Cropping Systems to Control Tropical Soil-Borne Pests in Dryland-Grown Taro

Performance Target 1: Determine resistance/ tolerance of newly introduced taro germplasm against root-knot nematodes.
This performance target was completed. The taro germplasm in the Thaipalm and Cho2000 collections were screened twice for resistance to or tolerance of root-knot nematodes. The Thaipalm collection contains 43 different accessions - 32 from Thailand, 6 from Nepal, and 5 from Vietnam. The Cho2000 collection contains 11 hybrids from parents originating from Indonesia, Papua New Guinea, Guam, Hawaii, Palau, Samoa and Thailand. No resistance to the root-knot nematode was found in the taro germplasm screened (Ortiz, 2005). Severe injury caused by M. javanica to roots and corm of a susceptible taro cultivar is shown in Figure 1. Potential candidates for further tolerance evaluation were Thailand cultivar 035 and hybrid 40 from a Hawaiian x Palauan cross, because these cultivars displayed a consistently low nematode reproductive factor (Rf) and high growth. In explanation, a reproductive factor of 1.5 indicates that 1.5 times the number of eggs of root-knot nematode Meloidogyne javanica was found at the end of the experiment relative to that added at the beginning.

Performance Target 2: Evaluate biomass potential of green manure crops, resistance to root-knot nematodes and Pythium.
This performance target was completed.

2a. Evaluation of Green Manure Species in Greenhouse. Of 22 candidate green manure species screened in the greenhouse during 2004-05, sorghum x sudangrass hybrids (Sorghum × drummondii) appeared to be the best green manure species for M. javanica control (Ortiz, 2005). They were very poor hosts to M. javanica, produced copious amounts of biomass, and grew well when inoculated with M. javanica and taro pathogen Pythium aphanidermatum combined or individually (Table 1). In particular, ‘Graze-all MST’ had a low nematode reproduction factor (Rf = 0.19).

2b. Evaluation of Green Manure Species on Island of Hawaii. Based on the greenhouse results, a preliminary field trial was planted in July 2004 along the Hamakua Coast of Hawaii to evaluate six promising green manure species. Yellow mustard (Sinapus alba ‘Ida Gold’), marigold (Tagetes patula ‘Nema-gone’ and ‘Golden Guardian’), Sorghum x Sudangrass (‘Sordan 79’ and ‘Tastemaker’) and sunn hemp (Crotalaria juncea) were compared to a weedy unplanted treatment and a weed-free weed mat treatment. The two marigold cultivars were grown in the greenhouse for seven weeks and then transplanted into the field. Other green manure crops were sowed directly into the field. Pests and diseases were observed. Dry matter was harvested after three months.

Bird-feeding was a problem on seeds of sorghum x sudangrass hybrids. Yellow mustard (Sinapis alba ‘Ida Gold’) plants were removed prematurely from the field due to heavy infestation by fungal pathogens, Rhizoctonia spp. and Pythium spp. Marigolds grew well after transplanting in the field, but this method was time-consuming and seeds of these two cultivars were very costly. Based on dry matter production in the field, economic considerations, and ease of management (e.g. direct seeding versus transplanting), the two most promising green manure species selected for further studies were sorghum x sudangrass hybrid ‘Sordan 79’ and sunn hemp (Figure 2).

Root-knot nematodes occurred in low population densities at planting in all green manure treatments and decreased at knockdown under all treatments. The green manure species tested were confirmed to be non-hosts to the root-knot nematode. Populations of the reniform nematode (Rotylenchulus reniformis) declined from planting to knockdown at levels ranging from 63% to 99% (Figure 3).

Performance Target 3: Determine the best management practices for green manure crops.
This performance target was completed.

3a. Green Manure x Growing Duration Trial on Island of Hawaii. One crop of nematode-susceptible buckwheat (Fagopyrum esculentum) was grown for 3 months, then plowed under and allowed to decompose for 1 month. Beginning in November 2004, two promising green manure crops, sorghum x sudangrass hybrid ‘Sordan 79’ and sunn hemp, were grown for 1, 2.5, and 4 months in Pepeekeo, Hawaii. A weed mat treatment and a weedy treatment were included for a total of eight treatments and four blocks. Root-knot nematode populations were low overall, although they differed significantly in green manure treatments after decomposition. The highest mean number of root-knot nematodes was found in 1-month sunn hemp and the lowest mean numbers in 4-month sunn hemp and 2.5-month ‘Sordan 79’ (Figure 4). Reproduction of reniform nematodes differed significantly in green manure treatments, with 4-month ‘Sordan 79’ supporting the highest reniform nematode reproduction (Figure 5). In contrast, 2.5- and 4-month sunn hemp was a poor host to the reniform nematode with lower reniform nematode reproduction compared to 4-month ‘Sordan 79.’

Glyphosate was applied to kill green manures and then taro planted 1 month later in March 2005. Soil samples were taken prior to planting of taro. Exchangeable potassium (K) differed significantly among treatments with the greatest level in 4-month ‘Sordan 79’ followed by 2.5-month ‘Sordan 79’, and the lowest level in the weed mat treatment. After 1 month of growth, tissue analyses of taro were conducted. Significant differences in leaf K were found in taro grown for 1 month, with the highest K concentration in taro grown after 2.5-month ‘Sordan 79’ and the lowest K concentration in taro grown after the weed mat treatment.

Visible treatment differences in growth of taro were evident at 3 months after planting (Figure 6). A representative taro plant was harvested in each plot at 5 months after planting and nematode populations associated with roots measured. Fresh weight of taro corms was negatively correlated to reniform nematode populations, with smaller corms having a higher nematode population (Figure 7). These results suggested that the reniform nematodes possibly contribute to yield reduction in taro.

Corm yields of taro were determined after 9 months of growth. Green manures grown for 2.5 or 4 months had a beneficial effect on the individual fresh corm weight of the subsequent crop of taro (Figure 8; Miyasaka et al., 2006). This beneficial effect could be due to: a) lower initial numbers of root-knot nematodes; b) lower reproduction of reniform nematodes; and/or c) greater exchangeable K in soil and in taro leaves, perhaps caused by slow release of nutrients during decomposition of green manure crops.

3b. Green Manure Species Trial on Island of Molokai. In 2003, soil samples showed a wide variation in nematode population densities in the field. To produce a uniform, high population density of nematodes, three crops of nematode-susceptible buckwheat, (Fagopyrum esculentum) were grown in the field during 2004 (Figure 9). An average of 6 root-knot nematodes was found per gram of buckwheat roots (Figure 9); however, root-knot nematode populations in the soil were low. After seven months of tilling to remove volunteer buckwheat seedlings, six green manure crops (sorghum x sudangrass ‘Graze-all MST’; sunn hemp; marigolds (Tagetes erecta ‘Scarletade,’ ‘Orange Deep,’ and ‘Scarletade Single;’ and buckwheat) plus a weedy control were planted in March 2005 for a total of 7 treatments. They were grown for approximately three months, then cut back with a flail mower, tilled under (Figure 10), and allowed to decompose for four months. Then, taro ‘Bun long’ was planted and grown for nine months.

No significant differences were found in numbers of root-knot nematodes at the start of the taro crop, nor mid-way through the taro crop. No significant differences in yield of taro were found when taro was harvested at nine months after planting. This lack of an effect of green manures on taro yields could be due to the long fallow period between buckwheat and planting of green manures, and that between green manure plow-down and planting of taro. Apparently, a long fallow period alone under the dry environmental conditions at the Molokai Applied Research Farm is sufficient to control root-knot nematode populations.

3c. Green Manure Species Trial on Island of Oahu. Four green manure crops (marigold Tagetes erecta ‘Orangeade’, marigold T. patula ‘Yellow Boy,’ sorghum x sudangrass hybrid ‘Sordan 79,’ and sunn hemp) were planted in January 2006 at the Waimanalo Experiment Station on the Windward coast of Oahu. A weedy control and weed-mat control also were installed at planting for a total of six treatments. During the 4 month growth period, green manures were cut back as needed with a flail mower to minimize seed-set. Then, green manures and the weedy control were cut down, tilled under, and allowed to decompose for 38 days. Treatment plots were tilled and disked prior to planting taro ‘Pauakea’. Taro was harvested after 10 months of growth. No significant differences in yield of taro were found between treatments.

3d. Green Manure Species Trial on Island of Maui. A similar experimental plan was followed as for Molokai and Oahu. Starting in April 2006, four nematode-resistant, green manure crops (sorghum x sudangrass ‘Graze-all’, marigold T. erecta ‘Scarletade’ and ‘Orangeade’, and sunn hemp)as well as nematode-susceptible buckwheat and a treatment containing naturally-occurring weeds were grown for three months. Then, all six treatments were plowed under and allowed to decompose for one month. Taro cultivar ‘Bun long’ was planted in August 2006 at a spacing of 0.3 x 1.2 m in plots containing a total of 60 plants. At three and six months after planting, two plants were harvested. At nine months after planting, a total of eight plants were harvested.

At three months after planting, taro grown after nematode-susceptible buckwheat had significantly lower dry weights of leaf blades, petioles, corms, and roots of the main plant compared to those grown after nematode-resistant sunn hemp (Table 2). In addition, the buckwheat treatment produced significantly lower dry weights of the sucker petioles compared to the sunn hemp treatment. Similar results were found at six and nine months after planting.

3e. Tillage of Green Manure trial on Island of Hawaii. Sorghum x sudangrass hybrid ‘Sordan 79’ was grown for approximately four months. In the Delayed-Tillage (DT) plots, ‘Sordan 79’ was killed with glyphosate and taro planted into the dead stubble after 45 days of decomposition. In the Conventional Tillage (CT) plots, the green manure crop was plowed, tilled (Figure 11), and after 45 days of decomposition three cultivars of taro (‘Bun long’, ‘Maui lehua’, and ‘Eleele naioea’) were planted. After three months of growth in the DT plots, soil was tilled and hilled around the base of taro plants.

After nine months of growth, fresh weight yield of taro corms with rotten portions removed was significantly greater in the CT plots compared to the DT plots (Figure 12). This yield difference was due to the significantly greater percent of corm rots in the DT plots, probably as a result of root damage caused by the delayed tillage that lead to invasion of corms by pathogens.

Performance Target 4: Assess the impact of green manure crops on soil bacterial, fungal, and nematode communities in the field.
This performance target was completed. Impact on nematode communities was reported above.

4a. Soil Microbial Bioassays. Soil samples were collected at the end of the green manure growing period (prior to incorporation), the end of the decomposition period (prior to taro planting) and/or at the end of the taro growing period (prior to harvest). Microbial population densities (colony forming units per 1g soil) were estimated for soil bacteria and fungi using standard culture-dependent bioassay techniques. Data were analyzed using the analysis of variance for repeated measures and treatment means were compared using the least significant difference (LSD).

4a.1. Green Manure x Growing Duration Trial on Island of Hawaii. Green manure crops provide nutrient resources that contribute to microbial blooms during decomposition. The soil bacterial densities increased, whereas the soil fungal densities decreased during the decomposition period (Table 3). The average treatment population densities tended to be significantly different at LSD0.05 using analysis of variance (ANOVA) for repeated measures. The microbial bloom typically expected from green manure decomposition was not observed, perhaps due to the long duration of one month between incorporation of green manures and sampling.

4a.2. Green Manure Species Trial on Island of Molokai. Similar to the results of the Green Manure x Growing Duration trial on the island of Hawaii, the soil bacterial densities increased while the soil fungal densities decreased during the decomposition period (Table 4). Under the fallow control, the soil microbial densities showed the least amount of change, whereas the bacterial and fungal densities changed the most under the ‘Graze-all MST’ and ‘Orangeade’ green manures, respectively. Similar to the results of the earlier trial, the microbial bloom expected from green manure incorporation was not detected, perhaps due to the long duration of four months between incorporation and sampling.

4a.3. Tillage of Green Manure trial on Island of Hawaii. Soil bacterial and fungal densities did not differ at the end of the decomposition period. The delayed tillage treatment, however, showed a higher incidence of taro corm rot than the conventional tillage treatment. The results of the bioassay did not correlate with incidence of taro corm rots, indicating that this measurement is not a good predictor for the effectiveness of taro pathogens.

4b. Soil Bacterial Community Profile. Total soil microbial DNA was isolated and the full-length 16S rRNA gene amplified using the bacterial primers P8F and P1492R. The amplified fragments were cloned using the TOPO-TA system and sequenced on an ABI 3700 DNA Analyzer.

4b.1. Impact of Tillage Practice on Soil Bacterial Community Profile. Soil samples for molecular analysis were collected at the end of the green manure growth period (H7) and at the end of the green manure decomposition period (H8). Four sequence libraries were constructed: conventional-till (CT) and delayed-till (DT) at the two time points noted above. The total number of clone sequences varied for each library as follows: 127 for H7-CT, 75 for H8-CT, 100 for H7-DT and 65 for H8-DT.

Unclassified bacteria that could not be placed even at the phylum level represented 44-71% of each sequence library and were grouped using FastGroupII. From the four libraries, 215 total unclassified clone sequences were placed into 210 unique groups; only two doublets and one triplet were found at the 95% sequence identity level. The high percentage of unclassified bacteria and unique (singleton) sequence groups indicated the extent of unassessed diversity and highlights the need for further characterization and classification of the soil microbial community.

Using the Library Compare tool at the Ribosomal Database Project II website, significant differences were detected between time points for the CT treatments for the order Sphingobacteriales, family Oxalobacteraceae and class Gammaproteobacteria at the 95% confidence level. The Sphingobacteriales and Oxalobacteraceae both comprise many species noted for their ability to digest chitin and other organic molecules, whereas the Gammaproteobacteria comprise many plant pathogens.

Three clones, representing 4.0% of the H8-CT community composition, showed >99% similarity to the family Oxalobacteraceae but could not be further classified to genus or species. Also in the H8-CT library, four clones (5.3% of the community) were classified to Sphingobacteriales, with three clones classified to genus (Chitinophaga or Hymenobacter) with >97% similarity. Interestingly, no sequences belonging to the Sphingobacteriales or Oxalobacteraceae were found in the H7-CT library which was 1.7X larger than the H8-CT library. For the Gammaproteobacteria, the community composition increased from 3.9% to 12% over the decomposition period in the CT treatment.

4c. T-RFLP Fingerprinting of Soil Bacterial Communities. Three field trials were selected for molecular analysis of the soil microbial community as affected by green manure treatments. The trials included a green manure species trial replicated on the islands of Molokai and Oahu and a green manure x growing duration trial established on the island of Hawaii. Soil samples that were collected at the end of the green manure growing period and at the end of the green manure decomposition period were chosen for molecular analysis. Briefly, the target gene (16S rRNA) was amplified from the total community DNA with the universal bacterial primers P8F and P1492R. The amplified product which contained one fluorescently labeled end was then digested with HaeIII restriction enzyme. The fluorescently labeled fragments (T-RFs) were separated by size on a Li-Cor IR2 DNA Sequencer to create the community fingerprint. Change in the community diversity was calculated with analysis of variance for repeated measures using the total number of bands detected at the two time points, with each band representing one or more different ribotypes. The Dice similarity coefficient (S) was also calculated to measure change in the community fingerprint.

4c.1. Green Manure x Growing Duration Trial on Island of Hawaii. Significant differences in the mean number of T-RF bands were found for green manure treatments, with 4-month sorghum x sudandgrass hybrid ‘Sordan 79’ exhibiting the greatest number of T-RF bands both pre- and post-decomposition (Table 5). The 2.5-month sunn hemp treatments displayed the greatest change in the soil bacterial community with a similarity coefficient (S) of 0.52±0.02. The 4-month ‘Sordan 79’ treatment showed the highest similarity over the decomposition period (S = 0.83±0.06). Significant differences were detected for soil microbial diversity among three duration times of sunn hemp, but no differences among the three duration treatments of ‘Sordan 79’. The 2.5 and 4 month growing periods for both sunn hemp and ‘Sordan 79’ showed a beneficial effect on final taro corm weights. The organic matter and nutrients added by the green manures are the likely main contributors to the beneficial effect but not the bacterial community diversity or population density.
4c.2. Green Manure Species Trial on Island of Molokai. The marigold ‘Scarletade Single’ treatment showed a decrease in community diversity over the decomposition period while the other treatments increased in diversity as measured by the number of T-RFs detected in the community fingerprint (Table 6). Sorghum x sudangrass hybrid ‘Graze-all’ and marigold ‘Orangeade’ showed the most increase on the number of bands detected by T-RFs, followed by sunn hemp and marigold ‘Scarletade Xanthophyll’. A significant difference was detected between these two groups, but not within each group.
In contrast, the bacterial bioassay data (Table 4, section 4a.2) showed an increase in soil bacterial population densities for all treatments. This discrepancy between the results of the microbial bioassay and that of T-RF analysis could be caused by: (1) the small number of microbial species (typically <1%) that can be grown in media used in the bioassay; and (2) the variation in number of species that are represented by the same band in the T-RF fingerprinting.

An increase in population density along with a decrease in community diversity indicates a shift in community evenness, i.e. a depression or bloom in a particular segment of the community. Marigolds have been shown to produce polythienyls that release singlet oxygen on decomposition. Singlet oxygen is a type of reactive oxygen species that can cause significant damage to cell structure and may impact non-target soil bacteria in addition to soil-dwelling nematodes.

4c.3. Green Manure Species Trial on Island of Oahu. The soil bacterial community showed a mean decrease in diversity over the decomposition period for all treatments except ‘Orangeade’ marigold which showed a very slight increase in diversity. The weed mat and fallow controls with little or no added organic matter also showed a decrease in diversity, implicating environmental factors in this decrease of soil microbial diversity. Sunn hemp showed the greatest decrease in community diversity and, accordingly, had the lowest similarity (S) score for community fingerprints pre- and post- green manure decomposition.

Performance Target 5: Conduct effective educational outreach to disseminate project information to taro growers.
This performance target was completed. A 5-minute video that shows the main highlights about green manures was produced by Mr. Dennis Miyahara from the College of Tropical Agriculture and Human Resources, University of Hawaii. Ms. Andrea Blas (GRA on this project) served as the narrator. The video can be viewed at http://www.ctahr.hawaii.edu/sustainag/downloads/greenmanure.wmv. It has been linked to the following web sites: Hawaii Sustainable Ag Cover crop webpage:
http://www.ctahr.hawaii.edu/sustainag/Database.asp
New Farmer: sustainable and organic production techniques:
http://www.ctahr.hawaii.edu/sustainag/NewFarmer/Production.asp#nematode
Organic Agriculture at CTAHR: links and resources (Hawaii specific)
http://www.ctahr.hawaii.edu/organic/resources.asp.

A field day was conducted on Molokai in June 2005 when the green manure crops matured. Also, a field demonstration was held on the use of a flail mower to chop green manure into smaller pieces to enhance decomposition, instead of using a rotary mower. In December 2006, Mr. Alton Arakaki conducted a demonstration workshop on cleaning and treating taro ‘huli’ (i.e. vegetative propagating material consisting of the upper 0.5 cm of corm plus the lower 30 cm of petiole) to minimize transferal of nematodes into a new field.