Refuge preferences and web trap lines: the potential for competitive displacement of endemic katipō Latrodectus katipo by the invasive false katipō Steatoda capensis (Araneae: Theridiidae).

doi:10.21203/rs.3.rs-4379057/v1

Competitive displacement is a form of interspecific competition. Here, we explore the potential for competitive displacement in refuges of the endangered katipō spider (Latrodectus katipo) by the invasive false katipō (Steatoda capensis) (Araneae, Theridiidae). We conducted experimental assays using artificial refuges to test individual preferences for refuge shape, surface, and height. We also tested how refuge type affects the number of web trap lines used for prey capture. Both species preferred triangular refuges over rectangular refuges, with no evidence of a preference for other refuge types. For reasons which remain unclear, individuals varied in their preferences for surface and height. Katipō spiders showed site fidelity in all three assays, while false katipō spiders only showed site fidelity in the shape and surface assays. However, there was also evidence of individual variation in site fidelity in the shape assay. Spiders constructed more web trap lines beneath the triangular refuges, potentially due to their preference for triangular refuges. We suggest that overlapping refuge preferences, but not web structure, may contribute to competition and the potential for competitive displacement of the katipō by false katipō. Although not exclusively preferred by katipō, introducing artificial triangular refuges in katipō habitats should be tested as a means of supporting the persistence and expansion of katipō populations.

Competition

artificial refuge

site fidelity

conservation

arthropod

web design.

Competitive displacement represents an intense form of interspecific competition, occurring when species compete for the same resources within a habitat, ultimately leading to the displacement of one species by another (deBach and Sundby 1963). This phenomenon has been documented across birds (Guillaumet and Russell 2022), reptiles (Kraus 2015), and mammals (Ferretti and Mori 2020). Competitive displacement has also been extensively documented in native and invasive arthropod populations, suggesting it is likely a widespread occurrence (Reitz and Trumble, 2002). Reviews and meta-analyses have revealed that several mechanisms can drive competitive displacement, including direct interference for limited resources (Bhuyain and Lim 2019), and pre-empting access to a resource (Gao and Reitz 2017).

A refuge is a crucial habitat resource, shielding organisms from both biotic and abiotic elements, and providing protection against predators and adverse weather conditions. In certain instances, species may exhibit a common preference for a particular refuge and prevent access by other species (Boon et al. 2023). Competitive displacement over refuges has been documented in diverse taxa, including fish (Shulman 1985), crustaceans (Savvides et al. 2015), and spiders. Notably, the invasive European hammock spider (Linyphia triangularis Clerck, 1758) (Araneae: Linyphiidae) competitively displaces the native bowl-and-doily spider (Frontinella pyramitela Walckenaer, 1841) from its web (Bednarski et al. 2010). The displacement of spiders from their web refuge is of significant concern, as it can disrupt their ability to effectively source food, defend themselves, and reproduce (Blackledge et al., 2003; Blackledge et al., 2007; Zevenbergen et al., 2008).

The katipō, (Lactrodectus katipo Powell, 1871) (Araneae: Theridiidae), belongs to the widow family and is endemic to Aotearoa, New Zealand, where they are restricted to sand dunes. Katipō are in serious decline, which is attributed to habitat loss and possibly also competition from invasive species (Hann 1990). Sand dunes are dynamic environments characterised by sparse vegetation, and extreme temperature variability (Maun 2009). Katipō exhibit preferences for refuges, choosing to construct their webs in dune plants such as scrambling pohuehue (Muehlenbeckia complexa) and pingao (Ficinia spiralis), as well as using driftwood and flotsam (Griffiths 2001). The decline of sand dunes across New Zealand by more than 55% since the 1950s (Ryan et al. 2023) has resulted in increasingly limited refuge resources for these spiders, making them more susceptible to competitive displacement by invasive species. We investigated the potential for competitive displacement of katipō by the invasive false katipō, (Steatoda capensis Hann, 1990) (Araneae: Theridiidae) in east Auckland sand dune ecosystems. The false katipō, believed to be introduced to New Zealand from South Africa, shares many traits with the katipō, such as a dark-coloured body and legs, and three-dimensional webs with trap lines leading to the ground to assist with prey capture (Sahni et al. 2012). However, the false katipō inhabits a broader range of habitats than the katipō (Hann 1990). False katipō are synanthropic and often inhabit areas in and around buildings, but they also inhabit sand dune ecosystems where they are found in the same types of refuges used by katipō, although our assessment of false katipō refuge preferences is currently limited to one South Island field study (Hann 1990).

Comparing refuge preferences of the katipō and the false katipō could provide valuable insights for implementing more effective conservation actions for the threatened katipō. If the two species share refuge preferences, the potential for competitive displacement of katipō by false katipō increases, and this information may guide the development of additional refugia in the limited sand dune habitat. Refuges exclusively preferred by katipō would offer options to limit the potential for competition from the false katipō. To explore this, we created artificial refuges (Cowan et al. 2021) and conducted three independent assays—shape, surface, and height—each with two refuge choices. We assessed refuge preferences for each species in each assay and evaluated individual refuge fidelity (the behaviour of remaining in or returning to previously used refuges; Merkle et al. 2022). Additionally, we compared the number of web trap lines built under each refuge type between the two species.

Given the specific habitat requirements of wild katipō, we predicted that katipō would have a preference in each refuge assay. In contrast, we predicted that false katipō, with broader habitat range, would show few preferences. We predicted that the katipō would have lower site fidelity than the false katipō, with the katipō relocating if the initial refuge choice is not ideal. We also predicted that both species would build similar numbers of trap lines due to the similarity in their web designs, as both species spin irregular tangled webs. However, we anticipated that the choice of refuge would impact the number of trap lines, with some refuge types being more suitable for building trap lines.

Collection and husbandry

Thirty subadult or adult female katipō spiders and thirty subadult or adult female false katipō spiders (6th instar or later) were collected from sand dunes in Te Aria Beach, Auckland, New Zealand (36°09'35"S 174°38'50"E), during the Austral spring and early summer months (August to December) of 2023. We searched dune plants such as scrambling pohuehue and kōwhangatara (Spinifex sericeus), naturally occurring logs, and long-term artificial lizard refuges (500mm-by-500mm onduline sheets) covering areas of bare sand among dune plants. We recorded the age, sex and species of each spider and placed them in temporary, individual 50ml plastic vials in an insulated container for transport to the captive laboratory facility at Massey University, Albany Campus.

In the laboratory, spiders were maintained in controlled environment rooms, with an average temperature of 19 ± 0.5 ºC and a relative humidity of 60 ± 11%. The lighting followed a 12 : 12-hour light : dark cycle, including a 1-hour ramp-up and ramp-down period. Individual spiders were housed in 500 ml cylindrical plastic containers (length × diameter × height; 100 × 100 × 150 mm). The container walls were textured by sanding as the spiders cannot climb smooth surfaces well. Ventilation was provided through small holes covered with fine mesh glued in place. A hole was made in the side of the container for feeding access and sealed with a tapered rubber stopper measuring 25 mm across the wide end, 20 mm across the skinny end, and 45 mm long. Before each assay, including between assays, spiders were offered a large (> 20 mm) mealworm (Tenebrio molitor) (Coleoptera: Tenebrionidae) in their housing container, and their webs were lightly sprayed with tap water daily. Offering one mealworm before each assay was sufficient to satiate the spiders.

Test procedure

Spiders were individually tested in clear acrylic arenas (150 × 150 × 150 mm) with two types of refugia inside. The refugia were constructed from cardboard shapes that were wedged into the arena, and contrasted variables that may be relevant to spiders when choosing a refuge in the wild (i.e., shape, surface, and height). A single-wall B-flute corrugated cardboard piece (30 × 70 × 3 mm) was positioned at the back of the arena between the two refugia, serving as a ladder for the spider to the refugia. The position of the refuge type to the left and right side of the arena was not randomized, however there is no evidence that this impacted our results.

We conducted three assays: a shape assay (n = 10 spiders per species) from August to October, followed by a surface assay (n = 10) from October to November, and a height assay (n = 10) in December. The order in which a spider was assigned to an assay depended on the field collection date. In cases where different assays were occurring simultaneously, spiders were randomly assigned to an assay. Each spider was tested twice and had a rest day between assays. Immediately prior to testing, spiders were weighed (± 0.01 mg) using an Ohaus Explorer semi-micro balance (EX125D) and released into the arena around dusk, at the start of their nocturnal activity period.

Refugia choices were recorded daily for seven consecutive days. Choice was recorded during the day, as the spiders do not switch refuges during the day. A choice was defined as the spider being within the borders of the cardboard refugia. If a spider was not within the refugia, it was recorded as ‘no choice’. After the final day, spiders were removed from the arena and reweighed. Cardboard refugia were used only once, and the acrylic surface of the arena was cleaned with soap and warm water between assays.

Shape assay

We tested whether katipō or false katipō preferred triangular or rectangular refugia. Refugia were made from a 435 × 150 × 3 mm strip of single-wall B-flute corrugated cardboard and configured into two distinct shapes: a triangle and a rectangle (Fig. 1A). The rectangular refugia measured 75 mm across, two walls measuring 85 mm each and two 90-degree angles. The triangular refugia had two sides measuring 95 mm each and one angle at 22.5 degrees. The rectangle shape was to the left of the arena, with the ladder at the centre rear of the arena.

Surface assay

We tested whether katipō or false katipō preferred a smooth or rough surface on the refuge cavity. Two refugia of identical size and shape were created from 245 × 150 × 3 mm smooth, single-wall B-flute corrugated cardboard, and rough, single-face B-flute corrugated cardboard with 2 mm high corrugations at a rate of 13 corrugations per 100 mm of length (Fig. 1B). For the rough refugia, the corrugation faced into the refugia, with the smooth side against the clear acrylic arena. The refuge shapes for the surface assay were the same as the rectangular refuge used in the shape assay. The smooth refuge was to the left of the arena, with the ladder at the centre rear of the arena.

Height assay

We tested whether katipō or false katipō preferred short or tall refugia. Height was measured from the top of the refuge to the base of the arena. The refugia were constructed from a 390 × 150 × 3 mm strip of single-wall B-flute corrugated cardboard configured into two heights. The short refuge was 95 mm high, while the tall refuge was 180 mm high (Fig. 1C). The internal cross-sections of the refugia were the same as in the rectangular refuge used in the shape assay; however, one wall on the same side of each refuge was 35 mm instead of 85 mm. The other wall was 85 mm. The short refuge was to the left of the arena, with the ladder at the centre rear of the arena.

Web trap lines

We counted the number of web trap lines under each refuge type at the end of each trial by turning off the lights in the laboratory. Using a focussed light source, we counted the total number of trap lines extending from the superior sheet of the web to the base of the arena. Trap lines were assigned to a refuge if they connected to the superior sheet of the web directly under the refuge cavity. Trap lines were identified by the presence of small liquid droplets along the silk thread.

Statistics

We used generalised linear mixed models (GLMMs) to analyse our results. We simplified each model using a backward stepwise approach based on Akaike's information criterion (AIC). AIC calculated the difference between models and their corresponding Akaike weights. Likelihood ratio tests were used to determine if katipō and/or false katipō differed in refuge choice, refuge fidelity, and number of trap lines. We compared the models with individual random effects against models without random effects while maintaining a consistent fixed effect structure. All statistical analyses were executed using the R statistical software package version 3.4.2 (R Core Team 2015). Models that incorporated random effects were fitted using the lme4 package (Bates et al. 2015). The lmtest package was used to calculate likelihood ratio test (Zeileis and Hothorn 2002).

To investigate whether katipō or false katipō preferred a refuge type we used GLMMs with a binomial distribution. Additionally, we assessed if choice remained consistent across replicates. A daily and a final day model were conducted for each refuge assay using the same model structure. The daily model analysed the spider’s daily refuge choice. The final day model analysed only the spider’s final (day seven) refuge choice. For each assay, the models were analysed twice: once using the daily model and once using the final choice model. A space was left on the data sheet on days when a spider made no choice. No instances of ‘no choice’ occurred in the final day of data collection. The models for the shape and height assays included species, replicate, and weight change (the difference in weight from the beginning to the end of the assay) as fixed effects, with individual identity as a random effect. The models for the surface assay included species, replicate, weight change and the interaction between species and replicate as fixed effects, with individual identity as a random effect. We also tested whether there was a preference for a refuge type in each assay using fixed intercept models.

To test katipō and false katipō site fidelity and if it was consistent between replicates, we used GLMMs with a Poisson distribution and the same backward stepwise approach. We also assessed site fidelity in each assay using number of switches as a metric. A switch was defined as a spider being in one refuge one day and a different refuge the next. The number of switches was summed for each spider for each replicate. The models for the shape and surface assays included species, replicate, and weight change as fixed effects, with individual identity as a random effect. The model for the height assay included species, replicate, weight change and the interaction between species and replicate as fixed effects, with individual identity as a random effect.

We used GLMMs with a Poisson distribution and the same backward stepwise approach to test whether katipō and false katipō built the same number of trap lines and if the number of trap lines were affected by refuge type or replicate. The number of traplines were compared using the final day refuge choice. The models for the shape assay and height assay included species, replicate, choice, and weight change as fixed effects, with individual identity as a random effect. The surface assay model included species, replicate, choice, weight change and the interaction between replicate and choice as fixed effects, with individual identity as a random effect.

To interpret our results, we followed the recommendations from Muff et al. (2021). P-values were categorized as follows: 1 to 0.1 indicated no evidence, 0.1 to 0.05 indicated weak evidence, 0.05 to 0.01 indicated moderate evidence, and 0.01 to 0.001 indicated strong evidence of an effect.

Shape Assay

We found no evidence that species or replicate influenced refuge choice (Fig. 2A, Table 1A). We found only weak evidence that weight change influenced refuge choice (Daily: P = 0.09; Final: P = 0.09, Table 1A). We found moderate evidence of a preference for the triangular refuge over the rectangular refuge in both species in the daily model and weak evidence in the final day model (Daily: P = 0.02; Final: P = 0.07, Fig. 2A, Table 1A). We found no evidence of individual variation in refuge shape choice. We found no evidence that species, replicate, or weight change affects site fidelity (Table 1A). We did however find moderate evidence of individual variation in site fidelity (P = 0.02, Table 1A). The highest number of switches recorded for an individual in an assay was four, and the lowest was zero. We found no evidence that species, replicate, or weight change affects the number of trap lines (Fig. 3A, Table 1A). We found strong evidence that the final refuge choice affected the number of trap lines, with more trap lines built under the triangular refuge than the rectangular refuge (P < 0.001, Fig. 3A, Table 1A). We found strong evidence of individual variation in the number of trap lines built (P < 0.001, Table 1A). The highest number of total trap lines produced by an individual of either species across both refugia in an assay from both replicates was twenty-nine, and the lowest was zero.

Surface Assay:

We found no evidence that species, replicate, or weight change influenced refuge choice (Fig. 2B, Table 1B). We found no evidence of a preference for rough or smooth refuge surfaces in either species (Fig. 2B, Table 1B). We found moderate evidence of individual variation in refuge surface choice (Daily: P = 0.03; Final: P = 0.03, Table 1B). We found no evidence that species, replicate, or weight change affected site fidelity (Table 1B). We also found no evidence of individual variation in site fidelity (Table 1B). The highest number of switches recorded for an individual in an assay was four, and the lowest was zero. We found no evidence that species, replicate, their interaction, or weight change impacts the number of trap lines (Fig. 3B, Table 1B). We found no evidence that the final refuge choice affected the number of trap lines (Fig. 3B, Table 1B). We found strong evidence of individual variation in the number of trap lines produced (P < 0.001, Table 1B). The highest number of total trap lines produced by an individual of either species across both refugia in an assay from both replicates was twenty, and the lowest was zero.

Height Assay:

We found no evidence that species, replicate, or weight change influenced refuge choice (Fig. 2C, Table 1C). However, we found moderate evidence that the interaction between species and replicate influenced refuge choice. We found moderate evidence that false katipō height choices differed between replicates. (Daily: P = 0.04; Final: P = 0.05, Fig. 2C, Table 1C). We found no evidence of a preference for short or tall refuges in either species (Fig. 2C, Table 1C). We found no evidence that species, replicate or weight change affect site fidelity (Table 1C). We also found no evidence of individual variation in site fidelity (Table 1C). The highest number of switches recorded for an individual in an assay was three, and the lowest was zero. We found no evidence that species, replicate, their interaction, or weight change impacts the number of trap lines (Fig. 3C, Table 1C). We found no evidence that the final refuge choice affected the number of trap lines (Fig. 3C, Table 1C). We found strong evidence of individual variation in the number of trap lines built (P < 0.001, Table 1C). The highest number of total trap lines produced by an individual of either species across both refugia in an assay from both replicates was nineteen, and the lowest was zero.

Table 1. Factors influencing refuge preference, the number of switches, and number of trap lines for the shape assay A, surface assay B, and height assay C.

A Shape	Full Model*				Fixed Intercept Model*				Refuge Fidelity				Refuge Trap Lines
Sample size per species	10				10				10				10
Random effects	Variance	SD	χ²	P					Variance	SD	χ²	P	Variance	SD	χ²	P
ID	0.87	0.93	0.42	0.52					0.78	0.88	5.60	0.02	1.70	1.30	65.00	<0.001

Fixed effects	β	SE	z	P	β	SE	z	P	β	SE	z	P	Β	SE	z	P
Intercept	2.26	1.75	1.29	0.20	1.13	0.62	1.84	0.07	0.34	0.86	0.40	0.69	-0.80	0.80	-1.00	0.32
Species	-0.94	1.02	-0.92	0.36					0.10	0.58	0.17	0.87	-1.22	0.68	-1.79	0.07
Replicate	0.21	0.82	0.26	0.80					-0.54	0.40	-1.35	0.18	-0.03	0.19	-0.18	0.85
Weight change	-28.9	17.26	-1.67	0.09					-8.37	9.00	-0.93	0.35	-0.78	5.65	-0.14	0.89
Choice													2.24	0.62	3.59	<0.001

Marginal R²	0.14				0.03				0.19				1.00
Conditional R²	0.20				0.04				0.20				1.00
AIC	52.60				51.80				107.10				176.50

*Because daily and final day refuge choices yielded similar results, only results from the final day refuge choice are shown here. See the supplementary materials for the daily refuge choices results.
P-values < 0.05 indicated in bold.

B Surface	Full Model*				Fixed Intercept Model*				Refuge Fidelity				Refuge Trap Lines
Sample size per species	10				10				10				10
Random effects	Variance	SD	χ²	P					Variance	SD	χ²	P	Variance	SD	χ²	P
ID	8.35	2.89	4.47	0.03					0.23	0.48	0.09	0.76	3.40	1.85	35.62	<0.001

Fixed effects	β	SE	z	P	β	SE	z	P	β	SE	z	P	β	SE	z	P
Intercept	1.63	2.30	0.71	0.48	0.62	0.84	0.74	0.46	-0.28	1.02	-0.28	0.78	0.51	1.45	0.35	0.72
Species	-1.11	1.95	-0.57	0.57					-0.28	0.68	-0.42	0.68	1.88	1.19	-1.58	0.11
Replicate	-1.12	1.23	-0.92	0.36					-0.85	0.60	-1.39	0.16	-0.38	0.50	-0.76	0.45
Weight change	25.63	21.98	1.20	0.24					6.77	4.97	1.36	0.17	0.40	11.03	-0.04	0.97
Choice													-42.65	965.45	-0.04	0.97
Replicate:Choice													20.07	482.72	-0.04	0.97

Marginal R²	0.18				0.11				0.11				1.00
Conditional R²	0.24				0.14				0.13				1.00
AIC	53.20				50.40				67.90				123.80

*Because daily and final day refuge choices yielded similar results, only results from the final refuge choice are shown here. See the supplementary materials for the daily refuge choices results.

P-values < 0.05 indicated in bold.

C Height	Full Model*				Fixed Intercept Model*				Refuge Fidelity				Refuge Trap Lines
Sample size per species	10				10				10				10
Random effects	Variance	SD	χ²	P					Variance	SD	χ²	P	Variance	SD	χ²	P
ID	2820.00	53.10	18.66	<0.001					1.43	1.20	2.82	0.09	3.40	1.85	35.60	>0.001

Fixed effects	β	SE	z	P	β	SE	z	P	β	SE	Z	P	β	SE	z	P
Intercept	11.64	7.21	1.62	0.11	-0.90	1.62	0.56	0.58	-3.22	1.90	-1.69	0.90	-0.51	1.40	-0.37	0.73
Species	-14.33	9.01	-1.59	0.11					-1.58	0.96	0.94	0.10	1.90	1.20	1.50	0.12
Replicate	-0.10	3.67	-0.03	0.98					0.64	0.73	0.72	0.38	-0.38	0.51	-0.74	0.46
Weight change	-6.95	39.00	-0.18	0.86					25.42	12.14	-2.09	0.46	0.40	11.00	0.04	0.97
Species:Replicate	13.69	6.77	2.02	0.04
Choice													-43.00	11739	-0.004	1.00
Replicate:Choice													20.00	5870	0.003	1.00

Marginal R²	0.41				0.22				0.30				0.96
Conditional R²	0.55				0.29				0.35				0.99
AIC	43.9				46.9				68.4				99.00

*Because daily and final day refuge choices yielded similar results, only results from the final refuge choice are shown here. See the supplementary materials for the daily refuge choices results.

P-values < 0.05 indicated in bold.

We found experimental evidence that the invasive false katipō may compete with the endemic katipō for refugia. Both species preferentially selected triangular refugia over rectangular refugia but showed no preferences in the surface or height assays. Individuals varied in their surface and height preferences but not in shape, the reason for which remains unclear. The lack of a shared preference in the surface and height assays might be due to insufficient variation in the height and surface choices used in our experiments. Future research might consider using triangular refuges (the preferred refuge shape by both species) with greater variation in surface and height choices, to determine shared preferences in these features.

Although assessing the drivers of refuge preference was beyond the scope of our study, such drivers are well-documented in other spiders (Glover 2013), including other Latrodectus species. Interestingly, the katipō and false katipō did not exhibit the same preferences in refuge type as the related brown widow spider (Latrodectus geometricus) (Araneae: Theridiidae) in North America where it is invasive. Invasive brown widow spiders use man-made refuges on the undersides of horizontal covers with acute angles, such as the undersides of plastic patio furniture and curved rims of rubbish bins (Vetter et al. 2021). These laboratory experiments revealed that brown widows displayed strong preferences for refugia options across multiple assays. This included in height/depth assays, where spiders preferentially selected 75mm and 100mm depths over 25mm and 50mm depths, as well as surface assays where spiders preferentially selected rough refuges lined with corrugated cardboard over smooth refuges (Vetter et al. 2021).

It is possible that katipō and false katipō may not demonstrate strong refuge preferences due to the need for plasticity in the dynamic nature of sand dune environments. These environments are susceptible to intense weather events that can limit refuge options. Therefore, narrow refugia preferences could severely limit an individual’s survival and reproductive opportunities, given the role of refugia in protection, prey capture, and reproduction for sit-and-wait predators (Scharf and Ovadia 2006). Nonetheless, selecting a particular refuge shape may be important for protection from predation. Triangular refugia potentially provide greater protection from predators compared to rectangular refugia. This is because the acute angle makes it more difficult for predators to access the spider, especially if it resides at the back of the refuge (Cloudsley-Thompson, 1995). This may explain the preference for triangular refuges we observed by both katipō and false katipō.

Our experiments considered site fidelity in the context of refuge choice compared between replicates, as well as refuge switching within assays. Like other spider species, the katipō and false katipō likely use spatial learning to find the most optimal refuge (Punzo, 2004) and may develop fidelity to a refuge. We did not aim to determine whether there was site fidelity in either species, as it is assumed to exist, given the behaviour’s presence in most other spiders, particularly other Theridiidae (Jakob et al. 2001). Katipō spiders showed site fidelity by using the same refuge in each replicate, while the false katipō did so in all but the height assay. The false katipō may have changed their refuge choice in the second height assay replicate to increase perceived fitness, but further investigation is required to understand why this behaviour was specific to this assay and this species. A third replicate may have offered further insight to see if the false katipō repeated their choice or if it was random. While we found no difference in site fidelity between species measured via number of refuge switches, we did, observe individual variation in the number of switches in the shape assay, which was the only assay where spiders showed a refuge preference. It is possible that the individual variation we observed in the shape assay was due to a preference for triangular refuges, but this relationship requires further investigation.

The number of traplines was higher within the preferred triangular refuges than the rectangular refuges. In the shape and height assays where there was no preference, trap lines were evenly spread across both refuge types. Our surface assay results contrast with findings from a similar experiment in brown widow spiders, where more trap lines were established on smooth than rough surfaces (Gutiérrez-Fonseca and Ortiz-Rivas, 2014). We found that the number of trap lines in each replicate was consistent, which was likely due to the spiders being satiated before each replicate. Zevenbergen et al. (2008) found that the western black widow (Latrodectus hesperus) (Araneae: Theridiidae) produce the same number of trap lines in each replicate when fed the same amount of food. We found evidence of some individual variation in the total number of trap lines built in each assay, similar to what DiRienzo and Montiglio (2016) found in western black widows. Interestingly, both katipō and false katipō produced a similar number of trap lines, suggesting that differences in web trap lines are unlikely to contribute to competition between the two species.

We acknowledge that our samples sizes are relatively small, and we interpret our results with caution. As katipō are a threatened species, we limited ourselves to collecting only thirty spiders from the wild, recognizing that removing katipō to the lab could be detrimental to their population. The use of artificial refuges by katipō in their natural habitat has been demonstrated by Costall and Death (2010), who used artificial cover objects for monitoring katipō. The refuges in our experiment were designed to compare factors that might influence refuge choice in the wild for both katipō and false katipō. We found that katipō and false katipō have overlapping refugia preferences which might lead to competition between the species. However, competition does not necessarily indicate displacement. Coexistence also remains a possibility. Further research should compare artificial refuges in the wild to natural refuges to determine preferences in both species, which would allow a better understanding of the value of artificial refuges in katipō conservation.

The critically endangered endemic katipō may be at risk of being outcompeted by the invasive false katipō. Our experimental assays showed that refuge selection, but not web design, may contribute to competition between the katipō and false katipō. However, further research is needed to confirm if displacement is occurring. Given that both species shared similar refuge selection behaviour in our experiments, we cannot yet recommend a specific refugia design that will provide katipō with shelters that are unlikely to be used by false katipō. Since both species produce a similar number of trap lines in different refugia, it seems unlikely that any artificial refugia type will offer katipō a prey capture advantage. Our study considers limited aspects of refuge choice in katipō and false katipō in the laboratory and further research should explore refuge choice in the field, including comparing artificial refuges to natural refuges.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Funding

This work was supported through the Massey University School of Natural Sciences Research Scholarship and 2022 Master’s Research Scholarship.

Author Contributions

All authors contributed to the study conception and design. Data collection was performed by James J. Roberts. Data analysis was performed by James J. Roberts and Anne Wignall. The first draft of the manuscript was written by James J. Roberts and all authors provided comments on previous versions of the manuscript. All authors read and approved the final manuscript.

Acknowledgements

We express our gratitude to The Shorebirds Trust and Tara Iti Golf Course for granting us access to the field site and for their accommodating assistance throughout the duration of our fieldwork. Special thanks go to Linda Guzik and Alex Flavell-Johnson. This project was carried out with the support and permission of the Department of Conservation and Auckland Council. We extend our thanks to volunteers Lily Walkington, Victoria Roberts, Lucy Casley, Jennifer Zhou, Daniel Cooper, Megan Dunnion, and Reshnu Sheela. We also thank Dr. Sheri Johnson for generously providing equipment to aid in the husbandry of the spiders.

Bates D, Mächler M, Bolker B, Walker S. 2015. Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67(1):1–48. https://doi.org/10.18637/jss.v067.i01
Bednarski J, Ginsberg H, Jakob EM. 2010. Competitive interactions between a native spider (Frontinella communis, Araneae: Linyphiidae) and an invasive spider (Linyphia triangularis, Araneae: Linyphiidae). Biological Invasions. 12(4):905-912. https://doi.org/10.1007/s10530-009-9511-7
Blackledge TA, Coddington JA, Gillespie RG. 2003. Are three-dimensional spider webs defensive adaptations? Ecology Letters. 6(1):13-18. https://doi.org/10.1046/j.1461-0248.2003.00384.x
Blackledge TA, Zevenbergen JM. 2007. Condition-dependent spider web architecture in the western black widow, Latrodectus hesperus. Animal Behaviour. 73(5):855-864. https://doi.org/10.1016/j.anbehav.2006.10.014
Boon JS, Keith SA, Exton DA, Field R. 2023. The role of refuges in biological invasions: A systematic review. Global Ecology and Biogeography. 32:1244–1271. https://doi.org/10.1111/geb.13701
Bhuyain MMH, Lim UT. (2019). Interference and exploitation competition between Frankliniella occidentalis and F. intonsa (Thysanoptera: Thripidae) in laboratory assays. Florida Entomologist, 102(2):322–328.
Cloudsley-Thompson JL. 1995. A review of the anti-predator devices of spiders. Bulletin of the British Arachnological Society. 10(3):81-96.
Costall JA, Death RG. (2010). Population monitoring of the endangered New Zealand spider, Latrodectus katipo, with artificial cover objects. New Zealand Journal of Ecology, 34(2):253–258.
Cowan MA, Callan MN, Watson MJ, Watson DM, Doherty TS, Michael DR, Nimmo DG. 2021. Artificial refuges for wildlife conservation: What is the state of the science? Biological Reviews. 96:2735-2754. https://doi.org/10.1111/brv.12776
Debach, P, Sundby RA. (1964). Competitive displacement between ecological homologues. Hilgardia, 34(5):105–166.
Dirienzo N, Montiglio PO. 2016. Linking consistent individual differences in web structure and behavior in black widow spiders. Behavioral Ecology. 27(5):1424-1431. https://doi.org/10.1093/beheco/arw048
Ferretti F, Mori E. 2020. Displacement interference between wild ungulate species: Does it occur? Ethology Ecology & Evolution. 32(1):2-15. https://doi.org/10.1080/03949370.2019.1680447
Gao Y, Reitz SR. (2017). Emerging Themes in Our Understanding of Species Displacements. Annual Review of Entomology. 62:165–183. https://doi.org/10.1146/annurev-ento-031616-035425
Glover N. (2013). 'The habitat preferences of web building spiders', The Plymouth Student Scientist. 6(1):363-375.
Griffiths JW. (2001). Web site characteristics, dispersal and species status of New Zealand's katipo spiders, Latrodectus katipo and L. atritus [Doctoral thesis, Department of Pest Management and Conservation, Lincoln University]. Lincoln University Digital Thesis. https://hdl.handle.net/10182/1680
Guillaumet A, Russell IJ. 2022. Bird communities in a changing world: The role of interspecific competition. Diversity. 14(10):857. https://doi.org/10.3390/d14100857
Gutiérrez-Fonseca PE, Ortiz-Rivas L. 2014. Substrate Preferences for Attaching Gumfoot Lines in Latrodectus geometricus (Araneae: Theridiidae). Entomological News. 123(5):371-379.
Hann SW. (1990). Evidence for the displacement of an endemic New Zealand spider, Latrodectus katipo Powell by the South African species Steatoda capensis Hann (Araneae: Theridiidae). New Zealand Journal of Zoology, 17(3):295-307. https://doi.org/10.1080/03014223.1990.10422937
Jakob E, Porter A, Uetz G. (2001). Site fidelity and the costs of movement among territories: An example from colonial web-building spiders. Canadian Journal of Zoology-revue Canadienne De Zoologie, 79:2094-2100. https://doi.org/10.1139/cjz-79-11-2094
Kraus F. 2015. Impacts from invasive reptiles and amphibians. Annual Review of Ecology, Evolution, and Systematics. 46(1):75-97.
Maun MA. 2009. The biology of coastal sand dunes. Oxford University Press.
Merkle JA, Abrahms B, Armstrong JB, Sawyer H, Costa DP, Chalfoun AD. 2022. Site fidelity as a maladaptive behavior in the Anthropocene. Frontiers in Ecology and the Environment. 20(3):187-194. https://doi.org/10.1002/fee.2456
Muff S, Nilsen EB, O'Hara RB, Nater CR. 2022. Rewriting results sections in the language of evidence. Trends in Ecology & Evolution, 37(3):203-210. https://doi.org/10.1016/j.tree.2021.10.009
Punzo F. 2004. The capacity for spatial learning in spiders: a review. British Arachnological Society. 13(3):65–72.
R Core Team. 2015. R: a language and environment for statistical computing. Available from: https://cran.r-project.org/.
Ryan C, Case BS, Bishop CD, Buckley HL. 2023. Ecosystem integrity of active sand dunes: A case study to implement and test the SEEA-EA global standard, from Aotearoa New Zealand. Ecological Indicators. 149. https://doi.org/10.1016/j.ecolind.2023.110172
Sahni V. Harris J. Blackledge T. 2012. Cobweb-weaving spiders produce different attachment discs for locomotion and prey capture. Nature Communications. 3:1106. https://doi.org/10.1038/ncomms2099
Savvides P, Louca V, Sfenthourakis S. 2015. Competition for shelter occupancy between a native freshwater crab and an invasive crayfish. Aquatic Ecology. 49(2):273–278. https://doi.org/10.1007/s10452-015-9522-7
Shulman MJ. 1985. Coral reef fish assemblages: intra- and interspecific competition for shelter sites. Environmental Biology of Fishes, 13:81–92. https://doi.org/10.1007/BF00002576
Scharf I, Ovadia O. 2006. Factors Influencing Site Abandonment and Site Selection in a Sit-and-Wait Predator: A Review of Pit-Building Antlion Larvae. Journal of Insect Behavior. 19(2):197–218. https://doi.org/10.1007/s10905-006-9017-4
Vetter RS, Penas LM, Hoddle MS. 2021. Laboratory refugia preferences of the brown widow spider, Latrodectus geometricus (Araneae: Theridiidae). The Journal of Arachnology. 44(1):52–57. https://doi.org/10.1636/J15-58.1
Zeileis A, Hothorn T. 2002. Diagnostic Checking in Regression Relationships. R News, 2(3):7–10. https://CRAN.R-project.org/doc/Rnews/
Zevenbergen JM, Schneider NK, Blackledge TA. 2008. Fine dining or fortress? Functional shifts in spider web architecture by the western black widow Latrodectus hesperus. Animal Behaviour. 76(3):823-829. https://doi.org/10.1016/j.anbehav.2008.05.008

SupplementaryMaterial.docx

Refuge preferences and web trap lines: the potential for competitive displacement of endemic katipō Latrodectus katipo by the invasive false katipō Steatoda capensis (Araneae: Theridiidae).

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods