Soil Biology & Biochemistry 33 (2001) 1959±1969 www.elsevier.com/locate/soilbio Earthworm excreta attract soil springtails: laboratory experiments on Heteromurus Nitidus (Collembola: Entomobryidae) Sandrine Salmon*, Jean-FrancËois Ponge Museum National d'Histoire Naturelle, Laboratoire d'EÂcologie GeÂneÂrale, 4, Avenue du Petit-ChaÃteau, 91800 Brunoy, France Received 11 September 2000; received in revised form 11 May 2001; accepted 15 June 2001 Abstract Microarthropods are often found more abundantly in soils with earthworms than in soils without. Earthworms probably create a favourable environment for microarthropods but few studies have aimed to explain this earthworm effect. The soil collembolan (Hexapoda) Hetero- murus nitidus, living in soils at pH . 5 only and thus rich in earthworms, is particularly attracted by earthworms in humus cores. The effect of earthworms on the distribution of H. nitidus can be mediated either by direct contact or by odour perception. Two experimental designs were used to determine the pathway of attraction. The ®rst set of experiments studied the effect of direct contact with earthworm excreta on the distribution of H. nitidus. The mixture of urine and mucus of the lumbricid earthworms Aporrectodea giardi and Alollobophora chlorotica signi®cantly attracted H. nitidus as compared to deionized water while fresh earthworm casts were not preferred to calcic mull made of older casts. The same experiment involving direct contact with mucus and methyl blue showed that Collembola sucked on mucus/urine, indicating that the interaction of Collembola and earthworms was at least partly trophic. The second experiment demonstrated that H. nitidus was attracted by the odour of Aporrectodea giardi at short distance. The odour of excreta (mucus, urine and casts) of Aporrectodea giardi also attracted H. nitidus but this attraction was weaker and did not occur constantly, possibly due to interactions with light and aggregation pheromones. We conclude that the prominent pathway by which earthworms could attract H. nitidus in the ®eld is through direct contact with earthworm mucus and urine. The acid-intolerant distribution of this species in the ®eld could be partly explained by a trophic interaction with some earthworm species. q 2001 Published by Elsevier Science Ltd. Keywords: Attraction; Collembola; Lumbricidae; Mucus; Olfactory signal 1. Introduction improve water availability, aeration, and pore size (Wickenbrock and Heisler, 1997; Devliegher and Collembolan communities vary in soils according to Verstraete, 1997; Subler and Kirsch, 1998), which may several factors, among them the presence of earthworms, attract microarthropods. In addition, earthworms increase high populations of which characterize mull humus forms the availability of a number of inorganic (Jeanson, 1972; with a moderately low to high pH (Satchell, 1967; Piearce, Scheu, 1987; Robinson et al., 1992) and organic compounds 1972). In fact, some studies have shown that the density and (Dubash and Ganti, 1964; Martin et al., 1987), favour some diversity of microarthropods increase in soils with bacteria (Kozlovskaja, 1969; Saetre, 1998) and increase or earthworms (Bayoumi, 1978; Marinissen and Bok, 1988; decrease fungal populations (Brown, 1995) on Hamilton and Sillman, 1989; Loranger et al., 1998). More- which Collembola may feed. These products may attract over, numerous Collembola have been observed directly Collembola or favour their abundance and diversity. wandering on giant earthworm bodies (Bouche personal The collembolan species Heteromurus nitidus is always communication) and in rearing beds of earthworms found in soils at pH . 5 that generally bear mull humus (Greenslade and Fletcher, 1986; Arbea and Jordana, (Ponge, 1993; Salmon and Ponge, 1999). This species 1988). However no attempt has been made to study direct seems to be particularly dependent on the presence of earth- positive effects of earthworms on the surrounding fauna. worms since mull humus is characterised by a high abun- Earthworms, through their action on the surrounding soil, dance of earthworms and particularly we demonstrated, from laboratory experiments, that Heteromurus nitidus preferred humus blocks with rather than without earth- * Corresponding author. Tel.: 133-1-604-79241; fax: 133-1-60-46-50- 09. worms (Salmon and Ponge, 1999). However, we know E-mail address: ssalmon@mnhn.fr (S. Salmon). neither the causes nor the pathway of this attraction. One 0038-0717/01/$ - see front matter q 2001 Published by Elsevier Science Ltd. PII: S0038-071 7(01)00129-8 1960 S. Salmon, J.-F. Ponge / Soil Biology & Biochemistry 33 (2001) 1959±1969 or many of the substances and microorganisms excreted by contained earthworm excreta (casts or mucus plus urine; see earthworms (see above) may attract Collembola. In fact, below), the other served as reference (calcic mull aggregates fungal-feeding Collembola were able to discriminate or water). Ten adult H. nitidus were taken randomly from between different fungal species and showed preferences the rearing boxes then introduced in each of the six Petri among them (Shaw, 1988; Thimm and Larink, 1995; dishes (replicates), half-way between the two half-disks. Kaneko et al., 1995). They recognized fungal odours and Their abundance was counted on each half-disk every were attracted or repelled by them (Bengtsson et al., 1988; 10 min for 140 min. Animals outside the half-disk areas Hedlund et al., 1995; Sadaka-Laulan et al., 1999). Some were ignored. All runs were established at ambient tempera- Collembola were also attracted by a mite species (Huber, ture (around 208C), under homogeneous light conditions 1979). In addition, earthworms are known to attract other (checked with LI 1000 Data Logger and LI-COR Radiation invertebrate species (Halpern et al., 1984; Morris and sensors). The two half-disks were equally moistened. A Pivnick, 1991). control experiment was performed with both half-disks The aim of this study was to determine the causes and the moistened with deionized water only, in order to verify pathway of attraction of the collembolan H. nitidus by two that some factor other than earthworm excreta, especially earthworm species. Is H. nitidus attracted at a distance by light, did not in¯uence the distribution of H. nitidus. This the odour of earthworms or does it remain in the vicinity of species is highly sensitive to light (Salmon and Ponge, the earthworms after direct contact? Earthworms and their 1998) and although light conditions were as homogenous excreta (mucus, urine and casts) were tested in odour experi- as possible a control experiment was necessary to verify this ments. In direct contact experiments, the effects of mucus point. plus urine, and casts, were studied. One experiment using For each replicate, means of 14 time-measures of earthworm mucus and urine and methyl blue was performed numbers of Collembola on each half-disks were calculated. to assess whether mucus was consumed by Collembola. The normality of the data was checked, and means of six The present paper focuses on two lumbricid species replicates in `earthworm excreta' were compared to means Aporrectodea giardi and Allolobophora chlorotica which in `reference substrate' by a paired t-test (Sokal and Rohlf, attracted H. nitidus in humus blocks (Salmon and Ponge, 1995). 1999). They were extracted from a calcic mull, a humus form in which H. nitidus is commonly found. The mechan- 2.2.2. Experiments with mucus and urine isms of attraction and the impact of earthworms on the Adult earthworms, eight A. giardi or 54 A. chlorotica, distribution of H. nitidus are discussed. were rinsed under tap water. They were then placed on moistened ®lter paper to void their guts and kept at 158C in darkness, in a water-tight plastic box, for 60 h. The ®lter 2. Materials and methods paper was renewed every 20 h. Three days after washing the worms, six half-disks of 2.1. Test organisms ®lter paper were saturated with mucus and urine by incubat- Specimens of H. nitidus were reared in laboratory ing them with the earthworms. For this purpose, half-disks cultures on moist Fontainebleau sand (pure ®ne quartz and earthworms, the two species separately, were placed for sand). They were fed with lichens and terrestrial microalgae 5 h in a water-tight enclosure at ambient temperature, in (Pleurococcus) from bark scrapings. Cultures were kept at darkness. Mucus and urine were tested together because 158C in permanent darkness. Each experimental run was both are excreted from the wall of earthworms (BoucheÂ, performed with new specimens. 1972). Six other half-disks were saturated with deionized Two earthworm species were sampled from a calcic mull water. H. nitidus could then choose between mucus plus in the laboratory park. Aporrectodea giardi is a large earth- urine or deionized water during 140 min as described worm (150 mm in length) belonging to the anecic category above. As H. nitidus is known to be especially sensitive to (BoucheÂ, 1972). Allolobophora chlorotica is smaller desiccation (Bauer and Christian, 1987), the two half-disks (50 mm) and classi®ed as endogeic. The extraction was were equally moistened (saturated) with deionized water or performed with 4½ formalin a few days before the start mucus plus urine. of each experiment, and in the mean time earthworms were kept in their original soil at 158C in darkness. 2.2.3. Experiments with casts 20 h before the start of the experiment eight adult A. 2.2. Direct contact experiments giardi or 60 adult A. chlorotica were rinsed under tap water and added to Petri dishes (é 14 cm) containing mois- 2.2.1. General protocol tened ®lter paper, to collect fresh casts. In order to condition Direct contact experiments were performed in six Petri reference substrates in the same manner, moistened ®lter dishes (é 8 cm) containing two half-disks (é 5 cm) of ®lter paper lining the bottom of Petri dishes (é 14 cm) was paper placed at 1.5 cm distance from each other. The dishes covered by calcic mull aggregates collected at the same were placed on a laboratory table. Only one of the half-disks time as earthworms. Earthworms and calcic mull were  S. Salmon, J.-F. Ponge / Soil Biology & Biochemistry 33 (2001) 1959±1969 1961 Fig. 1. Design of test-boxes used in odour experiments with one moistened area of ®lter paper as bottom substrate for Collembola. then kept at 158C in darkness, in a water-tight enclosure. On three Petri dishes (é 8 cm) containing three or four disks of the day of the experiment, fresh casts and calcic mull aggre- ®lter paper 1 day before the experiments started. Petri dishes gates were smeared separately on six half disks each by with earthworms were then placed in a watertight plastic means of a scalpel. Both types of half disk were deposited box in darkness at ambient temperature, for 20 h, in order in each Petri dish (replicate). Thereafter, choice experiments to impregnate the ®lter paper disks with earthworm excreta, were run as described above. H. nitidus could then choose i.e. mucus, urine and fresh casts. In the experiment where between casts and calcic mull material with similar colour the odour of the earthworm itself (not its excreta) was and consistency, the hemorganic horizon of calcic mull studied, ten A. giardi were rinsed and used directly. being formed of aged earthworm casts. The experimental design allowed volatile compounds from earthworm excreta to diffuse without any contact or 2.2.4. Experiments with methyl blue visual perception by Collembola (Fig. 1). All odour experi- A similar experiment with mucus plus urine of eight A. ments comprised ten replicates ( ˆ ten experimental cham- giardi was carried out on half-disks previously stained with bers). The experimental chambers consisted of rectangular methyl blue dye. Methyl blue was used to see whether plastic boxes 11.4 £ 8.4 £ 6.5 cm (l £ l £ h) divided into Collembola grazed and absorbed the mucus/urine secretion. two compartments, each containing a plastic vessel ®lled In each box the two half-disks were equally stained to with remoistened Fontainebleau sand. The top of one of prevent any differential effect of colour upon the distribu- the vessels was covered by a ®lter disk impregnated with tion of Collembola. Twelve half-disks were impregnated by A. giardi excreta. The other vessel was covered by a ®lter a methyl blue water solution (25 g l 21) and left to dry in disk impregnated with deionized water only to obtain a darkness for 6 days until the choice experiments started. similar moisture. The vessels were covered by a platform During the experiment H. nitidus could choose between on which 12 H. nitidus individuals were allowed to wander. two blue half-disks, the one impregnated with mucus plus The platform varied according to the type of experiment urine, the other with deionized water. (see below). Odour attraction experiments were accompa- A control experiment was performed in which each test nied by a control experiment, using the same apparatus but box contained two blue half-disks saturated with deionized without earthworm excreta. This control experiment was water. This control experiment allowed veri®cation of the designed to assess possible effects of other factors such as harmlessness of the dye and a contrast between the absorp- light on collembolan distribution. tive behaviour of H. nitidus in the presence of water only In the ®rst two sets of experiments, the platform for and water and mucus plus urine (test-experiment). H. nitidus consisted of a moistened sheet of ®lter paper At the end of these choice experiments Collembola which was ®xed to the box walls, 0.8±1 cm above the were transferred to ethanol (90%) and the colour of vessels (Fig. 1). The platform area was arbitrarily divided their gut contents was observed under a dissecting into two sectors by a drawn line. Above each vessel, the microscope. moist ®lter paper was perforated to allow the passage of lipophilic olfactory compounds through the suspended plat- 2.3. Odour attraction experiments form. Twelve adult or sub-adult H. nitidus (different for each replicate) were released over the suspended platform To test the attraction by the odour of earthworm excreta, (six in each sector) and each test box was placed immedi- 22 adult A. giardi were rinsed in tap water and then placed in ately in a dark enclosure. Light was present only during the 1962 S. Salmon, J.-F. Ponge / Soil Biology & Biochemistry 33 (2001) 1959±1969 Table 1 3. Results Numbers of H. nitidus on each half-disk in three choice experiments invol- ving a direct contact over 140 min (means of 14 times and six replicates). 3.1. Direct contact with earthworm excreta NS ˆ not signi®cant at 0.05 level; **, *** ˆ signi®cant at the 0.01 and 0.001 levels, respectively H. nitidus was signi®cantly more abundant on half-disks Content of half-discs Number of Results of impregnated with mucus and urine from A. giardi than on H. nitidus paired t test half-disks saturated with deionized water (Table 1). H. niti- (mean ^ SEM) (probability) dus was also more attracted by mucus and urine from A. Mucus and urine of A. giardi 6.04 ^ 0.34 0.0001*** chlorotica than by water (Table 1) and the attraction was Water 1.32 ^ 0.18 faster than for A. giardi (Fig. 2A). Preliminary experiments Mucus and urine of A. chlorotica 8.23 ^ 0.73 0.0063** in the same conditions but using earthworms previously Water 1.58 ^ 0.74 kept on moistened paper during 0, 12 or 36 h instead of 60 h (thus the gut was incompletely voided), gave similar Water (disk E1) 3.95 ^ 0.53 0.7889 NS Water (disk E2) 3.69 ^ 0.40 results (data not shown), con®rming the trend of the choice exerted by animals. The control experiment indicated the absence of an uncontrolled preference within the dish (Table short time spent for releasing (3 min) and counting (1 min) 1). Thus mucus and urine of both earthworm species animals. attracted H. nitidus individuals. Observations between In the ®rst series of experiments H. nitidus individuals successive measurements showed that generally Collem- were counted in both sectors every 10 min for 2.5 h. One bola did not redistribute between half-disks when in contact experimental set assessed the effect of the odour of earth- with mucus and urine from earthworms. Numerical changes worm excreta, the other was a control experiment (two were only due to new individuals coming in from other sectors `no odour'). The second series of experiments, places of the Petri dish. also comprising an odour and a control experimental set, When H. nitidus individuals were allowed to choose was performed in the same device but Collembola were between calcic mull and casts of A. giardi or A. chlorotica, counted every 30 min over 7 h in order to have an overview their density was slightly higher on casts but differences of longer-term effects. Results were analysed as for direct between both substrates were not signi®cant (Table 2, Fig. contact experiments using paired t-test. 2B). In the third series of experiments the apparatus was modi- In the presence of stained half-disks, H. nitidus indivi- ®ed to allow for a better balance in the distribution of duals did not show any preferences for a given area in the animals between the two sectors in the absence of an earth- control experiment (Table 3). All animals survived and were worm effect. Since the dispersal of Collembola is favoured distributed as in the case of undyed half-disks (Fig. 2C, by an increase in population density (Bengtsson et al., compare to Fig. 2A). Thus the methyl blue dye did not affect 1994), we reduced the size of the moist area by replacing the distribution and the behaviour of Collembola. In experi- the platform made of a ®lter paper sheet by two much smal- ments with water only, neither the gut contents nor ventral ler wet areas. The platform was made of a plastic sheet with tubes of Collembola showed any blue coloration, indicating two holes (é 3.2 cm) each covered with a perforated ®lter that in the absence of mucus they did not suck on the paper disk. Paper disks were moistened with an equal half-disks. When one of the two half-disks was saturated volume of deionized water to avoid the in¯uence of varying with mucus and urine from A. giardi then the corresponding humidity (Joosse and Verhoef, 1974). Twelve Collembola number of H. nitidus was signi®cantly higher on it (Table 3). were released in the space between the two moist areas. The abundance of animals on mucus and urine increased up Their number was counted on both ®lter paper disks each to 60 min and thereafter remained stable (Fig. 2C). Methyl 30 min over 5.5 h. A control experiment (a) with two areas blue did not disturb the attraction of H. nitidus by mucus and `no odour', was performed to verify the homogeneity of urine (compare with Fig. 2A). Forty per cent of individuals the distribution in the two areas in the absence of an exhibited a blue coloration in their gut, indicating that earthworm effect. An experiment (b) was carried out to Collembola sucked on methyl blue stained half-disks assess the odour of earthworm excreta as in the impregnated with mucus and urine. Two individuals (3%) previous device. A third experiment (c) was performed had only the end of their ventral tube blue stained. after replacing earthworm excreta by the earthworm itself: an individual of A. giardi was introduced in 3.2. Odour of earthworm and earthworm excreta one of both vessels, then the corresponding vessel was (casts 1 mucus 1 urine) locked up with a lid of perforated ®lter paper. H. niti- dus individuals were counted in each area at each time. The ®rst odour experiment (Table 4) showed that H. niti- Mean differences between both areas in the ten boxes dus was signi®cantly more abundant in the sector above (replicates) were compared in the same way as in previous earthworm excreta than in the sector without the odour of experiments. earthworm excreta, especially from 80 min onwards Fig. 2. Differences in the number of Heteromurus nitidus (means of six replicates ^ standard errors) between two half-disks of ®lter paper (`earthworm excreta' versus `reference substrate') in an experiment involving a direct contact over 140 min. (A) Experiment with mucus and urine versus water. (B) Experiment with fresh casts versus calcic mull. (C) Experiment with mucus and urine versus water in half-disks previously stained with methyl blue dye. 1964 S. Salmon, J.-F. Ponge / Soil Biology & Biochemistry 33 (2001) 1959±1969 Table 2 Table 3 Numbers of H. nitidus on each half-disk in two choice experiments invol- Numbers of H. nitidus on each half-disk in two choice experiments invol- ving a direct contact over 140 min (means of 14 times and six replicates). ving a direct contact over 140 min (means of 14 times and six replicates). NS ˆ not signi®cant at 0.05 level NS ˆ not signi®cant at 0.05 level; * ˆ signi®cant at 0.05 level Content of half-discs Number of H. nitidus Results of paired t Content of half-discs Number of H.nitidus Results of paired t (mean ^ SEM) test (probability) (mean ^ SEM) test (probability) Fresh casts of A. giardi 5.61 ^ 0.92 0.2049 NS Methyl blue 1 water (E1) 2.92 ^ 0.44 0.5606 NS Calcic mull 2.99 ^ 0.88 Methyl blue 1 water (E2) 3.39 ^ 0.65 Fresh casts of A. chlorotica 5.87 ^ 1.01 0.2661 NS Methyl blue 1 mucus and 7.51 ^ 1.08 0.0134* Calcic mull 3.49 ^ 0.90 urine of A. giardi Methyl blue 1 water 1.08 ^ 0.66 (Fig. 3A). 80 min is probably the time needed for Collem- bola to explore their milieu and for the odour to reach them. higher in the area `earthworm excreta odour' (Fig. 3C) but The aggregation towards the side with odour of earthworm that this preference was not signi®cant (Table 5). However, excreta increased with time up to the end of the experiment, the preference became signi®cant (P ˆ 0.005) when an i.e. 160 min (Fig. 3). The control experiment showed a little outlier replicate (where an unexpected aggregation occurred but not signi®cant clustering of animals which occurred in the area ªno odourº) was eliminated. indifferently in one of both sectors (Table 4 and Fig. 3A). When earthworm excreta were replaced by earthworms, Because a threshold level was not reached at the end of the mean abundance of Collembola was signi®cantly higher the 160 min experiment, a longer-lasting experiment was in the area with the odour of earthworms than in the area performed over 420 min (Fig. 3B). H. nitidus was signi®- without odour (Table 5 and Fig. 3C). Thus earthworms cantly more abundant in the `earthworm excreta odour' themselves, not defecating but rather excreting mucus and sector (Table 4) particularly from 60 min up to the end of urine, produced an olfactory signal, which was attractive to the experiment (Fig. 3B). Thus, the attraction of Collembola H. nitidus. The attraction occurred more rapidly with the by the odour of earthworm excreta became effective at the odour of earthworms than with that of their excreta. same time as in the shorter experiment and was stable over time, maybe reinforced by aggregation pheromones. In the control experiment the individuals tended to aggregate in 4. Discussion one or the other of both sectors, as indicated by the high level of standard errors (Fig. 3B). Collembola were slightly 4.1. Direct contact with earthworm excreta more abundant in one of both sectors, but this trend was not signi®cant. The mixture of mucus and urine excreted by the earth- The experiment using the device with smaller moist areas worms A. giardi and A. chlorotica strongly attracted H. gave results different from those obtained with a wider nitidus individuals when direct contact was possible. Earth- moist area. In the control experiment with water only the worm mucus is known to affect insect behaviour, either as mean difference between both areas was not signi®cant and an attractant (Morris and Pivnick, 1991), or as a repellent smaller than in the previous experiment and the standard (Laakso and SetaÈlaÈ, 1997). The earthworm Lumbricus error was reduced thus indicating a better balanced distribu- terrestris has been shown to excrete a cutaneous snake- tion (Table 5 and Fig. 3C). The experiment using earthworm attracting compound that acts after contact through the excreta indicated that the number of H. nitidus was always vomeronasal system of the snake (Halpern et al., 1984; Table 4 Numbers of H. nitidus in each sector of a moist area of ®lter paper, in the ®rst set of four odour experiments (means of respectively 16, 13 or 14 times and ten replicates). NS ˆ not signi®cant at 0.05 level; **, *** ˆ signi®cant at the 0.01 and 0.001 levels, respectively Experiments Odour occurrence in sectors Number of H. nitidus (mean ^ SEM) Results of paired t test (probability) Odour of earthworm excreta over 160 min Odour of A. giardi excreta 8.44 ^ 0.94 0.0078** No odour (water) 3.54 ^ 0.92 Control over 130 min No odour (water) 6.12 ^ 0.97 0.8703 NS No odour (water) 5.87 ^ 0.98 Odour of earthworm excreta over 420 min Odour of A. giardi excreta 9.86 ^ 0.66 0.0000*** No odour (water) 2.12 ^ 0.67 Control over 420 min No odour (water) 7.64 ^ 1.58 0.2139 NS No odour (water) 4.37 ^ 1.57  S. Salmon, J.-F. Ponge / Soil Biology & Biochemistry 33 (2001) 1959±1969 1965 Table 5 Numbers of H. nitidus in each of two distinct moist areas of ®lter paper, in three odour experiments over 330 min (means of 11 times and ten replicates). NS ˆ not signi®cant at 0.05 level; ** ˆ signi®cant at 0.01 level Experiments Odour occurrence in sectors Number of H. nitidus Results of paired t (mean ^ SEM) test (probability) Control over 330 min No odour (water) 5.35 ^ 0.76 0.6613 NS No odour (water) 4.93 ^ 0.54 Odour of earthworm excreta over 330 min Odour of A. giardi excreta 7.31 ^ 1.14 0.1252 NS No odour (water) 4.27 ^ 1.18 Odour of earthworm over 330 min Odour of A. giardi 8.03 ^ 0.70 0.0016** No odour (water) 3.38 ^ 0.70 Kirschenbaum et al., 1985; Wang et al., 1988). Attraction were repeatedly photographed under ¯ash. A preference of could be mediated by various nitrogenous molecules which H. nitidus for mucus plus urine was demonstrated with both are contained in the epidermal earthworm mucus (glycopro- A. chlorotica and A. giardi. Although these results teins, peptides, amino acids) and in urine (ammonia, urea) con®rmed the above demonstrated preferences, data are which mingles with mucus (El Duweini and Ghabbour, not shown here because unexpected preferences for a 1971; Cortez and BoucheÂ, 1987). given side were observed in the control experiment with The ®rst two experiments indicated that H. nitidus was water only. The ¯ash was probably responsible for this attracted more rapidly by mucus and urine of A. chlorotica bias as it affected the behaviour of the animals. Nevertheless than by that of A. giardi. This suggests that this kind of these choice experiments using methyl blue stained ®lter attraction varies according to the earthworm species. paper supported the absorption of mucus and urine by However the attraction could vary according to experimen- H. nitidus, since a blue coloration of gut contents was tal conditions and physiological state of earthworms, since observed only in the presence of A. giardi and A. chlorotica in the experiment with methyl blue the attraction by excreta excreta. of A. giardi was more rapid than without staining. In addi- Casts of A. giardi and A. chlorotica did not signi®cantly tion the sensitivity of Collembola to signals, like their attract H. nitidus. At ®rst sight this result is surprising since aggregational habit, could vary whether they are feeding earthworm faeces concentrate a number of nutrients like N or moulting (Joosse and Verhoef, 1974; Bengtsson et al., (Scheu, 1987; Parkin and Berry, 1994), Ca and P (Heine and 1994; Eisenbeis, 1982). Larink, 1993; Sharpley and Syers, 1976; Lunt and Jacobson, Some of the components of earthworm mucus or urine not 1944). Moreover H. nitidus ingest mainly invertebrate only constituted an attractant for H. nitidus but also induced faeces (Arpin et al., 1980; Salmon, unpublished). In our sucking of impregnated ®lter paper. The fact that collembo- experiments, the number of individuals was always greater, lan guts did not contain methyl blue in the absence of mucus although not signi®cantly, in fresh casts than in calcic mull and urine (control experiment with water only) gave indirect made of aged casts. Earthworm faeces contain intestinal evidence that earthworm epidermal excreta were actually mucus but most of the nitrogenous compounds within it absorbed by H. nitidus. This result means that one of the are reabsorbed in the foregut (Martin et al., 1987; Bernier, reasons for the attraction is that H. nitidus may feed on earth- 1998). Bouche et al. (1997) estimated that N excretion by worm mucus or urine. These excreta contain a number of epidermal mucus of earthworms exceeded that by faeces. nitrogen-rich molecules (see above) and epidermal mucus Thus our results reinforce the hypothesis that H. nitidus may also contains easily assimilable carbohydrates (Cortez and be attracted by easily available nitrogen-rich compounds BoucheÂ, 1987) which may be consumed by Collembola. such as amino-acids, proteins, urea and ammonia, which Two individuals found with the apex of their ventral tube can be used by hexapods in protein synthesis (Martin, 1979). stained blue had probably tried to absorb mucus by this way, unless staining occurred during absorption of water only. 4.2. Odour of earthworm and earthworm excreta Nevertheless this mode of absorption could be taken as negligible since the ventral tube is better known to allow Our experiments showed that H. nitidus was attracted by uptake of diluted salt solutions (Eisenbeis, 1982) rather than the odour of A. giardi and to a less extent by the odour of its that of high molecular weight compounds such as mucus excreta (mucus, urine and faeces) although this attraction and methyl blue. Earthworm urine could nonetheless be did not occur systematically. Ammonia and amino acids are absorbed by this way. volatile attractants for a fruit ¯y (Morton and Bateman, A set of choice experiments, not described in this paper, 1981; Bateman and Morton, 1981). Thus, some nitrogenous has been performed in darkness on isolated individuals, thus compounds present in earthworm excreta (see above) might preventing aggregation and interaction with light. Animals be volatile and attract Collembola at a distance. Fig. 3. Differences in the number of Heteromurus nitidus (means of ten replicates ^ standard errors) between two sectors (`earthworm excreta odour' or `earthworm odour' versus `no odour') in odour experiments. (A) Experiments over 160 and 130 min with two sectors in a moist area of ®lter paper. (B) Experiment over 420 min with two sectors in a moist area of ®lter paper. (C) Experiment over 330 min with two distinct moist areas of ®lter paper.  S. Salmon, J.-F. Ponge / Soil Biology & Biochemistry 33 (2001) 1959±1969 1967 The attraction was more ef®cient when H. nitidus was 5. Conclusion allowed to contact mucus and urine (attraction always signi®cant) than when it perceived only the odour of We proved that the mixture of earthworm mucus and excreta. In addition it was necessary to increase the number urine is a strong attractant for H. nitidus while the impact of replicates and individuals in order to detect the effect of of the odour of earthworm excreta on the distribution of this odour. The attraction after direct contact is the common case species is not very pronounced, and the contact with earth- in the interactions between earthworms and other animals worm faeces not ef®cient. We showed as well that H. nitidus (see above mentioned references). In our odour experiments absorbs earthworm mucus. The attraction to earthworm this attraction occurred at 1 cm distance. This is a short mucus, which lines burrow walls (Kretzschmar, 1987), distance compared to ®eld conditions, except if the olfactory could thus explain at least partly the ®eld distribution of signal can be transported by an air stream, like for instance H. nitidus according to a trophic interaction. This conclu- in an earthworm burrow. Moreover the olfactory attraction sion corroborates the results of ®eld studies from which H. of H. nitidus by A. giardi did not occur systematically in our nitidus had been found only in mull at pH . 5 (Ponge, 1993; experiments and some (but not signi®cant) aggregation Salmon and Ponge, 1999) that is a humus form character- arose in one of both sectors even in the absence of earth- ized by a high number and diversity of earthworms (Ponge worm odour. Thus in ®eld conditions the attraction by the et al., 1997). In mull humus at pH , 5 this collembolan odour of earthworm excreta probably interacts with one or species is absent (Ponge, 1993). A choice experiment several other factors. designed to assess the attraction of Collembola by mucus In another experiment (not presented here), Collembola and urine from earthworms living at pH , 5 will enable a exhibited no preference at all for the odour of earthworm better understanding of the distribution of H. nitidus in the excreta. We attributed this negative result to the fact that ®eld. excreta were not moist enough (with regard to other experi- mental runs). Variations from a replicate to another in the same experiment and from an experiment to another could Acknowledgements be due to the fact that the volatility of the chemicals involved in the attraction is weak as in the case for amino We wish to acknowledge Marc TheÂry for lending us light acids (Morton and Bateman, 1981). Another explanation, sensors and data recorders. We are grateful to Marielle compatible with the former, is that light may interact with Peroz and particularly Ceril Techer from the Laboratory, the odour of earthworm excreta. Although experiments were who assisted in sampling earthworms. carried out in darkness, test-boxes were submitted to ambient light for a short time during counting. In fact, light has a strong effect on the distribution of H. nitidus References (Salmon and Ponge, 1998). The effect of light may Arbea, J.I., Jordana, R., 1988. Nota sobre la presencia masiva de persist in darkness, probably following the deposition of Onychiurus folsomi Schaeffer (Collembola, Onychiuridae) en lechos aggregation pheromones (Krool and Bauer, 1987). Interac- de Eisenia andrei (Oligochaeta, Lumbricidae). Boletin de Sanidad tions between light and odour of earthworms or earthworm Vegetal, Plagas 14, 535±540. excreta could partially explain the absence of attraction by Arpin, P., Kilbertus, G., Ponge, J.F., Vannier, G., 1980. Importance de la odour in some replicates as well as the aggregation observed micro¯ore et de la microfaune en milieu forestier. In: Pesson, P. (Ed.). ActualiteÂs d'eÂcologie ForestieÁre. Gauthier-Villars, Paris, pp. 87±150. in the control experiment over 430 min. Our experiments Bateman, M.A., Morton, T.C., 1981. The importance of ammonia in also revealed that clustering interacted with the attraction to proteinaceous attractants for fruit ¯ies (family: Tephridae). Australian earthworm odour. H. nitidus produces an aggregation Journal of Agricultural Research 32, 883±903. pheromone which induces a decrease in its locomotory Bauer, T., Christian, E., 1987. Habitat dependent differences in the ¯ight activity (Verhoef et al., 1977; Verhoef, 1984; Krool and behavior of Collembola. Pedobiologia 30, 233±239. Bayoumi, B.M., 1978. Signi®cance of the microhabitat on the distribution Bauer, 1987). This effect depends on environmental factors of oribatid mites in a hornbeam-oak mixed forest. Opuscula Zoologica like water and food availability, physiological state and Budapest 15, 51±59. population density (Usher and Hider, 1975; Verhoef and Bengtsson, G., Erlandsson, A., Rundgren, S., 1988. Fungal odour attracts Nagelkerke, 1977; Leonard and Bradbury, 1984; Bengtsson soil Collembola. Soil Biology & Biochemistry 20, 25±30. et al., 1994). When individuals ®nd favourable conditions, Bengtsson, G., Hedlund, K., Rundgren, S., 1994. Food- and density- dependent dispersal: evidence from a soil collembolan. Journal of (e.g. in our moist experimental environment), together with Animal Ecology 63, 513±520. the odour of earthworm excreta and darkness, they cluster Bernier, N., 1998. Earthworm feeding activity and development of the and remain motionless. This probably lasts until they need humus pro®le. Biology & Fertility of Soils 26, 215±223. to seek food (Verhoef and Nagelkerke, 1977). As in the BoucheÂ, M.B., 1972. Lombriciens de France. Ecologie et systeÂmatique. present study, the pH preferenda of Folsomia candida I.N.R.A, Paris. BoucheÂ, M.B., Aladdan, F., Cortez, J., Hammed, R., Heidet, J.C., (Van Straalen and Verhoef, 1997) could not be determined Ferriere, G., Mazaud, D., Samih, M., 1997. Role of earthworms in the accurately because of interaction with a similar aggregative N cycle: A falsi®able assessment. Soil Biology & Biochemistry 29, process. 375±380. 1968 S. Salmon, J.-F. Ponge / Soil Biology & Biochemistry 33 (2001) 1959±1969 Brown, G.G., 1995. How do earthworms affect micro¯oral and faunal Lunt, H.A., Jacobson, H.G.M., 1944. The chemical composition of community diversity? Plant & Soil 170, 209±231. earthworm casts. Soil Science 58, 367±375. Cortez, J., BoucheÂ, M., 1987. Composition chimique du mucus cutane de Marinissen, J.C.Y., Bok, J., 1988. Earthworm-amended soil structure: its Allolobophora chaetophora chaetophora (Oligochaeta: Lumbricidae). in¯uence on Collembola populations in grassland. Pedobiologia 32, Comptes-Rendus de l'AcadeÂmie des Sciences de Paris, SeÂrie III 305, 243±252. 207±210. Martin, A., Cortez, J., Barois, I., Lavelle, P., 1987. Les mucus intestinaux Devliegher, W., Verstraete, W., 1997. Microorganisms and soil physico- de Ver de terre moteur de leurs interactions avec la micro¯ore. Revue chemical conditions in the drilosphere of Lumbricus terrestris. Soil d'Ecologie et de Biologie du Sol 24, 549±558. Biology & Biochemistry 29, 1721±1729. Martin, M.M., 1979. Biochemical implications of insect mycophagy. Dubash, P.J., Ganti, S.S., 1964. Earthworms and amino-acids in soil. Biological Review 54, 1±21. Current Science 7, 219±220. Morris, D.E., Pivnick, K.A., 1991. Earthworm mucus stimulates Eisenbeis, G., 1982. Physiological absorption of liquid water by oviposition in a predatory ¯y (Diptera: Anthomyiidae). Journal of Collembola: absorption by the ventral tube at different salinities. Jour- Chemical Ecology 17, 2045±2052. nal of Insect Physiology 28, 11±20. Morton, T.C., Bateman, M.A., 1981. Chemical studies on proteinaceous El Duweini, A.K., Ghabbour, I., 1971. Nitrogen contribution by live earth- attractants for fruit ¯ies, including the identi®cation of volatile worms to the soil. IVth Colloquium Pedobiologiae, 1970, September, constituents. Australian Journal of Agricultural Research 32, 905±916. Dijon, France INRA, Paris, pp. 495±501. Parkin, T.B., Berry, E.C., 1994. Nitrogen transformations associated with Greenslade, P., Fletcher, K.E., 1986. Collembola from earthworm rearing earthworm casts. Soil Biology & Biochemistry 26, 1233±1238. beds at Rothamsted including three new records for Britain. Piearce, T.G., 1972. Acid intolerant and ubiquitous Lumbricidae in selected Entomologist's Monthly Magazine 122, 143±144. habitats in North Wales. Journal of Animal Ecology 41, 397±410. Halpern, N., Schulman, N., Scribani, L., Kirschenbaum, D.M., 1984. Ponge, J.F., 1993. Biocenoses of Collembola in atlantic temperate grass- Characterization of vomeronasally-mediated response-eliciting woodland ecosystems. Pedobiologia 37, 223±244. components of earthworm wash. II. Pharmacology, Biochemistry & Ponge, J.F., Arpin, P., Sondag, F., Delecour, F., 1997. Soil fauna and site Behavior 21, 655±662. assessment in beech stands of the Belgian Ardennes. Canadian Journal Hamilton, W.E., Sillman, D.Y., 1989. In¯uence of earthworm middens on of Forest Research 27, 2053±2064. the distribution of soil microarthropods. Biology & Fertility of Soils 8, Robinson, C.H., Ineson, P., Piearce, T.G., Rowland, A.P., 1992. Nitrogen 279±284. mobilization by earthworms in limed peat soils under Picea sitchensis. Hedlund, K., Bengtsson, G., Rundgren, S., 1995. Fungal odour Journal of Applied Ecology 29, 226±237. discrimination in two sympatric species of fungivorous collembolans. Sadaka-Laulan, N., Ponge, J.-F., Roquebert, M.-F., Bury, E., Functional Ecology 9, 869±875. Boumezzough, A., 1999. Feeding preferences of the collembolan Heine, O., Larink, O., 1993. Food and cast analyses as a parameter of Onychiurus sinensis for fungi colonizing holm oak litter (Quercus turn-over of materials by earthworms (Lumbricus terrestris L.). rotundifolia Lam.). European Journal of Soil Biology 34, 179±188. Pedobiologia 37, 245±256. Saetre, P., 1998. Decomposition, microbial community structure, and Huber, I., 1979. Prey attraction and immobilization by allelomone from earthworm effects along a birch-spruce soil gradient. Ecology 79, nymphs of Womersia strandtmanni (Acarina: Trombiculidae). 834±846. Acarologia 20, 112±115. Salmon, S., Ponge, J.-F., 1999. Distribution of Heteromurus nitidus Jeanson, C., 1972. Etude microscopique de deÂpoÃts de fer, de manganeÁse et (Hexapoda, Collembola) according to soil acidity: interactions with de calcium dans un sol expeÂrimental, leur association avec des earthworms and predator pressure. Soil Biology & Biochemistry 31, microorganismes. Revue d'Ecologie et de Biologie du Sol 9, 479±489. 1161±1170. Joosse, E.N.G., Verhoef, H.A., 1974. On the aggregational habits of Salmon, S., Ponge, J.-F., 1998. Effect of light on a soil-dwelling springtail. surface-dwelling Collembola. Pedobiologia 14, 245±249. European Journal of Soil Biology 34, 199±201. Kaneko, N., McLean, M.A., Parkinson, D., 1995. Grazing preference of Satchell, J.E., 1967. Lumbricidae. In: Burges, A., Raw, F. (Eds.). Soil Onychiurus subtenuis (Collembola) and Opiella nova (Oribatei) for Biology. Academic Press, London, pp. 259±322. fungal species inoculated on pine needles. Pedobiologia 39, 538±546. Scheu, S., 1987. Microbial activity and nutrient dynamics in earthworm Kirschenbaum, D.M., Schulman, N., Yao, P., Halpern, N., 1985. casts (Lumbricidae). Biology & Fertility of Soils 5, 230±234. Chemoattractant for the garter snake: Characterization of Sharpley, A.N., Syers, J.K., 1976. Potential role of earthworms casts for the vomeronasally-mediated response-eliciting components of earthworm phosphorus enrichment of run-off waters. Soil Biology & Biochemistry wash. III. Comparative Biochemistry & Physiology, 82B, 447±453. 8, 341±346. Kozlovskaja, L.S., 1969. Der Ein¯uû der Exkremente von RegenwuÈrmen Shaw, P.J.A., 1988. A consistent hierarchy in the fungal feeding auf die Aktivierung der mikrobiellen Prozesse in TorfboÈden. preferences of the Collembola Onychiurus armatus. Pedobiologia 31, Pedobiologia 9, 158±164. 179±187. Kretzschmar, A., 1987. CaracteÂrisation microscopique de l'activite des Sokal, R.R., Rohlf, F.J., 1995. Biometry. 3rd ed. W.H. Freeman and Co, lombriciens endogeÂs. In: Fedoroff, N., Bresson, L.M., Courty, M.A. New York. (Eds.). Micromorphologie des sols. Soil Micromorphology. AFES, Subler, S., Kirsch, A.S., 1998. Spring dynamics of soil carbon, nitrogen, Paris, pp. 325±330. and microbial activity in earthworm middens in a no-till corn®eld. Krool, S., Bauer, T., 1987. Reproduction, development and pheromone Biology & Fertility of Soils 26, 243±249. secretion in Heteromurus nitidus Templeton, 1835 (Collembola, Thimm, T., Larink, O., 1995. Grazing preferences of some Collembola for Entomobryidae). Revue d'Ecologie et de Biologie du Sol 24, 187±195. endomycorrhizal fungi. Biology & Fertility of Soils 19, 266±268. Laakso, J., SetaÈlaÈ, H., 1997. Nest mounds of red wood ants (Formica Usher, M.B., Hider, M., 1975. Studies on populations of Folsomia candida aquilonia): hot spots for litter-dwelling earthworms. Oecologia 111, (Insecta: Collembola): causes of aggregations. Pedobiologia 15, 565±569. 276±283. Leonard, M.A., Bradbury, P.C., 1984. Aggregative behaviour in Folsomia Van Straalen, N.M., Verhoef, H.A., 1997. The development of a candida (Collembola: Isotomidae) with respect to previous bioindicator system for soil acidity based on arthropod pH preferences. conditioning. Pedobiologia 26, 369±372. Journal of Applied Ecology 34, 217±232. Loranger, G., Ponge, J.F., Blanchart, E., Lavelle, P., 1998. Impact of Verhoef, H.A., 1984. Releaser and primer pheromones in Collembola. earthworms on the diversity of microarthropods in a vertisol (Martini- Journal of Insect Physiology 30, 665±670. que). Biology & Fertility of Soils 27, 21±26. Verhoef, H.A., Nagelkerke, C.J., 1977. Formation and ecological  S. Salmon, J.-F. Ponge / Soil Biology & Biochemistry 33 (2001) 1959±1969 1969 signi®cance of aggregations in Collembola. Oecologia 31, earthworms of a proteinaceous chemoattractant to garter snakes. 215±226. Archives of Biochemistry & Biophysics 267, 459±466. Verhoef, H.A., Nagelkerke, C.J., Joosse, E.N.G., 1977. Aggregation Wickenbrock, L., Heisler, C., 1997. In¯uence of earthworm activity on the pheromones in Collembola. Journal of Insect Physiology 23, 1009± abundance of collembola in soil. Soil Biology & Biochemistry 29, 1013. 517±521. Wang, D., Chen, P., Jiang, X.C., Halpern, M., 1988. Isolation from