Earthworms occur all over the world in soils where there is adequate moisture to support them, except in arid sites such as deserts. In these soils, populations of earthworms are extremely variable, ranging from only a few individuals to more than 2000 per m2 so that the assessment of the size distribution and structure of earthworm populations and communities can be difficult. A particular complication is the seasonal changes in the numbers, demography, and vertical distribution of the populations especially in temperate regions and some parts of the tropics, so that for comparisons of earthworm communities, samples need to be taken at comparable times, using the same sampling methods.

5.1 Sampling Earthworm Population

To estimate earthworm population size, some method of quantitative sampling of earthworms in small defined sample areas is necessary. Sampling earthworm populations accurately can be extremely difficult because they are aggregated horizontally, and some have deep and relatively permanent burrows. Their distribution is vertically stratified, and it may be very difficult to separate them from many soils, and their life cycles are complex. Most of the methods that are used to sample earthworms have some shortcomings, and most estimates of field populations in the literature are likely to be underestimated because of the relative inefficiencies of the different sampling methods.

Earthworm numbers in soil samples can be assessed mechanically by counting, or the earthworms may be stimulated to move out of the soil in the field or from soil samples in the laboratory. Probably, a combination of these two approaches gives the best results. The various methods of assessing earthworm populations in the field and extracting earthworms from soil samples have been reviewed by Bouché (1972), Bouché and Gardner (1984), Lee (1985), Edwards (1991), Schmidt (2001a), Jiménez et al. (2006a), Rombke et al. (2006), Hendrix and van Vliet (2011), and Valckx et al. (2011). Population assessments can be divided into passive methods, where the earthworms are separated physically from soil onto the surface of the soil, and behavioral methods, where they are stimulated to emerge by physical or chemical stimuli.

5.1.1 Handsorting

Most early population studies involved digging up soil samples and sorting these by hand (Stockli, 1928). To enable accurate population estimates to be made, techniques were developed using cores or quadrats of soil of exact areas and depths [e.g., Bornebusch (1930), Ford (1935), Hopp (1947), Low (1955), Reynoldson (1955), Svendsen (1955), Wilcke (1955), Barley (1959a, b), van Rhee and Nathans (1961), El-Duweini and Ghabbour (1965)]. Zicsi (1958) noted that numbers of earthworms estimated per m2 decreased with increasing sample size and suggested that samples of 25 cm × 25 cm be used. In later work, Zicsi (1962) compared the efficiency of estimating populations of earthworms by handsorting samples of sizes 0.06, 0.25, 0.5, and 1.0 m2, taken with a square sampling tool. He concluded that 16 sample units per ha taken to a depth of 20 cm gave a good estimate of populations of medium-sized species. For larger worms and deeper burrowing species, he concluded that a larger area and deeper soil sample were required. Reinecke and Ljungström (1969) used four samples of 25 cm × 25 cm to sample earthworm populations in a South African pasture, and Rundgren (1975) four samples of 35 cm × 35 cm from Swedish soils.

Small earthworms may be missed by handsorting. Persson and Lohm (1977) suggested that individuals less than 200 mg live weight could often be overlooked. They suggested sampling to a depth of 30 cm in summer and a 60 cm depth in winter. Axelsson et al. (1971) suggested that specimens weighing less than 160 mg were often missed, and Reynolds (1973b) that earthworms less than 2 cm in length might not be counted because of their low visibility. Jiménez et al. (2006a) examined handsorting efficiency for earthworms sampled in tropical savannas and pastures in Colombia. They reported strong seasonal trends (Fig. 5.1) lowest efficiencies for all species (10−30%) occurring during the dry season when soils were difficult to sample. Monthly average efficiencies ranged from 30% to 44% for the smallest ocnerodrilid species to 80−100% for the epigeic Aymara spp. Nonlinear regressions of earthworm weight versus handsorting efficiency showed significant trends of increasing efficiency with increasing weight (Fig. 5.2) and provided correction factors appropriate for each earthworm species or group under study.

Fig. 5.1
figure 1

Seasonality of global efficiency of handsorting compared to washing and sieving of soil monoliths for Glossodrilus sp. (top) and Ocnerodrilidae sp. (bottom) in tropical savannas and pastures in Colombia. (Jiménez et al., 2006a)

Full size image
Fig. 5.2
figure 2

Efficiency of the handsorting method according to earthworm weight for the three species studied in both systems. Correlations were adjusted to the function y = axb and tested with ANOVA. ■ Glossodrilus sp. (y = 123.79 × 0.244, P < 0.05), ■ Aymara sp. (y = 305.18 × 0.607, P < 0.001), and ▲ Ocnerodrilidae sp. (y = 157.09 × 0.429, P < 0.05). The weight of Ocnerodrilidae was multiplied by 10 in order to plot the data in the same graph. (Jiménez et al., 2006a)

Full size image

Lavelle et al. (1999) reviewed a range of populations of earthworms in tropical agroecosystems. Lavelle (1978) sampled a mixed population of megascolecids and eudrilids in the Ivory Coast and recommended handsorting the top 50 cm of soil in 10 cm layers from 1 m2 quadrats, supplemented by washing and sieving 20 × 20 cm samples to recover smaller individuals and cocoons. Nelson and Satchell (1962) tested how many earthworms could be recovered from soil by handsorting, by introducing known numbers of worms into soil. They found that the smaller worms and dark-colored worms were often missed, and their numbers were consequently underestimated; when 924 worms were introduced to soil, 93% of all earthworms were recovered by handsorting, but only 80% of immature A. chlorotica and 74% of immature L. castaneus were found. They concluded that handsorting was satisfactory only for individuals of more than 0.2 g live weight. There were considerable differences in efficiency between individual sorters. Handsorting is particularly relevant when assessing populations of earthworm species that live close to the soil surface.

A number of factors influence efficiency of handsorting for earthworms, including species and body size of earthworms (i.e., seasonal phenology and population demography), root density (especially in grasslands), soil type, and a “human factor,” related to training of personnel and time spent on each sample (Schmidt, 2001a; Jiménez et al., 2006a).

5.1.2 Soil Washing

Cocoons and small earthworms are not recovered easily by handsorting or chemical stimulation. The only method that seems appropriate for these stages is a combination of washing and sieving soil samples, possibly followed by a flotation stage. For instance, Morris (1922) and Ladell (1936) used a method of washing soil samples with a jet of water through a series of sieves. Raw (1960b) handsorted samples from rather poor pasture soils and then washed the soil away from the same samples in a sieve of mesh 2 mm nested within another 0.5 mm mesh sieve standing in a bowl of water. The sieves were then immersed in a magnesium sulfate solution of specific gravity 1.2, and the earthworms that floated to the surface collected. Only 52% (84% of the weight) of the total earthworm populations collected were obtained by handsorting, and a further 48% were recovered by subsequent washing. From a heavy, poorly structured arable soil, 59% (90% by weight) and, from a light, well-drained soil, 89% (95% by weight) of the total numbers of earthworms in the soil were recovered by handsorting. Obviously, washing is more efficient than handsorting, and it also recovers cocoons, but the washing method takes much longer and often tends to damage the earthworms mechanically. A mechanized soil washing method, which involves rotating the containers in which the sieves stand, was much faster, less laborious, and suitable for most soils. Flotation after washing and sieving can improve the efficiency of separating earthworms from soil (Raw, 1960b; Gerard, 1967; Martin, 1976).

Walther and Snider (1984) reported that both cocoons and earthworms could be recovered efficiently from humus and soil by a combination of washing through sieves and handsorting. In this way, 97.7% of cocoons and 96.7% of earthworm biomass in the soil were recovered. Bouché and Beugnot (1972) used a modified washing method which involved soaking soil samples in 2% sodium hexametaphosphate to disperse clay particles, soaking in 4% formaldehyde to kill and fix earthworms, passing the mixture through a series of sieves and handsorting the earthworms in the last sieve. Judas (1988) described an apparatus that extracted earthworms efficiently from broad-leaved litter by washing and sieving. Washing/sieving methods are clearly valuable tools in estimating earthworm populations but tend to be time-consuming and laborious; however, they are essential in estimating the numbers of cocoons and very small, estivating, or dormant earthworms.

5.1.3 Electrical Methods

For many years, fishermen have obtained earthworms from soil for fish bait by attaching one lead of an AC electrical mains to a copper wire attached to a nonconducting handle and inserting it into the soil. Walton (1933) and Johnstone-Wallace (1937) first reported that such a technique could be used for sampling earthworm populations, and Doeksen (1950), who experimented further, suggested that a steel rod 8–10 mm in diameter and 75 cm long with an insulated handle was suitable as an electrode. He used 220–240 V at 3–5 A, and the strength of current was regulated either by a variable resistance or by inserting the electrode deeper into the soil. One to three electrodes could be used simultaneously.

The conductivity of the soil, and hence the current passing through it, depends on its moisture content and pH, but usually the current penetrates deep into the soil, bringing earthworms up from deep burrows; however, if the surface soil is dry, it may drive the earthworms downward instead. This possibility can be minimized by insulating the entire electrode inserted into the soil except its point. Earthworms usually emerge at distances between 20 cm and 1 m from the electrodes, but there is some danger that earthworms close to the electrode may be killed by the current. Nevertheless, this method does seem to be effective in sampling for deep-living earthworms in the field.

Satchell (1955) also used an electrical method with a 2 kVA generator that led to a water-cooled electrode inserted 46 cm into the soil, with a voltage of 360 V applied. The portable generator made the method more suitable for sampling earthworms in the field. Satchell considered that the main defect of the method was that of defining the exact limits of the volume of soil from which earthworms were recovered.

Edwards and Lofty (1975) described an electrical sampling method that consisted of inserting two electrodes, in the shape of forks with prongs 50 cm long, into the soil 1 m apart and passing the current from a 250 V diesel generator between the electrodes. The method was most effective when the current was kept to about 2–4 A. They found that another factor affecting the efficiency of the method was the pH of the soil through which the current was passed, more earthworms being recovered from soil with a low pH than from the same soil that was less acid.

Rushton and Luff (1984) used a circular electrode configuration. This was more efficient at extracting shallow-living than deep-burrowing species, and juveniles over adults, and extraction efficiency was correlated strongly with soil moisture content. Theilemann (1986) described an “octet” method in which the electrodes were arranged octagonally and current was alternated between opposing pairs of electrodes over time. Schmidt (2001b) used commercially built octet devices to sample earthworm populations in agricultural fields in Ireland and compared general population trends derived from electrical extraction, handsorting, and formalin extraction. The octet method yielded significantly higher earthworm numbers compared to formalin, although biomass estimates were not different. The octet and handsorting methods gave comparable earthworm numbers and biomass when soil temperature and moisture conditions were favorable; electrical extraction was less efficient than handsorting after soil was plowed or following drought, indicating the sensitivity of the octet method to soil conditions. Species composition of earthworm communities sampled by electrical and handsorting methods were similar for adults, but handsorting was more effective for the small Murchieona minuscule, whereas the octet method was more effective for Lumbricus festivus and L. terrestris in wheat-clover (Table 5.1). Eisenhauer et al. (2008) compared the octet method with mustard extraction for various earthworm groups in a grassland in Germany. They found that mustard was more effective on anecic earthworms, whereas both methods yielded low numbers of endogeic species, probably because of dry soil conditions. They concluded that the octet method did not adequately reflect the earthworm community composition, even after addition of water to the soil.

Table 5.1 Earthworm species composition (% of total numbers or biomass extracted)a in the conventional wheat and wheat-clover field at Lyons as assessed using the electrical octet method or soil sorting, averaged over all sampling dates from November 1995 to October 1997 (Schmidt, 2001)
Full size table

Despite more sources of variation than handsorting or chemical extraction, under favorable conditions, electrical extraction appears to be a viable earthworm sampling method, particularly in situations where soil disturbance must be minimized or where live earthworms uncontaminated by chemicals should be collected. Site-specific validation of the technique should be conducted for quantitative studies. The various methods of earthworm population sampling as aforementioned are still being explored to this day. da Silva et al. (2019) studied the populations in forestry plantations of Araucaria angustifolia and Pinus elliottii and native Atlantic forest in Southern Brazil extracted using handsorting and formalin in these three forest ecosystems. It was found that handsorting was a more efficient method of extraction than formalin extraction, particularly for the extraction of endogeic species. They also reported that formalin extraction collected primarily epi-endogeic species (Amynthas, Metaphire, Urobenus). The comparative results of these studies could provide a framework by which the sampling of earthworms may be optimized depending on the species or origin of the earthworm one intends to collect.

5.1.4 Chemical Methods

The first chemical extractant used to sample earthworm populations was mercuric chloride solution (1.7–2.3 liters of solution containing 15 mg HgCl2 in 18.25 liters of water) (Eaton & Chandler, 1942). Evans and Guild (1947a) used a potassium permanganate solution to bring earthworms to the soil surface (1.5 g per liter at a rate of 6.8 liters per m2) and later used this method in their population studies (Evans & Guild, 1947c; Guild, 1948, 1952). Jefferson (1955) used a solution of mowrah meal (the material remaining after oil is extracted from ground seeds of the Bassia tree, Bassia longifolia). Raw (1959) reported that a 0.55% formalin solution (25 mL of 40% formalin in 4.56 liters water applied to 0.36 m2 of soil surface) was very effective in bringing most species of earthworms to the surface. This may be because dilute formalin is less toxic to earthworms than potassium permanganate, which often kills earthworms before they reach the surface. The main disadvantage of these chemical methods is that they do not recover all species equally efficiently; those species with wide and deep burrows come to the surface much more readily than the species without such burrows (Bouché and Gardner, 1984). Baker (1983) reported that of 15 lumbricid species in a peat soil in Ireland, L. terrestris, D. octaedra, and Dendrobaena mammalis were sampled relatively efficiently by the formalin method. However, the formalin method seems to be relatively inefficient for sampling nonlumbricid species such as Diplocardia and Pheretima (Reynolds, 1976). Satchell (1969) also used the formalin method but recommended that much more solution should be used (three applications of 9 liters of 0.165–0.55% formalin per 0.5 m2). He pointed out that the efficiency of the method was seasonal, since both soil temperature and soil moisture content affect the number of earthworms coming to the soil surface, and he worked out a correction factor based on a regression analysis which would correct for the soil temperature at the time of sampling (Satchell, 1963), which he later modified (Lakhani & Satchell, 1970). However, numbers could not be corrected to account for earthworms that were aestivating and hence did not respond to the chemical. Chloroacetophenone was used to sample L. terrestris populations by Daniel et al. (1992). They reported that the efficiency of the method was improved by repeated application of the chemical on three successive evenings. Walther and Snider (1984) reported that 96% of lumbricids could be recovered from litter using 1 h of immersion in dilute (0.25%) formalin solution.

Because formalin can have negative impacts on plants and other soil biota, as well as deleterious effects on human health (Eichinger et al., 2007), alternative extractants have been considered in recent years. In particular, aqueous extract of mustard has shown earthworm extraction efficiency similar to that of other chemical extractants and has come into favor because of its minimal effects on human health and low phytotoxicity compared to formalin (e.g., Gunn, 1992: Chan & Munro, 2001; Zaborski, 2003; Leroy et al., 2008; Eisenhauer et al., 2008; Valckx et al., 2011). The efficacy of other eco-friendly extraction methods has also been explored, such as those which incorporate Allium cepa solution. A study carried out by Singh et al. (2018) compared formalin, AITC (Allyl isothiocyanate), and A. cepa found that compared to formalin and AITC, A. cepa solution was more efficient in extracting earthworms even across adult and juvenile earthworms. A. cepa solution expelled 66.15% and 53.30% more earthworms than AITC and formalin, respectively, showing that A. cepa solution exists a viable and eco-friendly alternative for the sampling of endogeic and anecic earthworms compared to common, more environmentally consequential methods. The effectiveness of any chemical extractant varies with earthworm species and activity, temperature, soil porosity, and soil water content, saturated soils being less likely to transmit extractant solutions deep into the soil. Comparisons with handsorting should be done before adopting chemical extraction techniques for quantitative population sampling.

5.1.5 Heat Extraction

Earthworms can be stimulated to emerge from intact soil samples by application of moderate heat, and this method, which has been little used in practice, may be useful in obtaining small surface-living species from matted turf. It involves using a container (55 × 45 cm) with a wire sieve 5 cm from its bottom. Soil quadrats cut from turf (20 × 20 × 10 cm deep) are placed on the sieve, immersed in water with 14 60 W light bulbs suspended above, and left for 3 h, after which earthworms can be collected from the bottom of the container, where they are driven by the temperature gradient.

5.1.6 Vibration Methods

In the United States, mechanical stimulation by vibration has been used commonly to collect species of acanthodrilids such as Diplocardia mississippiensis and Diplocardia floridana and the megascolecid Pheretima diffringens particularly in the southern United States. This technique (referred to locally as “grunting”) often takes the form of vibrating with a bow, a flexible rod, or wooden stake that has been inserted into the soil, producing the vibrations that pass through soil and bring the earthworms to the soil surface. Few experimental studies have been conducted using vibration sampling; Hendrix et al. (1994) used an equal-effort vibration technique to compare Diplocardia spp. populations in forest savannas in north Florida. In the same forests, Mitra et al. (2008) used geophones buried in linear and equidistant arrays around vibration sampling points to study amplitude decay and frequency composition of vibrations created by earthworm “grunting.” They found that seismic signals were broadband, with energy concentrated below 500 Hz (dominant frequency of 97.3 ± 11.7 Hz, n = 15) in the soil where earthworms surfaced. Numbers of earthworms emerging were correlated positively with signal strength (Fig. 5.3). Earthworm emergence in response to vibrations appears likely to be a reaction to subterranean predators such as moles (Catania, 2008; Mitra et al., 2009).

Fig. 5.3
figure 3

Distribution of recordings and worm emergence sites. (a) A map indicating the locations of surfaced worms (circles) during a single instance of grunting, geophones in a linear array (squares), and the wooden stake (star). The worm grunter is facing the linear array. Scale bar, 1 m. (b) Relative amplitude decay of vibrations recorded from a linear geophone array ( yZ42.998 eK0.3066x, R2Z0.9406). The inset shows the waveforms of four grunts recorded at 1.83 and 7.32 m. (c) The number of worms surfacing as a function of the distance from the stake. Data are plotted using sliding window comparisons (using 100 cm windows with 50 cm increments). (Mitra et al., 2009)

Full size image

In general, vibration techniques may not be effective for some taxonomic groups of earthworms (e.g., lumbricids, [Reynolds 1973a]) and are probably not suited to accurate quantitative measurements of populations. However, they may be useful for selective or comparative sampling of certain earthworm populations.

5.1.7 Counting Earthworm Casts

Some species of earthworms such as A. longa and Hyperiodrilus africanus have very distinctive surface casts, and many other species, such as A. caliginosa, have readily distinguishable surface casts (Chap. 10). Counting these casts cannot give valid assessments of earthworm populations, although Evans and Guild (1947b) reported that they obtained a close correlation between the numbers of casts deposited on the soil surface and the numbers of A. longa and A. caliginosa extracted from the soil under the casts. However, the presence of casts, which is seasonal and influenced by temperature and moisture, can be used as an easily applied index of earthworm activity, which in turn is a useful parameter of earthworm populations.

5.1.8 Mark and Recapture of Earthworms

The tagging of soil-inhabiting invertebrates with some kind of marker, releasing them into soil and then extracting them from soil samples, has been used as a method of estimating the size of many invertebrate populations. A few workers have used this method for assessing earthworm populations. Lowe et al. (2017) utilized visible implant elastomer (VIE) tags with the intent of exploring a simple in situ method of containment and retrieval, providing scope for further improvement of the tagging procedure, mortality assessment, and investigation of earthworm containment. Meinhardt (1976) stained earthworms with a water-soluble nontoxic green dye, which could stain the bodies of earthworms for several months, before returning them to the soil. Mazaud and Bouché (1980) used the same method to study earthworm dispersal rates and mortality in a pasture soil. Other workers have used radioactive isotopes such as 198Au to label L. terrestris (Joyner & Harmon, 1961), and Gerard (1963) inserted small pieces of 182Ta wire into the coelom of individuals of L. terrestris to label them and trace their movements. Bastardie et al. (2003) labeled L. terrestris and Nicodrilus giardi with coelomic injections of 60Co and were able to trace burrowing activity in artificial soil cores using scintillation detectors placed at angles around the cores. González et al. (2006) evaluated the use of a fluorescent elastomer injected into Pontoscolex corethrurus and tracked the marker for 4 months in populations held in field enclosures. Other methods have been detailed which may help the study of earthworm dispersal, a particularly challenging phenomenon to quantify.

5.1.9 Comparisons of Sampling Methods

A number of workers have compared the relative efficiency of extracting earthworms from soil by two or more methods. Svendsen (1955) reported that handsorting was much more efficient than using potassium permanganate extraction. Raw (1959) compared the use of formalin with that of potassium permanganate and handsorting. From one arable orchard, he obtained 59.7 earthworms from 0.36 m2 with formalin, 32.5 with potassium permanganate, and 47.5 by handsorting. Comparable figures for a grass orchard were 165.1 for formalin, 83.9 for potassium permanganate, and 280.0 for handsorting. Bouché (1969) compared handsorting with first applying formalin, then handsorting again to find earthworms that had not been extracted by the chemical. He reported that 55.4 earthworms per m2 were extracted with formalin and a further 273.4 per m2 by handsorting soil from the same area. This is an excellent way of improving the overall efficiency of the formalin method. Alternatively, handsorting can be done first to a depth of 10 cm and then formalin applied to the bottom of the handsorting pit (Bohlen et al., 1995a, b, c). Svendsen (1955) compared handsorting with potassium permanganate expulsion directly and obtained much better results with handsorting. Baker (1983) concluded that of 15 species in a reclaimed peat soil in Ireland, L. terrestris, D. octaedra, and Satchellius mammalis were those most efficiently sampled by the formalin method, whereas the other species, especially Aporrectodea spp., were sampled better by handsorting.

Bouché (1975) compared the efficiency of six methods of sampling earthworm populations: digging, handsorting, soil washing, sieving, flotation, and extraction with potassium permanganate or formalin. He stressed the importance of using behavioral characteristics in improving efficiency and calculated correction factors for the various methods in the soil he sampled. He compared three methods in a later study (Bouché & Gardner, 1984). Jiménez et al. (2006a) compared the relative efficiencies of handsorting in the field and wet sieving in the lab. They found consistently higher earthworm populations densities over all weight classes with the wet sieving method and calculated correction factors which could be applied to field collected samples.

Daniel et al. (1992) reported that populations of L. terrestris could be sampled equally efficiently by first using dilute formaldehyde or chloroacetophenone solutions and then handsorting to a depth of 110 cm. They improved the efficiency of the chemical methods by repeated application of the chemicals on three successive evenings. Gunn (1992) compared the use of mustard for estimating earthworm populations with using formalin, potassium permanganate, or household detergent and reported it to be better than formalin and as good as potassium permanganate. Barnes and Ellis (1979) concluded that formalin extraction worked better in direct drilled (no-till) soils than in cultivated soils. Callaham and Hendrix (1997) compared formalin and handsorting methods and reported that when small juvenile earthworms were most abundant (November-May), handsorting was superior to formalin extraction; for the remainder of the year, the two techniques showed similar efficiencies. As in other studies, formalin extraction was more effective than handsorting for the collection of anecic species (i.e., Lumbricus terrestris). Valckx et al. (2011) concluded that application of mustard suspensions of 3 g L−1 followed by two applications of 6 g L−1 was as effective as formalin or allyl isothiocyanate (the active ingredient in mustard used by Zaborski, 2003) in extracting earthworms from agricultural soil in Belgium. It should be noted that allyl isothiocyanate has a high mammalian toxicity. They found no significant difference in efficiency among extractants when combined with handsorting for total earthworm abundance, biomass, or species richness (Fig. 5.4).

Fig. 5.4
figure 4

Recovered earthworm numbers (m−2 _ standard error) (a), biomass (g m−2 _ standard error) (b), and species number (c) using extraction with allyl isothiocyanate (AITC), formalin, or mustard combined with handsorting (soil) sampling methods. Results are given per sample fraction and in total. (Valckx et al. 2011)

Full size image

Andriuzzi et al. (2017) used two methods to investigate earthworm community responses to land use. Higher abundance was obtained using the handsorting method than the formalin method, but the two methods still led to the same ecological conclusions, that earthworm biomass and density decreased with agricultural intensification and similar land use effects on earthworm ecological group proportions, age structure, and body size distribution, pointing to a relative loss of large-bodied earthworms with agricultural intensification, were observed.

There seems little doubt that handsorting or soil washing gives the best results for most species of earthworms and should be the standard against which other methods are compared or calibrated for routine use. However, handsorting is very time-consuming and does not work well for deep-burrowing species such as L. terrestris. Therefore, chemical extraction (mustard or formalin) seems the best compromise for species with deep burrows and can be combined with handsorting or electrical extraction of surface soils for maximum efficiency.

5.1.10 Number and Size of Samples for Population Assessment

Many different sizes of soil samples have been used to estimate earthworm populations, ranging from soil cores 20 cm in diameter to 50 × 50 cm and 1 m2 quadrats. Zicsi (1962) reported that the number of earthworms he recovered per m2 by handsorting decreased with the increasing size of the sample. The minimum sample area required depends very much on the density of a population in a particular site. For most purposes, 0.5 or 0.25 m2 quadrats seem to be a suitable size and are easy to handle. The optimum number of samples can be determined by pre-sampling and then basing the number of samples on the degree of precision needed for the estimate.

5.2 Size of Earthworm Populations

Earthworm populations can be expressed either in terms of numbers or weight (biomass) per unit area. The use of numbers is sometimes misleading because it does not differentiate between very small and large individuals or species, which have very different influences on soil processes. Biomass is often a preferable parameter but can also lead to misleading conclusions. It is probably best to express populations both as numbers and biomass, and most researchers on earthworms report their populations in this way. Reporting biomass on the basis of ash-free dry mass, determined by ashing earthworms in a muffle furnace at 400–500 °C, corrects for the mass of soil in the guts of sampled earthworms, allowing for more accurate comparisons between different investigations. This method is essential for estimating secondary productivity of earthworm tissue.

5.2.1 Numbers in Earthworm Populations

Relatively few workers have sampled the earthworm populations in a variety of habitats at the same time of year, so it is difficult to assess precisely which habitats can support the largest earthworm populations. Some typical populations that have been estimated by handsorting or formalin extraction are summarized in Table 5.2. These estimates are all very approximate because of large variations in efficiency of extraction and seasonal changes in numbers of earthworms. For instance, populations can range from less than 1 earthworm to more than 2000 per m2 (0.5–305 g per m2). There are almost always fewer earthworms in acid, mor soils, fallow soils, and moorlands than in good mull soils. The numbers of earthworms in regularly cultivated arable soils are usually very variable, and populations are intermediate in size between the more sterile habitats and those in pastures and natural grasslands which can support large numbers of earthworms. The populations in coniferous forests tend to be smaller and those in deciduous temperate forests and tropical forests rather larger than those in arable land. Clearly, there is great variability in earthworm populations, although grassland seems able to support larger populations than most other habitats, presumably due to the availability of large quantities of organic matter.

5.2.2 Biomass in Earthworm Populations

It is often difficult to obtain the live weight of large numbers of earthworms directly, since they have to be weighed soon after collection. Furthermore, specimens from field collections contain different amounts of gut contents which affect biomass measurements, and therefore it is necessary to account for this material or remove it either by allowing live earthworms to void their guts (e.g., overnight in moist paper towels) or by dissecting and removing gut contents from preserved specimens. “Fresh weight” or “wet weight” measurements face two further problems: Earthworms extracted under field conditions contain varying amount of water, and preserved specimens often lose biomass in preservative fluids (e.g., loss of mucus secretions in formalin; Lee, 1985). These factors must be considered in comparative studies across sites or seasons, or when different preservations techniques are used. For most purposes, it is best to convert live weight into dry weight, which may be accomplished by directly drying gut-voided specimens (e.g., freeze drying) or by using allometric relationships or regressions which relate body length or wet weight to dry weight (Lee, 1985; Hale et al., 2004). Satchell (1969) described a method for calculating the live weight of earthworms that had been kept in 10% formalin solution. He plotted a regression of the live weights of earthworms against their weights after being kept in 10% formalin, then oven-dried them at 105 °C, and reweighed them. He obtained the expression: 1 g dry weight = 5.5 g live weight. The gut contents of an earthworm may be as much as 20% of its total live weight, so this must be accounted for when estimates of population biomass are made. Collins (1992) calculated a regression model which related earthworm length to dry weight for lumbricids, using specimens from northern Wisconsin forests. Dynamics of earthworm population biomass are also affected by factors of the environment and serve as a great metric to determine the health of a soil ecosystem as is the case in Abail and Whalen’s study (2018) about the way earthworm populations increase when more crop residues are left upon the surface of soil, also contributing to individual earthworm biomass.

Many studies (e.g., isotopic studies of earthworm feeding ecology) require the measurement of carbon mass while also maintaining specimens in a condition suitable for taxonomic identification. Live specimens can be killed instantly in boiling water and divided into a posterior half for gut clearance and freeze drying for chemical analysis, and an anterior half for preservation in formalin or ethanol for later taxonomic study (e.g., Spain & Le Feuvre, 1996; Hendrix et al., 1999; Briones & Bol, 2003).

5.3 Population Structure: Age Distribution

Populations of most soil-inhabiting invertebrates tend to have a pyramidal age structure, with many more young individuals than mature ones at most times of the year; for instance, Raw (1962) reported that the proportions of L. terrestris individuals of different ages in his samples from an orchard were in the ratio of 8 mature earthworms to 13 large immatures and 31 small immatures. Not much information is available about such seasonal changes in age structure of populations of earthworms.

Evans and Guild (1948a) sampled 12 fields at Rothamsted in the United Kingdom for earthworm populations and compared the numbers of immature and adult earthworms of different species (Table 5.3). Clearly, these proportions are never in a fixed ratio and depend very much on the time of year when a population is sampled, so that after active breeding periods, the proportion of immature earthworms will be greatest, and at all other times relatively small. However, there is some indication that the structure of populations of different species may differ considerably. Seasonal changes in proportions of adult and immature worms of A. nocturna and A. caliginosa were reported by van Rhee (1967) and by Gerard (1960) for A. chlorotica (Fig. 5.5). Van Rhee (1967) compared the population structure of L. terrestris, L. castaneus, A. rosea, A. caliginosa, and A. chlorotica in four orchard soils in five successive years, and for all species, except L. castaneus, the immatures greatly outnumbered the adults.

Table 5.2 Numbers and weights of earthworms in different habitats
Full size table
Fig. 5.5
figure 5

Seasonal abundance of A. chlorotica in “Pastures” at Rothamsted 1958–1960. (Gerard, 1960)

Full size image

McCredie et al. (1992) reported that populations of A. trapezoids in Australia increased from 58 per m2 at the time of the beginning seasonal rains to 170 per m2 (88.6 g live weight per m2) after 10.5 weeks. Near the end of the wet season (in October), the population density was 37 per m2. At the time when the rains began, the population consisted of juvenile and semi-mature individuals. Clitellate earthworms were found 1 month later and predominated from August (10.5 weeks) to the end of the season. Cocoons occurred from August to October, and those incubated in the laboratory at 16 °C hatched on average after 42 days and produced about two juveniles each. Juvenile and immature earthworms collected from a quiescent state at the end of summer, matured within 1 month when reared in moist soil in the laboratory. Seasonal changes in population age structure have been reported for a number of lumbricid species in temperate regions (see Sect. 5.5). Aira et al. (2006) found that substrate C:N ratios significantly influenced age structure of E. fetida populations in vermicompost systems. Low C:N systems contained mostly (60%) mature earthworms with higher mean weight compared to high C:N systems which contained sevenfold higher population density consisting of 70% juveniles and hatchlings. These results suggest that there is greater allocation of energy to growth than to reproduction in substrates with low C:N and confirm previous studies that show the influence of dietary quality on earthworm population age structure (Domínguez et al., 1997).

There have been fewer studies of the age structure of non-lumbricid earthworm populations. In a thorough analysis of Millsonia anomala populations in savannas of western Africa, Lavelle (1971, 1974, 1978) reported a clear pyramid of age classes, with 25% of the population being 1–2 months old; 48.3%, 6–8 months; 18.3%, 11–12 months; 6.7%, 16–19 months; and 1.7% older than 24 months (Lavelle, 1971). Analyses of the age structure of other species are also reported in Lavelle (1978). Much of this is summarized in Lavelle et al. (1999). For the species Pheretima hupeiensis, in the northwestern United States, mature individuals were abundant in August, but many died after reproduction, and immatures were more numerous in September, although by November, the entire population became immature and remained so throughout the winter months (Grant, 1956).

Studies by Callaham et al. (2003) on the related megascolecid, Amynthas agrestis, in the southern Appalachian Mountains showed that populations in July and early August consisted entirely of small juveniles, which peaked in abundance in late August; in September, total numbers declined, but the abundance of adults peaked. No juveniles or subadults were collected after September, supporting the hypothesis that populations of these epigeic overwinter as cocoons (Uchida, 2004).

5.4 Population Structure: Spatial Distribution

Earthworms are not distributed randomly in soil. As with many other soil-inhabiting organisms, their species composition and population densities vary across landscapes and with soil depth, and they show varying demographic patterns over time (Rossi, 2003; Barot et al., 2007). Identifying spatial and temporal distribution patterns and the factors that control them has long been an active area of research in earthworm ecology (Lee, 1985).

5.4.1 Horizontal Distributions

Early studies by Guild (1952) and Murchie (1958) classified the possible factors that were likely to be responsible for variability in horizontal distributions of earthworms as:

  1. (i)

    Physicochemical (soil temperature, moisture, pH, inorganic salts, aeration, and texture)

  2. (ii)

    Availability of food (herbage, leaf litter, dung, consolidated organic matter)

  3. (iii)

    Reproductive potential and dispersive powers of the species. To which we add historical factors (including disturbance and colonization of new habitats)

Murchie (1958) concluded that no single one of these factors was likely to be solely responsible for the horizontal distribution, but rather the interaction of several or all of the factors, thereby providing suitable soil conditions for earthworms in some areas. Numerous studies have since shown the importance of variations in edaphic and microhabitat conditions on earthworm distributions (e.g., Nordström & Rundgren, 1974; Phillipson et al., 1976; Lavelle, 1983; Poier & Richter, 1992; Rossi et al., 1997, 2006; Nuutinen et al., 1998; Decaëns & Rossi, 2001; Hernández et al., 2002; Rossi, 2003).

While soil moisture may be the major physicochemical factor affecting local population densities of earthworms (Curry, 2004), several workers have also correlated population aggregations with available food supplies. For instance, in pasture, populations of D. octaedra and L. rubellus were aggregated significantly beneath dung-pats in spring (Boyd, 1957b, 1958), and pigmented species such as Lumbricus festivus, L. rubellus, L. castaneus, D. rubida, D. octaedra, and B. eiseni were shown to aggregate under dung-pats more than nonpigmented Allolobophora and Aporrectodea species (Svendsen, 1957). If they react so readily to manure, then it seems probable that variations in the distribution of other forms of organic matter may also influence earthworm distribution, e.g., total soil carbon (Hendrix et al., 1992), litter distribution (Curry, 2004), or carbon in coarse soil fractions (Rossi et al., 2006).

Earthworm population densities are often correlated with particular soil characteristics, but only a small proportion of their spatial variation may be explained by environmental heterogeneity (Barot et al., 2007). Satchell (1955) showed that L. castaneus and A. rosea were very aggregated in a relatively uniform pasture which had not been grazed. He suggested that aggregations might occur when earthworms are reproducing more rapidly than the offspring can disperse from the breeding site.

Satchell (1955) calculated indices of dispersion for adults and immatures which showed that the adults were nearly randomly dispersed but the immature were aggregated significantly. Hence, on this assumption, a species with distinct seasonal changes in numbers can be expected to pass from a very aggregated phase in the breeding season in early summer to a more randomly distributed phase in winter (Satchell, 1955). Barois et al. (1999) analyzed demographic and ecological characteristics of 26 tropical earthworm species and found that most species had aggregated spatial distributions. They concluded that intrinsic properties of high fecundity and small body size were associated with population aggregation, whereas species with low fecundity and large size were distributed more randomly.

Rossi (2003a, b) conducted a series of spatial statistical analyses of earthworm population distributions in a tropical savanna at Lamto, Côte d’Ivoire in western Africa. Earthworms were aggregated spatially, and the clusters (both patches and gaps) were distributed randomly across the 2025 m2 study area (Fig. 5.6). Distributions of the compacting species, Milsonia anomala were distinct from those of the decompacting eudrilid species, Chuniodrilus zielae and Stuhlmannia porofera. These distributions were relatively stable over the 2-year period of study and were correlated strongly with soil bulk density. Similar studies in pasture and native savanna “Llanos” in Colombia showed that Glossodrilus spp. and Andiodrilus spp. had opposite spatial distribution patterns, suggesting spatial exclusion between these two endogeic species, probably as a result of competitive interactions (Jiménez & Rossi, 2006; Jiménez et al., 2006b; Decaëns et al., 2009).

Fig. 5.6
figure 6

Map of clusters for Eudrilidae and Millsonia anomala in a grass savanna at Lamto, Côte d’Ivoire. Points contributing to a patch or a gap are represented by circles (local cluster index > 1.5) or squares (local cluster index < −1.5), respectively. (Rossi, 2003a)

Full size image

Dispersal rates of various earthworm species affect their spatial distribution, particularly for peregrine species which are well-adapted to colonize new areas (Lee, 1985). Hamblyn and Dingwall (1945) claimed that the rate of advance of the margin of populations of A. caliginosa, from inoculation points in recently limed grasslands, was of the order of 10 m per year, and they suggested that any more rapid horizontal dispersal was probably by cocoons carried in soil on agricultural implements, hooves of animals, feet of birds, or in streams. This conclusion has been supported by later studies. Many workers have introduced earthworms to new habitats and studied their rate of multiplication and spread. For instance, van Rhee (1969a, b) inoculated a new polder in the Netherlands with 4664 individuals of A. caliginosa and reported that they had multiplied to 384 740 individuals 1 year later. At the same site, 2558 individuals of A. chlorotica increased to 121 660 in the same period. He calculated a horizontal rate of spread of the population of A. caliginosa of 6 m per year and 4 m per year for A. chlorotica. Hoogerkamp et al. (1983) reported an annual horizontal dispersal rate of about 9 m for A. caliginosa and 4 m for L. terrestris. The spread of earthworms in inoculated polder soils was calculated by Stein et al. (1992) to be from 10 to 13 m per year. Stockdill (1982) inoculated A. caliginosa into New Zealand pastures at a spacing of 10 m and reported that the whole area became colonized after 8–10 years.

In general, it is likely that environmental heterogeneity as well as intrinsic properties of populations influence the spatial distribution of earthworm communities (Ettema & Wardle, 2002). Aggregated distribution patterns may result from population decline in certain areas due to disturbance (e.g., soil desiccation) or resource depletion; migration into areas with more favorable conditions (e.g., temperature, water, food availability); slow dispersal from actively growing populations (e.g., founder populations in newly colonized areas); competitive exclusion and self-organization due to predator/prey or parasite/host interactions; and gregarious behavior (Satchel, 1955; Dash, 1990; Jiménez & Rossi, 2006; Barot et al., 2007). More recent studies have also shown that factors influencing the horizontal distribution of earthworms continue to present themselves as more manifold and interdependent than has already been proven. Tajik et al. (2019) demonstrated that the horizontal distribution of earthworm abundance was affected by soil microbial respiration, land surface temperature, normalized difference vegetation index (a satellite observation of whether or not the observed target contains live green vegetation), Shannon index of trees, and soil moisture. The study itself further details the myriad of environmental factors which determine earthworm horizontal distribution which could play a more important role in distribution other than food availability and distribution.

5.4.2 Vertical Distributions

Different species of lumbricids in different ecological groupings inhabit different depth zones in the soil (Fig. 5.7), but the vertical distribution of each species changes considerably with the time of year. The seasonal vertical distribution of common British lumbricids has been studied by several workers. Species such as D. octaedra and B. eiseni live in the surface organic horizon of soil for most of the year. Allolobophora caliginosa, A. chlorotica, A. rosea, L. castaneus, and L. rubellus occur commonly within 8 cm of the soil surface, as do immature individuals of O. lacteum, O. cyaneum, A. longa, A. nocturna, and L. terrestris. Most adult and nearly mature individuals of O. cyaneum are in the top 15 cm, and although they have distinct burrows, these are usually temporary. Aporrectodea longa and A. nocturna have fairly permanent burrows which usually penetrate as deep as about 45 cm, but the deeper vertical burrows of L. terrestris commonly go down to a depth of 1 m and can penetrate as deep as 2.5 m. Gerard (1967) studied the changes in vertical distribution of common species of earthworms at different times of the year in England (Fig. 5.8). Most earthworms in his samples were below 7.5 cm deep in January and February, when the soil temperature was about 0 °C, but by March when the soil temperature had risen to 5 °C at a depth of 10 cm, most individuals of A. chlorotica, A. caliginosa, and A. rosea, and small- and medium-sized individuals of A. longa, A. nocturna, and L. terrestris, had moved into the top 7.5 cm of soil, although most of the larger earthworms were still deeper in the soil. From June to October, earthworms of most species were below the top 7.5 cm again, except for newly hatched individuals. In November, December, and the following April, most earthworms had returned to the top 7.5 cm. The two factors influencing movement to deeper soil seemed to be very cold or very dry surface soil. All species of earthworms (except L. terrestris) seemed to be quiescent in summer and mid-winter, and at both these times were deeper than 7.5 cm below the surface. More earthworms were quiescent in summer than in winter. Nearly all cocoons were found in the top 15 cm of soil, most being in the top 7.5 cm.

Fig. 5.7
figure 7

Vertical distributions of earthworms in a Rothamsted pasture. (Adapted from Satchell, 1955)

Full size image
Fig. 5.8
figure 8

The depth of six species of earthworms in monthly soil samples from January to December, 1959 (expressed as percentages for each species in each sample). Samples were taken in 7.5 cm layers to a depth of 30.5 cm, and sometimes deeper (up to 53 cm). (After Gerard, 1967)

Full size image

Seasonal changes in earthworm vertical distributions were also studied in Sweden (Rundgren, 1975). Dendrobaena octaedra and D. rubida occurred close to the surface throughout the year, but A. longa, A. caliginosa, A. rosea, and L. terrestris all penetrated deeper into the soil at some periods of the year (Rundgren, 1975). Earthworms of the genus Diplocardia in southeastern United States moved from a depth of about 10 cm in October, to about 40 cm deep in January, and return to the surface soil in spring, with a critical temperature for downward movement of approximately 6 °C (Dowdy, 1944). Millican and Lutterschmidt (2007) reported that populations of Dendrobaena invicta had seasonal vertical migrations but were quiescent from June to October in blackland prairie soils of eastern Texas. In the northwestern United States, earthworms of the species Pheretima hupeiensis were active in the 15–20 cm soil level and were near the surface only in September and March, and from November to February, they were deeper than 55 cm (Grant, 1956).

In Australia, A. trapezoides and A. caliginosa were most abundant in the surface layers of the soil (0–10 cm depth) for the 3–7 months from autumn to spring when the soils were most moist. During summer, most individuals were deeper than 20 cm below the surface, were inactive, and coiled tightly within spherical chambers (Baker et al., 1992a, b). In other studies (Baker et al., 1993a, b), A. rosea, A. trapezoides, Microscolex dubius, and Microscolex phosphoreus were most abundant when the soils were wettest, and during this period, the majority of earthworms were in the top 10 cm of soil. During the other months of the year, most earthworms were located at a depth below 10 cm, and mature adults with clitella were found only in winter and spring. Similarly, Gemascolex walkeri and O. cyaneum occurred predominantly in the top 10 cm of soil in Australian pastures, for 4–5 months each year, in autumn to spring, when the soils were wettest; during the dry season, they were much deeper (Baker et al., 1993a, b).

It seems that seasonal vertical migrations of earthworms occur in most parts of the world, and these are initiated by the environmental factors in the upper soil levels becoming unsuitable for earthworms to feed and grow satisfactorily. A positive correlation between the length of the lumbricid body and the depth of burrowing into soil was suggested by Piearce (1983). Data for several tropical earthworm species in Colombia showed no significant relationship between length/diameter (L/D) and weight/diameter (W/D) ratios of earthworm bodies and soil depth. However, when the small, epigeic ocnerodrilid species was removed from the analysis, correlations were significant; this small ocnerodrilid moves vertically through burrows of larger, deep-burrowing species, penetrating throughout the soil profile (Jiménez & Decaëns, 2000).

Reddy and Pasha (1993) investigated the influence of rainfall and temperature on the vertical distributions of earthworms in two semiarid grassland soils in India. The species were Octochaetona philloti (Michaelsen) (16–96 per m2) and Barogaster annandalei (Stephenson) (3.2–58.3 per m2). These earthworms migrated to deeper soil layers during winter and summer, O. philloti reaching a maximum depth of 40 cm and B. annandalei going down to 45 cm. They concluded that physical soil factors were more important collectively in influencing seasonal variations in populations and distributions than chemical factors. Usually those species that feed on or near the surface were dark-colored, whereas the subterranean species were predominantly paler in color.

The vertical distribution of earthworms and spatial patterns of earthworm assemblages in general are also influenced by the crops of the organic farming grass-clover rotation cropping sequence with burrows reflecting crops, fertilization, and tillage, all of which alter burrowing activity. Krogh et al. (2021) showed that grazing changed the burrowing depth reached by anecic earthworms. According to the study, cattle grazing favors the occurrence of burrows made by the anecics A. longa and L. herculeus while simultaneously decreasing the number of fine-medium macropores. Bottinelli and Capowiez (2020) proposed that the three standard categories used to classify earthworm ecological placement have been conflated with function and that the understanding of earthworm function, including that having to do with and affected by vertical distribution, relative to ecological categories ought to be reassessed. The seven ecological categories which include anecics, endogeics, and epigeics have often been reduced only to these three categories and considered as functional groups, tacitly using these to presume the way earthworms influence soil functioning. The uncritical acceptance of this categorization sometimes results in unexpected trends when species of earthworms are gathered in these categories to analyze burrow systems or cast properties for instance. This categorization bleeds over into the topic of the vertical distribution of earthworm populations and the dynamicity thereof which requires a functional understanding of earthworms that is not necessarily divorced from ecological niche but is understood as separate and influential upon one another.

5.5 Seasonal Populations and Activity

The numbers of earthworms and their degree of activity vary greatly during annual seasonal cycles, but not all earthworm ecologists have differentiated adequately between seasonal changes in numbers and changes in earthworm activity. This applies particularly to those researchers who used chemical extractant methods for determining populations; such techniques depend on the activity of the earthworms and quiescent individuals do not respond, so the numbers collected reflect both size of populations and their activity.

Evans and Guild (1947c) followed changes in numbers of earthworms in an old pasture field in England for more than a year, using a chemical sampling method, and they concluded that the two soil conditions that affected earthworm activity most were temperature and moisture (Fig. 5.9), although another important factor was the obligatory diapause that occurred from May to October, for the two species A. nocturna and A. longa, or the periods of quiescence or facultative diapause of A. chlorotica, A. caliginosa, and A. rosea during adverse conditions. The five species L. terrestris, A. rosea, A. chlorotica, A. nocturna, and A. caliginosa were most active between August and December and April to May (Fig. 5.10). Tiwari et al. (1992) also found a significant correlation between earthworm populations and temperature and moisture in a pineapple field.

Fig. 5.9
figure 9

(a) The seasonal changes in activity of all earthworms; “total of earthworms” = total number of earthworms in 10 samples; (b) the corresponding changes in soil moisture at 5 cm; and temperature at 10 cm, from March 1945 to June 1946. (From Evans & Guild, 1947c)

Full size image
Fig. 5.10
figure 10

The seasonal changes in activity of five species of earthworms. (After Evans & Guild, 1947c)

Full size image

Potvin and Lileskov (2016) reported earthworms’ sensitivity to dynamic environmental change, resulting response brought on by seasonal conditions and showed the difference in the effect of climate change on earthworm activity depending on earthworm niche. Seven years of data were collected on the activity of an anecic species (L. terrestris) and an endogeic species (Aporrectodea caliginosa complex), both of which were introduced, in two distinct soil/plant communities. Over those 7 years, Potvin and Lileskov (2016) recorded the distribution and activity state biweekly at a depth of 1.5 meters, tracked L. terrestris burrows using annually captured images and took data of the soil temperature and moisture. The activity and vertical distribution of earthworms were shown to be closely related to earthworm species and soil temperature in fall, winter, and spring. L. terrestris remained active through winter, while A. caliginosa complex was more likely to enter an aestivation period, or a period of dormancy. All earthworms decreased their activity substantially in the months of July and August when the temperature of the studied soil was highest and moisture was lowest compared to other months. Most of the burrows used by L. terrestris were used continuously and moved little over the course of the 7-year study, and it is posited that this could have created spatiotemporally stable hotspots of soil resources. The contrasting patterns of response observed in the study show that endogeic earthworms are more likely than anecic ones to adjust their states of activity in response to climate change-mediated shifts to soil moisture and temperature.

The amounts of seasonal activity of surface-casting earthworm species can also be assessed by the numbers of earthworm casts produced. Evans and Guild (1947c) found that the numbers of casts deposited, and the numbers of A. longa and A. nocturna that were obtained by potassium permanganate sampling, were correlated closely. Other workers have attempted to estimate seasonal activity by counting the numbers of earthworms on the surface at night, but this is very inaccurate because many species of earthworms only come to the surface when it is wet.

Gerard (1967) reported that in pasture soil in England, A. chlorotica, A. caliginosa, and A. rosea usually occurred within 10 cm of the soil surface, but when the soil temperature fell below 5 °C, or the soil became very dry, individuals of these species moved to deeper soil. Most cocoons were produced in late spring and early summer (Fig. 5.11); in hot, dry periods in summer, most species became inactive and were deeper in the soil. Callaham and Hendrix (1997) reported seasonal fluctuations of mostly European lumbricid populations in a bottomland forest in the southern Appalachian Piedmont, USA. Earthworm populations were relatively constant during the wet winter and spring but declined over summer and fall with increasing temperatures and lower soil moisture. Hopp (1947) believed that the most important factor influencing seasonal changes in numbers of worms in arable soils in the northern United States was the death of earthworms when the unprotected surface soil became frozen in winter. Seasonal changes in earthworm populations have also been ascribed to other causes; for instance, Waters (1955) suggested that flushes in the availability of dead root material and herbage debris were a main cause of increased numbers of earthworms, but this hypothesis is not generally accepted.

Fig. 5.11
figure 11

Seasonal production of cocoons by A. chlorotica and A. caliginosa. (After Gerard, 1967)

Full size image

Gates (1961) reported that earthworm activity in the tropics was also limited to certain seasons. In the monsoon tropical climate of Burma and the humid subtropical climate of India, earthworms are active mainly in the 4–6 months of the rainy season between May and October. The principal period of activity of Hyperiodrilus africanus in Nigeria occurred during May and June, which is the beginning of the wet season; thereafter, numbers lessened gradually until November, when very few were found. The numbers of cocoons produced also varied seasonally (Madge, 1969). In seasonally wet savannas of Côte d’Ivorie, West Africa, population densities of eudrilid and megascolecid earthworms were significantly related to rainfall (Lavelle, 1971; Tondoh, 2006) (Fig. 5.12). Millsonia anomala had phases of inactivity and quiescence occurring when soil water content fell below 7% (Lavelle, 1971). Interestingly, Tondoh (2006) found no correlation between rainfall and earthworm species richness; his analysis showed that niche overlap and pairwise species associations were random, suggesting no evidence of seasonal niche partitioning among earthworm species.

Fig. 5.12
figure 12

Monthly variation of the mean earthworm community abundance (ind m−2) and rainfall patterns in a tropical savanna Côte d’Ivoire. (Tondoh, 2006)

Full size image

Grant (1956) reported that Pheretima hupeiensis was most active in the northwestern United States during the summer months and retreated to soil levels below 55 cm deep from November to February.

By contrast, in the humid continental climate of the southeastern United States, this species is most active in the spring and autumn months. Hopp (1947) presented data on the seasonal earthworm population changes in Maryland, USA, which confirmed this, but since he only took samples to a depth of 7.5 cm, the proportion of the total population he extracted varied considerably with the time of year. Callaham and Hendrix (1997) reported seasonal fluctuation of mostly European lumbricid populations in a bottomland forest in the southern Apalachian Piedmont, USA. Earthworm populations densities were relatively constant during the wet winter and spring but declined over summer and fall with increasing temperatures and lower soil moisture.

Grant (1956) reported that P. hupehensis was most active in the northwestern United States during the summer months and retreated to soil levels below 55 cm deep from November to February. Other species of Asian pheretemoids (genus Amynthas) introduced into North America had distinct seasonality in population density and demography. In the southern Appalachian Mountains, Callaham et al. (2003) found only small juveniles of A. agrestis in July and early August, a peak in abundance of juveniles and subadults in late August, and a peak in abundance of adults in September as population densities declined to zero in late fall.

In grassland in Japan (Nakamura, 1968a, b), the greatest numbers of earthworms occurred in autumn, especially in October, and the numbers were very low in winter, particularly during January and February (Fig. 5.13). In Australia, populations of A. rosea, A. caliginosa, and O. cyaneum in pasture increased from May to July and decreased from July to October (Baker et al., 1993a). Soil temperatures were not low enough during the wet season (10 °C at 15 cm depth) to prevent breeding from occurring. The vertical distribution of the earthworms in these soils also changed during the year, so that in winter, most earthworms were in the top 15 cm and few were below 30 cm, whereas when soils began to dry out in spring, few earthworms remained in the top 15 cm. During summer, 60% of all worms were between 15 and 30 cm deep. The abundance of A. rosea, A. trapezoides, Microscolex dubius, and Microscolex phosphoreus was monitored monthly in lucerne and cereal fields in Australia (Baker et al., 1993b), and overall populations of about 303 per m2 were recorded. The highest numbers occurred in winter and spring. In a New Zealand pasture between 1951 and 1954, peak populations occurred in mid-winter (Waters, 1955; Barley, 1959a, b) (Fig. 5.14).

Fig. 5.13
figure 13

Seasonal variation in the population density of earthworms in alluvial soil grassland in Japan during the period May 1965 to April 1967. (After Nakamura, 1968a)

Full size image
Fig. 5.14
figure 14

Seasonal fluctuations in the abundance of earthworms found in pasture land at Palmerston North, New Zealand. W winter, Sp spring, Su summer, A autumn; weight; number. (From Waters, 1955)

Full size image

In addition to such seasonal changes in activity, many species of earthworms are distinctly diurnal in their activity. Experiments by one of the authors demonstrated clearly that L. terrestris tended to be most active from 6 p.m. to 6 a.m., although this varied with season. This diurnal activity is intrinsic and seems to be at least partially independent of temperature and light. Similarly, Butt et al. (2003) showed that surface activity of L. terrestris peaked 1 h after darkness and declined for 11 h until lights were turned on in the experimental chambers; both foraging and mating activity were affected by the diurnal cycle. Patterns of activity differ between species; for instance, Millsonia anomala has a marked diurnal rhythm of activity, with two maxima of casting at about midnight and 9 a.m. in moist soil (Lavelle, 1971).

Seasonal difference is an integral factor in determining earthworm behavior, arguably more so than other seemingly influential factors such as dominant tree species in a forest or subsection thereof. Bayranvand et al. (2017) showed that seasonal changes in earthworm populations and microbial respiration were observed under several forest species (Carpinus betulus, Ulmus minor, Pterocarya fraxinifolia, Alnus glutinosa, Populus caspica, and Quercus castaneifolia) in a temperate mixed forest in northern Iran. The authors observed that earthworm density/biomass varied seasonally but not significantly between tree species. Soil microbial respiration also did not differ between tree species and showed similar temporal trends in all soils under different tree species. Though tree species intimately affected soil chemical properties such as acidity, organic C, and total N content of mineral soils, the density/biomass and microbial respiration were not affected by tree species but were in fact controlled by the activity of trees and climate, exerting strong signs of seasonal dynamics.