Ecology and Behavior of the Giant Wood Spider

Ecology and Behaviorof the Giant Wood SpiderNephila maculata (Fabricius)

in New Guinea

MICHAEL H. ROBINSON

and

BARBARA ROBINSON

SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY • NUMBER 149

SERIAL P U B L I C A T I O N S OF T H E S M I T H S O N I A N INSTITUTION

The emphasis upon publications as a means of diffusing knowledge was expressedby the first Secretary of the Smithsonian Institution. In his formal plan for the Insti-tution, Joseph Henry articulated a program that included the following statement:"It is proposed to publish a series of reports, giving an account of the new discoveriesin science, and of the changes made from year to year in all branches of knowledge."This keynote of basic research has been adhered to over the years in the issuanceof thousands of titles in serial publications under the Smithsonian imprint, com-mencing with Smithsonian Contributions to Knowledge in 1848 and continuing withthe following active series:

Smithsonian Annals of Flight

Smithsonian Contributions to Anthropology

Smithsonian Contributions to Astrophysics

Smithsonian Contributions to Botany

Smithsonian Contributions to the Earth Sciences

Smithsonian Contributions to Paleobiology

Smithsonian Contributions to Zoology

Smithsonian Studies in History and Technology

In these series, the Institution publishes original articles and monographs dealingwith the research and collections of its several museums and offices and of professionalcolleagues at other institutions of learning. These papers report newly acquired facts,synoptic interpretations of data, or original theory in specialized fields. These pub-lications are distributed by mailing lists to libraries, laboratories, and other interestedinstitutions and specialists throughout the world. Individual copies may be obtainedfrom the Smithsonian Institution Press as long as stocks are available.

S. DILLON RIPLEY

SecretarySmithsonian Institution

S M I T H S O N I A N C O N T R I B U T I O N S T O Z O O L O G Y • N U M B E R 1 4 9

jicology and Behaviorof the Giant Wood SpiderNephila maculata (Fabricius)

in New Guinea

Michael H. Robinsonand Barbara Robinson

SMITHSONIAN INSTITUTION PRESS

City of Washington

1973

ABSTRACT

Robinson, Michael H., and Barbara Robinson. Ecology and Behavior of theGiant Wood Spider Nephila maculata (Fabricius) in New Guinea. SmithsonianContributions to Zoology, number 149, 76 pages, 30 figures, 11 tables, 1973.-—Investigations of the seasonal, reproductive, and population ecology of Nephilamaculata are reported in detail. In an investigation of feeding ecology over aone-year period, the discarded remains of the prey caught by a sample populationof ten adult female spiders were collected daily. These remains were identified(where possible) and the accumulated weekly discards from each spider weredried and weighed. Data from this study are analyzed, tabulated, and comparedwith the catches from insect traps located in the study area. The study includedan investigation of web structure, frequency of web renewal, and the number ofkleptoparasites associated with Nephila maculata.

Studies of behavior included courtship, mating, predatory behavior, andresponses to predators, sunlight, and rainfall. Courtship behavior included acomplex pattern of silk deposition by the male on the female, here reported forthe first time for the Araneida. The predatory behavior of N. maculaia is analyzedin terms of behavior units and behavior sequences and is compared with that ofother Nephila species, related species, and that of other araneids studied by theauthors.

The phenological aspects of the study are stressed and discussed.

OFFICIAL PUBLICATION DATE is handstamged jji a limited .number of initial copies and is recordedin the Institution's annual report, Sm^tnsonian Year. SI PRESS NUMBER 4779. SERIES COVERDESICN: The coral Montastrea cavemosa (Linnaeus).

Library of Congress Cataloging in Publication DataRobinson, Michael H.Ecology and behavior of the giant wood spider Nephila maculata (Fabricius) in New Guinea.(Smithsonian contributions to zoology, no. 149)Bibliography: p.1. Nephila maculata—Behavior. 2. Spiders—New Guinea. I. Robinson, Barbara, 1931—

joint author. II. Title. III. Series: Smithsonian Institution. Smithsonian contributions tozoology, no. 149.

QL1.S54 no. 149 [QL458.A65] 591'.08s [595'.44] 75-1905

For sale by the Superintendent of Documents, U.S. Government Printing OfficeWashington, D.C., 20402 - Price $1.45 (paper cover)

Contents

Page

Introduction 1The Species 2Techniques of Study 3The Study Area 7

Natural History and Ecology 7Web Studies 7

Web Location 7Web Structure 8Web Renewal 10Web Strength 12

Reproduction 12Males on Webs of Females 12Egg-laying . 13Hatchling Behavior and Dispersion 14Phenology of Population Increments 15

Prey and Prey Captures 17Taxonomic Range . 17Numbers in Taxa 17Temporal Distribution . 19Weights of Prey Remnants 21Comparison with Argiopc argentata 23Comparison with Insect-Trap Catches 24Rejections of Trapped Insects 27

Kleptoparasites 29Numbers and Nature 29Activities 30Colonization of Host Webs 32General Considerations 32

Energetics 33Phenology 33

Behavior . 34Terminology 34Sexual Behavior 34

Finding the Female 34Courtship 35Copulation 41Discussion 41

Predatory Behavior 44Behavior Units of Nephila maculata 44Behavior Units of Other Species 52Behavior Sequences 54Comparison with Argiope aemula . 59Comparison with Nephila clavipcs 60

iii

Page

Categorization of Araneids by Predatory Behavior 62Behavior of Prey in the Web . . 63Responses to Environmental Factors 64Defensive and Escape Behavior . 66

Discussion 67Phenology .. 67Behavior 69

Literature Cited 70Appendix 1: Activities of Male Nephila maculata during

Courtship Silk-deposition 73Appendix 2: Verbatim Notes on Selected Sequences Given

to Large Acridiids 75

Ecology and Behaviorof the Giant Wood SpiderNephila maculata (Fabricius)

in New Guinea

Michael H. Robinsonand Barbara Robinson

Introduction

Araneid spiders belonging to the genus Nephilaform a conspicuous element of the invertebratefaunas of both the Old and New World tropics.In addition, they extend into north and south tem-perate regions. The adult females of most speciesare large and build strong webs of golden silk thatare often of considerable size. McKeown (1963)cites examples of Nephila species catching and con-suming small birds and also mentions that thenatives of the South Pacific use the webs for catch-ing fish. An interesting account of the use ofNephila silk for fishing appears in a work of fictionby Olaf Ruhen, a New Zealander with wide exper-ience of the Pacific region. He states (1969:64)that the web is matted into a wad the size of ahuman thumb and flown over the surface of thesea suspended from a kite. There it is seized bygarfish which are unable to disentangle their teethfrom the silk and are caught. Surprisingly therehave been few studies of the biology of Nephilaspecies and attention has largely centered on thetaxonomy and zoogeography of this group. Fischer(1910a, b, c), Hingston (1922a, b, c; 1923), andThakur and Tembe (1956) have published short

Michael H. Robinson and Barbara Robinson, SmithsonianTropical Research Institute, Post Office Box 2072, Balboa,Canal Zone, Panama.

accounts of the natural history of Nephila maculata(Fabricius) in India. Bonnet (1929), Coleman(1948), Gerhardt (1930), Jager (1960), and Mc-Keown (1963) have written on various aspects ofthe behavior or ecology of other Nephila species.Peters (1954, 1955) studied the web construction ofNephila clavipes (L.).

Studies of the predatory behavior of Nephilaclavipes in Panama (Robinson and Mirick, 1971)led to the conclusion that this species was primitive(Robinson, Mirick, and Turner, 1969). At thisstage we thought that a more comprehensive studyof a Nephila species would be worthwhile andrewarding. We were able to make some observa-tions on Nephila and related Nephilengys speciesin West Africa during February and March 1970,and further observations on Nephila and Nephi-lengys in Madagascar and on Nephila and Herenniain India (March/April 1970). In New Guinea wefound that Nephila maculata was abundant in theWau valley (Morobe District) where we were tospend one year at the Wau Ecology Institute. Thespiders living in the arboretum of this institutewere protected from disturbance and we decided tomake a long-term field and laboratory study of thisspecies.

This paper is based on our studies in NewGuinea but also includes comparative data from

SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

the observations carried out in West Africa, Mad-agascar, and India, and unpublished data from thestudy carried out by Robinson and Mirick in Pan-ama. The New Guinea studies were basicallyecological and behavioral. Ecological studies ineluded observations on web structure and habitat,sampling of potential and actual prey, phenologicalstudies of reproduction and predation, investigationof population numbers of kleptoparasites, andrecording population fluctuations of adult and im-mature Nephila. The behavioral studies were main-ly concentrated on predatory behavior but alsoincluded studies of mating behavior and responsesto sunlight and rainfall. Some effects of behaviorare dealt with in the ecological section of thispaper because the underlying behavior patternswere not directly studied. This was the case withrespect to prey-rejection, web-renewal and egg-laying. Some studies that are referred to in thispaper, e.g., those on web adhesiveness and popula-tion fluctuations in Wau araneids, were carriedout in conjunction with our colleague Dr. Y. D.Lubin and are to be published separately. Thestudy of five species of Argiope that we carried outin New Guinea is referred to in the comparativesection on predatory behavior, but it is to be pub-lished in detail elsewhere.

ACKNOWLEDGMENTS.—We thank Dr. Y. D. Lubinfor her assistance during this study. She providedboth ideas and practical help. We are extremelygrateful to the staffs of the following institutionswho helped our work in Africa and India: Depart-ment of Zoology, University of Ghana, Legon;O.R.S.T.O.M. and the Institute de RechercheTropicale in Ivory Coast; O.R.S.T.O.M. in Mada-gascar; Bombay Natural History Society; Univer-sity of Delhi; Hindu University of Varanasi;Forestry Department of Assam; and, particularly,Dr. T. N. Ananthakrishnan in Madras.

In New Guinea we received much help from thestaff of the Forestry School at Bulolo, The LaeBotanic Gardens, and the entomologists of theDivision of Entomology laboratories at Bulolo. Ournative helpers, Rennie, Lik-lik Boy, and Polinowere superb naturalists and cheerful assistantsthroughout the study.

Our colleagues at the Smithsonian Tropical Re-search Institute, have, as always, contributed muchto the development of the ideas presented in this

paper. In particular we are grateful to Dr. A. S.Rand and Dr. H. Wolda for advice and help withthe analysis of data.

THE SPECIES

Fr. Chrysanthus identified the Wau material asNephila maculata. The species has reportedly awide distribution in the Old World tropics andsubtropics (Roewer, 1942; Bonnet, 1958). Roewer(1942:929) gives the distribution as "Ceylon, Indienbis China u. Australien," while Bonnet (1958:3077) adds "Afrique occidentale et Afrique aus-trale" and "PMexique." The records of the speciesfrom New Guinea have been detailed by Chrysan-thus (1959, 1960, 1971). The species occurs in allparts of the Wau valley where trees or bushes persistand we have recorded its presence to a height ofapproximately 2000 meters on the slopes of MountKaindi. Densities are often high in coffee planta-tions, where the species builds its webs in theflight paths of insects between the rows of plants.

Some details of the appearance of adult malesand females are evident from the photographs illus-trating this paper (Figures 17, 18, 19). The adultfemale varies in size from just below 40 mm inbody length to over 50 mm; most of this variationis due to variations in the length of the opistho-soma. Legs i are the longest, legs iv are usuallylonger than legs n and legs m are the shortest ataround half the length of legs i. See Table 1 forspecimen measurements, note that there is oftenasymmetry in leg lengths. We have records of theweights of adult females ranging from 2-4.25 gramsand think that the variations in adult weight prob-ably extend above and below this range (takenfrom a sample of 10). The adult female is dull anddark in coloration contrasting markedly with someother common Nephila species in this respect.Yaginuma (1969, pi. 29) gives a color illustrationof the Japanese form. The Wau specimens differfrom this in that the opisthosoma is completelyunmarked dorsally. The Japanese form evidentlyhas pale yellow lines and spots on this region.Dorsal markings of the opisthosoma are present onimmatures at Wau. They are not present on thepenultimate instar. Hingston's (1922a) descriptionof Nephila maculata in India refers to strikingcoloration of the dorsal surface of the abdomen.

NUMBER 149

TABLE 1.—Sizes of adult female Nephila maculata(- = leg missing.)

Spider12345678910Average

Wet Opistho-weight Length Prosoma soma

(gm) (mm) (mm) (mm)

Leg lengths (mm)II III IV

R R4.252.7252.0554.1343.5753.3873.802.933503.753.38

4140374048464740445043.3

1112101112131210111311.5

3028272936333530333731.8

828583

84808179858682.78

8086818583828282878483.2

6872677067717067716869.1

6671686968

6958727067.9

4140424041394138404140.3

39

403937403842394039.3

6971687068697069

69

687469566971

747168.9

Juveniles up to at least half the adult size (Figure1) possess the dense hairy "gaiters" on the distalportions of the tibiae of legs i, n, and iv that are

FIGURE 1.—Juvenile Nephila maculata showing "gaiters" onlegs (arrowed) that are lost in later stages. Length ca 25 mm.

characteristic of all stages of Nephila clavipes andseveral other Nephila species. The females of laterstages lose the gaiters as Thakur and Tembe (1956)have also noted. These authors include a figure ofan adult female showing dorsal lineation and spot-ting on the opisthosoma (1956, fig. 1). The opistho-soma of Wau adults is a dark slatey color (almostblack) with an overlying "bloom," or "patina," ofgold; the latter appears to wear off with time. Wefound a number of adult females in the Wau area,which had conspicuous pale yellowish brown legsin marked contrast to the black legs of the "nor-mal" form. One of these is illustrated in Figure 2,and a specimen is deposited with Fr. Chrysanthus.This form is a distinct one, and not merely a prod-uct of slow deposition of the appropriate pigmen-tation.

Adult males are very small, ca. 4-6 mm in bodylength and transluscent red to reddish orange incoloration, often with black markings on the tarsi.

TECHNIQUES OF STUDY

During the period May 1970 to May 1971 wecarried out daily observations on a sample popu-lation of adult female spiders present in the studyarea (described later); and from June 1970-April1971 we carried out the sampling of flying insectsin the same area. During daily observations wenoted the state of the web and its position; thepresence or absence of males and kleptoparasites,the presence or absence of prey, and evidence ofegg-laying as reflected in sudden reductions in thegrossness of the female opisthosoma and/or the


F K U U 2.—Revene color form of adult Nephila maculata in typical predatory position at hub.

presence of nearby egg-sacs. This recording involvedvisiting ten marked adult spiders daily. Prey trapswere placed under the webs of these spiders anddiscarded or rejected prey were collected daily forfurther examination and analysis. In addition tothis daily routine we carried out (with Dr. Y. D.Lubin) weekly counts of all the visible araneids inthree 100x2 meter transects, one of which borderedthe study area. These counts included counts ofadult and immature Nephila maculata. We alsocensused the kleptoparasites present in Nephilawebs outside the study area but within a 500 meterrange of its center.

Behavioral observations were carried out on adultspiders living in the study area but not formingpart of our sample population. The investigationson predatory behavior utilized similar techniquesto those employed by the senior author and co-workers in earlier studies and will not be describedin detail here. New techniques used for the firsttime in this study are described in the appropriatesection of the main text. Extensive use was madeof filming and film analysis, particularly in thestudy of mating behavior. We used Super 8 inmmovie film for this purpose. This allowed us to useconsiderably more portable equipment than was

NUMBER 149

possible in previous studies where we used 16 mmfilm. For filming behavior sequences we used aCanon 518 camera equipped with the manufactur-er's close-up lens. This enabled us to fill the framewith a portion of the body of the adult Nephilafemale, and, incidentally to use the viewing systemof the camera to watch events that were not clearlyvisible to the unaided eye. Using zoom optics forthis purpose provides an admirable technique forstudying the behavior of small animals from a con-venient distance (Robinson and Robinson, 1972a).All the black and white photographs were takenwith a 6 cm square single-lens reflex camera equip-ped with close-up attachments. The photography ofspider's webs is difficult (see Langer and Eberhard,1969) but can be facilitated by spraying the webwith matt-white aerosol paint and then interposinga black background between the web and surround-ing vegetation. All the accompanying photographsof web structure were taken by utilizing this tech-nique.

The rainfall data that we include in the pheno-logical analyses are derived from the records of theWau Ecology Institute and New Guinea Gold fie IdsCompany.

PREY TRAPS.—Our observations on the prey of

Argiope argentata (Fabricius) in Panama werecarried out by visiting ten webs at two-hourly inter-vals throughout the day and recording details ofthe prey present in the web. This method is ex-tremely tedious and time-consuming and does notpermit a determination of weights of prey exceptby extrapolation from weighed samples. Spidersare suctorial feeders and discard "trash packages"of prey remains after feeding. We decided to col-lect these packages for weighing and identification.To do this we placed meter square pieces of finenylon netting under the webs of the ten "sample"spiders and collected the remains each morning. Toprevent ants from removing prey debris we coatedthe prey trap attachment strings (one at each cor-ner) with tanglefoot. The method proved entirelysatisfactory. The traps caught the remains ofquite small prey (as small as 5 mm insects), theconspicuous remains of large prey, and the deadbodies of rejected prey items. The latter was a"bonus" item resulting from the fact that N. macu-la t a rejects many items of obnoxious prey aftera preliminary bite. (If heavy rain occurred during

the 24-hour interval between emptying the traps, itmay have washed some very small fragments ofprey through the mesh, but we think that thisfactor was a very small source of error in ourresults.)

The material from each trap was placed in aseparate vial each day, dried in a drying oven, andthe seven samples from each spider were examinedunder a binocular dissecting microscope at the endof each week. This process allowed us to make arecord of prey caught, on each day of the week.We also obtained a weekly total for the dry weightof prey remains discarded by each spider. (Wedecided that weighing the wet remains would notgive any useful information because of wide varia-tions in water content due to climatic conditions.)Sorting the material under a binocular microscopewas not easy since the masticatory process resultsin the extreme comminution of those insects thathave thin cuticles. It proved possible, however, toidentify most trash bundles to order and often tolower taxonomic divisions. The jaws and jumpinglegs of orthopterans served to identify the frag-mented remains of those insects and the color ofthe cuticular fragments was a further guide toseparation. Lepidopteran remains were character-ized by the presence of wing scales. The elytra ofeven the smallest coleopterans survived the com-minution of other parts. Compound trash parcelscould often be separated into their multiple con-stituents by counting elytra, jaws, and other hardparts, and the number of prey items thereby cal-culated. Occasional forceps marked the presenceof dermapterans and we were able to sort out somedipterans and hymenopterans on the basis of theirmembranous wings. Catches from the nearby insecttraps were useful in giving an indication of thepotential presence of some groups of flying insectsthat were periodically in abundance. These datahelped in the determination of dubious prey re-mains. Larger insects that were not caught by theprey traps (e.g., melolonthids and sphingids) wereseen at our Mercury vapor (Mv.) insect light andvery abundant flights were noted for comparisonwith the prey trap data.

The spiders were marked with Humbrol model-makers enamel paint using a color/position code toidentify individuals. The paint peeled off the sur-face of the opisthosoma after about one month so


it was necessary to re-mark the spiders every threeweeks. We usually had twenty marked spiders inthe study area; the ten sample spiders and ten "re-serves." The latter were used when one of thesample spiders died, disappeared or moved into aninaccessible position. Movement to an inaccessibleposition occurred very rarely since we were able toarrange the prey traps at most web sites (even whenthis involved some removal of surrounding vegeta-tion.) The fact that the spider builds its web acrossgaps in vegetation, in flight paths, normally en-sured that there was a fairly clear space beneaththe web in which to erect the prey trap.

We recovered the dead bodies of some spidersfrom the prey traps but lost most of our samplespecimens by "disappearance." Our very skilled andsharp-eyed New Cuinean assistants usually foundspiders that had merely moved long distances fromthe original web sites but many spiders disappearedwithout trace at some stage in the prey census. Wediscovered at an early stage in the study that veryfat spiders would disappear for one or two days andthen return, much depleted in size, to the same weblocation. These had produced egg-masses. As aconsequence of this observation we decided not toreplace missing sample spiders until they had beenabsent for three morning censuses. In time wenoticed that egg-laying absences were often precededby nonrenewal of the web and a fall off in catchesfor the days immediately prior to the absence, andthis factor helped in the interpretation of disap-pearances. Spider movements largely occurred atnight, and were noted the following morning, whenthe trap could be relocated. Thus there was verylittle disturbance of the data collection by move-ments from web site to web site. Individuals re-mained under observation for several months andprovided data on minimum periods of adultlongevity.

INSECT TRAPS.—To sample the availability ofpotential prey we used two techniques. We placedthree window-pane traps at sites in the study areathat were typical of web locations and collected theinsects from these, each morning, over most of thestudy period. We also used two sticky traps for ashorter period (necessitated by the late arrivalof our supplies of tangle-foot); these traps were alsoserviced daily.

We chose window-pane traps as a sampling de-

FIGURE 3.—Window-pane trap in study area.

vice since we considered these to be less conspicuousto flying insects than the opaque mass of a stickytrap. They, in fact, trapped insects not caught bysticky traps but failed to catch some insects knownto be in flight (from our observations of the spider'sprey and Mv. light catches.) These insects, includ-ing large beetles and moths, were not caught bythe sticky traps either. We have discussed elsewhere(Robinson and Robinson, 1970a:356) the problemsinvolved in sampling the prey available to largeorb-weavers and do not consider that these haveyet been solved.

The window-pane traps were constructed ofpieces of 24-ounce clear sheet grass 24 X 24 inchesin size (about 61 x 61 cm). This was mounted ina light wooden frame and its lower edge dippedinto a 3 cm tleep trough of water. The watercontained a wetting agent and a small quantity ofphenol. Figure 3 shows a trap in situ. We mountedthe traps on wooden legs with their centers about1 meter above ground level.

NUMBER 149

The sticky traps were made by smearing bothsides of a meter square of nylon mosquito nettingwith tanglefoot (I.C.I. Osticon) and mounting thenet on a light wooden frame. The traps wereerected with their bottom edges about 75 cm aboveground level.

Insects from the window-pane traps were pre-served in 70 percent ethanol and those from thesticky traps were collected in gasoline and trans-ferred to alcohol after the tanglefoot had beendissolved. The daily catches were examined month-ly, sorted taxonomically, and graded into sizeranges.

THE STUDY ARKA

The approximate boundaries of the study areaare shown on Figure 4. The area was originally acoffee plantation but, over a period of years, hasbeen converted into an arboretum by the activitiesof the Wau Ecology Institute (formerly BishopMuseum Field Station). Trees from Mount Kaindiand other areas in the Wau-Bulolo region havebeen planted and have reached a considerableheight above the remaining rows of coffee. Thelatter is picked periodically but was not pruned

FK;I!RK 4—I'lan of study area (hatched). (Black circles = largetrees, b — bamboo thicket, s^sticky trap, w = window-panetrap. Stiplcd area is transect 1.)

during our study year. Some of the tree and shrubspecies occurring in the area are as follows:

Rubiaceae: Gardenia species, Neoxandia species, Psychotriaspecies, PaveUa species

Euphorbiaccac: Phyllanthus species, Euphorbia pulcherrima,Macaranga species

Meliaceae: Disoxylon species, Toona sureniMoraceae: Ficus calopilina, Ficus dammaropsisSapindaceae: Dyclioneura species, Allophyllus speciesElaeocarpaceac: Elaeocarpus sphericus, Elaeocarpus dolicho-

stylusAnonaccae: Anona muricataSolanaceae: Datura CandidaAraucariaccae: Araucaria klinkiiSterculiaccac: Sterculia speciesMyrtaccac: Eugenia speciesLauraccac: Litsia speciesMelostomaceac: Astronia speciesRhamnaceac: Alphitonia incavaVerbenaceae: Callicarpa longifolia

Some idea of the height of the background vege-tation can be gained from Figure 3. The areasupports a diverse fauna including numerous arthro-pods that are predators of arthropods. Among theseare trap-building spiders of the families Uloboridae,Amaurobiidae, Pholcidae, Araneidae, Tetragnathi-dae, Agelinidae, Linyphiidae, and Theridiidae.Araneids included three species of Gasteracantha,two of Argiope, two of Leucaxige, one of Cyrto-phora, one of Arachnura, several araneus-like noc-turnal forms, Herennia ornatissima, Cyclosainsulana, and the, as yet, unidentified "New Guinealadder-web spider" (Robinson and Robinson,1972b). The theridiids include solitary, colonial,and kleptoparasitic forms. The trap-building spi-ders exist alongside a large hunting spider popula-tion and some sedentary raptors (dinopids). Inaddition there is a diverse assemblage of predatoryinsects including mantids and reduviids.

Natural History and Ecology

WEB STUDIES

Web Location

Webs of adult female spiders are almost alwayslocated with the bottom edge (lower foundationthread) above the herb layer, even when thisattains a height of more than 1 meter. A proportionof webs may be built with the prey capture area

8 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

partially overlapped by nearby vegetation but inmost cases webs span natural gaps in woody vege-tation. Hingston (1922a:643) notes that IndianN. maculata often builds its web where "it finds atunnel in the tangled growth where the insectspass to and fro." We found webs located in treesat heights of more than 5 meters above the groundand even sited on telephone and overhead powercables. Webs of males and immatures were dis-covered closer to ground level and sometimes withinthe herb layer. Back-to-back, web building, as notedin the case of N. clavipes (Shear, 1970), was uncom-mon in the Wau area.

Web Structure

In common with the webs of those other Nephilaspecies that we have seen, the webs of N. maculataare markedly assymmetric in that the hub is alwayslocated in the upper third of the web. Peters (1954,1955) has given details of the structure of the N.clavipes adult web, which is frequently incompleteabove the hub (see also Kaston and Kaston, 1953).The web of N. maculata differs in being almostalways complete above the hub, i.e., with the preycapture area intact in this region. Some sampledimensions of webs made by adult females (N = 10)convey the structural aspects of the web of N.maculata:

Range0-10

55-90

69-110

Viscid spirals above hubViscid spirals below hubRadii (counted halfway between

hub and periphery)

Compared with webs of adult males (N=5), thefemales' webs are appreciably larger (measurementsin cm):

Average5.7

72.5

86.9

Female Male

WidthHeight

Range Average Range Average62-100 84.1 7.8-11 9.479-116 96.2 10.4-13.1 11.7

In our sample of 3237 webs, those with barrierconstructions occurred in 640 instances (19.8% oftotal sample), of which 625 had dorsal barriersonly, and 15 had both dorsal and ventral barriers,but none had only ventral barriers. We note, there-fore, that barrier webs occur with maximum fre-quency below the orb, i.e., dorsal to the spider. Thestructural spiral is left in during construction of

the viscid spiral but since four or five viscid ele-ments are not interspersed between successiveelements of the temporary spiral, it does not pre-sent the "music-ruled" effect of the N. clavipes web.The above web dimensions show that the webs areoften more nearly circular than those of N. clavipes(i.e., with their greatest width not differing greatlyfrom their greatest height).

As in all the Nephila webs that we know of, thelower radii are often branched (see Kaston, 1964,for comments on evolutionary significance). Thehub is entire and surrounded by a spiral strengthen-ing zone. The hub region often appears to bebraced by a thread leading from the hub silk tosome nearby piece of vegetation. The viscid spiralis characteristically golden in color. Hingston(1922a:648, 1922b:912,1922c:918) gives details of thestructure and process of construction of the Nephilamaculata web. He notes that the radii are branchedand numerous (more than 100), that the temporaryspiral is left permanently in place and that a bar-rier web is built above the main sheet. He alsoobserved a marked asymmetry in the placement ofthe hub and quotes an example (1922c:918) wherea web had 130 turns of the viscid spiral below thehub and only 3 above.

The webs of males were found sporadically inthe study area and we have seen mature males witha web. They were all small and more like thenormal, almost symmetrical orbs of Araneus diade-matus.

The webs of immature females proved to be ofgreat interest for two reasons. First, a proportionwere found that had linear stabilimenta (Figure 5).These structures consist of a ribbon of silk placedabove and below the hub and consist of multi-strand silk, laid down between adjacent radii in azig-zag line. We never saw any stabilimenta in thewebs of mature females. As far as we are aware thisis the first and only record of stabilimentum build-ing by a Nephila species. We saw 14 such structuresin the course of examining hundreds of webs over aone-year period. Although we never saw the spiderconstruct a stabilimentum it is obvious that it isnot an accidental structure resulting from someanomalous aspect of web-building, but is addedalter the completion of the viscid spiral. It istempting to regard this aspect of N. maculata's webbuilding behavior as vestigial. Elsewhere (Robinson

NUMBER 149

FIGURE 5.—Juvenile Nephila maculata with stabilimentum.


and Robinson, 1970b) we have reviewed the varioustheories about the function of stabilimenta andconcluded that in some cases, at least, a camouflagesensu lato function, is improbable. This wouldcertainly seem to be the case with the rarely builtstabilimentum of N. maculala. (Nephila species doapparently construct some camouflage devices; bothN. maculata and N. clavipes deliberately hang plantmaterial and discarded prey debris in their barrierwebs above the hub. Nephila plumipes (Latreille)often hangs the corpses of prey in a line above thehub.)

The second interesting feature of some juvenilewebs is the form of the barrier webs. Note in Fig-ure 6 that the spider is standing on the main orband that there is a netlike large barrier web dorsalto the spider and a smaller, incomplete, similarbarrier web below the main orb. The structure ofthese barrier webs, an orb web structure with radiiand a structural spiral, has not been previouslydescribed. The hub of the barrier orb is drawnaway from the functional orb by an attachment lineso that the effect is one of a conical dome aboveand below the web proper. This highly organizedpair of structures is replaced at later stages by theapparently disorganized maze of lines that makeup the barrier web (s) of the mature spider. As inthe case of the stabilimentum we were not fortunateenough to see such barrier webs actually beingbuilt. It is, however, clear that they are not adven-titious structures built by other spiders. Thegreater development of the structure dorsal to thespider supports the suggestion that the barrier webmay protect the spider against its predators, in thiscase it would clearly prevent, or delay, attacks byspider-hunting wasps and parasitic dipterans. (Anumber of explanations of the function of barrierwebs are current among arachnologists but few havebeen published. Hingston (1922c) suggests that itprevents the escape of large prey from the main orb.He reasons that large insects may tear themselvesfree of the viscid spiral but that they are thenlikely to strike the barrier web and be driven backinto the main orb. This theory accounts for somefeatures of the structure of barrier webs but nottheir preponderance below the orb, i.e., above thespider's back.)

The plane of the N. maculata orb web seldomappears to be perpendicular but is usually inclined

to a greater or lesser extent. We did not attemptto measure angles of inclination but believe thatthe majority of webs fell within the range of 5°-30°deviation from the perpendicular. The spider as-sumes its predatory position on the undersurfaceof the web. The slope must facilitate transportationof prey on silk (the spider walks on the undersur-face with the prey hanging away from the web, inthe lower and larger portion).

Web Renewal

Our morning census revealed that webs are re-newed, entirely or in part, with greater frequencythan we anticipated. It was easy to distinguish re-newed parts because the edges of the renewed areaare conspicuous. In addition, the new viscid ele-ment in the renewed area looks fresh by contrastwith the old element, in which the viscous dropletsare often irregular and patches of adhering detritusare usually present.

Web renewal data from the daily census of 10adult female Nephila maculata are as follows:

Days female absent or on scaffold only: 413Total possible number of web days (3650-413) :3237A total of 2388 webs were lenewed or repaired during the

3237 days (each web averaging 1.4 days duration)

Number of websrenewed or repaired

1794435105

3113532

Total: 2388

After

24 hours48 hours3 days

4 days5 days6 days7 days9 days

Percentage of websrenewed or repaired

daily 75.1every 2 days: 185every 3 days: 4.3

within 3 days: 97.6

Nephila maculata cuts away areas of web duringheavy rainfall and renews them later, without com-pletely rebuilding the web. The spider cutsthrough the radii, close to the hub, on one side ofthe orb and an area of web collapses. (This be-havior can be induced by spraying a web with waterand may be a means of preventing the collapse ofthe entire web under the effect of a heavy load ofwater droplets.) When heavy rainfall occurs at thestage when the spider would be completing a new

NUMBER H<>

FIGURE 6.—Web of juvenile Sephila maculata showing highly organized barrier webs, spider isvisible on main orb (left center) , barrier web below spider (extreme left) is much smaller thanthat above spider (center to right, conical).

12 SMITHSONIAN COM RIBUTIONS TO ZOOLOGY

web, by laying down the viscid spiral, this stagemay be deferred until rainfall ceases. Other stagesup to the addition of the viscid spiral appear to becarried out during rainfall. Viscid spirals addedafter long periods of intermittent rainfall are oftenuntidy and incomplete.

As mentioned above we carried out, in collabo-ration with Dr. Y. D. Lubin, a number of experi-ments on variations in web adhesiveness with time.We used a modification of the technique used byEisner, Alsop, and Ettershank (1904) and tested theadhesiveness of standard sized lengths of N. macu-lata viscid spiral, at intervals, over a period ofseveral days. The viscid spiral, when protected fromrain, maintained a high level of adhesiveness formore than two days after it had been produced(Robinson, Robinson, and Lubin, in prep.). This

suggests that chemical deterioration of the adhesiveis not the factor necessitating the recorded frequen-cy of web renewal.

Web Strength

Dr. Y. D. Lubin (in litt.) has compared thestrengths of elements from the webs of Ar. maculatawith those of Cyrtophora moluccensis. The results

give no absolute values but show that the Cyrto-phora silk is much stronger than the Xcphila silk.The latter, however, is strong enough to retain in-sects in excess of 2 grams (large scarabaeoid beetles)until the spider can subdue them.

REPKOIH'CTION

Males on Webs of Females

Figure 7 shows graphically the number of malesrecorded on the ten study webs during the year andalso, for the second half of the study period, thenumber of males recorded actually on the body ofthe female during the morning census. We havealso plotted the numbers of males recorded on theweekly census of transect 1 (p. 4) during thesame period. Males are present on the webs offemales throughout the year and the numberincreases markedly from week 36 (commencing 20January 1971) and remains high until the end ofthe study period. The peaks in the number ofmales seen on the body of the female coincidewith the peaks in the totals of males recorded on afemale's web, during the period of high numbers.The greatest number of males recorded on transect

FICIRE 7—Weekly distribution of number of males found on webs of females in study areaand number of males seen on females themselves. Histograms show numbers of males censusedin nearby transect.

NUMBER 149 13

1 also occurs during the period from week 36 to 52.Figure 8 shows the numbers of adult female andimmature Ar. maculata recorded on transect 1, andit is interesting to note that the large populationof males on the study area occurs during a periodwhen the number of immature spiders is high onthe nearby transect (Figure 4).

Several males may attend a female at any onetime. Although we did not mark individuals wehave observed conspicuous individuals (naturallymarked by the loss of legs, or distinctive coloration)to be present on the same web for several clays.When not actually on the body of the female themales take up positions on bridge and framethreads, on elements of the barrier web (s), withinthe upper part of the prey capture area, and at thehul). Males were also observed on the webs of im-mature females, often when these must have beenat least two or more molts away from maturity.We have one record of a male mating with a femaleshortly after she had ecdysed, at a stage, in fact,when she was still hanging from her cast exoskele-ton.

Egg-laying

PERIODICITY.—During the year of study, the 10

adult female Nephila maculata produced an esti-mated total of 89 egg-sacs, which on a monthly basiswere counted as follows: May, 4; June, 9; July, 7;August, 8; September, 9; October, 9; November, 8;December, 6; January, 7; February, 9; March, 6;April, 7. These figures are only for cases of whatwe regard as certain instances of egg-laying. (Wescored egg-laying when we found a new egg-sac nearto a marked female that had been absent from aweb site for 1 to 2 days or when a female had beenabsent for this period and returned considerablyreduced in bulk although no egg-sac was located.Because of their camouflage, egg sacs are not alwayseasy to locale.) Spiders that simply disappearedfrom our sample may have laid eggs but are notscored. The estimate must, therefore, be regardedas conservative.

There was a strong tendency for female spidersto stop web replacement and repair some days priorto egg-laying. There was also a tendency for themto stop feeding in these circumstances. Althoughthese behaviors are not invariable they are apparentin the majority of cases as Figure 9 shows.

The old webs occupied by females on the pointof egg-laying were often functional, and the fall-off in prey capture was, therefore, in many cases,

50 MALES,PER WEEK. STUDY AREA

FIGURE 8.—Weekly distribution of numbers of adult and immature Sephila maculata on transect1. Horizontal black bar shows weeks when males were present.


number of

old wobi

1 2 3 4 5 6 7 8 9 10 11 12 13 14

-spiders obsont

FIGURE 9 Number of prey caught per day and numbers of new webs per day during the twoweeks prior to an egg-laying excursion. Based on data from 50 periods prior to egg-laying.

due to the spider not responding to potential preythat struck the web at this time.

All the egg-sacs that were found were similarin structure and golden green in color. Most of theegg-sacs were constructed under the shelter ofleaves in the upper branches of trees and shrubsnear to the parent female's web site. There, becauseof their coloration, they are quite well camouflaged.

DEVELOPMENT PERIOD.—Unfortunately we failedto obtain data on the development period of eggsin situ, and eggs that we brought to Europe to studyunder laboratory conditions failed to hatch.

Hatchling Behavior and Dispersion

After hatching the spiderlings remain in roughlyspherical aggregations around the egg-sac for up tonine days. During this time they respond to aircurrents by chopping on silk lines, but climb back

into the aggregation after the disturbance hasceased. Figure 10 shows an aggregation of spider-lings two days after emergence from the egg-sac.Presence of exuviae in some masses suggests that thespiders may molt at least once before dispersing.Lowry (in McKeown, 1963:150-151) described hatch-ing and the subsequent behavior of the hatchlingsin the case of N. plumipes, reporting that thespiderlings could remain within the egg-sac for aconsiderable period before emerging to form a loosecluster around it. She concludes that emergencemay be conditional on climatic conditions. Thakurand Tembe (1956:331-332) give data on the shapeof the egg-sacs of N. maculata near Bombay, India,and speak of the cocoon being placed in "a well-concealed place." Fischer (1910a) never found theegg-sacs of this species.

We did not see the process of dispersion but ona number of occasions noted the overnight disap-

NUMBER 149

FIGURE 10 An aggregation of newly emerged Sephila maculata spiderlings beneath a leaf.

pearance of clusters of spiderlings. Although wethink it probable that they may disperse by balloon-ing on silk threads, we have not seen aerial dis-persals of young in the tropics. Bristowe (1939:187)has commented that "it is probable that the youngof most species disperse themselves by aerial meth-ods in all countries where a sudden change inground temperature occurs in the morning, butmore observations are needed in the tropics beforewe can be certain that this means of dispersal is asimportant there as in temperate regions." Ourobservations at Wau and in Panama suggest thateggs are hatching at intervals over most of theyear and this fact probably explains the absenceof conspicuous massed migrations of ballooningspiderlings in these regions. Such migrations aretypically autumnal in the temperate zones (Bris-towe, 1939:182-201, for discussion and bibliogra-phy). We have certainly recorded the presence of

very young A', maculata, with webs, at distancesof over 50 meters from the nearest known egg-cluster. It seems a logical assumption that thesespiderlings arrived at such sites by aerial dispersion.

The group of freshly emerged spiderings of Ar.maculata shown in Figure 10 has produced a clearlyvisible complex maze of threads. Such clumps maycontain more than 1000 spiderlings. Mortality inthe early stages of development must be high.

Phenology of Population Increments

Although we did not monitor the presence ofearly stage spiderlings in the study area—an extra-ordinarily difficult habitat in which to count ani-mals less than 3 mm long—we did census theadjacent line transect (see page 4 and Figure -1).Here we recorded the presence of all trap-buildingspiders, scoring adult and immature spiders sep-


arately. It was not possible to measure the spidersduring the census so that records of immatures,refer to obviously immature spiderlings and includeseveral instars. All large increments of immatureswere found to be due to early-stage spiderlings.Figure 8 plots the outbreaks of immature Ar. macu-lata at this transect. There is a fairly regular incre-ment of spiderlings during most weeks of thecensus until week 25 (8 November 1970) when thenumber exceeds 15 for the first time. Thereafterit exceeds 15 until week 50 (2 May 1971). Duringthe period of this higher density of immatures thenumber occurring at the weekly census exceeded50 on 18 occasions. We think that Kcphila spider-lings are probably emerging over most of the yearin the study area and that the increments in themarginal transect area support this view. Thelarger increments in the wetter months suggestthat conditions may then be favorable for eitherhatching, or the survival of the spiderlings, orboth. The occurrence of peak populations ofmature males during this period supports the viewthat survival rates may be higher at this time, atleast for spiderlings up to the size of adult males.This matter is discussed later in an overall consid-eration of the phenology of the species at Wau. Thedata on egg-laying, on the presence of males infemale's webs, and on the occurrence of immatures

at the transect all suggest that reproduction in N.maculata at Wau, occurs throughout the year. NearBombay, India, where Thaktir and Tembe (1956)carried out their study the climate is much moremarkedly seasonal than at Wau. Figure 11 givesclimographs for Bombay, and for Bulolo, NewGuinea, and monthly rainfall data for Wau. Bulolois the nearest locality from which we have reliabletemperature data. Near Bombay the N. maculatapopulation is markedly seasonal. Thakur andTembe (1956:330) report that the young appearat the "end of August and early in September.Their bodies then measure 1 cm." These are ob-viously not newly emerged spiderlings which aresmaller than 3 mm in length. Thakur and Tembe(1956:331) also state that the females become

gravid "during Oc tober and November" after whicheggs are laid. The spiders become very rare by theend of December, almost absent in January and"no specimens at all were observed near Bombayduring other months of the year" (Thakur andTembe, 1956:330).

Allowing an extra month for the stages occurringprior to the 1 cm young seen in August, it wouldappear from this report that the spider is presentfor, at the maximum, seven months of the year andthat five months, at least, are spent as eggs or inthe egg-sacs. Reference to the climographs shows

* » MatM* rainfall

-I" -0-° - J 4 0 _L50

20

FIGURE II Above: Climograph for Bombay, India, and Btilolo, New Guinea; below, histograms

showing rainfall at study site during the study period.

NUMBER 149 1 /

that the spiders reappear about a month after thereturn of waim wet conditions (June) and becomerare during onset of dry colder conditions (Decem-ber/January).

PREY AND PREY CAPTURES

Taxonomic Range

Little is known of the taxonomic range of preycaught by Nephila species. Osorio and Moreno(1943) list 28 species of five orders of insects caught

by N. clavipes in Cuba.Thakur and Tembe (1956:332) state that near

Bombay the food of N. maculata "consists almostentirely of butterflies, moths, dragonflies and grass-hoppers." Hingston (1922c:918) suggests that themain prey are nocturnal "moths and flies."

Analysis of the data from prey traps and ourobservations over the whole of the Wau area con-firm that, as Turnbull (1960) reported for Linyphiatriangularis (Clerck), and we have reported forArgiope argentata (Fabricius) (Robinson and Rob-inson, 1970a), the spider is essentially opportunistic.Whatever is trapped at a particular web site andproves edible is prey. To illustrate: one observedspider included several winged phasmids, of twospecies, in its diet for a period of about three weeks.This is the only spider that we know of as takingphasmids in any quantity. Since we were alsostudying phasmids in this area we know that theweb of this spider was built next to the one tree inthe study area that was a food plant used by thetwo species. Thus the spider was in a position thatprovided it with a resource not practically availableto the others. Similarly one of the sample had,for some time, a web site close to a colony of largesocial wasps. For several weeks this spider caughtwasps of this species, which were not available toothers in the sample.

In general the taxonomic range of prey caughtby N. maculata at Wau was considerable. Analysisof the prey trap data did not permit division intotaxa lower than the ordinal level in most cases,since we were dealing with fragments and were notprepared to devote a major portion of time to theirdetermination of lower taxonomic categories. Nev-ertheless, some simple clues were usable; in deter-mining orthopterans we assumed that debris of aparticular green coloration represented tettigoni-

oids rather than acridoids, since the latter werecharacteristically brownish in coloration in thestudy area. We also knew the coloration of thepredominant butterfly species and were able toseparate these from moth debris when wing frag-ments were large enough. This background must beborne in mind when considering our data on thetaxonomic range of prey. (The dried prey frag-ments are available to any entomologist interestedin working with them, either as a whole or in aparticular order.)

Subsequent tables and figures show the datagrouped into seven orders and the categories "oth-ers" and "unidentified." The range of organismsincluded in the "others" category are as fol-lows—Araneidae: Gasteracantha species (1), Neph-ila maculata (3 males and 2 females); Salticidae(1); Lycosidae (1); Isoptera (2); Dictyoptera: Man-todea (2); Phasmatodea: Eurycnema species (1);Lepidoptera larva (1); and Dermaptera (1). Someof the easily identifiable subtaxa within the ordersare as follows—Odonata: Anisoptera; Hemiptera:Heteroptera, Pentatomidae and Homoptera, Cicadi-dae; Diptera: Tipulidae; Lepidoptera: Sphingidae;Coleoptera: Scarabaeoidea, Staphylinoidea, Elater-oidea, Cantharoidea, and Curculionoidea; andDictyoptera: Blattaria and Mantodea.

Numbers in Taxa

Table 2 shows the total numbers of prey in eachtaxonomic category for the whole year. We cannotreasonably allocate the unidentified remains toorders by using the proportions obtained in the caseof the identifiable remains because two groups,the Lepidoptera and Coleoptera, have featuresthat survive maceration. Insects of these orders arethus less likely to become "unidentifiable" than areinsects that are fragile and soft-bodied.

Odonates contribute only a small number to thetotal identified catch (12). Orthopterans are fairlylow in numbers (199) of which tettigonioids con-stitute a major proportion (ca. 75%). Grilloids (11)and Acridioids (30) contribute relatively smallnumbers but are probably useful food sources be-cause of their size when able to fly (i.e., when trap-pable by the spider's aerial web).

Hemipterans (273) are numerically ahead oforthopterans and the cicadas (139) are at least in

••*;*)

CM

nn

g I

=10)

II

,5

mac

CLV

z,"3

adu

ion

-Co

*)

mpl

8

-Pre

y

i

u

S

CM

1

CMCM

CM

4

O l

0 0

m

• < •

i n

O4

_

O l

oo

t o .

<o

I O

• « •

•n

0 4

- i

Pre

y

•

-

i

*

•

•

-

i

i

CM

-

1

1

CM

n

~ p

1

1

1

1

1

1

1

1

1

1

1

1

1

—

1

1

1

1

1

-

1

1

CM

c

IC

m « * —. 04 00 04

_ •»> CM

1 00 • •

| 00 to

A "

r> '

1 *

1 *

• * >

1 >«

1 f4

1 •«•

04 to

1 CM

1 00

1 >O

1 <O

1 <O

1 to

1 to

1 CM

1 • « •

1 to

1 —

1 04

1 •<•

1 1

1 04

Gry

lloi

dea

. .

..

Tet

tigo

nin

ldea

i

•

i

i

-

i

i

I

i

_

i

C 4

1

1

1

'

1

z«

-• m

1 CM 1

—

CM

1

-

t o

-

1

'

1

!

1

.

1

1

1

1

-

CM

" "

m

E!'i3

— I

m

1

I

I

1

-

1

-

1

1

I

t o

0 4

1

t

1

1

1

-

1

1

1

t -

•n

o>

t o

I

1

1

1

—

I

1 •

I 1

I 1

I P"

I 1

. -

I 1

I -

I '

I i

1 ""

1 CM

• •

1 1

1 —

Het

erop

tcra

..

Hom

opte

ra

.. •

Pir

arln

irlr

a

' -

1 tn

1 00

1 S

1 >OCM

1 toCM

1 s

1 CM

1 >nCM

1 O

1 m

I toCM

1 O

1 •*•CM

1 OCM

1 mCM

1 s1 a

1 toCM

1 OlCM

' S

1 OlCM

1 104

'51

1 S

1 S

1 ?

1 s

Lep

idop

tera

Vfn

tln

1 •»

1 CM

— CM

1 1

1 1

1 P"

1 <«<

1 CM

1 CM

1 »O

1 to

1 00

' 2

1 CM

1 —CM

Pt J

p« 00

1 <O

P< 1O

1 Ol

P" 04O4

1 O04

1 <C—

1 s

1 00

— tT

I£

if

i

i

i

i

i

CM

-

1

1

1

1

1

1

1

1

1

1

,

1

1

1

1

1

|

1

Nem

atoc

era

..

i

i i

- on i i

i 1 1 CM

- m i -

l — I —

t o

Bra

rhyc

era

and

~«

1

I

CM

1

CM

-»>

- « •

>o

<o

m

t o

1

1

1

1

1

- 1 1

1 1 1

1 1 1

I m i

i — i

1 - «04

1 O **04

— — 04

1 °° — ""

" 2 2

to oo ©04 ""

1 oo —' -04 —•

^ 2 §

04 Ol 2

1 04 Jg

*" - —

» — 04 to <O -O

I , .4- jO ,

1 <C •*

04 — -

1 |

1 —

1 • •

e< 1

1" 1 04 1OO

« — © 10 4

' ? '

" O 4 '

' f t -

1 04 CO 1

CM |

1 to

1 1

1 1 •

1 1

— 04 .

04 m <o .

i

Cyc

lorr

hap

ha

, , 0 0 ,

. 1 g ,

- 1 04 -

I I 0 4 -

~ — 04 —04

- 1 to 1oo

O««l«(i| |

Hym

enop

tera

..

.

» 1

!r

i I

Col

eopt

era

Scar

abae

oide

a .

i

i

i

i

i

i

-

i

1

i

i

i

i

CM

1

1

1

1

|

f

'

1 —

1

1

1

1

1

1

1

1

1

1

1

1

1

I

|

1

1

,

1

,

Can

thar

oid

ea

.

i

i

i

i

i

i

i

_

i

-

i

i

i

i

i

'

i

i

i

i

n

1

1 00 O S04 Ol g

CM

i oo io "5CM oo ° ;

1 00 to toCM S to

32 107

11.

33

2.4

~ J2 2 ww o .— *— C4 tO 3

to — **?

I to at to

•"Ust o

1 as2i e« :

• M0!

1 to O4 S

1 Ol to «

no*to

I to oo £*° 2 to

0 4

S6 129

2.60

5

1 — »O to'T to to

P. 04CM

1 Ol 00 00<• 04 —

— IO

1 m oo "

"221 K i i

04

1 oo tn fto to Ifl

t o

1 S5S— 1C

CM

• s s |1 S2g

1 04 © ioto 9 io— ^"

CM

•" £ oi on^" lO O— ^>

CM

' C4££

2 61 162

145

1.31

o>I K) K 4

"O ic «

I to ao «o

Oth

er o

rder

sU

nid

enti

fied

Tot

al n

um

ber

s ..

Tot

al

gram

s.

.. .

2

CM

>n

»n

o

O l

00r

?

<o• • •

»n•<•

•>«>

t o

CM

5>

o• •

Ol«n

00tn

t o

s

IOto

t n

•ntn

CM

r.

o

CM

Pre

y

_

i

,

i

i

i

i

1

i

i

l

—

i

-

i

1

i

i

i

i

i

!

1

1

R

|

1

1

•

1

1

,

|

1

|

1

1

«

1

1

'

1

1

1

1

1

1

1

1c

-

,

1

1

1

1

1

,

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

11

to O

— * "

l

-

i

-

I

'

1

•o

1

1

1

-

1

'

1

to

1

-

-

s•ao

1be

~ zh

i

i

:M

•o

1

I

I

I

i

1

I

1

I

I

CM

CM

R

T

'e

<

CM O to Olis r- to "2

1 oo to I

1 «N 1

to | — —

— 1 04 04

IO — 00 1

- 1 • * 1

— 1 — —

1 1 1 —

— 1 i

1 1 — «O

•o 1 1 to

1 1 04 |

— 1 1 00

1 1 — 0 0

1 1 1 01)

— 1 — CM

to i — r»

1 1 1 M

— 1 — <O

04 1 1 CS

<•• 1 1 «n

1 1 1 OS

1 to — «

1

Het

erop

tera

..

.H

omop

tera

..

..

' g

1 CM«9

1 o

1 at o

1 —0 4

i w

1 IOCM

1 <O

1 OCM

1 Olto

1 Olto

1 00to

I ««"

1 S5

1 S3

i <e

1 O

1 S3

i «

1 Ol

1 O)

Lep

idop

tera

Mot

hs

to ©— 04

1 IO

1 S5

1 00

1 S

— »o

— to

1 to

1 to

1 Ol

1 O

04 to

1 to

1 • •

— Ol

1 >o

— IO

1 Ol

1 ©

1 to

1 (C

1 in

1 <«•

1 04

1 I

Bu

tter

flie

s

-

-

-

'

i

.

1

•

i

i

,

i

i

1

1

i

1

1

i

i

i

!

1

— C M —

Bra

chyc

era

and

to —

1 04 P"

1 04 1O

1 04 04

1 1 1

1 1 -

1 1 * >

1 1 00

1 1 *o

1 04 1

1 1 1

1 1 - "

1 1 1

1 P« • •

1 1 • * •

I 1 —

— I to

1 —' 1

1 1 04

1 1 1

1 1 04

1 1 1

1 1 ~

Cyc

lorr

hap

ha

Hym

enop

tera

i

i

i

•

i

,

i

i

•

i

i

i

,

1

i

i

0 4

1

1

1

a

i

i

i

•

i

i

i

•

i

i

i

,

1

i

1

i

i

i

i

i

1399

t n

"

oCM

a

H

S

JM

0 0CM

s

to

9

CM

8

5

S

oCM

oto

179

—i

CM

1

1

|

I

CM

CM

• •

t o

t o

•««

CM

to

m

I O

(O IO04

c*

O l

<

I

t o

— —- 04

1 04

1 >O

1 <O

1 • « •

1 1

1 04

| |

1 1

| 1

1 1

— 1

• •

1 1

1 1

! |

— 1

- 1

1 1

1 1

1 -

Scar

abae

oide

a .

Ela

tero

idea

..

.C

anth

aroi

dea

..

MAN•M «M —

to

1 — ©

1 1 to

3.54

4

CM

04

5S

' " S I S

1 • a

1 — to

1 1 to

• I to

1 — OkN

1 ' s— I m

• " ' 04

O4 . -

1 — toto

P.P.JJ

1 ' S

' ' S

1 -*• o04

00 1 00

— 1 to

' ' 04

1 1 lO

1 ' S

1 1 to

1 1 •#

Dic

tyop

tera

Oth

er o

rder

sU

nid

enti

fied

04

149

2.02

7

** 00

K !

M tO

T *o

—

104

2.05

3

«4 K~* oo0 4

SJSw* O l

CM

991.

523

"I0 4

141

2.24

•*" <oCO 00

O l

8 2

Q if)

220 4

f —<5 —

•-

— tot o

912.

403

2

0110•1.

58

*~ "SIO 1O<o

n '

Tot

al n

um

be

Tot

al

gram

i

NUMBER 149 19

the same size range. Lepidopterans make up thesecond largest group of identified remains (1442)and constitute 33.3 percent of these (23.9% of allprey caught, including unidentified packages). Ofthe Lepidopterans, moths (1428) constituted thelargest number; this remains true even if we admitthe possibility of some misclassification in thiscategory. The preponderance of moths can beaccounted for by a number of factors, one of whichmay be a differential potential for web-avoidancebetween diurnal and nocturnal insects of this order.(There are undoubtedly differences in the numbersof diurnal and nocturnal lepidopterans in thestudy area and this point is discussed later). Dipterans rank third in total numbers (575) and couldbe heavily implicated in the unidentified category.Their importance in the prey-economy of the adultspider may be less than is suggested by their num-bers since very few of those caught in the study areaexceeded 50 mg in weight. Hymenopterans totaled184 and again may be well represented in the un-identified remains. Ants (74) were principally alatesand catches probably coincided with nuptial flights.Wasps (25) and bees (32) are low in numbers. TheColeoptera constituted 37.2 percent of identifiedprey and are the insects present in the largest totalnumber (1610). Of these beetles, those definitelyidentified as scarabioids (179) were probably theheaviest insects caught by the spiders during thisstudy. Seventeen Dictyopterans were identifiedamong the remains. Since small adult roaches werefrequently observed at night and represented 12.6percent of the total number of insects caught onthe sticky traps, they may be an important elementin the unidentified category. In Table 2 they arelisted as "Other orders."

It is very noteworthy that the relative proportionsof the major taxonomic groups that constitute thespider's prey vary very little between the ten indi-viduals that constituted our sample. (This relativeconstancy is even more remarkable when it is con-sidered that spiders nos. 1-10 were not the sameindividuals over the year.) Table 3 gives the pro-portions of the total catch that were made up byinsects of the orders Lepidoptera, Coleoptera, Dip-tera, Orthoptera, Hymenoptera, Hemiptera, andOdonata plus unidentified and "others." The great-est range of variation is in the "unidentified" cate-gory.

Temporal Distribution

Figure 12 shows the distribution of the propor-tions of prey numbers, by the same categories usedin Table 2, divided between the six driest and sixwettest months of the year.

The overall total for the six wettest months(October, November, December, January, Febru-ary, April) is less than that for the six driest months(March, May, June, July, August, September) butnot markedly so (2586 against 3453).

Numbers of orthopterans, dipterans, lepidopter-ans, hymenopterans, and coleopterans are allhigher in the dry months than in the wet months.Hemipterans are higher in numbers in the wetseason. The total weight of prey caught in the wet-test months exceeds that caught in the driestmonths—at least on the basis of the weight ofprey residues: 58.674 grams (47.5%) against 64.87(52.5%) for the wet months.

The fluctuation in weekly total catches for themajor (ordinal) categories can be seen from Figure

TABLE 3.—Percentage composition of prey per spider plus composition of insects caught by window-paneand sticky traps (total prey sample=6039)

Prey

OdonataOrthoptera ...Hemiptera ...Lepidoptera .DipteraHymenopteraColeoptera ...Other ordersUnidentified

Spider

1 2 3 49(N=6J9) (N=*

6 7 8(N=601) (N=48i) (N=6S8)

9 10 Mean W. P. Sticky(N=612) (\=69J) percentage trap trap

0.2 0.4 - 0.2 02 0.2 0.4 - 0.4 05 0.013.6 2.S S.9 2.2 2.0 4.0 4.7 3.8 4.6 2.5 3.3 0.8 1.04.4 4.8 3.7 3.7 4.6 5.7 3.6 3.9 7.6 42 4.5 7.5 8.1

245 25.9 20.1 23.9 23.0 23.2 21.2 21.2 22.9 305 23.9 23.1 1.79.5 9.3 8.2 10.1 10.0 9.3 7.7 8.9 12.8 7.4 9.5 26.6 11.22.4 3.0 2.7 2.9 3.3 2.9 3.8 4.4 2.8 2.9 3.0 9.6 5.0

28.6 25.5 30.6 23.6 24.8 23.0 29.7 32.9 25.4 19.6 26.7 26.5 55.60.S 1.4 0.7 - 0.8 0.8 0.2 0.7 - 0.6 0.54 5.8 17.2

26.4 27.7 29.7 33.3 31.1 30.7 28.8 23.7 23.8 31.9 28.3 - 0.2


.00%

5 0 -

•

ma.

LDIP.

i\ OIC. OIHIi

TOTALWIIOHT

FICURE 12.—Percentage composition of prey, by orders, divided between the six driest monthsand the six wettest months.

13. Note that the Odonata, Orthoptera, Diptera,Hemiptera, and Hymenoptera are absent fromsome weekly catches, whereas Lepidoptera andColeoptera are always present. Table 4 givesmonthly totals for the entire year and is comparablewith Table 2 of Robinson and Robinson (1970a:350). From this it will be seen that two orders, rep-resented by low total numbers, are absent fromidentified catches in some months. Thus odonatesare absent in January, March, April, October, andDecember. Dictyopterans are absent in April, June,July, and December.

Within some of the more numerous orders thereare absences of subgroups in certain months. Thuswithin the Orthoptera identifiable gryllids werepresent only in September, October, November,December, with more than half the total beingcaught in October (a wet month). Tettigonioidswere caught in all months, but three months (July,August, and September) account for more than halfthe total catch. The acridoids were absent in

identifiable remains for the months of February,April, May, August and October, and two months(March and November) account for more than halfthe yearly total. Some hemipterans were presentin all months, but the large and conspicuous cicadas(suborder Homoptera) were caught in all but one(October) of the six wettest months and only two(March and May) of the six driest months. Morethan half the total yearly catch of cicadas occurredin two months: April and November.

Lepidopterans were caught in all months but ofthe positively identified butterflies (13) eight werecaught in four of the driest months (March, May,July and August) and five in four of the wettestmonths (November, January, February and April).More than half the yearly catch of moths wascaught in five successive months (April, May, June,July, August).

Monthly totals of dipterans vary between 11(December) and 95 (August) and similar totals forHymenoptera between 7 (February) and 32 (May).

NUMBER 149 21

•Othinches

•Mks"S iS~FIGURE 13.—Weekly distribution at study site of numbers of

prey, by order, against rainfall.

Total catches lor coleopterans range from 99 inApril to 164 in January. The large scarabaeoidbeetles appear in large numbers in September, and172 out of the year's total of 179 are caught in thisand the succeeding six months. This period includesthe weeks of highest average weight of unit preyremains.

Numbers of unidentified prey remnants varythroughout the year and constitute 28.3 percent olthe yearly total. In the six driest months the totalis 1040 against 672 for the six wettest months. Thedifference, 368, is approximately 21 percent of thetotal. The highest number of unidentified remains,237, occurs in August and the lowest, 71, in De-cember.

Weights of Prey Remnants

The total weight of the prey remnants collectedin the prey traps was 123.544 grams. The weeklydistribution of the weights that make up this totalare given in Table 2 alongside the average weightper unit prey (obtained by dividing the weight bythe total number of prey items, identified and un-identified). What this figure may represent interms of live weight is discussed in the section onenergetics (page 33). It is more than half of theestimated live (wet) weight of prey caught by asample population of Argiope argentata in one year(Robinson and Robinson, 1970a:349).

Of this total weight 64.87 grams (52.5%) werethe contribution of the six wettest months and58.674 grams (47.5%) were collected in the sixdriest months. The highest weekly totals occur inOctober when, in two successive weeks (23 and 24),the totals exceeded 4.6 grams. Totals of over 4grams occurred in one week (19) in Septemberand one week (25) in November. In fact in the7 weeks commencing 23 September and ending 10November, the prey remnants totaled 29.563 grams,23.9 percent of the yearly total! This was also aperiod of high average weights of prey remnants(all averaging over .03 grams per unit). This sug-

gests that some large prey item is seasonally avail-able to the spiders. This is borne out by the factthat scarabaeoid beetles feature prominently in thecatches for these weeks. Melolonthids, in particular,were recorded in enormous numbers at light duringthis period and featured heavily in prey-trap re-


TABLE 4.—Monthly totals of insects caught by mature Nephila maculatasample (N=10)*

Prey Jan. Feb. Mar. Apr. May June July dug. Sept. Oct. Nov. Dec. Total Percent

OdonataOrthoptera

GrylloideaTettigonioideaAcridoidea

Subtotal ...Hemiptera

Heteroptera ....Homoptera ...Cicadidae

SubtotalLepidoptera

MothsButterfliesLarvae

SubtotalDiptera

NematoceraBrachycera &

Cyclorrhapha ...Subtotal

HytnenopteraAntsWaspsBees

SubtotalColeopteraScarabaeoideaElateroideaCantharoideaCurculionidea

SubtotalDictyoptera

BlattariaOther ordersUnidentified

Grand total .Percent caugh

per month

1125

319

27

841

8546

1

1014715

1

1164

1

S

1

16

20

1361

13731

7123

11

134

89

173

58

16

92

9452

12173

41

178

1192

11

2

34

12921

13256

3

47 31 52 59

1 1 2 61 1 - 18 5 10 8

1592

99

4 3 2 -4 2 1 2

97 142 13S 77440 479 505 419

59

1253

29

2291

273154

103

32

1193

122

136

31

101

1231

3

7

172

230 17267 62

4

626252

15141

32

11 3

151 177661 597

234

28

13

1402

14258

17

75

4182

15136

1

146 137

174

575

33

2

f>

1093

112

82

1395o523

12125

127

237625

231

134

52

1

8

94

94

39

241

64

137

30694121

!68

490

6II

1761

1

8

98

9882

61642

12

18

9760

1

113 158

32

167487

132

2010

35

38

1132

54

791

413

62

2108030

1

1

112

118

424

4154

14631

5161

66

80 668 111

II

251

8931111

106

71

337

121111

14730

199

623537

139273

142813

11442520

11

44

575

53257432

184

1395179

1121

4

1610

1715

17126039

.18

.182.43

.53.31.03.58.61

2.3

23.6532.02

23.888.16

.18

.739.52

.88

.411.23.53

3.05

23.1

2.96.18.35.07

26.68

2825

28.35

7.29 7.93 8.36 6.94 10.95 9.88 9.52 10.35 8.11 8.06 7.02 5.58'Numbers on row opposite ordinal name are of prey not identifiable to lower taxonomic levels.

mains. Their emergence is known to be facilitatedby wetting of the ground where they pupate, andthe rainfall in September followed a dry spell(Figure 13). The catch for the week commencing23 September (week 19) included 14 cockchaferswhich also figure prominently in the followingweeks 20-25 (15, 14, 11, 10, 11 and 5). The catchfor October also included five other scarabaeoidsof more than 20 mm in length. Other contributors

to the heavy average weight during this periodwere very large gryllids (page 20).

In all there was one week (week 48) with a totalof less than one gram of prey remnants (April1971), six weeks with between 1.0 and 1.5 grams,thirteen weeks with between 1.5 and 2.0 grams, 22weeks with between 2.0 and 3.0 grams, five weekswith over 3.0 and less than 4.0, and five weeks withover 4.0 and less than 5.0 grams.

NUMBER 149 23

We cannot assign weights to the numbers of preyin each taxonomic division since we were dealingwith prey remnants and it would have been far tootime-consuming to weigh each of the 6039 trashpackages separately. Similarly we were not able torecord size ranges of the prey items (as did Robin-son and Robinson, 1970a, when recording prey inthe webs of Argiope argentata). Thus there is nomethod of weighing samples and assigning approx-imation of weights to the numbers of prey in eachtaxon. We can make, however, some very crudeguesses about the effect of considering weight, onthe relative importance of the various categoriesof prey. Thus, in the case of the orthopterans, itwould be reasonable to assume that most of thosethat were caught would be alate adults that couldfly and therefore be more likely to move into webssited well above the ground than would juveniles,which can only jump. We would assume, therefore,that the relative importance of orthopterans as prey(3.3% of the total catch) is not likely to be dimin-ished because of below average weight. On theother hand few dipterans exceed 0.05 grams in liveweight and we would predict that if the weight ofthe 575 dipterans caught by Ncphila was knowntheir contribution to the total catch would be lessthan the 9.5 percent calculated from their numbers.Because the catch of hemipterans includes 139weighty cicadas (average weight probably above0.2 grams) we would expect the prey in this cate-gory to be more important than their 4.5 percentby numbers suggests. The two orders providing thepredominant numbers, Coleoptera and Lepidop-tera, are, we feel, likely to contribute numbers ofweighty prey and their predominant position islikely to be reinforced if we considered their con-tribution to the calories available to the spider.These considerations suggest that a ranking oftaxonomic categories in terms of "energy bundles"would read, in descending order: Coleoptera, Lepi-doptera, Orthoptera, Hemiptera, Diptera, Hyme-noptera, Odonata, and Dictyoptera.

Comparison of the monthly distributions of theweight of prey with the numbers of prey givesrise to some interesting observations. The weightsof prey start to rise in August whereas prey num-bers fall from this point until December. Theweights reach a peak in October and then fallthrough November to December but the December

total is still higher than all the other months ofrelatively low weights, except March. The Decem-ber low point in prey numbers is the lowest ofthe year. The rise in prey weights during thisperiod corresponds to an increase in rainfall tothe year's peak in October (which was also themonth with the largest number of days of rain).As stated earlier, the period of high weights andrelatively low numbers was one when heavy scar-abaeoid beetles were abundantly represented in thecatches and these probably account for the differ-ences cited above.

Comparison with Argiope argentata

As far as we are aware the only study of a spider'sprey over a one year period is that on Argiope ar-gentata (Robinson and Robinson, 1970a). Therehave been a number of excellent studies of the preyof temperate orb-weavers (see Kajack, 1965, for anextensive bibliography), but the subjects have allbeen spiders with a seasonal activity covering onlypart of a year. Argiope argentata differs from Neph-ila tnaculata in being largely diurnal, choosing openareas for its web sites, having an advanced patternof predatory behavior (Robinson, Mirick, andTurner, 1969), and being considerably smaller andlighter than N. maculata.

The area where studies of A. argentata were car-ried out has a much more seasonal climatic patternthan Wau. At Barro Colorado Island in Panamathere is a pronounced dry season of 3 to 4 months(total rainfall average 8.2", three months with lessthan 3" rain) and a distinct wet season of 8 to 9months, total rainfall average 98.9". This is nottrue of Wau where in only one month (August)did the total rainfall drop below 3"', and this wasexceptional.

There are major differences between the catchesof the two spiders in taxonomic content and intemporal distribution. In terms of numbers, hy-menopterans constitute the most important preyitem for Argiope with a total of 3304 out of 4672prey items. This compares with 184 out of 4327identified prey items in the case of Nephila. This is70.6 percent of the total prey caught by Argiope and3.0 percent of the total caught by Nephila. Interms of weight these hymenopterans constitute anestimated 18.1 percent of the total caught by Argi-


ope and are most unlikely to increase their rela-tive importance to Nephila if weight is considered.Orthopterans were the second most important preyitem, in numbers, to Argiope (11.4%) and consti-tuted only 3.3 percent, by numbers, of the Nephilacatch. In weight, orthopterans rise to the predomi-nantly important prey item for Argiope (55.2%)and are most unlikely to assume this importancefor Nephila. Of the orthopterans caught by Argiope,tettigonioids account for 57.4 percent, acridioidsfor 41.4 percent, and grylloids for 1.5 percent. Theequivalent percentages for Nephila are 73.9, 15.1,and 5.5, with 5.5 percent not assigned to super-family.

Coleopterans are the most important prey, nu-merically, for Nephila (26.6%) and are probablythe most important prey item in terms of weight.For Argiope they account for 5.8 percent of thecatch by numbers and an estimated 6.6 percent byweight. We have not seen Argiope feeding on largebeetles. Similarly, lepidopterans account for 23.9percent of the Nephila catch and only 2.9 percentof the Argiope catch, by numbers, and 5.4 percentby estimated weight. Of the lepidopterans caughtby Nephila 99 percent were classified as moths,0.9 percent as butterflies and 0.1 percent as larvae.In the case of Argiope 89.8 percent were butterfliesand 10.1 percent moths.

Of the other orders odonates were caught in smallnumbers by both species of spider, Nephila catching12 (0.2%) and Argiope 48 (1.03%, not 0.14% asgiven in Robinson and Robinson, 1970a: 349).Hemipterans occur in relatively small numbers inboth cases, Nephila catching 273 (4.5%) and Argi-ope 207 (4.4%). As stated earlier the relative im-portance of hemipterans to Nephila should perhapsbe increased if we consider that the cicadas areheavy, and therefore an important source of foodthat is belied by their relatively small numbers (139or 2.3% of the total catch by numbers), Argiopedid not catch large cicadas.

Dipterans account for 9.5 percent of the totalcatch of Nephila and 2.4 percent of the catch ofArgiope. Their importance to Nephila is probablyless than their numbers suggest because of theirsmall size.

Many of the differences between the taxonomiccomposition of the prey of the two spiders prob-ably result from differences in habitat, web siting,

and web structure. Of these the first two are prob-ably the most important. The two spiders are un-doubtedly catching insects that preponderate indifferent proportions in the two habitats, and thetwo zones within the two habitats. Robinson andRobinson (1970a:355-356) discuss the factors thatmay influence the catch of Argiope argentata atBarro Colorado Island. The influences acting todetermine the scope of the Nephila maculata catch,at Wau, are discussed above.

Differences in the seasonal patterns of the catchesof the two spiders are grossly relatable to differen-ces in the climatic pattern at the two localities. Theeffect of the lowland monsoon climate at BarroColorado is not entirely clear but there are sug-gestive data (Robinson and Robinson, 1970a:352-353). The dry season in Panama is clearly moreinimical to spider activity than the short dry inter-vals at Wau.

Comparison with Insect-Trap Catches

Kajack's (1965, 1967) studies of the food rela-tionships between spiders and their prey werebased on intensive studies of three species of smallaraneids. Prey caught by the spiders was sampledby removing prey items from the webs and theabundance of mobile insects was evaluated by usingsticky traps. Kajack found (1967:814) that "thefauna captured on these traps coincided with thosecaught in the webs so that it was possible to regardthe number of insects on sticky traps as an indexof abundance of the potential food of the spiders."The three species that Kajack studied, Araneusquadratus Clerk, A. cornutus Clerk, and Singahamata (Clerk) are all smallish spiders (Locket andMillidge, 1953), considerably smaller than Argiopeargentata, and very much smaller than Nephilamaculata. Kajack states (1967:813) that "very smallinsects which weigh hundredths or tenths of mg(from 0.05 to 0.1 mg) constitute the vast majority

of prey caught in spiders' webs." Thus die spidersKajack studied caught very much smaller prey thaneither Argiope argentata (Robinson and Robinson,1970a) or Nephila maculata. This point is relevantto a consideration of how insect sampling tech-niques can be used to assess the prey available tothe larger tropical orb-web spiders.

Pilot studies that we carried out in Panamasuggested that sticky traps would not adequately

NUMBER 149 25

sample the availability of prey to the larger orb-weavers. Many insects seemed capable of detectingthe presence of such hazards and avoiding them.This view was confirmed by some studies that wecarried out (in conjunction with W. Graney) toinvestigate the possibility that insects might de-tect the presence of apparently much less conspicu-ous traps such as spider's webs. After various trialswe eventually decided to watch the behavior ofinsects flying in the vicinity of a 10 X 5 foot lengthof double nylon mist netting that was erected inthe laboratory clearing at Barro Colorado Island.This was large enough, unlike a spider's web, toprovide a reasonable number of interceptions byinsects in an 8-hour observation period (dividedinto half-hour shifts). The net was presumablymore conspicuous than a web but much less con-spicuous than a sticky trap. Some insects provedquite capable of detecting its presence and madeavoidance maneuvers. Thus an insect would flytowards the net and then, when nearby, rise almostvertically, fly over the top, and proceed on its flightpath. Other insects blundered straight through thenet striking its threads, while others changeddirection after striking the net. It was interestingto find that the same species of insect would in oneset of circumstances avoid the net while in thecourse of another behavior it would hit it. Thuslarge wasps flying on a "beeline" would hit thenet, but when flying erratically from flower toflower they would miss it and take very obviousavoiding action. This suggested to us that someinsects might be able to avoid webs. We tried totest this further in New Guinea by mounting aNephila maculata web on the glass of a window-pane trap and operating this next to a normalwindow-pane trap. Each day we moved the trapsaround and tried thereby to eliminate any positioneffect. There were interesting differences betweenthe catches of these two traps over a 14-day period.Catches of insects belonging to the orders Diptera,Hymenoptera, Coleoptera, Lepidoptera, and He-miptera were all lower in the case of the trap withthe web mounted on the glass. The total number ofinsects involved (462 caught by the experimentaltrap, 505 by the control) is unfortunately too smallto enable us to determine whether there are anygroup* within orders that are less sensitive to

spider's webs than others. We are continuing thisexperimentation in other localities.

Our doubts about sticky traps led us to considerthe possibility that window-pane traps might givea better index of the food available to large spiderssuch as Nephila maculata. To test our Panama-based conclusions about the inadequacy of stickytraps we also used these at Wau. Data from thetraps are of interest in two respects Firstly theyenable a comparison to be made between thespecificity of the web and the specificity of thetraps. Secondly the trap data should reveal anycycling or periodicity of outbursts within specificgroups of insects, which should be valuable as acheck on fluctuations in the temporal pattern ofoccurrence of particular groups of prey.

In fact for the 43 weeks (3 June 1970-30 March1971) for which we have data from the window-pane traps the three traps caught more insects thanthe ten spiders: 7555 compared with 4850. Duringthe same weeks the window-pane traps caught morelepidopterans than the spiders: 1743 compared with1064, and more beetles: 2003 compared with 1342.The weekly analysis of numbers is given in Tables5 and 6. (All these figures deal with insects over4 mm in length; we decided that prey less thanthis size would pass straight through the web ofan adult Nephila). On the whole the majority ofinsects caught by the window-pane traps was smalland did not include large lepidopterans and beetlesof the sizes that were seen in the spider's webs.Table 7 details the size distribution of the in-sects caught by the window-pane traps.

In Figure 14a we have plotted the weekly totalsof prey from the prey traps and insects caught bythe window-pane and sticky traps. From this it willbe seen that there is no obvious correlation, positiveor negative, between fluctuations in the spider'sprey numbers and the numbers of insects caughtby the traps. We have worked our correlation co-efficients for total numbers of prey caught per weekby the spiders and insects caught by the window-pane traps. These give no basis for rejecting thehypothesis of no correlation between the two setsof results. We believe that there is no correlationbetween the two. Similar statistical results applyto comparing the weekly totals of lepidopteransand coleopterans caught by the spiders and thewindow-pane traps. Interestingly enough the pro-


TABLE 5.—Weekly analysis of insects caught in 3 window-pane traps

Prey

Odonata

HemipteraCicadoidea

DipteraHymenopteraColeoptera

Scarabaeoidea . . .ElateroideaCantharoidea . . .

Other ordersTotal

• No traps erected

3»

?18

.57fir.1749

116

4220

4

11

212

M571848

--9

3195

during

5

111

-411751845

13

1819

231

6

?8I

?7633827

2412

175

weeks 1

7

12|

inn1834

1413

206

8

141

46111

1326

25_7

236

and 2.

9

182

4fi011819

-712

195

10

151

51663731

56_9

224

11

18-

61fin2515

-5_4

179

Numbers on

12

7-

314520IC

17_3

130

13

201

44632115

46?2

178

14

20

4344

920

66?3

153

15

211

-40201021

G1_6

123

16

113

-5020II20

51_2

123

row opposite are of

17

14-

53331825

39_2

157

18

2162

4643

930

2512

158

19

_

83

3631164G

35_1

149

20

217

-62281719

3814

161

21

110

58451433

4?29

181

prey not identifiable

22

111

154262517

4824

53

to

23

1142

451914

33

56_

12151

24

,

6-

29221035

7415

120

25

210

2412420322622

14157

26

283

S3251437

633

19153

lower taxonomielevels.

Prey

Orlhopttra . . .Hemiptera . . . .

Cicadoidea . .Lepidoptera . . . .DipteraHymenoptera . .Coleoptera

Scarabaeoidea .ElateroideaCantharoidea

Dictyoptera . . . .Other orders . .

Total . .

19

SI.4

1811

75-

1479

10179

20

412-6

1512

72-

17697

160

TABLE

21

118-5

192044

-1057 '

14143

22

,

1711

2010

64-

156

• 72

144

6.—Weekly

23

411-4

184

44.

143

1217

131

24

_

8-6

129

47-27

118

110

25

15-6

137

52_489

11116

26

310

-2

2310

47-9

13156

138

analysis

27

310

1-

103

35.

19119

10111

28

12473

18II62

1118

1812

176

29

4831

125

58I

106

161

125

of

30

_

61

9162

35.3936

90

insects

31

_

1764

171540

211595

131

32

210--85

68-89

20I

131

caught

33

16I-

186

74-9

11197

152

34

182-

143

993

117

3513

198

35

-4-1

11

56145

184

105

i n

36

1(i1-

165

99298

145

166

two

37

_5-3

215

871

107

355

179

sticky

38

16-1

133

59-96

19-

117

59

14--

104

31-

131

221

87

traps

10 41

161293

42163

173

94

i4-186

27-

143

304

98

42

_6--

175

54-87

296

132

*

13

_9-1

117

37-64

186

99

44

_6--

173

491

105

213

115

49

15--

123

45-

139

153

106

Total2

32262

2460

396178

150213

269179446170

3533

Ordinal

Tola2

32-

28660

396178

---

1963446170

Per-cent

.06

.9-

8.11.7

11.215.04

---

55.5612.64.8

name are of prey not identifiable to lower taxonomie levels.

TABLE 7.—Size distribution of insects caught inwindow-pane traps*

Insects

OdonataOrthopteraHemiptera

HetcropteraHomoptcraCicadoidea

LepidopteraDipteraHymenopteraColeopteraOther orders

Total numbersTotal percent ...

<10mm 10-19mm 20-29mm > 30mm Total

129

208175

12451899586

1659368

616981.65

31

561458

45110812232672

123816.40

27

30

47

3

14

17

1451.92

113

.04

167

26421b88

174S2010

7222003441

7555

'Numbers on row opposite name are of prey not identi-fiable to lower taxonomie levels.

portions of the total catches that were constitutedby lepidopterans and coleopterans are similar whenwe compare the spider's prey with the window-pane

traps. The histograms in Figure lib show the per-centage of the total numbers constituted by nineorders of insects in the catches of the spiders,window-pane traps, and sticky traps. The differ-ences between the catches of lepidopterans andcoleopterans mentioned above are less than 1 per-cent. The sticky traps, on the other hand, caughtproportionally less lepidopterans (less than 1.7%)and proportionally more coleopterans (over 55.6%).These results indicate that very few of the unidenti-fied prey packages obtained from the spiders arelikely to be from lepidopterans or coleopterans; theproportions of these two orders are thus probablymore truely comparable with the trap data thanwith those for other orders, which might be alteredif the unidentified prey could be assigned to orders.Since we know that the spiders are catching largerinsects of several orders than are the window-panetraps, and catching them at different times, the

NUMBER 149 27

Prey

DipteraHymenopteraColeoptera

Scarabaeoidea . . .Elaceroidea . . . .Cantharoidea . .

DictyopteraOther orders

Total

27

6. 14

3. 47. 19. 12. 41

3557

. 162

28

4n

3VIV,1841

-1328

13173

29

•\1?

<;

n1242

18

105

11154

"0

f)

?*>4r>->n2358-6

1384

238

31

?">\

7riR

783069

16

1365

296

32

17

538->1901-2864

246

TABLE

33

53

•>r.112240

-5

1361

173

34

<>1

4fi1251-6

1567

171

5.— (Continued)

35

R1

?r>4fi1531

12787

151

36

<>63

497?1653

13738

223

37

15fl

11471233-5871

167

38

?

1''fi481652

16

1154

177

39

15

r.s451855

17567

230

40

I11

141341538

1676

24185

41

R1

%?133129

114

13152

42

144

?qu,1025

-4

125

14153

43

I1?4

17?17

1714

133

28132

44

|11

121?T

1336

-4

52

11130

45

81

11??9

22-

1252

20134

Total

67480

8817432010

7221491

14190308121320

7555

Ordinaltotal

67

56817432010

722

2003121320

Percent

89

7 5223 0726.6

9.56

26.511.64.24

similarity in proportions of lepidopterans andcoleopterans is in all probability merely a strangecoincidence.

The difference between the sticky traps and thewindow-pane traps in the proportions of these twogroups of insects is probably significant in terms oftrap specificity. The low catch of lepidopterans bythe sticky traps could be because these insectseither (1) can detect the presence of the stickytrap and avoid it more readily than they avoid awindow-pane trap, or (2) can escape more readilyfrom a sticky trap than from a window-pane trap.While (1) is possible even when the vast majorityof the lepidopterans involved is nocturnal moths(the level of illumination on many nights may stillpermit visual location or detection of traps) none-theless (2) is an attractive hypothesis. It is attrac-tive since both the spider's web and the stickytrap employ a similar method of initially arrestingthe insect in flight—an adhesive—and lepidopteransare known to have a high escape potential fromspider's webs (page 64) and should have a highescape potential from sticky traps. It is temptingto think that part of the difference between thespider-prey results and the sticky-trap results isdue to the active presence of the spider on its trap.

The differences in the proportions of beetlescaught by the spiders, the sticky traps, and thewindow-pane traps can be explained in anotherway. We can assume that the spiders and thewindow-pane traps both lose a number of beetlesthat are caught by the sticky traps. The spiders losesmall beetles that strike the web and are eitherignored because they are too light, or which are sosmall that they pass through the wide mesh of the

web. The window-pane traps presumably lose somebeetles that are strong enough to. crawl out of thewater after they have been knocked down by im-pact with the glass. The sticky traps catch bothsmall and fairly strong beetles. Very strong beetlesescape from them.

From these results and the considerations de-tailed on page 25, we are convinced that neithertype of insect-sampling device is adequate for thepurpose of assessing the availability of prey to largeorb-weaving spiders. Our light trap certainlybrought in some of the types of large prey that thespiders caught, but again it is known that the re-sponses of insects to ultraviolet light are confinedto certain groups of the many nocturnally activeanimals (Southwood, 1966:200-201, for commentson reliability).

Rejections of Trapped Insects

Rejection of distasteful or otherwise obnoxiousinsects occurs, in most cases, after Nephila has bitteninto the prey. It is then pulled out of the web andcast away from the plane of the web and falls tothe ground. This treatment, in most cases, damagesthe insect. When a prey is detected as obnoxiousbefore being bitten, it may be left for some timein situ without being bitten. It is almost alwaysremoved from the web, however, at some stageduring "housekeeping" activities (when particles ofplant material that blow into the web are alsoremoved). This type of rejection also involves roughhandling by the spider, and the prey is often pulledfrom the web in the jaws. Thus, in both cases, therejected prey may be damaged and wholly or partly


MR TRAPS

WIEKS

5 O -

0 , B .

• •Hyn

AhFIGURE 14.—a, Distribution of weekly totals of prey caught by spiders, and insects caught bywindow-pane traps and sticky traps; b, percentage composition of catches by spiders, by orders,window-pane traps, and sticky traps. Top portion of "others" category demarcates Dictyopterafrom the remainder.

immobilized as a consequence. From an early stagein our sampling of the spider's prey we were awareof the presence of numbers of dead but intact insectsin the prey traps. Conspicuous among these werelycid beetles, which are known to be obnoxious toa range of predators (Linsley, Eisner and Klots,1961), and it seemed probable that these lycids had

been caught and rejected by the spiders. Experi-ments showed that the spiders did in fact rejectlive lycids and we therefore decided to assumethat all the intact (=undigested) insects found inthe prey traps were rejects unless there were strongindications to the contrary. As it turned out thegreat majority of the insects so classified were apose-

NUMBER. 149 29

matic and therefore, unless they were Batesianmimics, obnoxious in some degree. (We are notsuggesting that the spider is responsive to the color-ation of the prey but merely that this is a confirma-tory piece of evidence.) Perhaps a proportion ofthe insects classified as rejects died in the vegetationabove the prey traps and fell in, but this seems tobe a very small possible source of error sincecorpses of known innocuous groups are extremelyrare.

Of the total 371 specimens rejected during thestudy year, their composition was as follows: Lepi-doptera, 16; Neuroptera, 18; Coleoptera: Lycidae,198; Hemiptera: Heteroptera, 12; Homoptera, 79;Diptera, 37; Nephila males, 2; Phasmatodea: Eury-cnema species, 2; Hymenoptera: Ichneumonoidea,1; Unidentified, 6. The 371 total represents 0.3 per-cent of all the insects found in the traps. Of therejects, coleopterans account for 53.3 percent, ofwhich the lycids account for the great majority.Hemipterans rank second (24.5%) of which themajority were homopterans. These have not yetbeen identified but were bright blue in color. Rothand Eisner (1962) do not list any identified repug-natorial secretions from the homopterans, butmany are known to produce waxy secretions whichmay be defensive. The heteropterans included pen-tatomids and reduviids, the former certainly pro-duce defensive secretions and many of the latterhave a painful bite. Neuropterans were present insmall numbers (18) as were lepidopterans (16).Lacewings are known to produce odors and someforms have eversible abdominal glands (Riek,1970:477). The "unidentified" 6 specimens includecockroaches, phasmids, and flies.

Lycids were rejected in all but six weeks of theyear and theii absence on these occasions seems tobe quite random. This is of interest since Turnbull(1960) showed that the spider Linyphia triangularis(Clerck) would accept prey items at some stagesin the season that were unacceptable at others. Wehave no data to suggest that lycids were acceptableat any stage.

Bristowe (1941:262-220) has surveyed the infor-mation available on food of spiders in great depth.He notes that coccids and aphids are either re-jected or "accepted with hesitation," but that otherhomopterans are for the most part accepted byspiders. He also states that lacewings are rejected

by most spiders. Among the Diptera he notes thatsciarids, cecidomyids and empids are sometimesrejected. Flies that we identified as bibionomorphs(i.e., closely related to the above flies) were cer-tainly among those rejected.

KLEPTOPARASITES

Numbers and Nature

Theridiid spiders of the genus Argyroides arefrequently associated with orb-web spiders. Theybuild their apparently unstructured complex ofthreads close to the plane of the host web and moveabout these and onto the host web. They have beengenerally referred to as inquilines. However thisterm has been used in such a variety of meaningsas to be imprecise and we think that the termkleptoparasite is more appropriate.

We have conducted studies of the relationshipbetween Argiope argentata and its kleptoparasites,and decided that since Nephila maculata seemed tohave a large number of these theridiids associatedwith its webs further study would be worthwhile.Fr. Chrysanthus has identified at least two speciesfrom the material that we sent him and otherspecimens remain unidentified. Kleptoparasites ofNephila maculata include at least the species Argy-roides argentatus Cambridge and Argyroides mini-aceus (Doleschall). Of these the former was alsofound (by Dr. Y. D. Lubin) in the web complexof Cyrtophora moluccensis, which suggests that itmay not be host-specific.

At the start of our study we counted the klepto-parasites present in 86 webs of adult N. maculatawithin an area of about 500 meters around thestudy area. We also counted the kleptoparasites onan equal number of webs belonging to immatureN. maculata whose sizes ranged from 20 to 38 mmin length. The total kleptoparasite "load" foradults was 236, and for immatures 129. Of theadult webs 23 out of 86 had no kleptoparasites(27%). The juvenile webs contained kleptopara-sites in all but 28 (32.5%) cases.

At each morning's web census we counted thenumber of kleptoparasites present in the ten Neph-ila samples. Figure 15 shows the fluctuation in theweekly totals. The variation is from 35-112, theaverage number per web per day is 0.92. This, whenone considers that such tiny spiders could easily


too

IIO 2O

I30

I4O

IBO

FIGURE 15.—Weekly fluctuations in numbers of kleptoparasites present on the sample Nephilamaculata webs.

be overlooked, is suggestive of a fairly regularrelationship. Fluctuations in the weekly totals ofkleptoparasites do not seem to be correlated withsuch factors as fluctuations in rainfall, number ofdays of rain, weight of prey, or numbers of prey.The maximum number of kleptoparasites that werecorded on any one web, on one day, was fifteen.

Activities

Unlike Argiope argentata and many other ara-neids, Nephila species do not store or leave food atthe capture site after it has been immobilized. Allprey is carried from the capture site and stored atthe hub region (Robinson and Mirick, 1971). Thisaspect of the predatory behavior of the host meansthat the kleptoparasites are not able to steal foodfrom the capture sites as has been described forthe kleptoparasites of A. argentata (Robinson andOlazarri, 1971:35). On the other hand the theridiidsoften move from their "waiting positions" aroundthe web and down to prey as it is being attackedby the host. We have seen the theridiids moving onthe surface of the prey and apparently feedingthere while the host is administering the bitingattack (which may last several minutes). When thehost starts wrapping the prey, after the bite, thekleptoparasites move off to their complex of threadsabove the surface of the host web, or below it on

their drag lines. This sort of feeding, during thehost's attack, is not without risk and we have seenthe kleptoparasites become enswathed in silk, along-side the prey, as the host wrapped the prey package.Because the maze of threads built by the klepto-parasites is in contact with the host web at variouspoints we suspect that the theridiids may be alertedto the presence of prey in the web by vibrationsproduced as a result of the activities of the hostand/or the struggles of the prey. Those theridiidsactually standing on the host web could be expectedto locate the source of the alerting vibrations morerapidly than those waiting on their own structuresoutside the web. We have not been able to demon-strate this.

In other cases, after the host has transported theprey to the hub and has commenced feeding, thekleptoparasites approach and move onto the prey,often climbing down the silk thread by which theprey is secured to the hub. They also behave in thesame way towards prey stored at the hub. Figure16 shows three kleptoparasites feeding on the sameprey as the host and one feeding on a stored preyitem. Stored prey usually become the object ofstealing sensu stricto. Kleptoparasites attach lines tostored prey and then move up into their own threadcomplex, dragging the prey with them. We haveseen prey several hundred times the weight of thetheridiids stolen in this way. Once the thread

NUMBER 149 31

FIGURE 16 Adult female, Nephila maculala, and four kleptoparasites (numbered) feeding on prey.


attaching the prey to the host web is severed thehost appears to be incapable of detecting its loca-tion. Nor does the host appear to be alerted bythe sudden loss of weight that occurs when the preyis removed from the web. On the other hand wehave seen hosts apparently searching for missingprey packages hours after they have been removed.The kleptoparasites give no convincing evidence ofcooperating to remove large items of prey althoughthey are apparently tolerant of each other as theymove around on the surface of large prey packages.

The host spider does not often react to the pres-ence of kleptoparasites when they are merely mov-ing about on its web. We have only one record,from many hours of observation, of the host leavingthe hub and attacking a kleptoparasite. This wasbitten and carried back to the hub where it waswrapped into the prey package on which the spiderwas already feeding. Robinson and Olazarri (1971:35) observed Argiope argentata attack and killtheridiids that were removing prey from capturesites. Host spiders, from time to time, becomeaware of the kleptoparasites moving about on theprey on which they are feeding and respond bymaking vigorous brushing movements around theperiphery of the prey with their legs. This is quitea different behavior to that ensuing from move-ments of the prey within the prey package. Whenthe latter occurs the spider usually continues feed-ing but occasionally rewraps the prey in silk.

Colonization of Host Webs

Kleptoparasites may remain in association witha host web for up to three or four days after thehost has gone. This lingering on could be adapativesince our studies have shown that temporary ab-sence of a host spider from a web site is not un-usual. In fact, spiders leave their webs, lay eggs,and then return to the same site—and sometimesthe same wel)—after one or two days absence. Alsoduring periods of heavy rain the spiders may spendone or two days out of their webs hanging undernearby leaves. New host webs are colonized withsurprising speed and we have recorded up to fivekleptoparasites in the web of an adult Nephila 48hours after it has moved into a web site that hadbeen unoccupied for several weeks. This occurredeven in areas where the host species had been com-pletely absent for months. If, as we have suggested,

the kleptoparasite is not strictly host-specific, itcould operate by moving onto any araneid web thatit detects and remaining there when conditions aresatisfactory. Presumably an intermittent supply offood would constitute a reward.

General Considerations

It is difficult to assess the role of the kleptopara-sites in the energy economy of the host. If weassume that a host may support four or five klepto-parasites throughout most of its adult life theircombined wet weight would certainly not exceed40 mg. Assuming that dry weight is 40 percent ofwet weight this gives a dry weight of spider at16 mg. Taking a metabolism figure of 27 caloriesper gram per day for spiders, based on the perhapsmore active lycosids (Macfadyen, 1963:234), we geta daily intake of 0.43 calories per day by the klepto-parasites. How much of this would be availableto the host, in their absence, is difficult to assess.All of the energy in food that was actually stolenwould be a debit. However, a proportion of theenergy taken from a source on which the host wasfeeding could be inaccessible to the host, since thekleptoparasites may be able, because of their smallsize, to extract food from prey structures such asthin legs, that the host cannot cope with becauseof the great size of its chelicerae. We have alsoobserved scavenging behavior by the kleptopara-sites, toward very small prey that become caughtin the web and which the host ignores. This typeof feeding may occur without loss to the host andcould contribute to satisfying the energy require-ments of the theridiids.

It is extremely difficult to assess the importanceof a possible loss of 0.43 calories per day to thehost, since this may vary according to the abun-dance, or lack, of prey from day to day.

Robinson, Mirick, and Turner (1969:492) sug-gest that Nephila clavipes does not store food atcapture sites in order to prevent it from beingstolen. If food were stored at these sites it mightbe particularly vulnerable to attack by kleptopara-sites whose activities would be difficult to detectbecause of the large area of the web. Our studiesshow that Nephila maculata still supports a con-siderable number of these parasites, even though,like N. clavipes, it stores food only at the hub.

NUMBER 149

ENERGETICS

From the dry weight of prey residues it is pos-sible to calculate, albeit crudely, the approximateenergy input to each "average" spider. We knowfrom our own tests that araneids can absorb up to78 percent of the wet weight of prey, such as do-mestic crickets. This leaves a minimum of 22percent to be accounted for in undigested material,unabsorbed water, and undigestible skeletal mate-rial. If we, therefore, assume that 20 percent of thewet weight of an insect is accounted for by undi-gestible cuticular structures, we can calculate backfrom our figures for the dry weight of prey-remainsto the original wet weight. We believe that thedegree of approximation involved is not greaterthan that involved in many calorific input calcula-tions, and that an assumption of 20 percent of thetotal wet weight for exoskeleton errs on the generousside. If the total dry weight of prey-remains for theyear, 123.544 grams, represents 20 percent of thewet weight of the insects from which they werederived (residual water having been driven off bythe drying process), then the original prey weighed617.7 grams. If we assume that 60 percent of thisis water (one of the widely used crude approxima-tions), then the dry weight of skeletal material plusfood is 247.08 grams. Deduct the weight of skeletalremains and the sources of calorific input become123.5 grams. If we assume an assimilation efficiencyof 46 percent the result is 56.8 grams per year dryweight intake, or 5.68 grams per spider, yielding0.015 grams per spider per day. Assuming that in-sect material averages 5000 calories per gram dryweight this gives 75 calories per spider per day, orfor a 3-gram spider, 25 cal/gram. This is less thanwe calculated for Argiope argentata—40 cal/grams(Robinson and Robinson, 1970a:354). However, inview of the approximations at several stages in thecalculation the differences are probably not im-portant (Macfadyen, 1963:234, gives a metabolismfigure of 27 cal/grams for Araneae).

PHENOLOGY

Spiders that have been studied from the stand-point of ecology have, so far, been highly seasonal(North Temperate studies) or present throughoutthe year but with a distinct period of reduced num-bers and reduction of prey intake during a marked

seasonal shift in climate (Argiope argentata atBarro Colorado Island). The Wau area is clearlyone where year-round activity is possible and in thetwelve months covered by our study no major varia-tions occurred at micro- or macro-climatic levels,at least within the study area. Adult and immaturefemale N. maculata were present throughout theyear and so were males. Egg-laying occurredthroughout the study period and at no stage didprey captures fall to a level that suggests seriousshortage of food. This is in marked contrast to thesituation at Barro Colorado where the spider pop-ulation fell to nil at one period of the dry seasonand there was a very pronounced fall in preycaptures preceding this.

Since the species is markedly seasonal in otherparts of its range (Thakur and Tembe, 1956) itseems reasonable to assume that the ultimate factormust be climatic. It is interesting to note that inthe woodland fringe habitats, represented bytransects 1 and 2, adult spiders were absent forlarge portions of the year. On the higher altitudetransect 3 (forest fringe) no adult female Nephilamaculata were recorded during the entire year al-though there was a regular occurrence of spiderlingsand immatures (Robinson, Lubin, and Robinson,in prep.). Since there were no major changes in thephytophysiognomy of these areas capable of affect-ing adult web sites during the study period, we areinclined to believe that they are susceptible togreater fluctuation in climatic conditions than thewoodland and forest habitats. They are thus afringe area for those spider species that requirewoodland or forest conditions. When conditionsfavor the survival of adult and immature N. macu-lata in the fringe areas, conditions in the adjacentforest should be optimal or nearly so. This hypothe-sis is borne out by the fact that the massive in-crease of immatures and males on transect 1, coin-cides with a similar increase in males, and maleson females, in the study area, and furthermoreoverlaps the period when the spiders in the studyarea were just past the period of maximum preycaptures (by weight). This period was one of rela-tively high rainfall.

Despite the overall lack of major seasonality inthe biology of N. maculata at Wau—reflected in thepresence of adults throughout the year, continualmating and egg-laying, and the results of the col-


lection of discarded prey—there arc temporal fluc-tuations in the composition of prey and the pres-ence of males that are of interest. The fact thatthere is a marked seasonality in the occurrence ofmelolonthids in the prey traps may be associatedwith the fact that the immature stages of theseinsects probably occur in grassland soils where thevariations in moisture content may be more ex-treme than in the woodland areas. Cicadas tooseem to occur most frequently in the wetter months.Other examples of seasonality in prey compositionhave been noted earlier (pp. 19-21). These and thevery great increase in the number of males foundon females' webs during the wetter months may befluctuations peculiar to the particular 12-monthperiod in which we carried out our study and muchmore extensive studies are, of course, required toelucidate long-term trends.

In considering the relationship between the va-riations in the biology of Nephila maculata thatoccur over time, it became apparent that attemptsto correlate them with simple climatic and/or bioticfactors are fraught with complications. Some ofthese are detailed later in the "Discussion."

Behavior

In the section on predatory behavior we haveincluded details of comparable behavior patternsthat we have studied in other species of Nephilaand the related genera Nephilengys and Herrenia.This inclusion of comparative material was carriedout to avoid the fragmentation of an essentiallysynthetic approach to the predatory behavior ofthese species. Such a procedure has lead to somerepetition, particularly since it seemed useful andnecessary to include cross references to other speciesin both the section on behavior units and that onbehavior sequences. We have included a final com-prehensive section to tie together the data on pred-atory behavior.

TERMINOLOGY

Except in the cases of newly described behaviorunits, terminology follows that of Robinson (1969)as expanded by Robinson and Olazarri (1971).

The term "courtship" is here used to includeall the regularly occurring interactions between the

adult male and the adult female that precede theact of insemination. Morris (1956:128) has definedcourtship as "the heterosexual reproductive com-munication system leading up to the consummatorysexual act," including in this definition precopula-tory behavior between the two sexes. The term"mating" is applied to varying ranges of sexualbehavior by different authors. With specific refer-ence to spiders Alexander and Ewer (1957:311)seem to use the term to cover both sperm induction(charging the pedipalps with sperm) and insemina-tion, although confusing matters by later referringto the "actual mating." Bristowe (1941:461-502)includes sperm induction, courtship, and copula-tion under the general heading of mating habits.Our inclination is to treat the term as synonymouswith copulation but to use the latter term in de-scriptions.

SEXUAL BEHAVIOR

Finding the Female

We have no data on how maies find females. Itseems reasonable to assume that they could simplylocate the females by locating their webs. Certainlythey tend to aggregate on these webs in considerablenumbers, and even spend long periods of time onthe webs of subadult females. Bristowe (1941:467)concludes "most, and probably all, male spiderspossess a sense which enables them to recognize thefemale's proximity without seeing her . . . [bytouching] the web she has spun either in the formof a snare or a retreat cell."

Our observation that males will remain in theweb of a subadult female after locating it (this isalso true of males of Nephila clavipes, Argiopeargentata, and other Argiope species) suggests thatsome property of the silk, rather than a specificchemical property of a female in reproductive con-dition, may mediate the first stage in the locationof a potential mate. On the other hand it remainspossible that all females produce a chemical signalof some kind, and that this is species-specific. Inthe course of constructing a web this substancecould be transferred to it and thereby label thestructure. Males will remain on a web for severaldays after the female has left it. This, like the simi-lar behavior of kleptoparasites, could be adaptive

NUMBER 149 35

since females often return to the same web afteran egg-laying excursion.

Courtship

The whole subject of courtship in spiders hasbeen the subject of considerable controversy, someof which involved hair-splitting quibbles aboutterminology. Bristowe and Locket (1926) proposeda theory of courtship that involved a dual func-tion; suppression of predatory behavior in the fe-male by "male recognition" and, stimulation of thefemale to the point of accepting copulatory at-tempts. Savory (1928) attacked this theory as beingtoo complex and attributing powers to the spiderthat it probably does not possess. He proposed toregard courtship as a chain of related instinctiveactions in which predatory urges become suppressedand physiological changes are induced in the par-ticipants. Crane (1949) takes an essentially simi-lar point of view on the main issues. Platnick(1971) argues that all these authors rightly stressthe role of courtship in suppressing cannibalistictendencies, but present courtship as a one-sidedactivity by which the male affects the female.Actually Savory (1928:220) noted that "the malehimself becomes more stimulated as courtship pro-ceeds." Platnick (1971) applies a releaser theoryderived from Tinbergen (1954) to spider courtship,suggesting that the behavior has four functions:synchronizing mating activities, orienting the in-dividuals, suppressing nonsexual tendencies, andinsuring species-specific mating. Morris (1956) inan extensive treatment of the function and causa-tion of courtship in vertebrates (mainly fishes)suggests that the conspicuous, all-absorbing natureof many courtship ceremonies makes them danger-ous to the participants and they must, therefore,have a strong selective advantage. He suggests fourfunctions: finding a mate, finding a mate of theright species, stimulating the mate, and synchroniz-ing reproductive arousal.

These categories differ from those of Platnick inincluding mate attraction (finding a mate) and notincluding the suppression of nonsexual tendencies(although that is implied in Morris's category ofarousing the mate). Bastock's (1967) survey ofcourtship suggests that there is no single answer tothe question of function and that this varies with

the circumstances. It is clear from Morris' (1956)description that attracting a mate may be accom-plished by using signals effective in one or severalsensory modalities. One of the most valuable ideasto come out of the reaction chain views of court-ship is that these, by the number of steps that theyinvolve, are fail-safe systems of insuring that matingoccurs between members of the same species (Mor-ris, 1956:130).

Descriptions of the courtship of Nephila speciesare limited. As far as we are aware they includestudies of Nephila madagascariensis by Bonnet(1930), Gerhardt (1933), Charezieux (1961), andsome observations on Nephila maculata by Fischer(1910a,c), Hingston (1923), and Thakur and Tern-

be (1956). There are also some notes on the court-ship of Nephila plumipes published by McKeown(1963).

Fischer (191Oa:527) describes the insertion of theembolus after the male had struck the genital aper-ture "by alternate, rapid pecks of the palps, muchas the fingers strike the keys of a typing machine,"a graphic and accurate description. He did not seethe process of sperm induction or any courtship.He notes that in one case there were five males on aweb and that a male climbed stealthily onto theabdomen of a female only to be brushed off by thefemale as he approached the "vulva." Hingston(1923:74) describes the males, the position theyadopt on the female web, their dependence onfood caught by the female, and their tactics inescaping from females that chase them. He alsodescribes a highly specialized position that thefemale adopts during copulation. In this the femaleleans away from the hub of the web by releasingall the legs on one side "the object being to layherself open so as to receive without obstructionthe advances of her mate."

Thakur and Tembe (1956:331) note that themales aggregate on the webs of females, and remainthere until the female is "ready for copulation."These authors give no details of the approach be-havior nor do they describe the act of copulationbut they do note that "the male does construct aweb on the back of the female, but this is notused for the transference of seminal fluid from thegenital organs to the receptaculum seminis situatedin the palpus." This paragraph shows that theseauthors saw, but largely ignored, a behavior that


constitutes a phase of courtship that is of longduration in most of the cases that we have seen.

Charezieux (1961:375) refers to a long periodspent by the male in tentative, hesitant, approachbehavior. This the "travaux d'approche" may last30-36 hours. During this time the male repeatedlyapproaches the female across the web, touches theapex of the female's abdomen and retreats, even-tually climbing onto the abdomen of the female,moving along its ventral face and copulating. Inthe approach phase the male was seen to vibratethe female's web "as if to check the responses ofthe female from a distance" ("comme s'il voulaitse rendre compte a distance des reactions de lafemelle.") Charezieux (1961:378) also states that thefemale eventually becomes torpid.

Mrs. Lowry's notes (in McKeown, 1963:147-148)on Nephila plumipes include the following points:the male may live on the female's web for weeksat a time; the male moves to the hub when thefemale vacates it and moves away when she returns;and he approaches the female along her drag linewhen she is feeding. The male eventually climbsonto the ventral face of the abdomen but retreatsif the female shows signs of activity. Lowry even-tually saw copulation occur when the female wastorpid and after the male had tapped her body inseveral places.

Our observations on the courtship of Nephilamaculate show that it contains elements of thepatterns described by these authors. Confusion overdescriptions of Nephila courtship may arise, notonly because of observational problems but alsobecause, as we discovered, some stages of courtshipmay be omitted by some males which behave inan "opportunistic" manner. The significance of"incomplete" courtship is discussed later. The bestcase of this that we have documented is the onewhere the male approached a female that was inthe process of molting to the adult stage, and matedwith her while she was still "drying out" and hang-ing from her cast skin. Approach in this case wasaccomplished in less than one minute and insem-mination was over shortly afterwards. We have theentire sequence recorded on film. Crane (1949:176)records a similar case in salticids; the female laterproduced fertile eggs.

Other examples of opportunism include behaviorsequences where the male climbs onto the female

and copulates while she is feeding, and omits thetypical precopulatory behavior. We have two rec-ords of this type of mating and the absence of silkon the body of the female confirmed that the "nor-mal" precopulatory behavior had been omitted.Figure 17 shows a male in the process of copulationwhile the female is feeding, but in this case themale had carried out the lengthy normal precopula-tory behavior.

In all the other cases of courtship that we sawthe procedure was essentially the same, althoughthere were some differences of detail. We have neverseen the entire process from start to finish, buthave seen all the stages and filmed them on manyoccasions. Essentially there are three fairly distinctstages: approach and movement onto the female,complex silk deposition on the female, and move-ment toward the cpigyne leading to copulation.

The approach stage may occur when the femaleis at the hub, where she may or may not be feeding,or it may occur at a prey capture site when thefemale is attack-biting the prey. The male ap-proaches the female on either side of the web, i.e.,either from the upper surface "facing" the femaleor along the lower surface on which she is standing.If the male arrives at the female on the uppersurface he shuttles through the web to the side onwhich she is located. Then follows a period oftouching which may vary in interval from a fewseconds to several minutes. At this time the malemay touch the tarsi of any of the female's legsand/or her mouthparts and pedipalps. This phaseends when the male climbs onto the female or re-treats. Once on the female the second stage com-mences.

The second stage of laying down silk threads onthe female may occur intermittently for very longperiods and the male may spend the greater part ofseveral days standing on the female. During thistime he periodically indulges in complex "weaving"or thread deposition activities. We have one recordof two hours continual activity at this stage, whichwas the longest period of uninterrupted observationthat we carried out. To observe the details wewatched the pair through the lens system of a cinecamera equipped with a telephoto zoom lens, towhich we attached a supplementary close-up lens.This allowed us to watch the pair from a distance,and, at the same time, vary the field of vision from

NUMBER 149

FIGURE 17.—Profile view of copulation in Nephila maculata. One em bolus of the male is inserted(A) . The gusset of silk that the male has attached from the abdomen of the female to herthorax is dearly visible (B) . Silk can also be seen on the leg bases.


one occupied by the entire female to a narrow viewof only part of her body. Such observation is verytiring and two hours was really beyond the limit.The notes on this session are given in Appendix 1.Essentially the male lays down silk threads betweenthe bases of the female's legs and between theprosoma and the dorsal basal edge of the opistho-soraa. Nearly all of the bouts of silk deposition,observed during the 2-hour session, were initiatedby the male moving down from dorsal surface ofthe female's abdomen, where he had been standingmotionless. Before moving onto the thorax themale attached silk to the raised fronto-dorsal edgeof the abdomen. The distinct dabbing movementsof the male spinnerets against the female markedthe start of a deposition bout. The male attachedsilk lines between the abdominal edge and positionson the dorsal surface of the prosoma. He returnedfrequently to the abdomen to repeat the process.Threads were placed on the opisthosoma at variousdistances from the raised anterior edge of thestructure and then stretched to points on the pro-soma at varying distances from the waist. Theywere stretched parallel to the long axis of thespider and also at angles such as left opisthosomato right prosoma. Mixed in between these move-ments were those involving the leg bases. In thesethe male moved either from the opisthosoma, or,more frequently, from the prosoma, onto the coxa-trochanter region of one or other of the eight legs.He then attached lines to the leg base immediatelybeneath it. Attachments were made to the upperand lower surfaces of the basal joints and fromthese to leg bases on both adjacent and oppositeoutsides of the prosoma. In the latter case thelines traversed the dorsal surface of the prosoma.

The most interesting postures adopted duringthis lashing process occurred when the male, fromabove, passed lines onto the lower (ventral) sur-faces of the legs. To do this the male would archhis body and twist it sideways between the legs sothat it moved like a shuttle. On occasion the malemoved entirely beneath the leg bases while attach-ing threads. The results of this silk deposition canbe seen in Figures 17 and 18, which are of two differ-ent adult females. The view in Figure 17 shows thelateral profile of the two spiders (male + female)and, on the female, a sheet of silk threads can beseen curving from the anterior edge of the opistho-

soma to the posterio-dorsal region of the prosoma.Also visible in this photograph is a complex ofthreads extending well up onto the femur of leftleg n and other threads connecting the base of rightleg i to the basal region of the right chelicera.Figure 18 is a dorsal view of a pair of Nephila andshows silk lines between all the leg bases and alsothe complex of threads on the dorsal surface of theprosoma. Figure 19 shows another pair of spiderswith the male in the act of attaching silk to theleft lateral aspect of the female's opisthosoma.Some lines are attached from the leg bases to theventro-lateral margins of the prosoma, but the malespends very little time beneath the female duringthe process of silk deposition.

During the process of intermittent silk depositionthe male rests fairly frequently. The rest attitude isalmost always assumed on the dorsal surface of thefemale's abdomen and usually above the raised an-terior edge (where there is a fairly conspicuous paleline). While they are resting males may tap at thebases of the female's legs, or hold onto these withthe feet of legs i and II. Contact in these cases maypossibly be with the silk between the leg bases ofthe female. We have records of males assumingspecial postures when the female moves about ina precipitate manner (as she may do during a pred-atory excursion). One male consistently respondedto the movement of the female by flattening himselfagainst the steeply sloping anterior declivity of theabdomen and in this position could be seen clingingto the silk "shrouds" on the thorax. Such a positionwas also adopted by other males for which we havea less complete record. We are inclined to believethat this is the most secure and sheltered positionthat the male can adopt and that it may help tosecure him against disturbance by the female.

We saw few instances in which the female ap-peared to respond to the activity of the male.These rare responses only occurred when the malewas on the ventral surface of the legs. In this situa-tion several females made brushing movementsdirected at the males (by bringing the tibia of athird leg along the underside of their prosoma).These movements were similar to those made inapparent response to the activities of kleptopara-sites. Copulation follows the stage of silk deposi-tion, and approaches to the female genital orificemay occur intermittently during silk deposition

NUMBER 149

FIGURE 18.—Male on dorsal surface of the abdomen of a temale Kephita maculata. Silk linesare visible on the thorax of the female leading to the leg bases and abdomen.


FICIRE 19.—Male Xephila maculata in the act of attaching silk to the right fourth leg of thefemale. Silk produced by previous deposition activities is clearly visible on all leg bases. Notethat at least five kk-ptoparasitcs are visible on this photograph.

NUMBER 149 41

bouts. We have film records of three successfulcopulations and a still photograph of another (Fig-ure 17).

Approaches to the region of the epigyne are madeby two basic routes. The male most frequentlymoves on the dorsal surface to the apex of theopisthosoma and then walks forward down theventral surface, halting just behind the margin ofthe epigyne. There is a tuft of longish stiff hairsimmediately anterior to this furrow and these mayact as a marker for locating the appropriate region.The male can also circle around the anterior regionof the opisthosoma to reach the same point, andfrequently retreats from the ventral surface by arapid circling move. When descending the ventralface of the female opisthosoma he is clearly be-layed on the drag line which is attached somewherenear the abdominal apex. Presumably the drag lineis also used when circling, but it is not then clearlyvisible.

Copulation

A typical copulatory position is shown in Figure17. From this it is obvious that only one embolus isinserted at a time and the other is simultaneouslyraised above the long axis of the male's body. Theswollen distal portion of the pedipalps are, at thisstage, black and shiny, the emboli are long andextremely narrow in their distal region. The em-bolus of the left palpus is clearly visible in itsinserted position in the photograph. Insertion maybe very brief and is often a matter of seconds. Suchbrief insertions may represent unsuccessful at-tempts. At other times they may last for over aminute. The removal movements are conspicuous.The structures sometime appear to stick so thatthere is a distinct recoil upon the embolus beingfreed. Insertion is preceded by the spider beatingupon the female genital region with very rapidalternating movements of the pedipalps. Analysisof film suggests that this beating is done with theemboli reflexed backwards.

The palps are used alternately, several times,with pauses between bouts of insertion. Bristowe(1941:491) suggests that this technique is used bythe majority of spiders.

Charezieux (1961:377) shows a pair of Nephilatnadagascariensis in a copulatory position exactly

comparable to that adopted by N. maculata. Henotes (1961:378) that the male beats on the epigynewith his palps ("II frappe alors de ses bulbes copu-lateurs l'epigyne, les actionne rapidement tous lesdeux"). He does not state whether one or two palpsare involved at the same time, although this may beimplied. Charezieux also refers to the brevity ofcopulation and also to the fact that it may berepeated several times.

The female makes no attempt to attack the maleduring or after copulation (in marked contrast tothe behavior of several New Guinea Argiope spe-cies; Robinson and Robinson, in prep.). Thakurand Tembe (1956:331) state that copulation occursat irregular intervals for a day or two, and that thefemale does not attack the male, but that "he justdies a natural death, apparently due to exhaustion."They do not describe the act of copulation.

Discussion

The most interesting aspect of the courtship be-havior of Nephila maculata is the complex threaddeposition, which the male normally undertakes.This behavior was not described by Fischer (1910a,b,c) Hingston (1922a,b,c, 1923) or Thakur andTembe (1956), although the latter authors presum-ably are referring to the results of this processwhen they state that the male builds a web on theback of the female (1956:331). Preliminary studiesthat we have carried out on Nephila clavipes suggestthat silk binding is absent from the mating be-havior of that species.

It is not described for N. madagascariensis, andpresumably would not have been missed by theauthors that studied this species. As far as we areaware there are no published accounts of otheraraneids indulging in such complex silk depositionduring courtship. Bristowe (1929:320-321), how-ever, records that during the courtship of Metascgmentata the male may begin to wrap the femaleas though she were prey. The female tears herselffree and Bristowe assumes that the behavior is theresult of multiple motivation and that althoughthe female is not restrained by the silk she maybe stimulated by the male's actions. Bristowe (1929)and others following his generalizations based on alimited number of observations on a limited num-ber of species (e.g., Platnick 1971:40-41), state


that courtship in araneicls involves the building ofmating threads onto which the female is coaxedduring mating. Descriptions of the construction ofmating threads, e.g., by Meta segmentata (Bristowe,1929:320-321), mention that the male walks back-wards and forwards over the female trailing aseries of threads that are eventually to become asingle composite copulation thread (=matingthread) on which copulation occurs. It is not clearfrom the description exactly how the copulationthread is oriented in relation to the female, wheth-er it is dorsal or ventral. If it is laid down by themale walking across the dorsal surface of the spiderthen there is some similarity between this and thebehavior that we have detailed for Nephila macu-late Our own observations on the mating behaviorof Argiope aemula show that the male moved veryextensively over the dorso-lateral surfaces of thefemale piior to copulation (Robinson and Robin-son, in prep.). These movements had some of thecharacter of the Nephila maculata binding move-ments but did not include discernible silk attach-ment movements. The existence of such wanderingmay be of interest in considering the evolution ofcourtship behavior in araneids. In passing it isworth noting that in Thomisids the male may, asin the case of Xysticus lanio C. L. Koch, tie thefemale to the ground with silk. This process, de-scribed by Bristowe (1931:1406) is presumably notentirely to restrain the female since she frees her-self after copulation has occurred.

Perhaps the most interesting question about thebinding behavior is: what is its function? At firstsight it would seem logical to assume that it func-tions in some way to restrain the female so that themale can safely copulate. This is suggested, albeitindirectly, by the fact that the behavior is omittedon occasions when the male can catch a female atthe molting stage or while she is feeding. Despitethe fact, however, that the leg bases are the subjectof a great deal of binding activity, the female iswell able to rush off on predatory excursions fromthe hub after the silk has been heavily deposited.Her running ability does not, therefore, seem to beimpaired. It may be that the binding, particularlythe deposition of threads from the opithosoma tothe prosoma, on the dorsal surface only, preventsthe female from bending at the waist. The malewould then be able to copulate without the danger

of being picked off by the female. This explanationof the binding behavior only occurred to us afterwe had seen Argiope aemula mate, and in the proc-ess pick off the male from her ventral surface. Thisobservation was made at the end of our stay in NewGuinea and we were not then able to carry outany experiments on the efficacy of the binding inpreventing the Nephila female from moving at thewaist. However, we have since carried out experi-ments with the females of Nephila clavipes andgluing threads on dead females effectively preventsbackward ventrally directed movements of theprosoma. This evidence is certainly not critical butis suggestive. The unrestrained spider is able tobend at the waist to an angle of over 90 degreeswhen not in contact with a substrate (presumablyby contracting muscles associated with the waistjoint) and further than this when it can use thelegs to pull the thorax towards the abdomen. Itcan also achieve a more than 90 degree flexure byallowing the abdomen to drop toward the thoraxin a ventral direction, a situation in which theaction of internal muscles is presumably supple-mented by gravity. This degree of bending could besufficient to enable the female to pick off the malein certain circumstances, i.e., by seizing one of thelegs projecting beyond the waist joint onto theprosoma of the female. This can be visualized byexamining Figure 17.

If the act of mating were to be accomplishedrapidly, once the male had assumed a copulatoryposition, there would not seem to be any adaptiveadvantage in insuring his postcopulatory survivalby an elaborate precopulatory behavior pattern. If,however, a binding function of silk depositioninsures that the male can safely achieve repeatedinsemination bouts it could be of survival value.

The immunity of the male to attack by thefemale, a point that we have observed in all thematings that we have seen, could be a direct con-sequence of the binding behavior as suggested aboveor it could be that the female is not restrained bythe silk but by some change in her internal stateconsequent on the males' behavior. There are alarge number of cases of complex courtship patternswhich have the effect (among others) of loweringhostility between the sexes or reducing predatorydrives that might otherwise be directed against themale. In this particular case the restraint is not

NUMBER 149 43

exercised during the approach stage, which mayalso be one during which the male is at risk, butactually during the process of copulation. Signalingsystems may exist that enable the male to identifyhimself as nonprey to the female during the ap-proach stage. The web-strumming that the male ofNephila madagascariensis performs during his ap-proach (Charezieux, 1961) would seem to belongto this category. We did not observe such behaviorin the case of Nephila maculata. (We suspect thatthe males of N. maculata may be light enough notto trigger a predatory response as they move acrossthe web.)

There are still other possible explanations, interms of function, of the complex binding be-havior. The structure could be a sperm web. Weconsider this unlikely because it is constructed in amanner which suggests unnecessary complexity forsuch a function. A sperm web would presumablynot need to encompass the leg bases in such athorough way, nor would the complex diagonalmembers laid down from opisthosoma to prosomaseem to be necessary. However, since we did notobserve sperm induction we cannot entirely ex-clude this possibility. A further possibility is thatthe structure provides a foothold for the male sothat he can cling securely to the female until themoment for copulation arrives. This too seems afairly remote possibility since the male could,presumably, achieve a secure position on the femaleby attaching his drag line to some part of her bodyand gripping its far from smooth surface with histarsal claws. At the most he could achieve a verysecure foothold with a modest skeleton structure ofsilk lines rather than the complex structure that hedoes build.

The foothold theory gains no direct support fromour observations on Nephila clavipes. These showthat the male is able to move about the surface ofthe female with facility for long periods, and re-main there during fairly violent activity (duringprey wrapping) without a silken foothold structure.A primary foothold function for the silk couldthus be justified only by assuming that the surfaceof Nepliila maculata afforded considerably less foot-hold than that of Nephila clavipes.

There is no reason to assume that the silk struc-ture has only one function. Therefore, it could

function as a restraining "harness," a platform offootholds, and a sperm web.

All the authors who have described Nephila mat-ing refer to the passivity of the female duringcopulation. It seems worth reiterating that court-ship activities may function principally to inducesuch a state, which may be due to internal ratherthan external restraint.

The proportion of matings that we observed inwhich the stage of silk deposition was omitted donot raise major problems in explanation. Be-havioral investigation of courtship patterns over awide range of organisms has shown that stages ina species-specific courtship pattern may be omittedin a proportion of instances, while at the sametime successful mating ensues. Such exceptions areoften explained in terms of variations in individualreadiness to mate or reproductive drive, etc. Ethol-ogists argue about the validity of drive concepts(Hinde 1966:139-147), but accept the fact thatanimals do not always respond to the same stimulussituation in the same way. Morris (1956) showshow courtship patterns that can be idealized intoreaction chains do, in reality, have a degree ofvariability in both the presence of, and orderingwithin, their component units of behavior. Thusone explanation for the opportunistic males ofNephila maculata (which copulate without firstgoing through the normal prior phase of silk depo-sition on the female) would be that they are highlymotivated to copulate and that this internal statein some way suppresses the preliminary elements ofbehavior. Alternatively males could be regarded ascapable of recognizing situations in which certainstages of mating can be omitted as unnecessary.Our observations do not provide any basis for de-ciding between these and other possible alternatives.

A further general problem remains. Alexanderand Ewer (1957:312) in their discussion on theorigin of mating behavior in spiders argue that the"double process" of spider mating (stages of sperminduction and sperm transfer) confers a selectiveadvantage in that "it permits the male to preparefor the mating beforehand, and, when in the prox-imity of the female, to be as quick as possiblewith the actual insemination." They postulate thatsperm induction may have evolved through anintermediate stage of spermatophore production, asin pseudoscorpions and scorpions, and that the pro-


tospiders may have produced spermatophores thatbecame redundant when web-spinning arose. Toour view, however probable this may be, it doesnot bear on the problem raised by the matingbehavior of Nephila maculata, where although themale may be prepared to insemination beforehandhe nevertheless spends a great deal of time actuallyin contact with the female before copulation occurs.

PREDATORY BEHAVIOR

Behavior Units of Nephila maculata

In this section we describe the behavior units inorder, from those involved at the initiation of apredatory excursion to those preceding the onsetof feeding at its successful conclusion.

BEHAVIOR PRIOR TO CONTACT WITH THE PREY.—

Predatory Position at the Hub: The normal waitingposition at the hub is shown in Figure 2. Note thatall the leg pairs are separate (compare with thepredatory position of Argiope argentata describedby Robinson and Olazarri, 1971:2) and extended ina more or less bilaterally symmetrical pattern. Nor-mally all the tarsi are in contact with web members,but under the influence of strong incident sunlightthis position may be considerably modified (p. 64).The spider rests on the underside of the web, atthe hub, with the body more-or-less parallel to theplane of the web. The drag line is attached to theweb just above the hub. The waiting stance of allthe species of Nephila that we have seen is funda-mentally similar, even to the disposition of thelegs and the angle that they form to the body.

Plucking-. After there has been an impact of someobject with the web the spider orients towards thegeneral direction of the prey (or other object) andthen may either approach directly or make pluck-ing movements. These are usually carried out withboth legs i, which may be extended almost parallelto each other in the direction of the prey (Figure20). In effect the spider jerks the web members onwhich the tarsi of legs i are placed, without break-ing tarsal contact with the web. Such pluckingmovements may be repeated as the spider approach-es the prey across the web. As in the case of Argiopeargentata, plucking movements occur most com-monly if the prey is immobile after it strikes theweb. Approach plucking may be abandoned andthe spider accelerate towards the prey if the latter

becomes active after a plucking bout. Radii lyingwithin a fairly wide sector around the prey may beplucked before the spider leaves the hub on a pred-atory excursion. When this happens the spidereventually approaches along the radial directionleading to the prey. We have never seen the spiderhave to make directional corrections to the left orright after it has set out on an approach pathwayinitiated by plucking. We have seen such correc-tions made when the spider has run towards theprey without prior plucking. A full discussion ofthe possible function (s) of plucking is given inRobinson and Olazarri (1971:5-6) and their conclu-sions are supported by our observations on Nephilamaculata. This species often carries out wide-sector plucking after returning to the hub followinga predatory excursion. It will also perform pluckingmovements from a prey capture site if other preybecome entangled nearby. Such prey may then beapproached without the spider first returning to thehub. We have not seen Argiope species behave inthis way.

Approaches to the Prey: Approaches to prey maybe rapid or slow and deliberate. Rapid approachesare made by the spider running across the under-surface of the web and occur most frequently whenthe prey is active. In such cases the spider may evenovershoot the position of the prey along the radialpathway by several inches (this occurred in 3 outof 20 presentations of live dragonflies to adult fe-male spiders). After overshooting, the spider eitherturns or backs up to the prey. Approaches maystart slowly and accelerate if the prey becomesactive during the approach. Slow approaches maybe punctuated by pauses during which the spiderstops and plucks, or may be continuous, whenplucking with one leg i may form part of a strideforwards.

Very large prey may be approached slowly evenwhen they are active. When the spider nears suchprey it may raise legs i and one or both legs II offthe web and flex them dorso-posteriorly. The re-sultant stance looks very defensive. It may alsoinvolve a dorsally directed concave curvature ofthe entire body brought about by flexion at thewaist. A similar posture was noted by Robinson andMirick (1971:128) in the case of Nephila clavipes.A very similar posture is adopted if the spider istouched lightly from above when it is at the hub.

NUMBER 149 45

FICURE 20.—Female Nephila mmculaU plucking as the advances toward prey (not shown).


All approaches to prey involve the production of adragline by which the spider remains attached tothe web at some point. The spider at the hub hasa dragline attachment to the web at the hub region,but it frequently reattaches the dragline beforemaking a predatory excursion. It seldom attachesthe dragline at other points on the approach path-way. Exceptions to this occur when the prey isparticularly heavy, then the spider can be seen tomake the characteristic dabbing movements of thespinnerets against the web, as it closes with the prey.

Not all objects striking the web are approachedand particularly heavy objects may elicit an escaperesponse. In the escape response the spider runsup from the hub to the upper frame members, to-wards nearby vegetation. Escape responses aredescribed in detail later.

BEHAVIOR ON CONTACT WITH PREY.—Touching

and Palpation: In most cases when the spider con-tacts the prey it touches the prey with one, or both,legs i and continues its approach until it is stand-ing with the anterior prosoma above part of theprey body. The tarsi of the leading legs touch theprey and the approach gait is unbroken until thespider assumes the attack stance. However, insome cases, when the prey is touched with the tarsiof the outstretched leading leg(s) the spider haltsso that it is approximately a leg i length away fromthe prey (8-9 cm). It may then back up the webuntil it is more than this distance from the prey,and wait out of contact, or even retreat to thehub. We saw such behavior only in the case of veryheavy prey (over 1.5 grams) and then only in a pro-portion of cases. When halted the spider mayraise legs i and II off the web and assume the pecu-liarly curved attitude described above. Eventuallythe spider either retreats to the hub (infrequently)or advances slightly and taps the surface of theprey with the tarsi of legs i. The tapping move-ments, are slow, deliberate, and often repeated.Verbatim notes of attacks on large acridiids inwhich bouts of tapping occurred are reproducedin Appendix 2.

Tapping is often followed by prolonged tarsaltouching and then most frequently by the bite andback-off attack behavior described below. However,in a proportion of cases, touching is followed by aslow advance to the assumption of an attack stanceover the prey and in this case, as the spider moves

forward, the pedipalps may be extended and con-tact the prey during the advance. In some casesthey may be lowered onto the surface of the preyonly when the spider assumes its attack stance. Thispedipalpal contact (=palpation) is almost certainlymade during the normal touching approach but isnot then conspicuous because the advance to anattack posture merges without pause into attackitself and a separate palpation stage cannot be dis-tinguished. Exceptions to this have been observedin attacks on pentatomids and reduviids whenthere is sometimes a pause between the assumptionof an attack stance and attack. We presume that theodors produced by these insects may affect thespider and contribute to the abnormal pause thatallows palpation to be seen as a distinct stage.Robinson (1969) found that Argiope argentatapalpated prey that were covered in spider silk in avery distinct way and assumed that they did thisbecause they received an unusual stimulus fromsuch prey.

ATTACK BEHAVIOR.—Nephila maculata attacks allprey with the jaws and does not use silk as anattack weapon. Three types of attack are discern-ible.

Long Bite: The greater proportion of prey isdealt with by a sustained or long bite. The spideropens the chelicerae (by lateral movement), stoopsand inserts the fangs into the prey. The bite isthen sustained at its initial point of entry for along period. If the spider is dealing with a preythat is vibrating or kicking, the point of insertionof the fangs may be the point of initial contactand therefore apparently random in location. Some-times this bite may be ineffectual, as for instancewhen the spider bites into the wing of a moth andthe fangs simply pass through the structure. Suchbites are corrected and the spider moves the pointof insertion, in steps, until it contacts more sub-stantial tissues. Stepwise movements of the pointof insertion of the fangs may occur in bites that areinitially delivered to the body of the prey, andthese are more difficult to account for in terms offunction and causation. Thus in 12 out of 25encounters with live dragonflies the bite was even-tually sustained at a point on the thorax just be-hind the head. Since bites were moved to thisposition from other locations in 6 of the 12 casesthere could be some element of choice in the exact

NUMBER 149 47

location of the bite on the body of the prey. (Thisregion of the prey plus the thorax in general maybe a particularly conspicuous target in contrast tothe abdomen which is long and very narrow.)

The spider could presumably determine thelocation of this target at the touching stage. Inaddition, the wings are attached to the bulbousthorax and vibrations of the web produced bywing movements could lead the spider to this re-gion. (Four bites were initially aimed at the wingbases). Hingston (1922c:920) states that N. macu-lata always bites the thorax of the prey first; thus,"she behaves as though she knew the anatomy of herprey." Bites at some heavily sclerotised insects maynot penetrate at first attempt and the spider maymove along the prey, attempting an insertion, untila vulnerable region is found. In presentations ofinelolonthid beetles we found that the spiders ap-peared to be unable to penetrate the smooth andcurved elytra. Sustained bites at these beetles wereoften made at the lateral or apical margins of theelytra where the soft ventral surface of the abdomenapparently offered a more easily penetrable bitesite. Attempts to bite into the elytra were madeand the impact of the fangs as they repeatedlyglanced off the cuticle could be heard distinctly.

Bite and Back-off'. As described earlier, and byRobinson and Mirick (1971) in the case of Nephilaclavipes, attacks that are preceded by slow ap-proaches and made after leg raising and prelimi-nary tapping of the prey are often carried out bythe bite and back-off behavior. This entails a rapidlunge at the prey followed by a short bite and sub-sequent retreat. After the first bite the spider re-treats to a position usually within about 10 cm ofthe prey. Lunges, bites, and retreats may be re-peated a number of times. We have a record of 9cycles of attack over a period of 80 seconds whenthe prey was a large melolonthid beetle and oneof 13 such attacks, in 2 bouts, for an attack on a5 cm acridiid. The bites are rarely sustained formore than a few seconds. We suspect that they sel-dom exceed 5 seconds, and most of them are con-siderably less. They are usually carried out withlegs i and II raised back over the prosoma of thebody of the spider and the spider's body arched.Prior to the attack there are often intention move-ments, during which the spider spreads the jawsto a pincer-like extent while swaying forwards

and then backwards on the spot where it is standing.After a series of attacks the spider moves close to

the prey and delivers a sustained bite. This may bepreceded by tapping and palpation. We have seena series of short bites followed by wrapping be-havior without the occurrence of a sustained bite.Even a sustained bite may be interrupted by theprey struggling vigorously.

Seize and Pull-out: Very small and light prey aresimply seized in the jaws and pulled from the web.Thus in the presentation of 50 live blowflies (25-50mg) the bite time was one second or less in 4 casesand this was followed by a pull-out time of similarduration. On a purely arbitrary basis we canseparate this behavior from the other attacks onthese flies where bite times of up to 18 seconds wererecorded (average for all 50 attacks 5.16 seconds).The distinction between a short bite and seizure inthe jaws is thus somewhat arbitrary. A less arbitrarycriterion for distinguishing between the categories"seize in jaws" and "bite" would be possible if itcould be determined whether venom was usedwhen the prey was simply seized in the jaws andpulled from the web. The distinction between sei-zure in the jaws and biting in Argiope and otheradvanced araneids is made operationally and issimple, since prey that are bitten are always wrap-ped at the capture site. This is not the case withNephila species since prey that are given sustainedbites of several minutes may subsequently be re-moved from the web by being pulled out (Robin-son, Mirick, and Turner, 1969; Robinson and Mir-ick, 1971).

REMOVAL OF PREY FROM CAPTURE SITE.—Pulling

Out: At the termination of a biting attack thespider almost always attempts to pull the prey fromthe web. This behavior occurs in all the Nephilaspecies that we have observed and Robinson andMirick (1971; 129) have described it in the caseof Nephila clavipes. Hingston (1922c:920) describesthis behavior as tearing the prey from the viscidattachments. Without releasing its cheliceral hold(i.e., bite) the spider pulls outwards on the prey,often repeatedly flexing and extending its legs inthe process. If the prey becomes freed it is thencarried back from the capture site to the hub. Some-times if it becomes only partly freed, the spidermay hang away from the web wholly or partiallysuspended on its drag line and pass its legs between


the prey and the web. This motion frees the preyof the remaining attachments essentially by strip-ping the viscid element away from the prey in a"teasing" movement. Damage to the web resultingfrom the process is surprisingly slight. In mostcases, a few viscid spiral elements may be brokenand the golden droplets of adhesive stripped offothers, but large holes are rarely produced. Thiscontrasts sharply with the results of the cutting-out behavior carried out on prey that are wrappedat the capture site.

(Records of prey capture sequences given in thenext major section of the description of predatorybehavior show that the pull-out technique succeedswith certain types of prey. With other types of preyit is either replaced, after failure, or, omittedaltogether.)

Wrap in situ: Very large prey, as well as preythat for other reasons cannot be freed by pulling,are wrapped at the capture site while still enmeshedin the web. Such prey are cut from the web, partlyduring the wrapping process and partly when it iscomplete. Nephila species wrap relatively slowly(when compared with species of the genera Argi-ope, Araneus, Eriophora, and others). The spiderstands on the prey and casts skeins of silk onto itby alternate movements of the hind limbs. Thesemovements are easily counted, unlike the veryrapid movements of Argiope species. The prey be-comes covered with a complex of silk skeins butnot completely enswathed in a layer of silk (anothercontrast with the advanced spiders). The prey isnot normally rotated about an axis while silk islaid down directly from the spinnerets, a behaviorfrequently performed by Argiope argentata.

The wrapping bout is interrupted frequently asthe spider cuts entangling web elements that re-sist the passage of the wrapping silk over its sur-face. Cutting is done by bringing the cheliceraedown to the strand or by pulling the silk up to thejaws on the tarsus of a leg, and also by apparentlysnapping it with a foot. Towards the end of awrapping bout the spider may have turned through180 degrees and be hanging with its dorsal surfacepartly oriented towards the web, while the preyis suspended below the web on just one point ofattachment. Figure 21 shows this remarkable wrap-ping stance. The consequence of this style of wrap-ping is that the resultant prey package invariably

hangs below the web. Attack wrapping by manyother species of araneid, on the contrary, often re-sults in a prey package that lies securely in theweb plane and has to be cut out of the web beforetransportation (Robinson and Olazarri, 1971, figs.3,7).

Free-wrapping: As in the case of Nephila clavipes(Robinson and Mirick, 1971:135-136) some preythat are freed from the web by pulling are wrappedbefore transportation. This form of wrapping re-sults in the prey being trussed into a large fairlycompact prey package and is applied to bulky preythat can nevertheless be freed from the web with-out being wrapped in situ. The freed prey may beheld in the jaws until the first skeins of silk arepassed over it by movements of legs iv; it is thenheld with legs in until it is trussed in silk. Thespider is normally hanging below the web planewhile it performs the free-wrapping, in the mannershown in Figure 21.

TRANSPORTATION TO THE HUB.—Carrying in Jaws:Light prey are carried to the hub in the jaws, asis the case with most araneids. Nephila maculatacommonly exhibits a behavior that we have onlyseen occasionally in other Nephila species, and havenot noted in other araneids. Prey from the largearea beneath the hub of the web are carried in thejaws as the spider backs up the web hanging headdownwards. This backing movement is illustratedin Figure 22. In effect the spider fends itself offthe web surface with walking movements of legsi, II, and HI while climbing "hand over hand" upthe drag line with the tarsi of legs rv. In cases wherewe were able to watch the process at close range, itbecame obvious that the spider was in fact gather-ing up quite considerable lengths of the drag lineinto a loose hank as it made its ascent backwards.The process is fairly slow compared with merelyturning and walking forwards to the hub. Thislatter behavior does occur in the spider's repertoireand ensures a more speedy resumption of its preda-tory position at the hub. Thus in 50 presentationsof live moths 42 spiders backed up the web, and 7turned and walked or ran back to the hub. Theseseven averaged 4.4 seconds for the return journey,whereas the average time for those that backed upthe web was 34.8 seconds (detailed analysis in sec-tion on behavior sequences). What are the possibleadaptive advantages of this behavior? A spider

NUMBER 149 49

FIGURE 21.—Wrapping posture of Nephila maculata (based on film records)


FIGURE 22.—Nephila maculata backing up the web to hub.Prey (carried in jaws) omitted for clarity. Note that thedragline is being ascended backwards and collected in thetarsus of the left fourth leg. (Based on successive frames offilm records).

carrying a load in its jaws and ascending the under-surface of an inclined web has the center of gravityin a position that is more favorable to a trouble-free ascent than a spider walking forwards with preyin its jaws. The question of why N. maculataascends backwards and N. clavipes (for instance)does not is difficult to resolve.

Carrying on Silk: The other basic method ofcarrying prey back to the hub is on a silk threadsuspended from the spinnerets. The suspensorythread may be suported by being held by one orboth legs iv, or may, less commonly, be without

secondary support. Large prey are carried back tothe hub in this way. Because of the slope of theweb, prey carried on silk hang away from the webplane and seldom become entangled in the webduring transportation.

Wrapping in Transit: Prey, particularly thosecarried forwards in the jaws, occasionally becomeentangled in the web during transportation. Whenthis happens the spider makes pulling movements,and, if these fail, briefly wraps the prey and thenproceeds to carry on silk as described above.

BEHAVIOR AT THE HUB WITH PREY.—Carried inJaws: Prey carried in jaws are almost always wrap-ped in silk on arrival at the hub. The spider usuallywraps in a head-up position comparable to thatshown in Figure 21, and turns to achieve thisposition if it arrives at the hub in a different orien-tation. Before wrapping, the spider usually attachesits drag line to the hub silk. Wrapping at the hubis carried out in the slow manner that it characteris-tic of the genus. Small prey, carried from the upperregion of the web, are often wrapped when thespider is in a head-down position (as it will bewhen arriving from above the hub). Wrapping inthis case is achieved by the spider swinging theabdomen into an almost horizontal position withthe anterior prosoma close to the web and theapex of the opisthosoma well away from the web.In this position the long legs rv can make wrappingmovements without fouling the web.

Carried on Silk: When the spider arrives at thehub with prey suspended behind it on a silk threada process of attaching this thread to the hub regionis carried out. The spider dabs the spinneretsagainst the hub silk at the point of arrival, therebymaking the first attachment of the prey to the hub.It then turns until it assumes the head-down posi-tion normally adopted when the spider is at thehub. As it circles to the left or right it attaches itsdrag line, at intervals, to the hub. As it does thisthe leg iv on the side opposite the direction of turn-ing is used to stretch the drag line between thepoints of attachment (Figure 23). The stretchingmovement is essentially similar to the one thatmost araneids use when laying down the viscidspiral between the radii during web building. Werecorded the direction of turns made by spidersduring this rundgang and the number of attach-ment points made with different types of prey.

NUMBER 149 51

FIGURE 23.—Silk attachment during the rundgang. The silkline leading from the prey package has been attached atA, B, c. Between attachments, as between c and D, the spiderstretches the silk with movements of the appropriate legiv, as shown.

These data are given in the section on behaviorsequences.

Prey at the Hub: If the spider was feeding atthe hub before catching the prey with which it isreturning there is some modification of the be-haviors described above. If the previous prey waslarge the new prey will almost always be attachedat some distance from the old, often above it, andthe spider will resume feeding on the old preyafter the process of attachment. If the previousprey was small, and/or partially consumed, the newprey and the old prey are frequently wrapped to-gether into one bundle and the spider then feedson the compound package.

Assumption of a Feeding Posture: After thespider has suspended the prey package from thehub it frequently backs a few centimeters up the

web and reattaches its drag line before moving backto the central region of the hub. Before commencingfeeding the spider may undertake grooming activi-ties in which the legs are rubbed together and par-ticular attention given to the tarsal regions. Thetarsi of legs i and n are almost always drawnthrough the chelicerae, one by one, after a preda-tory sequence, and the other tarsi may also betreated in this way. After such grooming, andsometimes a variable period of inactivity, thespider often plucks the radii over a wide sectorand then reaches out with one of the first legs untilit contacts the thread on which the prey is sus-pended. It hauls on this thread until it can graspthe prey package and manipulate it. The periodof prey package manipulation varies considerablyin duration and involves some or all of the firstthree pairs of legs. The short third legs are usedfor manipulation beneath the spider while theflexed legs I and n operate in front of the prosoma.During the manipulation the prey package ismoved backwards and forwards, rolled over andtwisted from side to side. The spider makes shortbites at the prey as it is manipulated. Eventuallythe spider sustains a cheliceral insertion at onepoint and places legs i and n back on the web intheir normal extended positions; one or both legsHI may still hold the prey, but this is unusual. Atthis stage we assume that the spider is feeding.

Feeding Behavior: We have not watched theentire process of feeding, but examination of dis-carded prey remains allow certain deductions to bemade about the processes involved. The remains ofheavily sclerotized prey items, such as beetles, con-sist of entire exoskeletal shells. These suggeststraightforward suctorial feeding; however, theelytra and membranous wings of beetles are oftenfound separately. We believe that these may benipped off and discarded during the feeding proc-ess. The discarded remains of orthopterans arefrequently in a highly comminuted state suggestingthat at some stage in the feeding process thespider chews up the prey into smaller portions.This process may facilitate the digestion of preymaterial that is otherwise not readily accessibleto enzymes that are simply pumped into the interiorof a perforated but entire insect. Very soft insects,such as flies, can be seen to be kneeded between


the chelicerae during feeding, and are highly com-pressed when dropped as trash parcels.

Behavior Units of Other Species

In Ghana, mainly in the Legon area, we encoun-tered two species of Nephila: Nephila constrictaand Nephila turneri (the latter a large and strikingspider), as well as Nephilengys cruentata. We pre-sented a variety of prey items to the Nephila species,and also to Nephilengys cruentata. We concentrat-ed attention on the latter species since the spider isin many ways similar in appearance to an Argiopeand builds a cocoon-like retreat that is continuouswith the web. We encountered Nephilengys againin Madagascar, at a forest site near Perinet andcarried out further observations there. We firstencountered Herrenia species at Manaas in Assamand later encountered the genus at Periyar inKerala, South India. The spider builds elongatewebs close to large tree boles and from this habitand its appearance we decided that it must be aNephilengys. We eventually found very similarspiders building webs in similar situations at Wau.Specimens of these were identified by Fr. Chrysan-thus as Herrenia ornatissima. Since this genus is inthe Nephila group we carried out studies of itspredatory behavior.

All the Nephila group spiders that we studieden route to New Guinea possessed broadly similarbehavior units. All attacked a wide range of preyby biting. All except Herrenia ornatissima madepull-out attempts after biting and all tapped largeprey from a distance before biting. The Nephilaand Nephilengys species showed the type of biteand back-off attack on large prey that we havedescribed above for Nephila maculata. We were un-able to induce Herrenia ornatissima to attack verylarge prey and did not see bite-and-back-off be-havior in this case. None of the spiders could beinduced to attack prey by wrapping, although wepresented types of prey (acridiids, beetles, dragon-flies etc.) that Argiope species consistently attackin that way. We observed wrapping at the capturesite (after biting) in all the species, and this wasperformed with slow alternate movements of legsiv as in Nephila maculata. All the species, exceptHerrenia ornatissima wrapped prey, on occasion,that had been freed from the web by pulling. Allthe species hung prey at the hub and did not store

it at the capture site. Nephilengys cruentata assum-ed an extraordinary position in its retreat (Figure24) facing outwards towards the web, but with itsdorsal surface towards the web that lies above theretreat. Spiders returning to the retreat with preyenter the retreat head first and turn over and to-wards the entrance when they are inside or at theentrance. This considerably complicates the be-havior of wrapping prey at the feeding site (hub orretreat) and also complicates the process of hangingprey at the retreat entrance (equivalent to hangingit at the hub in the other species).

FIGURE 24.—Relationship of the retreat of Nephilengys cruen-tata to the main orb and the position of legs iv in relation tothe retreat. Note that these legs contact the dorsal surface ofthe retreat which is continuous with the web surface.

Herrenia ornatissima builds its elongate websclose to tree boles, often less than 2 cm away fromthe bark, and rests in a cuplike silken depression atthe top of the web, close to or touching the bark.This again poses problems since at this place thespider only has access to the upper surface of theweb. It is interesting to note that the dorsal sur-face of this spider is beautifully camouflaged andthat it is very difficult to pick out against a back-ground of bark (Figure 25). The specific name may

NUMBER 149 55

FIGURE 25.—Male (above) and female (feeding) Htrrenia omatissima on surface of web. Theweb is built very dose to the surface of a tree and is elongate. The female has a camouflageddorsal surface.


be derived from the very bright orange-red spoton the ventral surface that is not visible when thespider moves about the upper surface of the web,as it normally does.

We did not observe, in the other species ofNephila, the process of backing up the web withprey in the jaws that is so characteristic of thecarry-in-jaws technique of Nephila maculate.

Table 8 summarizes the main features of thebehavior units of these species.

TABLE 8.—Units of predatory behavior( + present, - absent, ? not known)

Behavior unit

Legs reflexed approachTappingBite and back offAttack biteAttack bite with leg raising..Pull outWrap in situCut outFree wrapAttack wrapCarry in jawsCarry on silkWrap in transitWrap at hubWrap at retreatLeave prey in situ

I i

i 15 s

+++++++++

+++rare+?

that figure prominently in the diet of the Wauspiders (p. 18) and others of interest for com-parative purposes (Table 9).

TABLE 9.—Behavior sequences given by N. macu-lata to four types of prey (durations in seconds)

PREY WEIGHT (mg)mid

ATTACX UNITS

Bitemsd

Pull outmsd

Wrap and cut out

m

sd

TRANSPORT UNITS

Carry in jawsForward

msd

Back-upmsd

Wrap at hubmsd

Totalmsd

*New Guinea. • Panama. * Ghana and Ivory Coast. ' Ghanaand Madagascar. * New Guinea and India.

Flies Moths Katydids DragonfliesN=50) (N=50) (N=50) (N = 30)

34.8 137.3 129.6 110.911.2 56.3 30.3 21.2

5.1 15.8 16.6 44.53.8 9.5 10.0 31.6

2.5 16.4 16.2 52.22.1 13.9 17.2 24.2

lone none* none (N=21)101.238.6

(N=7) (N=8) (N=S) On silk(N=23)

102 4.4 4.0 23.69.3 1.9 - 11.5(N=43) (N=42) (N=47) (N=7)58.7 34.8 42.8 30.720.3 22.7 18.4 17.5(N=6) (N=S5) (N=7)

.26.3 44.0 615 61.510.1 15.0 19.7 11.6

65.029.1

105.343.9

166.645.2

238.861.5

* One free wrap, one wrap in transit.m = mean, sd=standard deviation.

Behavior Sequences

We follow Robinson and Olazarri (1971) in herepresenting descriptions of the behavior sequencesgiven to various types of prey. The responses givento each type of prey are outlined separately andillustrated in a standard form. Comparisons be-tween the types of sequences given to differenttypes of prey are largely confined to the concludingpart of this section.

The prey types chosen for replicated presenta-tions to the free-living adult spiders represent prey

Sequences with Live Flies: Fifty blowflies werepresented to SO spiders. Where the same spider wasused for two observations at least one day inter-vened between presentations. The results are sum-marized in Figure 26. All the flies were bitten forshort periods, pulled from the web and carried inthe jaws to the hub. On 7 out of 50 occasions thespider turned at the capture site and ran or walkedback to the hub facing the direction of movement.In all the other cases the spider backed up the webin the manner described earlier. Only 6 of the 50captured flies were wrapped immediately on arrival

NUMBER 149 55

at the hub. Only four spiders plucked the webbefore rushing down to attack the prey—a reflectionof the fact that the majority of flies vibrated ina sustained manner after striking the web.

FIGURE 26 Predatory sequences upon live flies. The widthof the arrows is proportional to the number of responses inthe direction indicated.

Five spiders overshot the mark, running beyondthe prey, as they made their attack. This led tosome delay in biting. The sequences are simple andrapid; run to prey, bite, pull out, carry in jawsbacking up the web, feed.

Temporal Aspects: The attack and removal fromthe web stages of the sequence are of short duration(bite average 5.1 seconds, range 1-18 seconds; pull-

out average 2.5 seconds, range 1-13 seconds). Thetotal durations are long, due to the slow nature ofthe backing-up process when the prey is carried inthe jaws. The average back-up time was 58.7 sec-onds, range 21-106 seconds, while the average turn-and-run time was 7 seconds, range 3-13 (or 10.2seconds if the one spider that turned and walkedto the hub is considered in addition).

The total sequence durations for those sequencesin which the spider turned before returning to thehub are all less than two-thirds of a minute, theothers average well over one minute.

Sequences with Live Moths: Fifty live moths werepresented to 28 spiders. The moths were of differ-ent species but fairly similar in size and weight. Allwere bitten at the capture site and then pulled fromthe web, without capture-site wrapping. Eighteenspiders plucked the web from the hub before mak-ing a predatory excursion. Four spiders had distinctstages of touching or palpating the prey beforebiting. Transportation was predominantly effectedby backing up the web with the prey in the jaws(42 out of 50), seven spiders turned and ran backto the hub with the prey in their jaws. One spiderwrapped the prey after it stuck in the web duringin-jaws transportation.

Fifteen prey items were not wrapped at the hubafter transportation, two of these had been wrappedprior to or during transportation (1 free-wrap).

The form of the behavior sequences given to livemoths is summarized in Figure 27.

Temporal Aspects: The attack and pull-outphases are of longer duration than in the equiva-lent stages of the attack on flies. Bites averaged15.8 seconds; range 6-58 seconds, and pull-out unitsaveraged 16.4 seconds, range 1-70 seconds. Trans-portation times were again high in the case ofspiders that carried the prey in their jaws whilebacking up the web (average 34.8 seconds, range3-118). Those that turned and ran back to the hubaveraged 4.4 seconds; range 2-8. After wrapping,


CUT OUT

FIGURE 27.—Predatory sequences upon live moths.

15 spiders turned right at the hub and 20 turnedleft. Wrap at hub averaged 26 leg movements perspider and 2.1 silk attachments in the rundgang.

Wrapping at the hub after transportation (70%of all cases) accounted for an average of 44.0seconds.

Sequences with Live Katydids: Fifty live katy-dids were presented to 32 adult spiders. These were

bitten and pulled out in all cases. All except threeof the katydids were subjected to being carried inthe jaws as the spider backed up the web to thehub. Three katydids were carried by the spider inthe jaws as it ran to the hub after turning at thecapture site. All the spiders wrapped the prey at thehub after transportation. The sequence for smallkatydids was thus very simple, as shown in Figure28. Plucking occurred in 11 cases out of 50. It isinteresting to note that free-wrapping did not occurin any of these sequences and that prey were notwrapped at the capture site.

Temporal Aspects: Bite durations were slightlylonger than those given to moths, averaging 16.6seconds. This difference is not significant (statisticalanalysis in Table 9). The pull-out time averaged16.2 seconds, slightly less than that given to moths.Transportation times were high—mean 42.8 sec-onds—for those spiders that backed up the web withprey in their jaws. The three spiders that turnedand ran back to the hub averaged 4 seconds fortransportation. All the prey were wrapped at thehub after transportation and this behavior unitaccounted for a mean time of 61.2 seconds. Totalsequence durations averaged 166.6 seconds; this ishigher than in the case of moths. Thirty-threespiders turned left at the hub and 17 turned right.Silk attachments during the rundgang averaged 3.3and leg movements during wrap at hub averaged43.8 per spider.

Sequences with Live Dragonflies: The sequencesgiven to live dragonflies were characterized by thehigh proportion of wrap and cut-out units thatoccurred. Dragonflies are much more bulky thankatydids in the sense that a greater area of insectstructure become enmeshed and adherent to theweb, because of the body length and the extendedwings. At the same time the insects do not pulleasily from the web. This contrasts with mothswhich have a large wing area in contact with theviscid spiral. We therefore interpret the high pro-portion of wrapping at the capture site, whichoccurs after extensive pulling-out attempts, as aresponse to the problem of prey removal (Robinsonand Mirick, 1971). Twenty-one out of 30 dragonflieswere wrapped at the capture site after pulling-outattempts. Pulling-out attempts were successful inremoving prey from the web in only nine cases.Free-wrapping behavior occurred after successful

NUMBER 149 57

MTf

PUU OUT

f m -WHAP j CARRY

WRAP

AT

HIM

MAN»UtATt

FIGURE 28.—Predatory sequences upon live katydids.

pull-out attempts in one case and after cutting outin two cases. Pulling-out movements resulted in thespider severing the head of the dragonfly in onecase. The head was then transported to the hubas though it were the entire prey. The spider re-moved and transported the remainder of the

dragonfly several hours later. Wrapping occurredduring transportation in one case. There were thusfour cases of wrapping not obviously associatedwith the removal of the prey from the web or itsstorage at the hub (three cases of wrapping at thecapture site after removal from the web and onewrap in transit). Wrapping at the hub occurredin every case where the spider freed the prey bypulling, i.e., in all those cases where the prey wascarried to the hub without having undergonewrapping behavior. Transportation was predomi-nantly carried out by the spider suspending theprey on silk behind its body and walking forwardsback to the hub. Twenty-three spiders carriedprey in this way although one of these started bybacking up the web with the prey in its jaws. Theprey in this case became entangled in the web,was wrapped in situ, cut out, and then carried onsilk. Figure 29 summaiizes the results.

Temporal Aspects: Bite durations averaged 44.5seconds, persistant pull-out attempts after bitingaveraged 52.2 seconds (only nine of these were suc-cessful). Wrapping at the capture site averaged101.2 seconds in duration. Total sequence time forlive dragonflies averaged 238.8 seconds. This meansequence time is higher than that involved in anyof the sequences reviewed so far despite the factthat the dragonflies were not heavier, on average,than the moths or katydids (Table 9). The sevenspiders that wrapped at the hub averaged 61.5seconds for this behavior, four turned left and threeturned right.

Sequences with Large Acridiids: We made 25presentations of very large acridiids to an equalnumber of spiders. These insects were between ca.45 mm and 50 mm in length and between 2 and 3grams in weight. They had large mandibles andspiny posterior legs. They thus were potentiallydangerous and capable of vigorous escape and de-fense activities. We expected that they would pre-sent considerable problems to a spider that did nothave an attack-wrapping behavior. They did. Thespiders were successful in completing a predatorysequence in 13 out of the 25 cases. Twelve acridiidsescaped at some stage in the spider's attack se-quence or without being attacked. The successfulsequences were all highly complex and the partialsequences of the unsuccessful spiders that attackedthe grasshoppers were also complex. The complex


FIGURE 29.—Predatory sequences upon live dragonflies.

sequences cannot be given in diagrammatic form buta selection of verbatim field notes are given in theappendix. In general the sequences showed com-plexity at the approach, attack, and removal fromthe web stages. Approach was of the hesitant typedescribed in the section on behavior units. In morethan half the cases approach was preceded by re-

treat from the hub away from the point of impactof the prey. This form of defensive behavior makesadaptive sense when objects hitting the web arelarge and/or heavy. Attack units involved theemployment of the bite-and-back-off technique withnumerous repetitions of the short bite. Retreatingaway from the prey between bites was a fairly reg-ular occurrence. Early bites were often directedat appendages rather than the body proper and thisagain indicates the temerity with which the spiderapproaches such large prey. Prey removal alwaysinvolved extensive wrapping even when the degreeof attachment to the web had been greatly reducedby the tearing consequent on the prey's struggles.A summary of the major features of the complexsequences is given below.

Temporal Aspects: Because of the repetition ofmajor units in the attack and approach systems, aswell as extensive prey-packaging operations, theduration of the total sequences is consistently high.In extreme cases the duration exceeds three-quar-ters of an hour. The data (below) on both struc-tural and temporal aspects of the sequences supportthe analysis given above.

UNITS OCCURRING IN THE 13 SUCCESSFULSEQUENCES

Sequential order of units Proportion of occurrence1. Retreat from hub before attack 9/132. Plucking before attack 7/133. Slow approach ± leg raising 11/134. Tapping prey before attack 8/135. Bite and back-off attack 11/13*6. Sustained bite ± prior 5 12/137. Wrap in situ, interrupted by rest periods 13/138. Carry on silk 13/139. Hang on hub 13/13

• Maximum numbei of interrupted bites = 13 in twosessions, 3 followed by 10, more than 8 minutes later.

DURATIONS (MINIMA AND MAXIMA) FOR THE 13SEQUENCES (in minutes and seconds)

Sequential order of units Range1. Approach (from retreat position or hub) 1:10-7:402. Bite and back off 0:54-14:103. Sustained bite 0:40- 8:304. Wrap and cut out 1:14-13:405. Transportation 0.30- 3:10

Total Sequence Time:# 9:40-47:01• Includes time spent retreating and rest periods, which are

not included in units 1-5.

Sequences with Live Melolonthids: At certaintimes of the year melolonthid beetles constitute an

NUMBER 149 59

important element of the prey captured by free-living spiders. Initial observations suggested thatthese bulky, comparatively heavily-armored insects,are not easily subdued by the biting attacks ofNephila maculata. We, therefore, presented livingmelolonthids to the spiders and concurrently pre-sented a similar number to the much smallerArgiope aemula. The latter species builds its websin the herb layer and is much less likely to catchthe beetle in large numbers. It is, however, a spiderthat has attack wrapping as its major predatorystrategy and a comparison of the behavior of thetwo species towards the beetles was of interest. Forease of comparison we include details of the Argi-ope predatory sequences in this section, under aseparate heading.

Ten melolonthids were presented to both speciesof spider. Nephila maculata attacked, in all cases,by biting attacks, but these proved to be morecomplex than any that we had previously observedin that they did not conform to the "typical"pattern. Thus the bite-and-back-off pattern wasemployed in five out of ten presentations but inonly one case was it followed by a sustained bite.In four out of the five cases the spider wrapped theprey after a series of bite-and-back-off units. This ismost unusual. In the five other presentations thespiders made bites at the prey without withdrawingfrom contact between bites, but in only two caseswere a series of such bites followed by a sustainedbite; in the three others the spider then proceededto wrap the beetle. Thus contrary to the patternobserved in presentations of other insects, wrappingoccurred in the absence of a sustained bite in allbut three of the cases. Examination of the detailednotes on the sequences shows that the spiders hadextreme difficulty in penetrating the hard, smoothand rounded cuticle of the insect and biting at-tempts were often unsuccessful in penetrating thecuticle effectively. In this situation the spider seemsto have the capacity to switch in the wrap behaviorunit. Reexamination of the detailed notes on thebehavior of N. maculata towards large acridiidsshows that in one or possibly two cases the spidercommenced wrapping after a long series of bite-and-back-off units had terminated in a relatively shortsustained bite; at the time we missed the possibleimplications of this behavior.

In only one case did the spider attempt to pull

the prey out of the web prior to wrapping it insitu and then cutting it from the capture site. Thebeetles, unlike the acridiids, did not struggle vigor-ously during the attacks and made no attempt toraise the elytra and beat the wings (p. 63). As aconsequence they were not largely freed from theweb when the spider prepared to transport them.All the prey were transported to the hub on silklines and not carried in the jaws.

Temporal Aspects: Total sequence times werelong in duration and average over 12 minutes.Operations at the capture site (approach, attack,and packaging) occupy more than half the sequencetime in nearly all cases.

Comparison with Argiope aemula

Nephila maculata ranges from 2 to 4 grams inweight and is much heavier than Argiope aemula(0.5-0.9 grams). Despite this fact Argiope takesless time to attack heavy melolonthid beetles (0.5-0.8 grams in weight). Durations (in minutes andseconds) of predatory sequences are given below.Those for Argiope are for times spent actually incontact with the prey and include periods of rest-on-prey (ROP) (two sequences) but not periods ofrest-on-hub (ROH). The latter period is omittedbecause the spider is at that stage capable of initiat-ing new attacks and is not occupied with predatorybehavior. Nephila does not have a rest-on-hub be-havior. The Nephila attacked all the prey by bitingwhereas Argiope attacked all the prey by wrapping.

Arigiope aemula4:21 ROH6:40 ROH4:13 ROH5:15 ROH9:18 ROP8:47 ROP4:05 ROH3:24 ROH6iS9 ROH

11:32 ROH

Nephila maculata7:57

10-588:17

31:10

28:4112:159:18

13:0811:3210:29

Robinson, Mirick, and Turner (1969) have arguedthat attack wrapping gives the spider the potentialof leaving the wrapped prey at the capture sitewithout having to wait there until it is sufficientlysubdued to be safely cut from the web and trans-ported. It can thus reduce the time spent away


from the hub by interrupting a predatory sequencebefore transportation and leaving this process untilafter a period at the hub monitoring the web. Inthe sequences given to live melolonthids the twothat did not involve rest at hub stages were longerthan all but one of those that did.

The wrapping units were extensive and punctu-ated by short bites or biting attempts. The en-swathement of the prey was very extensive andproduced layers of silk that were very thick incomparison to the effects of wrapping by Nephila.In fact the wrapping behavior of Argiope looksmuch more effective than that of Nephila. It in-volves synchronous (or nearly so) movements ofthe two fourth legs, which cast broad skeins of silkover the prey in rapid succession. On the other

hand the leg movements of Nephila are slow andalternating. They carry relatively narrow skeins ofsilk onto the body of the prey.

All the beetles were transported on silk behindthe spider; in two cases the insects were wrappedagain at the hub after being hung. Argiope appear-ed to be able to bite the wrapped prey withoutdifficulty, although its chelicerae are much smallerand less heavily sclerotized than are those ofNephila.

Comparison with Nephila clavipes

Table 10 summarizes the main features of theattack sequences of Nephila maculata and N. clav-ipes for a number of types of prey. Robinson and

TABLE 10.—Comparisons of sequence units and durations (in seconds) of N.maculata and N. clavipes •

LepidopteraN. maculata N. cltcviprs

on moths, N=50) (on butterflies, N=20)

OrthopteraN. maculatn N. clavipes

(on katydids, N=50) (on crickets.N=20)

Dragon Biet.V. maculata N. clavipes

(N=30) (N=10)Prey weight (ing)

msd

ATTACK UNITS

Bitemsd

Pull-outmsd

Wrap, & cut outmsd

137.356.3

15.89.5

16.413.9

TRANSPORT UNITS

Carry on silkmsd

Carry in jaws

msd

Wrap at hubmsd

Totalmsd

102.4SS.0

29.418.6

47.633.6

1S5.7 (N=7)25.9

none 16.0 (N = 7)S.7

Forward Backward4.4 (N=8) 54.8 (N=42) 2851.9 22.7 16.4

129.630.3

16.610.0

162175

231525.8

29.317.6

12.826.9

110.9212

44531.6

522242

142.737.1

63.723.7

61.655.3

47.0 (N=l)

19.0 (N=l)

Forward Backward4.0 (N=3)42.8 (N=47) 112

18.4 10.5

101.2 (N=21) • • 75.6 (N=7)38.6 18.4

44.0 (N=35)15.0

105.343.9

99.6 (N=13)38.6

228.947.9

61.2 (N=50)19.7

166.645.2

50.719.8

141.347.5

23.6 (N=23)115

30.7 (N = 7)175

615 (N=7)11.6

238.861.5

24.9 (N=7)16.1

53.0 (N=3)

68.7 (N=3)22.9

210.061.3

• Trigona (stingless bees) smallest prey of N. clavipes not strictly comparable with fliesAH sequences are: Bite. Pull out, Carry in jaws, Wrap at hub. Mean total sequence time =maculata on flies.

• • Two were wrapped after removal from the web (see text).

presented to N. maculata (N=50).70, sd = 15. See Table 9 for N.

NUMBER 149 61

Mirick (1971) give further details of the predatorybehavior of N. clavipes, including a summary dia-gram. Generally the similarities between the twospiders are more striking than the differences. Dif-ferences between sequence durations are discussedbelow. The other main differences lie in the formof the behavior units and have been detailed in thatsection.

In terms of unit and sequence durations compari-sons are complicated by the fact that the insectspresented to the two species were not similar insize (including weight), morphology, or species. Inorder to minimize size differences we have selectedsequences of N. clavipes predatory behavior thatwere given to insects that were the closest in sizeto those used in presentations to N. maculata. Thisselection has considerably reduced the sample sizesfor N. clavipes, all of which are now smaller thanthose for N. maculata. This factor may furthercomplicate the comparison. In addition, it must beremembered that Nephila maculata is larger, heav-ier, and has longer legs than N. clavipes. UnlikeNephila clavipes, which transports prey in the jawsby walking forwards to the hub, N. maculata backsup the web, in most cases, when carrying prey inits jaws.

Despite these differences, and complicating fac-tors, there are overall similarities in some sequenceand unit durations—and in the composition of thesequences. Thus, with dragonflies of roughly similarweight, overall sequence times are comparable, evento range and deviation! Bite and pull-out durationsare similar, but the smaller Nephila clavipes showsmuch more variance in pull-out durations. Trans-portation times for prey carried in the jaws arehigher in the case of N. clavipes, moving forwardsup the slope of the web, than for N. maculata mov-ing backwards. This difference is not statisticallysignificant but suggests that our inference, thatbacking up may be advantageous if prey is bulky,could be correct. The possibility that it could bedue to the longer stride of N. maculata is to someextent discounted by the very close similarity ofthe durations of carry-on-silk transportation. Bothspiders wrapped this prey in situ more frequentlythan they successfully pulled dragonflies from theweb.

Comparison of sequences that N. clavipes gave todomestic crickets and N. maculata gave to the katy-

dids show that these were similar in duration de-spite the fact that the crickets were heavier. Onlyone cricket was not removed from the web by pull-ing and no katydids were wrapped in situ. In thiscase the process of backing up the web with preycarried in the jaws (N. maculata) was a lengthyproccess compared with the process of transporta-tion by walking forward with the prey in the jaws,carried out by the smaller N. clavipes.

In the case of butterflies, Nephila maculata wasable to free these, in all cases, by pulling themfrom the web, whereas N. clavipes wrapped in situ28 percent of the butterflies, before cutting themfrom the web. This fact can be related to the dif-ferences in free space, beneath the two species, thatis available for the process of removing bulky preyby pulling. N. clavipes spent considerably moretime, on average, wrapping prey at the hub aftertransportation and this again may be related tothe problem of passing silk over bulky prey whenthere is a relatively small amount of free spacebeneath the spider. The mean total sequence dura-tion for the treatment of butterflies by N. clavipesis over twice that taken by N. maculata. This is atleast partly affected by the long duration of thepull-out, wrap-in-situ, and wrap-at-hub units, inthe case of N. clavipes.

When dealing with small prey of less than .050grams both spiders give similar sequences of basi-cally similar mean durations. Nephila clavipes at-tacked stingless bees (Trigona species) by a simplebite, pull-out, carry-in-jaws, wrap-at-hub procedure.Nephila maculata attacked slightly larger Calli-phora-like flies in this way, although it backed upthe web in 86 percent of the cases. Although N.clavipes wrapped all the Trigona, on, or shortlyafter, arrival at the hub, N. maculata wrapped only12 percent of the flies within one minute of arrivalat the hub. Flies may be relatively much smaller inrelation to the jaws of N. maculata than Trigonaare to N. clavipes, and there could thus be lessneed to secure them by wrapping to prevent thembeing lost during further attacks.

Very large orthopteran prey elicited similar at-tack behavior from the two species (see Robinsonand Mirick, 1971, and above). We did not presentlarge beetles to N. clavipes but suspect that theywould probably encounter the same problems in

62

handling these prey as N. maculata (and probablyin a more exaggerated form).

Categorization of Araneids by Predatory Behavior

Table 11 gives, in summary form a broad cate-gorization of the predatory behavior of those ara-neids for which we have data. There are essentiallytwo major groups of predatory techniques employedby araneids and the distinction between these twogroups is made on the basis of the presence orabsence of attack wrapping. Further distinctionsor subdivisions of the predatory techniques can bemade within the two major groups on the basis ofthe employment of other forms of wrapping be-havior. Such categorization would follow that madeby Robinson, Mirick, and Turner (1969) on thebasis of a much smaller sample of araneid species.We have, so far, found no araneid that shows thebehavior of the hypothetical "Stage 1" constructedby Robinson, Mirick and Turner (1969:499). Thisstage forms one of the alternative sequences foundin most of the Nephila group spiders but does notoccur as the sole predatory technique. We wouldpredict that if it occurs as a unique technique inany species existing at present, that species would


probably be a small spider within the Nephilagroup or a spider specializing on prey much smallerthan itself and possibly building a close-meshed webof lightweight silk. In such a situation prey couldbe bitten, pulled from the web, and wrapping needonly occur at the hub (where it would serve thefunction of allowing the spider to store prey caughtduring one predatory excursion while it made fur-ther attacks on subsequent prey.)

Robinson, Mirick and Turner (1969:500) notedthat some spiders with a isiephila-type attack be-havior were able, as a consequence of post-immobilization wrapping, to leave prey in the webat the capture site and return to the hub to feed onalready caught prey. They thereby were able to omitthe transportation of newly caught prey. This sub-category of a Nephila-type predatory technique wassuggested by studies of one Gasteracantha speciesand two Micrathena species in Panama. Studies ofGasteracantha species in New Guinea confirm thatthey too store prey at the capture site (Y. D. Lubin,pers. comm.).

The occurrence of pull-out behavior in Nephilaspecies as an almost invariable first approach toremoving prey from the web has been remarkedupon as possibly primitive behavior. We know that

TABLE 11.—Predatory behavior of araneid spiders classified on the basis of princi-pal attack units; sub classification based on prey storage strategies

(US = United States, P = Panama, NG = New Guinea, A = Africa, I = India)

Attack biting used exclusivelyNo attack wrapping

AH prey stored at hub Some prey left in situafter attack

Nephila clavipes (P) Gasteracantha cancriformisN. maculata (NG) (P) • • •N. turneri (A) G. theisi (NG)N. constricta (A) Micrathena schreibersi (P)Herrenia ornatissima (NG)

Nephilengys cruentata (A & I) • •

Attack wrapping a major predatory componentAttack biting reserved for some prey items

Prey stored at hub Some prey left in situafter attack

Cyrtophora moluccensis (NG) • Argiope argentata (P)A. savignyi (P)A. florida (US)A. trifasciata (US)A. aurantia (US)A. picta (NG)A. aemula (NG)A. reinwardti (NG)A. aetheria (NG)Araneus marmoreus (US)

• Y. D. Lubin, pers. comm.• • Site of prey storage not determined.

•••This species is now known to attack wrap very large prey (determined after manuscript at press).

NUMBER 149 63

it occurs also in Nephilengys species. The size ofprey that can be removed from the web in this waymay ultimately depend on the relative length ofthe spider's legs i, u, and iv, which determines themaximum distance between the spider's body andthe web surface. This, in turn, determines the freespace available for removing and manipulatingprey. Other factors, presumably including the de-formability of the web plane may restrict thepracticability of using the pull-out technique forremoving prey. Herrenia ornatissima may be unableto use the pull-out technique on larger prey becauseits legs are relatively short (compared to those ofNephila species) and also because deformations ofthe web plane could result in parts of it beingpushed against the nearby tree trunk and adheringthereto. Spiders that build close-meshed webs maynot be able to free prey by the pull-out techniqueunless they immobilize them very rapidly, i.e., be-fore the prey's struggles enmesh it inextricably.Nephila species are remarkably fast in their move-ments across the web, this too may be a consequenceof having relatively long "sprinters" legs.

The occurrence of wrapping after a series ofapparently abortive biting attempts in the attackson very large, or heavily armored, prey by Nephilamaculata is oi interest for two reasons. First, itsuggests that wrapping behavior is not necessarilytriggered by the stimuli that the spider receivesas a consequence of a successful attack bite, asPeters (1931) claimed for Araneus diadematus. Itcould, however, still be triggered by the act ofbiting. In cases where the chelicerae do not pene-trate there is no possibility of the spider testingthe edibility of the prey by taking in materials fromwithin the insect cuticle. In such cases the decisionto proceed with the predatory act must be based onthe perception of stimuli derived from the outsideof the prey. Secondly, the occurrence of wrappingbehavior in the absence of prior successful bitingcan be regarded as a possible first step in the evolu-tion of attack wrapping. As such it is of primeinterest.

BEHAVIOR OF PREY IN THE WEB

In the course of presenting a wide variety of preyto Nephila maculata we made a large number ofincidental observations on the behavior of the prey

items after they had become enmeshed. These areof some interest since they indicate different es-cape potentials of prey when the spider is present inthe web. Since the prey were mostly thrown intothe web by the experimenters their initial orienta-tion may be unnatural and this could affect theirsubsequent behavior. However, if some of theseprey show escape movements that are wholly orpartly effective, there is good reason to assume thatprey arriving in a more natural manner would be,at the very least, as likely to escape.

In general only very small prey and very largeprey tend to escape if the spider is present andreacting normally. Otherwise, the approach to theprey is so rapid that the spider is attacking beforethey prey has had time to free itself. Lepidopteransare an exception.

Of the very large prey the acridiids kick vigor-ously when in the web and these movements,together with their weight, cause extensive tearingof the web so that they slip downward towards thelower margin and may then escape. In addition, ourdata suggest that a vigorously kicking, large acridiidis approached much more cautiously than an inertone and this may gain the insect escape time. Fur-thermore, even when an attack has started (and isat the protracted bite-and-back-off stage) the spidermay be repelled by vigorous kicking and the preymay be able to escape. We have long suspected thatorthopteran regurgitants may affect spider silk buthave not investigated this matter. Most large beetlesstruggle when in the web and may slip to the edgeand escape. Their relatively short legs have a muchless dramatic tearing effect on the web, but themovements that they produce may help to loosenthe attachment of the prey from the viscid element.It is noticeable that they close their wings andelytra in the web, thus presenting a compact smoothmass. In general large beetles seem to have a muchsmaller surface area/weight ratio than large acri-diids and therefore, per unit weight, have less sur-face in contact with viscid droplets. They thus slipout of contact with adhesive fairly rapidly. Afterwatching the wrapping attacks of Argiope aemulaon beetles, we are struck by the fact that relativelysmooth cuticle presents a much more difficult sur-face on which to deposit enswathing layers thandoes the cuticle of orthopterans. If viscid dropletsadhere less strongly to shiny surfaces, some beetles


could have a higher escape potential than others—beetles and insects.

We have no records of lepidopteran escapesfrom our prey-analysis studies. When throwingmoths into the web during the studies on predatorysequences we found that a number did not adheresufficiently to remain in the web until the spiderarrived. In these cases vigorous fluttering move-ments apparently facilitated escape.

Very small prey items, e.g., nematocerans andsmall hemipterans were ignored by the adult femaleNephila even when struggling. Some of these arecertainly consumed by kleptoparasites, but somemay escape. We suspect that when rain wets theweb small insects that have not been attacked maybe liberated.

RESPONSES TO ENVIRONMENTAL FACTORS

SUNLIGHT.—Many tropical insects that assumemore-or-less immobile resting attitudes by daychange their orientation in apparent response tochanges in the direction of incident sunlight. Thusmany orthopterans minimize heat absorption byaligning themselves with their long axis parallel tothe sun's beams. Kettlewell (1959) noted this inthe case of Brazilian katydids, and Uvarov (1965)has reviewed reports of this phenomenon in acri-diids. There are some reports of spiders assuminga special orientation with respect to the sun'sposition. Thus Pointing (1965) showed that thelinyphiid Frontinella communis (Hentz) wouldorient to the sun's disc in both horizontal and ver-tical planes unless it was in the shade or at webtemperatures below 30°C.

Krakauer (1972) has studied the thermal respon-ses of Nephila clavipes and found that at tempera-tures of above 35°C the spider orients the tip ofits abdomen towards the sun and thereby reducesthe radiant heat load on its body. In addition hefound that the spider may also employ evaporativecooling by extruding liquid from the chelicerae.Krakauer"s experimental technique was to placeheat lamps above the middorsal area of the abdo-men of captive spiders, but he did not experimentwith spiders in the field. He does however suggest,on the basis of field records, that Nephila clavipesbuilds its web oriented in a plane normal to thearea of maximum insolation and that the tilt of

the web to the vertical faciliates postural thermo-regulation by reducing the angle at which thespider has to hang in order to point at the sunwhen this is at its zenith.

Our attention was independently drawn to theresponse of Nephila maculata to bright sunlightwhen we found a mature female standing with herlong axis more or less at right angles to the web,facing away from the sun. This complex posture isshown in Figure 30, and is the reverse of that fig-ured by Krakauer (1972, fig. 2). Despite the differ-ence in position the total effect is presumably thesame, i.e., the minimum surface area is exposedto radiant heat. We assume that positions such asthis are not so economic of energy as those involv-ing aiming the abdominal apex at the sun butbecome necessary because of web orientation.

To investigate further orientation to the sun wemounted a circular mirror on a tripod and thenmanipulated the apparent position of the sun bydirecting sunlight onto spiders that were otherwisein the shade. In this way we elicited responses to"impossible" positions of the sun. Such positions,e.g., when the sun was apparently shining upwardsfrom below the web elicited orientations that madeadaptive sense, the spider aligned itself as though itwere minimizing heat absorption from this improb-able heat source. We were also able to inducemovements of spiders across the hub region throughmore than 90 degrees of arc in a few minutes andthus replicate the effects of several hours of azimuthshift.

We did not have the necessary equipment tomonitor temperature at the surface of the spiderbut black bulb temperatures were in excess of 30°Cwhen we made our observations. Clearly spidershave a capacity of postural thermoregulation, whichincludes positions that are more complex thanwould be necessary if the web were always simplyoriented to be normal to the area of maximuminsolation. In fact many of the postures may involvethe spider in energy-consuming attitudes, but pre-sumably these need not be adopted for very longperiods in most tropical regions where uninterrupt-ed sunlight (except in dry season periods) may beunusual. Nephila maculata seldom builds out inthe open and may thus be shaded for long periods,even when there is uninterrupted sunlight in openareas. We are strongly inclined to think that web-

NUMBER 149

FIGURE SO.—Ncphila maculata orienting to strong sunlight, which is passing through the webfrom the right of the picture. This spider has lost the left leg i and right leg it.


site choice is much more likely to be determined byavailability of supporting structures (and preda-tory success) than by the necessity for economizingin energy expenditure on postural thermoregula-tion.

One point that emerges from our observations asworthy of further investigation is the extent towhich heat and visible light interact in mediatingthe orientation of thermoregulating spiders. Itseems entirely possible that temperature may triggerthe response of postural thermoregulation (as sug-gested by the work of Pointing, 1965, and Krakaeur,1972) while the actual orientation may be based onresponses to visible light. We made no attempt todetermine whether the spider's eyes were involvedin the sunlight response, but are currently doingso in experiments with Nephila clavipes.

RAINFALL.—Pointing (1965:73,75) has noted thathis spiders responded to heavy precipitation bymoving to positions in the barrage of threads abovethe sheet web. There they assumed head-downpositions until the rainfall ceased. However theyresponded to less heavy raindrops—but heavy rain-fall—as they responded to sunlight, i.e., by present-ing the minimum surface area to the rainfall. Fromour observations on Argiope argentata in Panamawe know that some spiders may adopt special rainfallpostures if they remain in their webs and do notseek shelter under nearby vegetation. Thus Argiopeargentata hangs away from its sloping web so thatthe body is almost perpendicular and legs i and n,which are always directed anteriorly, are off theweb. In the rainfall posture these two pairs of legsare held outstretched in line with the midlateralplane of the body and at a fairly acute angle tothe long axis, i.e., they are held more anteriorlythan in the normal resting attitude. This positioncould be interpreted as minimizing the cross-sectional area exposed to the rain—assuming thattropical rains fall more-or-less vertically—or thatthe spider in this position maximizes the flow ofwater off its body surface with the anterior appen-dages forming a sort of drip-tip. We have sinceseen this behavior in Nephila clavipes, and Leu-cauge species. In the latter case the spider hangsalmost vertically, head down, from its horizontalweb. Charezieux (1967) has described a rainfallposition adopted by Nephila madagascariensix.

Nephilia maculata responds to heavy rain by

cutting away areas of web (p. 10) and may thenretire to shelter under the leaves of nearby vegeta-tion. Continuous fairly light rain can evoke theweb-cutting response, but the spider may then re-main at the hub of the partially dismantled web.When it does this it assumes a position somewhatsimilar to that assumed by Argiope argentata. Fig-ure 20 of N. maculata essentially shows the basicattitudes of the legs when the spider is respondingto rainfall. The rainfall posture differs from theplucking only in that the spider hangs away fromthe inclined web, more or less in a vertical position,but with the abdominal apex tucked in against theweb.

DEFENSIVE AND ESCAPE BEHAVIOR

Nephila maculata exhibits behavior that can becategorized as defensive because of its form and/orcontext. Touching the spider lightly on its dorsalsurface (unprotected by the web) evokes a charac-teristic posture. In this the prosoma is raised ante-riorly and the total length of the body assumes aconcave dorsal orientation. The posterior margin ofthe prosoma thus becomes pressed against the ante-rior margin of the opisthosoma at the waist joint.In addition, the first two pairs of legs are flexedand raised off the web. This posture may be a de-fense against aerial predators or parasites. The bodyflexure could protect the possibly vulnerable waistand position the body so that the legs can be usedto fend-off insects attacking from above. The raisedlegs may constitute a barrier to such attacks sincethey are held over the prosoma. Because these legsare not gripping the web they may be instantlyavailable for defensive movements.

Escape movements occur in response to disturb-ance of the barrier web or to violent movementswithin the orb. The spider invariably runs upwardsonto upper frame members and often into theshelter of surrounding vegetation. Whether or notthe escape run terminates on a frame thread or invegetation seems to be largely dependent on thestrength of the initial disturbance and whether ornot it persists. We have not been able to quantifythis. The form of the escape run is often distin-guishable from a predatory run because it includesdorsoventral "bouncing" of the body.

NUMBER 149 67

Discussion

This study spanned one year, on a day to daybasis, and touched on many aspects of the biologyof Nephila maculata. Despite its wide scope, it isonly an introductory investigation into the bio-nomics of the species and thus we can raise indetail the more general questions that the studyprovokes and outline some of the principles thathave emerged.

PHENOLOGY

In terms of seasonal patterns of predation exertedby large trap-building spiders we have added datafrom a much less climatically variable area thanthat where we carried out our study of the prey ofArgiope argentata (Robinson and Robinson, 1970a).Differences between the results of the two studiesare not, unfortunately, ascribable to the differencesin climatic pattern alone. Two different techniqueswere used to determine the number, nature, andtemporal distribution of the prey caught by twospiders of distinctly different ecologies. From dataon population dynamics, in addition to the datafrom the prey-capture studies, we feel confident thatNephila maculata in the more equable (than Pan-ama) climate at Wau, is able to subsist throughoutthe year without major perturbations. This raisesa question that seems to bedevil many studies ofanimal phenology in the tropics. How typical wasthe year during which the study was carried out, orhow consistent are long-term patterns of climate?Indications from Barro Colorado suggest that theremay be major differences between the dry seasonsin the amount of rainfall, to name only one factor.Thus the dry season after our study of Argiope ar-gentata in Panama (Robinson and Robinson, 1970a)was much wetter, and the present dry season (1972)has been distinguished by an abnormally highJanuary rainfall. At Wau there are indications ofvery considerable fluctuations in the annual dis-tribution of rainfall. Thus, in May 1970 the rainfallat our study site was 2.8 inches and in the Mayfollowing our study the rainfall was 7.7 inches.Brookfield and Hart (1966) give coefficients of vari-iaton (cv) for the monthly rainfall at the Waunumber 1 station, based on 28 years of completedata. All months show more than a 30 percent coeffi-cient and four months (January, May, August, and

October) show more than 50 percent. These authorsalso give the standard deviations for monthly rain-fall at this station, which also reflect the greatvariability. Brookfield and Hart (1966:14) defineareas of high variability for annual means as thosewith a coefficient of variability greater than 25 per-cent. Wau number 1 has an annual cv of 15 per-cent. The annual variation may be much less im-portant to some organisms than the short termvariations.

What aspect of rainfall may most directly affectweb-building spiders is presently unknown. Theoperation of a trap that is damaged by rainfallcould be affected by the number of discrete rainfallperiods per day, the intensity of the rainfall, thenumber of successive days of rain, and so on. Wehave plotted web renewal and prey-capture figuresagainst some of these aspects of rainfall, but thereare no simple correlations. Similarly we need toknow much more about the effect of the variousaspects of rainfall on the different developmentalstages of the spider before we can assess the effectsof this factor on the biology of the species. Smallwebs produced by immature stages may be moresusceptible to damage by rain than the strongerwebs of adults; but, on the other hand, they may bebuilt in more sheltered sites. Such sites may beavailable to immatures with small webs whereasadults may only be able to build their webs in moreexposed sites. The complexities are immense. Manybiologists accept climatological data from one pointin an area and use this to interpret biological datafrom the entire area. Perhaps this is a valid tech-nique for some organisms and some aspects of theirbiology. It is interesting to note that the threeWau rainfall stations, separated by less than 1kilometer in horizontal distance, and 150 metersin height, show very considerable differences in thetemporal pattern of precipitation.

Since our data on web adhesiveness suggest acorrelation between this factor and relative hu-midity, periods of rainfall that do not result inraindrop damage to the web may be more favorableto the spider than periods when the efficiency ofthe adhesive droplets could be adversely affected bylow humidities. This is yet another factor to beconsidered as part of the complex interactionbetween the arachnid and climate. It is usual todelimit seasons by rainfall and temperature criteria


and many of the climographs used by biologistsplot these two factors (Richards, 1957). D. Leston ofthe University of Ghana (pers. comra.) has sug-gested that hours of sunlight may be crucial tomany organisms and, if these are considered, sometropical areas with two grossly distinct seasons interms of rainfall and temperature then have threeseasons. Since many orb-web spiders have responsesto sunlight plus radiant heat, variation in hoursof sunlight may affect them directly as well as indi-rectly.

Web-building spiders could also be affected indi-rectly by the influence of climatic factors on thestructure of their habitat. Web supports for the im-mature stages of spiders, such as Nephila maculate,may depend on the presence of ephemeral non-woody vegetation (herbs and forbs). These maycycle in abundance during the year in many parts ofthe range of the species. This is certainly true, fromour own observations, of the Bombay site whereThakur and Tembe (1956) studied the species.Even such apparently obscure factors as a periodof heavy production of wind-dispersed seeds couldaffect the success of spiders, since their webs maytrap very large numbers of them and become con-spicuous and ineffective. Large-scale leaf-fall mayhave a similar effect and also increase the amountof sunlight reaching the area of the forest in whichthe spiders build their webs. Apart from increasingthe drain of energy involved in thermoregulation,increased illumination could then render the websand spiders more conspicuous to predators, prey,and parasites.

Our phenological studies were primarily concen-trated on variations in the weight and nature ofthe prey items with time. The data are too crudeto permit anything more than guesses about theirsignificance. This is a complex field. It seems pos-sible that there may be differences between adultfemale Nephila in the caloric input necessary tomaintain a regular egg-laying routine (whateverthat may turn out to be). Thus there are certainlyconsiderable differences in the size of adult females,other than those of expandable parts such as theopisthosoma. Small females presumably have a low-er caloric requirement than adults of large size. Wehave no baseline from which to calculate the max-imum rate of egg-sac production. It seems quitepossible that the minimum interval between egg

masses may be fixed physiologically and that theanimals can adjust to variations in food supply byaltering the number of eggs in the egg sac andextending the period between layings. Certainlythe increase in the size of the opisthosoma prior toegg laying is very striking and may represent aconsiderable increase in weight. We have no ideaif there is an upper limit to food intake that couldbe achieved under normal conditions in the field.Questions about the relationship between food sup-ply and egg production could, perhaps, best beanswered by a laboratory study with captive speci-mens.

The fact that the total weights of prey remnantsvaried, on a weekly basis, between less than onegram to nearly five shows that there were (at leastrelatively) lean weeks and rich weeks. Detailedanalysis of the data shows that lean weeks are leanweeks for all spiders. Thus in week 47 (commencing14 April 1971) the total catch was .958 grams.No spider was responsible for more than 210 mg,5 spiders caught more than 100 mg, 5 spiders caughtless than 100 mg. In rich week 24 (commencing28 November) with a total of 4.670 grams of foodremains, two spiders accounted for more than 1gram of remains each, two spiders exceeded 500 mgin rejected remains, three spiders exceeded 200 mgand the remainder fell below this. Thus the spreadon the good weeks is greater than that on the leanweeks.

The spiders should, presumably, be attempting tomaximize their catches and could be expected toshow some evidence of smoothing the variations inthe abundance of potential prey by behavioral ad-justments. Thus they could attempt to compensatefor low catches at one place by moving the locationof their webs. In addition, they might feed on someorganisms, during periods of low catches, that theywould reject when food was abundant. Our datado not show a positive correlation between preda-tory success and stability of web site, nor, unfor-tunately, a negative correlation between movementsof web sites and low returns of prey. The influenceof other factors could mask such effects. Nor havewe any evidence that the spiders accept some typesof prey, during periods of relative food shortage,that are rejected at other times. Despite this, wethought that the effects of the spider's behaviormight result in there being less variability in the

NUMBER 149

numbers of the prey that they caught than therewas in the numbers of insects caught by thewindow-pane traps. This proved to be the case;but since the means of the two samples are quitedifferent this evidence is, at the most, only sug-gestive.

Turning to the phenology of population incre-ments, it is obvious that it is necessary to conductdetailed studies on the developmental rates of thetwo sexes, under different feeding regimes, and atdifferent times of the year, before we can begin toelucidate this problem. Collection of egg sacs lead-ing to estimations of the fertility of the eggs, thedevelopmental period and factors that influence it,conditions favoring eclosion and subsequent dis-persion, are all necessary and practicable furthersteps in the study.

Studies carried out in the relatively benign cli-matic regimes of upland tropical areas like Wau,should provide some clues about how widely dis-tributed species manage to cope with the moreextreme climatic regimes in other parts of theirrange. (Tropical is here used in the sense of Rich-ards (1957:135), i.e., it applies to the area that lieswithin the isotherm of 20°C mean annual tempera-ture.) We can ask whether the adaptations whichenable N. maculata to exist under the very differentclimatic regime at Bombay already exist in theWau population. Evidence from the fringe habitatsin the Wau area suggests that they may not.

Levy (1970) has suggested that the life cycles ofspiders may fall into two broad categories basedon the developmental rates of the two sexes. In oneclass are the majority of spiders studied so far(primarily temperate region), in which spiders ofboth sexes that derive from the same cocoon reachmaturity at the same time. In the second class arethose spiders in which the sexes mature at substan-tially different rates (siblings could not mate). Levy(1970:534-535) suggests that there are two strate-gies that would insure a coincidence of adult malesand females in the latter case. These are that co-coons should be produced throughout "long periodsof the year," or that there should be two adultphases in the year with males of one phase coinci-dent with females of the previous phase. He suggeststhat, in the case of Nephila madagascariensis, thetwo phase system may operate. Bonnet's (1930) datashow that males mature in approximately a quarter

of the time necessary for females to mature. Ourdata show that there are mature males and femalespresent throughout the year and copulation plusegg laying is occurring throughout the year. InBombay the same species has an apparently onephase reproductive cycle (like a temperate species).This aspect of adaptive reproductive strategy on thepart of Nephila maculata could either be a resultof the direct influence of different climatic regimesor be due to genetic differences in the reproductivebiology of the spiders in different parts of theirrange.

BEHAVIOR

Nephila maculata is the second tropical memberof this genus whose courtship behavior has beenstudied. The discovery of the complex silk-deposi-tion activities of the male adds a new element toihe picture of araneid courtship behavior. General-izations about araneid courtship have, so far, beenalmost entirely based on studies of north temperateforms. We think that further comparative studiesof tropical forms are essential if the evolution ofcourtship behavior is to be understood. The pos-sibility that silk deposition may be a primitivecondition cannot presently be rejected. The occur-rence of this or similar behavior in other speciesand genera may have been missed by students,because of the practical difficulties involved in see-ing large numbers of courtship sessions in seasonal,and widely dispersed, temperate forms.

Our observations on the predatory behavior ofNephila maculata, other Nephila species, andNephilengys and Herrenia, as well as the study ofArgiope species in New Guinea, show that thesespecies conform to the overall categorization ofaraneid predatory behavior provided by Robinson,Mirick, and Turner (1969).

The experiments with large melolonthids as preyshow that the absence of attack wrapping in theNephila group of spiders places them at a disad-vantage, compared with spiders that have attackwrapping, in capturing large prey. This raises thequestion of why the Nephila species have not beenout-competed by Argiope species (or other ad-vanced spiders). No existing Argiope species appearto occupy quite the same niche as the Nephilaspecies. Furthermore all the Nephila species that


we have seen build large strong and relatively fine-meshed webs that are operated day and night. Otherlarge spiders that build webs of comparable size(e.g., Eriophora fuliginea, see Robinson, Robin-son, and Graney, 1972) operate them principally atnight. At night, moths may constitute a large pro-portion of the prey and attack biting is the mosteffective strategy for dealing with these insects.There is probably an advantage in sheer size whenoperating a large web. (The immatures, probablyin the same zone of the forest habitat as Argiopespecies, have at some stage, webs that are similarin size, and they may be more close to niche over-lap at this stage).

The sun-orientation behavior that we have de-scribed raises some interesting questions since, asfar as we know, it does not occur in Argiope speciesthat build webs in positions that are more fre-quently exposed to direct sunlight. There are anumber of possible explanations of this difference.Argiope species may orient their webs so that thespider does not receive radiant heat on its black,ventral surface but on its reflectant dorsal surface.What evidence we have does not support this the-ory. There are very considerable differences in thesurface to volume ratios between the smallish, flat-tened, Argiope species and the large cylindricalNephila species, which must have less surface areaper unit volume than the Argiope species. Thisshould lessen the heat load that they have to com-pensate for but increase the problem of heat dissi-pation once they become heated. The brain, situ-ated in the prosoma, may be an organ that isparticularly sensitive to high temperatures and it isworth noting that all the Nephila species that wehave seen have silvery (possibly reflectant) dorsalcoloration of the prosoma.

Stabilimentum building by immature Nephila

maculata, the first records of this phenomenon forthe genus, raises interesting questions. It seems un-likely that this implies any close relationship withthe genera of spiders that presently build ribbonstabilimenta. If it was evolved independently andis now a vestigial character in Nephila maculatawhat was its original function? Similarly the con-struction of orb-like barrier webs by juvenile N.maculata is a newly described piece of behavior andmay have important implications from the stand-point of defense against predators, suggesting thatthe immature stages that build such barriers maybe particularly susceptible to attack by flying preda-tors. It might also be a clue about the origins ofbarrier webs in general.

We have since discovered that in Panama im-mature Nephila clavipes build complex structuredbarrier webs that are drawn away from the mainweb at their centers and are conical. These aresimilar in gross structure to those described forN. maculata (p. 10), but do not have an orderlyspiral. They are most complete above the dorsalsurface of the spider where there may be two dis-tinct plane structures. Above the main orb, ventralto the spider, there is usually one plane structureand a second supporting complex. The existence ofa triple barrier (two plane structures and one sup-porting structure) above the spider suggests thatthis complex may be defensive. The spider mustbe protected from below by the main orb as wellas the secondary structures. Full details of thejuvenile barrier webs of Nephila clavipes, withphotographs, are to be published separately.

We suspect that almost any long-term studiesof tropical spiders would reveal suggestive, new,information about their behavior and ecology. Thisis part of the fascination of spiders.

Addendum

Since this study was at press, new information has become available on two subjects dealtwith herein: (1) Stabilimentum building by Nephila species (pp. 8-19, herein). It has beendiscovered that, on rare occasions, Nephila clavipes builds linear ribbon stabilimenta. Detailsand comments on the possible evolutionary significance of these structures, will be given by us("The Stabilimentum of Nephila clavipes and the Origins of Linear Stabilimenta in Araneids,"Psyche, in press). (2) Postural thermorcgulation by Nephila species (pp. 64-66, herein). Therepertoire of thermoregulatory postures of Nephila clavipes is now known to include ananalogous posture to that described herein and illustrated in Figure 30. A detailed analysis ofthe thermoregulatory behavior of N. clavipes is in preparation ("Adaptive Complexity: TheThermoregulatory Postures of the Golden-web Spider Nephila clavipes at Low Latitudes.").

NUMBER 149 71

Literature Cited

Alexander, A. J.. and D. W. Ewer1957. On the Origin of Mating Behavior in Spiders. Amer-

ican Naturalist, 91:311-317.Bastock, M.

1967. Courtship: A Zoological Study. London: HeinemannEducational Books Ltd.

Bonnet, P.1929. Les Araignlcs exotiques en Europe, II: Elevage a

Toulouse de la grande Araigne'e fileusc de Mada-gascar ct considerations sur l'Araneiculture (Pre-miere Partie). Bulletin de la Sociiti Zoologique deFrance, 54:501-523.

1930. Les Araignees exotiques en Europe: II. Elevage aToulouse de la grande Araignee fileusc de Mada-gascar et considerations sur 1' Araneiculture (Deux-ieme Partie) Bulletin de la Sociiti Zoologique deFrance, 55:53-77.

1958. Bibliographia Araneorum, 2 (N-T) :4. Toulouse,France: Douladore.

Bristowe, W. S.1929. The Mating Habits of Spiders, with Special Ref-

erence to Problems Surrounding Sex Dimorphism.Proceedings of the Zoological Society of London,1929: (2):309-358.

1931. The Mating Habits of Spiders: A Second Supple-ment. Proceedings of the Zoological Society of Lon-don, 1931: (4): 1401-1412.

1939. The Comity of Spiders. Volume 1. London: RaySociety.

1941. The Comity of Spiders. Volume 2. London: RaySociety.

Bristowe, W. S., and G. H. Locket1926. The Courtship of British Lycosid Spiders and its

Probable Significance. Proceedings of the Zoologi-cal Society of London, 1926(2) :317-347.

Brookftcld, H. C, and D. Hart1966. Rainfall in the Tropical Southwest Pacific. Depart-

ment of Geography Publication G/3, AustralianNational University, Canberra.

Charczieux, B.1961 Contribution a l'etude du comportement sexuel de

l'Argiopide Nephila madagascariensis observe1 dansson pays d'origine. Bulletin de la Socie'te Zoologiquede France, 86:372-379.

1967. Comportement de l'Araignee Nephila madagascar-iensis Vinson au moment des chutes de pluie. An-nales de la Faculti de Sciences de VUniversite deMadagascar, 5:117-120.

Chrysanthus, Fr.1959. Spiders from South New Guinea, II. Nova Guinea,

new series, 10:197-206.1960. Spiders from South New Guinea, III. Nova Guinea

Zoology, new series, 3:23-42.1971. Further Notes on the Spiders of New Guinea, I:

(Argyopidae). Zoologische Verhandelingen, 113:1-52.

Colcman, E.1948. An Interesting Spider. Victorian Naturalist (Mel-

bourne) , 64:206.Crane, J.

1949. Comparative Biology of Salticid Spiders at RanchoGrande, Venezuela, Part IV: An Analysis of Display.Zoologica, 34:159-214.

Eisner, T., R. Alsop, and G. Ettershank1964. Adhesiveness of Spider Silk. Science, 146:1058-1061.

Fischer, C. E. C.1910a. Pairing of the Spider Nephila metadata Fabr. Jour-

nal of the Bombay Natural History Society, 20:526-528.

1910b. Further Note on the Spider Nephila maculata.Journal of the Bombay Natural History Society,20:528.

1910c. Further Note on the Spider Nephila maculata Fabr.Journal of the Bombay Natural History Society,20:887-888.

Gerhardt, U.1930. Biologische Beobachtungen an Nephila madagas-

cariensis Vins. Sitzungsberichte der Geselbchaftnaturforschender Freunde zu Berlin, 8:358-360.

1933. Neuc Untersuchungen zur Sexualbiolgie der Spin-nen, insbesondere an Arten der Mittelmeerlanderund Tropen. Zeitschrift fur Morphologic undOkologie der Tiere, 27:1-75.

Hinde, R.1966. Animal Behaviour. New York: McGraw Hill.

Hingston, R. W. G.1922a. The Snare of the Giant Wood Spider (Nephila

maculata), Part I. Journal of the Bombay NaturalHistory Society, 28:642-649.

1922b. The Snare of the Giant Wood Spider (Nephilamaculata), Part II: The Continuation of the Snare.Journal of the Bombay Natural History Society, 28:911-917.

1922c. The Snare of the Giant Wood Spider (Nephilamaculata), Part III: Further Lessons of the Nephila.Journal of the Bombay Natural History Society,28:917-923.

1923. The Giant Wood Spider (Nephila maculata).Journal of the Bombay Natural History Society,29:70-76.

Jager, S. de1960. Nephila, a Highly Intelligent Spider. African Wild

Life, 14:205-208.Kajack, A.

1965. An Analysis of Food Relations Between the Spiders—Araneus cornutus Clerck and Araneus quadratusClerck—and Their Prey in Meadows. EkologiaPolska, A, 13:717-768.

1967. Productivity of Some Populations of Web Spiders.Pages 807-820, in Secondary Productivity of Terres-trial Ecosystems, II. Warsaw: Panstwowe Wydawn-ictwo Naukowe.


Kaston, B. J.1964. The Evolution of Spider Webs. American Zoologist,

4:191-207.Kaston, B. J., and E. Kaston

1953. How To Know the Spiders. Dubuque, Iowa: Wm.C. Brown, Co.

Kettlewell, H. B. D.1959. Brazilian Insect Adaptations. Endeavour, 18:200-

210.Krakauer, T.

1972. Thermal Responses of the Orb-weaving SpiderNephila clavipes (Araneae: Argiopidae): AmericanMidland Naturalist, 88:245-250.

Langer, R. M., and W. Eberhard1969. Laboratory Photography of Spider Silk. American

Zoologist, 9:97-101.Levy, G.

1970. The Life Cycle of Thomisus onustus (Thomisidae:Araneae) and Outlines for the Classification of theLife Histories of Spiders. Journal of Zoology (Lon-don), 160:523-536.

Linsley E. G., T. Eisner, and A. B. Klots1961. Mimetic Assemblages of Sibling Species of Lycid

Beetles. Evolution, 15:15-29.Locket, G. H., and A. F. Millidge

1953. British Spiders. Volume 2. London: Ray Society.Macfadyen, A.

1963. Animal Ecology: Aims and Methods. London: Pit-man.

McKeown, K. C.1963. Australian Spiders. Sidney: Angus & Robertson.

Morris, D.1956. The Function and Causation of Courtship Cere-

monies. Fondation Singer-Polignac, Colloque Inter-national sur I'lnstinct (Paris), 261-266.

Osorio, J. M., and A. Moreno1943. Contribuciones de Zool6gica Economica, I: "La

Arafta Gran Tejedora" Nephila clavipes (L). Me-morias de la Sodedad Cubana de Historia Natural,17:53-57.

Peters, H. M.1931. Die Fandhanglung der Kreuzspinne (Epeira diade-

mata L.) Experimented Analysen des Verhaltens.Zeitschrift fur Vergleichende Physiologie, 15:693-748.

1954. Estudios adicionales sobre la estructura de la redconcentrica de las aranas. Communicaciones del In-stituto Tropical de la Vniversidad de El Salvador,1:1-18.

1955. Contribuciones sobre la etologfa y ecologfa com-parada de las araflas tejedoras tropicales. Com-unicaciones del Instituto Tropical de la Vniversidadde El Salvador, 1-2:37-46.

Platnick, N.1971. The Evolution of Courtship Behaviour in Spiders.

Bulletin of the British Arachnological Society, 2:40-47.

Pointing, P. J.1965. Some Factors Influencing the Orientation of the

Spider Frontinella communis (Hentz.) in Its Web(Araneae: Linyphiidae). Canadian Entomologist,97:69-78.

Richards, P. W.1957. The Tropical Rain Forest. Cambridge, England:

Cambridge University Press.Riek, E. F.

1970. Neuroptera. Pages 472-494 in The Insects of Aus-tralia. Melbourne: Melbourne University Press.

Robinson, M. H.1969. Predatory Behavior of Argiope argentata (Fabri-

cius). American Zoologist, 9:161-173.Robinson, M. H., Y. D. Lubin, and B. Robinson

In prep. Population Studies of Wcb-building Spiders atWau, New Guinea.

Robinson, M. H., and H. Mirick1971. The Predatory Behavior of the Golden-web Spider

Nephila clavipes (Araneae: Araneidae). Psyche, 78:123-139.

Robinson, M. H., H. Mirick, and O. Turner1969. The Predatory Behavior of Some Arancid Spiders

and the Origin of Immobilization Wrapping.Psyche, 76:487-501.

Robinson, M. H., and J. Olazarri1971. Units of Behavior and Complex Sequences in the

Predatory Behavior of Argiope argentata (Fabri-cius): (Araneae: Araneidae). Smithsonian Contri-butions to Zoology, 65:1-36, figures 1-16.

Robinson, M. H., and B. Robinson1970a. The Prey Caught by a Sample Population of the

Spider Argiope argentata (Araneae: Araneidae)in Panama: A Year's Census Data. Zoological Jour-nal of the Linnean Society of London, 49:345-357.

1970b. The Stabilimentum of the Orb-web Spider Argiopeargentata: An Improbable Defence against Preda-tors. Canadian Entomologist, 102:641-655.

1972a. Techniques for the Observation of Spider Be-haviour. Bulletin of the British Arachnological So-ciety, 2:58-59.

1972b. The Structure, Possible Function and Origin ofthe Remarkable Ladder-web Built by a New Guin-ea Orb-web Spider (Araneae: Araneidae). Journalof Natural History (London), 6:687-694.

In prep. Comparative Studies of the Mating Behaviour ofOrb-web Spiders, 1: Argiope species.

Robinson, M. H., B. Robinson, and W. Graney1972. The Predatory Behaviour of the Nocturnal Orb-

web Spider Eriophora fuliginea (C. L. Koch)(Araneae: Araneidae). Revista Peruana de Ento-mologia, 14:304-315.

Robinson, M. H., B. Robinson, and Y. D. LubinIn prep. Long-Term Changes in the Adhesiveness of

Spiders' Webs.Roewer, C. Fr.

1942. Katalog der Araneae. Bremen.Roth, L. M., and T. Eisner

1962. Chemical Defences of Arthropods. Annual Reviewof Entomology, 7:107-136.

NUMBER 149 73

Ruhen, O.1969. Scan the Dark Coast. London: Hodder and Stough-

ton.Savory, T. H.

1928. The Biology of Spiders. London: Sidgwick andJackson.

Shear, W. A.1970. The Evolution of Social Phenomena in Spiders.

Bulletin of the British Arachnological Society, 1:65-76.

South wood, T. R. E.1966. Ecological Methods. London: Methuen 8c Co. Ltd.

Thakur, M. K., and V. B. Tembe1956. Bionomics of the Giant Wood Spider, Nephila macu-

lata Fabr. Journal of the Bombay Natural HistorySociety, 55:350-334.

Tinbergen, N.1954. The Origin and Evolution of Courtship and

Threat Display. Pages 233-250 in Evolution as aProcess. London: Allen and Unwin.

Turnbull, A. L.1960. The Prey of the Spider Linyphia triangularis

(Clerck) (Araneae: Linyphiidae). Canadian Jour-nal of Zoology, 38:859-873.

Uvarov, B.1965. Grasshoppers and Locusts, 1. Cambridge, England:

Cambridge University Press.

Yaginuma, T.1969. Spiders of Japan in Colour. Osaka: Hoikusha Pub-

lishing Co. Ltd.

Appendix 1

Activities of Male Nephila tnaculata during Courtship Silk-deposition

(Notes on two-hour observation session; time intervals omitted, except for long rest periods; silk attachment occurs at all pointslisted)

Start: male moves from dorsal surface abdomen(DSA) onto base of third leg left (3L), to surfaceabdomen, right then left, to 4R then across thoraxto 4L, 3L, back to DSA right, pause 2 min.+, DSAleft, to 4L, to 4R, 3 and 2R rapidly, then 4L, 3L,2L, back to DSA right, then 3R, 2R, 1R, DSA left,pause 1.40 min., then to 4L taps leg base, 3L tapsleg base DSA right, three silk attachments, then4L, 3L, 2L, to DSA left one attachment, 4R, 3R,2R, 1R, then DSA left, DSA right, then 4R, 3R, 2R,4L, 3L, 4R, 3R, 2R two attachments, DSA andpause left, then right, rest 1.20 with legs 1, 2, onbases of female leg 4R, then move to halfway upfemale DSA, then back to abdominal margin right,4L, 3L, 2L then between 2-1L, DSA left, pause2.10 on outer edge, touching legs of female, thenDSA right, pause 3.15, 4L, 4R, 3R, between 3-2R,4L, DSA center, 4L, DSA right pause 2.25, DSA lefttwo attachments, 4L, 3L, 2-1L, 4R, 3R, 3-2R, DSAright pause 1.00+, move halfway up right DSA,pause 2.05, 4R, 3L, 4-3R, 3-2R with two attach-ments below leg bases, DSA right 4.00, DSA edgeright taps bases of female legs 2 and 3R (apparent-ly feeling beneath female's legs, 3.30, 3L, 3-4L.4R, 3-2R, 4L, DSA right, then against raised ante-rior edge of abdomen as female attacks coccinellidbeetle beetle rejected, female backs up web to hub,male remains on scarp of abdomen, female strokesthe base of her legs 4, 3, 2L with tarsi of oppositelegs 2, 3, as though feeling the silk deposited bythe male, silk now clearly visible as a gusset at theleg bases, male inactive for 2.30 min., female plucksweb in response to gust of wind, male moves to4R, underneath 4R, 3R, 2R, 1R, 2R, then DSAmedian position, 4L, 3L, DSA median, DSA right,4L, DSA right, rests 2.00+, 4R, 4L, 2L, 3L, all well-up from leg bases, DSA right in upper half abdo-men, 2R, 4L, 3-2R, DSA center, 4L, 3-2L, DSAcenter about 8mm from edge, male feels between

2-3R, 3-4R with legs 1 and 2, DSA right 1.12 then4R, 3R, 2R, DSA left, 4L under, 3L under, DSAright pauses 40 sec. touching 4R with legs, then4-3L, 2L, 4R, 3R, 2R, 1R, 4L, 4R, DSA right rest1.00+3R, 2R, 4L, 4R, 3-2R, 2-1R, DSA right 3-2L,2L, 1L, 4R under, SR, 2R, DSA left rest 2.32, 3-2L,2-3R, 4L under, 3L under, 4R, 3-2R, DSA right,DSA left, rest 2.05 tapping 4R, 3R, 4-3R, 2-1R,DSA right backs up to abdominal apex and attachesline, returns to edge tapping occasionally 2.45,down to 4R, 3L, 3-2L, 4-3R, 3-2R, DSA left resttapping 4, 3L 1.43, 2-3R, 3-4L, 3-2L, 1-2L, 4R,3R, 2R, all under, DSA midline 1.00+, moves un-derneath female abdomen sideways, on top again,4L, 4R, 1R, pedipalps!, 4L, DSA right, sidles tounderside of abdomen advances down, then, ontop, 1L, pedipalp L, 1R, 2R, 2-3R, 2-1L, DSAright, pauses 2.20, then sidles under abdomen andquickly back, again under approaches epigyne tapswith pedipalps, taps again, alternating movements,Pinsert right pedipalp momentarily, then left, leftstill down, standing with right legs 1, 2, on baseof female legs—2 minutes insertion ???, male shut-tles up onto DSA after observer accidently touchesweb foundation line, motionless, touches 4R, 4L,1-2R, 4R under, 1-2R under, DSA right 1.00+, 4R,3R, 4R under, 3-2R, 2-1R, DSA left rest less than1.00, 4-3L, 2-1L, 3-2R, 4-3R, DSA midline thenup to apex abdcmen, insect strikes web femaleresponds immediately, male stands on scarp ofabdomen—is this a response to female's movement?,prey escapes rest 2.00, male moves up to DSA fromscarp, female cleans pedipalps, rests 2.12, thenspends 5.00+ on DSA intermittently tapping legbases of female . . . [MHR's observation periodended, BR takes over].

Down to 3R, DSA median 1.10, 4L, 3L, 3L, DSAright, pauses 2.40, cleaning feet. 2R, 3R, 1R, 2Runder, DSA center 3.10, 2L, 3L tapping, 3R-4R,

74

NUMBER 149 75

2-3L, DSA right rest 1.40, 3-4L, 3-4R, 2-3L, 3-4R,DSA left, 2-3-L, DSA right, rest 2.10, 1-2L, pedi-palp L, 3-2R under, 2-3L, 4R, DSA median 2.20,2-3R, DSA center rest 1.40, 2-3L, 3-4R, 2-3R, 3-4L, 3-2L, 3-4R, 3R, DSA center rest less than 1.0,chelicerae R, 1R, chelicerae L, 3-2L, 3R, 2L, 3R,2L, 3R, 3R, 4R, DSA center rest less than 1.00,apex of abdomen, down ventral surface of abdo-men, shuttle to DSA center, 4R, 3R, 3L, 4R, 3R,

4L, 3L, 3R, 4R, 2R, 3-4L, DSA right rest 2.18, apexof abdomen, down dorsal surface to scarp right,rest 1.10, web blows in wind, male stationary.Female cleaning and moving abdomen as if tryingto remove male, swings off web so that all legs onone side can scrape at male.

End of session: 2 hours; time spent in silk at-tachment, approximately 49.55 minutes.

Appendix 2

Verbatim Notes on Selected Sequences Given to Large Acridiids

(See pages 57-58, for background)

BehaviorPresentation 1: 50 mm acridid

PluckSlowly down to preyTouchMove down y2 in., tap with legs i, raise

these back off web over thorax in de-fensive posture

Lunge forward, bite, back-offMotionlessEdge forward until pedipalps contact preyLunge forward over prey, bite, back offMotionlessLunge, biteSustained bite on second right leg of preyMove bite location, in short steps to right

thoraxShort biting attempts jaws edgewise to

tough thoraxReturn to biting prey leg 2RShift bite to junction of leg 1R and

thoraxSmall bites at thorax edgeWrap prey (x82)Cut out and pullWrap (x28)Cut outWrap (X16)Cut outCarry on silk, supported by both legs iv

then left ivAt hubTurn right, attach carry-silk 3 timesPause and clean, feet, chelicerae, pedi-

palpsPull up prey, handle, then feed.

Total sequence time 28:42

Elapsed time(min.'.sec.)

0- 0:090:10- 0:180:18- 4:154:15- 7:40

7:40- 7:427:4S- 9:209:21- 9:289:20- 9:329:33-11:50

11:51-11:5711:58-15:0515:05-15:10

15:11-15:18

15:19-17:3017:31-21:25

21:26-22:2022:21-23:1523:16-23:3423:35-24:1224:13-24:5624:57-25:0825:09-26:3926:40-27:10

27:1027:11-27:3827:39-28:40

28:41-28:42

Presentation 4: 50 mm acridtid(The absence of a substantial bite in this

be noted.)

Retreat to 18" above hub on upper foun-dation thread

Edge slowly down to hubSlowly approach prey, legs i k II (x ' )

raised off web, pluckHalt 8" from preyResume slow approachTouch with legs iTouch with legs II as wellPalpateLunge, bite, back-offPause, motionlessLunge, bite, back-offPauseLunge, bite, back-offPauseLunge, bite, back-offPauseLunge, bite, back-offPause about 3" from prey touching it

with 2 legs i, one leg IIMove dose to prey and palpate along

dorsal surface thorax to headTurn left and wrap (x 8 6 ) cutting

intermittentlyCut along upper surface of prey and

attach line to intact web abovePalpate down to headCut out belowWrap (Xl6)Cut aboveWrap (X8)Cut final attachmentsCarry on silk with support from both

legs rvAt hubTurn right at hub and attach silk 4

timesManipulate preyFeed

sequence should

0:01- 2:02

2:03- 2:182:19- 3:10

3:11- 3:173:18- 3:513:52- 3:553:56- 4:104:11- 4:224:23- 4:254:26- 4:344:35- 4:384:39- 4:464:47- 4:504:51- 5:075:08- 5:105:11- 5:165:17- 5:255:26-12:40

12:41-13:12

13:13-16:25

16:26-16:42

16:43-17:0817:09-17:2717:28-17:5117:52-17:5717:58-18:0518:06-18:3518:36-19:15

19:1619:17-19:23

19:24-20:1820:18

•& U. B. GOVERNMENT PRINTING OFFICE. 1*7» 8t«-*X4/*2

Total sequence time 20:18

76

Publication in Smithsonian Contributions to Zoology

Manuscripts for serial publications are accepted by the Smithsonian Institution Press, sub-ject to substantive review, only through departments of the various Smithsonian museums. Non-Smithsonian authors should address inquiries to the appropriate department. If submission isinvited, the following format requirements of the Press will govern the preparation of copy.

Copy must be typewritten, double-spaced, on one side of standard white bond paper, withW'2," top and left margins, submitted in ribbon copy with a carbon or duplicate, and accompa-nied by the original artwork. Duplicate copies of all material, including illustrations, should beretained by the author. There may be several paragraphs to a page, but each page should beginwith a new paragraph. Number consecutively all pages, including title page, abstract, text, litera-ture cited, legends, and tables. The minimum length is 30 pages, including typescript and illus-trations.

The title should be complete and clear for easy indexing by abstracting services. Taxonomictitles will carry a final line indicating the higher categories to which the taxon is referable:"(Hymenoptera: Sphecidae)." Include an abstract as an introductory part of the text. Identifythe author on the first page of text with an unnumbered footnote that includes his professionalmailing address. A table of contents is optional. An index, if required, may be supplied by theauthor when he returns page proof.

Two headings are used: (1) text heads (boldface in print) for major sections and chap-ters and (2) paragraph sideheads (caps and small caps in print) for subdivisions. Furtherheadings may be worked out with the editor.

In taxonomic keys, number only the first item of each couplet; if there is only one couplet,omit the number. For easy reference, number also the taxa and their corresponding headingsthroughout the text; do not incorporate page references in the key.

In synonymy, use the short form (taxon, author, date:page) with a full reference at theend of the paper under "Literature Cited." Begin each taxon at the left margin with subse-quent lines indented about three spaces. Within an entry, use a period-dash (.—) to separateeach reference. Enclose with square brackets any annotation in, or at the end of, the entry.For references within the text, use the author-date system: "(Jones, 1910)" and "Jones(1910)." If the reference is expanded, abbreviate the data: "Jones (1910:122, pi. 20: fig. 1)."

Simple tabulations in the text (e.g., columns of data) may carry headings or not, but theyshould not contain rules. Formal tables must be submitted as pages separate from the text, andeach table, no matter how large, should be pasted up as a single sheet of copy.

Use the metric system instead of, or in addition to, the English system.Illustrations (line drawings, maps, photographs, shaded drawings) can be intermixed

throughout the printed text. They will be termed Figures and should be numbered consecu-tively; however, if a group of figures is treated as a single figure, the components should beindicated by lowercase italic letters on the illustration, in the legend, and in text references:"Figure 9b." If illustrations (usually tone photographs) are printed separately from the text asfull pages on a different stock of paper, they will be termed Plates, and individual componentsshould be lettered (Plate 9b) but may be numbered (Plate 9: figure 2). Never combine thenumbering system of text illustrations with that of plate illustrations. Submit all legends onpages separate from the text and not attached to the artwork. An instruction booklet for thepreparation of illustrations is available from the Press on request.

In the bibliography (usually called "Literature Cited"), spell out book, journal, andarticle titles, using initial caps with all words except minor terms such as "and, of, the." Forcapitalization of titles in foreign languages, follow the national practice of each language.Underscore (for italics) book and journal titles. Use the colon-parentheses system for volume,number, and page citations: "10(2) :5-9." Spell out such words as "figures," "plates," "pages."

For free copies of his own paper, a Smithsonian author should indicate his requirementson "Form 36" (submitted to the Press with the manuscript). A non-Smithsonian author willreceive 50 free copies; order forms for quantities above this amount with instructions for pay-ment will be supplied when page proof is forwarded.

Ecology and Behavior of the Giant Wood Spider

Documents

Transcript of Ecology and Behavior of the Giant Wood Spider