Earthworm species occurrence in agroecosystems in the Midlands, KwaZulu-Natal, South Africa

Little is known about the species composition of earthworms in agroecosystems in South Africa even though earthworms provide soil ecosystem services and are useful biological indicators of changes in the habitats. Given the land use and management impact biodiversity, the aim of this study was to document earthworm species that occur under cultivated land in the KwaZulu-Natal Midlands. A survey of nine farms that practise conservation agriculture was carried out between 2018 and 2020. Twelve earthworm species belonging to four introduced families: Acanthodrilidae (Dichogaster bolaui), Rhinodrilidae (Pontoscolex corenthrurus), Lumbricidae (Aporrectodea caliginosa, Aporrectodea rosea, Aporrectodea trapezoides, Lumbricus rubellus, Octolasion cyaneum, Octolasion lacteum), Megascolecidae (Amynthas aeruginosus, Amynthas corticis, Amynthas gracilis, Amynthas rodericensis) and juveniles from an indigenous family Tritogeniidae were recorded from cultivated fields. The type of crop (habitat) affected both species richness and abundance of earthworms significantly. However, post hoc results showed differences in species richness between the soya and the maize only, with greater species richness in the maize. Our results demonstrate that habitat type has a major influence on communities of earthworms in agroecosystems.


Introduction
Terrestrial ecosystems benefit immensely from soil organisms (Nielsen et al. 2015). Knowledge on soil fauna has increased and more attention has been given to these taxa in recent years ) and earthworms' contribution to soil ecosystem and soil fertility has been documented. Earthworms contribute to ecosystem services by converting organic matter into rich humus in the form of casts. Earthworms improve soil fertility and quality, influence soil formation, improve soil nutrient availability, stabilise the soil, increase soil porosity, improve water infiltration and increase the overall health of the soil Jouquet et al. 2006;Brown et al. 2018).
Food production relies on healthy soils and there is an urgent need to understand biodiversity and biophysical regulations of soil fertility better . Therefore, access to accurate taxonomic information of soil organisms is essential. Unfortunately, taxonomists trained to identify soil fauna are in decline . The adoption of environmentally friendly and sustainable agriculture is therefore long overdue because of rapid increase in human population, climate change and deteriorating soils (Kassam et al. 2009;Delgado and Gantzer 2015).
Earthworm populations tend to do better in no-till systems (Bartz et al. 2013;Santos et al. 2018), hence earthworms are widely used as soil health indicators (Nadolny et al. 2020). A recent review in Brazilian no-tillage agriculture highlighted that the no-till system promotes earthworm populations (Demetrio et al. 2019). However, according to Santos et al (2018), although work has been done in South America on no-till systems, knowledge of earthworm diversity in agroecosystems is still limited.
In South Africa, few studies have documented earthworm species in agroecosystems. The studies that looked at the occurrence of earthworms in agricultural ecosystems reported that peregrine species were dominant (Visser and Reinecke 1977;Reinecke and Visser 1980;Dlamini et al. 2001;Haynes et al. 2003;Simonsen et al. 2010). Tillage is known to affect endogenic and anecic earthworm diversity and abundance negatively (Reinecke and Visser 1980) unlike no-tillage agriculture (Peigne et al. 2009;Hutcheon et al. 2001). According to Dlamini et al. (2001) and Haynes et al. (2003), earthworms in agroecosystems have not been studied adequately in South Africa. After Nxele (2015) recorded indigenous earthworms in sugar-cane fields that had been under no-till for more than twenty years, we hypothesised that more indigenous species occur under no-till agriculture. As such, a study to document earthworm diversity under cultivated fields was initiated in 2018 in minimum tillage or no-till agroecosystems.

Study sites
Nine farms in KwaZulu-Natal Midlands (Fig. 1) were sampled for earthworms. Each farm had either sugar-cane, maize, soya, ryegrass pasture, mixed species pasture or a mixture of crops (Table 1). All the farms practise conservation agriculture; however, the farms have been under no-tillage for a different number of years.

Earthworm sampling
Earthworms were collected quantitatively and qualitatively. The quantitative method follows that of Nxele et al. (2015) and Bartz et al. (2014) with slight modification on plot sizes. Sampling was carried out in one hectare with nine sampling points; adjacent sampling points were 30 m apart. Earthworms were collected by digging out 50 cm x 50 cm × 20 cm soil monoliths. The soil was hand-sorted for earthworms in large plastic trays (50 cm × 40 cm × 5 cm). The holes were filled back with the soil after removing the specimens from the soil. Active searching under stones and logs was part of qualitative sampling. Specimens that were collected were washed and narcotised using 20% ethanol solution. Some specimens were preserved in absolute ethanol for DNA analysis. The remaining specimens were fixed in 4% formalin for at least 24 hours before  being preserved in 75% ethanol. Studies of the internal anatomy were conducted after dorsal dissections. The KZN museum database and the following literature: Plisko 2001Plisko , 2010Ljungstom 1972;Michaelsen 1899Michaelsen , 1908Michaelsen , 1913Reynold and Reinecke 1976;Zicsi and Reinecke 1992;Reinecke 1977 andPickford 1937 were used to obtain information on distribution. All new material is deposited in the Oligochaeta collection in the KwaZulu-Natal Museum (NMSA).

Data analyses
Data analysis was per crop type regardless of which farm it came from. Species richness and abundance datasets were analysed in R using the generalised linear mixed models (GLMMs) because data were not normally distributed. The lme4 package (Bates et al. 2015) was used when calculating GLMMs. The Poisson distribution was the best fit for the species richness dataset, while the negative binomial distribution was the best fit for the species abundance dataset (Bolker et al. 2009). In the models, the type of crop (maize, pasture, soya and sugar-cane) was the fixed factor, while the random factor was the farm. The multcomp package (Hothorn et al. 2008) was used to determine the similarities and/or differences between pairs of crops. Remarks. In the current study, this species was found in sugar-cane, maize and soya fields. It was the common species in the sugar-cane fields although not in high numbers. In soya, it was found with megascolecids and lumbricids. Remarks. This species occurs in numerous sites, both natural and agricultural. In natural sites, Plisko (2001) reported this species in grasslands, forests, natural bushes and near rivers. In cultivated fields, Pontoscolex corenthrurus has been collected under most crops including maize, sugar-cane, banana, avocado, citrus; some specimens were collected in pine and gum plantations, some even from vegetable gardens (Plisko 2001). This species has been used in experiments at different institutions. Pontoscolex corenthrurus has also been collected in polluted sites in KZN (Plisko 2001). The current collection was in sugarcane on one farm only, although we had expected to collect the species in other areas in the KZN Midlands because of its wide occurrence. Remarks. In the current collection, Aporrectodea caliginosa was collected in mixed species pasture. This species is closely related to A. trapezoides and they are found together mostly; however, caliginosa is less common (Plisko 2010). This is the first report of this species in this region. ( Remarks. Aporrectodea trapezoides has common occurrence compared to A. caliginosa. In the current study, this species was collected together with A. caliginosa on cultivated land with mixed species pasture. The mixed species field is under minimal tilling with little disturbance of the top soil from time to time. Previously, Aporrectodea trapezoides was collected from natural habitats that include forests, natural bush, cultivated fields and along rivers (Plisko 2010 Remarks. This species has been collected in two areas in the EC and WC only. The EC record is from a private garden soil (Plisko 2010), whilst the WC record is from a forest in Kirstenbosch. The new material is from a mixed species pasture and the species occurred in high numbers. There is a high possibility that the distribution of this species in agroecosystems is wider than what is known because not much sampling in cultivated fields has been done. Remarks. The species has been found in grasslands, agricultural fields and nature reserves in KZN and GP only. It is unclear why there are no records from other South African provinces although sampling has been done in these provinces. In the present study, Amynthas aeruginosus was collected under maize, soya, sugar-cane and mixed species pastures. Remarks. The species has been reported from almost all over South Africa; it is apparent that this species occupies a wide range of biotopes. It is common in the upper layers of soil, mostly in rotting litter in plantations (Plisko 2010) and decomposed sugar-cane reeds (Ljungström 1972). In the present study, Amynthas corticis was collected from maize fields only. Remarks. Amynthas rodericensis is common in natural and agricultural fields.

Discussion
This checklist adds to the knowledge of the species composition of earthworms on farms in South Africa. These data will contribute to future studies on the importance of earthworms in agriculture. Commercial farmers, who were the first to adopt the no-till system in KZN own some of the farms that we sampled. The earthworms that we collected from cultivated fields were introduced, except specimens of Tritogenia. Tritogenia specimens were collected from a ryegrass pasture field in Karkloof, which is near a veld (natural grassland), so it is possible that the indigenous species are recolonising the pasture in this particular instance. However, Tritogenia have been collected from ryegrass pasture in the past from KZN Midlands (Haynes et al. 2003) making this a second record of Tritogenia in ryegrass pasture in the KZN Midlands. Visser and Reinecke (1977) collected small numbers of Tritogenia in cultivated fields at Mooi River. Two species, Aporrectodea caliginosa and Octolasion cyaneum, were recorded for the first time in KwaZulu-Natal. Aporrectodea caliginosa is morphologically similar to Aporrectodea trapezoides with the external difference being tubercula pubertatis on two separate tubercles 31 and 33 for caliginosa, but continuous bands on 31-33 for trapezoides. The colour of the two species is also different with trapezoides mostly dark brownish-grey, whilst caliginosa is mostly pale . Octolasion cyaneum has been collected from the Cape in two sites, one being the garden where imported flowers were planted and at a resort (Plisko 2010). It is likely that this species may have a wider range than previously reported; however, more sampling is needed to confirm this for agroecosystems.
According to Visser and Reinecke (1977), introduced earthworm species are more abundant in most parts of South Africa. The presence of introduced species in agricultural fields is consistent with Walsh et al. (2013) who recorded alien earthworms, dominated by a lumbricid, Aporrectodea trapezoides, from across wheat growing fields. Earthworm communities were dominated by introduced species in a study by Manono and Moller (2015) in New Zealand pastures. Similarly, lumbricids were dominant in the present study; they were collected in almost all crops, except in sugar-cane and soya. Amuza et al. (2021) assessed the presence of earthworms in agricultural crops in Romania and found that the lumbricid, Aporrectodea caliginosa nocturna, was the dominant species. Reinecke and Visser (1980) conducted a study in the Mooi River irrigation field to look at the effect of land use and fertilisers on earthworms. The lumbricids, A. trapezoides and E. rosea, had high population densities, which suggested that, in South Africa, the common earthworms in cultivated fields were lumbricids (Reinecke and Visser 1980). Advantages of introduced earthworms is that they adapt easily to different environments and they may reproduce parthenogenetically (Visser and Reinecke 1977;Ljungström 1972;Reynolds and Reinecke 1976) resulting in rapid increase in numbers.
Southern African indigenous earthworms tend to vanish almost immediately after the land is used for agricultural purposes (Reinecke and Visser 1980). However, the current collection of Tritogenia in ryegrass pasture, the results of Haynes et al. (2003), Visser and Reinecke (1977) and Nxele (2015), suggest that, with sustainable land use, it is possible that the indigenous earthworms will re-colonise cultivated fields.
The type of crop affected both species richness (p = 0.02) and abundance (p < 0.001) of earthworms significantly (Table 2). Species richness (p = 0.04) and abundance (p < 0.001) of earthworms were significantly higher in the pasture than in the soya; this observation contradicted Manono and Moller (2015) who reported lower species richness in pasture. Geographical differences in our study and that of Manono and Moller (2015) could be the reasons for differences in our results. According to Manono and Moller (2015), food supply determines earthworm abundance and agricultural practices may affect the availability of organic material in the soil. The farms, on which we sampled, also plant cover crops, which increase the availability of organic matter in the soil. However, the change in plant communities may affect the quality and quantity of organic matter that maybe available to earthworms (Hubbard et al. 1999;Bohlen et al. 1997); this may explain why the perennial pasture, which has no plant rotation, had more earthworms.
Our results also demonstrated significantly greater species richness (p = 0.02) and abundance (p < 0.001) of earthworms in the maize than in the soya. These results are similar to those of Amuza et al. (2021) who reported greater abundance of earthworms in maize than in soybean. All the maize fields had old maize residues on the ground, as well as short stubs, which would have provided a continuous food supply for earthworms.
As introduced earthworms have been collected in all biotopes in South Africa (Plisko 2010) and from agricultural fields (Dlamini et al. 2001;Haynes et al. 2003;Reinecke and Visser 1980), it is not surprising that introduced earthworms were collected from all the farms on which we sampled. There are gaps in our knowledge of the species composition of earthworms in different agroecosystems. As such, more extensive sampling during different seasons is necessary in order to gain more insight into the taxonomic diversity and distribution of earthworms in agroecosystems in South Africa.