Sumatra disease of clove usually affects productive trees over 10 years of age (Hadiwijaya, 1983). Externally the initial symptom of Sumatra disease of cloves is unseasonal yellowing of leaves followed by leaf-drop from the tips of branches high in the crown (Figure 2). However, the leaves may also wilt suddenly and turn brown, but stay attached to the branch. Affected twigs turn reddish brown and progressively die back (Figures 2a,b). Symptoms typically progress to lower branches until the whole crown is affected, and the tree dies within 6–18 months (Bennett et al., 1985). Artificial inoculation of R. syzygii subsp. syzygii on 3 months old S. aromaticum seedling leads to symptom appearance beginning with leaf yellowing and drying at 28 days after inoculation and the death of the seedling 56 days after inoculation (Danaatmadja et al., 2009).

Internally, the newly formed wood adjacent to the cambium becomes discolored a pale grayish-brown. When cut infected branches often produce a milky white to pale brown bacterial ooze from the cut surface (Figure 2c). The discolouration of the xylem can be traced down the trunk into one or more major roots (Bennett et al., 1985).

Sumatra disease of clove was initially observed in 1975 (Waller and Sitepu, 1975) and further reported in 1985 (Bennett et al., 1985). The disease affects S. aromaticum and some species of Myrtaceae including some indigenous species in native forests in Indonesia (Lomer et al., 1992), such as Syzygium aqueum (Eden–Green et al., 1992). The disease, which was initially confined to the Indonesian provinces of Sumatra and West Java, has now spread to Central Java and East Java and causes economic losses of up to 5–10% per year (Tjahyono, 2013).

Initially Sumatra disease of clove was assumed to be caused by nutritional disorders (Hadiwidjaja, 1956), mineral toxicities (Finck, 1972) as well as disease causing agents such as Ralstonia solanacearum [Xanthomonas solanacearum, Pseudomonas solanacearum (Hidir, 1973)], Phytophthora spp. (Djafaruddin et al., 1979), leaf spot fungi Phyllostictina sp. (Kranz, 1976), and the fastidious-xylem limited bacteria (XLB) (Bennett et al., 1985; Hunt et al., 1985) prior to the causative agent being taxonomically described as Pseudomonas syzygii (Roberts et al., 1990). Although R. syzygii subsp. indonesiensis strains have been isolated from the roots and lower trunk of trees only R. syzygii subsp. syzygii can systemically colonize and kill S. aromaticum trees.

Ralstonia syzygii subsp. syzygii is included as one of the xylem-restricted or xylem-limited bacteria, which live in xylem cells or tracheary elements of plants (Purcell and Hopkins, 1996). Similar to other diseases caused by xylem-restricted bacteria, R. syzygii subsp. syzygii is transmitted by insect vectors that feed on xylem sap (Bové and Garnier, 2002). The tube-building cercopoid (Hemiptera), Hindola fulva was found to be the natural insect vector in Sumatra whereas H. striata (Hemiptera: Machaerotidae) has been described as the primarily vector in Java (Eden–Green et al., 1992). Vector transmission of R. syzygii subsp. syzygii is persistent with a short latent period between acquisition and transmission of the pathogen (Eden–Green et al., 1992). Lomer et al. (1992) suggested that Sumatra disease of cloves may have been transferred to clove trees from an unknown forest hosts because of the localized initial distribution of the disease and the corresponding localized distribution of the vector species. If this hypothesis is correct then the pathogen may have a wider host range than has been previously identified (Purcell and Hopkins, 1996).

Sumatra disease of clove has a distinct pattern of disease expression and distribution in the field. Seedlings less than 2 years old are rarely affected, with trees over 10 years of age being the first to show symptoms and the first to die. The disease advances on a broad front, at an estimated rate of 1–2 km per year and then disappears for years until young trees mature and the cycle repeats (Bennett et al., 1985). As would be expected of an insect transmitted disease the disease spreads in a jump-spread pattern and spreads rapidly in all directions uphill, downhill, and across rivers (Bennett et al., 1985). Bennett et al. (1985) also reported that the rate of spread and symptom expression was partially affected by altitude possibly due to the lower temperature at higher altitudes where rapid decline symptoms are most commonly observed.

Symptoms of Blood Disease are quite similar to Moko disease caused by insect-transmitted strains, the male flower bud and peduncle discolor and shrivel, the fruit pulp shows a reddish dry rot (Figure 3), and the vascular tissue throughout the plant exhibits a reddish discoloration, which emits reddish-brown bacterial ooze when cut (Sequeira and Averre, 1961; Buddenhagen, 1962). Moko disease of Musa spp. is caused by R. solanacearum strains which belong to phylotype II of the R. solanacearum species complex (Fegan and Prior, 2006). The older leaves of blood disease-infected Musa spp. become yellow, followed by wilting, necrosis and collapse; younger leaves turn bright yellow before becoming necrotic and dry. The pathogen rapidly colonizes the entire plant, and suckers will also wilt and die (Eden-Green, 1994b).

The banana blood disease pathogen is disseminated in soil and water and on farm tools, and has been hypothesized to enter host roots through natural openings or wounds (Gäumann, 1921, 1924). Gäumann (1921) reported that the pathogen survived in soil for at least a year in infested plant residues and infected fruits after entering the plant through its roots. Infested soil, tools, and vehicles move the pathogen within plantations, and movement of infected fruit and planting material enable long-distance spread. Insects that visit Musa spp. inflorescences, particularly those of cultivars with dehiscent bracts and an ABB genome (a plantain with one set of chromosomes donated by Musa acuminata and two by Musa balbisiana), can spread the pathogen rapidly over great distances (Mairawita et al., 2012). R. syzygii subsp. celebesensis has been isolated from the insect species Trigona minangkabau (Mairawita et al., 2012) and Erionota thrax (Suharjo et al., 2008). Erionota thrax has been observed to visit banana flowers 2–3 times a day (Suharjo et al., 2008). The bracts of ABB/BBB genotype Musa spp. are non-persistent and as they fall off they leave abscission scars which provide sites for pathogen entry into the vascular tissue of the plant. Also the male buds of the highly susceptible cultivar “Pisang Kepok” appear to be particularly attractive to insects such as wasps, bees and flies possibly because the male flower nectar has a high sugar content (Setyobudi and Hermanto, 1999).

The transmigration of people from Java to less populated islands in the country appears to be associated with the spread of the disease. R. syzygii subsp. celebesensis is thought to have originated on Selayar Island near Sulawesi, as the disease was first reported in the early 1900’s after the introduction of dessert bananas (Eden–Green, 1994a). The disease spread to Java in the late 1980’s and has become common on local M. paradisiaca cultivars in Sulawesi (Stover and Espinoza, 1992). Unfortunately, the pathogen has spread to most of the larger Indonesian islands, with average yield losses exceeding 35% (Supriadi, 2005), and has also been reported on the island of New Guinea (Davis et al., 2001).

The host range of R. syzygii subsp. celebesensis is not as wide as the R. solanacearum strains causing Moko and Bugtok diseases on Musa spp. Baharuddin (1994) showed that Heliconia sp. and Strelitzia reginae, both relatives of the Musaceae, are susceptible to R. syzygii subsp. celebesensis as are Canna indica, Datura stramonium, Asclepias currassiva, and Solanum nigrum. Unlike R. solanacearum strains causing Moko disease, R. syzygii subsp. celebesensis is not pathogenic on S. lycopersicum and Solanum melongena seedlings (Cellier and Prior, 2010; Eden–Green, 1994a; Supriadi, 2005). Strains of R. syzygii subsp. celebesensis are also not able to infect Arachis hypogaea, Capsicum sp., Nicotiana tabacum, S. tuberosum, and Zingiber officinale (Baharuddin, 1994).

Ralstonia syzygii subsp. indonesiensis causes disease in a number of solanaceous plants in Indonesia and other countries in Asia but has also has been isolated from clove plants in Indonesia (Table 1). The disease symptoms caused by R. syzygii subsp. indonesiensis strains on solanaceous crops are no different from those described in the past for R. solanacearum (Kelman, 1953). The external symptoms of the infected plants are wilting, stunting and yellowing of the foliage with the disease progressing until the plant completely collapses from wilt (Figure 4). Internally the vascular tissue becomes progressively discolored in the early stages of infection with portions of the pith and cortex becoming involved as disease develops until complete necrosis occurs.

In Japan R. syzygii subsp. indonesiensis strains have only been isolated from S. tuberosum (Horita et al., 2014). The R. syzygii subsp. indonesiensis strains isolated from S. tuberosum in Japan have been reported to show varying levels of pathogenicity for S. lycopersicum, A. hypogaea and N. tabacum (Suga et al., 2013). However, none of the strains tested by Suga et al. (2013) were pathogenic for S. melongena. A R. syzygii subsp. indonesiensis strain isolated from S. lycopersicum in Indonesia (PSI07) was found to not to cause wilting of Cucumis melo, Anthurium andraeanum, Musa spp. and S. tuberosum but retained pathogenicity for S. lycopersicum (Ailloud et al., 2015).

There is little literature describing the epidemiology and ecology of R. syzygii subsp. indonesiensis strains. However, as is commonly described for other R. solanacearum species complex strains, the bacterium is reported to be able to survive in the field for long periods of time (Suga et al., 2013). Unlike R. solanacearum strains which cause brown rot of S. tuberosum R. syzygii subsp. indonesiensis strains have been shown to only cause disease in S. tuberosum in tropical but not temperate conditions (Cellier and Prior, 2010). In comparison to R. pseudosolanacearum strains Habe (2016) showed that R. syzygii subsp. indonesiensis exhibited high pathogenicity for S. tuberosum at a lower temperature (26°C) than R. pseudosolanacearum strains (28°C).

All known clove varieties appear equally susceptible (Bennett et al., 1985). As indicated above the spread of the disease in the field is via an insect vector, Hindola spp., and most probably contaminated agricultural tools. Therefore local agricultural departments in Indonesia recommend that agricultural tools used for field work should be disinfected between uses, infected plants should be eradicated and insecticide should be applied to minimize the spread of the disease by insect vectors. Although several insecticides have been tested without effective control, aldicarb and carbofuran granules have provided effective control from 7 to 35 and 28 to 217 days after treatment, respectively (Stride and Nurmansyah, 1991). It has further been suggested that if resistant rootstocks could be identified for grafting the degeneration of the roots may be controlled (Bennett et al., 1985). Antibiotics have been shown to be useful in controlling the disease in mature trees for short periods but their use as an effective disease management tool is not recommended (Hunt et al., 1985). Biological control of the pathogen using antagonistic bacterial endophytes and rhizobacteria has also been suggested. Dwimartina et al. (2017) found that endophytic strains of Bacillus subtilis subsp. subtilis and rhizobacteria which produce indole acetic acid and dissolved phosphate could inhibit the growth of R. syzygii subsp. syzygii. However, the agropolitical challenges of S. aromaticum cultivation in Indonesia has made the application of any disease management strategies difficult (Baharuddin, 1994).

Some natural enemies of Hindola spp that may be potential as biological control of the insect vector were identified. Nuhardiyati et al. (1990) reported that H. fulva was found in the population with the unknown species of Hindola, Hindola sp. in Bengkulu, Sumatra, and Indonesia. Stylops sp. was found to parasitize the nymphs and adults of Hindola spp. The nymph of family Tettigoniidae was also found and assumed as the predator of Hindola sp. nymphs (Nuhardiyati et al., 1990). On the other hand, Hemipterian parasitoid paratized the nymph and eggs of Hindola spp. (Balfas et al., 1990). A member of the genus Acmopolynema was the parasite of Hindola spp. in Java, Indonesia (Balfas et al., 1990). This parasitoid of the insect vector of R. syzygii subsp. syzygii is a potential natural biological control agent of the insect vector. Balfas et al. (1990) reported that 30% of Hindola spp. eggs collected from clove and 60–80% of eggs collected from the clove related tree, Xanthostemon chrysanthus, were parasitized by Acmopolynema.

It is thought that all edible Musa spp. may be susceptible to blood disease as no Indonesian cultivars of Musa spp. have been found to be resistant (Gäumann, 1921; Supriadi, 2005). However, some tolerance to blood disease has been reported to occur that may be a source of genetic material for resistance breeding (Supriadi, 2005).

Restricting the movement of planting material from infected areas has been successful in limiting the spread of the disease. A quarantine imposed by the Dutch to limit the spread the disease from Sulawesi was effective for over 60 years until the disease eventually spread to Java around 1987. From this point onward the pathogen has spread rapidly over the Indonesian archipelago and more recently Malaysia.

Removal of the male flower has been found to be effective in controlling the spread of disease as has the use of cultivars that abort the male bud (Hermanto et al., 2013). The cultivar “Pisang Puju,” an acceptable resistant Musa paradisiaca variety from Sulawesi, and “Pisang Sepatu Amora” may be suitable because these cultivars abort the male bud, blocking insect transmission (Hermanto et al., 2013).

A combination of basic quarantine and sanitation practices has been suggested to reduce the spread of blood disease (Davis et al., 2001). These measures include prohibiting movement of Musa spp. plants or plant parts out of infected areas, using disease free-planting materials, removing male buds immediately after the emergence of the last fruit, pesticide application as soon as the symptoms appear to reduce vector related spread and sterilizing the knives for harvesting.

Biological control of blood disease was suppressed by the application of endophytic bacteria including Bacillus sp and Bacillus subtilis isolated from Musa troglodytarum (Hadiba et al., 2010; Laturapeissa et al., 2014). Since the disease is insect-spread through the bacterial contaminated body of insects visiting Musa spp., therefore the biological control of insects associated with Musa spp. is potential for limiting the rate of blood disease spread. Cosmopolites sordidus, the Musa spp. and plantain root and corn borer insect, is not only potential in damaging banana plantation, but also increases the spread rate of blood disease (Subandiyah et al., 2005). This insect pest was reported to be effectively controlled by the application of Steinernematid nematodes (Figueroa, 1988) and Beauveria bassiana (Fancelli et al., 2013).

Habe (2016) identified that there is a degree of resistance in S. tuberosum cultivars to R. syzygii subsp. indonesiensis. Suga et al. (2013) assessed pathogenic differences between R. pseudosolanacearum and R. syzygii subsp. indonesiensis strains against several Japanese varieties of S. tuberosum and breeding lines and indicated that R. syzygii subsp. indonesiensis strains show high virulence to the breeding lines carrying bacterial wilt resistance conferred from the wild Solanum sp., Solanum phureja, which is resistant to R. pseudosolanacearum strains. Several wild species of S. tuberosum such as S. phureja, Solanum stenostomum, and Solanum commersonii have been used as genetic resources to breed for resistance to bacterial wilt worldwide, and certain new S. tuberosum varieties with a high level of resistance have been identified. However, the high levels of resistance of these new S. tuberosum varieties have only been confirmed against R. pseudosolanacearum or R. solanacearum strains. R. syzygii subsp. indonesiensis strains have not been commonly used as targets in breeding for bacterial wilt resistance, most probably because this organism has only recently been identified as a taxonomic group within the R. solanacearum species complex and the restricted distribution [Japan, Korea, the Philippines, India, Indonesia, and Australia (Arwiyanto et al., 2015)] of this organism. Investigations to determine the pathogenicity of R. syzygii subsp. indonesiensis strains toward for new S. tuberosum varieties are needed to address the issue of variability in pathogenicity among different strains. Furthermore, the identification of new genetic resources for breeding need to consider resistance to R. syzygii subsp. indonesiensis strains in the future.