Ballooning: Data from Spiders in Freefall Indicate the Importance of Posture

7 Pages·1992·3.5 MB·English

by  Robert B Suter

)
1992. TheJournal ofArachnology 20:107-113
BALLOONING: DATA FROM SPIDERS IN FREEFALL
INDICATE THE IMPORTANCE OF POSTURE
Robert B. Suter: Department ofBiology, Vassar College, Poughkeepsie, New York
USA
12601
ABSTRACT. Ballooning, the aerial displacement ofa spider caused by friction between air and the spider
with its silk, has considerable ecological importance but remains poorly understood as a mechanical process.
Thestudiesreportedhereprovideinsightintothemechanicsofballooningbywayofexperimentsinvolving(1
stroboscopic measurement ofthe rates offall ofspiders slowed by known lengths ofsilk and (2) direct mea-
surement ofthe draggenerated by moving air at the surface ofspiders as those spiders change their postures.
Theterminal velocitiesofspiders trailingsilk, derived from theirrates offall at known distances from release,
were remarkably variable. The variability could not be attributed entirely to silk length and spider mass, and
in somecases silklengthwasnoteven statisticallycorrelatedwith terminal velocity. Dragmeasurements made
onlivingspidersrevealedthatthevariabilityinterminalvelocitiescouldbeexplainedbyvariabilityinposture.
Some spiders, particularly small ones, travel allowed one to predict roughly its terminal ve-
greatdistancesbyballooning.Theecologicalcir- locityatanyknownlengthofballooningsilk(Su-
cumstances under which ballooning occurs are ter 1991).
nowwell-documented(e.g., Richter 1970, Vugts Theexperimentsreportedbelowconstitutean
& VanWingerden 1976), anddataderived from extension ofthat work, by testing the true ter-
newly developed techniques (Greenstone et al. minalvelocitiesoffallingspidersagainstthepre-
1985) should increase our understanding ofthe dictions generated in the earlier work.
temporal, spatial, and geographical parameters
of ballooning. The physics of ballooning, in a METHODS
theoretical sense, has also been well elucidated, Spiders.—The great majority of ballooning
both very generally (Vogel 1981) and in consid- spiders have masses below 2 mg (Greenstone et
erable detail (Humphrey 1987). In contrast, our al. 1985, 1987),althoughmuchlargerspidersare
empirical information about the physics ofbal- found among the aerial plankton (Coyle et al.
looninghasbeenlimitedtodeductionsfrom mi- 1985).Becauseofthatsizedistribution,onlysmall
crometeorological,behavioral, andmorphomet- spiders were used in this study. In freefall trials,
ric observations of ballooning (e.g., Dean & five spiders were subjects: three were unidenti-
Sterling 1985, Eberhard 1987, Greenstone et al. fiedhatchlingtheridiids(0.18,0.18,and0.2mg),
1987), and a single experimental study (Suter onewas an unidentified immature theridiid (1.8
1991). mg) and one was an immature Uloborus glo-
In that earlier paper, I reported on the drag mosus (Walckenaer) (Uloboridae) (1.0 mg). In
producedbysilkandbywholespidersunderthe the drag measurement trials, both subjects were
laminar flow conditions ofa wind tunnel. The unidentifiedtheridiids,oneahatchling(0.18 mg)
most useful results of the study were a set of andoneanimmature(7.2 mg). Thespiderswere
equationsrelatingtheweights ofspidersand the capturedfromtheirwebsinAugustandSeptem-
lengthsoftheirballooningsilkstotheirterminal ber 1991 andwere maintained in plasticvials at
velocities and to the velocities ofrising air nec- 100%R.H.forthefewhourstofewdaysbetween
essary to make them airborne. Because the ex- capture and testing. They were not fed during
periments involved spiders of several families, theircaptivityandwereweighedwithin24hours
many sizes, and a variety ofshapes, the results oftesting.
were quite general and therefore directly appli- Freefall.—Figure 1 shows the apparatus used
cable,inasense,onlytospidersofaverageshape to measure the acceleration and velocity ofspi-
anda particularposture. Nevertheless, fora liv- ders falling through still air in a dark room. In
ing spider ofa particular weight, the equations eachtrial,aspiderwasinducedtoattachitsdrag-
107
108 THEJOURNALOFARACHNOLOGY
TMAX
dak 3200)wasusedtorecordthespider’s
motion. The height ofthe nichrome wire loop
nichrome
was adjustable, which made it possible to alter
the distance a spider fell before it appeared in
front ofthe camera.
Each spider in these trials was tested repeat-
edly, at different silklengthsandat differentdis-
tancesfrom itsrelease. Theresultingvelocityvs.
distance lines were plotted together with theo-
retical curves derived from drag measurements
made on spiders and their silk in wind tunnel
experiments reported elsewhere (Suter 1991).
Manyofthevelocityvs. distancelineswerelong
enough to confirm that their shapes conformed
tothetheoreticalcurves(e.g.. Figs. 3-5). I there-
fore assumed that all of the lines represented
segments of similarly shaped curves, and esti-
mated the terminal velocities by extrapolation
in cases in which the final three data points did
notindicaterelativelyconstantvelocity(e.g.,Figs.
6-7).
Drag measurement.—Data derived from the
freefall trials suggested that variation in spider
posture strongly influenced the rates of fall of
spiders trailing draglines. Because posture was
Figure 1.—Apparatus used to measure the velocity difficult to ascertain from the images produced
ofaspiderfallingthroughstillairwhiletrailingaknown during freefall trials, I designed the apparatus
lengthofdraglinesilk. Electricalcurrentappliedtothe shown in Fig. 2 to allow direct observation of
nichrome wire loop at the top released the silk and posture and simultaneousmeasurement ofdrag.
spider. The falling spider, illuminated by stroboscope Theapparatusconstitutedaminiaturewindtun-
flashes at 100/s, was photographed as it fell past the nelinwhichairneartheaperturewasinlaminar
scroboscope and camera (not shown). The nichrome flow and had velocity (measured by a hot wire
wire loop could be adjusted to any height, allowing anemometer)thatcould becloselycontrolled(at
experimental manipulation ofthe distance the spider
airspeedslessthan0.2m/s,velocitiesvariedless
fell before beingphotographed. than 0.008 m/s). The tests reported here took
place at an air velocity of0.07 m/s.
linetoasmallnichromewireloop(diameter:0.5 Each spider in these trials was attached at the
cm). At some point during its subsequent de- posterior surface of the abdomen (with cyano-
scent, electrical current was fed to the wire loop acrylate glue) to the end ofa wire 1.4 cm long
which incandesced rapidly and released the spi- (diameter: 17.8 iim). The other end ofthis wire
der trailing its dragline (a hot-wire anemometer was connected with paraffin to the end ofa gal-
indicated no measurable air flow, at 1 cm below vanometerneedle(see Fig. 2). An opaqueplastic
thewire loop, duringthe first0.5 saftertheloop target (1.2 x 2.0 mm) mounted on the galva-
wasenergized). Dim lightfocusedonarulernear nometerneedlepartiallyinterruptedthebeam of
the path ofthe dangling and then falling spider a HeNe laser, so that very small movements of
allowedanobservertonoteandrecordthelength theneedleweredetectedaschangesintheamount
ofsilkwithwhichthespiderhadbeensuspended oflight fallingupon a photodetector. DC output
justpriortoitsrelease. Acamerawith itsshutter from the photodetector was converted to a fre-
open, mounted so that it focused on both the quency modulated signalbypassingitthrough a
FM
rulerand the path ofthe falling spider, recorded voltage-controlledoscillatorcircuit, and the
the position ofthe spider during each flash ofa signal was recorded on one audio channel ofthe
strobe (set at 100 flashes/s). Because ofthe very videotapethatwasrecordingtheactivitiesofthe
briefdurationofeachflash, highspeedfilm (Ko- spider.
SUTER-BALLOONING: POSTUREANDTERMINAL VELOCITY 109
AirSource
nSSWS
.SSAW
sSSSW
vSASAS
vvvS\SSSSSSSSSWA\ HcNe
v\SSS\SSA%A\ Laser
\SSSSS
^SS^S^
sS\S%\
AirSource
Galvanometer
Figure 2.—Side view(a) andtopview(b)oftheapparatususedto measure draggeneratedbya livingspider
in a laminar air flow while recording changes in its posture. Changes in the posture ofthe spider (above the
middle arrow in a, and inset) caused changes in friction between the movingair and the spider’s surface, and
that change in drag caused movement ofthe needle. The change in position ofthe needle was detected as an
alteredamountoflightfallingonthephotodetector(inb).BoththeFM-encodedsignalbearingdraginformation
and the video signal bearingposture information were recorded forlateranalysis.
110 THE JOURNALOFARACHNOLOGY
Distance from Release (cm) Silk Length (cm)
Figures3-12.—Eachgraphontheleftcontainsdatapertainingtoadifferentspiderasitfell,repeatedly,through
still air. In each graph, the top curve represents the velocity ofan object falling through a vacuum; the other
twocurves represent velocitiescalculated from Suter(1991) ofa spiderofthe samemass asthetestspiderand
with the shortest length oftrailing silk used in the tests shown (middle curve) or the longest length oftrailing
silk (bottom curve). The shorter curves in each graph represent the velocities ofthe test spider in separate
freefalls: becauseoftheexperimental setup(see textand Fig. 1), curvesegmentsoriginatingnearertothe origin
are from trials in which a relatively longlength ofsilk acted as the balloon, and segments originatingat longer
distances from release are from trials with shorter silk lengths. Each graph on the right shows the relationship
between silklengthand terminal velocityofeach testspiderin Figs. 3-7. Thetestspideridentities, masses, and
SUTER-BALLOONING: POSTUREAND TERMINALVELOCITY 111
I calibrated the apparatus at the end ofeach
trial by first turning off the air flow and then
giving the spider a series of coils of very fine
nichromewire (diameter: 2.5 iim), each ofmea-
sured length (and therefore of known weight).
Fortunately the spiders were very cooperative:
each grabbed and held on to each wire coil, ma-
nipulating it for several seconds then dropping
it, allowing me to record (as above) the effect of
each weight on the output ofthe photodetector.
Because output from the photodetector was re-
corded on the same tape that carried postural
information in the form ofa video signal, sub-
sequentanalysisoftherelationshipbetweendrag
and posture was facilitated.
RESULTS Figure 13.—Draggeneratedonthesurfacesofather-
Spiders with trailing silk draglines passed the idiidhatchling(0.18mg).Thehorizontallinesindicate
dragatdifferentpostures:(a)alllegsextendedandwav-
camera lens at velocities that were in part de-
ing;(b)alllegslooselyheldagainstthebody;(c)alllegs
pendent on how farthey had fallen. Figures 3-7 tightly held against the body. Note that a very small
show the velocities of these falling spiders as changeinposture(frombtoc)resultedinasubstantial
functionsofdistancefrom therelease point(i.e., reduction in drag.
the elevation ofthe spider when its fall began).
Each figure presents the data from multiple re-
leasesofthesamespider,allplottedagainstthree was by no means perfect. Significant inverse re-
theoretical curves. Those theoretical curves rep- lationships between silk length and terminal ve-
resent (1) the velocity of a spider falling in a locity were found for both ofthe 0.18 mg ther-
vacuum (no drag), (2) the velocity ofa spider of idiid spiderlings(Figs. 8, 9) but notforthe other
the same mass as the test animal, trailing the threetestanimals(Figs. 10-12),allofwhichwere
minimum length ofsilk used by the test spider larger. The substantial deviations from the ex-
duringthetrials(dragcalculatedfromSuter 1991), pected relationship led to the tentative conclu-
and (3) the velocityofa spiderofthe same mass sion that posture must play a prominent role in
as the test animal, trailing the maximum length influencing terminal velocity.
ofsilk used by the test spider during the trials Measurements ofdrag on spiders mounted in
(drag calculated from Suter 1991). Of the five a laminar air stream are shown in Figs. 13, 14.
spiders tested in this study, only the immature Relativelysmallchangesinpostureforbothvery
uloborid produced data that consistently corre- small and somewhat larger spiders caused as
sponded closely with expectations derived from much as a 10-fold change in drag, even at the
the wind tunnel experiments. very low air velocity used in this test.
Because the distance from release varied in-
DISCUSSION
versely with silk length, individual curves near
the origins of Figs. 3-7 represent spiders with Aspiderfallingthroughstillairacceleratesun-
relatively longdraglinesand curves farfrom the tilthedragproducedbyairflowingpast itsbody
origins represent spiders trailing relatively short and trailing silk just equals the pull ofgravity:
lengths of silk. Sometimes a spider fell more at that point, acceleration is zero and the spider
slowlywhentrailingalongerdraglinethan when isatitsterminalvelocity.Thusterminalvelocity
trailing shorter draglines, but this relationship isafunctionofthemassofthespiderandvarious
velocity vs. silk length relationships were as follows: Figs. 3 & 8, hatchling theridiid, 0.18 mg, R = 0.90, P <
0.01; Figs. 4 & 9, hatchlingtheridiid, 0.18 mg, R = 0.88, P < 0.05; Figs. 5 & 10, hatchlingtheridiid, 0.20 mg,
R = 0.39, P > 0.05; Figs. 6 & 11, C/. glomosus, 1.0 mg, R = 0.59, P > 0.05; Figs. 7 & 12, immature theridiid,
1.8 mg, R = 0.49, P > 0.05.
9
112 THEJOURNALOFARACHNOLOGY
0 10 20 30 40 50 60 70
Time
(s)
Figure 14.—Draggenerated on the surfaces ofa 7.2 mg theridiid. The postures shown at three points in the
graph are computer-enhanced images from individual video frames.
characteristics ofthe surfaces ofthe spider itself is sometimes a significant component in the de-
and its silk. The relevant characteristics ofthe termination ofterminal velocity (Figs. 8, 9), it
surfaceshavebeen discussedin theoryby Hum- by no means acts alone: even at a single silk
phrey (1987), and I have discussed elsewhere length in an individual spider, terminal velocity
(Suter 1991) their relative importances as de- can varybya factorof2.6 (e.g., Fig. 10: at 5 cm,
ducedfromwindtunnelexperiments. BothIand 78 vs. 29 cm/s). This variability must arise,
Humphrey acknowledged a role of posture in therefore, from the single uncontrolled variable
determining drag, but both ofus considered its in this system, the posture ofthe falling spider
role to be secondary, Humphrey because sim- relative to the direction ofmotion.
plifyingassumptionsneededtobemadetomake Theinfluenceofpostureonthedragdeveloped
the calculations tractable, and I because ofmy by a spider in a moving air stream is clearly
use ofdead spiders in the wind tunnel experi- evidentin Figs. 13, 14. Even whatappearsto be
ments. The results reported here make clear the asmallchange,pullingthelegstightlytothebody
needtoreassesstheroleofpostureinballooning. from a posture where the legs are slightly more
Figures 3-7 showthatindividualspiderstrail- loosely held, can result in a reduction in dragof
ing silk sometimes fall much faster and some- 0.17 at a very low air velocity (0.07 m/s),
timesmuchslowerthanpredicted.Becausethese andthatisforaspiderwithaweightof1.76 nN.
large differences can occur between trials of a At much higherair velocities like those reached
single spider (e.g., Figs. 3, 7), they cannot be by a falling spider of the same size (e.g., from
attributed to differences in morphology. All of Fig. 3, 50cm/s),theeffectofthisposturalchange
the silk used by the spiders was dragline, the would be much greater [using Equation 4 from
structure of which is known and relatively in- Suter(1991),thereductionindragduetoacom-
variant (Suter 1991), so that the differences in parableposturalchangeat50cm/swouldbe 1.1
velocities cannot be attributed to variations in mN] and would easily account for the terminal
silkstructureeither. Finally,althoughsilklength velocity differences seen in Figs. 8-12.
SUTER-BALLOONING: POSTUREAND TERMINALVELOCITY 113
The importance ofposture as an influence on looners (Araneae: Atypidae, Ctenizidae). J. Arach-
terminal velocity in ballooning spiders must, of nol., 13:291-296.
course,varywiththeamountofsilkthatisused: Dean, D. A., & W. L. Sterling. 1985. Size and phe-
when a spider uses only a short length of silk, nologyofballooningspidersattwolocationsineast-
theinfluenceofposturewillbegreaterthanwhen EbeerrhnarTde,xaWs., GU.SA.198J.7.AraHcohnwols.p,id1e3r:s1i1ni1t-i1a2te0.airborne
silkisverylongandlittle ofthedragon the silk/ lines. J. Arachnol., 15:1-9.
spider system is developed by the spider’s body Greenstone, M. H., C. E. Morgan & A. L. Hultsch.
and appendages. Unfortunately, almost nothing 1985. Spider ballooning: development and evalu-
is known about the amount ofsilk actually used ationoffieldtrappingmethods(Araneae).J.Arach-
byballooningspiders.Thatabsenceofdatameans nol., 13:337-345.
that the importance ofposture in the travels of Greenstone, M. H., C. E. Morgan, A. L. Hultsch, R.
ballooning spiders (even in their ability to con- A.Farrow&J.E.Dowse. 1987. Ballooningspiders
trol theirelevation) cannot yet be assessed. inMissouri,USAandNewSouthWales,Australia:
familyandmassdistributions.J.Arachnol., 15:163-
ACKNOWLEDGMENTS
170.
I am indebted to Katherine J. Suter who as- Humopnhsrpeiyd,erJ.baAl.lCo.oni1n9g8.7.OecFolluoigdima,ec7h3a:n4i6c9-c4on7s7t.raints
sistedinthedatacollection,toWilliamG. Eber- Richter, C. J. J. 1970. Aerial dispersal in relation to
hard and Matthew H. Greenstone for their crit- habitat in eight wolfspider species {Pardosa, Ara-
icalreadingsofthemanuscript,andtoFernando neae, Lycosidae). Oecologia, 5:200-214.
Nottebohm who generously provided me with Suter, R. B. 1991. Ballooning in spiders: results of
both laboratory and office space at The Rocke- wind tunnel experiments. Ethol. Ecol. & Evol., 3:
feller University Field Research Center, Mill- 13-25.
brook, N. Y. The work reported here was sup- Vogel,S. 1981. LifeinMovingFluids.WillardGrant
ported in part by an equipment grant from the Press, Boston.
Beadle Fund ofVassar College. Vugts, H. V., & W. K. R. E. Van Wingerden. 1976.
Meteorological aspects of aeronautic behavior of
LITERATURECITED spiders. Oikos, 27:433-444.
Coyle,F. A., M.A. Greenstone,A. L. Hultsch&C. E.
Manuscript receivedDecember 1991, revisedFebruary
Morgan. 1985. Ballooning mygalomorphs: esti-
1992.
matesofthemassesofSphodrosand Ummidiabal-