promoting some crops over others and of promoting the use of agrochemicals. All these events have had an effect on what an individual farmer is capable of doing on his/her farm and on how he/she tries to accomplish his/her objectives. In particular, the incentives scheme seems to influence the crops grown in this region, which are principally plantain, root and tuber crops, and coffee. The greatest assistance provided by the government is the subsidization of the workers’ salaries. Subsidized fertilizers and other agrochemicals are also offered by the scheme, although some crops such as cabbage do not qualify for the incentives. When the linear programming model was used to remove the influence of the incentives scheme, the composition of the crop types needed to maximize farmer objectives (profits), changed and less money was made. Plantain was no longer the dominant crop and those crops with a lower labor requirement were chosen by the model. In part, the incentives scheme maintains a way of life in this Central Region and keeps people employed but on the other hand it limits crop options and discourages some forms of innovation. Most farmers are content to continue planting the crops expected of them. When the data generated from the on-farm IPM experiment with Trichogramma was included in the model, it was quickly apparent that the associated additional costs precluded it from being chosen. The easiest way for this ‘IPM cabbage’ to be included in the model was by slightly increasing the selling price of the cabbage produced. The price of the Trichogramma purchased from California and the increased labor requirements were the two main reasons why cabbage grown, using this method of DBM control, was not feasible. This use of a linear programming (LP) model to assess IPM strategies in a farming system is considered unique. Once the necessary software becomes available at a reasonable price, the use of LP models could become an important tool in the planning, research and implementation of IPM strategies.

Participation Summary

Educational & Outreach Activities

Participation Summary:

Education/outreach description:

Two seminars were presented:

Jan. 2004. Mid-Florida Research & Education Centre, University of Florida, Apopka. "Host preferences in Trichogramma and how understanding the dynamics of a farming system may improve IPM research."

Mar. 2004. Dept. of Crop Protection, Mayaguez Campus, University of Puerto Rico.
"Host preferences of the egg parasitoid Trichogramma and its control of the Diamondback moth in cabbage."

Four manuscripts have been prepared and are in review for publication in refereed journals:

Host preferences of Trichogramma pretiosum (Hymenoptera: Trichogrammatidae) and the influence of prior ovipositional experience on the parasitism of Plutella xylostella (Lepidoptera: Plutellidae) and Pseudoplusia includens (Lepidoptera: Noctuidae) eggs. Richard W. H. Pluke and Gary L. Leibee

The parasitism and direct parasitoid-induced mortality of Plutella xylostella (Lepidoptera: Plutellidae) and Pseudoplusia includens (Lepidoptera: Noctuidae) eggs by three species of Trichogramma (Hymenoptera: Trichogrammatidae). Richard W. H. Pluke and Gary L. Leibee

Host preferences, parasitism levels and effect of plant development on parasitism of Plutella xylostella (Lepidoptera: Plutellidae) and Pseudoplusia includens (Lepidoptera: Noctuidae) by Trichogramma pretiosum (Hymenoptera: Trichogrammatidae) in field plantings of cabbage in Puerto Rico. Richard W. H. Pluke, Gary L. Leibee, and Angel L. Gonzales Rodriguez.

Scientific Note-An easy and dependable method of collecting diamondback moth eggs in bulk.
Richard W. H. Pluke and Gary L. Leibee

Project Outcomes

Project outcomes:

This work contributes on many different levels. From an applied perspective, we have shown that Trichogramma can have an important role as part of an IPM or biological control effort of DBM in crucifer crops. Aside from Trichogramma’s inherent ability to parasitize DBM in the field, we have identified some of the trophic interactions involved. We show that it is important to understand a host species’ ovipositional behavior and a parasitoid’s searching pattern, in the context of a host plant’s changing physical environment. This work affirms that biological agents are not easy substitutes for insecticides and that they should not be discarded just because they don’t give consistent results. This lack of consistency is used to discredit the use biological control agents, especially in commercial settings. We argue that more work needs to be done to understand the particularities of host ovipositional behavior and parasitoid searching for different host/parasitoid systems. Once these can be understood in the context of plant/crop growth, we may well achieve the desired ‘consistency’ of product. In the case of Trichogramma and DBM control, it may mean that Trichogramma should only be recommended when cabbage plants are small and DBM eggs are oviposited near the ground. At later plant growth stage, other biological agents may then become more applicable.

The farming systems and linear programming work may help reassess how the Central Region farmers are best served by the governmental incentives program. The data collected will also help agricultural researchers frame their work so that it has the best chance of being compatible and therefore adopted by the region’s farmers. Hopefully the use of linear programming models will become common practice in farming systems research and in particular, in IPM research.

Farmer Adoption

The previously described government incentives scheme seems to influence the crops grown in this region of Puerto Rico, which are principally plantain, root and tuber crops, and coffee. The greatest assistance provided by the government is the subsidization of the workers’ salaries. Subsidized fertilizers and other agrochemicals are also offered by the scheme, although some crops such as cabbage do not qualify for the incentives. When the linear programming model was used to remove the influence of the incentives scheme, the composition of the crop types needed to maximize farmer objectives (profits), changed and less money was made. Plantain was no longer the dominant crop and those crops with a lower labor requirement were chosen by the model. In part, the incentives scheme maintains a way of life in this Central Region and keeps people employed but on the other hand it limits crop options and discourages some forms of innovation. Most farmers are content to continue planting the crops expected of them.

When the data generated from the on-farm IPM experiment with Trichogramma was included in the model, it was quickly apparent that the associated additional costs precluded it from being chosen. The price of the Trichogramma purchased from California and the increased labor requirements were shown to be the two main reasons why cabbage grown using this method of DBM control would not feasible.


Areas needing additional study

  • There needs to be more work done to determine whether Trichogramma pretiosum, when released from ground stations, does tend to search the plants from the bottom up. It may be that the commercial Trichogramma used in the experiments had lost their motivation or ability to fly to the plants. Alternatively, this behavior could be explained by the relatively high winds that prevailed during the experiments.

    More experiments need to be done, with cabbage of different ages, to determine if there is a pattern of DBM oviposition that follows the apical growth of the plant.

    Based on work done in China, it may be worth conducting the field experiments again, using Trichogrammatoidea bactrae. In our laboratory work, this species showed greatest overall parasitism (when including the soybean looper). It would be interesting to see if this translates to increased parasitism in the field.

    The use of commercially produced biological control agents in Puerto Rico is greatly prohibited by the lack of local producers. All biological control agents have to be sourced from outside the country, at high costs. A study into the feasibility of locally-produced biological control agents needs to be done. This study should include a survey of those pest species amenable to biological control, an evaluation of the biological control agents that potentially could be beneficial to agriculture in Puerto Rico and a logistical/economic analysis of what would be needed to set up a local production facility.

    An assessment of the present incentives program of the Puerto Rican Dept. of Agriculture is needed. Changes should be sought that would allow greater flexibility in the system. There could also be an initiative to support IPM or organic practices, part of which could include finding markets that are willing to pay a premium for agricultural produce grown without agrochemicals.

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.