*Article* **A Comparative Study of Field Nematode Communities over a Decade of Cotton Production in Australia**

#### **Oliver Knox 1,\*, David Backhouse <sup>1</sup> and Vadakattu Gupta <sup>2</sup>**


Received: 23 December 2019; Accepted: 13 January 2020; Published: 15 January 2020

**Abstract:** Soil nematode populations have the potential to indicate ecosystem disturbances. In response to questions about nematode interactions with soilborne diseases and whether genetically modified cotton altered nematode populations, several fields in the Namoi cotton growing area of Australia were sampled between 2005 and 2007. No significant interactions were observed, but nematodes numbers were low and postulated to be due to the use of the nematicide aldicarb. Aldicarb was removed from the system in 2011 and in 2015 funding allowed some fields to be resampled to determine if there had been a change in the nematode numbers following aldicarb removal. No significant changes in the total nematode numbers were observed, implying that the removal of aldicarb had little impact on the total nematode population size. However, an increase in plant parasitic nematodes was observed in both fields, but the species identified and the levels of change were not considered a threat to cotton production nor driven solely by altered pesticide chemistry. Additionally, greater numbers of higher order coloniser-persisters in the 2015 samples suggests that the current cotton production system is less disruptive to the soil ecosystem than that of a decade ago.

**Keywords:** axonchium; helicotylenchus; tylenchorhynchus; pratylenchus; reniform; vertosol; gossypium

#### **1. Introduction**

The use of pesticides often courts controversy and remains an issue that often results in political intervention [1,2]. Changes in the regulatory processes of both the EU and the United States EPA brought about a decision from Bayer to halt production of aldicarb, a nematicide developed in the 1970s, by 2014 and for complete removal of the product by 2018 [3]. Aldicarb was utilised on a range of crops, but primarily in Australia in cotton, sugar cane and citrus [4,5].

Australian cotton systems have historically been without the nematode related production issues experienced by other cotton producing nations [6], although the presence of the reniform nematode, *Rotylenchus reniformus* [7], in the Theodore production area of Queensland highlights that this status can change. As a consequence of this, aldicarb was not registered for nematode control, but for early season control of aphids, mirids, jassids, mites, wireworms and thrips that aldicarb's systemic activity offered whilst retaining beneficial populations [8]. Control of these early season pests following the removal of aldicarb from Australia in 2011 has been provided either through the optional use of neonicotinoids, in the form of Cruiser® (active ingredient (a.i.) thiomethoxam, Syngenta) [9], or through the continued or adopted use of the organophosphates and carbamates, such as phorate and carbosulfan, respectively. The impact of neonicotinoids on entomopathogenic nematodes has

been reported to have limited impact on reproduction [10,11], which might imply limited effects on other free living soil nematodes [10,11]. The organophosphate and carbamates are known to have nematicidal activity particularly against reniform, lesion and root-knot nematodes [12,13], but existing work has been on sandy soils, not in clay vertosols. Additionally, impacts beyond the targeted pest nematode population have either not been undertaken [12,14] or found no difference [13].

Adoption of the synthetic pyrethroids to control of wireworm and mirids offers protection to above and below ground herbivorous damage, however, their impact on nematodes is negligible [5,15]. This assumption is based on the facts that no deleterious effects from synthetic pyrethroids have been found on entomophathogenic nematodes [16–18]. However, when pyrethroids were introduced to aquatic systems nematodes flourished [19], although *Daptonema trabeculosum* was found to be sensitive to permethrin [15].

In the USA, aldicarb has been replaced in the cotton production system with either Avicta® seed treatments (a.i. abamectin, thiamethoxam, mefenoxam and fludioxanil, Syngenta) in possible conjunction with Velum® (a.i. fluopyram and imadicloprid, Bayer CropScience) or the use of Vydate® (a.i. anticholinesterase, DuPont). At present, these products are not licensed for Australian cotton where rotations and management conditions to promote rapid cotton establishment are the predominant forms of nematode control [20,21].

In our initial nematode work in the Namoi in between 2005 and 2007, the low numbers of recovered nematodes (<5 nematodes/g soil) were hypothesised as being due to the systemic use of aldicarb [22,23]. This assumption was based on the impact aldicarb has on free living nematodes in culture and under carrots [11,24]. However, despite being initially developed as a nematicide, aldicarb has been rarely studied, in relation to free living nematodes [11], does not affect free living nematodes under potato [25] and we could find no published evidence of its impact under cotton rotations. With changes in funding, movement of staff and the removal of aldicarb in 2011, we were unable to test our hypothesis directly, instead resampling fields in in the upper and lower Namoi valley in 2015, which were originally sampled in 2005 and 2007 and for which nematode community analysis had been undertaken [22,26]. The nematode communities were assessed and compared between the sampling years to determine if the nematode numbers had increased with the removal of aldicarb and if there had been changes in the nematode population structure. The results are discussed within the context of the potential for effects on the Australian cotton production system and the ecological significance of the observations.

#### **2. Materials and Methods**

#### *2.1. Soil Characteristics and Nematode Sampling*

Field A: In July 2005 and June of 2007, a field in the lower Namoi (field A) was sampled as part of investigations into non-target effects of genetically modified (GM) cotton on soil microbiology. The field soil is a grey vertosol, 52% clay, pH 8.2 and 200 m above sea level. The mean annual maximal temperatures is 26 ◦C and minimum 12 ◦C and the area receives 660 mm of summer dominant rainfall. In the field, samples were collected from under each variety being cultivated, resulting in 16 samples in 2005 and 12 in 2007, with sites evenly spaced along 180 m of the plant line. Approximately one kilogram of topsoil was taken to a depth of 15 cm at each site from under mature cotton. In March, 2015, this field was resampled when it was again under cotton, using field maps of the 2007 trial to return to approximately the same location except that only six samples were taken from the plant line at equidistant points from the tail to head ditch with the field having been planted under only one variety. This field had been in a cotton–wheat rotation, with cotton planted in October of every even year. Aldicarb had been applied at cotton sowing at a standard rate of 7 kg Temik®/ha (1.05 kg a.i.) for thrips control with the final application made in October of 2010. In 2012 and 2014, phorate was applied with cotton sowing as 6 kg Thimet®/ha (600 g/ha a.i.). Neither chemical was used in the wheat phase of the rotation.

Field B: In late October of 2005, soil was sampled from a field in the upper Namoi (field B) as part of an investigation into nematode interactions with verticillium wilt. This field is a black vertosol, 65% clay, pH 8.5 and 270 m above sea level. Mean maximum and minimum temperatures are 12 and 27 ◦C, respectively, with the area receiving roughly 640 mm of summer dominant rain. One kilogram of surface soil to a depth of 15 cm was recovered from the plant line of cotton seedlings. Briefly, sample points were established from both the Northern and North-Western corners of the field by walking a 20 m by 10 row transect into the crop and taking a sample. The transect walk was then repeated until six samples had been gathered from each entry point. In March, when the field was under mature cotton and again in June of 2015 after picking and root cutting, we collected samples close to the original sampling points, based on field notes and discussions with the farmer. This field had predominantly been under a cotton–cotton–wheat rotation since 1988, although sorghum had been introduced in place of wheat in 2009, 2013 and 2014. Aldicarb had been applied as Temik® at 7 kg/ha in every year that cotton was sown, resulting in aldicarb application in 13 out of 28 years, with the last application in 2011.

Cultivations varied between fields due to differences in the rotations, but both had been subjected to pupae busting, a minimal cultivation to a depth of 10 cm at least 30 cm either side of the plant line, post cotton crop harvesting and had been subjected to bed reformation in the spring prior to cotton planting.

#### *2.2. Soil Analysis*

In all cases, field sampled soil was placed in plastic bags and returned in a chilled ice box to the laboratory. In the laboratory, the samples were sieved through a 2 mm sieve and a 300 g subsample was sent within 48 h of samples being taken in the field to Biological Crop Protection (Moggill, Queensland, Australia) for nematode community analysis. Briefly, the soil moisture content was determined gravimetrically and 200 mL of soil was weighed and used to establish Whitehead trays for nematode extraction. Nematodes were subsequently recovered from the water solution within the trays and assessed to determine nematode abundance. A sample of approximately 120 nematodes from the count were identified to genus and, in the case of the plant parasitic nematodes, to species where possible to facilitate community compositional analysis [27]. Recovered nematode data were analysed both as recovered numbers and as the number of nematodes present per gram of dry weight equivalent of soil to mitigate moisture and soil porosity differences.

#### *2.3. Root Tissue Analysis*

Roots were collected from all samples during the sieving process and the root tissue was cleared using the NaOCl and acid fuchsin method of Byrd et al. [26,28]. Roots were spread over a 1 cm gridded Petri dish and examined under a stereo microscope (20 to 45 x magnification) for the presence of nematodes.

#### *2.4. Community Comparisons and Statistical Analysis*

The nematode community data from the 2005, 2007 and 2015 field samples were tabulated. Comparative analyses for the free living nematodes and between the plant parasitic nematode types were conducted on either raw or percentage compositional data, respectively, with multiple Wilcoxon rank-sum tests between all possible pairwise comparisons. Significance in differences of the median values was taken at the level of *p* < 0.05/x, where x represented the number of groups within any series of pairwise comparisons. This decision was based on the existence of small sample sets for each field and a lack of normality of the data. The nematode channel ratio (NCR) [29] was calculated from the bacterial and fungal trophic group composition of the samples. Additional community composition and change was assessed using the Nematode INdicator Joint Analysis (NINJA) web based program [30] with probability of similarity of mean outcomes assessed with ANOVA, with significance

taken at *p* < 0.05. This on-line tool was also used to generate maturity index (MI), Plant Parasitic Index (PPI), enrichment (EI) and structural indexes (SI) for the samples [31,32].

#### **3. Results**

#### *3.1. Soil Sample and Total Nematode Comparisons*

The 200 mL soil samples had an averaged dry weight equivalent of 126.5 g (stdev = 4.5, *n* = 30) for field A and 134.5 g (stdev = 8.8, *n* = 18) for field B over the period of assessment with no apparent statistical difference between weights with sampling time or field, however, moisture content varied between 24% and 35%. The total number of nematodes recovered per 200 mL of soil ranged from 267 to 2944, with an average of 1194, mode of 371 and standard deviation of 609 and standard error of 85. Analysis of the total recovered nematodes did not indicate any significant difference in nematodes/g assessed either within fields, between years or in combination (Table 1), but were detected for many nematode ecological indexes and footprints (Table 2), primarily due to changes in the nematode population structure recorded in 2015 in field B.

**Table 1.** Mean nematode counts of total free living nematodes, per g dry weight equivalent of soil and the percentage of plant parasitic from 200 mL soil Whitehead tray recoveries of samples collected in cotton fields A and B in the Namoi valley. The percentage contributions of the stunt (*Merlinius* and *Tylenchorhynchus* spp.), lesion (*Pratylenchus* sp.) spiral (*Helicotylenchus* sp.) and dagger nematodes to the plant parasitic nematodes within samples and years are given. Similarities in the plant parasitic population are assessed with Wilcoxon rank-sum tests and significantly similar medians are indicated with the same upper case letter.


\* statistical analysis in cases where the nematode was previously not detected assumes a 0 value in the samples of those years. No detection within the samples is indicated by 'nd' and 'ns' indicates no significant difference.

**Table 2.** Summary mean, standard deviations (SD) and corresponding ANOVA *p* values from the Nematode INdicator Joint Analysis (NINJA) of the field analysed samples from 2005, 2007 and 2015 in field A and 2005 and 2015 in field B.



**Table 2.** *Cont.*

#### *3.2. Plant Parasitic Nematode Populations*

The percentage of the nematode population representing plant parasitic nematodes had not changed in field A and was reflected in the PPI scores for the field, which averaged 2.38, 2.56 and 2.09 for 2005, 2007 and 2015, respectively. However, the PPI had significantly (*p* < 0.001) increased in field B from 2.29 in 2005 to 3.18 in 2015. Additionally, the composition of plant parasitic nematodes, in terms of the abundance of specific parasitic genera, revealed changes in both fields. For example, in the field B there was and remained no evidence of lesion nematodes (*Pratylenchus* sp.), but a significant decrease in stunt (*Merlinius* and *Tylenchorhynchus* spp.) and an increase in spiral (*Helicotylenchus* sp.) nematodes was observed. In field A, spiral nematodes were not observed in 2005 and 2007 samples, but were found in the 2015 samples at >0.2% of the total nematode population. Stunt nematodes were significantly (*p* < 0.001) higher in both fields in 2005 than in other sampling years, whilst the proportion of lesion nematodes increased with time in field A (Table 1). Data on the abundances of the ectoparasites, semi-endoparasites and migratory endoparasites as their % composition of the herbivore assemblage implied that within field A the migratory endoparasites increased as the ectoparasites were reduced, whilst in field B the semi-endoparasties appeared to have replaced the migratory endoparasites (Figure 1).

**Figure 1.** The percentage of the migratory endoparasites (e.g., *Pratylenchus*, white), ectoparasitic (e.g., *Tylenchorhynchus*, grey), semi-endoparasitic (e.g., *Helicotylenchus*, black) and ectoparasitic (e.g., *Xiphenema*, dashed) feeding types of the herbivorous nematodes assemblage identified from 200 mL soil samples from field A in 2005, 2007 and 2015 and field B in 2005 and 2015. Number of samples and time of year differed between years with error bars representing the standard error of the means.

#### *3.3. Nematode Community Assemblages*

Community analysis with NINJA indicated that there was significant (*p* < 0.05, ANOVA) difference in the maturity, plant parasitic, enrichment and structural indexes and the herbivore, fungivore, bacterivore and omnivore footprints within the assessed field material (Table 2). The changes in the assessed community reflected these differences in terms of shifts in the relative proportions of omnivore, predatory, bacterivores, fungivores and herbivorous nematodes present (Figure 2) as well as in changes to the composition of the herbivorous nematode assemblage (Figure 1). Whilst changes in the structural and enrichment status of the samples were both significant (Table 2), graphical representation of the data (Figure 3) supported an improvement in maturity of the analysed ecosystem rather than nutrient enrichment, due to an increase in the number of higher order coloniser-persisters in the samples. This was particularly evident for field B between 2005 and 2015 (Figure 3). NCR analysis indicated similar scores between fields, but that the 2005 samples had a lower ratio than the populations of subsequent samples in both fields (Table 2).

**Figure 2.** Average percentage of the total recovered omnivorous (white), predatory (black), bacterivorous (light grey), fungivorous (stripped) and herbivorous (dark grey) nematode feeding types as identified from the evaluation of ~120 nematodes from each sample (*n* ≥ 6) from field A and field B over each year of sampling. Error bars represent the standard errors of the means.

**Figure 3.** Food web analysis of nematode community assemblages from field A, sampled in 2005 (-), 2007 ( ) and 2015 (-) and field B sampled in 2005 ( ) and 2015 ( ). The Enrichment index parallels with the nutrient enrichment whilst the structural index correlates with the maturity of the ecosystem.

#### *3.4. Root Tissue Observations*

No nematodes were observed within the cleared and stained root tissue from the 2015 samples, which was in contrast to the observation and recovery of *H. dihystera* from roots in field B and of Rhabditea and Aphelenchidea from roots in field A in 2005.

#### **4. Discussion**

In general, abundances of total nematodes in soil supporting Australian cotton systems, as observed in 2005 and 2007 [26,33], are considered low [34]. In addition to this, Australian cotton production systems have not reported nematode issues, with the exception of the recent and localized occurrence of the reniform nematode [7], and this was partly attributed to the widespread use of aldicarb in cotton [22,23]. Aldicarb has a highly variable half-life in soil that ranges from a few to 408 days, with more rapid detoxification occurring in anaerobic soils [4,5]. In Australian cotton soils, the half-life is thought to be about a week in surface soils, due to high soil temperatures and the repeated fluctuation between aerobic and anaerobic soil conditions from flood irrigations [4]. With aldicarb absent from these cotton fields for several years, residual compound and active metabolites from historic applications should have fallen below effective levels [35]. In an attempt to discern if this hypothesis was correct, two fields, roughly 160 km apart, in the Namoi valley, which had nematode community data from 2005 and 2007, were reassessed in 2015.

Although observations from the 2015 sampling indicated that significant changes in the composition of nematode communities were occurring (Figure 3), the total numbers of nematodes supported within the vertosols had not changed (Table 1). This was taken as indication that aldicarb had not imposed a limitation on the population size as initially hypothesised, which is in keeping with other work where pesticide changes had not altered nematode population size, but had been associated with a change in species richness [36,37]. Whilst the implications of other variations in the assessed fields' management systems, such as differences in clay content, irrigation strategies, rotational histories and periods of fallow, could not be investigated from the field records available, it was noted that between the two fields the frequency of fallows occurring post wheat and prior to the return to cotton in the rotation varied [38]. Periods of long fallow of over 7 months in Australian grains production systems, which can incorporate cotton, have been previously reported as causing a reduction in the free living nematode population and altering the nematode channel ratio [29] in favour of a fungal dominated decompositional community [39]. However, the populations analysed in these fields indicated a move to more bacterially dominated decompositional communities over time (Table 2).

Whilst we saw little change in the total free living nematode population across our samples, changes in the nematode community composition were noted in the herbivorous assembly in field B whilst herbivorous nematodes remained unchanged in field A. In a study in Slovakian, maize fields increasing insecticidal chemistry to five times the recommended dose did not significantly alter the nematode communities, but season of assessment did [37]; however, this trial did not interrogate other management decisions. When comparing results from these cotton fields to existing studies [37,39,40], it becomes apparent that there is a requirement for further systematic interrogation of the production systems in order to identify the drivers of nematode community change.

Changes in cotton production practices are also possible causes of the observed differences in the nematode communities within these fields over the last decade [41]. Since 2005, both farms have experienced drought that has seen both differences in the amount and quality of the water used for cotton irrigation in different years, which could have influenced nematode communities [42]. There has also been a change in the preferred cultivar material from cultivars based on the Sicot 189 family in 2005/7 to those of Sicot 74 and 75 in 2015 along with changes in pesticide use and nutrition management [41], which has included the loss of aldicarb from the Australian cotton production system. Additionally, sampling was not possible around the time of aldicarb removal from the system due to funding, staffing movements and that sampling across the two fields occurred at different times

within the cotton phase of the rotation, due to weather constraints that were unavoidable. These issues further highlight that gaps exist in our knowledge of nematodes within Australian vertosols over temporal periods.

Knowledge about the long-term changes in nematode communities due to changing crop management practices would help in the development of options to avoid unexpected threats in addition to providing insights into the ecology of soil fauna in production systems with multiple crop, chemical and physical factors potentially influencing abundance and composition [43]. So whilst the main drivers of nematode community change remain elusive, the nature of the differences between fields and study periods highlighted the continued need for vigilance and the imposition of the 'come clean, go clean' farm hygiene strategy, as currently promoted throughout the Australian cotton industry. This strategy is required to continue to limit the spread of potential problem nematodes, such as the reniform nematode, which is causing cotton production issues in Theodore [7], but remains undetected in New South Wales (NSW). However, the presence of *H. dihystera* within field A and *Xiphenema* sp. in field B in the 2015 samples was noted as neither had been previously detected there. Whilst it is possible that these nematodes were not previously observed due to scarcity, the possibility that they were introduced through soil movement on contaminated machinery over the intervening decade remains plausible.

Changes in other members of the herbivorous nematode population were also noted. *T. ewingi*, was still isolated from both fields, but in field B *T. ewingi* was significantly reduced as a percentage of the plant parasitic population due to an increase in soil recovery of *H. dihystera* (Table 1). This change was hypothesised as being due to rotational differences, which included the incorporation of sorghum into the rotation of field B. This hypothesis was based on both *Tylenchorhynchus* and *Helicotylenchus* spp. being known to survive on wheat [43] and having both been recorded on wheat and sorghum in Australia [44]. Additionally, in a >20 year experiment involving continuous sorghum there was little impact on *Tylenchorhynchus* spp., but incorporation of sorghum straw resulted in a significant increase in the number of *Helicotylenchus* spp. recovered [45], which mirrored the observed change in field B.

The isolation of *H. dihystera* within field B was also noted to have changed over the decade. *H. dihystera* was first observed in Australian cotton roots collected from field B [26], but was absent from the soil samples in 2005. However, these observations were reversed in 2015 with *H. dihystera* only observed in soil. This observation could possibly be linked to the difference in the time of sampling [37] and a reduction in the number of samples, but might also be a function of the maturity of the cotton roots. More likely though is *H. dihystera* ability to feed on sorghum as either an endo or ectoparasite [45,46] and that sorghum was planted into the field B rotation in three of the previous five years to the 2015 sampling.

The other plant parasitic nematode shift considered to be of note was that of the lesion nematode, mostly *P. thornei*, which remained absent in field B, but had significantly increased in numbers in field A. Although still not considered an issue for cotton production in Australia, establishment of a population of around the levels found in 2015 without appropriate management could become an issue for grain crops grown in rotation with cotton [47,48].

Out with the changes in the plant parasitic populations, there was an increase in general maturity index of the community in the 2015 soil samples, suggesting an increase in the abundance of higher order coloniser-persister (C-P) nematodes. This change was particularly evident with the increase in the numbers of *Axonchium* sp., although it was echoed to a lesser extent in other nematodes with C-P scores of >3 [31]. The *Axonchium* nematodes increased from 0.14% to 0.28% of the population in field A, but in field B they increased from 0.42% to 17% of the total population and in some samples represented 40% of the total free living nematode population. Members of the genus *Axonchium* pose something of enigma, because the lack of a clearly identifiable mouth part makes them hard to assign to a specific trophic group. This has seen *Axonchium* associated with either bacterial, root hair and therefore plant parasitic or predatory feeding patterns [49,50]. Given the increase in these nematodes in field B, it would be prudent to establish the exact feeding strategy of these nematodes, as changes in

assignation of feeding strategy to a fungivore or omnivore, rather than an herbivorous ectoparasite, increased the maturity and structural index, whilst reducing the plant parasitic index for field B. However, altering the assigned feeding type for *Axonchium* had little to no impact on either the channel or enrichment index and no effect on field A analysis, where they were less abundant in the samples.

From a production stand point, the apparent rise in plant associated and parasitic nematodes could be seen as grounds for concern, especially in the absence of any chemical or cultivar control options, but at the same time the increase in the maturity index of the populations (Figure 3), partly though changes in predatory nematodes, could be indication of more persistent and stable populations that might self-regulate any potential production threat [37,42]. Although most of the samples still exemplify a state of degradation, based on the quadrat in which they occur [32], there does appear to be a trend toward a trajectory in both enrichment and structural indexes (Figure 3). This observation implies that between 2005 and 2015 the examined cotton production systems are moving toward more opportunistic bacterial feeding strategies, based on the enrichment index, whilst the improvement in the structural index implies a less disturbed soil food web and improved trophic interactions [51]. However, nematodes of the higher order trophic groups, which drive these developments, are known to be easily disrupted by soil cultivation [51], making this a potentially unreliable control mechanism under existing cotton production strategies that still involve some form of tillage.

In general, these observations indicate a continuing change in the nematode populations in the Australian cotton fields sampled, probably due to changes in soil management, rotational variation and seasonal environmental conditions [37,41,45], whereas the impact from pesticides is perhaps not as important as originally hypothesised [22]. However, the scale of the current assessment highlights a need for more intensive sampling and for an improved understanding of the genera present. Whilst changes in the herbivorous nematode populations in these NSW fields implies limited current threat to cotton production in these areas, the risk of movement of the reniform nematode from Queensland and the absence of available nematicidal chemistry would caution that continued monitoring and vigilance is warranted.

**Author Contributions:** O.K. undertook the field sampling, analysis and manuscript preparation. D.B. provided project delivery assistance, technical and editorial support. V.G. assisted with the 2005 and 2007 sample analysis, strategy for the 2015 analysis and manuscript editorial support. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was undertaken as part of the activities of the Cotton Hub at UNE with funding provided by the University of New England and the Cotton Research and Development Corporation under UNE1403 and UNE2001. The Initial surveys were conducted with funding from the Cotton Catchment and Communities CRC and CRDC with assistance from staff at CSIRO and NSW DPI. Nematode extraction and analysis was conducted by Biological Crop Protection, Moggill, Queensland for all samples other than root material.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Communication* **Distribution and Restricted Vertical Movement of Nematodes in a Heavy Clay Soil**

#### **Oliver Knox 1,\*, Katherine Polain 1, Elijha Fortescue <sup>1</sup> and Bryan Gri**ffi**ths <sup>2</sup>**


Received: 24 December 2019; Accepted: 1 February 2020; Published: 4 February 2020

**Abstract:** A large part of Australia's broad acre irrigation industry, which includes cotton, is farmed on heavy clay Vertosols. Recent changes in nematicide chemical availability, changes in rotations and the observation of the reniform nematode in central Queensland has highlighted that we need to improve our understanding of nematodes in these soils. We undertook preliminary investigations into distribution by depth under a cotton-cotton and cotton-maize rotation as well as vertical movement experiments in microcosms to better understand nematode distribution and movement in heavy clay soils. Analysis revealed that field populations decreased with soil sample depth, but there were also differences between rotations. In microcosm experiments, vertical movement of nematodes in these heavy clay soils was restricted, even in the presence of plant roots and moisture, both of which were hypothesised to improve nematode migration. The results imply that crop rotation currently remains a plausible option for nematode control, and that we still have a lot to learn about the ecology of nematode populations in Vertosols.

**Keywords:** *Gossypium*; *Zea mays*; vertisol; reniform

#### **1. Introduction**

In 2007, several experiments were undertaken within the Namoi valley cotton production area of New South Wales (NSW), Australia. These experiments were looking for interactions between genetically modified cotton and the soil biota [1], as well as the potential for an interaction between nematodes and the verticillium wilt [2,3], which is a production issue in the valley. At that time, there was no known nematode issue affecting Australian cotton production, although some potentially pathogenic nematodes were isolated [4,5], but these were in low numbers and possibly controlled by flood irrigation and the use of aldicarb [6].

Changes in funding and relocation of staff meant that continued monitoring was not possible; however, in 2014, a reversal in circumstance meant sampling, albeit to a limited extent, was recommenced. During the break in monitoring several changes occurred in the production system [7], with the removal of aldicarb and a shift to rotations that included maize being of note [8,9]. Additionally, *Rotylenhus reniformis* had been associated with yield losses around the Theodore area of central Queensland [10], which acted as a reminder of the importance of the Australian cotton industries 'come clean, go clean' policies [11]. The impact of reniform in Theodore also highlighted an industry requirement for more information on our nematode populations if we were to attempt to avoid the issues that were experienced in the USA. In the USA, reniform spread across almost half of the cotton fields of Alabama, Louisiana, and Mississippi in 50 years, reducing the yields by up to 20% [12,13].

We asked two questions to address some of the current unknowns, with regard to the Australian cotton production system. One was whether the inclusion of maize into the cotton rotation could affect the distribution of nematodes in the soil profile? The second was, do nematodes have the potential to move up a soil profile under favourable conditions? We undertook a combination of field core assessments and glasshouse based recolonization studies to address these questions. The results of these experiments are presented and discussed.

#### **2. Materials and Methods**

#### *2.1. Soil Sites and Characteristics*

Vertical distribution of nematodes, with regard to rotation, was recovered from soils taken from field C1 at the Australian Cotton Research Institute (ACRI), Narrabri, NSW. The soil is an alkaline dark grey clay Vertosol (approximately 66% clay) with a known decreases in soil carbon down the profiles [14]. The rotation on the site has previously been explained in detail [15], and cores were taken to a depth of 1 m in January of 2017 with a portable coring rig [16] from within the cotton-cotton and cotton-maize rotations when both rotations were planted to cotton. Cores were returned to the University of New England (UNE), where they were divided into 0–15, 15–30, 30–50, 50–70, and 70–100 cm depths and nematodes were extracted using a passive recovery technique [17] prior to enumeration. Other field parameters, such as cropping history and planting dates, were gathered from field records at the time of sampling.

The soils gravimetric water content (GWC) was assessed by comparing the weight of a field fresh sample with the resultant weight after drying to a constant mass at 105 ◦C. The dry weight bulk density was calculated from the mass of the soils that were recovered from the core while assuming no compaction during sampling.

Two soils were used in the vertical movement experiments. The first, designated 'Kirby', was collected from UNE's Kirby farm and it was a sandy loam (grey Chromosol [18]); 73% sand, 12% silt, and 14% clay with a pHH2O (1 to 5 in water) of 5.4. The second soil, 'Cotton', was collected from a cotton property near Moree, NSW and it was a clay soil (black Vertosol [18]); 9% sand, 16% silt, and 74% clay with a pHH2O of 8.2.

#### *2.2. Soil Sterilization for Vertical Movement*

Soil was autoclaved in 1 kg amounts at 20% GWC in open bags for one hour at 121 ◦C, at 1.5 bar and with the process repeated three times, with a 24 h break between the commencements of each autoclave cycle. Upon the completion of the sterilisation process, the autoclaved aliquots were combined into a sterile polypropylene bag and then left for two weeks in an open aseptic environment. After this time, three samples were taken from the soil and screened for nematode presence using passive extraction.

#### *2.3. Microcosm Design*

The microcosms were made from an unplasticised polyvinyl chloride (uPVC) pipe with an internal diameter of 50 mm. The pipe was cut into 40 cm lengths, which were then cut longitudinally to allow for the microcosm to be split lengthwise to facilitate soil recovery. The bottom of the microcosm was held together and sealed with a 50 mm uPVC end cap and the top of the tube with a 50 mm uPVC pipe to pipe joining collar. The cut edges of the pipe were sealed with tape to prevent water loss and splitting under expansion of the soil. Under experimental conditions, the microcosms were supported in plastic crates, which carried up to 16 microcosms.

#### *2.4. Microcosm Packing*

The microcosms were packed, so that sterile and non-sterile soil was represented in all combinations within the experiments as either a top (0–15 cm) or bottom (15–30 cm) treatment. This meant that there was; Kirby top: Kirby bottom, Kirby sterile top: Kirby bottom, Kirby top: Kirby sterile bottom, and Kirby sterile top: Kirby sterile bottom with the same combinations for the Cotton soil. The soils were packed to generate a dry bulk density of 1.4 g/cm3, which was achieved by weighing the required

mass of soil for each half of the microcosm and adding one-third of the mass at a time before tamping the tube five times on the bench to get the required compaction. An internal 15 cm mark was present in each tube to assist with packing to the desired bulk density. After either the bottom or the tops of the tubes were packed water was added to the presenting surface to raise the gravimetric water content of the soil to 20%.

#### *2.5. Planting and Watering*

Into the planted microcosms two seeds of wheat, variety Gregory, were planted to a depth of 1 cm and then the tops of all the microcosms were overlaid with 20 mL of 4 mm polypropylene beads to reduce evaporation. The initial starting weight of each established microcosm was taken and the GWC maintained by weight every Monday, Wednesday, and Friday of the experiment duration with the addition of variable amounts of rainwater to within 0.25 g of starting weight.

In a second experiment, a flood irrigation for half of the planted and unplanted microcosms was conducted two weeks after establishment by adding 50 mL of rain water to each of the identified microcosms. This was calculated as being sufficient water to raise the GWC to 35%, which had been established as being equivalent to −10 kPa.

#### *2.6. Recovery and Nematode Counting*

The microcosms were destructively sampled four weeks (28 days) after sowing wheat. The above ground plant height was recorded and the plant shoot material excised. Fresh weight was determined and the samples were dried for 48 h at 80 ◦C to determine the dry weight. Plastic beads were recovered from the top of the microcosms and then the tape and top and bottom caps were removed. The microcosms were opened in a large tray and the depth of visible root growth recorded. Soil was then recovered from 5 to 10 cm and 20 to 25 cm depths. A proportion of this soil was recovered to an aluminium tray to determine the GWC and approximately 10 g was weighed into a 50 mL centrifuge tube for nematode recovery [17].

#### *2.7. Results and Analysis*

Excel was used to tabulate results and interrogate data for correlation coefficients (r) generation. GenStat was used to undertake analysis of variance (ANOVA) of the measured variables, with Tukey's comparison test used to determine differences between multiple means with significance assumed to occur at the *p* < 0.05 level. Outcomes were graphically presented.

#### **3. Results**

#### *3.1. Vertical Distribution*

The total free living nematode populations were observed to decrease with depth under both the cotton-cotton and cotton-maize rotations with the overall population decline fitting the equation <sup>y</sup> <sup>=</sup> <sup>−</sup>0.0928x<sup>3</sup> <sup>+</sup> 0.8549x<sup>2</sup> <sup>−</sup> 2.7682x <sup>+</sup> 4.6508, with a correlation of r = 0.99. There was no significant difference between the rotations (*p* = 0.07), but there was a difference with depth (*p* = 0.001). An interaction between depth and rotation (*p* = 0.02) was observed with a larger nematode population in the cotton-cotton rotation between 30 to 70 cm than that recovered from under the cotton-maize rotation (Figure 1).

There was a good correlation between soil gravimetric water and nematode recovery from the cotton-maize rotation (r = 0.87), but not for cotton-cotton (r = 0.28). Both of the systems had good correlation between soil bulk density and the average number of nematodes (r = 0.80 and 0.84), with nematode abundance following a negative exponential curve as the bulk density increased.

**Figure 1.** The mean free living nematodes per gram of soil recovered from a Vertosol supporting a cotton-cotton (white) and cotton-maize (grey) rotation to varying depths. Error bars represent the standard error of the means (n = 3) and the asterisks (\*) indicate a significant difference between rotation and depth (*p* = 0.02).

#### *3.2. Vertical Movement*

The initial nematode populations were enumerated at 6.1 and 9.8 nematodes/g for the Kirby and Cotton soils, respectively. Examination of the soils, post sterilisation recovered no live nematodes in the Kirby soil, but the Cotton soil had 1.1 nematode/g (11% of the original nematode population) still alive after three rounds of autoclaving.

In both microcosm experiments, the wheat roots reached the bottom of the columns in the Kirby soil (30 cm), but only managed an average depth of 20.6 cm in the Cotton soil. Despite this, the wheat biomass was significantly higher (*p* < 0.001) in the Cotton soil than the Kirby soil, with means of 0.3 and 0.1 g, respectively.

In the first microcosm experiment, there was no significant difference in the nematode recovery between the Kirby and Cotton soils (*p* = 0.32), the top and bottom of the microcosms (*p* = 0.33), and whether wheat was planted or not (*p* = 0.11). Despite not being significant, nematode recovery, being expressed as a ratio of the control, implied movement up into sterile Kirby soil in both the presence and absence of wheat (Figure 2a). The average ratio of nematodes in sterile Cotton soil did not get above 1 in upper sterile Cotton soil, which implied a lack of upward movement (Figure 2a). In the bottom of the microcosms, there was a trend for increased nematode recovery in both sterile Kirby and Cotton soils, but only when wheat was planted (Figure 2a), despite the maintained 20% gravimetric water content.

In the second microcosm experiment, imposing flood irrigation on the Cotton soil significantly increased the number of recovered nematodes (*p* = 0.07), with 2.7 as compared to 1.63 nematodes/g for irrigated and GWC maintained soil, respectively. There was no significant difference in nematode recovery from either top or bottom of the microcosm (*p* = 0.39). Planting wheat had no significant effect on nematode recovery (*p* = 0.41), although the nematode recovery ratio increased above 1 for both irrigation treatments in the absence of planted wheat (Figure 2b).

**Figure 2.** The ratio of nematodes recovered from sterilised soil situated either above or below non-sterile soil, compared to those recovered from a completely sterile treatment. A ratio of more than 1 (for upward movement, lighter shades) and −1 (for downward movement, darker shades) indicates an increase over the control. Kirby (Yellow) and Cotton (Brown) indicate where soil was sourced with (**a**) looking at the impact of sowing wheat (diagonal black shading) on nematode recovery, while (**b**) is the analysis of the impact of a maintained versus flood irrigation treatment (black dashed border) only in the Cotton soils.

#### **4. Discussion**

Farming systems are prone to change and the Australian cotton production system is no exception. However, the focus of these changes are often on either crop productivity or chemical and physical properties of the soil [19], with less attention being given to the soil biology [1], despite the fact that most, if not all, of our production diseases and pests are biological. We attempted to address some

simple questions relating to nematodes in these systems in the face of the first observations of reniform nematode causing problems in Australian cotton [10] and the loss of potential chemical controls [8].

Our initial focus was on whether nematode populations declined with depth and whether rotations could influence their distribution. Our results indicated that there was a decline with depth as well as differences between rotations. Given that we sampled at a time when both of the rotations were growing cotton, we believe it would be safe to assume that the dissimilarity in the recovered nematodes/g between 30 to 70 cm (Figure 1) occurred due to rotational difference. With cotton being a tap rooted eudicot and maize a fibrous rooted monocot, a probable driver for changes in the nematode numbers between these depths is rooting patterns [20,21] in combination with these roots persisting post-harvest [22]. Root exudation and decomposition both have the potential to alter the soil microbiology [23], which, in turn, would directly influence both the nematode community composition and size [24]. Differences in the field management that are associated with the different rotational crops, such as fertilizer regimes, cultivation, and stubble management, could also be altering the soil microbial community and in turn the nematodes [25]. In keeping with this, cotton and maize roots are known to differentially alter the soils' abiotic properties [9], thus potentially altering the nematode population densities, which was supported with the observed correlations between nematode numbers, soil moisture, and bulk density. What a change is abundance does not address is whether it is also associated with a change in the population's trophic groups? Unfortunately, limitations on the volume of soil in our microcosms, our inability to remove all of the nematodes from the Cotton soil with autoclaving and the recoveries of only one to two nematodes/g from the recolonized soil, there was insufficient numbers to confirm this. However, with known pathogenic nematodes in these soils and a potential industry threat identified elsewhere, the difference in nematode abundance in soil from under the different rotations adds support for rotational crops remaining one of the few strategies available at present for nematode control in Australian cotton system [26,27].

Having observed a difference between the rotations, we postulated whether there was potential for nematodes to move vertically within these soils. Vertical nematode movement has been previously reported, notably for several plant parasitic nematodes that recolinise and recover from populations that reside deeper in the soil after crop protection control measures, such as nematicide application, have been implemented [28,29]. However, this work was undertaken on lighter soils than the Vertosol soil being investigated here [29]. Water is known to play a key role in both nematode movement and shaping community structure [30–32], and so we initially kept our soils at a moisture level that should have facilitated nematode movement [30]. However, in our limited and short term experiments, nematode movement either up or down in a heavy clay Vertosol appeared to be restrictive (Figure 2). In addition, we included the planting of wheat as a treatment factor, while assuming that the presence of growing roots might encourage nematode movement [33], but we observed no significant movement in response to plant roots (Figure 2). While surprising, it has been previously reported that the vertical distribution of roots does not always correlate to nematode movement or abundance [34]. While our microcosm experiments imply limited nematode movement and recolonisation potential in Vertosols, there are a number of caveats to consider prior to deriving any generalisations regarding nematode movements in these heavy clay soils. Firstly, our system was only run for four weeks, a relatively short period of time in a cropping cycle, we had limited replication and our Vertosol columns were not exposed to repeated flooding and drying cycles, as experienced under field conditions, but kept constantly moist. Finally, we did not work on the soils containing the reniform nematode due to quarantine concerns, but, given the potential for nematodes to behave differently, could not rule out the potential for *R. reniformis* to recolonise Vertosols from depth after flooding [10,28].

Accordingly, whilst these studies were preliminary, it is apparent that we still have much to learn about the diversity, potential threats, activity and importance of nematodes in Australian Vertosols, which themselves are challenging to work with. Within these heavy clay soils, the potential to use crop selection as a control strategy remains [26,27]. In the face of a reduction in available chemical controls [8], this strategy may continue to be one of the few mitigation options other than preventing

nematode movement from infected fields [10] in the first place by maintaining 'come clean, go clean' practices [11].

**Author Contributions:** O.K. and B.G. undertook sampling, analysis and manuscript preparation. K.P. assisted with soils analysis and E.F. assisted with soil sampling and nematode enumeration. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was made possible with funding from the CRDC under UNE1403 and UNE2001. The University of New England (UNE) contributed via the GRASS programme, which supported Elijha, Katherine's PhD and in-kind and cash to UNE1403 and UNE2001.

**Acknowledgments:** Access to the trial sites and farms is gratefully acknowledged as is glasshouse support from Mick Faint at UNE.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
