Convergent setal morphology in sand-covering spiders suggests a design principle for particle capture: Electronic Supplementary Material
Rebecca P. Duncan, Kellar Autumn, Greta J. Binford*
Department of Biology
Lewis & Clark College
0615 SW Palatine Hill Rd
Portland, Oregon 97219
* Author to whom correspondence should be addressed. , phone (503)768-7653, fax (503)768-7658
Methods
(a) Taxon sampling and rationale for haplogyne outgroups
We surveyed hair morphology on the carapace of haplogyne outgroup taxa that were selected to represent a broad phylogenetic range (ESM figure 1), including the most basal family and at least one family in most major clades. The goal was to enable estimation of the timing of evolutionary origin of hairlettes in this lineage. Our taxon sampling is denser among taxa closely related to Sicarius. In the genus Loxosceles we sampled representatives of all but one currently described species group.
(b) Specimen collection and care for non-Sicarius Haplogynes
All spiders were collected in the field at the locations indicated in ESM table 1. More detailed collection localities are available from RPD or GJB. Loxosceles, Drymusa and Scytodes were reared in the lab as described for Sicarius and Homalonychus, but were not kept in sand.
(c) Sample preparation for comparative analyses of setal morphology and sand attachment
As Loxosceles, Drymusa and Scytodes were important genera for other research, we used molted carapaces from these genera to conduct all analyses (as described for Sicarius and Homalonychus). For all other outgroups, we isolated the cephalothorax from spiders previously preserved in 75% ethanol, subjected them to three washes in 100% molecular grade ethanol (at least three hours per wash), and dehydrated them in a CPD2 critical point dryer (Ted Pella). After critical point drying, we mounted samples on SEM stubs using either colloidal silver liquid (Ted Pella) or superglue (Devcon) and let the adhesive set overnight. We dusted samples in fine sand (grade 0; ESM figure 3) and applied acceleration as described before sputter coating and imaging in the SEM.
(d) Determining the particle sizes that attach to spiders
We separated sand bought from a hardware store into three size categories using a soil sieve with 3 screens (Hubbard Scientific; mesh sizes 35, 60, 120). Categories were named grades 0-2 according to size, where grade 0 contained the finest particles and grade 2 contained the coarsest. We took small samples of each and measured the particle sizes from images in ImageJ (NIH) to generate a particle size distribution as follows: For grades 1 and 2 we cast a small sample of particles into a Petri dish and imaged under a dissecting microscope (Nikon SMZ1500) with a digital camera (Nikon Coolpix 995). For grade 0 we obtained a small sample on a SEM stub with double-stick tape (Scotch), sputter coated it and imaged the sample by SEM because particles were too small to measure accurately under a dissecting microscope. We took images at three magnifications (500x, 1,000x and 7,000x) to show the full range of grain sizes in the sample. We captured images at 1,000x and 7,000x in regions within the 500x images. For all samples, we took the first image at a section of the Petri dish or SEM stub where the particles were far enough apart to distinguish the edges of each grain. We captured subsequent images at an equal number of frames from each.
To calculate the size distribution of particles from the images, we measured the length at the longest part, and the width at the widest part perpendicular to the length measurement of each grain. For grade 0 sand, sand grains >m in length were measured from images at 500x magnification and sand grains between 0.5 and 10m were measured and counted from images at 1,000x magnification. The smallest sand grains were too small to measure accurately, so sand grains under 0.5m were counted and not measured from images at 7,000x magnification. We divided the number of sand grains by the area of each image to estimate the number of sand grains per square mm and used these estimates to generate a frequency distribution of sand grains in three categories: >10m, 0.5-10m, and <0.5m.
To determine which size range of sand could attach to the spiders, we placed live, freshly molted and undusted Sicarius and Homalonychus representatives of various life stages and from various populations in boxes containing different grades of sand. We monitored sand capture by self-dusting over the course of four days, using a different individual for each assay. We visualized spiders under a dissecting microscope (Nikon SMZ1500) and assessed whether sand capture had occurred and if any cuticle was still visible. We took images with a digital camera (Nikon Coolpix 995) before and 24 and 96 hours after placing spiders in sand. We assessed the degree to which ceramic microspheres attach to Sicarius using the same protocol.
ESM Table 1. Species with representatives whose carapace setal morphologies were surveyed by SEM. N is the number of individuals surveyed for each species.
Species / Collection locality (GPS information available upon request) / N / ReferenceHomalonychus selenopoides / USA: Pima Co, AZ, Organ Pipe Cactus National Monument (three localities) / 5 / Roth 1984
Homalonychus selenopoides / USA: Pima Co, AZ,.Tucson Mountain State Park / 1 / Roth 1984
Homalonychus theologys / USA: San Diego co, CA, Anza-Borrego Desert State Park (two localities) / 2 / Roth 1984
Homalonychus theologus / USA: Imperial co, CA, SW of Hwy S2 / 1 / Roth 1984
Homalonychus theologus / USA: Palm Springs, CA, Boyd Research Center, Deep Canyon / 3 / Roth 1984
Sicarius sp. / SOUTH AFRICA: Northern Cape Province, Oorlogskloop Nature Reserve / 2
Sicarius albospinosa / NAMIBIA: Brandberg Mountains / 1 / Purcell, 1908
Sicarius albospinosa / NAMIBIA: Gobabeb Research Station / 3 / Purcell, 1908
Sicarius albospinosa / NAMIBIA: Twyfelfontein, Wundergat / 1 / Purcell, 1908
Sicarius dolichocephalus / NAMIBIA: Ruacana Falls / 3 / Lawrence, 1928
Sicarius damarensis / NAMIBIA: Munsterland cave, Near Outjo / Lawrence, 1928
Sicarius rupestris / ARGENTINA: Rio Negro, Paso Córdoba / 3 / Holmberg, 1881
Sicarius terrosus / ARGENTINA: San Luis, Parque Sierra de las Quijadas / 3 / Nicolet, 1849
Sicarius patagonicus / ARGENTINA: Neuquin, Arroyito Dam / 1 / Simon, 1919
Sicarius peruensis / PERU: Tumbes, 19km North of Mancora / 1 / Keyserling, 1880
Sicarius peruensis / PERU: Quebrada Mogollon / 1 / Keyserling, 1880
Loxosceles spinulosa / NAMIBIA: Ruacana Falls / 1 / Purcell, 1905
Loxosceles spinulosa / NAMIBIA: Windhoek, Arebusch Campsite / 2 / Purcell, 1905
Loxosceles spinulosa / NAMIBIA: Munsterland cave, Near Outjo / 1 / Purcell, 1905
Loxosceles spinulosa / SOUTH AFRICA: Northwest Province, Borakalo National Park / 1 / Purcell, 1905
Loxosceles vonwredei / NAMIBIA: Uisib farm, Otjozondjupa / 5 / Newlands, 1980
Loxosceles laeta / ARGENTINA: Buenos Aires / 3 / Nicolet, 1849
Loxosceles intermedia / ARGENTINA: Parque Nacional el Palmar / 1 / Mello-Leitão, 1934
Loxosceles blanda / USA: Carlesbad, NM / 3 / Gertsch and Ennik, 1983
Scytodes sp. / PERU: Junin, Vitoc / 1
Scytodes sp. / PERU: Cuzco, Quillabamba / 2
Scytodes sp. / PERU: Lambayeque, Olmos, Along Rio Marañón, near El Mayo / 1
Scytodes sp. / PERU: Pevas / 1
Drymusa serrana / ARGENTINA: San Luis, Merlo / 4 / Goloboff & Ramirez, 1991
Diguetia canites / USA: Pima co, AZ, Brown Canyon, Baboquivari Mountains / 3 / McCook, 1889
Plectreuris tristis / USA: Pima co, AZ, Tucson Mountain Park / 1 / Simon, 1839
Plectreuris tristis / USA: Black Canyon City, AZ / 1 / Simon, 1839
Dysdera crocata / USA: Multnomah co, OR, Portland / 2 / Koch, 1838
Kukulkania arizonica / USA: Pima co, AZ, Catalina mountains, Rillito wash / 1 / Chamberlin & Ivie, 1935
Kukulkania arizonica / USA: Pima co, AZ, Catalina mountains, Sabino cree / 1 / Chamberlin & Ivie, 1935
ESM Figure 1. Phylogenetic tree of Haplogynes. Phylogenetic relationships between Sicarius and other haplogyne taxa based on morphology. Asterisks mark the outgroups represented in sand attachment and setal morphology comparative analyses. Tree is based on Platnick et al., 1991
ESM Figure 2. Sand associates less with setae lacking hairlettes but still clumps densely in setose regions of the carapace. To establish the role of setal morphology in sand adhesion we dusted the cephalothorax or carapace of 8 haplogyne outgroups with fine sand, applied acceleration as described in methods and compared them to dusted Sicarius and Homalonychus using SEM. Sand adhered to all outgroups, but largely associated with regions containing dense setae. Sand clumping occurred on setose regions of the cuticle even when particles did not associate strongly with individual setae (e.g. Loxosceles spinulosa, Plectreuridae, Filistatidae). The degree to which particles adhered to individual setae varied, but was generally much less than the degree to which they adhered to Sicarius and Homalonychus setae. Particles were often trapped between or under setae (arrows, Scytodes, Drymusa, Dysderidae, Plectreuridae) and associated more strongly with shorter setae that lay close to the cuticle than with longer setae (arrow, Drymusa). n = 2 Filistatidae, 2 Dysderidae, 2 Plectreuridae, 3 Diguetidae, 4 Drymusa serrana, 5 Scytodes sp., 5 Loxosceles spinulosa, 5 Loxosceles vonwredei.
ESM Figure 3. Particle size distributions of grades 0, 1 and 2 sand. We separated sand into three size categories using a soil sieve with three screens (mesh sizes 35, 60 and 120; Hubbard Scientific) and characterized the particle size distribution of each category by measuring sand grains in a sample of each one. Categories were named grades 0-2 from finest to most coarse particle sizes. Grade 0 sand was used in all comparative analyses of sand attachment.
ESM Figure 4. Fine particles attach to Sicarius and Homalonychus. We placed live Sicarius in each size category of sand and in ceramic microspheres (3m; mean diameter = 40m) and allowed them to self-dust. We placed live Homalonychus in grades 1 and 0 sand. We monitored them over the course of 4 days to determine the range of particle sizes that could stick to them. (a) Grade 2 sand did not adhere (n=4) to Sicarius. Grade 1 sand adhered to Sicarius (n=7), but did not completely mask the cuticle after 4 days. Only grade 0 sand and the microspheres totally covered Sicarius (n=4 for each) to the point that none of the cuticle was visible. (b) In Homalonychus, grade 1 sand adhered (n=4), but a substantial amount of cuticle was visible after 4 days. Grade 0 sand completely covered the cuticle in most individuals, leaving only small areas where the colour of the cuticle showed through in one (n=4). In all individuals dusted with grade 0, sand failed to cover a small region on the posterior slope of the carapace (arrowhead).
1