social behavior in amblypygids, and a reassessment of arachnid social patterns

23 Pages·2007·1.28 MB·English

2006. The Journal of Arachnology 34:399–421
SOCIAL BEHAVIOR IN AMBLYPYGIDS,
AND A REASSESSMENT OF ARACHNID SOCIAL PATTERNS
Linda S. Rayor and Lisa Anne Taylor1: Department of Entomology, Cornell
University, Ithaca, New York 14853, USA. E-mail: [email protected]
ABSTRACT. Aggregation,extendedmother-offspring-siblinginteractions,andcomplexsocialbehaviors
are extremely rare among arachnids. We report and quantify for the first time in Amblypygi prolonged
mother-offspring-sibling associations, active aggregation, and frequent ‘‘amicable’’ (tolerant, nonaggres-
sive) tactile interactions in two species: Phrynus marginemaculatus C.L. Koch 1840 (Phrynidae) and
Damon diadema (Simon, 1876) (Phrynichidae). Sociality is characterized by frequent contact and toler-
ance, and infrequent agonism until sexual maturity in D. diadema and into adulthood in P. marginema-
culatus. Weexperimentallyexaminedpotentialbenefitsandcostsaffectingaggregation:riskofpredation,
preferredhabitatsandpreyavailability.Onlyincreasedpredationriskdecreasednearest-neighbordistances
and increased maternal vigilance. Individuals aggregated on a variety of surface textures and locations
thatvarieddaily,ratherthanaggregatingonlyonpreferredmicrohabitats.Manipulationofpreyabundance
had no affect on the tendency to aggregate.
Patterns of parental care, duration of association, and the presence of social traits found in the most
social taxa of non-spider arachnids are reviewed. Species in most arachnid orders havetransientparental
care with defense of eggs, a brief period of association with newly emerged young prior to independent
foragingandexplosivedispersalfromthenatalnest.Moreprolongedsociality,withlong-termassociations
among mothers-offspring-siblings is rare and is only described in a few species in the Amblypygi, Scor-
pionida, Pseudoscorpionida, and Acari. All such species have subsocial origins, but current use of the
term subsocial is overly broad and we propose a more restrictedterminology for clarity.
Keywords: Whip spiders, aggregation, maternal care, ontogeny, predation risk, foraging
Sociality in arachnids is extremely rare. mother-offspring-sibling associations that last
Herewereportprolongedassociationbetween for brief periods (spiders: Kim 2000; Schnei-
mothers and their immature offspring, active der 1995; scorpions: Polis & Lourenco1986),
aggregationamongsiblings,andextensiveso- ‘‘subsocial’’ associations that last until sexual
cial contact in two species of captive ambly- maturity (spider mites: Saito 1997; Mori &
pygids,PhrynusmarginemaculatusC.L.Koch Saito 2005; amblypygids: this manuscript), to
1840 (Phrynidae Blanchard) and Damon dia- the complex multiple adult societies of the
dema (Simon 1876) (Family Phrynichidae), most social of the colonial and cooperative
that differs from the solitary behavior previ- spiders (reviews in Buskirk 1981; Aviles
ously described for amblypygid adults. Sub- 1997; Whitehouse & Lubin 2005) (see Tables
sociality, or an association between mothers 1–3 for details). Along with describing pat-
and their offspring or among siblings prior to ternsofsocialinteractionsinamblypygids,we
reaching sexual maturity, has not been previ- propose expanding and modernizing the defi-
ously reported in the Amblypygi. Although nitions of subsociality to better interpret the
arachnid sociality at all levels of complexity social and evolutionary implications that al-
is uncommon, sociality has been applied to low these rare predatory species to success-
describe behaviors ranging fromtransientear- fully live in groups.
ly parental care (Laniatores opilionids: Ma- Beyond studies of the dramatic courtship
chado & Raimundo 2001; Machado 2002; and fighting behavior of adult amblypygids
Uropygids: Schmidt 2003), to ‘‘subsocial’’ (Alexander1962;Weygoldt&Hoffman1995;
Weygoldt 2000, 2002) relatively little work
1Currentaddress:SchoolofLifeSciences,Arizona hasbeendoneonotheraspectsofamblypygid
State University, PO Box 874601, Tempe, AZ behavior. Most studies suggest that adult am-
85287. blypygids aregenerallysolitaryandintolerant
399
400 THE JOURNALOF ARACHNOLOGY
Table1.—Generalpatternsofearlyparentalcareforeacharachnidorder.Duetoalackofinformation,
Palpigradi and Ricinulei were excluded. Where informative, T indicates the pattern is typical for the
members of the order, C indicates that this is common among some Families, and R indicates that only
a few rare species exhibit the alternate pattern. Numbers refer to citations in all threeTables:1(cid:2)Avile´s
1997; 2 (cid:2) Brach 1978; 3 (cid:2) Buskirk 1981; 4 (cid:2) Cloudsley-Thompson 1977; 5 (cid:2)Coddingtonetal.1990;
6 (cid:2) Evans 1998; 7 (cid:2) Harvey 2003; 8 (cid:2) Kim 2000; 9 (cid:2) Kullmann 1972; 10 (cid:2) Machado 2002; 11
Machado & Oliveira 1998; 12 (cid:2) Machado & Raimundo 2001; 13 (cid:2) Machado & Vasconcelos 1998; 14
(cid:2) Mahsberg 2001; 15 (cid:2) Mori & Saito 2005; 16 (cid:2) Polis & Lourenco 1986; 17 (cid:2) Polis & Sissom1990;
18 (cid:2) Punzo 1998; 19 (cid:2) Ramachandra& Bali 1990; 20 (cid:2) Rowland 1972; 21 (cid:2) Rayor & Taylor 2006;
Arachnid order
Parental trait Araneae Amblypygi Uropygi Schizomida
Carry egg sac C T: 28 T: 19, 23 20
Live young
Guard eggs C
Youngest instar carried R T: 28 T: 23 20
First mobile instar guarded C 21 23
No Associationwith 1st instar C
Association with 1–2 instars prior to explosivedispersal R T: 28 19, 23 7, 20
predators. No work has evaluated whether that ‘‘Within favorable habitats, [Phrynus as-
these traits are also characteristic of juvenile peratipes Wood 1983, from Mexico] individ-
andsubadultamblypygids.AccordingtoWey- uals tend to aggregate, and numerous speci-
goldt (2000) ‘‘Whip spiders are not social an- mens could be collected from a single cave.’’
imals,infacttheyarenotevengregarious.On It is common in outhouses or slit latrines
the contrary, they avoid each other or react throughout the neotropics, to find multiple
aggressively when encountering a congener.’’ amblypygids in proximity to one another,
Other researchers have also concluded that probably to take advantage of the abundant
adultamblypygidsaresolitaryanimalsbothin insect prey at such sites (Rayor pers. obs.). In
the field and laboratory (Gray & Robinson Florida, P. marginemaculatus are frequently
1986; Alexander 1962). Alexander(1962)ob- associated with lightning downed trees with
servedonlyaggressiveandthreateningbehav- activetermitecolonieswhereanumberofam-
ior between adult conspecifics of both sexes blypygids maybefoundapproximately30cm
in Damon variegatus (Perty 1834; from cen- apart at regular intervals along the termite
tral and southern Africa) and Phrynus bar- trails (T. Gearheart pers. comm).
badensis (called Admetus barbadensis) (Po- Female amblypygids lay eggs and carry
cock 1894; from Barbados). She concludes them in a brood pouch underneath their ab-
that much of the intraspecific behavior that domen and the newly emerged young climb
does not lead up to courtship ‘‘consists of onto their mother’s abdomen for the duration
threateningcontestsinwhichthedelicatefeel- of their first instar (first incomplete juvenile
ers [whips] areused as ‘weapons’.’’Otherac- or pullius, Canard & Stockmann 1993; prae-
counts describe similar adult male-male con- nymphae, Weygoldt 2000). Alexander (1962),
flicts (Weygoldt & Hoffman 1995; Weygoldt Weygoldt (2000), and we have observed that
2000). if one of the immobile, helpless young falls
In contrast, Weygoldt (1977) found that from the mother’s back during the first instar,
Heterophrynus longicornisButler1873(Fam- and cannot get back on, the mother will do
ily Heterophrynidae) from the Brazilian Am- nothing to help, and may even eat this indi-
azon were especially tolerant of one another. vidual. From observations of this behavior,
Individuals were often found in male-female Alexander (1962) concludes ‘‘that there is no
pairs, along with a variable number of im- maternal behavior towards young [amblypy-
mature individuals, up until the fourth or fifth gids] that leave their perches on the back of
instar. In one case seven adults were found the mother’’ (Alexander 1962). However,
inhabiting a large log. Quintero (1981) notes Mahsberg (2001) describes similar cannibal-
RAYOR& TAYLOR—SUBSOCIALITYIN AMBLYPYGIDS 401
Table 1.—Extended. 22 (cid:2) Saito 1997; 23 (cid:2) Schmidt 2003; 24 (cid:2) Schneider 1995; 25 (cid:2) Shivashankar
1994;26(cid:2)Tizo-Pedroso&Del-Claro2005;27(cid:2)Weygoldt1969;28(cid:2)Weygoldt2000;29(cid:2)Whitehouse
&Lubin2005;30(cid:2)Zeh&Zeh1990.Note:Thereisextensiveconfusionoverthedifferentdevelopmental
patterns and terminology related to arachnid development (Canard&Stockmann1993).Inthearachnids,
after emergence from the egg, in most there is a first instar (incomplete juvenile) that does not feed
independently and is essentially immobile, while the second instar is capable of feeding independently
and is mobile. For each order, we have consistently referenced the patterns of behavior from the first
mobile, independent instar as defined by Canard & Stockmann (1993).
Arachind order
Opiliones Solifugae Pseudoscorpiones Scorpiones Acari
C: 18 T: 26, 27
T: 17
R: 12 R: 15, 22
27 T: 17
R: 12 17 R: 15, 22
T: 12 T: 4, 18
R: 12 R: 4 T: 2, 26, 27, 30 T: 14, 17 R: 22
ism of first instars that have fallen from the ofthesespecies.Ourpreliminaryobservations
mother’s abdomen among the social emperor of P. marginemaculatus and D. diadema sug-
scorpions, Pandinus imperator Koch 1841, gested that these animals not only tolerated
and suggests that maternal cannibalism atthis one another, but consistently aggregated and
stageweedsoutdeformedindividualsinorder interactedfrequentlywithextensivewhipcon-
to allocate energy towards other young with tact. Here we document the tendency to ag-
superior reproductive value. We consider Al- gregate, how social dynamics change with
exander’s (1962) conclusions about the limits maturity, and the role of the whips in medi-
ofmaternalbehaviorbasedsolelyondamaged ating social contact among amblypygids. In
individuals during the first instar that cannot addition, we investigated the following three
survive independently to be flawed. questions to evaluate possible costs and ben-
Alexander(1962)concludedthatafterleav- efits that individuals may experience by ag-
ing their mother’s back, young amblypygids gregating. First, do individuals form aggre-
were ‘‘clearly independent’’ of their mother. gations merely to gain access to particularly
Gray & Robinson (1986) describe laboratory favorable microclimates within our experi-
observations of a small Australian species, mental cages? Second, given that amblypy-
Charinus pescotti Dunn, 1949, that also dis- gids are obligate predators that potentially
perse from the mother after undergoing their compete for prey (Weygoldt 2000), do differ-
firstmolt.Weygoldt(2000)statesthatafterthe ences in prey abundance effect individuals’
young leave the mother’s abdomen (in a lab- tendency to aggregate? Third, since aggrega-
oratory setting) they ‘‘begin their life without tion is often associated with reduced risk of
any further maternal care.’’ However Wey- predation(Alexander1974),doesthepresence
goldt (2000) indicates that ‘‘in most species of a potential predator or a disturbance affect
the nymphal instars are not aggressive to- group size or proximity to other members of
wards each other; moreover, the motherisnot the group? Finally, we review the patterns of
aggressivetowardsheroffspring.Growingan- early parental care and more complex social
imals become increasingly aggressive shortly behavior known for each of the non-Araneae
before reaching maturity but the adults are arachnid orders, and make proposals about
more tolerant of each other.’’ characterizing social behavior in these spe-
In this paper, we expand upon Weygoldt’s cies.
(2000) observations and describe in detail
METHODS
prolonged mother-offspring-sibling associa-
tions in two species of captive amblypygids, Study animals.—Two species of amblypy-
as well as associations between adults in one gidswereusedinthisstudy.Male,female,and
402 THE JOURNALOF ARACHNOLOGY
Table 2.—Patterns of association in the most social species known for each arachnid order. Citations
are given as in Table 1. The numbers of species (n) known to display this type of association are listed
for the non-Araneae arachnids. Spiders are represented by a few token species that represent particular
features.TheAcariareonlyrepresentedbytheseventetranychidspidermiteswhosepatternsofassociation
Arachnid order
Social trait Araneae Amblypygi Uropygi Schizomida
Associationpast2ndinstar,for1–2additionalinstars 8, 29
Association part-way through development 3, 24
Association up to sexual maturity 6 Dd, 21; n (cid:2) 1
Associatepast sexual maturity-multipleadults 1, 29 Pm, 21; n (cid:2) 1
spermatophore stalk voucher specimens of D. dema closely match the allometric changes in
diadema have been deposited in the Cornell body and palp width reported by Weygoldt
University Insect Collection and at the Smith- (1999). Mother-offspring-sibling interactions
sonian National Museum of Natural History. in D. diadema were observed in five clutches
Voucher specimens of P. marginemaculatus produced by three adult females. Here we re-
have been deposited in the Cornell University port spatial and behavioral changes that oc-
Insect Collection. Extensive video documen- curred during the first year until sexual ma-
tation of both species’ social interactions and turity for the five groups, but most of the
the exploratory response to lizards are avail- experimental data reported in this paper were
able through the Cornell Laboratory of Orni- collectedfromGroup4(seeGroupdetailsbe-
thology,MacaulayLibrary.Videovouchersare low). Spatial dynamics were observed in a
archived in the Macaulay Library and can be group of ten unrelated adult D. diadema.
found at: http://animalbehaviorarchive.org (or In total, nine amblypygid groups of differ-
from the author). These videos can be located ent ages and housing conditions were ob-
through an Advanced Search of the Notes for served.Initialclutchorgroupsizeisgivenfor
‘‘RayorAmblypygidSociality’’or‘‘Rayoram- each group. All ages for immatures relate to
blypygid predator exploration’’ or by species the start of the second instar when the young
name, Rayor, & behavior. descended from the mother’s abdomen and
P.marginemaculatus(Phrynidae)isasmall wereindependentlymobileandarereferredto
species common in southern Florida, often asthe‘‘emergencedate.’’Theemergencedate
foundonhouses,underboards,logs,trash,un- at the start of the second instar is comparable
der the bark of dead trees, and limestone out- to the point when young, newly mobile uro-
crops (Muma 1967, Hebets pers. comm.). All pygids and scorpions descend from their
specimens used in this study were collected mother’s back. The groups were: Group 1
from pine and oak flatlands near Fort Myers, (clutch size n (cid:2) 12) and Group 2 (n (cid:2) 18)
FL by Todd Gearheart between January 1998 each consisted of a single P. marginemacu-
and August 2000. Interactions between P. latus mother and her offspring (emergence in
marginemaculatus mothers and offspring late August 1998; observed September 1998–
were observed in two clutches from emer- April 1999). Group 3 was a mixed colony of
gence through eight months. Adult spatial in- unrelated young, subadults and adults of P.
teractions were observed in a captive colony marginemaculatus (group size n (cid:2) 29: fe-
of 29 unrelated adult and subadult P. margi- males (cid:2) 13, males (cid:2) 8, undetermined sex (cid:2)
nemaculatus. 7; observed August 2000–May 2001). Group
The second species, D. diadema (Phryni- 4consistedofoneD.diademamotherandher
chidae), is found in coastal forests and caves offspring (clutch size n (cid:2) 18; emergence No-
in Tanzania and Kenya (Weygoldt 1999). vember 1999; observed November 1999–
AdultspecimenswerecaughtinsouthernTan- March 2001). Groups 5–8 each consisted of
zania near the Usubara Mountain Range by D. diadema mothers housed with their off-
local collectors and sold for export (Somma spring from May 2001 through October2002.
& Gearheart, pers. comm.), and all young Group 5 (n (cid:2) 32; emergence 24 April 2001)
were born in captivity. Body sizes of D. dia- and Group 6 (n (cid:2) 38; emergence 24 May
RAYOR& TAYLOR—SUBSOCIALITYIN AMBLYPYGIDS 403
Table 2.—Extended. have been well studied, and omit species whose social interactions are anecdotal
(Saito, pers. comm.). For the amblypygids, we have indicated the categories in which Damon diadema
(Dd) and Phrynus marginemaculus(Pm) fit into the scheme based on the data in this paper.
Arachnid order
Opiliones Solifugae Pseudoscorpiones Scorpiones Acari
14; n (cid:2) 8 22
14, 25; n (cid:2) 1 15, 22; n (cid:2) 3
14, 16 22
2, 26, 30; n (cid:2) 3 16; n (cid:2) 2 22; n (cid:2) 4
2001; same mother as Group 4). Group 7 (n withinthecages werepresumedtohaveequal
(cid:2) 50; emergence 18 November 2001). Group exposuretolightortohumanactivitieswithin
8 (n (cid:2) 20; emergence 28 April 2002) born to the room. Behavioral and spatial observations
the Group 5 female, and resided with their were made at various times throughout the
mother and subadult Group 5 siblings. Group day and night, typically in the dark under red
9 consisted of unrelated mature adults of D. light.Additionalbehavioralobservationswere
diadema(n(cid:2)9:6females,3males;observed recordedinalmosttotaldarknessusingaSony
from August 2000–February 2001). When digital camcorder (model DV-TRV9), with
two females in a group oviposited, they were ‘‘nightshot’’ infrared lighting.
removed from the group cage. AmblypygidWhips.—Thesensoryandso-
Housing and Diet.—All animals were cial lives of amblypygids are clearly centered
housed in clear plastic or glass aquarium cag- on the first pair of legs (or whips), which are
es with a substrate of vermiculite and potting extensively used for odor discrimination (He-
soil to maintain humidity. Cage sizes for P. bets & Chapman 2000), spatial location, and
marginemaculatus were: Groups 1 and 2—10 tactile contact between individuals (personal
(cid:3) 10 (cid:3) 18 cm; Group 3—50 (cid:3) 25 (cid:3) 30 cm. observation).Thewhipsaremodifiedintothin
CagesizesforD.diademawere:Groups4and antenniform sensory structures that can mea-
6—50 (cid:3) 25 (cid:3) 30 cm; Groups 5, 7, and 9— sure three to six times the length of the body.
50 (cid:3) 26 (cid:3) 42 cm. Large sheets of cork bark The whips are covered with sensitive che-
were tilted along the vertical glass walls of mosensory and mechanosensory setae (Foelix
each cage to provide a climbing surface with et al. 1975; Foelix & Hebets 2001; Foelix et
easy observer visibility. P. marginemaculatus al. 2002), and are capable of extremely deli-
were strongly thigmotactic and preferred to cate movements approximately 340(cid:4) around
rest in the tight space between two vertical theaxisoftheirbodies(pers.obs.).Whipcon-
surfaces, in this case, the space between the tactamongindividualswasacharacteristicas-
glass walls of the cage and the surface of the pect of amblypygid behavior in aggregations.
cork bark. D. diadema were less thigmotactic It was not always possible to discern whether
but also favored the vertical surfacesbetween the whips actually made contact or simply
the glass walls and the cork bark. Inallcases, passedwithinmillimetersofeachother—both
the cages and sheets of bark were sufficiently were considered to be contact. Social inter-
large that all individuals could distribute actions appear to be mediated through whip
themselves far apart from other individuals contact. Unlike the extremely rapid, directed
within the cage on suitable habitat. All ani- flicking motions with a vertical component
malswerefedappropriatesizedcrickets(Ach- that characterize whip movements in aggres-
eta domesticus, Gryllus bimaculata or G. sive interactions, particularly in intrasexual
oceanicus) ad lib several times per week (ex- conflicts (seeWeygoldt2000, 2002),thewhip
cept during periods of experimental manipu- movements involved in amicable social inter-
lation of their diet). All cages were placed on actions were typically long, comparatively
turntables, so that they could be rotated and slow repeated movements of one individual’s
individuals in all positions in the cages could whips down the length of the others whip on
be viewed with minimal disruption. Because a primarily horizontal plane. Throughout this
of the regular rotation of cages, all positions research,weconsidered‘‘withinwhiplength’’
404 THE JOURNALOF ARACHNOLOGY
ed
aracteriz(Dd)an Acari 15,2215,2215,22 15,2215,2215,22 twmohaekbteehecaronnthtaaepcptinrwodpiivtrhiidaotuenaelvsaawnrioeatrbheleecr,lfoaosnredeeevnsaoslueugnahttiiantlog-
ha ly, as a measure of tolerance between neigh-
enusedtocmondiadem Scorpiones 14,16,2514,16,2514,16,2514,16,2514,16,2514,16,2514,16 16 bpoboyfotrgdesiyn.dsApfawasrrittsichlllioutnhssaewtrnrahtitespoidgmleebptnhylgyetrhFtihoagenfuidorrentsweouha3cni–pho8sit,.nhgea‘‘rmWmwbohelyrirpee-
ea
bD length’’ was defined as the length of a single
s
hathavenwhich corpione 30303030 30 wfthrohemiipr clfiavugilenlysgaaesnxtithmeenaydlesedxt.hterWonudhgeidhpstthhweeiergrlweashmsipsesiaduseussirneogdf
eaturestegoriesi Pseudos 2,26,2,26,2,26,2,26,262,26,2,30 27 deoxidauslvpicaraoeldipfurecoremsdomcroamrrekpteradoraaicbntldievivemildyeuaamsluser.aeBssu.ortehdmfreothm-
Theoccurrenceoffaveindicatedthecat Arachnidorder pilionesSolifugae 10,13 10,13 Ts(8cgGpo)rooraMrcduodthiepuoeaadspttlnehsrrtgedhem1erlyeda-ianOntlniaeoovfdmvcfeisae2fipttrc)iormosdiannooenisngvtdnhe-eoSeDlPfaori-b.np.aolodmlfimltinfhaesgaipednnrrretdgIaimniinlivngnatitedeia(rmaumGansaeccadrl,toascinuswouiipnnlbesasslait.tmr4n—hue–ges--
1.h O 5, 5, ple taken no more than once daily at a ran-
e
pecies.CitationsareasinTableTable2.Fortheamblypygids,wonthedatainthispaper. Schizo-AmblypygiUropygimida Dd,Pm,21Dd,Pm,21Dd,Pm,21 Dd,Pm,21 dsfgetalsabheoaornoroneemoluomtsidyghuttnipthablpdhwytlrelscoiiyrn,elnioucgbrngoifieht,efnrcowmdhowgswtsiehnoreeiovsioeetlntithriuyhdhteiwapoiuncepnrwraoeroit-llrireaihy.mofednre.feetmIoftIndomrhsnoibebnpehdotePshyroityeov.ihetwrontiirehvwmmd-ngroeemud’eaesrdfagr.fraelwwsrginsopnAoiyhhnnrdwwutiiieciyppnpevimeotegsrihlluaed,aeeitgcnucnnonschrguaggioolblttwntuswahshhl,sptci,hreuinaeoodiirsingpneefrf-,
snd times a week from the time they emerged
niddiase ae 99 though 4 months old (Total sessions when
Table3.—CharacteristictraitsofthesocialarachsocialityareidentifiedforthesocialspeciesindicatePhrynusmarginemaculus(Pm)fitintotheschemeb SocialtraitArane Toleranceofgroupmembers1,3,Tendencytoaggregate1,3,Overlappinggenerationsofkin1Preysharing29Cooperativeconstructionoruseofretreat1Non-agonisticcommunicativebehaviorCooperativedefenseTemporaryaggregationsofadultsandsubadults29 25swrefps2fsybttsaheroppips,,e6ooooetcetraaar)Audinnrmnoiott.tenimoaiiinsrsen(cid:2)aaagdCdd(cid:2)esolll8wpeuesioot2gdddmrtdlhhfPmo4etmf7haayeeoelw1.ttt)benDiprnohaarr,;.maoaaganew.mcDccmGrrsGreithdooapnhdyinaurirtdbllaasioccorhorlltiailtreeigeshunuuseinydoc,ccaientdppgneaalttlttgissdeeohhmnsemvpd6edd.buemrc4iwaaaa,::oodsaeI–tcasteutluaGGninaph7hicarsgaonsrrvaeulcleDh(cid:2)bsoo(tcr’alovc,iosGTsduuaat.alaeeiuoatpptlll4rolnhurcdretttooa2nvoaau14esiccusla;slartlt,,aaphaieadwtnnesGottsmitdeneiioeenuer9om:g(cid:2)(cid:2)mrspdintgronoelasnelseage(unbietr16Tsoi,eauepts(cid:2)oc48otoensnlenwoon;;aaftdtsdr7tealt3GGiveaol,iodllwwe2vnsarrrtftncas)eshiteoothheei.thetleaeuuoeheeesnts(cid:2)sariyooppnkndes-r--t
RAYOR& TAYLOR—SUBSOCIALITYIN AMBLYPYGIDS 405
individuals toward one another was sketched our calculations. Amblypygids do not have
at onerandomly chosen time.Orientationwas scopulaandwereunabletoclimbonthesides
determined by measuring the vector anglebe- of the glass or plastic cages, so this area was
tween the direction of the young (the orien- not considered to be part of the potential us-
tation of the midsagittal plane extended ante- able space. We calculated the expected num-
riorly through the longitudinal body axis ber of individuals for each piece of bark,
toward the palps) and alinedrawntowardthe based on the size following the logic above,
closest part of the mother’s body using a pro- and then used chi-square goodness of fit tests
tractor. The mean vector angle of orientation to evaluate whether the amblypygids were
towards each adult female was calculated us- randomly or evenly distributed on the bark or
ing the Rayleigh test (Batschelet 1981). whether there was evidence of aggregation.
Whip contact among individuals was a Costs and Benefits of Aggregation.—Ma-
characteristic aspect of amblypygid behavior nipulationofSpatialandTexturalUniformity:
in aggregations. If social interactions are me- To determine whether individuals aggregated
diated by whip contact, we would expect in- to gain access to the most favorable or pro-
dividuals to position themselves closer than tected locations in the cages, we created an
oneextendedwhip-distanceapart.Toevaluate environment in which all surfaces were pre-
this, we compared mean whip length for am- sumed to be texturally and spatially uniform
blypygids of a given age with mean nearest by manipulating bark textureandunifyingthe
neighbor distance at that age for D. diadema position of the bark relative to the glass sides
using t-tests. Whip lengths were measureddi- of the cages. Both P. marginemaculatus
rectly with dial calipers and nearest neighbor (Group 3) and D. diadema (Group 4 at 14
distances were measured with a ruler through months old) were tested in ‘‘uniform’’ envi-
theglasswallsoftheobservationcage,ordis- ronments. The original cork bark was re-
tances were calculated from x-y coordinates. moved and replaced with uniform rectangular
In addition to the spatial data collected,ex- pieces of plywood (5 mm thick)approximate-
tensive behavioral observations were made ly the same size as the four walls of the cage.
throughout their lives. Observations and rec- The plywood was placed 2 cm from and par-
ords were made through direct observation allel to the walls, so that every position inthe
and by video, in order to gain a better under- cage was equally favorable in terms of tex-
standing of general behavior patterns of the ture, distance from the cage wall, and relative
amblypygids. These qualitative descriptions light levels.Withthe‘‘uniform’’plywoodset-
are reported where relevant. up, location and distance between individuals
Adult Dynamics.—Spatial data, including were recorded. To evaluate the animals’ spa-
group association, nearest neighbor distances, tial distribution, the total area of the plywood
whip length, and whether the animals were was visually divided into rectangular areas of
courting, were collected over time on the un- comparable size. If the animals were distrib-
relatedD.diademaadultsinGroup9forcom- uted at random, the expected number of in-
parison with immature individuals. dividuals found in each of these areas of the
Evidence of Aggregation in Imma- plywood would be proportional to the size of
tures.—To determine whether there was evi- that area. In Group 3 (P. marginemaculatus),
dence of aggregation by D. diadema (Group the total area oftheplywoodwasdividedinto
4at10monthsold),werecordedthelocations 12 sections of equal area. In Group 4 (D.dia-
ofindividualsonthecorkbarkwithinthecag- dema), where the individuals were signifi-
es once dailyatonerandomlychosenpointin cantly larger and typically formed larger ag-
timefor 19 days. Wepredictedthatifindivid- gregations, the total area was divided into six
uals were evenly spaced throughout the cage sectionsofequalarea.Chi-squaregoodnessof
or distributed at random, the number of indi- fit tests were used to determine if individuals
vidualsfoundoneachpieceofbarkatagiven were randomly or evenly distributed or if
timeshould be proportional totheareaofthat there was evidence of aggregation.
piece of bark. Since thigmotactic individuals Manipulation of Food Abundance: To de-
preferred to rest in the vertical space between termine if aggregation might increase the
the bark and the glass walls of the cage, we competition for food, we manipulated food
used only the area on this side of the bark in abundance fortheadultP.marginemaculatus.
406 THE JOURNALOF ARACHNOLOGY
For this experiment, Group 3 (n (cid:2) 29) was Little is known about the predators of am-
divided up into two equivalent groups, and blypygids (Weygoldt 2000; Hebets 2002). As
housed in separate 3.8-l glass cages. From 15 a first attempt to see if there is a reaction to
September through 22 October 2000, Group an animal that could potentially be apredator,
3A (n (cid:2) 15) was fed daily so that there were we introduced a generalist insectivorous,
always at least ten live uneaten cricketsinthe semi-arboreal, male lizard (Anolis carolinen-
cage at any given time. Group 3B (n (cid:2) 14) sis) as a potential predator into the cages of
was deprived of food during this time: only Groups 3, 4, and 7 in a second experiment.
two small crickets were put into the cage at The interactions between the lizard and am-
the beginning of each week of this period. blypygids were observed and videotaped for
From 25 October through 20 November, the 1 hour in each cage before the lizard was re-
feedingscheduleswerereversedsothatGroup moved. This lizard species is sympatric with
3AwasfooddeprivedandGroup3Bwaswell Phrynus, and similar in size to insectivorous
fed. Observations were made several times chameleons,gekkos,skinks,oragamidlizards
per week (n (cid:2) 20 observations) and both lo- which may prey on D. diadema in Tanzania
cationanddistancesbetweenindividualswere (where our adult specimens were collected)
recorded. For each individual, the linear dis- (Conant & Collins 1998; Spawls 2002). The
tance to the nearest neighbor was determined lizard measured 16.1 cm from nose to tip of
on each observation day. The mean nearest- tail, while the amblypygids ranged in size
neighbordistancewasthencalculatedforeach from the smallest specimens of P. margine-
individual during each experimental period maculatus (body length (cid:2) 6 mm) to the larg-
and a paired t-test was used to compare these est adult specimens of D. diadema (body
variables during the food deprived and the length (cid:2) 42 mm). Although the lizard may
food surplus phases of the experiment. have had difficultyinkillingthelargestofthe
ResponsetoDisturbanceandtoaPotential D. diadema adults, its size would have been
Predator: Since reduced predation risk is one sufficient for it to easily attack juvenile D.
of the predicted benefits of sociality for most diadema or any of the P. marginemaculatus.
animals (Alexander 1974; Krause & Ruxton
RESULTS
2002), we predicted that young amblypygids
would benefit by the presence of their mother Mother-Offspring-Sibling Interactions.—
or other amblypygids through active defense, From emergence through four months, more
earlier warning, or through aggregation. To young P. marginemaculatus were found ag-
determine the response of amblypygids to po- gregated near their mother or in several
tentially dangerous disturbances or predators, groups of closely associated siblings than
weconductedtwoexperiments.Inthefirstex- were found solitarily (Figs. 1–4). Patterns of
periment, we examined the response of a association changed little during those first
mother and her offspring to a disturbance in four months.
which the experimental cage was gently rat- Young P. marginemaculatus in proximity
tledforapproximately15seconds.Werecord- to their mothers oriented their bodies towards
ed the siblings nearest neighbor distances and her body significantly more than expected
group size for two clutches of D. diadema (Rayleigh test: Group 1: r (cid:2) 0.5704, angle (cid:2)
(Groups 5 and 6) 5 minutes before the distur- 58.5,P(cid:5)0.001,n(cid:2)67;Group2:r(cid:2)0.3672,
bance and then again 2 minutes after it. Dis- angle (cid:2) 49.6(cid:4), P (cid:5) 0.003, n (cid:2) 69). Such ori-
turbance trial replicates were done one or entation indicates that the young are attentive
more days apart between. Datawerecollected to the presence of their mother. Similar di-
for Group 6 at 1.5 and 8 months (n (cid:2) 4 rep- rected orientation is seen in courting adults
licates at each age), and Group 5 at 2 months (Weygoldt 2000).
(n (cid:2) 4 replicates). The data were analyzed Our observation of these clutches of P.
using a mixed model with age group (3 age marginemaculatus young did not extend until
levels)andtreatment(beforeandafterthedis- theyreachedsexualmaturityduetomassmor-
turbance) as fixed effects and replicate for tality from mites at nine months. We believe
each group as a random effect, followed by a that P. marginemaculatus reach sexual matu-
multiplecomparisonanalysisusingtheTukey- rity at between 12 to 18 months old. We have
Kramer adjustment. casual, unquantified observations of two ad-
RAYOR& TAYLOR—SUBSOCIALITYIN AMBLYPYGIDS 407
directly to a third group of five young on the
opposite side of the cage and repeated the in-
teraction for several minutes, beforereturning
to sit in the middle of the first group. In each
case, the young oriented towards the mother
andweretactilelyinteractive.Theinteractions
were deliberately initiated and appeared to be
affiliative behavior between the mother and
her offspring.
Immature D. diadema (Groups 4–8) re-
mainedcloselyassociatedandinteractivewith
their siblings until they reached sexual matu-
Figure 1.—Mean number ((cid:6)SD) of Phrynus rity between 13 to 15 months old (Fig. 5–9).
marginemaculatus young in Groups 1 and 2 found For the first few weeks to a month after de-
in association with their mother, in sibling groups,
scending fromthemother’sdorsumtheyoung
or solitarily. Data from Group 1 (n (cid:2) 14 observa-
clustered tightly around their mother (see
tions)andGroup2(n(cid:2)26observations)werecol-
Figs. 17, 18). The young surrounded their
lected on an approximately weekly basis from age
1 month to 4 months. mother, sat beneath her, or hid in smallcracks
on the cork bark nearby. As theymatured,the
young distributed themselves more widely
ditional clutches suggesting that when young
through the space available in the cage but
are born into group colonies with multiple
continued to aggregate in sibling groups or in
adultspresent,theyoungwerenotharmedand
groups near the mother, with only a few in-
readily interacted with other adults. However,
dividualsfoundsolitarilyuntilreachingsexual
inbothcoloniesmother-offspringgroupswere
maturity (Figs. 8, 9).
more labile than in the solo cages. Soon after
The mother’s physiological state affected
the young emerged, the mother moved away
thedurationandamicabilityofthemother-off-
from her offspring, while the young dispersed
springassociation.Basedonourobservations,
fromthenatalsiteinsmallgroupstoassociate
forseveralweekspriortomoltingsomeofthe
with older individuals throughout the colony.
maturefemalesactivelymovedawayfromher
Mothers of both species studied interacted
offspring, reduced interaction rates, and in
frequently with their offspring; often the
Group 4 became agonistic. In the wild, pre-
mother would sit in the middle of a group of
molt behavior could be a time when the fe-
her offspring and stroke their bodies with her
malepermanentlyleavestheyoung.Incaptiv-
whips. The following is a typical observation
ity, the females reinitiated ‘‘amicable’’
of P. marginemaculatus with three-week-old
(tolerant,nonaggressive)interactionswithoff-
young: Initially the adult female stood alone
spring and returned to rest in the middle of
on a section of vertical bark. She made a di-
young soon after molting.
rected walk into a group of ten closely asso-
Constant whip contact among immature
ciated offspring and gently stroked them with
amblypygids characterized aggregations. For
her whips. The young moved to surround and
immature D. diadema, the average nearest
orient to her, and stroked her in return, touch-
ingherwhip,pedipalps,andlegs.Oftheseten neighbor distance was significantly less than
young, the mother made individual contact the extended length of the whips, indicating
with seven of them over (cid:7)4 minutes. Al- that individuals positioned themselves close
though the young initially had been sitting enough for tactile interactions with their
close together, slowly waving their whips, neighbors(Figs.10,11).Nearestneighbordis-
once the adult female joined the group the tances among individuals in groupswerecon-
youngsters’ whip movements quickened so sistently small relative to the size of the ani-
that most of the young contacted one another mals. Although nearest neighbor distances
aswellasthefemale.Thenthemotherwalked increase with the increasing size of the ani-
directly to a separate group of twoyoungsters mals, the tendency to aggregate and the size
five body lengths away, engaged in mutual of groups remained fairly constant (Figs. 10,
stroking for (cid:7)30 seconds. Next she walked 11).
408 THE JOURNALOF ARACHNOLOGY
Figures2–7.—Photographsoftypicalamblypygidassociationsatdifferentages.Notetheextentofwhip
contact among individuals and the closed, non-aggressive positions of the palps. 2. Three Phrynus mar-
ginemaculatus young (Group 2) oriented to and interacting with their mother at age 2 months. Two
additional individuals are under the adult female’s left legs. 3. P. marginemaculatus adults (Group 3) on
uniform plywood bark. 4. Amicable interactions among 9-month old P. marginemaculatus from Groups
1and2(distinguishedbythepresenceorabsenceofawhitedotonthecarapace)thathadbeencombined
2dayspreviously.5.Damondiademayoung(Group5)atage6months.6.AdultfemaleD.diademaand
13ofher10-montholdoffspring(Group6).Notethatthegroupwasnotlimitedtothesmallspaceshown,
but had 4,446 cm2 of suitable bark throughout the cage. 7. Damon diadema offspring at age 13 months
(Group 4). The male on the left (wider, thinner palps) molted to sexual maturity before his three sisters.