Distribution of Root-Lesion and Stunt Nematodes, and Their Relationship with Soil Properties and Nematode Fauna in Sugarcane Fields in Okinawa, Japan

Sugarcane cultivation in Japan has not yet focused on suppressing plant-parasitic nematodes. For proper nematode management, it is essential to know the spatial distribution of economically important plant-parasitic nematodes and free-living nematodes that play important roles in terrestrial ecosystems. We aimed to reveal nematode fauna and soil properties in 85 sugarcane fields of three major sugarcane producing islands in Japan, and to examine their relationship by using the mixed-effect model and by visualizing the spatial distributions using the inverse distance weighting (IDW) approach. The nematode community structures were analyzed by non-metric multidimensional scaling (NMDS). Among plant-parasitic nematodes in sugarcane, the root-lesion nematodes (Pratylenchus sp.) and the stunt nematodes (Tylenchorhynchus sp.) were widely distributed in these islands, yet the abundance and the species varied geospatially. Soil pH was significantly correlated with the abundance of Pratylenchus and Tylenchorhynchus species. The abundance of Pratylenchus and Tylenchorhynchus species were significantly correlated with soil pH. The abundance of Pratylenchus was significantly correlated with the abundance of free-living nematodes, the number of free-living nematode species, and exchangeable cation K+, as were the abundance of Tylenchorhynchus to the clay content and that of non-Tylenchorhynchus. This study also revealed that the three islands had different nematode faunas, which were explained especially by soil pH, texture, and exchangeable basic cations.


Introduction
More than 310 species of 48 genera of plant-parasitic nematodes have been reported in sugarcane [1], and these nematode genera include Helicotylenchus, Hoplolaimus, Meloidogyne, Paratylenchus, Pratylenchus, Rotylenchulus, Tylenchorhynchus and Xiphinema. Among these, the root-lesion nematode (Pratylenchus sp.), the root-knot nematode (Meloidogyne sp.), and the dagger nematode (Xiphinema sp.) are potentially and between 124 • 6 14.1" E and 124 • 14 44.1" E) in Japan (Figure 1; generated by leaflet package [17] in R). They were the top three sugarcane producing islands in the 2018-2019 season in Okinawa [8]. We separated Okinawa-North and Okinawa-South, because their major soil types are different. The major soil types of each region are red-yellow soil (local name in Japan: Kunigami-Mahji) in Okinawa-North and Ishigaki; dark-red soil (local name: Shimajiri-Mahji) in Miyako; and dark-red soil plus marlitic terrestrial regosols (or grey terrace soil, local name: Jahgaru) in Okinawa-South [18,19]. Soil sampling was conducted in the early fall at the end of September and in the middle of October in 2016 in Okinawa-North and South, in mid-August in 2017 in Miyako and in the early July in 2018 in Ishigaki. The 20-year averages of precipitation and day-time temperature [20] were 2152 ± 425 (standard deviation) mm annum −1 and 22.9 ± 0.3 • C (the north of Okinawa); 2210 ± 439 mm annum −1 and 23.4 ± 0.3 • C (the south of Okinawa); 2086 ± 317 mm annum −1 and 23.9 ± 0.3 • C (Miyako); and 2129 ± 381 mm annum −1 and 24.6 ± 0.3 • C (Ishigaki).
Agronomy 2020, 10, x FOR PEER REVIEW 3 of 17 the top three sugarcane producing islands in the 2018-19 season in Okinawa [8]. We separated Okinawa-North and Okinawa-South, because their major soil types are different. The major soil types of each region are red-yellow soil (local name in Japan: Kunigami-Mahji) in Okinawa-North and Ishigaki; dark-red soil (local name: Shimajiri-Mahji) in Miyako; and dark-red soil plus marlitic terrestrial regosols (or grey terrace soil, local name: Jahgaru) in Okinawa-South [18,19].   [21] published by Geospatial Information Authority of Japan [22] and developed using leaflet package in R.

Soil Sampling
Soil was collected for nematode and physico-chemical analyses in a total of 85 sugarcane fields (18,31,20, and 16 fields in Okinawa-North, Okinawa-South, Miyako, and Ishigaki, respectively). For nematode analysis, the soil was collected with an auger (3 cm diameter) at 0-30 cm depth within 10-15 cm of the base of 5 randomly selected sugarcane plants in each field. The soil was passed through a 5 mm aperture sieve to remove rocks and debris, mixed well and kept at room temperature for no more than two days before nematode extraction. The soil moisture was determined using 100.0 g fresh soil dried in an oven (60 °C × 48 h). For physico-chemical analysis, the soil was collected at a depth of 0-30 cm at 5 randomly selected spots within 20-30 cm of the sugarcane ridge in the same field as for the nematode analysis. The collected soil samples were air-dried, sieved to pass through a 2 mm mesh screen, and stored at ambient temperature until analysis. The maps are based on the blank map [21] published by Geospatial Information Authority of Japan [22] and developed using leaflet package in R.

Soil Sampling
Soil was collected for nematode and physico-chemical analyses in a total of 85 sugarcane fields (18,31,20, and 16 fields in Okinawa-North, Okinawa-South, Miyako, and Ishigaki, respectively). For nematode analysis, the soil was collected with an auger (3 cm diameter) at 0-30 cm depth within 10-15 cm of the base of 5 randomly selected sugarcane plants in each field. The soil was passed through a 5 mm aperture sieve to remove rocks and debris, mixed well and kept at room temperature for no more than two days before nematode extraction. The soil moisture was determined using 100.0 g fresh soil dried in an oven (60 • C × 48 h). For physico-chemical analysis, the soil was collected at a depth of 0-30 cm at 5 randomly selected spots within 20-30 cm of the sugarcane ridge in the same field as for the nematode analysis. The collected soil samples were air-dried, sieved to pass through a 2 mm mesh screen, and stored at ambient temperature until analysis.

Nematode Extraction, Identification and Enumeration
Nematodes were extracted in triplicate from 20 g subsample of each well mixed soil sample to evaluate nematode fauna using the Baermann funnel method (room temperature, 72 h [23]). The soils after the Baermann extraction were further set in the sugar-flotation nematode extraction method to extract remaining nematodes following Jenkins [24] with some modifications [25]. Nematodes were counted and identified based on their morphological characters under a stereomicroscope (BX53, Olympus, Tokyo, Japan). Up to 6 root-lesion nematodes and stunt nematodes were randomly selected from soils of each field and identified using the real-time PCR method (P. zeae and T. leviterminalis by the method in the previous study [26] and P. coffeae and T. annulatus using unpublished real-time PCR primers). Thresholds of Pratylenchus spp. and Tylenchorhynchus spp. for severe damage against sugarcane were set at 80 (and 40) and 100 (and 50) nematodes 20 g soil −1 , respectively, based on previous studies [25,27]. Other plant-parasitic nematodes identified in this study were Helicotylenchus sp., Rotylenchulus spp. (R. reniformis and R. borealis), Hoplolaimus sp., Meloidogyne sp., Paratylenchus sp., Xiphinema sp., and ring nematodes. Among these plant-parasitic nematodes, Meloidogyne sp. and Xiphinema sp. may be potentially harmful [1] for sugarcane, however, their presence in the investigated fields was low and therefore were not examined in this study. Free-living nematodes were counted in total and classified separately for the first 100 individuals or all, if total numbers were less than 100, based on their morphological characters [28,29] and general feeding habits [30]. The classification of free-living nematodes in the current study covered Acrobeloides, other Cephalobidae, Rhabditis, other Rhabditidae, and the other bacterial feeders. The other feeding types including fungal feeders (Aphelenchus, Aphelenchoides, Filenchus and Ditylenchus), mostly omnivorous nematodes in the Dorylaimida order and predator nematodes were also recorded. Soil DNA was extracted from 0.25 g of oven-dried (60 • C, 48 h) and homogenized soil in duplicate [31]. The DNA solutions dissolved in 100 µL Tris-EDTA buffer (10 mmol liter −1 Tris-HCl, 1 mmol liter −1 EDTA, pH 8.0) were diluted 10-folds with RNasefree water and used for real-time PCR assays for confirming the root-lesion and stunt nematode species existing in field samples using the same primer sets described above.

Physico-Chemical Properties of Soil
Soil pH and electric conductivity (EC) were determined in deionized water (H 2 O) and 1 M KCl at a soil-to-solution ratio of 1:5. Soil particle size distribution was determined by the wet-sieving and pipet method [32]. Total carbon (TC) and total nitrogen (TN) were quantified by the dry combustion method with an NC analyzer (SUMIGRAPH NC TR22, Sumika Chemical Analysis Service, Ltd., Osaka, Japan). Available phosphate (Available-P) was evaluated by the Truog method [33]. A batch extraction with ammonium acetate (1 M, Ph 7.0) was performed to measure the cation-exchange capacity (CEC) and exchangeable basic cations (K + , Na + , Ca 2+ , and Mg 2+ ), which were determined using atomic absorption spectroscopy (A-A-640-01; Shimadzu, Kyoto, Japan).

Free-Living Nematode Biodiversity and Maturity Indices
Several important indices were used to assess the biodiversity of free-living nematodes in sugarcane fields, i.e., the Shannon-Wiener index (H'), to evaluate the diversity of soil free-living nematode fauna, the Simpson's D to assess dominance of abundant taxa, species richness (Margalef index), and evenness J' to evaluate using genus and family levels as taxa classification [13]. Maturity index [34], maturity index (colonizer-persister (cp) value 2-5) [35], and maturity index (Cephalobidae adjusted) [29] were also used to gauge the condition of the soil ecosystem.

Statistical Analysis
For nematode abundance, percentages of each free-living nematode species, diversity and maturity indices, and soil physico-chemical properties, multiple comparisons (Tukey-Kramer's method) were conducted among four different sites (Okinawa-North and South, Miyako, and Ishigaki). Stepwise regression analysis was used to narrow down the potential predictors among physico-chemical parameters, nematode biodiversity indices and non-target nematode abundance to explain the abundance of the target nematodes, Pratylenchus and Tylenchorhynchus nematodes. The variance inflation factor (VIF) was used to detect the presence of linear relationships among variables. Then, linear mixed model analysis was performed using the parameters remained (pH (H 2 O), K + , the abundance of the free-living nematodes, and the number of free-living nematode species for Pratylenchus, and pH (H 2 O), the abundance of non-Tylenchorhynchus nematodes, and the clay content for Tylenchorhynchus) as fixed effects with site (Okinawa-South and North, Miyako, and Ishigaki) as a random effect to control for variation in categorical explanatory variables among different locations of soil collection. The statistical analysis was conducted in R v.3.6.1 (R Core Team, Vienna, Austria) [36] using multiple comparisons (multicomp library [37]) and linear mixed-effects models (lmer library of lme4 package [38]) for all variables (square-rooted). Further, in order to understand the spatial pattern of Pratylenchus and Tylenchorhynchus distribution together with these parameters, geospatial mapping was conducted in each island by the inverse distance weighting (IDW) method using gstat package [39,40] in R. Non-metric multidimensional scaling (NMDS) using Bray-Curtis dissimilarities [41] was applied for the nematode community analysis using the nematode classification in the Section 2.3 above. The result of NMDS analysis highlighted the difference among regions and pH levels (high: ≥8.3: the 1st quartile; middle: <8.3: the 2nd and 3rd quartiles; low: <5.4: the 4th quartile of all the 85 data) using vegan package [42] and shown in graph using ggplot2 package [43] in R. Then the physico-chemical variables were fitted to the NMDS as environmental parameters.

Nematodes in Sugarcane Fields
Plant-parasitic nematodes were significantly (p < 0.05) more in Okinawa-North than in Miyako and Ishigaki, and more than 80% of the plant-parasitic nematodes in each region were species of Pratylenchus, Tylenchorhynchus and Helicotylenchus (Table 1). Species identification of individual nematodes using the real-time PCR method showed that all the root-lesion nematodes were either P. zeae or P. coffeae, and all the stunt nematodes were T. leviterminalis or T. annulatus. The abundances of Pratylenchus and Helicotylenchus nematodes were significantly (p < 0.05) higher in Okinawa-North than in the other 3 regions, while Tylenchorhynchus was more in Okinawa-South than Miyako and Ishigaki. Nematodes of the genus Meloidogyne were not populous in all four regions and found in one field each in Okinawa-North and South and Ishigaki, and 7 fields (35%) in Miyako. Rotylenchulus sp. was detected at 28% and 68% of the studied fields in Okinawa-North and South, respectively, and 100% and 94% in Miyako and Ishigaki, respectively. The abundance of Rotylenchulus sp. also tended to be more in Okinawa-South and Miyako than in Okinawa-North and Ishigaki. Xiphinema sp. was found only in 2 fields in Okinawa-South. The abundance of free-living nematodes was significantly (p < 0.05) more in Okinawa-North than in Okinawa-South and Ishigaki (Table 1). Nematodes of the genus Acrobeloides accounted for around 50% of all free-living nematodes in Okinawa-North and South, and around 15% in Miyako and Ishigaki, while other Cephalobidae did for 15% to 19% with no significant difference among the regions ( Table 2). The proportion of Rhabditis sp. was significantly (p < 0.05) more in Okinawa-North and South than in Miyako and Ishigaki, yet other Rhabditidae had no significant difference among the regions. The proportion of other bacterial feeders was significantly (p < 0.05) more in Miyako and Ishigaki than in Okinawa-North and South. The proportion of Filenchus sp. was significantly (p < 0.05) higher in Ishigaki than in the other regions. The proportions of Dorylaimida and predator nematodes were significantly (p < 0.05) more in Miyako and Ishigaki than in Okinawa-North and South (Table 2).

Spatial Distribution of Pratylenchus and Tylenchorhynchus
The root-lesion nematodes (Pratylenchus) were detected in all the sugarcane fields except 3 fields in Okinawa-North and South. The dominant root-lesion nematode species was P. zeae in Okinawa-North and Ishigaki, while it was P. coffeae in Okinawa-South and a mixture of P. zeae and P. coffeae in Miyako. There were more than 80 root-lesion nematodes 20 g soil −1 observed in 28%, 5%, and 6% of the fields in Okinawa-North, Miyako and Ishigaki, respectively, and more than 40 root-lesion nematodes 20 g soil −1 in 72%, 13%, 20%, and 6% of the fields in Okinawa-North, Okinawa-South, Miyako, and Ishigaki, respectively ( Figure 2). The stunt nematodes (Tylenchorhynchus) were detected in all the sugarcane fields except 2 fields in Okinawa and Ishigaki. The dominant stunt nematode species was T. leviterminalis throughout most of the four regions, while T. annulatus were detected in Okinawa-North, in the most of Miyako and occasionally in Ishigaki. There were more than 100 stunt nematodes 20 g soil −1 observed in 54% of the fields in the four regions, and more than 50 stunt nematodes 20 g soil −1 in 74% of the fields (Figure 3).

Soil Physico-Chemical Properties
Values of pH (in both H 2 O and KCl) and CEC were significantly (p < 0.05) lower in Okinawa-North and Miyako than in Okinawa-South and Ishigaki (Table 3). The clay content was significantly (p < 0.05) lower in Okinawa-North and Miyako than in Okinawa-South and Ishigaki, while the sand content was significantly (p < 0.05) higher in Okinawa-North and Miyako than in Okinawa-South and Ishigaki. Soil moisture, exchangeable K + , and Na + were significantly (p < 0.05) higher in Miyako than the other three regions, and total C and total N were higher in Miyako than Okinawa-South and Ishigaki. Exchangeable Ca 2+ was significantly (p < 0.05) higher in Okinawa-South than the other three regions (Table 3).   [21] published by Geospatial Information Authority of Japan [22].   [21] published by Geospatial Information Authority of Japan [22].

The Relationship between Soil Properties and Abundance of Pratylenchus and Tylenchorhynchus
The mixed-effect model among the four regions to test selected variables to explain the abundance of Pratylenchus showed that pH (H 2 O), K + , the abundance of free-living nematodes, and the number of free-living nematode species were significant (p < 0.05). All the parameters except the abundance of free-living nematodes negatively affected the Pratylenchus abundance. The mixed-effect model for Tylenchorhynchus showed that pH, the abundance of non-Tylenchorhynchus nematodes, and clay content were significant (p < 0.05). These parameters except clay content positively affected the Tylenchorhynchus abundance. The inverse distance weighting (IDW) prediction figures confirmed the clearer relationship for the abundance of Pratylenchus and Tylenchorhynchus against pH (H 2 O) in the different directions (Figures 4 and 5). The other parameters, exchangeable cation K + , the abundance of free-living nematodes, and the number of free-living nematode species also showed the relationship with the abundance of Pratylenchus to a certain degree, as did clay content and the abundance of non-Tylenchorhynchus nematodes with that of Tylenchorhynchus.

Diversity of Free-Living Nematodes
The values of Shannon-Weiner index (H ), evenness J , the number of free-living nematode species, and species richness tended to be higher in Miyako and Ishigaki than in Okinawa-North and South, while Simpson's D did in the opposite direction (Table 4). All the three maturity indices showed significantly (p < 0.05) lower values in Okinawa-North and South than in Miyako and Ishigaki (Table 4).

Nematode Fauna in Sugarcane Fields
NMDS generated the 3-dimensional graphs (Figure 6), which was suitable for this analysis with the stress value of 0.157 (the 2-dimensional NMDS showed unclear results with the stress value over 0.2). The result showed clear difference between Okinawa (North and South in the blue oval) and the other islands (Miyako and Ishigaki) but not among the levels of pH ( Figure 6A). The graph also presented the environmental factors, pH, clay, silt, K + , Na + , and Ca 2+ , which showed significant (p < 0.01) difference. The second graph ( Figure 6B) showed differences among Okinawa (both North and South: in the blue oval), Miyako (in the yellow oval), and Ishigaki (in the red oval). The graph included the significant (p < 0.01) environmental factors, pH, clay, silt, sand, K + , Na + , and Ca 2+ . Further, the third graph ( Figure 6C) showed that Miyako and Ishigaki tended to be separated in which Okinawa-North and Okinawa-South were leaning toward Ishigaki and Miyako, respectively. The environmental factors, pH, clay, silt, sand, total C, Ca 2+ , Mg 2+ , and CEC were significant (p < 0.01) in this graph. Among the environmental parameters in all the dimensions together, soil pH and exchangeable basic cations (K + , Na + , and Ca 2+ ) showed significant difference (p < 0.001) as the soil chemical properties, so did (p < 0.01) the clay, silt and sand contents as the soil physical properties. Moisture, total C, total N, Mg 2+ , and CEC were also significantly (p < 0.05) different in the NMDS analysis.

Discussion
Species of Pratylenchus and Tylenchorhynchus were ubiquitously detected in all the four regions in Okinawa. The result was consistent with the previous study [25], which reported that Pratylenchus and Tylenchorhynchus were found in all 15 studied fields in Kitadaito island, another major sugarcane producing region in Okinawa. The survey in the current study found that Pratylenchus nematodes were more than threshold or half of that in most of the fields in Okinawa-North, so were Tylenchorhynchus in Okinawa-North and South. The abundance of Pratylenchus sp. was influenced by pH, K + , the abundance of free-living nematodes, and the number of free-living nematodes species. Pratylenchus sp. abundance was negatively correlated with the pH among these parameters, suggesting that Pratylenchus may be more populous in lower pH areas (e.g., Okinawa-North) than in higher pH areas (e.g., Okinawa-South). The result was consistent with the previous study [44] reporting inverse relationship between the abundance of Pratylenchus and pH. With regard to K + , the result was consistent with the previous study [45], which reported a positive correlation of K + and plant-parasitic nematodes. Also diverse free-living nematodes can suppress plant-parasitic nematodes possibly through the increase of predator nematodes shown in Miyako and Ishigaki. Yet, another parameter, the abundance of free-living nematodes, was positively correlated with the ‡ †

Discussion
Species of Pratylenchus and Tylenchorhynchus were ubiquitously detected in all the four regions in Okinawa. The result was consistent with the previous study [25], which reported that Pratylenchus and Tylenchorhynchus were found in all 15 studied fields in Kitadaito island, another major sugarcane producing region in Okinawa. The survey in the current study found that Pratylenchus nematodes were more than threshold or half of that in most of the fields in Okinawa-North, so were Tylenchorhynchus in Okinawa-North and South. The abundance of Pratylenchus sp. was influenced by pH, K + , the abundance of free-living nematodes, and the number of free-living nematodes species. Pratylenchus sp. abundance was negatively correlated with the pH among these parameters, suggesting that Pratylenchus may be more populous in lower pH areas (e.g., Okinawa-North) than in higher pH areas (e.g., Okinawa-South). The result was consistent with the previous study [44] reporting inverse relationship between the abundance of Pratylenchus and pH. With regard to K + , the result was consistent with the previous study [45], which reported a positive correlation of K + and plant-parasitic nematodes. Also diverse free-living nematodes can suppress plant-parasitic nematodes possibly through the increase of predator nematodes shown in Miyako and Ishigaki. Yet, another parameter, the abundance of free-living nematodes, was positively correlated with the abundance of Pratylenchus, though a negative relationship between total free-living nematodes and P. crenatus was reported [16]. The exact reason on the different results was not known.
The abundance of Tylenchorhynchus was affected by pH, clay contents and the abundance of non-Tylenchorhynchus nematodes. Unlike Pratylenchus, the abundance of Tylenchorhynchus was positively correlated with pH. This relationship was consistent with the previous study [46] reporting that Tylenchorhynchus maximus increased with positive relationship with pH. The current study showed inverse relationship between the abundance of Tylenchorhynchus and the soil clay content. On this point, the previous study [47] reported differently on Tylenchorhynchus clarus in soils with 4 to 32% clay, possibly due to the excessive clay contents in the fields of the current study and the difference of species, yet the exact reason was unknown. The nematode abundance other than Tylenchorhynchus had a positive relationship with the abundance of Tylenchorhynchus. This can indicate that the more nematodes were found, the more Tylenchorhynchus. CEC did not show significant influence on Tylenchorhynchus sp. (and Pratylenchus described above) in this study. Norton et al. [48] also reported that none of the plant-parasitic nematodes were significantly correlated with CEC.
Species of Helicotylenchus were the third most populous plant-parasitic nematodes found in this study. Though Berry et al. [49] discussed that H. dihystera should provide sustainable means of reducing the effect of more pathogenic species, no clear relationship was found between H. dihystera and the target plant-parasitic nematodes (Pratylenchus and Tylenchorhynchus) in this study. Meloidogyne species were rarely found at the studied sites and this was possibly due to the clayey soil texture, which may not be suitable for Meloidogyne inhabitation [1]. The previous study [50] reported R. reniformis was found from more than 50% of fields (various crops) in Okinawa, though the current study showed Rotylenchulus detection of over 90% of the fields in Miyako and in Ishigaki, respectively, and around 30% and 70% in Okinawa-North and South, respectively. The result suggested that difference in crops and studied fields might cause the inconsistent results.
This study also revealed that the species of Pratylenchus and Tylenchorhynchus were different among the regions. The combination in Okinawa-North was primarily P. zeae and both T. leviterminalis and T. annulatus, while that in Okinawa-South was P. coffeae and T. leviterminalis. In Miyako, the combination was the mixture of P. zeae and P. coffeae, and that of T. leviterminalis and T. annulatus. Cadet and Spaull [1] reported that the abundance of P. zeae was greater in clay soils with some exception in Australia and South Africa. Pratylenchus zeae was most numerous in Okinawa-North, however, the soil in this region was not necessarily most clayey compared to the other regions. Though Miyako's soil was most clayey among the regions, the abundance of Pratylenchus was less than in Okinawa-North (if we simply compared these two regions without considering other differences) and the genus was a combination of P. zeae and P. coffeae. The data reported in the other study [3] for Australian sugarcane fields showed some tendency of inverse relationship between the ratio of clay in soil and the abundance of P. zeae. This may suggest that the clay contents may influence Pratylenchus species suitable for particular sugarcane field with appropriate level of clay contents for the specific species. The level of pH may also influence the species of Pratylenchus as the previous study [51] reported that raising the soil pH from 5.0 to 6.9 in which alfalfa was grown increased the abundance of P. penetrans and greatly reduced the abundance of P. crenatus. The same study commented that temperature also affected the reproduction rates of these species. Since both P. zeae and P. coffeae were increased in the 25-30 • C range [52], in which the studied fields in the current study were, the difference of pH can influence more on the reproduction of both species than temperature. Stirling et al. [53] discussed that total C and total N from crop residues and roots may increase the nematode trapping fungi, which may have been one of the suppressive factors of plant-parasitic nematodes in sugarcane soils. Our study, however, did not clearly show the relationship of total C and total N to the abundance of the root-lesion and stunt nematodes. Since the current study did not cover the nematode trapping fungi, the exact reason was unclear on this point.
The free-living nematodes tended to be separated between the group of Okinawa (North and South) and that of Miyako and Ishigaki. This tendency led to the similar combination in diversity and maturity indices. The result suggested that the free-living nematode faunas were similar within each group. Though the exact mechanism was not understood, geographical and climatic differences may play the major role in this grouping. Okinawa-North and South are in the same island, while Miyako and Ishigaki are the islands next to each other (just 50 km away). Therefore, the temperature and precipitation are very similar within each group. Overall, the trends of the three maturity indices were consistent among the regions. The result was not in line with the previous study [29] conducted in northern climate zones in Japan. Rhabditis and other Rhabditidae affected the value of MI (cp2-5) and Cephalobidae does that of MI (Cephalobidae adjusted). Since the compositions of these nematodes to the total free-living nematodes were consistent across the regions, the adjustments made no apparent difference on the indices of each region. The maturity index values of Okinawa-North and South were lower than those of Miyako and Ishigaki primarily because of the different composition of the high c-p value Dorylaimida to the total free-living nematodes.
Further, many of the soil physico-chemical properties (i.e., pH, clay and sand contents, Ca 2+ , Mg 2+ , and CEC) tended to be separated between the group of Okinawa-North and Ishigaki and that of Okinawa-South and Miyako, while the other parameters did not. Such difference may be due to the different soil materials, causing the different soil types of those two groups. Though no detailed mechanism was identified, such geographical separation of the physico-chemical properties may result in the difference in dominant Pratylenchus species in each group. Such separation coupled with that of the free-living nematodes mentioned above may be involved in the clear differences of nematode communities (1) between the group of Okinawa-North and South and that of Miyako and Ishigaki; (2) among Okinawa-North and South, Miyako, and Ishigaki; and (3) between the group of Okinawa-South and Miyako and that of Okinawa-North and Ishigaki, shown in the NMDS analysis. This analysis reinforced that pH and soil texture may impact not only the plant-parasitic but also the free-living nematodes. The result was basically consistent with the previous study [15], which reported that the nematode community may be a well reflected indicator of the soil properties such as soil moisture, pH, and organic soil layer. Yet, further study is needed for more comprehensive understanding of below-ground microbial ecosystem.

Conclusions
This study revealed both Pratylenchus and Tylenchorhynchus species were widely distributed in sugarcane fields in Okinawa, Japan. Pratylenchus sp. was highly abundant and may be a yield limiting factor especially in Okinawa-North, so was Tylenchorhynchus sp. in all the regions examined. It was confirmed that the major species of the root-lesion nematodes were P. zeae and P. coffeae and that those of the stunt nematodes were T. leviterminalis and T. annulatus. It was also found that soil pH influenced the abundance of both Pratylenchus and Tylenchorhynchus nematodes in different directions. Then, the abundance of free-living nematodes, the number of free-living nematode species, and exchangeable cation K + were shown to be important parameters to the abundance of Pratylenchus species, as were the clay content and non-Tylenchorhynchus nematode abundance to that of Tylenchorhynchus. Further, this study revealed that the three islands (the four regions) had different characteristics of the nematode fauna, and the difference was explained especially by pH, physical properties (soil texture) and exchangeable basic cations. This study provided basic data for proper management of the plant-parasitic nematodes and free-living nematodes in sugarcane fields in Okinawa to improve their productivity.