JOURNAL OF MORPHOLOGY 206:211-216(1990)

The Refationshipof Book Lungand Tracheal Systems in the FamilyUloboridae BRENT D. OPELL nia Po lv t ec hni c l rx t i t ute and s t at e u niue r s it v Bf::I r,i;,', ltt#;Y l ;#F

ABSTRACT The book lung surface areas of representatives of six genera were measured and divided by spider prosomal and leg mass and by the product of prosomal length and width to obtain two indexes of book lung area. Both indexesare greatest in speciesthat have the simplest tracheal systemsand smallest in those with the most extensivetracheal systems.When speciesare ordered phylogenetically and the lung area indexes of their hypothetical ancestorscomputed, a transformational increasein lung area is indicated. Neither the lung area of a speciesnor its phyloge- netic position is significantly related to the generalmoisture regime of its habitat. The number of leavesin a book lung is positively correlated with both spider sizeand total lung area. However, leaf number is not correlated with tracheal development and showsno transformational change.

The Tracheospira (seruu, Platnick, '77) con- branch to enter the legs. Tracheae entering the tains that have two respiratory compo- prosomaeof and have nents: a pair of book lungs that oxygenate the diameters that are, relative to combined proso- hemolymph and a tracheal system that carries mal and leg masses,nearly twice those of Waitk- oxygen directly to tissues. A variety of spider era artd (Table 1). This difference is tracheal patterns have been documented and attributed to the fact that they more actively '87; their adaptiveness discussed (Bromhall, monitor and manipulate their reduced prey cap- Opell,'87a). However,the relationship between ture webs(Opell'87a,b). Orb-weavers of the gen- these patterns and book lung development has era and have tracheae not beenstudied. that are restricted to the opisthosoma (Opell, Spider book lungs residein the anterior, ven- '79). Opell ('79) reported that the tracheolesin tral region of the opisthosoma. Each lung is Octonoba extend into the prosoma, but none formed of a stack of elongate,flattened, cuticle- appeared in cross sections through the pedicels lined leaves,between which hemolymph circu- of North American used in lates.The air spacebetween the two lamellae of this study (unpublished observations). each leaf is maintained by cuticular struts and Comparisons of the relative book lung surface communicateswith a common atrium that opens areasin these taxa make it possibleto determine '67; to the outside via a spiracular slit (Levi, how tracheal development,activity patterns, and '76; '80; Moore, Anderson and Prestwich, Hex- phylogenetic position influence book lung devel- ter,'82). The width of the interlamellar air space opment. If spiders with more extensive tracheal within eachleaf showsa negativeallometric rela- systems have greater book lung surface areas tionship with spider mass, €Lnapparent adapta- than those with more weakly developed tra- tion to reduce the distance over which oxygen cheae,this would suggestthat both systems act must diffuse to reach the hemolymph (Anderson in consort to supply a spider's increasedoxygen and Prestwich,'80). demands.If more extensivetracheal systemsare Members of the family exhibit the associatedwith smaller book lung surface areas, full range of tracheal development found in spi- this would indicate that the tracheae meet more '87). ders (Opell,'79,'87a; Bromhall, The orb- acute or more specific oxygen demands (such as weavinggenera and Uloborus, the tri- those imposed by web-monitoring behaviors) angle-weaving genus Hyptiotes, and the and, therefore have assumed a greater role in irregr.rlar-web-weavinggenus Mi agranxnxopesre- meeting a spider's total oxygen demands. Any tain the family's plesiomorphic pattern, charac- transformational changesin book lungs of these terized by two stout tracheal trunks that extend speciesshould be apparent when their relative through the pedicel, into the prosoma, and surface areasare ordered phylogenetically.

o 1990WILEY-LISS. INC. TABLE 1. Comparbonof tracheal d.etelnpneat,spider size, booh lungsurface areas,and. interhmellar ah apaNewidth.s (nean t standarddcoiatiaa [sampla size]) Do not enter prosoma Enter prosoma Tracheae, Orb-web Orb-web Reduced-web web type, species P. tingerw O. sinensis W. waitakeriensis U. glomosus M. animotus H. cauatus Pedicle tracheal area'/ prosomal and leg mass 0 0 410 448 76t 840 Carapacelength (pm) 1,460t 158(10) 1,484* 126(10) 1,252r 88 (5) 1,364t 49 (9) 1,507t 112(10) 1,017t 69 (7) Carapacewidth (pm) 1,358t 145(10) 1,394t 134(10) 1,168t 70 (5) 1,171t 45 (9) 1,0?3t 57 (10) 1,151t 32 (7) Prosomal and leg mass (mg) 2.84t 0.95(10) 3.43t 1.25(10) 2.L7t 0.26(7) 1.91t 0.30(9) 2.38t 0.53(10) 1.49t 0.26(7) Book lung leaves 29.4r 6.3(10) 30.9 t 6.7 (9) t7.2x2.8(5) 229 t 2.0(7) 22.4 x3.8(9\ r5.7 t 7.7(7) Number of lung leaves/ prosomal and leg mass (mg) r0.4x.2.4 (r0) 9.2 r 2.2(9) 8.2t 2.5(5) tr.9 t 2.2 (7) 9.9t 1.0(9) 10.0t 1.9(7) Book lung area (mm') 6.97t 2.55(10) 7.47+ 2.76(9) 2.53 t0.47 (7) 3.77t 0.61(8) 2.98t 0.91(10) 2.08t 0.31(7) Lung area (mm')/ prosomal and leg mass (mg) 2.35r 0.51(10) 2.13r 0.33(9) 1.19r 0.31(5) 1.92t 0.31(8) 1.26t 0.19(10) 1.43x 0.29(7) Lung area Atm')/ ctuapacelength x width 0.83* 0.20(10) 0.87t 0.17(9) 0.4a t 0.08(5) 0.59 t 0.09(8) 0.46t 0.10(10) 0.45 t 0.07(7) Live spider mass (mg) L4.52t 3.63(20) 9.93 r 4.65(45)'z 6.39 t 3.36(37)'? 6.76 t 3.06(42)2 Predicted interlamellar air space width (rrm) 1.6 1.5 1.3 1.3 Measured interlamellar air space width (rrm) 1.5r 0.1(5) 1.3t 0.1(5) 1.2t 0.1(5) 1.5t 0.3(5) 1.5t 0.2(5) r.2 t 0.2(5) '87a. 'Mean valuesfrom Opell, '87b. 'Mean valuesfrom Opell, SPIDER BOOK LUNG SURFACE AREAS 2L3

METHODSAND MATERIALS ment leacheslittle material from tissue.has little Adult females, representing 6 of the 18 de- effect on mass,and coagulateshemolymph, pre- scribed uloborid genera were studied: the orb- venting its loss when a specimen's pedicel is weavers Waitkera waitakeriensis (Chamber- severed.Prior to weighing, specimenswere blot- lain, '46) from northern New Zealand, Uloborus ted and air dried for 30 seconds.The massesof glomosus(Walckenaer, 1841) from Virginia, Oc- eight Miagramnxopes animotus so treated aver- (range tonoba sinensis (Simon, 1880) an introduced aged103 % 95-t05 % ) of their live weights. Asian speciescollected in Virginia, and Philo- Specieswere assignedto one of three groups of increasingtracheal phy- ponella tingena (Chamberlinand Ivie, '36) from development,one of four groups, Panama and ;the irregular-web weaver logenetic and one of three habitatgroups (Fig. M iagrammop es animotus (Chickering,'68)from of increasingmoistureavailability 1). Within limited Hyptiotes Puerto Rico; and the triangle-web weaver Hyp- the context of this study, and Miagrammop es Philoponella Oct onoba tiotes cauatus(Hentz, 1847)from Virginia. These arrd artd are sister taxa, making it inappropriate to assign specieswere identified using the following revi- one member of the pair a more advanced status sions (Muma and Gertsch,'64; Chickering, '68; than the other. One factor analysis of variance Opell,'79;and Yoshida,'80). (ANOVA) and Kruskal-Wallis K-sample rank- The anterior region of the opisthosoma con- ing (K-W) tests were used to determine if tra- taining the book lungs was removed, cleared in cheal type and habitat explained the differences hot l0 (weight/volume) NaOH to remove non- % in relative book lung surface area. chitinous tissue, rinsed in water, and stained to Phylogenetic tracts were evaluated using make the thin book lung leavesvisible. One book ANOVA tests, K-W tests, and transformational lung from in eachspecimen was dissected a drop analysis.The latter approach infers from termi- its isolated leavesflat- of mounting medium and nal (living) taxathe states of hypothetical Ernces- tened under a coverslip. tral taxa and determines if these states exhibit leaves lung The surface areasof the of a book directional changesindicative of an evolutionary were measuredwith a digitizing tablet, summed, trend (Lauder, '81; Huey, '87; Huey and Ben- and this value multiplied by four to obtain a nett,'87). Hypothetical ancestorsare used in an spider's total book lung surface area. The width attempt to neutralizethe unique selectiveforces of the air spaceseparating a leafs lamellae was that have operated on each terminal taxa since photo- measured from compound microscope its origin. I used the following schemeof iterative graphsof a folded leaf (Anderson and Prestwich, '80). averaging to determine the states of hypotheti- The shapeof a species'booklung leaveswas cal ancestorsHr-Hn. reconstructed from outline drawings of one of the three largest leavesof five specimensof each H1: Hyptiotes + Miagrammopes/2. species.A midline was drawn from a leafs tip to Hz: (Ht + Waitkera) + (Hr + Uloborw) the center of a line connecting the corners of its + (Waitkera + Uloborus)/6. base.The perpendicular distance from this mid- He: (Hr + Uloborus) + (Uloborus + Octonoba) line to the leafs left and right margins was mea- + (Hr + Octonoba)/6. sured at 15 equidistant points (Fig. 2). Mean (Uloborus (Octonoba valueswere used to reconstruct leaf shapes. Hs: + Octonoba)+ + Philop.) Two indexesof spider sizewere used for com- + (Uloborus + Philop.)/6. parisons: 1) the product of prosomal length and width and 2) combined prosomal and leg masses. RESULTS The former index accounts for differences in Table 1 presents mean values for '84) the six spe- uloborid prosomal shape (Opell, and the cies studied. Habitat differences do not con- latter for differencesin a spider's feeding and the found the interpretation of these results. Nei- mass of forming eggs in its opisthosoma. For ther the tracheal patterns of the six study species example, in Octonoba sinerxis, prosomal length nor the two indexes of relative book lung surface is poorly correlated with live weight (r:0.60, areaare affectedby habitat (ANOVAP> 0.18). P > 0.09, N : 9) but more strongly correlated Likewise, phylogenetic position has no affect on with combined prosomal and leg mass (r : 0.94, the habitat of a species(ANOVAP < 0.35). P<0.0003,N:9). The three tracheal groups of increasing tra- Specimens of Philoponella tingena were cheal development are characterizedby decreas- stored in ethanol and transferred to buffer one ing book lung surface area. When book lung day before being weighed. All others were fixed surface area is divided by combined prosomal in 3% glutaraldehyde/S% formaldehyde and and leg mass,the mean ratios for tracheal groups stored in sodium cacodvlate buffer. This treat- 1-3 are 2.25, I.64, ffid 1.33 mm2/mg, respec- 2L4 B.D. OPELL

WAITKERA ULOBORUS OCTONOBA PHILOPONELLA

TRACHEAL PATTERN 2 HABITAT 1 PHYLOGENETIC POSITION 3

2.1gmm2 / no olapm2 /pi'2 .1 1O.5/ mg .35mm2/nro o,46pm2 / pri2 1.80 mm2 /mq to.o /mq 0.64 pm2 / pr2 10.4 / mg

1.4gmm2/ ma o.sopm' / sttr' to.o / mg

Fig. 1. Phylogenetic, tracheal, and habitat categoriesof soma,2) smaller tracheaeenter the prosoma and legs,3) stout the genera studied. To the right of H,-H' estimatpd values tracheaeenter the prosoma and legs (Opell'79; '87b). Habitat are given for: 1) book lung area relative to prosomal and leg categories: 1) dry temperate shrubbery and forest under- mass (upper), 2) book lung area relative to the product of brush, 2) moist temperate forests and buildings, and 3) wet '40; prosoma length and width (middle), and 3) the number of tropical forests (Comstock Forster '67; Opell '79,'82, '87a,b: book lung leaves relative to prosomal and leg mass (lower). Peasleeand Peck'83; David Court and Denis Gibbs. Tracheal categories 1) tracheae restricted to the opistho- personal comrnunication). tively. When book lung surfacearea is divided by When all six speciesare considered,the num- the product of prosomal length and width, the ber of leavesper book lung correlates positively mean ratios for t^rachealgroups 1-3 are 0.85, with spider prosomal and leg mass (r: 0.76, 0.53,and 0.45 p.m'/pmz,respectively. These in- P < 0.0001, N : 47) and with total book lung verserelationships betweenboth indexesof book area (r : 0.91,P < 0.0001,N : 47). Therefore, lung surface area and tracheal development are mass-specificindexes (Table 1) were used in significant (ANOVA, K-W P < 0.0001),as are comparisonsof leaf number. A transformational those between orb-weavers with different tra- analysis of the number of leaves per prosomal chealpatterns (tracheal groups 1 and 2;ANOVA, and leg mass shows an average change of only K-W P < 0.0013). S% (FiS. 1). Likewise, ANOVA tests show an Transformational analysesof changesin hypo- insignificant effect (P > 0.51) of the three tra- thetical ancestorsHr-Ho show increasesin both cheal and four phylogenetic categories on the indexesof lung area (Fig. 1). The ratio of lung ratio of book lung leaves to prosomal and leg area to prosomal and leg mass increases by a mass. mean of 209band the ratio of lung areato proso- The shapeof the six species'booklung leaves mal length and width increases by a mean of is similar, although those with greater book lung 24%. These results are supported by compari- areasand more leaveshave narrower leaves(Fig. sons of book lung surface areas acrossphyloge- 2). When the mean value of a leafs 16 width netic groups. The mean ratios of lung area to measurementsis divided by its length, the result- prosomal and leg mass for phylogenetic groups ing index is inversely related to both the number 1-a (Fig. 1) are: 1.19,1.33, 1.92, and 2.25 mm'/ of a specimen's book lung leaves (r: -0.55, mg, respectively.The mean ratios of lung Eueato P < 0.002, N : 30) and to its book lung area the product of prosomal length and width for (r : -0.4I, P.0.024, N : 30). However,this these phylogeneticgroups are: 0.31, 0.45, 0.59, index of leaf width is not correlatedwith a speci- and 0.85 pm' / pm', respectively.The interactions men's combined prosomal and leg mass between phylogenetic position and both indexes (P < 0.11). of lung area are significant (ANOVA, K-W P < 0.0001).When only the four orb-weavers DISCUSSION werecompared, phylogenetic position alsosignif- In the Uloboridae, increasedbook lung area is icantly affected both indexes of book lung sw- achieved by increasesin leaf number and area. facearea (ANOVA, K-W P < 0.0008). Tracheae and book lungs are complementary SPIDER BOOK LUNG SURFACE AREAS 2L5

,r:i.

j

...... i."" i.'.....i....o' 0.26+ 0.03 0.26t 0.02 0.30t 0.03 0.3110.03 0.28+ 0.03 0.31+ 0.03 Philopnella Oclonoba Waitkera Ulobrus Miagrammopes Hyptiotes

Fig. 2. Reconstructed shapes of book lung leaves. The are expressedrelative to their lengths. Below each leaf is its lengths of the leaves have been made equal and their mean mean width-to-length ratio followed by the standard devia- right and left widths at eachreference point along the midline tion ofthis ratio. respiratorymechanisms;when one system is bet- eral factors may contribute to this trend. These ter developedthe other provides lessof a spider's spiders may spin wider cribellar capture threads total oxygen. The inverse relationship between that more securely hold prey and, therefore, re- these two systems suggeststhat their develop- quire less rapid and less vigorous behavior to ment is governed not only by a spider's total subdue prey. They may possessmore aciniform respiratory demands, but also by the specificity spigots, responsible for producing the wrapping of thesedemands. silk used to swath struggling pr€y, and therefore, The tracheae of Hyptiotes cauatus and Mia- be able to thoroughly wrap prey with fewer leg gramnxopesaninxotus extend into the prosoma movements. and legs (Opell, '87a), where they can directly Across spiders of different sizes, book lung meet the oxygen demands imposed by these area is directly related to metabolic rate (Ander- spiders' active web monitoring and manipula- son,'70;Anderson and Prestwich,'82).However, '78; '86; tion tactics (Lubin et al., Lubin, Opell, this measurement does not include the area of '87b).These species'smallerrelative book lung fine tracheoles acrosswhich oxygen diffuses di- areasprobably supply their non-locomotor respi- rectly into tissuesand, therefore, underestimates ratory demands. In contrast, the tracheae of total respiratory surfacearea, particularly of spi- Octonobasinensis andPhiloponella tingena are derswith well developedtracheal systems.There- confined to the opisthosoma,requiring all loco- fore, it is inappropriate to conclude that ul- motor oxygen demands to be met by he- oborids with the greatest book lung areas also molymph-borne oxygen. These spiders' book have the highest metabolic rates. lungs must meet both locomotor and non-loco- On average, Anderson and Prestwich's ('80) motor respiratory demands. Not surprisingly, formula: width of interlamellar air space : 0.90 these orb-weavershave the largest relative book body mass0'2tpredicts to within L0% the mean lung areasof the six speciesstudied. interlamellar separationof the four specieswhose Within the Uloboridae, there is a transforma- adult massesare known (Table 1). Considering tional increase in book lung area. This is seen that the standard deviations of the six species' despite the fact that Hyptiotes and Miagranl- measuredinterlamellar separationsaverag e L2% nxopes have more well developed tracheae to of their me€unvalues, the generality of Anderson accommodatethe active web-monitoring behav- and Prestwich'smodel is supported bythis study. iors their reduced-websdemand. In these gen- Orb-weavers of the families Uloboridae and era, increased behavioral involvement compen- Araneidae have similar relative book lung sur- sates for the reduced material (silk) invested in face areas. The four large araneids studied by capture webs. The inverse may be true in orb- Anderson and Prestwich ('82) lAcanthepeira weaversthat demonstrate an increasingreliance stellata, Argiope aurantia, Mecynogea lemnis- upon book lungs to supply their respiratory cata, and Metazygia wittfeldae; mean mass of needs.In these spiders,oxygen demand appears about 382 mg] had a mean value of 0.54 mm2 of to become less acute in more derived taxa. Sev- book lung area per milligram of live mass. The 2t6 B.D. OPELL uloborid Octonoba sinensis,with a poorly devel- Hentz, N.M. (1847)Descriptions and figures of the Araneides oped tracheal system similar to those found in of the United States.Boston J. Nat. Hist. 5:M3478. Hexter, S.H. (1982) Lungbook microstructure in Tegenaria the four araneid species, has a value of 0.51 sp. Bull. Br. Arachnol. S