Zoomorphology(1987) 101 :255-259 Zoomorphology O Springer-Verlag 1987

The influenceof web monitoring tactics on the trachealsystems of spiders in the family Uloboridae(Arachnida, Araneida)

Brent D. Opell Departmentof Biology.Virginia Polytechnic Institute and State University, Blacksburg, VA 24061USA

Summary.Uloborids that spin reducedwebs more actively respiratory water loss or to provide a greater and more monitor them than thosethat constructorb webs.H.1'ptiotes direct supply of oxygento prosomaland leg muscles(Levi use both their first and fourth legs to tensetheir triangle- and Kirber 1976). The former advantage is probably of webs,whereas Miagrammopes rely principally on their first particular benefitto small spiders(Levi 1967)and the latter legsto monitor andjerk the threadsof their irregularwebs. to largespiders with highermetabolic rates or more energet- The respiratory systemsof thesespiders include tracheae ic prey capturetactics (Anderson 1970;Cloudsley-Thomp- that extendinto the prosoma,bifurcate, and enter the legs. son 1957;Wilson and Bullock 1973;Davis and Edney 1952; To determineif the legsresponsible for activeweb-monitor- Dresco-Derouet1960). Extensivetracheal systemsboth ing tacticshave more extensivetracheal supplies, the total provide a more direct supply of oxygen to the prosoma crosssectional area has beencomputed of the tracheaeen- and preventthis supplyfrom beinginterrupted during peri- tering the legsof mature femaleorb web and reducedweb ods of activity when fluid exchangebetween opistosoma uloborids. Each leg's value has been divided by the cross and prosoma is interrupted(Anderson and Prestwich1975). sectionalarea of the trachealtrunks that enterthe prosoma. Previous studiesof tracheal developmentin arachnids Theseindexes reveal no significantdifferences between the have attempted to correlate major differencesin tracheal relative tracheal suppliesof the orb weaversinvestigated patternswith size,habitat, and activity patterns.Most of (Waitkera w,aitkerensis,Tangaroa beattvi, (Jloborusglomo- thesecomparisons are betweenmembers of different fami- srzs).But the first, third, and fourth legs of H. cattatu.sand lies and even dilferent orders.None attemptsto quantify the first legs of M. animotusand M. pinopu.shave greater trachealdevelopment. The purposeof this studyis to com- relativetracheal supplies than those of the three orb weav- pare the trachealsystems of similar sizedspiders that belong ing species.Relative to leg volume. the first and fourth to the same lamily and have the same grade of tracheal legs of H. cauatushave the greatestand the first legs of development,but that dilfer in their activity patterns.This Miagramntopesspecies the next greatesttracheal supplies. comparison provides a conservativetest of the hypothesis When tracheallengths are considered.these differences in that tracheal evolution reflects change in respiratory de- potentialoxygen supplies remain, showingthat area differ- mand. If this hypothesisis correct, not only should the encesdo not simply compensatefor differencesin the dis- most activespiders have the most well-developedtracheal tancesover which oxygen must diffuse. These differences systems,but these differencesshould be expressedin the are leg-specihcand not species-specific,and uloborids with legsthat are specificallyresponsible for their greateractivi- the most extensivetracheal supplies are found in moist habi- ty. tats. Thus the observeddifferences are best explained as The Uloboridaeprovide a good opportunityto evaluate adaptationsto meet the greater oxygen demands of legs the associationof tracheal developmentand activity pat- responsiblefor active web-monitoring tactics and not as tern. Although the full spectrum of tracheal patterns is adaptationsto reducerespiratory water loss. found within this family (Opell 1979),the plesiomorphic and most common pattern is characterizedby a pair of stout tracheal trunks that enters the prosoma and divides A. Introduction into smaller tracheae that extend into the appendages (Fig. 1). Within the legs, further branching suppliestra- The respiratorysystems of spidersexhibit much diversity. cheolesto the muscles(Figs. 2, 4).This pattern is found In the ground-planof Araneida two pairs of book lungs in both orb weavinggenera whose members simply hang are present,but in all membersof the large subtaxon Tra- from the hubs of their webs (Opell and Eberhard 1984) cheospira(Platnick 1977), tracheaereplace the posterior while waiting for prey as well as in generawhose members pair of book lungs.Within this group and evenwithin some more actively monitor their reducedwebs and more force- families,e.g., Hahniidae (Forsrer 1970), Linyphiidae (Blest fully rnanipulatethem when a prey is caught (Lubin 1986; 1976;Millidge 1984,1986), Uloboridae (Opell 1979),the Lubin et al. 1918;Opell 1982,1984a; Peters 1938). Despite trachealsystem exhibits varying degreesof development, differencesin web rnanipulation,all uloboridsexhibit simi- rangingfrom fine tracheaethat are restrictedto the opisto- lar prey wrapping and handling behaviors(Lubin 1986). somato stout trachealtrunks that extendinto the prosoma Lacking poison glands (Opell 1979).they artack-wrapen- wherethey bifurcateand enter the appendages. snaredprey, but neither bite nor grapple with struggling More extensivetracheal systemsare thought to reduce prey. 256

Tangaroa beattyi opell, 1983, from the caroline Islands, and uloborus glomoszs (walckenaer, rB41) from virginia. The triangle-web species Hyptiotes ceuatus (Hentz, lgql) was collected from Virginia, the"single-line-weaver,, Mia_ gratnmopes cuimotu.r chickering, 1968, from the center for Energy and Environment Research'sEl verde, puerto Rico field station, and an undescribed member of the Miagrant- tnopes aspinatus speciesgroup (opell l9g4a) from the orga- nization for Tropical studies' La Selva, costa Rica field station. Mean carapace lengths of these species are: 1312 pm, 760 pm, 1372 pm, 906 prn. 1512 pm, and 1138 pm, respectively. waitkera and Tangaroa exhibit the most primitive grade of genitalic development found in the uloboridae and uloborus has the most advanced grade found in orb weaving genera that retain an extensivetra- cheal system (opell 1979). Hltptiotes and, Miagrantmopes are sister taxa (Opell 1984b). Tracheal lneasurements were taken from plastic-embed- ded speci'ens cross sectioned with a Sorvall JB-4 micro- tome. Prior to sectioning, specirnens were frxed in 3% glu- taraldehydel3% fonnaldehyde in 0.1 M sodium cacodylate buffer (pH 7.3) at 20"-26o C for 12-18 h, rinsed in 0.1 M sodium cacodylate buffer, dehydrated through a graded se- ries of acetone, and embedded in Spurr's epoxy resin. 1 pm thick cross-sectionswere stained with 1% toluidine blue. An index of relative tracheal developrnent was obtained by dividing the total cross sectional area of the tracheae entering a leg by the cross sectional area of the tracheal trunk serving that half of the prosoma (Fig. 1). Each of the two tracheal trunks entering the prosoma serves only one side of the body, and so right and left legs of the same specimen were treated as separate samples. The diameters of tracheal trunks entering the prosoma and of tracheae entering leg coxae (Fig. 3) were measured at x 500 with the ocular micrometer of a compound microscope. In one U. glomosas specimen, damage prevented me f,rom measur_ ing the legs on one side of the body. To avoid using dupli_ cate measurements of looped tracheae, measurements were taken during the course of a careful posterior-to-anterior examination of serial sections.If a trachea's perpendicular-measured profile deviated noticeably from a circle, I its diameter in at least three axes and used the average diameter 2 for calculations. If a trachea was obliquery sectioned,I used the maximum width of its lumen as its diameter. From 1, 2. The opistosomaland prosomal (Fig. Iqr. l) and first leg these measurements, I computed the cross sectional area (Fig.2) trachealsystems of Lllobctrusgromosus as seenin dorsal of each tracheae and then summed the cross sectional area viewsof a sodium hydroxide-clearedlemale specimen of tracheae serving each leg. To estimate the volume of the first and fourth leg muscle If tracheal development reflects oxygen demand, the legs served by the tracheae of these legs, I determined the mean used by these reduced web uloborids to monitor and oper- volumes of the right legs of five mature female specimens ate their webs should have more extensive tracheal r.tppli.t of each species except T. beattyi. The volume of each leg than the legs of orb weavers with the same type of tracheal article was determined by multiplying its mean cross sec- pattern. To test this hypothesis I compared the tracheae tional area by its length. The diameters used to compute of three orb web species, one triangle-web species whose this cross sectional area were obtained as follows: coxa, members use both their first and fourth legs io tense their mean central width and height; trochanter, mean central reduced webs (Opell 1982,1985, 1987) and two,.single-line_ width and height; femur, mean width and height at proxi_ web " species that rely on their first legs to moniior and mal 25o/o of length, center, and distal 25% of length; pa_ manipulate their webs (Lubin 1986; Lubin er al. l97g). tella, mean central width and height; tibia, same -eusure_ ments as femur, metatarsus, mean width and height at B. Materials and methods proximal 33oh and distal 33"/o of length; tarsus, same mea- surements as metatarsus. only mature female specimens were used in this study. The The cross sectional area of the tracheae serving a leg three orb weaving species studied were: waitkera w,aitker- is not the only lactor influencing the availability of bxygen ensis (Chamberlain, 1946) from northern New Zealand. to the leg muscles. If oxygen moves principally by diffuii,cn, 257

Figs.4.Fig.3,Cross-sectionthroughtheoriginoftherightfrstcoxaofUloborusg/otrorrr,showingtwolargeandtwosmall tracheae erfering the feg Fig.4. First leg muscle fibers of Miagrantnopes animotus and a Jargeand small tracheole supplying them with oxygen. 7 tracheaor tracheole;MFmuscle fiber; G first leg ganglion ofsubesophagealganglion; N nervesenervating the fiIst leg; M mitochond a

Trble l. Samplc size, number of tracheae entering each leg coxa, total cross sectional area of tracheae seNing each leg, and first and fourth leg volumes. Mean and (standard deviation) are presented

Waitkera Tangaroa Uloborus Hyptiotes Miagrammopes Miagrammopes wailkerensis betttttti glotnosus c0Dalu.l Qttimolus sp.

Leg I : Spiders/sides 214 112 214 112 316 Tracheae 2or3 1 6or9 2or3 1 1or2 Area (pm2) 331 (4s) 128 (28) 28e (23) 387 (37) 891(10e) s9't (122)

'1 tA Leg 2: Spiders/sides 214 112 )t4 Ll- 112 316 Tracheae 1or2 1 4or6 2 6 13 Area (trrm2) 133 (3e) 62 (1e) 157 (s1) 1se (40) 421(13) 161(43) Leg 3: Spiders/sides 112 112 316 Tracheae 1 1 4 2 1 1 Area (pm2) 110 (36) 38 (1s) 101(31 ) 1s3 (30) 83 (11) 88 (33) Leg 4: Spiders/sides )t7 112 214 112 316 Tracheae 1or3 1 8or14 3or5 1 13 Area (pm2) 312 (16) e2 (23) 383 (33) s06 (100) 416(2s) 302(14)

Total leg areas (pm2) 8e0 (24) 318(85) 855(e) 12s2(122) 1811(107) 1148(166) Pedicle: Spiders/sides 214 112 214 214 112 316 Total tracheae 2 2 2 2 2 2 Area/tracheae (pm2) 1e48(1e8) 120 (64) 1s4e(131) 1747(3se) 33ss(138) 2461(3s7) Leg I volume (104 pm3) 62798(1 1 483) 48822(1e43s) 27438 (2020) 104188(15300) 46e0s (5404) Leg4 volume (104pm3) 43327 (833e) 33069(12133) 2e37s(6264) 402s3 (417s) 36387(119s3)

the speed of its movement is directly proportional to the C. Results square root of tracheal length (Anderson and Prestwich 1980; Schmidt-Nielsen 1979). Therefore, a short, small di- Table 1 presentsmean leg cross sectionalareas and vol- ameter trachea could supply as much oxygen as a long, umes, Table 2 indexesof the relative tracheal suppliesto large diameter trachea. To evaluate this interaction, I di- the legs of each species.The indexespresented in Table 3 vided the total cross sectional area of tracheae entering account for differencesin the lengths of tracheaeserving each leg by the square root of the distance from the spiracle the legs.From Table2, four trends are apparent:(1) the to the leg base (Fig. 1). I estimated the latter by measuring first and fourth legs of orb weavershave similar relative the distance from the tracheal spiracle to a point on the trachealsupplies and thesesupplies exceed those of the sec- sternum midline perpendicular to each coxa and added to ond and third legs;(2) reducedweb uloboridshave better this the distance from this sternal point to each leg's coxa. developedfirst leg tracheaethan do orb weavinguloborids; 258

Table 2. Indexes of each leg's tracheal supply relative to the area of tracheae entering the prosoma and, for first and fourth legs, total leg volume x 104 pm3. Mean and (standard deviation) of the former index are presented. The latter index is derived from means presentedin Table I

Waitkera Tangaroa Llloborus Hyptiotes Miagrammopes Miagrammopes waitkerensis beattyi glomosus cauatus unimotus sp.

Leg 1: Area/pedicle area 0.17(0.03) 0.18 (0.05) 0.18(0.03) 0.23 (0.03) 0.27 (0.02) 0.25 (0.09) Volume/trach. area 190 169 71 117 79

Leg 2: Area/pedicle area 0.07 (0.02) 0.0e(0.03) 0.10 (0.04) 0.10 (0.04) 0.13 (0.001) 0.07 (0.02)

Leg 3: Area/pedicle area 0.06 (0.01) 0.0s(0.03) 0.06 (0.02) 0.0e (0.01) 0.03 (0.002) 0.04 (0.02)

Leg 4: Area/pedicle area 0.16(0.03) 0.13 (0.04) 0.24 (0.04) 0.2e (0.04) 0.12(0.01) 0.13(0.02) Volume/trach. area 139 86 58 97 121

Total les trach. area 0.47 (0.05) 0.45(0.16) 0.s2(0.04) 0.71 (0.12) 0.54 (0.01) 0.48 (0.13) 0.5 Pedicle trach. area

Table 3. Mean estimatesof the lengths of tracheaeserving cach leg and indexesof potential oxygen availabiTity.yalnes ln parenthesesarc the standard deviationsof distancemeasurcments. Cross sectionalareas used in thesecalculations are those means presentedin Trble I

Waitkera Uloborus Hyptiotes Miagrammopes Miagrammopes waitkerensis glomosus caDatus animotus sp.

Leg I: Distance (pm) 3072 (326) 2s72(s38) 2088 (230) 3772 (464) 2804 (248) Area//distance 5.97 5.70 8.47 14.51 11.28

Leg II: Distance (pm) 2196 (330) 2312(s06) 1800(1e7) 3388 (471) 2s80 (220) a a- a aa Area//distance 2.52 J.Z I ).-) I | .zJ 3.11

Leg III: Distance (pm) 2536 (305) 2004 (483) ts72 (209) 30e6(44s) 2348 (207) Areall/ distance 2.18 2.26 3.86 1.49 1.82

Leg IV: Distance (pm) 2204 (273) 1664 (472) 1240 (161) 2668 (3ee) 2028 (218) Arealy' distance 6.65 9.39 14.3"7 8.05 6.71

(3) in Hyptiotes the fourth legs have the most highly devel- and fourth legs.In M. animotusthere is no significantdiflfer- oped tracheal systems, whereas in Miagrammopes the first ence between tracheal suppliesto the second and fourth leg tracheae are most highly developed; and (4) the proso- legs. mal tracheal system of H. cauatus is more extensive than When interspecific differencesare afialyzedwith a two- those of the other species. tailed /-test, there is no significant differencebetween the When intraspecific dilferences in relative tracheal sup- relative tracheal supplies to the legs of orb weavers,al- plies are analyzed with a two-tailed /-test, there is no signifi- though the greatersupply to U. glomosusfourth legs (P: cant difference (P>0.05) between the first and fourth legs 0.051)closely approaches statistical significance. When rela- of each of the three orb weavers. By contrast, these values tive tracheal values for the three orb weaving speciesare differ significantly (P< 0.05) in H. cauatus, where the fourth pooled and compared with those of reduced web species leg has a relative tracheal supply that is 1.26 times greater using a two-tailed r-test, the tracheal suppliesof all but than that of the first leg, and in the two Miagrammopes the secondleg of H. cauatusare significantlygreater (P < species,where the first legs have tracheal supplies that aver- 0.05)than thoseof the orb weavers.All M. animorzslegs age 2.09 times greater than those of the fourth legs. Only differ significantlyfrom those of orb weavers,but only the in Miagrammopes do relative tracheal supplies to the second third legs of M.sp.differ significantlylrom thoseof orb and third legs differ significantly. In both species the third weavers.However, P values for all but the second pair leg has a smaller relative tracheal supply than the second. of M. sp. legs are lessthan 0.09, indicatingthat they do In all species but M. animotus the second and third leg not differ fundamentally from those of M. animotus.In tracheal supplies are smaller than those of both the first Miagrammopes only the first pair of legs have consistently 259

greater relative tracheal supplies tl-ranthose of orb weavers. Tangrtroabeattyi specimcns. Paula E. Cushing,Amy D. Ware, and The third and fourth legs of Miagrammopes have smaller ElizabethM. Reinoehlsectioned specimens and assistedwith mea- relative values than those of orb weavers. surements.National Science Foundation grants No. DEB-8011713 and BSR-S407919to the author supportedthis study.

D. Discussion References The results support the hypothesis that web monitoring AndersonJF (1970)Metabolic rates of spiders.Comp Biochem tactics have influenced uloborid tracheal development. Physiol 33:51 72 causingmore actively monitoring spidersto have more well- [A] Anderson JF, PrestwichKN (1975) The fluid pressurepump of developed tracheal supplies to those legs that are responsi- spiders(Chelicerata, Araneae). Z Morphol Tiere81:257 277 ble for this activity. Indexes of potential oxygen availability AndersonJF, PrestwichKN (1980)Scaling of subunitstructures (Table 3) show that cross sectional differences are not sim- in book lungsof spiders(Araneae). J Morphol 165:167-174 ply compensatory changes related to differences in tracheal Blest AD (1976) The trachealarrangement and the classification lengths. The first legs of H. cauatu,rand the two Miagram- of linyphiidspiders. J Zool 180: 185-194 mopes specieshave the greatest values, despite the fact that Cloudsley-ThompsonJ (1957)Nocturnal ecologyand water regula- H. cauatus has the shortest tracheae and Miagrammopes tions of the British cribellate spidersof the genus Ciniflo. J Linn London 43:133-152 have the longest. Likewise, leg proportions do not affect Soc Davis ME, Edney EB (1952) The evaporation of water from the leg volume served by each 1 p t of tracheal cross-sec- spiders.J Exp Biol29:571-582 (Table tion 2). The smallest values (rnost extensivesupplies) Dresco-DerouetL (1960)Etude biologiquecompar6e de quelques are found rn H. cauotus which has short, stout legs and espdcesd'Araignees lucicoles et troglophiles.Arch Zool Exp in the two Miagrammopes specieswhich have long, slender Gen 98:271-354 legs. ForsterRR (1970)The spidersof New Zealand:IIL Otago Mus Two observations suggest that tracheal dilferences are Zool Bull3:1-184 not primarily adaptations to reduce respiratory water loss. Levi FIW (1967) Adaptations of respiratory systemsof spiders. First, there is no clear association of tracheal development Evolution21 :511 583 Levi HW, Kirber WM (1976) in with habitat moisture: T. beattyi, H. cauatus, and the two On the evolution of tracheae arachnids.Bull Brit Arach Soc3:187 188 Miagrammopes species are found in forest habitats with Lubin YD (1986)Web building and prey capturein Uloboridae. greatest the moisture and W. waitkerensis and U. glomosus In: Shear WA (ed) Spiders:webs, bchavior,and evolution. in more exposed habitats (personal observations, personal Stanford University Press,Stanford, pp 132-171 communications from James Berry, Joseph Beatty, and Da- Lubin YD, EberhardWG, MontgomeryGG (1978)Webs of Mia- vid Court). Second, tracheal development is leg-specific grammopes(Araneae: Uloboridae) in the neotropics.Psyche rather than species-specific,as would be expected if it were 85:1 23 primarily a responseto water conservation. MillidgeAF (1984)The taxonomyof the Linyphiidae,based chiefly Likewise, differences in relative tracheal supply to the on the epigynaland trachealcharacters (Araneae: Linyphiidae). Bull Brit Arach legs cannot be attributed to differences in tracheal branch- Soc6:229-267 Millidge AF (1986) A revision of the tracheal structuresof the ing patterns. For example, the first legs of both H. cauatu.s Linyphiidae(Araneae). Bull Brit Arach Soc2:57-61 and W. w,aitkerensishave two to three tracheae, but the Opell BD (1979) Revision of the generaand tropical American former has significantly greater relative cross sectional area. speciesof the spiderfamily Uloboridae.Bull Mus Comp Zool Similarly. a single trachea serves the first legs of both 7. Harv Univ 148:433 549 beattyi and M. animotus, but the latter has et significantly Opell BD (1982)Post-hatching development and web production greater relative cross sectional area. These observations sug- of Hyptiotes calutus (Hentz) (Araneae: Uloboridae). J Arach gest that tracheal branching patterns are conservative and 10:185-191 that increased oxygen demands are met principally by in- Opell BD (1984a\Phylogenetic review of the genusMiagrantmopes (Araneae: Arach 240 creasesin tracheal diameters. sensulato Uloboridae).J 12:229 Opell BD (1984b)Comparison of the carapacefeatures in the fami- The differences in tracheal development documented in ly Uloboridae(Araneae). J Arach 12:105-114 this study strongly suggest that reduced web uloborids ex- Opell BD (1985)Force exerted by orb-weband triangle-webspiders pend more energy in monitoring their webs than do orb of the family Uloboridae.Can J Zool 65:580-583 web uloborids. The specific way in which the tracheal sys- Opell BD (1987)Changes in web-monitoringforces associated with tem adapts to these demands suggeststhat the spiders' basal web reduction in the spider family Uloboridae. Can J Zool (resting) metabolic rates may not differ. In light of Ander- 65:1028-1034 son's (1970) findings that the resting metabolism of spiders Opell BD. EberhardWG (1984)Resting posturcs of orb-weaving correlates highly with their book-lung surface areas, this uloboridspiders. J Arach 11:369-376 (193S) hypothesis predicts that there will be little difference in the Petcrs HM Uber das Netz Der Dreieckspinne,Hyptiotes paradoxus.Zool Anz 121:49 59 relative book lung surlace areas of the spiders studied here. Platnick NI (1977)The hypochiloidspiders: a cladisticanalysis However, the less active web monitoring tactics of orb with notes on the Atypoidea (Arachnida, Araneae).Am Mus weavers may impose less acute oxygen demands that can Novitates2627:1-23 be more easily met by oxygen disolved in the hemolymph. Schmidt-NielsenK (1979)Animal physiology:adaptation and en- If this is the case, orb weaving uloborids may rely more vironment,2nd edn. CambridgeUniversity Press,London extensively on their book lungs and, consequently, have Wilson RS, Bullock J (1973) The hydraulic interaction between relatively larger book lung surface areas. prosomaand opistosomain Amaurohius.fbrox(Chelicerata, Ar- aneae).Z Morphol Tiere74:221 230 Acknowledgements.David Court collectedand fixed Waitkera w'ttit- kerensisspecimens and JosephBeatty and James Berry collected ReceivedJanuarv 7- 1981