Testing soil nematode extraction efficiency using different variations of the Baermann funnel method

Nematodes are increasingly used as powerful bioindicators of soil food web composition and functioning in ecological studies. Todays’ ecological research aims to investigate not only local relationships but global patterns, which requires consistent methodology across locales. Thus, a common and easy extraction protocol of soil nematodes is needed. In this study, we present a detailed protocol of the Baermann funnel method and highlight how different soil pre-treatments and equipment (soil type, amount of soil, sieving, filter type) can affect extraction efficiency and community composition by using natural nematode communities. We found that highest nematode extraction efficiency was achieved using lowest soil weight (25 g instead of 50 g or 100 g) in combination with soil sieving, and by using milk filters (instead of paper towels). PCA at the family level revealed that different pre-treatments significantly affected nematode community composition. Increasing the amount of soil increased the proportion of larger-sized nematodes being able to overcome long distances. Sieving is suggested to break up soil aggregates and, therefore, facilitate moving in general. Interestingly, sieving did not negatively affect larger nematodes that are supposed to have a higher probability of getting bruised during sieving. The present study shows that variations in the extraction protocol can alter the total density and community composition of extracted nematodes and provides recommendations for an efficient and standardized approach in future studies. Having a simple, cheap, and standardized extraction protocol can facilitate the assessment of soil biodiversity in global contexts.

Building on many previous studies testing nematode extraction efficiencies [13,16,17], 82 we provide a detailed protocol to extract soil nematodes, e.g., for global assessments, focusing 83 on a consistent methodology and high extraction efficiency, which can be used by many 84 laboratories worldwide. By doing so, we highlight how common soil pre-treatments and 85 equipment (sieving, amount of soil, filter type) can affect extraction efficiency and 86 community composition and should therefore be considered in future studies. 87 Generally, soil samples are homogenized before extraction by sieving [18]. Mesh sizes 88 from 1 mm up to 5 mm are commonly used, whereby small mesh sizes require gentle sieving to not bruise nematodes [12]. Mortality can occur when samples are handled roughly, and loss 90 of species after sieving was shown to be species-specific [13,19]. By contrast, 91 homogenization breaks soil aggregates and is assumed to facilitate the movement of 92 nematodes through the soil, likely resulting in higher nematode extraction efficiency. 93 However, despite the inconsistency of mesh sizes used in different studies, the consequences 94 for extraction efficiency and comparability of results have not been tested before. 95 In ecological long-term experiments or in pristine habitats, destructive samplings like 96 taking soil cores often are strongly limited to prevent destruction of the plots, and only small 97 amounts of soil may be available for nematode extraction. Using large amounts of soil, on the 98 other hand, may also reduce extraction efficiency as less mobile nematodes are discriminated 99 [12,17]. Thus, exploring the role of the amount of extracted soil for nematode extraction 100 efficiency is required to provide general recommendations. 101 Different permeable filters are used to separate nematodes from soil. Most often 102 cotton-wool milk filters are used, but also cheesecloth, filter paper, or paper tissue are 103 suggested. However, knowledge of the influence of different filters on extraction efficiency is 104 missing. Regarding the availability of materials and costs, we test milk filters and common 105 paper towels in this study.

106
A well-chosen combination of the settings described above may help to increase 107 nematode extraction efficiency and to avoid potential biases of different extraction protocols.

108
In this study, we evaluated different settings of the Baermann-funnel method by varying 1) 109 different sieving mesh sizes, 2) different amounts of soil, and 3) two different filter types to 110 investigate the consequences for the total amount of extracted nematodes and for nematode 111 community composition. In addition, two very different soil types, i.e., loamy and sandy soil, 112 were used to enable us to make general recommendations.

116
We tested the effects of four variables in soil nematode extraction in a full-factorial 118 design: two soil types (loamy and sandy soil), three soil sieving treatments (2 mm mesh size, 119 5 mm mesh size, no sieving), three amounts of extracted soil (25 g, 50 g, and 100 g fresh 120 weight), and two permeable filters (milk filters and paper towels). All treatments were 121 replicated five times resulting in 180 samples. and C-to-N ratio 15.7. Clay content was 14%, silt content 41%, and sand content 45% [21].

126
The sandy soil was taken from the Kreinitz Experiment, a tree biodiversity experiment in 127 Zeithain, Saxony, Germany [22]. Soil samples were taken at a distance of around 10 m to the 128 experimental plots from 0 to 20 cm depth. Soil pH was 5.5, carbon concentration 1.1 %, 129 nitrogen concentration 0.1%, and C-to-N ratio was 11.4. Clay content was 2%, silt content 130 5%, and sand 94%.

131
Before any treatments were applied, soil was gently mixed. Afterwards, a fraction of the soil 132 was sieved with a mesh of 2 mm or 5 mm. One fraction of the soil was not sieved, but roots 133 and stones were removed by hand to correctly evaluate soil weight [12]. Three different 134 amounts of fresh soil were used for extraction: 25 g, 50 g, and 100 g, representing a thickness 135 of the soil during extraction of about 1, 2, and 4 cm, respectively. Finally, two different filter 136 types were used: commonly used milk filters (Sana, type FT 25) and paper towels (ZVG Baermann funnel apparatus 140 The Baermann funnel apparatus consisted of a funnel with an inner diameter of 11 cm. 141 Using a different diameter is possible but should be constant among and within studies. A 12 142 cm piece of silicone hose was attached to the funnel ending and fixed with hot glue to prevent 143 water leakage. The tube was closed with a squeezer clip at the end of the silicon hose. The

144
Baermann funnel apparatus was installed in a horizontal position without buckling of the 145 silicon hose. Soil of a certain amount was filled into circular PVC tubes (7 cm in diameter) 146 with a mesh of 250 µm at the bottom, allowing nematodes to traverse the mesh. The mesh 147 was covered with a filter (milk filter or paper towel) to prevent soil particles to enter the 148 nematode solution (Fig 1). To obtain clean samples, we used a large piece of the filter 149 material to prevent soil particles to enter the nematode solution from the side. This, however, 150 increases evaporation and water has to be added if necessary to prevent that the soil falls dry.  The weight of the empty PVC tube including the filter and a label to identify the 158 sample was noted. Fresh soil of 25 g, 50 g, or 100 g was filled into the PVC tubes. The exact 159 weight has to be noted to get soil water content and relate nematodes to g dry soil. The height 160 of the soil volume was 0.9 cm, 1.9 cm, and 3.8 cm for 25 g, 50 g, and 100 g of fresh soil 161 weight, respectively. Afterwards, the PVC tube with soil was inserted in the funnel.     Table 1. Nematode families extracted from loamy and sandy soil. List of nematode families extracted from loamy (L) and sandy (S) soil using 212 the Baermann funnel method with assigned c-p classes after Bongers (1990) and Bongers and Bongers (1998) and trophic groups after Yeates et al.,

213
(1993). In addition, assignment to r-and K strategists. R-strategsits were nematodes from c-p class 1 and 2, whereas c-p classes 3 to 5 are classified Microlaimidae. Occurrence describes in which soil types nematodes occurred. Using mean relative occurrence of nematode families in loamy and 217 sandy soils were used to assign nematodes to abundant (up to 5%), medium (5-1%), and rare (below 1%) families. Means ± SD of nematode 218 families are given for samples displayed in the PCA using different extraction treatments, i.e., using i) 25 g of fresh soil sieved at 5 mm and ii) no 219 sieving, respectively, as well as using 100 g of fresh soil sieved at 5 mm (n = 3). Taxa were sorted by overall mean of extracted nematodes. Nematode extraction efficiency was affected by a significant four-way interaction of 225 soil type, sieving, amount of soil, and filter (Table S1). To better identify specific treatment 226 effects, datasets were divided by soil type as nematode densities were mainly affected by soil 227 type. Mean ± sd nematode density in loamy soil was 17.5 ± 12.4 nematodes g -1 dry soil 228 compared to 1.4 ± 0.9 nematodes g -1 dry soil in sandy soil. The combination of treatments was 229 of importance for nematode extraction efficiency in the loamy soil as indicated by the three-230 way interaction, but this was not the case in the sandy soil (Table 2). In loamy soil, highest nematode extraction efficiency was achieved using 25 g of fresh 240 soil sieved with a mesh size of 5 mm and using milk filters (42.2 ± 7.7 nematodes g -1 dry soil 241 weight), with milk filters generally increasing nematode extraction efficiency (Fig 2a).

242
Sieving 25 g of soil with a smaller mesh resulted in only slightly fewer nematodes (40.4 ± 5.6 243 nematodes g -1 dry soil weight; -4% in comparison to sieving at 5 mm) and did not differ 244 significantly from sieving with 5 mm. In contrast, nematode extraction efficiency was 245 significantly lower when the soil was not sieved. No sieving of 25 g fresh soil resulted in 31.8 246 nematodes g dry soil -1 , that is, 25% fewer nematodes compared to highest number of 247 extracted nematodes. In sandy soil, sieving was not of significant importance (  Fig 2c). Generally, paper towels significantly reduced overall nematode efficiency by 65% in 261 loamy soil and by 34% in sandy soil. Increasing the amount of soil decreased nematode 262 extraction efficiency from 25 g to 50 g by 30% and from 25 g to 100 by 61% in loamy soil, 263 whereas in sandy soil the reduction was 42% and 60%, respectively (Table 3).  To assess the consequences of sieving and soil weight on nematode community 100 g fresh soil compared to 25 g fresh soil (Fig 4a). No sieving did not increase the amount 284 of K-strategists, i.e., larger organisms that are supposed to be more likely to be damaged by 285 sieving (Fig S1). However, in samples using 100 g loamy soil, more large ( Fig S1) and rare In the present study, we found that the combination of different extraction treatments 315 significantly affected nematode extraction efficiency. Although treatment combinations were 316 of different importance in loamy and sandy soil, overall highest numbers of extracted nematodes were observed when using milk filters and the lowest amount of soil, i.e., 25 g 318 fresh soil. Sieving was important in the loamy soil, while it did not matter if soil was sieved 319 with 2 or 5 mm mesh size. Although not sieving soil yielded highest nematode numbers in the 320 sandy soil, no significant difference was found between sieving and not sieving. Therefore, to 321 achieve high nematode extraction efficiency in different types of soil, it is recommended to 322 use small amounts of soil in combination with sieving and using milk filters.

323
Sieving was of higher importance in the loamy soil than in the sandy soil. Loamy soil 324 has more stable soil aggregates than sandy soil, which is why we suggest that breaking up soil

342
This may be the result of a higher probability that 100 g soil contains more rare species.
Moreover, rare species often are large in body size, which is why they may have a higher 344 chance to be extracted from larger amounts of soil, as mentioned above. However, the soil 345 volume of 100 g samples also selected against the majority of other nematode families. In 346 summary, using the combination of treatments that resulted in highest nematode extraction 347 efficiency may select against some rare species but may better reflect total densities. To

352
Using milk filters resulted in a significantly higher number of extracted nematodes 353 than using paper towels. Paper towels are supposed to adsorb water, whereas milk filters are 354 supposed to filter a solution. The fabric of paper towels is probably chosen such that the fibers 355 will take up water, and this paper structure may hamper nematodes to pass the paper towel.

356
Instead of using paper towels as an alternative for milk filters for biodiversity assessments,  sieving), type of filter (milk filter and paper towel), soil weight (25 g, 50 g, and 100 g fresh