Crop Protection 34 (2012) 119e126

Contents lists available at SciVerse ScienceDirect

Crop Protection journal homepage: www.elsevier.com/locate/cropro

Limited efficacy of guava interplanting on citrus greening disease: Effectiveness of protection against disease invasion breaks down after one year Katsuya Ichinose a, *, Nguyen V. Hoa b, Doan V. Bang b, Do H. Tuan b, Le Q. Dien b a b

Tropical Agriculture Research Front, Japan International Research Centre for Agricultural Sciences, 1091-1 Maesato-kawarabaru, Ishigaki, Okinawa-ken 907-0002, Japan Southern Horticultural Research Institute, Long Dinh, Chau Thanh, Tien Giang, Viet Nam

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 November 2010 Received in revised form 14 November 2011 Accepted 27 November 2011

No sustainable management practices have been established yet for citrus greening disease (CG). Here we show the efficacy of interplanting guava on CG and its limitation. In 2004, four farmers in southern Vietnam found fewer occurrences of CG in their orchards with guava interplanting than in other orchards without it. The efficacy of guava interplanting was evaluated from field assessment in 93 citrus orchards in southern Vietnam. The CG infection was lowest in orchards where either chemical control with both non-neonicotinoid insecticides and neonicotinoids or the interplanting with guava was performed. Three field experiments were then carried out investigating guava interplanting. In these experiments, no citrus trees in orchards interplanted with guavas were infected by CG for over one year and a few months, while about 20% of trees were infected during the same period in orchards without guavas. There were significantly fewer psyllids in guava interplanted orchards in the first year, but the insect increased thereafter. Almost all trees were infected by CG after two and a half years irrespective of the presence of guavas, indicating that guava interplanting was effective for one year at most. Guava interplanting reduced invasion by the vector but failed to regulate its subsequent generation succession. Ó 2011 Elsevier Ltd. All rights reserved.

Keywords: Citrus nobilis Cultural control Diaphorina citri Huanglongbing Interplanting Psidium guajava

1. Introduction Citrus greening disease (CG), commonly known as huanglongbing, is a lethal disease of citrus, and no effective controls have yet been established for this disease (Halbert and Manjunath, 2004; Bové, 2006). In Asian countries, the pathogen that causes CG is Candidatus Liberibacter asiaticus, which is transmitted by the psyllid vector, Diaphorina citri Kuwayama (Da Graça, 1991; Jagoueix et al., 1994; Hung et al., 2004). Once trees are infected by the pathogen, they decline within several years (Yang et al., 2006). If intensive chemical control is not undertaken, this disease spreads quickly in orchards. In southern Vietnam, unless there is appropriate management, more than 50% of trees in orchards can become infected in a period of two to three years, and such orchards are often terminated within three to five years (Ichinose and Kano, 2006). In early 2004, four farmers in Cai Be, southern Vietnam, presented an unusual story to the staff at the Southern Horticulture Research Institute of Vietnam: These farmers had performed guava interplanting in orchards of Citrus nobilis Loreilo for three years. The cultivar used, named King Mandarin, is known to be sensitive to CG (Gottwald et al., 1989; Koizumi et al., 1997). The farmers found few * Corresponding author. Tel.: þ81 980 82 2308; fax: þ84 980 82 0614. E-mail address: [email protected] (K. Ichinose). 0261-2194/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.cropro.2011.11.023

CG occurrences in their orchards, whereas neighboring orchards without guavas were severely damaged by this disease irrespective of the cultivar. The farmers had started guava interplanting in the hope of obtaining earlier income from guavas, and they did not intend for the guava trees to provide CG control. In this region, it usually takes about two years after planting for citrus trees to yield fruits, but it usually takes half a year to a year for guavas to yield their first fruit. Hence, farmers can anticipate earlier income by the interplanting of guavas in citrus orchards. Hearing from the four farmers about the guava interplanting, we constructed three hypotheses regarding the efficacy of guava interplanting: 1) the citrus might receive a substance or substances from the guava that has lethal or regulatory effects on the CG pathogen; 2) guava interplanting could prevent the invasion and succession of the vector, resulting in lower risk of the pathogen invasion and spread in the orchard; and 3) the effect of guava could be accidental or could be the result of other confounding factors, and scientifically designed experiments would then fail to reveal the guava efficacy. According to the first hypothesis, guava should produce substances that would affect the pathogen. Microbial substances are known in the guava (Mohamed et al., 1994; Arima and Danno, 2002; Gonçalves et al., 2008). These or other unidentified substances with antibacterial activity are emitted either into the air

120

K. Ichinose et al. / Crop Protection 34 (2012) 119e126

or into the soil or both, and are absorbed through the aerial body or the root system of the citrus trees. Once transported in the citrus, they could affect the growth or survival of the CG pathogen. We tested this hypothesis indirectly by two analyses in a field study. The study involved the field assessment of CG infection in as many types of orchards as possible. In some orchards, young trees that had been assumedly produced from diseased mother trees by marcotting (propagation by layering) would have been planted. If guava contained substances that were lethal or regulatory against the pathogen, remedy effects would appear. For example, some farmers started citrus orchards by planting young trees produced by marcotting from mother plants that had already been infected by CG. Guavas would have been interplanted in some of these orchards and not in others. The proportion of infected trees would then be lowered in some years in the former than in the latter. In the second point, indirect evidence would also be obtained from a comparison of the longevity of diseased citrus trees after CG infection between plantations where guavas were absent versus where they were present. If guavas had any antibacterial activity against the CG pathogen, diseased trees would live longer in orchards with guava interplanting than in those without guavas. The second hypothesis was tested directly in field studies designed for this report. We established orchards either without or with guava interplanting, and compared the occurrence of CG and the psyllid population dynamics between these orchards up to three years after planting. These studies were all carried out in the Mekong Delta Region of Vietnam. Based on the results obtained, we discuss the efficacy of guava interplanting in preventing CG and the limits of guava interplanting as a management strategy. 2. Methods and materials 2.1. Guava interplanting in old orchards Since guava interplanting had been intensively performed until 2004, field assessment to test the remedial effect of guava interplanting was very limited. We visited a total of 93 orchards of Citrus nobilis, from May to November 2005 in southern Vietnam and asked farmers whether they planted young trees with a diseasefree certificate or those without such a certificate, and what management strategies they employed regarding CG. At that time, no farmers attempted guava interplanting for the management of CG. However, orchards with guavas interplanted between citrus trees were recorded as ones being managed by guava interplanting. The management was distinguished into six categories: no specific management for CG (33), chemical control by either non-systemic or systemic insecticides, chemical control by these two types of insecticides (27, 10 and 13, respectively), and biological control using weaver ants, Oecophylla smaragdina F. (2), or guava interplanting (8). The weaver ant has been traditionally used as a biocontrol agent for citrus cultivation in southern Vietnam (Van Mele et al., 2002). Among the eight orchards with guava interplanting, six orchards were planted with young trees produced by grafting, and two with ones produced by marcotting. Their mother plants from which the trees were produced appear to have already been infected by CG. During the questioning of farmers, it was not elucidated exactly when they implemented any management. Hence, it is assumed in this study that farmers performed management since planting the orchard. Also, it was impossible to know exactly when trees were produced. Thus, the tree age was estimated from the year for planting. For the test of CG infection, we randomly selected 10 to 20 trees and collected three leaves from each tree in each orchard. These leaves were brought to a laboratory for detection of the CG pathogen by PCR. Detection of the CG pathogen was attempted

qualitatively by PCR, for which DNA was extracted from midrib of individual leaves by CTAB. Two primers OI1 and OI2c which were specific in Candidatus Liberibacter asiaticus were used for PCR (Jagoueix et al., 1994). The sucrose synthase gene obtained from satsuma mandarin, C. reticulata Blanco, was used as an internal control (Urasaki et al., 2007). Trees with a pathogen-positive result were defined as infected. 2.2. Field experiment in Cai Be Two orchards were established, one with no guavas and one with guava interplanting (NI and GI, respectively), in Cai Be located in the Mekong Delta Region of Vietnam, about 100 km southwest of Ho Chi Minh. Both orchards were about 1500 m2 in size. In both orchards, there were some citrus trees, which were eliminated one month before the experiment was commenced. The shortest distance between these two orchards was about 10 m. There were two citrus orchards near NI; one was 5 m to the north and one 15 m to the east. The proportions of CG-infected trees in these nearby orchards were 100 and 25%, respectively. GI was surrounded by three orchards, one 5 m to the west (100% trees infected by CG), one 5 m to the east (54%), and the other 5 m to the south (25%). The dominant cultivar in these neighboring orchards was C. nobilis, mixed with several oranges, pomelos, lemons, limes or sweet mandarins. The land owners of our fields told us that these orchards had been planted four to five years before our study. In April 2005, 15 trees were randomly selected in these orchards, and five leaves were collected to test the CG infection of the trees by PCR. The CG infection proportion in these adjacent orchards had increased from 80 to 100 % by the end of this study. The C. nobilis trees used were all produced by the Southern Horticultural Research Institute of Vietnam (SOFRI) under the condition of no invasion of either the pathogen or the vector. This institute supplies citrus planting material with disease-free certificates to the public. Soil mounds, 60-cm wide and 30-cm high, were made at the location where citrus planting orchards were to be planted about one month before the planting. The distance between the centers of two adjacent mounds was 2.5 m. In NI, no crops other than C. nobilis were planted, and the distance between two adjacent C. nobilis trees was 2.5 m. In GI, guava trees (about 1 m high) had already been planted with 2.5-m distances both between rows and between trees one year before our experiment, and mounds were made at the midpoints between neighboring guava trees. This provided a distance of 2.5 m between two C. nobilis trees and 1.25 m between a C. nobilis and guava. The cultivar of guava planted in GI was ‘xa ly nghe’, which had been traditionally cultivated in the Mekong Delta in Vietnam. On each mound, 0.3 kg calcium, 0.5 kg phosphate and 5 kg of organic fertilizers were supplied and mixed with soil. On 24 May 2005, we planted a total of 81 young trees in NI and 104 in GI. The same amount of these fertilizers was applied every half year after the planting. After the planting of the trees, 50 g synthetic fertilizer (NePeK: 14, 14, 8) was applied on each mound once a month. On the day of planting, the trees were cut at a height of 50e70 cm above the ground surface to induce flushing. This pruning on the day of planting is recommended by public organizations and is commonly performed by farmers in southern Vietnam. No insecticides were applied to control psyllids. Water was supplied as needed. This field experiment was continued until July 2008. After planting, new shoots, adult psyllids and nymph colonies on individual trees were counted once a month in the two orchards, but twice a month for the first two months. Psyllid adults were counted individually, and nymphs were counted as a group on individual shoots. Thus, the number of nymph colonies on a tree corresponded to that of shoots of the tree infested by the psyllid

K. Ichinose et al. / Crop Protection 34 (2012) 119e126

nymphs. Tree deaths were recorded at each sampling time. Tree growth was estimated from trees randomly selected in each orchard, 12 in GI and 15 in NI. The height, width and depth of the canopy of these trees were measured once every month, and the canopy size was approximated by the product of these three measurements. We collected three leaves randomly from each of these trees in order to detect the CG pathogen by PCR. When we found any symptomatic leaves, we collected them selectively. These leaves were brought to a laboratory and kept in a freezer at 30  C until the analyses. 2.3. Field experiment in Dong Thap The efficacy of guava interplanting was examined also in Dong Thap, about 150 km southwest of Ho Chi Minh. The orchard, 3000 m2, was divided into six quadrats, four making up 2500 m2 of the total area and two making up 500 m2. In the former four quadrats, guavas, ‘xa ly nghe’, had been planted about one year before the planting of the C. nobilis trees. Mounds for the planting of C. nobilis were prepared as in the experiment in Cai Be, and 207 trees of C. nobilis were planted on 27 July 2007. In the latter two quadrats, 45 C. nobilis trees were additionally planted on 7 January 2008. The difference in the planting time was due to social issues. All C. nobilis trees used for this experiment were purchased at a public store. The infection of the trees by CG was individually examined by PCR before planting, and no CG infection was detected. Since visits to this orchard were limited for foreigners, assessments were not allowed to be performed as frequently as in Cai Be. In each visit, new shoots and psyllid adults and nymph colonies were counted on 120 trees, which had been randomly selected in the orchards just after the planting. Symptoms of CG (Bové, 2006) on individual trees as well as death were observed on all planted trees. Since the land owner was hesitant to continue to manage this orchard without insecticide after one year, the counts of new shoots and psyllids were given up and only CG symptom observations were performed thereafter. 2.4. Field experiment in My Luong In My Luong, 100 km southwest of Ho Chi Minh, one orchard without guavas and eight orchards with guava interplanting were established. The guava cultivar used was ‘khong hat’. Both C. nobilis trees and guava ones were produced by the SOFRI. Since many young guava trees had been infested at purchase by the nematode Meloidogyne enterolobii Yang & Eisenback (Iwahori et al., 2009), nematicide (Ethoprophos, 20%, ADC Co. Ltd. in Ho Chi Minh) as in the instruction by the company had to be used to control the nematode on the guava before and after planting. This nematicide, the commercial name of which is NOKAP, was the only nematicide commercially available in Vietnam when the studies were carried out. The size of these orchards ranged from 1000 to 10000 m2. Guavas were planted first in the eight orchards in December 2006, and C. nobilis trees were planted in all orchards in January 2008. Before the planting of the C. nobilis trees, no CG infection was confirmed in them by PCR, as in the former experiments. After the planting, 20 trees were randomly selected in each orchard. CG symptoms on each tree were observed in the first and second years. In this experiment, circumference was measured to represent the tree growth, since pruning and training of trees were carried out in order to increase the yield of fruits. The circumference of the trees was measured at about 1 cm above the upper limit of the grafting of a scion. Since the land owner of the control orchard was reluctant to continue the cultivation of citrus in the beginning of the third year, the experiment was terminated that year.

121

2.5. Extracts of guava with organic solvents for the test of deterrence to psyllids On 8 May 2005, 1 kg of leaves of the guava plant, cultivar ’xa ly nghe’, was collected and brought to the laboratory. The leaves were dried under shaded conditions for one week, and the weight was reduced to 560 g. The dried leaves were submerged in 10 l of hexane for one month. Afterward, the solvent was filtered by filter paper and the remaining material was kept in 10 l of acetone for one month. The solvent was filtered as the hexane solvent and the remaining sample was subsequently submerged in 10 l of methanol. Again, the solvent was filtered. These filtered solvents were concentrated to 100 ml (100 times concentration) and kept in a refrigerator. When the above extracts were used, they were diluted 100 times for bioassay tests, which were performed in November to December 2005. One day before each test, a sufficient number of psyllids were collected from fields and kept in a laboratory at natural temperatures with no food supply for starvation for about 24 h. Leaves of C. nobilis that had not been treated with insecticide were obtained from trees maintained in a greenhouse on the day of the experiment. One of the three extracts was sprayed on each leaf in the laboratory. After these leaves were dried (for 30e45 min), plaster of Paris was bedded on the bottom of a transparent plastic box (300 mm wide, 150 mm deep and 150 mm high) at a 1-cm thickness for maintenance of enough humidity for both the leaves and psyllids released in the box. Before the plaster became solid, three leaves treated with the same extract were vertically inserted on the plaster. After the plaster became solid, 10 psyllids were released in this box, and the number of psyllids on each leaf was counted 24 h after the release. All observations were made in the laboratory with an air conditioner set at 28  C and fluorescent lights. This experiment was replicated five times. The two least-selected extracts were determined by the mean numbers of psyllids on the leaves and used for the second experiment. Leaves of C. nobilis sprayed with either none or one of the two extracts were examined as the first experiment (six replicates). The extract that was selected with a lower frequency was used for a further experiment, in which leaves of C. nobilis were treated with either only the solvent used for the extract, the extract, or nothing. These leaves were examined as described above (seven replicates). 2.6. Data analyses Means and SE values were calculated in all the measurements of this study. For the experiment in old orchards, the annual CG infection rate in each orchard was calculated as a geometric mean for the year after planting. The mean of the rate was compared between the six treatments defined as above. In the field assessment study, two types of the guava interplanting orchards were recognised as for whether the orchards had been started with the planting of either young trees produced by grafting or those by marcotting. The numbers of these orchards were six and two, respectively. In the field study in Cai Be, the effect of the guava interplanting on the growth of citrus trees was examined by the comparison of the canopy size between NI and GI. The mean of the size was calculated for each orchard at each sampling time, and compared between the two orchards over all samplings by the KolmogoroveSmirnov test. The influence of the guava on the citrus was also evaluated by the shoot numbers. The numbers of new shoots of the three defined classes, small to large, on each tree was summed for each class for each tree over all the sampling times, and the summed numbers were between two orchards by a multivariate

122

K. Ichinose et al. / Crop Protection 34 (2012) 119e126

analysis of variables, MANOVA. A hypothesis that guava interplanting would reduce the occurrences of infected trees by CG through the regulation of the psyllid population in the orchard was examined by comparing the numbers of psyllid adults and nymph colonies between NI and GI. However, the numbers of new shoots would be possible a confounding factor in the difference of the psyllid numbers. Psyllids are attracted to new shoots (Yasuda et al., 2006; Hall et al., 2008; Ikeda and Ashihara, 2008), and accordingly the more a tree would flush new shoots the more psyllids it would be infested by. This possibility could be avoided if the numbers of adults and nymphs were divided by the number of all new shoots for individual trees at each sampling, and the ratios of adults and nymphs to shoots were calculated for individual trees. Seasonality would be another confounding factor, which would be avoided by summing the number of new shoots, adults and nymph colonies in individual trees over sampling times. The total numbers of adults and nymphs were divided by the total number of new shoots, and the ratios were transformed into square roots in order to approximate the data distribution to a normal distribution for the conditions of the use of parametric methods. The efficiency of guava interplanting on the delay of CG infection was evaluated by the period during which trees had been from the CG infection, and the period in unit of month was compared between the orchards by ANOVA. Since CG infection leads the tree to die within several years (Yang et al., 2006), the curing effect of guava was expected to detect the period from the first infection of a tree by CG until its death was longer in guava interplanted orchards than in non-interplanted orchards. This hypothesis was tested by comparing the survival period of trees after the first infection by CG between NI and GI in Cai Be by ANOVA. For the experiment in My Luong, the effect of guava interplanting on the CG infection was evaluated by the difference in the frequencies of non-symptomatic trees and symptomatic ones between the control orchard and guava interplanting orchards. For this purpose, the numbers of non-symptomatic and symptomatic trees in guava-interplanted orchards were summed for each year, and the proportions were compared by a chi square test with those in the control orchard. For the extract experiments, the numbers of psyllids found on leaves were divided by the number of psyllids initially released (10). These numbers were transformed to square roots for the comparison between treatments by ANOVA. 3. Results 3.1. Guava interplanting in old orchards The annual CG infection rate was lowest in the orchards with the treatment with both non-neonicotinoids and neonicotinoids and in

those with guava interplanting (Table 1). The infection rate in biocontrol orchards was higher than in other orchards. The annual CG proportion did not differ significantly between the six management methods: no management, chemical control by non-systemic, control by systemic insecticides, control by insecticides of both types, biological control by weaver ants, and guava interplanting (F5, 87 ¼ 0.76, P ¼ 0.58). The difference in the annual CG infection rate was marginally significant between the two types in guava interplanting (t ¼ 2.58, df ¼ 6, P ¼ 0.042). 3.2. Guava interplanting in Cai Be Trees grew bigger in GI than in NI throughout the experiment in Cai Be (Fig. 1A), and the canopy size differed significantly between the two orchards in the course of the experiment (D ¼ 0.583, P < 0.001). MANOVA performed on new shoots of three classes showed a significant difference between the two orchards: Pilai’s trace was 1.01 (F39, 513 ¼ 6.63, P < 0.001). New shoots were counted at any sampling time through the experiment in both orchards, but their numbers were usually more in GI than in NI (Fig. 1B). Trees did not show a distinct seasonality in flushing in the year, but roughly repeated flushing in a two-month cycle. The total numbers of both psyllid adults and nymph colonies counted on individual trees were less in GI than in NI until the second year, but they increased in both orchards after April 2007 (Fig. 2). No significant differences in these numbers were detected by MANOVA: the Pilai’s trace was 0.012 (F2, 182 ¼ 1.10, P ¼ 0.33). The study period was divided into two periods, one before March 2007 and the other after April 2007, and the data were analysed for each period. The psyllid numbers were compared for each of adult and nymph between the orchards by univariate ANOVA. Both the numbers of adults and nymphs were significantly different in the first period before March 2007 (F1, 183 ¼ 8.77, P ¼ 0.03; F1, 183 ¼ 10.63, P ¼ 0.001), but neither in the second after April 2007 (F1, 152 ¼ 3.60, P > 0.05; F1, 152 ¼ 0.02, P > 0.05). The ratio of the numbers of adults and nymph colonies to those of new shoots were also compared in the following three terms: all sampling times, samplings before March 2007, samplings after April 2007. The ratio was higher in NI than in GI in any term (Table 2). MANOVA revealed a significant difference in the data for all terms: over all samplings (Pilai’s trace ¼ 0.113, F3, 18 ¼ 11.6, P < 0.001); before March 2007 (Pilai’s trace ¼ 0.084, F3, 182 ¼ 8.34, P < 0.001); and after April 2007 (Pilai’s trace ¼ 0.045, F3, 152 ¼ 3.60, P ¼ 0.030). The CG pathogen was first detected in November 2005 in NI, and in September 2006 in GI, thus one year and two months earlier in NI (Fig. 3A). In the first two years, the CG infection proportion remained higher in NI than in GI, but there were no apparent differences in the proportion between the two orchards in the third

Table 1 Years after planting and the mean (SE) of annual CG infection rate in orchards of C. nobilis studied from May to November 2005 in southern Vietnam. In these orchards, vector psyllids were controlled using one of six management strategies: no practice for the control of psyllids (No control), use of only non-neonicotinoids (Non-neonicotinoid), use of only neonicotinoids (Neonicotinoid), use of both types of insecticides (Both), no use of insecticides but use of weaver ant (Biocontrol), no use of insecticides but use of Oecophylla smaragdina, and no use of insecticides but use of interplanting of guavas (Guava). Control method

No control Non-neonicotinoid Neonicotinoid Both Biocontrol Guava Guavaa a b

No. of orchards

33 27 10 13 2 6 2

Years after planting

CG-infection rate (%/year)

Meanb

SE

Min

Max

Meanb

SE

Min

Max

5.83 4.31 3.20 4.58 2.50 2.88 3.00

0.89 0.43 0.20 0.33 0.50 0.13 0.00

1 2 2 2 2 2 3

20 10 4 7 3 3 3

3.98 3.06 3.54 2.66 6.28 2.53 3.10

0.95 0.18 0.17 0.45 2.49 0.17 0.11

1.22 1.37 2.80 0.00 3.79 1.71 2.99

33.30 4.64 4.11 7.34 8.77 2.71 3.22

In these orchards, planted trees were produced by marcotting. Geometric means are calculated for this table.

K. Ichinose et al. / Crop Protection 34 (2012) 119e126

A Orchard NI GI

No. of adults (/tree)

Crown volume (m3)

20

1.5

10

Orchard NI GI

1.0

0.5

1.0

B No. of nymph colonies (/tree)

No. of new shoots (/tree)

A

0.0 Jun 05 Dec 05 Jun 06 Dec 06 Jun 07 Dec 07 Jun 08

0 Jun 05 Dec 05 Jun 06 Dec 06 Jun 07 Dec 07 Jun 08

60

123

Orchard NI GI

40

20

B Orchard NI GI

0.5

0.0

0 Jun 05 Dec 05 Jun 06 Dec 06 Jun 07 Dec 07 Jun 08

Jun 05 Dec 05 Jun 06 Dec 06 Jun 07 Dec 07 Jun 08

Fig. 1. Mean (SE) canopy size (A) and number of new shoots (B) of citrus trees planted in two experimental orchards without/with guava interplanting (NI and GI, respectively).

Fig. 2. Mean numbers (SE) of psyllid adults (A) and nymph colonies (B) per tree in two C. nobilis orchards with/without guava interplanting in Cai Be, southern Vietnam.

year. The mean period until the first detection of the pathogen was significantly longer in GI than in NI: 26.8 (SE ¼ 2.20) months versus 16.5 (SE ¼ 2.2) months, respectively (F1, 25 ¼ 7.66, P ¼ 0.01). The survival period of trees from the first detection of the pathogen until death did not differ significantly between the two orchards (F1, 11 ¼ 0.83, P ¼ 0.38): 5.7 (SE ¼ 2.2, n ¼ 6) months in NI and 8.5 (SE ¼ 2.0, n ¼ 7) months in GI. Almost all trees in both orchards were infected by January 2008.

(SE ¼ 0.0) in the planting month (July 2008), 3.3 (SE ¼ 0.5) five months after planting (December 2008), 5.5 (SE ¼ 0.6) eight months later (March 2009), 8.6 (SE ¼ 1.2) 10 months later (May 2009), and 10.0 (SE ¼ 1.0) 11 months later (June 2009). The invasion of psyllids occurred within one month after planting, and adults were counted more or less at any sampling time. Nymphs were found at only the first sampling time, and none was detected thereafter. The mean densities of adults and nymphs were always smaller than 0.1/tree. The first CG symptoms were observed about a half year after planting both in quadrats without guavas and in those with guava interplanting (Fig. 3B). The CG proportion reached 30% in one year in the former, whereas it took one and half years to reach this

3.3. Field experiment in Dong Thap The mean numbers of shoots on 120 trees in the quadrats with guavas in Dong Thap were increased with sampling time: 0.0

Table 2 Mean numbers (SE) of new shoots, psyllid adults and nymph colonies per tree per sampling from May 2005 to July 2009 in two orchards without/with guava interplanting (NI and GI, respectively) in Cai Be, southern Vietnam. Orchard Trees

Total

Numbers/tree/sampling New shoots

NI GI Until March 2007 NI GI After April 2007 NI GI

Ratio to shoot Adult

Nymph

Adult

Nymph

Mean

SE

Mean

SE

Mean

SE

Mean

SE

Mean

SE

81 104

3.32 10.83

0.23 0.31

0.10 0.08

0.02 0.01

0.06 0.06

0.01 0.01

0.115 0.056

0.012 0.006

0.084 0.045

0.012 0.005

81 104

3.53 9.76

0.22 0.33

0.03 0.01

0.01 0.00

0.04 0.01

0.01 0.00

0.043 0.012

0.008 0.003

0.045 0.006

0.011 0.002

55 99

4.45 12.87

0.40 0.50

0.29 0.17

0.07 0.03

0.13 0.13

0.03 0.02

0.220 0.079

0.026 0.009

0.119 0.065

0.018 0.008

124

K. Ichinose et al. / Crop Protection 34 (2012) 119e126

CG-infected trees (%)

100

higher in the control orchard than in orchards with guava interplanting in the first year (Table 3). In the first year, the infection proportion was all below 10% in orchards with guavas, and 45% in the control orchard. In the second year, the proportion varied from 25 to 100 %, mostly around 50e60 %, in the former and 60% in the latter. Thus, in the second year, although the tree size in the orchard without guavas was still smaller than that in orchards with guavas, the difference in the proportion of CG-infected trees between orchards was reduced from the previous year to this year. The test showed that significantly more trees were infected by CG in the control orchard in the first year (c2 ¼ 41.79, P < 0.001, Yeat’s correction), but no significant differences were detected in the second year (c2 ¼ 0.55, P ¼ 0.46).

A Orchard NI GI

50

0

3.5. Experiments on guava extract

Jun 05

Symptomatic trees (%)

100

Dec 05

Jun 06

Dec 06

Jun 07

Dec 07

Jun 08

B No guava Guava interplanting

50

0 Jul 07

Jan 08

Jul 08

Jan 09

Jul 09

Fig. 3. Proportion of trees infected by citrus greening (CG) in the orchard without/with guava interplanting in Cai Be (A), and in quadrats without/with guava interplanting in Dong Thap (B). The CG infection in Cai Be was determined by PCR and that in Dong Thap by observation of symptoms.

The mean numbers of attracted psyllids were 1.9 (SE ¼ 1.1) on leaves with methanol extract, 2.0 (SE ¼ 1.4) on those treated with acetone, and 0.6 (SE ¼ 0.6) on those treated with hexane. Although these numbers were not significantly different between treatments (F2, 12 ¼ 0.47, P ¼ 0.63), the lower two extracts, acetone and hexane, were selected for the second experiment. More psyllids were observed on leaves with the acetone extract, 2.9 (SE ¼ 0.9), than on those with the hexane extract, 0.0 (SE ¼ 0.0), while leaves without extract attracted most psyllids, 6.8 (SE ¼ 0.8). The difference in the psyllid number was significant between treatments (F2, 15 ¼ 24.2, P < 0.001). The difference between acetone and hexane was significant (Tukey’s test, P ¼ 0.03). Finally, we compared leaf selection by psyllids between leaves treated either by no extract, by hexane solvent only, or by hexane extract. No psyllids were seen on hexane extracts, 0.0 (SE ¼ 0.0), and more psyllids were seen on control leaves, 6.5 (SE ¼ 1.9), while the psyllid number on leaves treated with only hexane solvent was intermediate, 3.3 (SE ¼ 2.4). The number of psyllids was significantly different between the treatments (F2, 18 ¼ 24.22, P < 0.001), and the leaves treated with hexane extract were least selected (Tukey’s test, P < 0.001). 4. Discussion

proportion in the latter. Irrespective of the interplanting treatment, more than 80% of trees showed symptoms of CG in all quadrats in the two years after planting. 3.4. Infection of trees by CG in My Luong The experiment in My Luong showed that the tree size was smaller and the proportion of trees with CG symptoms was much Table 3 Means of circumference of trees (SE) and the number of symptomatic trees in orchards with/without guava interplanting in My Luoung, southern Vietnam one (2008) and two years (2009) after planting of C. nobilis seedlings. In both years, 20 trees were randomly selected in each orchard for the measures in this study. Symptomatic trees (%)a

Orchard no Interplanting Circumference (mm) 2008

1 3 4 5 6 7 8 9 10 a

Guava Guava Guava Guava Guava Guava Guava Guava Control

2009

2008

2009

Mean

SE

Mean

SE

No. (%)

No. (%)

81.9 68.3 46.8 50.3 85.9 84.8 52.1 71.0 40.1

5.2 4.4 5.7 4.9 5.3 5.3 4.2 1.7 4.3

100.4 99.8 79.3 62.2 94.8 127.5 116.4 112.6 63.2

5.4 5.7 5.0 2.8 6.6 7.1 8.3 2.9 2.1

0 1 0 2 0 1 0 0 9

20 8 10 13 12 8 5 6 12

(0) (5) (0) (10) (0) (5) (0) (0) (45)

(100) (40) (50) (65) (60) (40) (25) (30) (60)

Numerals in parentheses represent the proportion of symptomatic trees.

4.1. Limit to the efficacy of guava Our field experiment in two orchards in Cai Be showed that guava interplanting effectively protected citrus trees from CG infection for the first year and four months. Since the CG pathogen can be detected a few to several months after the inoculation (Lopes et al., 2009), the protection by guava interplanting may have completely protected the orchard from invasion by the disease for about one year. Without guavas, the infection percentage reached about 20% in about a half year in the non-guava interplanted orchard (NI). However, the infection proportion in the guava interplanted orchard (GI) increased as much as that in NI in two and half years. This finding indicates that guava interplanting could be effective for the management of CG for one year, but its efficacy would be weakened or lost in two years. Similar results were obtained in the other two experiments in the other locations. The efficacy of guava interplanting on CG was thus limited to one year and a few months at most. The effect of guava on the occurrence of CG infection seems to have been a result of the protection of new shoots from infestation by psyllids. This effect is evident from the analyses of the ratio of the numbers of psyllids to new shoots in NI and GI in Cai Be. New shoots were always more abundant in GI than in NI throughout the experiment. Although psyllids are more likely to be attracted to new shoots (Tsai et al., 2002; Yasuda et al., 2006; Ikeda and Ashihara, 2008), the number of psyllids per tree did not differ significantly between the two orchards. The lack of this difference

K. Ichinose et al. / Crop Protection 34 (2012) 119e126

evidently resulted in a smaller ratio of the number of psyllids to that of new shoots in GI compared to NI. The decrease in the ratio is indicative of the efficacy of guava interplanting on the invasion/ dispersion of psyllids. Despite the efficacy, GI was invaded by as many psyllids as NI after the second year, thus leading to no difference in the infection proportion in the third year. A similar infection pattern of trees by CG was seen in the quadrats with guava in Dong Thap. These results indicate that efficacy of guava interplanting in heavily infected areas could be guaranteed for at most one and a half years after planting. Although the transmission efficiency of the pathogen by the psyllid is not known in detail, some estimates suggest that one psyllid would transmit the pathogen to the host plant by sucking for 1 h (Halbert and Manjunath, 2004; Yang et al., 2006). Perennial crops such as fruits perpetually face the risk of infestation by pests, and the risk could be additive from year to year. The tolerance level for psyllids thus should be set at nearly 0 for transmissible diseases on these crops. This is quite different from pests that induce feeding damage. If the psyllid were not the vector of CG, the damage that it causes would be trivial except for the effect on very young trees (Gottwald et al., 1989; Yang et al., 2006). The control level of other psyllids that cause feeding damage in crops is at a higher level than one individual per plant, and considerable time is required for serious damage to occur (e.g., Mensah and Madden, 1992; Liu et al., 2006). In contrast, since CG may be transmitted by the psyllid in a very short time as described above, an effective method for the control of the psyllid should ideally eliminate psyllids on the crop instantly and maintain this effect for a long time. This explains the limit in the efficacy of guava interplanting for the management of CG. 4.2. Mechanism of guava interplanting on the reduction of CG Before this study we had two hypotheses about the efficacy of guava interplanting on CG: either the guava had a remedial effect on diseased trees or the guava had a protective effect from vector invasion or dispersion by interception. If the first hypothesis were correct, a curing effect of guava interplanting on infected trees would have been seen in the two cases in this study: 1) The percentage of infected trees in guava orchards where tree propagation for citrus was achieved by marcotting from infected mother plants would have been as low as that in guava interplanting orchards where young trees produced by grafting had been planted, or at least should have been lower than in orchards with no control method, and 2) the survival period of infected trees, defined as the time of the first detection of the CG pathogen until their death, should have been elongated more in GI than in NI. The proportion of trees infected by CG in old orchards was higher in orchards with trees produced by marcotting than in orchards with grafted trees. The survival period showed no significant difference between NI and GI. These results do not support the hypothesis of a remedial effect of guavas. The second hypothesis on the protective effect seems to be more probable, and it is supported by the results in the Cai Be orchards of fewer psyllids per tree until March 2007 in GI than in NI. Ratios of the psyllid number to the new shoot number were smaller in GI than in NI throughout the experiment, indicating that fewer psyllids invaded GI early in the experiment. This should have contributed to lower proportions of CG-infected trees, especially in the early stage of the experiment when trees were small and young and fewer psyllids were present. However, no apparent differences in the psyllid number per tree were detected between NI and GI after April 2007. The differences in the tree per capita infestation by the psyllid may be related to the relative tree size of the citrus trees compared to the guava trees. After the second year, the citrus trees were generally much bigger than the guava trees, possibly diluting the guava effect per capita of citrus. In turn, the efficacy of guava

125

interplanting may be involved in the relative tree size of guava trees compared to citrus trees. Seven hypotheses have been suggested so far to explain the occurrence of pest reduction by interplanting, and a more general hypothesis has been proposed by Finch and Collier (2000): once pests are attracted to the general area of the correct host plants, insects land non-selectively on any green matter, irrespective of the appropriateness of the plant as a host (Hori, 1999; Morley et al., 2005; Bukovinszky et al., 2007). Thus, if host plants are monocultured, insects can reach the correct hosts more easily and accurately, and if host plants are not monocultured the converse is true. The rate at which pests land on correct host plants may depend on the presence of interplanted plants that interrupt the visual detection of hosts by the insect, thus reducing the success of the pest invasion (Floater and Zalucki, 2000). Although visual interruption was not examined in our experiments, the results of our experiments on guava extracts suggest that some volatiles of guava could also play a role in the psyllid reduction by functioning as repellents against the psyllids. A protective effect of chemical substances in interplanted plants is known in some crops (Miklasiewicz and Hammond, 2001; Held et al., 2003; Kong et al., 2005), while it has not been confirmed in others (Finch et al., 2003). If these substances can be applied in citrus orchards, the risk of CG infection could be reduced further in combination with guava interplanting or insecticide use. Citrus psyllids are known to move frequently between orchards (Boina et al., 2009) but rarely travel over long distances (Hayashikawa et al., 2007; Arakawa and Miyamoto, 2007), although inter-insular migration might be possible in rare cases (Sakamaki, 2005). Various invasion distances may be realised by the psyllid in the Mekong Delta. Such migration patterns of the psyllid may make it difficult to protect orchards from psyllid invasions by the use of only a repellent. Aside from the efficacy of guava interplanting in the reduction of CG, guava also may facilitate the growth of citrus, possibly due to physiological effects as well as a windbreak effect on the citrus. Since CG weakens citrus trees and induces them to be less tolerant of other diseases (Chang and Bay-Petersen, 2003), the protection of trees from CG, especially in the initial period of growth, will provide the citrus with a better chance to remain healthy and produce a greater fruit yield. Physiological interaction of guavas with interplanted crops is known in the sunflower (Bheemaiah and Subrahmanyam, 2004). In our field experiment in Cai Be, the owner farmer of the field told us that the yield in 2007 in GI was 10 times greater than that in NI, at 300 kg/1000 m2 versus 30 kg/ 1000 m2, respectively, and the trees grew faster and flushed more new shoots in GI. Since data on yield were not systematically obtained in this experiment, e.g., yield per tree, harvesting times, quality of fruits, and so on, an economic evaluation of guava interplanting was not possible. However, farmers in southern Vietnam have been performing guava interplanting with an expectation that yields from both the C. nobilis and guava in guava interplanting orchards promise to be greater than those from only C. nobilis in non-interplanting orchards. In particular, guava provides yields in about a half year after planting, while C. nobilis cannot yield fruits until one and a half years later. This earlier yield seems to make an important contribution to the farmer’s economy. The difference in the time of the first cropping between guava and citrus is one of the biggest advantages of the guava interplanting. Acknowledgments This study was financially supported by the collaborative research project for control of citrus greening in the Mekong Delta Region of Vietnam between the Japan International Research Centre for Agricultural Sciences and the Southern Horticultural

126

K. Ichinose et al. / Crop Protection 34 (2012) 119e126

Research Institute of Vietnam. We thank Dr. N. M. Chau and Dr. L. T. T. Hong for their official support of our research. Dr. K. Yasuda, Dr. M. Koizumi, and Dr. T. Kano provided us with helpful suggestions for our experiments. Dr. H. Iwahori identified the nematode that infested guavas in My Luong, southern Vietnam. References Arakawa, K., Miyamoto, K., 2007. Flight ability of Asiatic citrus psyllid, Diaphorina citri Kuwayama (Homoptera; Psyllidae), measured by a flight mill. Res. Bull. Pl. Prot. Jpn. 43, 23e26. Arima, H., Danno, G.I., 2002. Isolation of antimicrobial compounds from guava (Psidium guajava L.) and their structural elucidation. Biosci. Biotechnol. Biochem. 66, 1727e1730. Bheemaiah, G., Subrahmanyam, V.R., 2004. Effect of green-leaf manuring on productivity of sunflower (Helianthus annuus) under guava (Psidium guajava) - based agrihorticultural system. Ind. J. Agr. Sci. 74, 515e520. Boina, D.R., Meyer, W.L., Onagbola, E.O., Stelinski, L.L., 2009. Quantifying dispersal of Diaphorina citri (Hemiptera: Psyllidae) by immunomarking and potential impact of unmanaged groves on commercial citrus management. Environ. Entomol. 38, 1250e1258. Bové, J.M., 2006. Huanglongbing: a destructive, newly emerging, century-old disease of citrus. J. Plant Pathol. 88, 7e37. Bukovinszky, T., Tréfás, H., van Lenteren, J.C., Vet, L.E.M., Fremont, J., 2007. Plant competition in pest-suppressive intercropping systems complicates evaluation of herbivore responses. Agr. Ecosys. Environ. 102, 185e196. Chang, W.N., Bay-Petersen, J., 2003. Citrus Production: a Manual for Asian Farmers. FFTC, Taipei, Taiwan, p. 85. Da Graça, J.V., 1991. Citrus greening disease. Ann. Rev. Phytopath. 73, 1358e1363. Finch, S., Billiald, H., Collier, R.H., 2003. Companion planting e do aromatic plants disrupt host-plant finding by the cabbage root fly and the onion fly more effectively than non-aromatic plants? Entomol. Exper. Appl. 109, 183e195. Finch, S., Collier, R.H., 2000. Host-plant selection by insects e a theory based on ‘appropriate/inappropriate’ landings’ by pest insects of cruciferous plants. Entomol. Exper. Appl. 96, 91e102. Floater, G.J., Zalucki, M.P., 2000. Habitat structure and egg distributions in the processionary caterpillar Ochrogaster linifer: lessons for conservation and pest management. J. Appl. Entomol. 37, 87e99. Gonçalves, F.A., Andrade Neto, M., Bezerra, J.N.S., Macrae, A., de Sousa, O.V., Fonteles-Filho, A.A., Vieira, R.H.S.F., 2008. Antibacterial activity of guava, Psidium guajava Linnaeus, leaf extracts on diarrhea-causing enteric bacteria isolated from seabob shrimp, Xiphopenaeus kroyeri (Heller). Rev. Inst. Med. Trop. S. Paulo 50, 11e15. Gottwald, T.R., Aubert, B., Xue-Yan, Z., 1989. Preliminary analysis of citrus greening (Huanglongbing) epidemics in the People’s Republic of China and FrenchReunion Island. Phytopathology 150, 687e693. Hall, D., Hentz, M.G., Adair Jr., R.C., 2008. Population ecology and phenology of Diaphorina citri (Hemiptera: Psyllidae) in two Florida citrus groves. Environ. Entomol. 37, 914e924. Hayashikawa, S., Suenaga, H., Matsuhira, K., 2007. Invasion of Asian citrus psyllids, Diaphorina citri into citrus orchards in residential areas. Kyushu Plant Prot. Res. 53, 99e102. Halbert, S.E., Manjunath, K.L., 2004. Asian citrus psyllids (Sternorhyncha: Psyllidae) and greening disease of citrus: a literature review and assessment of risk in Florida. Fla. Entomol. 87, 330e353. Held, D.W., Gonsiska, P., Potter, D.A., 2003. Evaluating companion planting and non-host masking odors for protecting roses from the Japanese beetle (Coleoptera: Scarabaeidae). J. Econ. Entomol. 96, 81e87. Hori, M., 1999. Role of host plant odors in the host finding behaviors of aphids. Appl. Entomol. Zool. 34, 293e298.

Hung, T.H., Hung, S.C., Chen, C.N., Hsu, M.H., Su, H.J., 2004. Detection by PCR of Candidatus Liberibacter asiaticus, the bacterium causing citrus huanglongbing in vector psyllids: application to the study of vector-pathogen relationships. Plant Pathol. 53, 96e102. Ichinose, K., Kano, T., 2006. Citrus greening disease and its management in the Mekong Delta Region of Vietnam. Shokubutsuboeki 60, 302e307 (in Japanese). Ikeda, K., Ashihara, W., 2008. Preference of adult Asian citrus psyllid, Diaphorina citri (Homoptera: Psyllidae) for Murraya paniculata and Citrus unshiu. Jpn. J. Appl. Entomol. Zool. 52, 27e30 (in Japanese with English summary). Iwahori, H., Truc, N.T.N., Bang, D.V., Ichinose, K., 2009. First report of root-knot nematode Meloidogyne enterolobii on guava in Vietnam. Plant Dis. 93, 675. Jagoueix, S., Bové, J.M., Garnier, M., 1994. The phloem-limited bacterium of greening disease of citrus is a member of the a subdivision of the proteobacteria. Int. J. Sys. Bacteriol. 44, 379e386. Koizumi, M., Prommintara, M., Linwattana, G., Kaisuwan, T., 1997. Epidemiological aspects of citrus huanglongbing (greening) disease in Thailand. JARC 31, 205e211. Kong, C., Hu, F., Xu, X., Zhang, M., Liang, W., 2005. Volatile allelochemicals in the Ageratum conyzoides intercropped citrus orchard and their effects on mites Amblyseius newsami and Panonychus citri. J. Chem. Ecol. 31, 2193e2203. Liu, D., Johnson, L., Trumble, J.T., 2006. Differential responses to feeding by the tomato/potato psyllid between two tomato cultivars and their implications in establishment of injury levels and potential of damaged plant recovery. Ins. Sci. 13, 195e204. Lopes, S.A., Bertolini, E., Frare, G.F., Martins, E.C., Wulff, N.A., Teixeira, D.C., Fernandes, N.G., Cambra, M., 2009. Graft transmission efficiencies and multiplication of ‘Candidatus Liberibacter americanus’ and ‘Ca. Liberibacter asiaticus’ in citrus plants. Phytopathology 99, 301e306. Mensah, R.K., Madden, J.L., 1992. Feeding behaviour and pest status of Ctenarytaina thysanura (Ferris and Klyver) (Hemiptera: Psyllidae) on Boronia megastigma (Nees) in Tasmania. J. Aust. Entomol. Soc. 31, 71e78. Miklasiewicz, T.J., Hammond, R.B., 2001. Density of potato leafhopper (Homoptera; Cicadellidae) in response to soybean-wheat cropping system. Environ. Entomol. 30, 204e214. Morley, K., Finch, S., Collier, R.H., 2005. Companion planting e behaviour of the cabbage root fly on host plants and non-host plants. Entomol. Exper. Appl. 117, 15e25. Mohamed, S., Hassan, Z., Hamid, N.A., 1994. Antimicrobial activity of some tropical fruit wastes (guava, starfruit, banana, papaya, passionfruit, langsat, duku, rambutan and rambai). Pertanika J. Trop. Agri. Sci. 17, 219e227. Sakamaki, Y., 2005. Possible migration of the Asian citrus psyllid, Diaphorina citri Kuwayama (Homoptera: Psyllidae) between and within islands. Kagoshima Univ. Res. Center. Pacific Islands, Occasional Papers 42, 121e124 (in Japanese with English summary). Tsai, J.H., Wang, J.J., Liu, Y.H., 2002. Seasonal abundance of the Asian citrus psyllid, Diaphorina citri (Homoptera: Psyllidae) in southern Florida. Fla. Entomol. 85, 446e451. Urasaki, N., Uehara, T., Kawano, S., 2007. Highly reliable polymerase chain reaction assay for citrus huanglongbing (citrus greening disease) using the sucrose synthase gene as an internal control. Jpn. J. Phytopathol. 73, 25e28 (in Japanese with English summary). Van Mele, P., Cuc, N.T.T., Van Huis, A., 2002. Direct and indirect influences of the weaver ant Oecophylla smaragdina on citrus farmers’ pest perceptions and management practices in the Mekong Delta, Vietnam. Int. J. Pest Manage. 48, 225e232. Yang, Y., Huang, M., Andrew, G.C.B., Xia, Y., Ouyang, G., Xiong, J., 2006. Distribution, biology, ecology and control of the psyllid Diaphorina citri Kuwayama, a major pest of citrus: a status report for China. Int. J. P. Manage. 52, 343e352. Yasuda, K., Kawamura, F., Ohishi, T., 2006. Location and preference of adult Asian psyllid, Diaphorina citri (Homoptera: Psyllidae) on Chinese box orange jasmine, Murraya exotica L. and flat lemon, Citrus depressa. Jpn. J. Appl. Entomol. Zool. 49, 146e149 (in Japanese with English summary).