Which web to invade? Argyrodine kleptoparasites differentiate amongst architecturally different host webs

Biodiversity Data Journal 13: e172146 doi: 10.3897/BDJ.13.e172146 Research Article Which web to invade? Argyrodine kleptoparasites differentiate amongst architecturally different host webs Ingi Agnarsson , Fi-Niaina Ramahefarison , Heiða Hlín Matthíasdóttir , Leyla Kudari , Magnús Máni Dagsson , Nína Guðrún Baldursdóttir , Ragnhildur Sara Bergsdóttir , Rakel Bærings Halldórsdóttir , Snorri Björn Magnússon , Snædís Huld Björnsdóttir , Matjaž Gregorič ‡ Faculty of Life and Environmental Sciences, School of Engineering and Natural Sciences, University of Iceland, Reykjavik, Iceland § 2Department of Entomology, National Museum of Natural History, Smithsonian Institution, Washington DC, United States of America | 3Department of Entomology, University of Madagascar, Antananarivo, Madagascar ¶ Research Centre of the Slovenian Academy of Sciences and Arts, Jovan Hadzi Institute of Biology, Ljubljana, Slovenia # Postgraduate School ZRC SAZU, Ljubljana, Slovenia Corresponding author: Ingi Agnarsson ([email protected]) Academic editor: Emma McCarroll Shaw Received: 15 Sep 2025 | Accepted: 19 Nov 2025 | Published: 15 Dec 2025 Citation: Agnarsson I, Ramahefarison F-N, Matthíasdóttir H, Kudari L, Dagsson M, Baldursdóttir N, Bergsdóttir R, Halldórsdóttir R, Magnússon S, Björnsdóttir S, Gregorič M (2025) Which web to invade? Argyrodine kleptoparasites differentiate amongst architecturally different host webs. Biodiversity Data Journal 13: e172146. https://doi.org/10.3897/BDJ.13.e172146 Abstract Kleptoparasitism, the theft of resources from another organism, is a widespread foraging strategy amongst animals. Many Argyrodinae cobweb spiders (Theridiidae) are obligatory kleptoparasites that have abandoned web building, relying instead on webs of larger hosts (kleptotany). Theory predicts that kleptoparasites are not randomly distributed amongst host webs and prior studies indicate that web architecture (size and tenure) and prey availability impact kleptoparasite abundance and host choice. This study is the first to examine argyrodine biology in Madagascar and provides an insight into a multi-species endemic community of spider kleptoparasites and their in-situ distribution across contrasting host webs: Nephilingis (Nephilidae, large nocturnal orb weavers), Caripetella (Pisauridae, large nocturnal sheet web builders) and Anelosimus ‡,§ | ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ¶,# © Agnarsson I et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. (Theridiidae, small cathemeral social spiders constructing 3D tangle webs). We found a striking diversity of nine kleptoparasite species in six genera that, remarkably, are not conspecific with the five species that were previously known from all of Madagascar. Kleptoparasite species composition and abundance varied across the three hosts: some appeared host-specific, while others are versatile. In general, argyrodine kleptoparasites evidently discriminate amongst hosts, but differ in the degree of host preference. At the community level, we speculate that species assembly into host webs involves a complex interplay between host preference and species competitive hierarchy. Future field research should investigate this system as a community of multiple interacting species to gain a more comprehensive understanding of the rules that may govern the assembly of diverse kleptoparasites into equally variable host webs. Keywords Argyrodinae, host choice, kleptoparasitism, kleptotany, spiders Introduction Species interactions range from indirect to direct and from mutually beneficial to hostile. Symbiotic relationship have been defined as “an intimate interaction between different organisms, where at least one of the parties is obligatorily dependent on the association as a part of its life history”. (Leung and Poulin 2008 p. 107). Symbiosis can be mutualistic, commensal or parasitic (Campbell and Reece 2005, Odum and Barrett 2005, Leung and Poulin 2008). Kleptoparasitism is a widespread exploitative foraging strategy involving the theft of extrinsic resources that are of potential value to the host and involves balancing the energy costs of searching for food against the benefits and costs of stealing resources (Brockmann and Dawkins 1979, Vollrath 1979, Elgar 1993, Iyengar 2008, Agnarsson 2025). While facultative kleptoparasitism is common (e.g. Iyengar (2008)), natural selection has engrained this strategy in some groups, where host-selection may be a more critical foraging decision than the exact nature of the stolen resource. Precious little, however, is known about host choice in many kleptoparasitic organisms like spiders (Agnarsson 2025). Obligate kleptoparasitic spiders are largely kleptotanic, in that they abandon web building and rely on foraging symbiotically in the webs of larger hosts (Vollrath (1984), Vollrath (1987), Whitehouse (1986), Elgar (1993), Grostal and Walter (1997), Whitehouse (2011), Su (2012), Su and Smith (2014), Agnarsson (2025)). The majority of spider kleptoparasites belong to the subfamily Argyrodinae (Theridiidae), a versatile group employing diverse strategies to exploit various host species (Agnarsson 2004). Amongst their hosts are: 1) large nephilid orb weavers such as Nephila, Trichonephila and Nephilingis, that seem to be preferred hosts in most subtropical and tropical environments (Elgar 1993, Whitehouse et al. 2002, Agnarsson 2025); 2) large araneid orb weavers such as Argiope, Cyrtophora and Gasteracantha; 3) large sheet weavers including various agelenids, the pisaurid Caripetella, the zoropsid Tengella and others; 2Agnarsson I et al 4) social and colonial spiders (mostly Araneidae, Eresidae and Theridiidae) that build complex silken networks and 5) some smaller species (notably Theridiidae and Linyphiidae) that are locally abundant (Vollrath 1979, Eberhard et al. 1993, Leborgne et al. 1998, Hénaut 2000, Miyashita 2002, Lloyd and Elgar 2006, McCrate and Uetz 2009, Kuntner and Agnarsson 2010, Silva and Sierwald 2013, Agnarsson 2025). Argyrodinae spiders are speciose (Exline and Levi 1962, Agnarsson 2004, Vanuytven et al. 2024, World Spider Catalog Version 26 2025, Pett et al. 2025) and vary in size, morphology and tactics to obtain resources from host webs (Elgar 1993, Whitehouse 2011) including pilfering small insects, stealing wrapped prey, feeding on digested prey alongside the host, opportunistic predation on the host, its offspring and eggs and even the consumption of host web silk (Whitehouse 1986, Vollrath 1987, Elgar 1993, Miyashita et al. 2004). Their hosts are also diverse, differing in size, innate aggression (solitary vs. communal/social), web architecture and prey choice (Agnarsson 2025). Fundamental foraging and game theories predict that kleptoparasitic species do not distribute randomly amongst potential host webs (MacArthur and Pianka 1966, Gotelli 2008, Iyengar 2008, Gregorič et al. 2024). The limited field observations corroborate these predictions: 1) distribution of kleptoparasites is non-random amongst hosts; 2) large and predictable webs are favoured by kleptoparasites, with large webs, smaller webs and those that have short duration in space and time being disfavoured; 3) larger webs have more kleptoparasite individuals and species and 4) kleptoparasite species range from versatile generalists to relative host specialists (Vinson 1863, Exline 1945, Vollrath 1987, Elgar 1993, Whitehouse et al. 2002, Agnarsson 2003, Agnarsson and Kuntner 2005, Whitehouse 2011, Gregorič et al. 2024). However, most prior studies focus on a single kleptoparasite or a single host species, while many do not identify the kleptoparsites to the species level, complicating comparisons (Whitehouse 1986, Cangialosi 1990, Eberhard et al. 1993, Whitehouse and Jackson 1993, Grostal and Walter 1997, Tso and Severinghaus 1998, Hénaut et al. 2005, Baba et al. 2007, McCrate and Uetz 2009, Moura et al. 2020, Straus and Avilés 2022, Straus and Avilés 2023, Peng et al. 2024, but see Whitehouse (1988)). Amongst the few exceptions, Shinkai (Shinkai 2007) provided a list of host records for eight kleptoparasitic species from Japan, Grostal (Grostal 1999) provided notes on the ecology and the range of hosts of five Argyrodes species in Australia, Straus and Avilés (Straus and Avilés 2018) explored the distribution of three morphospecies of Faiditus amongst webs of two social and one subsocial Anelosimus species and Fernandez-Fournier and Avilés (2018) explored the impact of sociality and web size on the composition of kleptoparasite communities. Here, we investigate the distribution of an endemic spider community: multiple argyrodine species amongst three hosts in the small Analamazaotra National Park in Madagascar. The three host species are amongst the most abundant web builders in the Reserve, but contrast starkly in web architecture and behaviour (Fig. 1, see Material and Methods). Our study offers the first insight into argyrodine kleptoparasite biology in Madagascar and is the first to explore in-situ snapshot distribution of a species-rich kleptoparasite community amongst architecturally diverse host webs. Which web to invade? Argyrodine kleptoparasites differentiate amongst architecturally ... 3 Material and methods Study system The hosts all build long lasting webs, but differ in web architecture, size, activity patterns and innate aggression (Fig. 1). Nephilingis livida (Vinson, 1863) builds a sticky orb web and Caripetella madagascariensis (Lenz, 1886) ( Lenz 1886) makes a horizontally domed sheet made of dry silk (Yu et al. 2025), both being nocturnal. Anelosimus spiders are subsocial, cathemeral and build three-dimensional tangles of dry silk. Analamazaotra is home to about 10 Anelosimus species ( Agnarsson et al. 2015) that we sampled Figure 1. Study system. A, B Nephilingis livida host, female (A) and vertical orb web (B); C, D Caripetella madagascariensis host, female in web (C) and horizontal domed sheet web (D); E, F Anelosimus vondrona Agnarsson & Kuntner, 2005 host, female with egg sac in web (E) and social 3D tangle web (F); G-I, kleptoparasites in action. Famakytta analamazaotra from this study, pair in copula in N. livida web (G), Famakytta sp. from Montagne d'Ambre, male approaching freshly hatched N. livida spiderlings (H) and Argyrodes kleptoparasites stealing prey from Nephilingis borbonica (Vinson, 1863) in Réunion (I). Total length of females (front of carapace to spinnerets) N. livida ~ 20 mm, C. madagascariensis ~ 18-25 mm, A. vondrona ~ 5 mm, argyrodines ~ 3-6 mm. Web size (longest axis), N. livida orb ~ 40-80 cm high, C. madagascariensis sheet ~ 30-60 cm wide, A. vondrona basket ~ 10-25 cm width, up to ~ 30 cm tangle above. Photos A, C, D, I by M. Kuntner, B, E-H by I. Agnarsson.  4Agnarsson I et al indiscriminately, along with a few webs of the architecturally similar subsocial pisaurid Dendrolycosa sampled. Survey of kleptoparasites in host webs Analamazaotra National Park is a montane rainforest reserve in eastern Madagascar (25.5 km , between 900-1050 m elevation, centre ca. 18°56'10''S, 48°25'42''E, Fig. 2), where we sampled during 21-28 May 2024. Samples were taken from Nephilingis livida, Caripetella madagascariensis, Anelosimus vondrona, A. nazariani Agnarsson & Kuntner, 2005, A. may Agnarsson, 2005 and Dendrolycosa sp. (Fig. 1, Table 1). The host species are endemic to Madagascar, except N. livida that also occurs on northern Indian Ocean islands (Kuntner and Agnarsson 2011). Caripetella and Anelosimus ( Agnarsson and Kuntner 2005, Agnarsson et al. 2015, Agnarsson et al. 2016) were mainly collected along the 2.5 km long Indri 1 trail loop that spans the western portion of the Reserve (Fig. 2). Most Nephilingis samples were taken at the forest edge from the grounds of the Feon'ny Ala cottages (Fig. 2). Host Fam Arg Del Ade Lrab Lrat Lrin Ltia Neo Total Nephilingis 72 66.1% 2 1.8% 0 0% 1 0.9% 23 21.1% 5 4.6% 5 4.6% 0 0% 1 0.9% 109 Caripetella 0 0% 0 0% 1 0.8% 8 6.3% 78 61.4% 5 3.9% 21 16.5% 13 10.2% 1 0.8% 127 Anelosimus 0 0% 0 0% 0 0% 0 0% 32 50.8% 12 19% 0 0% 4 6.3% 15 23.8% 63 The kleptoparasities were captured with pooters from ~ 20 webs of N. livida and C. madagascariensis and placed in live collection vials and used for preliminary translocation experiments described below, before being preserved in ethanol. The entirety of 87 subsocial colonies (75 Anelosimus and nine Dendrolycosa) were collected in plastic bags and later dissected, with all the web inhabitants preserved in ethanol. Taxonomic identification and molecular studies were subsequently performed at the University of Iceland. Translocation experimental pilot To examine the feasibility of species translocation for future research we translocated kleptoparasites between N. livida and C. madagascariensis webs, previously cleared of native kleptoparasites. In 17 translocation experiments, four to seven foreign kleptoparasitic spiders were placed, feet first, on to threads of the outer edge of webs of 2 Table 1. The distribution and abundance of the ten kleptoparasite species found amongst the studied three host webs (Faana, Famakytta analamazaotra; Ar, Argyrodes cf. argyrodes; Del, Deelemanella helmscahani; Ade, Argyrodella ampingamena; Lrab, Lokitandroka rabesahala; Lrat, Lokitandroka ratsimanga; Lrin, Lokitandroka rinha; Ltia, Lokitandroka tiana; Neo, Neospintharus matjazkuntneri). See Suppl. material 2 for cell-shading that marks the relative abundance of kleptoparasite species for each host web and does not compare their relative abundance amongst host webs. Which web to invade? Argyrodine kleptoparasites differentiate amongst architecturally ... 5 each cleared host web: 54 native Caripetella kleptoparasities on 10 Nephilingis webs and 34 native Nephilingis kleptoparasities on seven Caripetella webs. We then monitored the behaviour and movement of kleptoparasites. Morphological identifications and DNA barcodes All samples were preserved in 95% ethanol in the field. For details on morphological identification, DNA extraction, sequencing and analyses see Pett et al. (2025). Statistical analyses To investigate the association between kleptoparasitic species and host web types, we used Fisher’s Exact Test in RStudio (RStudio Team 2020) using simulation with one million replicates (code Suppl. material 1). Data resources Distribution data are provided in Table 1, species descriptions and DNA barcoding data are published in Pett et al. (Pett et al. 2025. Figure 2. Analamazaotra National Park (25.5 km ) is located in eastern Madagascar, see inset map. The edges of the Park are clearly visible by changes in vegetation, evident by dark green forest cover, surrounded by lighter green vegetation and brownish human developments. The research was done along the approximately 2.5 km Indri trail loop (circuit Indri 1) and on the grounds of the Feon'ny Ala cottages. Scale bar in the lower right corner is 1 km.  2 6Agnarsson I et al Results Survey of kleptoparasites in host webs and species identification A total of 299 argyrodine individuals were sampled (Table 1, Suppl. material 2) belonging to nine species from four known and two new genera, all treated in Pett et al. (2025): Argyrodes cf. argyrodes, Deelemanella helmscahani, Argyrodella ampingamena, Famakytta analamazaotra, Lokitandroka rabesahala, L. ratsimanga, L. rinha, L. tiana and Neospintharus matjazkuntneri. Distribution of kleptoparasitic spiders amongst host webs Species composition and abundance of kleptoparasitic spiders differed amongst host webs (Table 1, Fig. 3, Suppl. material 2). Fisher's Exact Test with simulated p-values based on one million replicates revealed a significant difference in kleptoparasite species association with the three host web types (p < 0.01). Figure 3. Distribution and overlap of kleptoparasitic species found on different host webs. Fam, Famakytta analamazaotra; Arg, Argyrodes cf. argyrodes; Del, Deelemanella helmscahani; Ade, Argyrodella ampingamena; Lrab, Lokitandroka rabesahala; Lrat, Lokitandroka ratsimanga; Lrin, Lokitandroka rinha; Ltia, Lokitandroka tiana; Neo, Neospintharus matjazkuntneri. Bar plots mark the relative abundance of kleptoparasite species for each host web and does not compare their relative abundance amongst host webs.  Which web to invade? Argyrodine kleptoparasites differentiate amongst architecturally ... 7 Translocation experimental pilot Of 54 Caripetella native kleptoparasitic spiders introduced on to Nephilingis webs, 15 (28%) left the web entirely (likely all Lokiandroka), 36 (67%) stayed in the host web vicinity, connected to or within 5 cm of the host’s structural thread. Three were lost. Nephilingis attacked and killied two introduced kleptoparasites. Of 34 Nephilingis native kleptoparasites introduced on to seven Caripetella webs, three (all Famakytta analamazaotra) (9%) left the web entirely, seven (20.5%) stayed in the vicinity of the host web and 24 (70.5%) remained on the host capture area. Caripetella showed no response towards introduced kleptoparasites. Discussion Obligatory kleptoparasitism is a widespread ecological strategy buttressed on the adaptive advantage of stealing resources gathered by another organism, foregoing the cost of independent resource acquisition (Iyengar 2008). In spiders, kleptoparasitic interactions are diverse and involve complex ecological and evolutionary dynamics (Vollrath 1987, Elgar 1993, Whitehouse 2011). Recent research highlights the role of spiders as "superhosts" for obligate kleptoparasites and the versatile strategies spider kleptoparasites utilise to harvest resources (Agnarsson 2025). Host selection by kleptoparasites is driven by traits, such as web size, complexity, persistence and host behaviour, all contributing to the multi-species assembly process (reviewed in Agnarsson (2025)). However, our knowledge of these ecosystems is incredibly sparse: of the about 210 likely kleptoparasitic Argyrodinae species, about half are unknown, most of the remainder are known only by the putative host they were collected from, while 15 or so species are well documented. Our study exemplifies the value of intense surveys in poorly-known areas. The first inventory of the distribution of argyrodine kleptoparasites amongst host webs in Madagascar revealed a hitherto unknown ecosystem with multiple novel players and interactions. The study is also the first to survey the hosts in question. Nephilingis livida is a large, mostly nocturnal spider constructing a huge and sticky vertical orb web (Gregorič et al. 2011). Despite the size and ubiquity of N. livida, the only kleptoparasite that has been recorded in its web is A. argyrodes (see Agnarsson (2025)). Caripetella madagascariensis is an exclusively nocturnal giant nursery web spider that was documented only recently in scientific literature to construct a prey capture web (Yu et al. 2025). Her web, one of the world’s largest horizontal domed sheets, is thus an unstudied ecosystem. The cathemeral Anelosimus and Dendrolycosa form subsocial spider communities that construct three-dimensional tangle webs, whose numerous inhabitants — a mother and her offspring — are barely larger than the kleptoparasites (Agnarsson and Kuntner 2005). While Argyrodes sp. and Neospintharus sp. have been haphazardly recorded from Malagasy Anelosimus, surveys are lacking and no kleptoparasites have been documented from Dendrolycosa. 8Agnarsson I et al We found a high diversity of nine argyrodine kleptoparasite species, all of which are new to science (see Pett et al. 2025). This discovery aligns with prior knowledge of Analamazaotra as rich in spider diversity, for example, as a world hotspot of the cosmopolitan Anelosimus spiders (Agnarsson et al. 2015, Agnarsson et al. 2016) and old world Caerostris spiders ( Gregorič et al. 2011). The study area represents less than 1/200,000 of this vast island, yet we sampled double the diversity of kleptoparasitic argyrodine spiders hitherto known from all of Madagascar, underscoring the remarkable biodiversity of Madagascar and how even short-term studies can reveal new insights into this globally distributed ecosystem. Kleptoparasite species composition and abundance differed amongst host webs. Some kleptoparasites appear to be highly selective, for example, three out of ten species occurred exclusively in webs of a single host, including the most common kleptoparasite in Nephilingis webs, Famakytta analamazaotra (Fig. 3). Even amongst species that utilised all three hosts (e.g. Lokitandroka rabesahala, L. ratsimanga), their relative abundances generally tended towards a particular host species or web type (Fig. 3). From the perspective of host species, Caripetella and Nephilingis shared an argyrodine community of five species, but in different abundances and each hosted additional species not found in the other. These large webs, thus, represent distinct habitats for kleptoparasites. Similarly, while each of these share a community of four argyrodine species with the sub-social Anelosimus/Dendrolycosa webs, the latter stand out in harboring relatively few species, yet hosting the vast majority of Neospintharus matjazkuntneri. Detailed studies have been undertaken on a dozen or so kleptoparasite species, for example, Argyrodes antipodianus (Whitehouse 1986, Whitehouse 1988, Whitehouse 2011), Faiditus ululans (e.g. Cangialosi (1990)) and Neospintharus trigonum (e.g. Larcher and Wise 1985, Cangialosi 1997); or sometimes on co-existing species pairs like Argyrodes elevatus and Faiditus caudatus (Vollrath 1979, Vollrath 1984, Vollrath 1987). The kleptoparsitic communities of certain regions (e.g. Shinkai (2007)) like Japan and a few host species are relatively well studied, including Cyrtophora citricola, Trichonephila clavipes and Metepeira incrassata (Agnarsson 2025). However, understanding host and kleptoparasite interactions and species assembly at the community level is at its infancy. Our study is amongst very few that provide a snapshot of the distribution of a diverse and endemic community of kleptoparasites across contrasting webs of similarly endemic hosts within a geographically restricted area. Our findings align with the most inclusive prior study, where Fernandez-Fournier and Avilés (Fernandez-Fournier and Avilés 2018) found that, amongst two social and two solitary spider species, host species traits, like social structure and web architecture, significantly influence their kleptoparasite community. Our findings contribute to an emerging picture of dynamic kleptoparasite communities influenced by the traits of their potential hosts. Our data do not indicate whether all kleptoparasite species found in a particular host web use these as preferred resource sites (Whitehouse 2011). Regardless, it seems clear that different host species provide distinct habitats. For example, only 4/9 kleptoparsite species were documented in Anelosimus webs, all in low abundance, averaging less th Which web to invade? Argyrodine kleptoparasites differentiate amongst architecturally ... 9 Suppl. material 2: Table S1 Authors: Agnarsson et al. Data type: Table Brief description: Image of Table 1 with colours highlighting species abundances amongst host webs. Download file (73.70 kb) 16 Agnarsson I et al