Wolf spider burrows from a modern saline sandflat in central Argentina: morphology, taphonomy and clues for recognition of fossil examples

Previous descriptions of modern wolf spider burrows

The first work unequivocally related to burrows of wolf spiders was “The castle –building spider” from Illinois (USA) published by Hancock (1899). This paper describes in detail the burrows produced by Geolycosa domifex (Hancock, 1899) (= Lycosa domifex), explaining important aspects such as materials and the methods of construction. Geolycosa domifex burrows are described as vertical shafts, unless obstacles cause some deviation (Fig. 1A). Ratcliffe & Fagerstrom (1980), in their widely cited work on traces found in Holocene floodplains, described spider burrows in general (assigned to Ctenizidae, Antrodiaetidae, Theraphosidae and Lycosidae) as simple or branched tunnels, sometimes with side chambers that are separated from the main tunnel by hinged doors (Fig. 1B). Burrows of Geolycosa xera archboldi (McCrone, 1963) and G. hubbelli (Wallace, 1942) from Florida, USA, are illustrated as vertical shafts showing a gradual transition between the shaft and the terminal chamber (Figs. 1C–1D) (Carrel, 2008). Geolycosa missouriensis (Banks, 1895) burrows from Mississippi, USA, are described as vertical forms, narrower at the surface and broader near the bottom, sometimes with a conspicuously enlarged chamber at the bottom (Fig. 1E) (Suter, Stratton & Miller, 2011). Geolycosa sp. burrows from India, exhibited a contrasting morphology in comparison with previous records of wolf spiders. These burrows were complex with a U-shaped form, two chambers (one located at the entrance and the other at the end of the burrow), and shallow hollows described as drainages or prey traps (Fig. 1F) (Chikhale, Santape & Bodkhe, 2013). Albín, Simó & Aisenberg (2015), reported different burrow morphologies produced by Allocosa brasiliensis (Petrunkevitch, 1910) from Uruguay, linking these variations in the morphology to the development stage and sex of the spider that produce them. These authors described burrows with a simple vertical shaft and a terminal chamber produced by adults, shallow capsules by virgin females, and Y-shaped burrows by male juveniles (Fig. 1G). Hils & Hembree (2015), through experimental neoichnological studies, recorded four burrow morphologies produced by Hogna lenta (Hentz, 1844) (Lycosidae): vertical shafts, vertical shafts with a terminal chamber, sub-vertical shafts, and Y-shaped burrows (Fig. 1H). Geolycosa vultuosa (Koch, 1838) burrows from Albania are characterized as vertical to subvertical, slightly curved or straight shafts with a basal chamber, showing either a gradual transition between the shaft and the basal chamber or a well delineated chamber (Vrenozi & Uchman, 2015). In a taxonomic revision of the halotolerant wolf spider genus Tetralycosa Roewer, 1960 (Framenau & Hudson, 2017), the burrows of three species (T. alteripa McKay, 1976, T. williamsi Framenau & Hudson, 2017, and T. eyrei Hickman, 1944) were described. Tetralycosa burrows are vertical shafts with an offset (a curvature) at mid-depth, which are later modified by backfilling the part above the curvature and creating a new burrow oriented in the opposite direction (Fig. 1I) (Framenau & Hudson, 2017). Allocosa senex (Mello-Leitão, 1945) burrows from Uruguay are also simple vertical shafts with a downward widening (Fig. 1J) (Foelix et al., 2017). Finally, the burrows of Trochosa hispanica (Simon, 1870) from Albania (Fig. 1K) were described as simple, vertical shafts with a terminal chamber (Uchman, Vrenozi & Muceku, 2017).

From the previous account, it is clear that the most common form in wolf-spider burrow are almost vertical cylinders with a rounded end that increase progressively in width downward, vertical shafts with a terminal chamber, and Y shaped burrows. Hasiotis & Bourke (2006) also suggested that horizontal burrows systems with a pustulose ornamentation are produced by spiders, however, the illustrated burrow system (Hasiotis, 2002, p. 114, figure B) is typical of surface burrows produced by Grillotalpidae (e.g., Chamberlain, 1975). Figure 1 also highlights that the burrows produced under experimental conditions (Fig. 1H) contrast markedly with the remaining ones excavated in natural conditions.

Ocurrence of Pavocosa sp. burrows

In early spring (October 2016) abundant burrow entrances of similar size were observed in the sandflat surface. Spider burrows were found in a sparsely vegetated sandflat (0 to 10% of plant coverage), with the only presence of the small halophyte shrub Heterostachys ritteriana Ungern-Sternberg, 1876 (Fig. 4A). The burrows were simple vertical and silk lined forms (Fig. 4B), appearing either open and covered with a thin ring of silk (Fig. 4C) or partially closed with a plug of silk and sediment pellets (Fig. 4D). Surrounding the burrow (in a radius of up to 64 cm) abundant small spherical sediment pellets were observed (with a density of up to 290 pellets/m²) (Fig. 4F), and at this time no casts were made. In late spring (December 2016) burrow density was lower, and they were restricted to a small area on the edge of the saline lake with sparse vegetation at the boundary with the bare sandflat. A total of eight casts were obtained, five were inhabited burrows, while the remaining were abandoned. The inhabited burrows showed up to two sacs of eggs in the lowermost part (Fig. 4E). During the field trip conducted in summer (February 2017), very few burrows were observed, all open and partially filled with some sand, and they seemed to be uninhabited for a long time. At this time only one uninhabited burrow was casted.

Sandflat sediments

The pit dug in the saline sandflat where the burrows occur was 60 cm deep (Fig. 5A). The uppermost bed (# 1) is 13 cm thick and mainly composed of poorly-sorted pale yellowish brown (10 YR 6/2) silty sand containing 0.9% of carbonate (Figs. 5B, 5C). The lower 5 cm of bed 1 exhibits thin diffuse evaporite laminae and a mud lamina. This bed contained the studied Pavocosa sp. burrows. Bed 2 (7 cm thick) is poorly-sorted moderate yellowish brown (10 YR 5/4) silty sand, with massive structure and 0.8% of carbonate. Bed 3 (5 cm thick) is very poorly-sorted, dark yellowish brown (10 YR 4/2), silty sand with massive structure, containing 1.4% of carbonate and small (2 mm) gastropod shells comparable with Heleobia (Stimpson, 1865). The 27 cm thick bed 4 is very poorly-sorted, massive, moderate brown (5 YR 4/4), sandy silt containing 0.6% CO₃. The 6 cm thick lowermost bed (# 5), is mainly composed of fine-grained, pale yellowish brown (10 YR 6/2) sand with abundant carbonate cement that matches with the water table. Field work was conducted in rainy days, however, the water table was well below the bottom of Pavocosa sp. burrows (about 40–45 cm below the bottom of the burrows).

Producer of the burrows: Pavocosa sp.

Although the genus Pavocosa (Roewer, 1960) was never reviewed, and its composition was recently questioned (Toscano-Gadea & Costa, 2016), the inclusion of the material studied as an undescribed species of Pavocosa was possible through the comparison of the males and females of Pavocosa gallopavo (Mello-Leitão, 1941) (Figs. 6A, 6C), the type species of the genus. The male holotype of P. gallopavo (MLP-15065) and females from MACN collection were examined and they share with Pavocosa sp. (Figs. 6B, 6D) the presence of deep furrows on the atrium, parallel to the median septum of the female epigyne and the coloration pattern (Figs. 6A, 6B), characteristics probably diagnostic of the genus (L Piacentini, personal observations). The enlarged posterior eyes in Pavocosa sp. and the shape of the genitalia are clearly distinctive from P. gallopavo. The fangs of specimens captured inside the burrows (n = 3) are separated about 3.9–4.6 mm (Fig. 7H). The environmental distribution of Pavocosa is little known, although it seems to prefer bare patches in sandy grassland soils (L. Piacentini, pers. obs., 2014–2017).

Additional material of the described species from Córdoba (Salinas Grandes, 29°50′39″S, 64°40′16″W), Santiago del Estero and San Luis (Pampa de las Salinas; 32°12′19″S, 64°39′13″W) were recorded from MACN-Ar collection (23503, 23505 to 23513, 24096, and 38710), all from saline environments. The burrows of Pavocosa sp. from Córdoba (A Peretti, C Mattoni and M Izquierdo, pers. comm., 2008) and San Luis (M Ramírez, pers. comm., 2016) are very similar to those described in this work.

Pavocosa sp. burrows

The inhabited burrows (n = 5) (Figs. 7A–7E) are simple, vertical and circular shafts with an inclination of the main axis of 72°–88°(average: 80°), the length ranges from 115 to 130 mm (average: 120 mm). The diameter gradually increases from an upper narrow neck that is 12 to 15 mm wide (average 14 mm) and 5–8 mm long (average 6 mm), to a maximum diameter in the lower half ranging from 18 to 28 mm (average 23 mm). The outline of the entrance and cross-section of the maximum diameter of the burrows are subcircular. In average, the widest part of the burrow is 64% larger than the neck. The burrow cast surface of five burrow casts exhibits sparse ornamentation in the form sets of two linear parallel ridges (Figs. 7F–7G) about 2.8–4.4 mm long (average 3.4 mm, n = 16) and 2.2–4.5 mm wide (average: 3.4 mm, n = 14) aligned oblique to perpendicular (range: 42°–89°, average: 64°, n = 14) to the main axis of the burrow. The Supplemental Information contains interactive PDF files of each of the Pavocosa sp. burrow casts.

Modified Pavocosa sp. burrows

Uninhabited Pavocosa sp. burrows (n = 4) display some kind of modification in their overall form (Figs. 8A, 8B, 8F, 8G) (see Supplemental Information for interactive 3D models of each cast). All are composed of a highly inclined shaft (range: 78°–87°; average: 84.5°), with an upper constriction and an average maximum diameter ranging from 15 to 22 mm (average 19 mm). Three types of modifications were identified. (1) Subcylindrical burrows (108–116 mm long by 15–22 mm wide) with a subhorizontal expansion in the middle part forming an ”umbrella” (Figs. 8A–8B). The shaft boundary exhibit scarce ornamentation in the form of sets of two linear parallel ridges similar to those of the inhabited Pavocosa sp. burrows. The “umbrella” structure shows an oval to lobed shape in the plan view (Figs. 8C–8D), with a minimum diameter of 47–54 mm and a maximum diameter of 59–66 mm. The “umbrella” surface exhibits an ornamentation in form of small (1.4 mm in diameter) rounded knobs (Fig. 8E). The burrow bottom is rounded or partially filled with sediments. (2) Subcylindrical burrow about 116 mm long and 21 mm wide with two smaller burrows (8 mm of diameter) arising at the bottom of the larger burrow (Fig. 8F). (3) A third form is a 143 mm high and 101 mm wide funnel that ends in a 24 mm wide cylindrical shaft with an oblique bottom (Fig. 8G). The surface of the funnel exhibits sets of two parallel ridges (about 21 mm long and 9.2 mm wide) running oblique to the major axis (Fig. 8H).

Clues for identification of wolf-spider burrows in the fossil record

Pavocosa sp. produces open burrows with distinctive features as cylindrical, nearly vertical, silk lined shaft showing a gradual downward widening, a neck in the top and a rounded end, the entrance sometimes plugged with a cap of silk and sediment pellets, and a particularly distinctive surface ornamentation on the burrow margin. Most of these features are shared with other wolf spider burrows documented in the literature (Fig. 1) (Hancock, 1899; Ratcliffe & Fagerstrom, 1980; Carrel, 2008; Suter, Stratton & Miller, 2011; Albín, Simó & Aisenberg, 2015; Hils & Hembree, 2015; Vrenozi & Uchman, 2015; Foelix et al., 2017; Uchman, Vrenozi & Muceku, 2017). In particular, the presence of a neck and downward widening seem to be a common feature in wolf spider burrows found in natural settings. For Pavocosa sp. burrows this widening is about 64%, whereas it is 52% for Trochosa hispanica (Uchman, Vrenozi & Muceku, 2017).

Another highly distinctive feature of Pavocosa sp. burrows is their surface ornamentation in the form of two short parallel ridges oblique to perpendicular with the burrow axis that appear in most burrow casts (Figs. 7F–7G). Although this surface ornamentation was not recorded in some casts, probably due to the presence of the silk lining, all the burrow casts with delicate preservation of the surface texture exhibit these paired ridges. This feature was not identified in previous studies of wolf spider burrows and is potentially related to the burrowing technique used by Pavocosa sp. Spiders use two main mechanisms of excavation: (1) by pushing and compressing sediment using the pedipalps (Hils & Hembree, 2015) and (2) by scraping the soil with help of fangs from chelicerae (Stokes, 1884; Suter, Stratton & Miller, 2011; Hils & Hembree, 2015; Foelix et al., 2016). Although we have not observed Pavocosa sp. during digging, the sets of two linear parallel ridges observed on the surface of the better preserved burrow casts are similar in form and shape with the arrangement of fangs of collected specimens. The distance between fangs (3.9–4.6 mm) overlaps with the distance between ridges within a set (2.2–4.5 mm). Thus we propose that excavation in Pavocosa sp. involves the use of fangs, as in the type 2 excavation mechanism mentioned above.

Silk lined burrows are unique in spiders and essentially impart stability in soft substrates to prevent collapse (Ratcliffe & Fagerstrom, 1980; Foelix et al., 2017; Hils & Hembree, 2015). The presence of organic matter in the form of a silk lining increase the potential of preservation of wolf spider burrows (Uchman, Vrenozi & Muceku, 2017), well above those of all others arthropods that in habit the same environment.

Spider burrows may be modified by reoccupation or predation, as well as by environmental changes. Reoccupation of abandoned lycosid and mygalomorph burrows by lizards, centipedes, moths, wasps, beetles and ants have been documented (e.g., Fellows, Fenner & Bull, 2009). Ants have been also observed invading occupied wolf spider burrows with the purpose of prey piracy (Marshall, 1995). However, it has not been documented if the reoccupation results in any change in the morphology of the burrow. Common spider burrow disturbances caused by predation includes those produced by pompilid wasps that prey the spider and occasionally dig a tunnel perpendicular to the spider burrow (Gwynne, 1979; Costa, Pérez-Miles & Mignone, 2004), and excavation of the upper part of the burrows by armadillos (Suter, Stratton & Miller, 2011).

Most of Pavocosa sp. burrows are susceptible to go through a large amount of disturbances, including those caused by the reoccupation by another organism (Figs. 8A–8B and 8F) and predation of the dweller (Fig. 8G). Two kinds of burrow modifications observed during this study are tentatively related to reoccupation of burrows: those with an expansion in the middle part as a kind of “umbrella” (Figs. 8A–8B) and those with smaller excavations at the bottom of the burrow (Fig. 8F). Even if we cannot discard an inorganic origin (i.e., evaporite leaching) for the “umbrella” structure seen in some casts, it is highly reminiscent of oval to lobed ant nest chambers (Tschinkel, 2003). Although no ants were recorded when making the casts, they were commonly seen in the sandflat surface constructing nests within vertebrate footprints and abandoned burrows, presumably to avoid the hard efflorescent salt crust of the sandflat surface. The producer of the smaller burrows at the bottom of Pavocosa sp. burrow is unknown. Funnel shaped burrows (Fig. 8G) are similar to the probing marks related to predation by small mammals, and similar structures are described in the literature including Sarzetti & Genise (2011) from northern Argentina, Suter, Stratton & Miller (2011): Fig. 2), and Platt (2014), the latter two from Mississippi, USA. Small mammals found in this area with similar behaviours are the armadillos and skunks. The more likely producer is a small armadillo as suggested by the size of the funnel and most importantly by the presence of sets of two large ridges in the cast surface (compare Platt, 2014), interpreted as scratch marks (Fig. 8H).

Preservation of burrows in the margin of saline lakes, including those of wolf spiders, is affected by environmental factors like early cementation by evaporites and swelling of expansive clays during flooding (e.g., Scott, Renaut & Owen, 2010). Cementation by evaporites favors preservation, whereas wetting and drying cycles of swelling clays can destroy the burrows.

Both the original Pavocosa sp. burrows and those modified by reoccupation or predation can be compared with known ichnogenera. The simple vertical forms are grossly comparable with Skolithos (see Alpert, 1974; Schlirf, 2000); some significant differences are the presence of a constriction or neck, the downward widening and the surface texture. These features are potentially significant ichnotaxonomicaly (Schlirf & Uchman, 2005), although no proposed ichnotaxon match them. Slight variations in burrow diameter are allowed in Skolithos (Alpert, 1974; Schlirf, 2000), although the observed differences in Pavocosa sp. burrow diameter are significant and repetitive. There are a few examples of ornamented Skolithos, all of them from continental settings and tentatively assigned to insects or spiders, but they are not comparable to that observed in Pavocosa sp. burrows (Bromley & Asgaard, 1979; Schlirf, Uchman & Kümmel, 2001; Netto, 2007). These ornamented Skolithos exhibit indistinct striations, except for the example described by Netto (2007) that display horizontal striae forming a circular ring. In consequence, there is no known fossil burrow with all the features described for the studied wolf spider burrows.

Modified Pavocosa sp. burrows with an “umbrella” if fossilized can be confused with Daimoniobarax (Smith et al., 2011); in particular, the umbrella is comparable with chambers and the vertical burrow of the spider is comparable with the shaft connecting the chambers in Daimoniobarax. A potential difference is the considerably larger diameter of the burrow connecting the chambers that averages 40% of chamber diameter in the modified Pavocosa sp. burrow and 10% in Daimoniobarax (Smith et al., 2011). The modified Pavocosa sp. burrow with smaller burrows arising from the bottom can be confused with a downward bifurcation as seen in rhizoliths (Klappa, 1980), a roughly similar rhizolith was figured by Melchor, Genise & Miquel (2002), Fig. 3B). Finally, funnel shaped burrows can be compared with several ichnogenera including Monocraterion (Torell, 1870); Conostichnus (Lesquereux, 1876); Rosselia (Dahmer, 1937); Conichnus (Männil, 1966); and Cornulatichnus (Carroll & Trewin, 1995) (see also Platt, 2014). A fundamental difference with these ichnogenera is the lack of large paired surface ridges, as seen in the predated Pavocosa sp. burrow. Further differences are: (1) Monocraterion shows smaller radial burrows arising from the central funnel (Jensen, 1997); (2) Conostichnus exhibits a duodecimal symmetry and transverse and longitudinal ridges and furrows (Pemberton, Frey & Bromley, 1988); (3) Rosselia is a bulbous structure with a concentrically laminated fill (Schlirf, Nara & Uchman, 2002); (4) Conichnus exhibits a rounded apex and common chevron-like fill (Pemberton, Frey & Bromley, 1988); and (5) Cornulatichnus has a well-developed lining (Carroll & Trewin, 1995). Conical sedimentary structures of inorganic origin can also resemble Pavocosa sp. burrows modified by predation. Buck & Goldring (2003) identified two main inorganic processes that produced conical sedimentary structures: collapse and dewatering. The former is distinguished by V or U shaped downwarping of lamination and the latter by deformed lamination and massive zone at the base of the cone (Buck & Goldring, 2003). These features allow distinction from the predated (i.e., funnel-shaped) Pavocosa sp. burrow, that would have a massive fill.

Burrowing spiders belong to Mesothelae and Opisthothelae (Coddington, 2005). Although Mesothelae dates back to the Late Carboniferous, the only known burrowing group (Liphistiidae) has no fossil record (Dunlop, Penney & Jekel, 2017). Within Opisthothelae, burrowing spiders are found in the Middle Triassic to Recent Mygalomorphae, which includes the tarantulas and trapdoor spiders, and in the Cretaceous to Recent Lycosoidea (included in Araneomorphae) that comprises the wolf spiders (Dunlop, 2010; Dunlop, Penney & Jekel, 2017). The oldest putative example of Lycosoidea comes from Turonian beds of Botswana (Selden, Anderson & Anderson, 2009), which is close to the age of the superfamily suggested by phylogenetic studies (70 Ma, after Garrison et al., 2016), although most fossil records are from the Paleogene to Recent (Dunlop, Penney & Jekel, 2017). In addition, phylogenetic studies on web type suggest that the spider common ancestor likely foraged from a subterranean burrow, mostly sealed by a trapdoor (Garrison et al., 2016). In consequence, the record of spider burrows can be traced back at least to the Middle Triassic (and probably to the Late Carboniferous) and lycosid burrows to the Late Cretaceous or Cenozoic.

The use of fossils to calibrate molecular phylogenies is an uprising topic in spider biology (Planas, Fernández-Montraveta & Ribera, 2013; Wood et al., 2013; Moradmand, Schönhofer & Jäger, 2014). The absence of reliable fossil record, such as in Lycosidae (Penney, 2001), is an important impediment and the potential identification of wolf spider burrows on the fossil record, with the clues provided herein, can be a useful alternative source of data.

Environmental distribution of Pavocosa sp. burrows

The sediments of the sandflat containing the Pavocosa sp. burrows reflect the interaction between the nearby eolian and lacustrine settings. The two upper beds are essentially sandy deposits with a mixture of dominant fine sand and silt (samples S1 and S2; Fig. 5). The dominance of the coarse fraction (fine sand), poor sorting and the frequency distribution are comparable with those of modern interdune deposits (e.g., Ahlbrandt, 1979). Poorly defined laminae with evaporites in bed 1 are interpreted as the result of capillary rise and precipitation from brines. The sandy nature of the material where Pavocosa sp. excavated the burrows and their location 40 cm above the water table suggests preference for well-drained substrates. In contrast, the lowermost silty beds (samples S3 and S4; Fig. 5) are interpreted as dominantly lacustrine deposits, on the basis of the fine grain size and the presence of gastropods shells. Heleobia is a very common extant gastropod in South America recorded in estuarine and continental settings, including saline lakes (see review in Cazzaniga, 2011). In consequence, the logged section reflects the migration of the parabolic dune towards the northeast over the Gran Salitral lacustrine sediments (for a more detailed interpretation of dune deposits see Melchor et al., 2012). The presence of abundant Pavocosa sp. burrows in the well-drained sandflat deposits of the Gran Salitral and similar occurrences reported in the literature (e.g., Hudson & Adams, 1996) suggest that wolf spider colonization of saline lakes occur preferentially during dry periods of the lake.

Wolf spiders (Lycosidae) are one of the most successful spider families distributed in most of the habitats around the world (World Spider Catalog, 2017). Lycosids display a wide range of prey-capture strategies from web builders to burrow-dwellers or vagrant species. The use of burrows in wolf spiders can be in some cases obligatory, temporary in male juveniles, and as brood care in females (Logunov, 2011), or merely facultative in absence of objects as a rock that serves as a retreat. In general, burrows in wolf spiders are related to open areas of xerothermic habitats with sparse or no vegetation (e.g., sandy seashores, dune heaths, limestone areas and desert nanophanerophyte steppe) (Logunov, 2011). Some wolf spider species have specific habit preferences, as is the case of halotolerant species that inhabit the surface of salt lakes, most of them included in Tetralycosa and other species as Lycosa salifodina (McKay, 1976) from Australia (Hudson & Adams, 1996; Framenau & Leung, 2013), and two other Argentinian species including Pavocosa sp. In particular, Pavocosa sp. has been documented in saline lakes of Cordoba, Santiago del Estero, San Luis and La Pampa. In consequence, it is likely that the described burrows are typical of saline environments.