Methodological adaptation for nematodes extration in forest soils of the southern Mata Atlântica

The objective of this work was to clarify whether the method to extract nematodes from European soils is suitable for forest soils and litter in the eastern of Paraná state, Brazil, and whether nematode abundance differs between sites with different ecosystems and levels of human interference. The study sites were situated in the coastal area of the Serra do Mar, near the town of Antonina, in Eastern Paraná, Brazil. Cobb’s sieving and decanting method was more appropriate than ISO method, since extraction effi ciency was higher and intrasample variability was signifi cantly lower. In order to achieve an extraction effi ciency higher than 90%, Cobb’s method was modifi ed. For the extraction of nematodes from litter, the Baermann funnel, with an extraction time of 48 hours, yielded an extraction effi ciency higher than 90%. Nematode abundance in litter was higher than in soil. The mean number of individuals extracted from the litter increased with the age stage of the forest sites sampled, and there was no difference in the number of individuals in the soil of the four forest sites. Mean nematode abundance in soil in banana plantations was about twice as high compared to the banana-palmito mixed stands and to the forest sites.


Introduction
Mata Atlântica (Brazilian Atlantic Forest) is among the most diverse and most threatened ecosystems of the world.Its forests originally covered about 1.3 million km 2 , corresponding to 15% of the land surface of Brazil, extending for more than 3,000 km along the eastern Brazilian coast, between 6°S and 30°S (Fundação SOS Mata-Atlântica, 1993).The region has been largely impacted by colonization, which resulted in extensive deforestation, conversion into farmland and urbanization.Consequently, forest has been reduced to about 7% of the original area, or less than 100,000 km 2 (Schäffer & Prochnow, 2002).In the remaining forested area, one of the largest remnants is found at the coast of Paraná.It encloses a mosaic of environments at elevations ranging from sea level to more than 1,500 m a.s.l.As a consequence of the geomorphologic and climatic variations, the dense ombrophilous forest (Veloso et al., 1991) is divided in T. Moser and S. Frankenbach Pesq. agropec. bras., Brasília, v.44, n.8, p.975-980, ago. 2009 four subformations (lowland, alluvial, submountain and mountain forest), each presenting distinct fl oral and structural characteristics (Instituto Brasileiro de Geografi a e Estatística, 1992).In addition, small agroforestry sites are scattered in the same area.
Nematoda are amongst the most numerous soil animals, also showing a high diversity.However, due to their small size and complex extraction from soils, they are relatively rarely studied.In forest soils of tropical regions in general, and of South America in particular, almost nothing is known about free-living nematodes.Most taxonomic effort have been addressed towards important plant parasites (Huang & Cares, 2006).
The main objectives of the present study were to clarify whether the method to extract nematodes from European soils is suitable for forest soils and litter of eastern Paraná, and to identify whether nematode abundance differs between sample sites with different ecosystems and levels of human interference.

Materials and Methods
Abundance and distribution of nematodes were studied in four forest sites of different ages and six agroforestry sites (three banana monocultures and three mixed banana-palmito plantations), located in Brazilian Mata Atlântica, in 2008.The investigation was part of the German-Brazilian cooperative project Solobioma (Project Solobioma, 2009), which aims to assess the ecosystem quality of secondary forests in this region of eastern Paraná and their potential to conserve biodiversity and related ecosystem services.
The study sites were located in the coastal plain area of the state of Paraná, in the municipalities of Antonina and Guaraqueçaba.Soils were characterized by low pHCaCl 2 (3.5-4.3),variable C/N-ratios (15-40), and medium organic matter contents (2-5%), in the upper 20 cm soil layer (Ferretti & Britez, 2006).Details of the site characteristics, including exact UTM coordinates are shown in Table 1, and further information on the sample sites can be found in Römbke et al. (2009).
The climate of the coastal region of Paraná can be described as mesothermic subtropical humid, corresponding to the Cfa type, according to Köppen´s classifi cation (Schröder, 2000;Strahler & Strahler, 2005).Mean annual temperature is above 18°C, and monthly precipitation over 60 mm.Frost rarely occurs in areas from sea level up to 700 m a.s.l.(Instituto Paranaense de Desenvolvimento Econômico e Social, 2001).Annual rainfall in the region varies between 2,000 and 3,000 mm (Roderjan & Kunyoshi, 1988) and has seasonality.Lower rainfall occurs from the end of autumn to winter (April to August), and higher rainfall during the warmer Brazilian summer (September to March) (Instituto Paranaense de Desenvolvimento Econômico e Social, 2001).
Nematodes were sampled in four forest age stages (initial herbaceous, young arboreal, medium stage secondary forests, and old-growth forest, one site each, on sites no.32, 35, 38 and 41, in Table 1) on Cambisols; as well as in two agroforestry systems (banana and banana-palmito, three sites each, no.46-51, in Table 1) on Gleysols.The forest stages represent phases of the natural regeneration process, covering patches of forests differing in age (3 to >80 years) and thus anthropogenic infl uence.The forest regeneration sites were in the Itaqui reserve owned to and administrated by the regional NGO Society for Wildlife Research and Environmental Education (SPVS).The two agroforestry systems were located just outside, at a place called Rio Pequeno, about 20 km north of Antonina.Both were plantations, grown with banana or a mixture of banana and palmito (Euterpe edulis) plants.
Table 1.Characteristics of the 10 study sites in eastern Paraná, Brazil.Site numbers correspond to the Solobioma project sample sites (Römbke et al., 2009).Pesq. agropec. bras., Brasília, v.44, n.8, p.975-980, ago. 2009 At each site, fi ve plots were established, and ten samples (subsequently mixed and reduced to fi ve composite samples) were taken with a soil corer (2.5-cm diameter; 10-cm depth).Additionally, at the forest sites, litter samples were taken.Nematodes were extracted from soil through the modifi ed sieving and decanting method of Cobb (Cobb, 1918;S'Jacob & Van Bezooijen, 1984;Southey, 1986), and the ISO method (International Organization for Standardization, 2005).For the litter samples, Baermann funnel extraction (Baermann, 1917)  For the extraction of free-living nematodes from soil, according to Cobb's modifi ed decanting and sieving method, each composite sample was thoroughly mixed, and a subsample of 50 g fresh weight was steeped in 400 mL tap water for approximately 15 min.Then, soil and water were stirred for 60 s; and after 15 s of sedimentation, the supernatant was carefully decanted in a collecting plastic bowl.This procedure was repeated twice, each time with 400 mL tap water.The remaining sediment was discarded, and the combined nematode suspension was poured through a cascade of several sieves with decreasing mesh sizes (1,000, 350, 175, 100 and 45 µm).To improve cleaning, the nematode suspension was poured fi ve times through the 45 µm sieve.The nematodes on the 45 µm sieve were rinsed with tap water and transferred in a separate plastic bowl.The collected supernatant of the nematode suspension was carefully decanted.The remaining nematodes were poured into an extraction sieve containing a cotton-wool fi lter and placed in a shallow tray fi lled with tap water.Within the following 24 hours, the nematodes actively moved through the cotton-wool into the tray.This nematode suspension was poured through a small 20 µm sieve to separate the nematodes from the water.The nematodes were carefully rinsed from the 20 µm sieve into a counting dish; thereafter they were counted.
For the extraction of free-living nematodes from soil by the ISO method, a subsample of 50 g fresh weight of the mixed composite sample was also taken.Plastic bowls (diameter ca.20 cm, height ca. 10 cm) were carefully fi lled up with tap water until empty plastic sieves (diameter ca. 15 cam, mesh width ca.0.5 mm) were completely covered by the water.The samples were put into the sieves, and were, if necessary, carefully broken apart by hand.The bottom of the sieves did not reach the bottom of the bowls.At the end of the extraction period, after 24 hours, the sieves were removed and the soil was discarded.The water including the nematodes was slowly and carefully decanted from the bowls and poured through a small 20 µm sieve to separate the nematodes from the water and the sediment.Thereafter, the nematodes were counted.
For the extraction of free-living nematodes from litter, by the Baermann funnel method, litter samples (10 g fresh weight) were carefully cut into ca. 1 cm seized pieces.Thereafter, each sample was wrapped in a piece of cheesecloth, forming a loose ball.Funnels (slope of the sides approximately 45°) with a piece of soft silicone rubber tubes attached to the stem, closed with a squeezer clip, were placed in a funnel stand.The funnels were fi lled with tap water until it reached up to ca. 1 cm below the rim.The cheesecloth containing the sample was hanged in the funnel so that the sample was totally submerged, without touching the bottom of the funnel.After a period of 24 hours, the nematode suspension was tapped by opening the squeezer clip carefully, and the nematodes were counted at fi rst time.After further 24 hours, the nematode suspension was tapped again and the nematodes were counted a second time.
All nematodes were counted alive, under an inverse microscope (magnifi cation 160 x).The counted samples were transferred into 5 mL polypropylene tubes.For the taxonomical identifi cation of the nematodes, the samples were heated up to 60°C and conserved in formaldehyde (4%).Until identifi cation, the samples were stored in a refrigerator at 4-6°C.Identifi cation at least to family level (trophic groups) and further data assessment as maturity index (Bongers, 1990) are to be completed.

Results and Discussion
Cobb versus ISO extraction effi ciency were compared, and the mean number of individuals extracted from fi ve soil samples of medium secondary forest by Cobb's method was slightly but consistently higher than those extracted by the ISO method (Table 2).However, the Pesq. agropec. bras., Brasília, v.44, n.8, p.975-980, ago. 2009 difference was not statistically signifi cant.The coeffi cient of variation, as a measure of the variability between samples, did not differ between the extraction methods.
Afterwards, Cobb's method was compared to ISO's method, focusing on intra-sample variability.The coeffi cient of variation, for the mean number of nematodes of four subsamples of one particular soil sample of medium secondary forest, extracted by the Cobb method, was less than half as high as the one determined through the ISO method (Table 2).This difference was statistically signifi cant (t-test; two-sided, p<0.05).For both methods, it can be stated that intra-sample variability was clearly below inter-sample variability.
Since extraction effi ciency was higher, and intra-sample variability was signifi cantly lower, Cobb's sieving and decanting method was preferable to the ISO's method, and was used for all further extractions.However, due to the small sizes (diameter of 20-35 µm) of a great portion of the nematodes extracted from the Cambisols, the last extraction step (pouring the nematode suspension over a 20 µm gauze) should be repeated at least twice to retain more than 90% of the total number.Finally, the optimum extraction time of litter samples was determined.Nematodes of fi ve litter samples of the medium forest site were extracted using Baermann funnels and counted after 24 hours, and again after 48 hours.On average, only 10.8% of the nematodes found in the fi rst count (24 hours) were determined in the second count (48 hours).This result was confi rmed in a second group of fi ve litter samples, from the old growth forest site: the second count yielded 9.2% of the fi rst count.Therefore, for all further extractions, a total extraction time of 48 hours with two counts after 24 and 48 hours was regarded to be suffi cient to extract more than 90% of the total number of nematodes and, at the same time, to check the extraction effi ciency.
At the four forest sites, on average 511.2-781.3nematodes per 100 g soil dry weight were found.Their abundance did not differ signifi cantly between the four forest sites (Figure 1).In general, nematode abundance in litter was much higher than in soil.Between 7,687.5 and 29,194.0individuals per 100 g litter dry weight were found (Figure 2).Their mean number in litter increased with the age stage of the forest sampled.However, in the litter of the sites with intermediate age   Pesq.agropec.bras., Brasília, v.44, n.8, p.975-980, ago. 2009 (advanced and medium stage), no difference was found.Nevertheless, the difference in abundance between the young and old growth forests was statistically signifi cant (t-test; two-sided, p<0.05).
The mean nematode abundance in soil, at the two different agroforestry systems, differed markedly.The abundance was more than twice as high at the banana plantations, compared to the banana-palmito mixed stands (Figure 1).However, this difference was not statistically signifi cant, due to the high variability in nematode numbers, particularly in the three banana monoculture sites.The mean abundance determined for the banana-palmito mixed stands was similar to the ones of the four different forest sites.Litter could not be evaluated, since there was not enough material available on the agroforestry sites.
The nematode abundance, determined for the forest soils, was similar to the ones observed in literature (Peterson & Luxton, 1982;Bloemers et al., 1997).However, sampling techniques, sampling depth and extraction methods often vary in different projects and, sometimes, account for differing results (Mulder et al., 2005).Therefore, comparisons of nematode densities between studies should be made very cautiously.
The results from the forest sites (no difference in soil, but increasing numbers in the litter with increasing age) should be verifi ed by assessing at least further two sites of each stage.Even more importantly, the nematodes sampled so far should be identifi ed at least to family level, in order to determine the percentage of the trophic groups.This information should be used to calculate the maturity index (Bongers, 1990), and to clarify whether their diversity at the different sites differs or not.Finally, the usefulness of nematodes for the biological classifi cation and assessment of these soils must still be ascertained.

Conclusions
Cobb's modified sieving and decanting method is more appropriate than the ISO method, for the extraction of nematodes from tropic forest soils, presenting extraction efficiency higher than 90%.
2. The Baermann funnel method, with an extraction time of 48 hours, yields extraction efficiency higher than 90% and is suitable for the extraction of nematodes from litter of tropical forests.
3. The mean number of nematodes extracted from soil does not differ between the four forest sites, whereas the mean number of individuals extracted from litter increases with the age stage of the forest sites sampled.
4. Mean nematode abundance in soil in banana plantations was about twice as high compared to the banana-palmito mixed stands and to the forest sites.

Figure 2 .
Figure 2. Nematode number in litter (individuals per 100 g of litter dry weight), in forests of four different age stages: H, young; A, advanced; M, medium secondary forests; F, old-growth forest (one site each).Mean±standard deviation (n = 5).

Figure 1 .
Figure 1.Nematode number in soil (individuals per 100 g of soil dry weight), in forest of four different age stages: H, young; A, advanced; M, medium secondary forests; F, old-growth forest (one site each); and in two different agroforestry systems: B, banana; BP, banana-palmito (three sites each).Mean±standard deviation (n = 5).
was used.Since these methods were developed for temperate soils, additional experiments were conducted: a comparison of the extraction effi ciency of the Cobb and ISO method; a comparison of intra-sample variability of these methods; and a comparison between different extraction times for Baermann's method.

Table 2 .
Means, standard deviations (SD) and the coeffi cient of variation (CV) of nematode number (individuals 100 g -1 soil dry weight) extracted from a medium secondary forest soil by Cobb's or ISO method.