A spider that decorates its web perpendicular to the web plane

Abstract

This paper examines and describes web decorating behavior by juveniles of the orb-weaving spider, Trichonephila antipodiana. Decorations consist of a three-dimensional ‘tube’ of silk line scaffolding within which detritus, prey items and moult exuviae are laid perpendicular to the web plane, extending from the central sticky web forward and backward to the dorsal and ventral barrier webs. Complementarity with barrier web construction, combined with the vulnerability of this ontogenetic stage, suggest that the primary function of the decorations is as an anti-predator device. Various lines of evidence support the view that the fundamental structural purpose of the decorations is perpendicular extension: (1) median web residency time was significantly greater for juveniles with decorations ≥ 1 cm in size; (2) detritus was the primary contributor to decoration length; and (3) examination of the in-laying technique for prey inclusions found them to be ‘unnecessarily’ dismembered and stretched out along the decorative plane. Although there is probably some degree of interdependence between predator avoidance and prey catching success, extension as a property fits most consistently within a defensive interpretation of function in so far as it confers either distance from and/or interference with predators.

Keywords:

• barrier web
• defense
• detritus
• exuviae
• ontogeny

Introduction

The use of web decorations by spiders has been reported for numerous species and at least two families (Araneidae, Uloboridae) (Simon 1895). The tally is undoubtedly incomplete, but it is sufficient to indicate that decorations provide an intriguing portal into the ‘extra-web’ engineering and behavioral toolkit of many web-building spiders. In material terms, decorations simultaneously show diversity and conservatism of form depending on phylogenetic, ontogenetic, and ecological context. In behavioral terms, web decorations have generated a rich literature of experiment and observation exploring the functional significance of these decorations (e.g. Eberhard 1990; Herberstein et al. 2000; Blackedge et al. 2011 and references therein). It would be jumping the gun to say that either the engineering or ethology has been comprehensively decoded, but insights continue to accumulate.

Hingston (1927) was the first to attempt a morphological classification of these structures, describing eight primary character classes with a total of 22 specific types. In structural terms, this is still the most comprehensive attempt at a taxonomy of forms. Later studies and reviews have tended to move away from proximate characters to focus more on ultimate causation, aiming, in particular, to try and flesh out the functional origins and purpose of these structures within an evolutionarily consistent perspective. This is particularly reflective of how the structures of some genera – primarily Argiope Audouin, 1826 and Cyclosa Menge, 1866 – are both sufficiently well-known and available enough as test models to offer many opportunities for probing ethological and evolutionary questions (e.g. Eberhard 1990; Herberstein et al. 2000; Blackedge et al. 2011 and references therein).

However, one of the properties of diversity is to provide alternative perspectives on what is well known. In other words, if a new iterative term is added to the equation, how do its differences and similarities connect with perspectives that have already been acquired? This question is given focus here, by the examination of a previously undescribed and unknown form of web decoration. These are the decorations of juveniles of the nephilid spider, Trichonephila antipodiana (Walckenaer, 1842). Although materially identical to other species that utilize combinations of silk and detritus for the decorations they integrate into the web (e.g. Tan and Li 2009; Tseng and Tso 2009); structurally the form of decorations made by T. antipodiana has one property that diverges greatly from the pre-existing morphological catalog of forms: these decorations are laid perpendicular to the web plane. All other known and described forms of organized decoration work on an in-laying principle in which either silk and/or other materials are embedded across the plane of the web. An exception to this is the tuft decorations of barrier webs in Thelacantha brevispina (Doleschall, 1857) (see Tseng et al. 2011) which are outside the sticky web but which, in that the barriers are made of disparate lines of silk may still be considered to be laid along the vertical plane.

Although some adult specimens of Trichonephila Dahl, 1911 are known to hang prey larders outside their webs in addition to silk stabilimenta inside their webs, these structures are architecturally unsophisticated and unorganized – and indeed, are mostly referred to in prey management terms, rather than decorative terms (e.g. Harvey et al. 2007). On the other hand, the decorations described here that are produced by T. antipodiana juveniles have distinct structural and morphological elements that are archetypally representative of the concept of spider decoration as organized material manipulation (Figure 1).

Figure 1. Three examples A–C, of the perpendicular web decoration template of Trichonephila antipodiana juveniles.

Although T. antipodiana is one of the most prevalent representatives of the genus in Southeast Asia, its biology remains largely unknown. General biological observations were recently summarized by Harvey et al. (2007). With the exception of Zhang et al.’s (2012) interesting work on the properties of adult silk, no real quantitative research has been carried out on the species. This paper describes a field study carried out to describe the deployment and explore the potential function(s) of these decorations. It was hypothesized that this unusual perpendicular form serves as an anti-predator device, where, depending on the extent of decoration, it may provide both a physical barrier and/or decoy to ventral and dorsal attack as well as potentially a medium for ‘imperfect’ crypsis.

Methods

Research was carried out at Huay Yai district, Banglamung, Chonburi province in central eastern Thailand. Taxonomic nomenclature follows the World Spider Catalog (2019). Species identity of T. antipodiana was confirmed by field observations of adults in the same locality; rearing a juvenile female; as well as by subsequent examination of the genitalia of males and adult females collected from neighboring Rayong province. The adult female is readily identified by the presence of a conical protuberance on the anterior of the sternum (Harvey et al. 2007).

Fieldwork was carried out over a 4-month period in 2014 within a known habitat matrix of ca. 1–2 km² consisting of woodland, clearings, orchards, and managed vegetation. The habitat was patrolled regularly in search of juvenile webs. Whenever a new web was found its location was mapped using site descriptors and landmarks. Webs were then visited every 2–3 days and monitored until their relocation or disappearance to provide a measure of population web residency duration. A total of 90 spiders were monitored. Basic web descriptors were collected for each web: height above ground (cm), length (cm), width (cm), and compass orientation.

Web decorations were monitored over time. Length (cm), determined using digital calipers, was used as a measure of decorative size with separate and combined measures for forward and backward extension. Forward position was noted as the side of the web on which the spider was resting. Decorations were photographed and analyzed for descriptive information about structure and to determine compositional contents. Three classes of web inclusion were noted: detritus, prey and moult exuviae. The latter two inclusion classes were defined as items with discernible morphology that corresponded to prey and spider exuviae respectively; detritus was defined as any vegetative material and any amorphous material which contributed to decorative size but had no discernible morphology.

For comparisons of web residency, the duration (in days) of web occupation was compared between webs with decorations ≥ 1 cm and <1 cm in length. Although in strict quantitative terms, this represents an arbitrary delineation point, it is based on biological reasoning: (1) it offers a means to delineate decorations that differ by an order of magnitude in length (e.g. 0.5 cm vs. 5 cm); and (2) it is informed by a recognition that decorations must exceed juvenile body size (ca. 1–3 cm length; ca. 0.5–1 cm width) if they are to offer any form of distance from or interference to predators. Decorations ≥ 1cm in length are denoted as ‘functional’ decorations.

Data for residency time and decoration size were not normally distributed and could not be transformed. Statistical comparisons were therefore made with non-parametric tests using a Kruskal–Wallis test and a Mann–Whitney test, respectively. For comparisons of decoration composition, exuviae were excluded from statistical analysis. Exuviae comprise a minor and occasional component of decorations so were not included because: (a) they do not contribute to the decorations in a structurally significant manner; (b) they would detract from the precision of non-parametric comparisons by yielding significant differences that confuse the main distinction between prey and detritus, particularly since non-parametrics do not allow for post-hoc compartmentalization of the contributions of each component to statistical differences.

Results

Figure 2 shows an example of a juvenile web decoration to show web context of the decorations, in particular, the spatial relations with other web elements. Figure 3 shows close-ups of a decoration web to delineate aspects of decoration structure, in particular the silk scaffolding lines that enable perpendicular extension: (a) a ‘tube’ of horizontal silk lines that hold the decorative material; and (b) sets of diagonal silk lines both above and below the decoration ‘tube’ that hold it in place. Table 1 summarizes the parameters of webs. Web residency of juveniles was compared between webs with and without ‘functional’ decorations (≥ 1 cm in length). A Kruskal–Wallis test using decoration size as a factorial delineator found a significant difference in the median number of days webs were occupied, with ‘functional’ webs occupied for longer time periods: n = 43, 47; df = 1; H = 4.47; p = 0.034 (adjusted for ties) (Figure 4A).

Figure 2. Juvenile web decoration of Trichonephila antipodiana within web context, highlighting interrelationship between other web elements, in particular barrier webs and sticky web.

Figure 3. Progressive magnifications A–C, of web decoration of Trichonephila antipodiana juvenile to delineate scaffold structures that form the tube of the decoration itself and provide support for it.

Figure 4. Deconstructed aspects of the decorations of Trichonephila antipodiana: A, the relationship between web residency and functional decorations; B, the relative composition of decorations by size of detrital and prey items.

Table 1. Dimensions and orientations of webs of T. antipodiana juveniles.

Web parameter	Measures
	Mean	SE
Web height (cm)	136.66	4.63
Web width (cm)	9.02	0.37
Web length (cm)	9.29	0.45
	N	E	S	W
Compass orientation (% distribution)	32.22	5.55	30	32.22

Table 2 compares the contributions of web, prey, and exuviae to the composition of all web decorations. ‘Functional’ decorations were analyzed for the relative contributions of prey and detritus to decoration perpendicular extension (length). A Mann–Whitney test found significant differences between the median length of detrital and prey inclusions in ‘functional’ decorations where n = 47, 47; W = 2578.3; p = 0.0088 (adjusted for ties) (Figure 4B). Macro-photographic analysis of decorations showed that decorative inclusions were manipulated using an in-laying technique that maximized perpendicular extension. Figure 5 shows examples of prey inclusions dismembered and stretched out across the plane of the decorations to which they contribute.

Figure 5. Examples A–E of prey inclusions from the web decorations of Trichonephila antipodiana demonstrating their post-mortem manipulation [fragments A-E unidentified; fragment F is the remains of a weaver ant, Oecophylla smaragdina Fabricius, 1775].

Table 2. Composition and occurrence of web decorations in the webs of T. antipodiana juveniles.

	Decorations	No decorations
Relative occurrence (total)	76 (84.4%)	14 (15.6%)
	≥ 1cm	<1cm
Relative occurrence (functional)	47 (52%)	43 (48%)
	Dorsal	Ventral	Combined
Relative distribution of decorations in relation to web aspect	40 (53%)	10 (13%)	26 (34%)
Mean decoration length (cm)	1.84 ± 0.26	1.48 ± 0.76	3.30 ± 0.42
	Detritus	Prey	Exuviae
Mean decorative composition of all decorations (cm)	1.37 ± 0.19	0.84 ± 0.14	0.08 ± 0.02
Mean percentage composition of all decorations (%)	48	44	8
Mean decorative composition of functional decorations (cm)	2.03 ± 0.26	1.23 ± 0.21
Mean percentage composition of functional decorations (%)	61	39

A few direct observations of predation, preparatory moves toward predation, and attempted predation were also made opportunistically during the fieldwork. All involved aranaeophagy, in particular, the targeting of juvenile T. antipodiana by adult spiders from two other Families: Salticidae and Oxyopidae. Two salticids were observed invading webs and catching juveniles, one salticid was observed in a reconnoitering position studying a web and its occupant that subsequently disappeared. Salticids were not captured so could not be identified. One additional observation was made of an oxypid, identified visually but not captured, as an adult male Oxyopes sunandae (Tikader, 1970), chasing a juvenile across the bridge line of its web.

Discussion

The web decorations constructed by T. antipodiana juveniles are unusual in structure. They deviate from known spider decorative forms by a single fundamental structural characteristic: perpendicularity. Close-up examination of web construction show that these decorations are incorporated into web architecture through two sets of novel web elements: a ‘tube’ of silk lines that carries the decorative material and a network of silk struts that hold and brace the decoration in place both from above and below its plane of extension.

Architecture is complemented by the behavioral ecology of decoration construction. Web residency was found to be greater in spiders with ‘functional’ (≥ 1 cm length) decorations. The composition of decorations was dominated by detritus. This excludes the possibility of the decorations being primarily aimed at attracting prey via a ‘carrion effect’. It does not exclude prey attraction as a component of the decoration, but it does rule it out as a fundamental functional explanation. Extension as a property – and an aim of the decoration – is supported by the use of detrital material, as it is these, largely, foreign inclusions, that enable the decorations to achieve mass and dimension. Examination of the in-laying technique of decorative materials further supports the view that extension is the intended function of the decorations. This is particularly highlighted by the use of prey items. Although the lining up leaves or leaf fragments in line with decorations might be skeptically considered ‘accidental’, or at least insufficiently evidential, the arrangement of prey inclusions is less easily explained away. Spiders feed on prey by sucking out their contents either immediately after capture or after storage as wrapped prey bundles. Prey wrapping is a compression technique. Prey feeding is done at a single point. For prey to subsequently be dismembered, torn apart, and spread across the decoration can only be the product of post-mortem manipulation.

Taken together, both the architecture and behavioral ecology of these decorations support the view that the decorations are functionally designed to extend along the perpendicular plane of the sticky web. Such extension fits with a predator-avoidance interpretation of the decorations. Extension provides both physical interference for and distance from dorsal and ventral attack. Likewise, the decorations may also provide some degree of ‘imperfect’ crypsis for spiders in their webs, with their juvenile ‘prey image’ confused or diluted by apparent contiguity with the decorations (see Hershaft 1968 for a discussion of the strategic rationale behind the deployment of ‘imperfect decoys’).

A protective interpretation is also consistent with the ontogenetic (vulnerable life stage) and web (web already utilizes barrier webs – i.e. evidence of pre-existing utilization of interference defenses) contexts of T. antipodiana juveniles. However, it is noted that, additional opportunistic field observations emphasize that predator avoidance cannot work at all spatial scales simultaneously: a strategic weakness of these devices is that they are ineffective against arthropods working on similar spatial and visual scales – specifically aranaeophagic hunting spiders.

It is not realistic – even with laboratory manipulations, let alone field studies – to construct full-proof functional explanations of web decorations. The sheer diversity of both predictable and unpredictable environmental contexts means that there is always leeway for competing and interactive hypotheses of function to have traction. For example, although the link between residency time and decoration length seems obvious, interactive factors are also readily inferred: the link between web residency and habitat food supply has long been known (Hodge 1987). Likewise, in many situations there must be some degree of biological interdependence between prey attraction and predator avoidance – both are often considered as competing hypotheses of function, but both contribute obviously to the fitness and the success of each other (e.g. Tan et al. 2010). In T. antipodiana juveniles, for example, the decorations are dependent at least partially on prey capture success – energy availability for construction and prey capture viability of web sites, for example, being two pre-requirements for decoration construction and maintenance (e.g. Li 2005). Interestingly, Blamires et al. (2010) recently found that prey interception rate in the webs of Trichonephila clavata (L. Koch, 1878) was greatest in webs without constructions, but retention rate was greatest in webs with barrier webs and carcass decorations.

The decorations may not be optimized as carrion traps, but with the irregular presence of prey being processed and with the detritus potentially offering perch sites for flying invertebrates, they may well attract prey. Nonetheless, if extensibility is considered as the main attribute of these decorations, it is not an attribute that in and of itself contributes in any specific way to prey attraction, whereas it is an attribute that specifically confers protective properties via interference and distancing. Thus, the primary function of these decorations appears to be as an anti-predator device, but environmental context and variability, suggest that this function fits into a complex and holistic array of factors that determine fitness and survival.

Disclosure statement

No potential conflict of interest was reported by the author.