Earthworm Communities in Native Savannas and Pastures from A

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Digital.CSIC

1 Reproductive strategies of three native earthworm species from the 2 savannas of Carimagua (Colombia) 3 4 5 Juan J. Jiménez1†, Ana G. Moreno1 & Patrick Lavelle2 6 7 8 1 Departamento de Biología Animal I (Zoología). Facultad de Biología. Universidad 9 Complutense. 28040. Madrid - SPAIN. 10 2 Laboratoire d’Ecologie des Sols Tropicaux ORSTOM. 32 Av. Henri Varagnat, 93143. 11 Bondy - FRANCE 12 13 14 † Corresponding author 15 Phone: + 34-91 394 49 55 16 Fax: + 34-91 394 49 47 17 E-mail: [email protected] 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 This paper is also a portion of the Ph. D. thesis submitted by the first author.

33 Summary 34 35 New data on the reproductive biology of three native glossoscolecid species, 36 Andiodrilus n. sp., Glossodrilus n. sp. and Martiodrilus carimaguensis (Jiménez y 37 Moreno, in press) from the Colombian Llanos are showed. The study was conducted at 38 CIAT-CORPOICA Carimagua Research Station (Meta, Colombia). Two contrasting 39 systems were studied: a native savanna in which no management was performed and a 40 17-yr old grazed pasture. Earthworms and cocoons were hand-sorted from one metre 41 square monoliths combined with washing-sieving techniques. 42 The seasonal dynamics of reproduction varied among the species considered. Both 43 Glossodrilus n. sp. and M. carimaguensis displayed one reproductive period per year, at 44 the end of the rainy season. On the other hand, Andiodrilus n. sp. showed two 45 reproductive periods per year, one at the beginning and the other at the end of the rainy 46 season. Cocoons were laid at different depths till a maximum of 50 cm. 47 The larger the size of the species the larger the cocoon it produced. The relation between 48 cocoon weight and adult weight was the same for Andiodrilus n. sp. and Glossodrilus n. 49 sp., ca. 6%, whereas for M. carimaguensis this value has been the highest obtained up to 50 date for any tropical and temperate earthworm, i.e., 16%. 51 Cultures under controlled conditions were performed to determine the incubation time 52 and the number of individuals hatched by cocoon. 53 54 55 Key words: Earthworms, Reproductive strategies, Cocoons, Life cycles, Vertical 56 distribution 57

58 Introduction 59 60 Reproductive patterns, as time of cocoon deposition, number of cocoons 61 produced, their location in the soil profile, incubation time and hatching rate, are 62 considered basic biological features regarding the life cycles of earthworms. These 63 patterns have been described by numerous authors for both temperate and tropical 64 earthworms; however, the number of species studied till date is still low. Studies on any 65 of one or several aspects of the reproductive strategies are those of Stephenson (1930), 66 Walsh (1936), Evans and Guild (1948), Gerard (1967), Satchell (1967), Bouché (1972), 67 Reynolds (1973), Nowak (1975), Phillipson and Bolton (1977), Lavelle (1971; 1978; 68 1979), Senapati (1980) and Garnsey (1994) among others. 69 70 Bouché (1977) and Lavelle (1977) have defined a close relationhip between 71 reproductive strategies and ecological categories of earthworms. Epigeic worms, living 72 and feeding on litter, ussually come across environmental stress through cocoons. 73 Endogeic earthworms, living and feeding in the soil, show a resistance form named 74 quiescence during unfavourable conditions and enter rapidly into reproductive period 75 when conditions become suitable. Finally, anecic earthworms, living in the soil but 76 feeding in the surface, employ two mechanisms to avoid mortality within the 77 population: diapause, that may be different for adults and juveniles and cocoons at the 78 end of the rainy season (Jiménez et al., 1998). 79 80 There is a very scarce information regarding basic knowledge on the biology and 81 ecology of earthworm species from the Neotropical region as only a few studies have 82 been undertaken. Németh (1981) studied the earthworm fauna from the tropical forests 83 of Venezuela, Fragoso (1993) in natural and disturbed environments in the South of 84 Mexico, Feijoo (unpubl.) in the hillsides from Valle del Cauca in Colombia and Muñoz- 85 Pedreros (1997) in Chile. In the Eastern plains of Colombia any study on earthworm 86 communities began before 1993 when the International Centre for Tropical Agriculture 87 (CIAT) and the European Community established a cooperation to evaluate the 88 dynamics of earthworm communities in the natural savannas and their possibilities of 89 management in the near future. 90

91 In an earlier paper (Jiménez et al., 1998) some data about the reproductive 92 strategy of M. carimaguensis were shown. In this paper a more precise detail of life 93 cycle and the dynamics of the reproduction is given for three native species from the 94 eastern plains of Colombia. 95 96 97

98 Materials and Methods 99 100 Study site 101 102 The study area is located at the CIAT-CORPOICA (Centro Internacional de 103 Agricultura Tropical and Corporación Colombiana de Investigación Agropecuaria 104 agreement) Carimagua Research Station, in the well-drained isohyperthermic savannas 105 of the Eastern Plains of Colombia (4 37' N and 71 19' W and 175 meters altitude). 106 Average annual rainfall and temperature are about 2280 mm and 26°C respectively, with 107 a dry season from December to March. Soils are of two types: predominant low-fertility 108 Oxisols in the upland (“altos”) and Ultisols in the low-lying (“bajos”) savannas. The

109 former are characterized by their acidity (pH (H20) 4.5), a high Al saturation (> 90%) 110 and low values of exchangeable Ca, Mg and K. Chemical factors that contribute to acid- 111 soil infertility and subsequent effects on plant growth are complex and include Al 112 toxicity, low content of available P and low rates of N mineralization (Rao et al., 1993). 113 Two different and contrasting systems were evaluated: A native savanna where 114 the predominant plant species are Andropogon bicornis, Gymnopogon sp., Panicum 115 spp., Trachypogon spp. and Imperata sp., and a 17-yr old grazed improved pasture that 116 associates an exotic african grass, Brachiaria decumbens cv. Basilisk and a tropical 117 forage herbaceous legume species, Pueraria phaseoloides CIAT 9900 (“kudzu”). 118 Stocking rates for the latter system are 1 cattle ha-1 in the dry season and 2 cattle ha -1 in 119 the wet season, whereas no management is performed in the former. 120 121 Earthworm sampling 122 123 The experimental designed was carried on the basis of a randomly stratified 124 sampling procedure. In both systems, a defined area of 90x90 m was isolated and 125 divided into regular quadrats of 10x10 m. Two physical methods were performed during 126 17 months (from april 1994 to september 1995, except june 1994): hand sorting and 127 washing-sieving of respective five monthly 1x1x0.5 m and ten 20x20x20 cm soil 128 monoliths (after Lavelle, 1978) within five randomly-chosen quadrats in each plot. 129 Monoliths were split into 10 cm layers to assess the vertical pattern and cocoons

130 collected from each layer washed in water and carried to the laboratory. See Lavelle 131 (1978), Jiménez et al (1998) and Jiménez (in press) for sampling details. 132 133 Incubation of cocoons 134 135 Cocoons obtained in field samples were weighed and deposited in Petri dishes 136 with moistened filter papers. Each day cultures were remoistened and hatched 137 individuals weighed. The temperature in the laboratory was maintained between 26 and 138 28º C. 139 140 Results 141 142 Mating could never be observed (this probably happens by night but I was 143 warned not to go outside the station for security reasons) so it has been assessed from 144 indirect methods. The reproductive period may be related to some morphological 145 features, a prominent clitellum in adults and cocoons extracted in the field. The time of 146 the year by which the adults show a conspicuous clitellum varied among species: 147 Andiodrilus n. sp. At the beginning of the rainy season (april-may), M. carimaguensis in 148 the middle (july-august) and Glossodrilus n. sp. before the entry of the dry season 149 (september-october). 150 151 Cocoons: Morphology, size and incubation time 152 153 The cocoons produced by Glossodrilus n. sp. are whitish, slightly spherical 154 (Table 1), with the end sharp-pointedly extended. Their average size was 3.0x2.2 mm 155 and 6.2 mg of fresh weight in vivo. Any of the 79 cocoons collected in the field could 156 emerge under laboratory conditions, so incubation time, number and weight of newly 157 hatched worms could not be determined for this species. 158 159 Cocoons of Andiodrilus n. sp. are also spherical, whitish translucent with a mean 160 diameter of 5 mm and a fresh weight of 78.9 mg. An unique zygote can be observed 161 through the cocoon with an incubation period of 13 days and a fresh weight of 55.7 at

162 hatching. Almost 50% of the overall amount of cocoons did not hatch in the laboratory 163 and 71% of the total weight of the cocoon corresponded to earthworm weight. 164 165 The cocoon of M. carimaguensis was yellowish, ovoid in shape, rather large 166 (23.6 x 14 mm) and weighing on average 1.808 mg. In each cocoon the number of 167 embryos emerged was two (one individual emerged from only one cocoon) weighing 168 760 mg (Figure 1). The average incubation time for this species, considering all the 169 cocoons, was 23.5 days and 80% of the cocoon weight was given by the worms. 170 Considering all the cocoons cultured under controlled conditions the hatching ratio was 171 73.9%. 172 173 Seasonal dynamics 174 175 The seasonal dynamics of cocoon deposition, both number and location in the 176 soil profile, should help us to establish the dynamics of the reproductive period as well 177 as the adaptive strategies employed by the different species. Glossodrilus n. sp. and M. 178 carimaguensis showed an unique reproductive period whilst Andiodrilus n. sp. seemed 179 to display a bimodal dynamics. 180 181 Glossodrilus n. sp. laid its cocoons at the end of the rainy period (Figure 2). The 182 first cocoons in the savanna appeared in october 1994 till a maximum peak in january 183 when all the population, totally inmatures, were inactive. In the pasture, on the contrary, 184 the first cocoons were found in september 1994 with a maximum number obtained in 185 february. Mainly the total number of cocoons hatch at the onset of the following rainy 186 season. 187 188 Cocoons of Andiodrilus n. sp. were found at two different periods along the 189 rainy season (Figure 3). In the savanna cocoons were collected at the beginning, with a 190 maximum number, and in the middle of the rainy season. In the second year of the study 191 two peaks appeared in the same months, though number of cocoons were lower than 192 that obtained in the first year. In the pasture a high number of cocoons were obtained 193 compared with the savanna. A peak was also observed in may 1994 followed by another

194 cocoon deposition in october. By the second year the higher values appeared in may- 195 june and august. 196 197 M. carimaguensis deposited its cocoons at the end of the rainy season and the 198 maximum number obtained for this species was in september-october (Figure 4). In the 199 following year some cocoons were obtained in march what confirms that some cocoons 200 do not hatch until the next season. Only one cocoon was collected in the savanna by july 201 1994, so comparisons with this system are not possible. 202 203 Vertical patterns 204 205 The depth at which cocoons were laid has been assessed from the total number 206 of cocoons collected in each stratum (Table 2a, 2b). Andiodrilus n. sp. and Glossodrilus 207 n. sp. are topsoil species so most of the cocoons were found in the first 20 cm of soil 208 profile. The mean cocoon deposition depth for Andiodrilus n. sp. was 6.5 cm and 5.6 cm 209 in respectively savanna and pasture systems (Figure 5). 210 211 The mean depth for Glossodrilus n. sp. was 8.8 cm in the savanna and 12.4 cm in 212 the pasture (Figure 6). These data were referred to the number of cocoons obtained by 213 washing/sieving technique; since some cocoons were obtained below 20 cm depth in the 214 hand-sorting method a correction in the percentage of cocoons found was needed. For 215 this reason, we have considered that 95% of the total number of cocoons were located in 216 the first 20 cm and 5% in the 20-30 cm stratum in the savanna. Data form one metre 217 square samples showed that 85% of the cocoons were placed in the 0-20 cm stratum and 218 15% in the 20-40 cm stratum. Probably a little fraction of cocoons was placed deeper 219 than 20 cm in the soil and they could not be obtained when applying the washing- 220 sievieng technique.That is why data obtained from handsorting of one metre square 221 samples were considered as valid assessment. 222 223 The vertical distribution of cocoons for M. carimaguensis could only be 224 determined in the pasture as only one cocoon was obtained in the savanna. The cocoons 225 were found at an average depth of 26 cm reaching a maximum of 50 cm. More than 226 50% of the cocoons were placed in the 20-30 cm stratum (Figure 7).

227 228 Cocoon size versus adult size 229 230 According to Lavelle (1978) the relationship between the weight of cocoons and 231 adults has been assess in order to determine the amount of energy, expressed as adult 232 weight percentage, that is invested in the development of cocoons (Table 3). 233 234 The higher the size of the adult the larger the cocoon it forms (Figure 8), 235 although the relationship between the variables considered was the same for two species 236 differing in size: Andiodrilus n. sp. and Glossodrilus n. sp. However, this value rose up 237 to 16% of adult weight for M. carimaguensis. 238 239 Fecundity 240 241 Estimates of fecundity are based on field data from july 1994 to june 1995. It 242 was necessary to use a correction factor to recalculate density of adults for Glossodrilus 243 n. sp. as it has been proved that almost 50% of the population was lost when hand- 244 sorting method was employed (Jiménez, unpubl.). The higher values of fecundity, i.e. 245 number of cocoons produced by adult by year, appeared for Glossodrilus n. sp. This 246 species produced 9.9 cocoons.adult-1.year-1 in the savanna and 13 cocoons.adult-1.year-1 247 in the pasture. Andiodrilus n. sp. and M. carimaguensis presented lower values, 0.61 and 248 0.50 for the former in respectively savanna and pasture systems and 0.25 for the latter 249 (actually, this values was 0.49, the product of 0.25 by 1.91 individuals per cocoon). 250 251 Discussion 252 253 Factors affecting reproductive activity of earthworms are both temperaure and 254 moisture in the soil (Evans & Guild, 1948). Michalis and Panidis (1993) concluded that 255 both factors affect the reproductive pattern for Octodrilus complanatus and similar 256 observations were made by Reinecke and Kriel (1980) for the compost worm E. fetida. 257 At Carimagua the strong seasonal climatic variations restrict cocoon deposition to only 258 8 months as the rest of the year the earthworm community is inactive at all. In general, 259 two different strategies appear:

260 261 1. Continuos deposition of cocoons for Andiodrilus n. sp. 262 2. Cocoon deposition at the end of the rainy season and diapause or death of 263 adults for respectively M. carimaguensis and Glossodrilus n.sp. 264 265 Cocoon size is not always correlated with adult size (Edwards & Bohlen, 1996). 266 However , Lavelle (1981) found a clear positive relationship between the size of the 267 adults and the cocoons produced for temperate and tropical earthworms. In this study we 268 agree with the latter, the larger the adult the bigger the cocoon. But we must not draw 269 any conclusion regarding this matter; the high reproduction investment made by M. 270 carimaguensis may be the result of a more complex strategy before its inactivation. The 271 value obtained for this species has been the highest to date, and that is why it is an 272 outlier in Figure 8. So cocoon size should be also related to adaptive strategies for those 273 species living in ecosystems with a strong seasonality. 274 275 Some of our results agree with others obtained in the tropical region. Lavelle 276 (1971; 1978) reported that the hatching ratio for M. anomala under laboratory conditions 277 is almost 70% and that some dehydrated cocoons by the dry period hatch at the onset of 278 the rainy season. In this study we found lower hatching ratio values for a similar species 279 in both size, 5mm diameter and 50 mg weight and quiescence pattern., Andiodrilus n. 280 sp. 281 282 Most of unhatched cocoons showed fungal infection over its surface. This could 283 determine the low values of hatching in the laboratory, although we cannot assure if 284 these fungi were the cause or the result of cocoon non-viability, as it has been 285 mentioned by Senapati (1980) for Indian worms. 286 287 It seems to exist a clear relationship between the number of cocoons produced 288 and its location in the soil profile (Satchell, 1967). He concluded that species in a high 289 risk mortality environment have a greater rate of cocoon production. In this sense 290 Glossodrilus n. sp. produced the larger amount of cocoons and its location is preferably 291 in the first 10 cm of soil (Jiménez, unpubl.). Many of these cocoons will not hatch and 292 will die, a typical r-strategy species (Pianka, 1970). On the contrary, M. carimaguensis

293 produced much fewer amount of cocoons, having the ability to descend deeper into the 294 soil to aestivate, a K-strategy species. 295 296 Differences in cocoon production between the savanna and pasture may be partly 297 explained by external factors, such as quality of organic matter input. The dynamics of 298 cocoon production showed that earthworms mainly laid their cocoons at both the 299 beginning and end of each wet season, when fluctuation of soil moisture figures are 300 higher in the soil profile. 301 302 There is a population survival through cocoons during the summer. If cocoons of 303 Glossodrilus n. sp. have not hatched before the entry of the dry season, cocoon 304 development is delayed and hatching will take place at the onset of a new wet season. 305 This pattern is similar to that observed by Lavelle (1978) for M. anomala. 306 307 Acknowledgements 308 309 This study was included within the STD-3 European Macrofauna Project and the 310 Tropical Lowlands Program at the International Centre for Tropical Agriculture, CIAT 311 (Cali, Colombia). 312 313 314 References 315 316 Barois, I., Ángeles, A., Blanchart, E., Brossard, M., Fragoso, C., Jiménez, J. J., 317 Martínez, M. A., Moreno, A. G., Lattaud, C., Lavelle, P., Rossi, J.-P., Senapati, 318 B. K., Giri, S & Tondoh, J. (1996). Ecology of species with large environmental 319 tolerance and/or extended distributions. In: P. Lavelle (Ed.). Conservation of Soil 320 Fertility in Low-Input Agricultural Systems of the Humid Tropics by 321 Manipulating Earthworm Communities. Macrofauna Project Final Report. 322 Bouché, M. B. (1972). Lombriciens de France. Ecologie et Systematique. INRA Publ. 323 72-2, Paris. 671 p. 324 Bouché, M. B. 1977. Stratégies lombriciennes. Écol. Bull. 25, 122-132.

325 Edwards, C. A. & Bohlen, P. J. (1996). Biology and Ecology of Earthworms. 3rd edition. 326 Chapman and Hall, London, UK.426p. 327 Evans, A. C. & Guild, W. J. McL. (1948). Studies on the relationships between 328 earthworms and soil fertility. IV. On the life cycles of some British Lumbricidae. 329 Ann. Appl. Biol. 35, 471-484. 330 Fragoso, C. (1993). Les peuplements de vers de terre dans l’est et sud-est du Mexique. 331 These de Doctorat, Université Paris VI. 228 p + annexes. 332 Garnsey, R. B. (1994). Seasonal activity and aestivation of lumbricid earthworms in the 333 Middlands of Tasmania. Aust. J. Soil Res. 32, 1355-1367. 334 Gerard, B. M. (1967). Factors affecting earthworms in pastures. J. Anim. Ecol. 36, 235- 335 252. 336 Jiménez, J. J. y Moreno, A. G. Martiodrilus carimaguensis (Oligochaeta, 337 Glossoscolecidae), Una nueva especie de lombriz de tierrra para Colombia. 338 Megadrilogica, (in press). 339 Jiménez, J. J., Moreno, A. G., Lavelle, P. and Decaëns, T. 1998. Population dynamics 340 and adaptive strategies of Martiodrilus carimaguensis (Oligochaeta, 341 Glossoscolecidae), a native species from the well- drained savannas of 342 Colombia. App. Soil Eco.???? 343 Jiménez, J. J., Moreno, A. G., Decaëns, T., Lavelle, P., Fisher, M.J. & Thomas, R. J. 344 Earthworm communities in native savannas and man-made pastures of the 345 eastern plains of Colombia. Biol. Fertil. Soils (in press). 346 Lavelle, P. (1971). Recherches ecologiques dans la savane de Lamto (Côte d’Ivoire): 347 Production annuelle d’un ver de terre Millsonia anomala, Omodeo. La terre et la 348 vie, 2-71, 240-254. 349 Lavelle, P. (1977). Bilan energetique des populations naturelles du ver de terre 350 geophage Millsonia anomala (Oligochetes-Acanthodrilidae) dans la savane de 351 Lamto (Côte d'Ivoire). Geo-Eco-Trop 1 (2), 149-157. 352 Lavelle, P. (1978). Les vers de terre de la savane de Lamto (Côte d'Ivoire): peuplements, 353 populations et fonctions dans l'ècosystème. Thèse de Doctorat, Paris VI. Publ. 354 Lab. Zool. E.N.S., 12, 301 pp. 355 Lavelle, P. (1979). Relations entre types écologiques et profils démographiques chez les 356 vers de terre de la savane de Lamto (Côte d’Ivoire). Rev. Ecol. Biol. Sol 16, 85- 357 101.

358 Lavelle, P. (1981). Stratégies de reproduction chez les vers de terre. Acta Oecologica 2 359 (2), 117-133. 360 Michalis, K., and Panidis, S. (1993). Seasonal variation in the reproductive activity of 361 Octodrilus complanatus (Oligochaeta, Lumbricidae). Eur. J. Soil Biol. 29 (3-4), 362 161-166. 363 Muñoz-Pedreros, A., Ruiz, E., Poblete, C. Y Santelices, M. (1997). Aspectos de la 364 biología reproductiva de lumbrícidos silvestres (Oligochaeta: Lumbricidae) en el 365 sur de Chile. Rev. Chil. Hist. Nat. 70, 101-108. 366 Németh, A. (1981). Estudio ecológico preliminar de las lombrices de tierra (oligochaeta) 367 en ecosistemas de bosque húmedo tropical en San Carlos de Río Negro, 368 Territorio Federal Amazonas. Tesis de grado, Universidad Central de Venezuela. 369 92 p. 370 Nowak, E. (1975). Population density of earthworms and some elements of their 371 production in several grassland environmments. Ekol. Pol. 23, 459-491. 372 Phillipson, J. And Bolton, P. J. (1977). Growth and cocoon production by 373 Allolobophora rosea (Oligochaeta: Lumbricidae). Pedobiologia 17, 70-82. 374 Pianka, E. R. (1970). On r- and K-selection. Am. Nat. 104, 592-597. 375 Rao, I. M., Zeigler, R. S., Vera, R. And Sarkarung, S. (1993). Selection and breeding for 376 acid-soil tolerance in crops. Upland rice and tropical forages as case studies. 377 Bioscience 43, 454-465. 378 Reinecke, A. J. And Kriel, J. R. (1980). Influence of temperature on the reproduction of 379 the earthworm Eisenia fetida (Oligochaeta). S. Afr. J. Zool. 16 (2), 96-100. 380 Reynolds, J. W. (1973). Earthworm (Annelida: Oligochaeta) ecology and systematics. 381 In: D. Dindal (Ed.) Proceedings 1st Soil Microcommunities Conf., Syracuse, 382 N.Y., October 1971. Pp. 95-120. 383 Satchell, J. E. (1967). Lumbricidae. In: A. Burges and F. Raw (eds.). Soil Biology. pp 384 259-322. Academic Press, London. 385 Senapati, B. K. (1980). Aspects of ecophysiological studies on tropical earthworms. 386 Distribution, population dynamics, production, energetics and their role in the 387 decomposition process. Ph. D. Thesis, Sambalpur University, India. 154 p. 388 Stephenson, J. (1930). The Oligochaeta. Oxford, Clarendon Press. 389 Walsh, A. S. (1936). Lumbricus lore. Turtox Nevs 14, 57-60. 390

391 Table caption 392 393 Table 1. Main biological features of cocoons obtained for three species studied1 394 395 Tabla 2a. Vertical distribution and average depth at which cocoons were collected 396 in the savanna system 397 398 Tabla 2b. Vertical distribution and average depth at which cocoons were collected 399 in the pasture system 400 401 Table 3. Investment of individual weight into cocoon weight for three species from 402 Carimagua 403

404 Table 1. Main biological features of cocoons obtained for three species studied1 405 Glossodrilus n. sp. Andiodrilus n. sp. M. carimaguensis Cocoons studied 79 88 46 Morphology Spherical Spherical Ovoid Size (mm) 3,0x2,2 6 23,6x14 Fresh weight in vivo (mg) 6,2  1,4 78,9  22,2 1.808  414,5 (4,6-8,1) (30-130) (890-3020) Incubation time (days) ? 12,8  6,3 (1-28) 23,3  12,9 (1-48) Individuals per cocoon ? 1,02  0,15 1,91  0,29 Weight of newly hatched 52 55,7  22,9 760  219,4 individuals (mg) (20-120) (270-1760) Individual weight / 80,6 71 79,7 Cocoon weight (%) Hatching rate under 0 48,3 73,9 controlled conditions (%) 406 1 Mean  standard deviation (minimum and maximum within brackets) 407 2 Data from individuals in the field 408

409 Tabla 2a. Vertical distribution and average depth at which cocoons were collected 410 in the savanna system 411 Stratum (cm) Species Average depth 0-10 10-20 20-30 30-40 40-50 Glossodrilus n. sp.* 8.7 67.5 27.5 5.0 - - Andiodrilus n. sp. 6.5 85.0 15.0 - - - M. carimaguensis ------412 * Data from washin and sieving technique 413 414 415 Tabla 2b. Vertical distribution and average depth at which cocoons were collected 416 in the pasture system 417 Stratum (cm) Species Average depth 0-10 10-20 20-30 30-40 40-50 Glossodrilus n. sp.* 12.4 48.2 36.8 7.5 7.5 - Andiodrilus n. sp. 5.6 94.1 5.9 - - - M. carimaguensis 26.0 1.6 18.0 54.1 21.3 4.9 418 * Data from washin and sieving technique 419

420 Table 3. Investment of individual weight into cocoon weight for three species from 421 Carimagua Species Adult weight (mg) Cocoon weight (mg) Cocoon w./Adult w. (%) Glossodrilus n. sp. 110.6 (n = 32) 6.2 (n = 60) 6.2 Andiodrilus n. sp. 1.340 (n = 26) 78.9 (n = 88) 5.9 M. carimaguensis 11.240 (n = 29) 1.808 (n = 46) 16.1 422 423

424 Figure caption 425 426 Figure 1. Box-plot of weight of juveniles of M. carimaguensis showing the mean 427 (dotted line), 5% and 95% percentiles. 428 429 Figure 2. Monthly average number of cocoons (N m-2) collected in both systems for 430 Glossodrilus n. sp. (data from washing-sieving technique). 431 432 Figure 3. Monthly average number of cocoons (N m-2) collected in both systems for 433 Andiodrilus n. sp. (data from hand sorting of 1m2 monoliths) 434 435 Figure 4. Monthly average number of cocoons (N m-2) collected in the pasture for 436 M. carimaguensis (data from hand sorting of 1m2 monoliths) 437 438 Figura 5. Yearly average vertical distribution of cocoons for Glossodrilus n. sp. in 439 the savanna (grey) and in the pasture (black ); (washing-sieving technique). 440 441 Figura 6. Yearly average vertical distribution of cocoons for Andiodrilus n. sp. in 442 the savanna (grey) and in the pasture (black); (hand-sorting method). 443 444 Figure 7. Yearly average vertical distribution of cocoons for M. carimaguensis in 445 the pasture system (Hand-sorting method). 446 447 Figura 8. Relationship between fresh weight (Log natural scale) of cocoons and 448 adults for 38 species from tropical () and temperate sites () plus three 449 species from Carimagua () (p<0.01). Data were taken from Lavelle (1981) 450 and Barois et al (1996). 451 452

1400

1200

1000

800

600

400

Juvenile weight (mg)

200

0 Lower individual Higher individual weight weight

453 454 Figure 1. Box-plot of weight of juveniles of M. carimaguensis showing the mean 455 (dotted line), 5% and 95% percentiles. 456 457

458

2200 Savanna 2000 Pasture 1800 1600 1400

-2 1200 1000

N x 5m N x 800 600 400 200 0 M94 A M JL A S O N D J F M A M J JL A S95 Months

459 460 Figure 2. Monthly number of cocoons collected in both systems for Glossodrilus n. 461 sp. (data from washing-sieving technique). 462 463

50 Savanna Pasture

-2 30

N x 5m 20

0 M94 A M JL A S O N D J F M A M J JL A S95 Months

464 465 Figure 3. Monthly number of cocoons collected in both systems for Andiodrilus n. 466 sp. (data from hand-sorting method) 467

-2

N x 5m

0 M94 A M JL A S O N D J F M A M J JL A S95 Months

468 469 Figure 4. Monthly number of cocoons collected in the pasture for M. 470 carimaguensis (data from hand-sorting method) 471

0-10

10-20

20-30 N = 302 N = 895 30-40 X = 8.7 cm X = 12.4 cm

40-50

Stratum (cm) 50-60 Savanna 60-70 Pasture

70-80

100 80 60 40 20 0 20 40 60 80 100 472 Population (%) 473 Figura 5. Yearly average vertical distribution of cocoons for Glossodrilus n. sp. in 474 the savanna (grey) and in the pasture (black ); (washing-sieving technique). 475 476 477

0-10

10-20

20-30 N = 20 N = 51 30-40 X = 6.5 cm X = 5.6 cm

40-50

Stratum (cm) 50-60 Savanna 60-70 Pasture

70-80

100 80 60 40 20 0 20 40 60 80 100 478 Population (%) 479 Figura 6. Yearly average vertical distribution of cocoons for Andiodrilus n. sp. in 480 the savanna (grey) and in the pasture (black); (hand-sorting method). 481

0-10

10-20

20-30

30-40

40-50 N = 61 Stratum (cm) 50-60 X = 26.0 cm

60-70

70-80

100 80 60 40 20 0 20 40 60 80 100

482 Population (%) 483 Figure 7. Yearly average vertical distribution of cocoons for M. carimaguensis in 484 the pasture system (Hand-sorting method). 485 486

8 Mca 7 6

5 And 4

3 r = 0.884 Glo 2

Cocoon weight (mg) weight Cocoon 1 0 345678910 11 Adult weight (mg)

487 488 489 Figura 8. Relationship between fresh weight (Log natural scale) of cocoons and 490 adults for 38 species from tropical () and temperate sites () plus three 491 species from Carimagua () (p<0.01). Data were taken from Lavelle (1981) 492 and Barois et al (1996). 493 494