AGGREGATIONS OF MALLOS AND DICTYNA ( ARANEAE , DICTYNIDAE ) : POPULATION CHARACTERISTICS By R

Generally web spiders are solitary animals. Usually only a single spider occupies a single web, with the relatively common exception of webs jointly occupied by a male-female pair or females with their young. In most species, occupied webs are usually not connected to other occupied webs by silk lines. However, aggregations of various types and degrees occur in some species (reviewed by Buskirk, 1975; Krafft, 1970; and Shear, 1970). These include cases in which occupied webs are in close proximity (sometimes touching), but occupied by single spiders, as well as cases in which single webs are occupied by groups of spiders of all sex/age classes. The species involved are sometimes referred to as “social spiders” (e.g., Burgess, 1978; Kullmann, 1972). At present relatively little is known about the social adaptations of these spiders. Some of the first things we would like to know about animals that occur in aggregations are the size, composition and spacing characteristics of naturally occurring groups. Data concerning these basic social characteristics will be presented in this paper for several species of dictynid spiders. The Dictynidae are a group of web-building cribellate spiders, most of which have body lengths less than 5 mm. In the closely related genera Mallos and Dictyna, there are species with different types of social organization. Most are solitary; at least three are communal, territorial; and at least one is communal, non-territorial. Each individual web of a solitary species consists of a mesh in which prey is captured and a nest within the mesh. Communal, territorial


Introduction
Generally web spiders are solitary animals.Usually only a single spider occupies a single web, with the relatively common exception of webs jointly occupied by a male-female pair or females with their young.In most species, occupied webs are usually not connected to other occupied webs by silk lines.However, aggregations of various types and degrees occur in some species (reviewed by Buskirk, 1975;Krafft, 1970;and Shear, 1970).These include cases in which occupied webs are in close proximity (sometimes touching), but occupied by single spiders, as well as cases in which single webs are occupied by groups of spiders of all sex/age classes.The species involved are sometimes referred to as "social spiders" (e.g., Burgess, 1978;Kullmann, 1972).At present relatively little is known about the social adaptations of these spiders.Some of the first things we would like to know about animals that occur in aggregations are the size, composition and spacing characteristics of naturally occurring groups.Data concerning these basic social characteristics will be presented in this paper for several species of dictynid spiders.
The Dictynidae are a group of web-building cribellate spiders, most of which have body lengths less than 5 mm.In the closely related genera Mallos and Dictyna, there are species with different types of social organization.Most are solitary; at least three are communal, territorial; and at least one is communal, non-territorial.Each individual web of a solitary species consists of a mesh in which prey is captured and a nest within the mesh.Communal, territorial Present address: Department of Zoology, University of Canterbury, Christchurch 1, New Zealand.
Manuscript received by the editor June 10, 1978.
species (A/. trivittatus Banks, Dictyna c Banks, D. albopilosa Franganillo) live in web complexes.Web complexes consist of variable numbers of web units, each consisting of a nest and a mesh; and web units are connected to each other by strands of silk in the interstitial web.Evidently, web units are treated by the occupants as defended territories (Jackson, 1978a).In the communal, non territorial species ( M. gregalis Simon), groups of spiders liv communal webs which consist of surface sheets with holes leading into the interior.The interior contains nests, tunnels and supporting lines connecting the sheet with the substrate beneath (see Burgess, 1976;Diguet, 1909aDiguet, , 1915)).Communal webs are not divided into defended units as in web complexes.More detailed information concerning social organization, web characteristics, and habitats is provided elsewhere (Jackson, 1978a).

DATA FROM NATURAL POPULATIONS
For census areas, locations were chosen which seemed to have particularly high densities of spiders.The census areas for Mallos niveus O. P. Cambridge were one-dimensional (horizontal lines).All other census areas were rectangular.For D. calcarata and Censuses 3-6 for AT.trivittatus, complete web complexes of relatively small size were used; and the boundaries of the census areas were chosen to closely correspond to the edges of the web complexes.The distance from each spider in the complex to the nearest conspecific in the vicinity but not in the complex was greater than twice the diameter of the complex in each of these census areas.For Census No. 1 and 2 for M. trivittatus, each area censused was part of a large web complex.
The corners of each census area were marked with pieces of tape.Each web or web unit was completely searched, a process that destroyed the web unit.The occupants of the units were recorded; and a piece of tape, with a code number, was placed at the location of each occupied nest before it was destroyed.The distance of each of these nests to its nearest neighboring nest was measured.
The exact locations of spiders on the sheet and interstitial webs proved excessively difficult to measure accurately.This was because they were frequently obscured from view by silk and debris; and the process of locating spiders inevitably disturbed them, leading to changes in their positions before they could be recorded.Only the component of the web (nest, mesh, interstitial web) occupied by the spider could be recorded with confidence.When counting the number of spiders per web unit, those in the interstitial web were included in the record of the web unit to which they were in closest proximity.
Spacing and density data were not collected for D. albopilosa.The web complexes of this species were particularly difficult to disassemble in the field because they were three-dimensional and located on the leaves, stems and roots in dense growths of herbaceous plants.Those of D. calcarata and M. trivittatus, in contrast, were two-dimensional and usually built on more accessible flat surfaces.
At San Anton Falls (Morelos, Mexico), 23 occupied web units of D. albopilosa were searched, and the occupants were recorded, with no attempt to assess density or spatial relations.In addition to the spiders in the census areas, the occupants of another 48 web units of D. calcarata and 152 of M. trivittatus were recorded.Each of these was in the general vicinity of the census areas.
Data for M. trivittatus were collected in June; those for M. niveus, D. calcarata, and D. albopilosa were collected in September.These species seem to feed predominantly in the late afternoon and early evening (Jackson, 1977a).Data in this study were collected at mid-day, when the spiders were less active.

LABORATORY STUDIES -• Mallos gregalis
In the laboratory, M. gregalis in large communal webs were maintained on a diet of houseflies ( domestica), provided at approximately 5-day intervals.Temperature was maintained at approximately 24° C, and the light cycle was approximately 13 L: 11D.These large colonies were begun from fewer than 200 spiders collected by Burgess (see Burgess, 1976) in Guadalajara, Mexico, 3 yr previous.The webs in the laboratory were on plants and other objects, and they were not enclosed.
Two census methods were used.1.Small discrete communal webs and relatively discrete portions of larger communal webs were completely disassembled in the spring and early summer.All occu pants were recorded; and the condition, structure, and size of the webs were recorded.2. The surface of a large communal web, at least 2 years old, was censused with 30cm transects in the fall.Each transect consisted of a string laid haphazardly onto the web, and all

Psyche
[March spiders within 1 mm of the string were counted.The web was arbitrarily divided into 3 sections.Two transects were placed onto each section on each census-day; and there were 11 census days, each 2 to 3 days apart.A "oensus" was slfefmed as the sum of the spiders found on the two on a given, section? on a given census-day.
A r e a s covered by a single jM ^m ous web complex.The bottom, of theAEflvert was covered by water; and in the immediate, vicinity ^ the water, web units were scarce, Slffixteting fhise parwSflthé culvert, the web complex was conservatively estimated to bfe ?9 m2 in area.Census Areas NoifTf' and 2 were eadh ilppfeAhe culvert, within 2 m of opposite entranced and 1.5 m above p # bottom of the ci^^rt.Census Area No. 1 was 56 cm X 50 fm; No. 2, 60 ^( ^5 6 cm.jpl^ÉÔ^^Af^a No. 3 was a web complex (54 cm X 42 cm) 30 fel'CgWlhe culV^rJ,, On a large boulder (ca. 2 m high) beside the crcekjgfj^ffi boulder was overhanging the ground, forming a shallow cave (ca.l;tt from ground to top; 0.5 m from entrance to back).The web complex was on the upper surface of the cave.

Mallos niveus
Census Area No. 4 (56 cm X 32 cm) and No. 5 (56 cm X 56 cm) were web complexes on rock ledges near Estes Park (Colorado, USA).Each was 1 to 2 m above the ground.Census Area No. 6 was a web complex (124 cm X 34 cm) on the upper surface of a shallow

Dictyna calcar at a
Census Area No, 1 (24 cm X 24 cm) and No. 2 (54 cm X 46 cm) were web complex#!pn the external walls of buildings in Chapala (Jalisco, Mexico),

M. t f t t a t u s and
were relatively similar (Table 1).Since the M.'WÈÊÊ^ aggregations were linear (one-dim^^wnal) it is problematic to ip|npare densities in this species with the demsilfe^Sund in the other species.The in the census areas were oriented s i l l that the shorter distance %as vertical.-ÿi^Éaifli^S^S^WWable 1 were obtained by multiplying thé 'ïèr%ih Si ili'Ilaggre^âlMSs ^lly 3 j^n, the mean widi'% ^ ' ^L ^#fP ^^ebs found On buiIdihg^(Ji §|^i$fr, 1978a).M. gregalis tends to liV# in higbpf densities.The webs censused by MSilod No. lM 'able 2) had a mean densityQ t^®[).66spiders per cm2 and per cmMffi Based on manf sours spent in the Turkey Creek, it seems that the c^feMes obtained in Cft^us Nos. 1 and 2 for M. irPH9$eitus were represent|ttj^&f , the large Simplex in the culvert.Using the mean ofihe twO^^nlitifeîpr spider^ (14.23 per 1000 cm2), webs (9.1# per 1000 ^i ^| a n |'4f^ppfcd \^te^8.79per 1000 cm2) as estimates for the densities fdf; the complex as a whole, this web complex coptfthledi 10,200 spiders, 7,700 webs, and 6,500 occupied webs.This was bffijjsflffi large'sf Aveb complex of this or any other species in this study.
Web units of M. tnvittatus tend to be 5 cm X 4 cm; those of D. calcarata tend to be 3 cm X 2 cm; and webs of M.
on walls of buildings tend to be 4 cm X 3 cm (Jackson, 1978a).In the culvert on E. Turkey Creek, the interstitial area between web units tended to be quite small, since the mean for nearest neighbor distance was not much greater than the diameter of web units (Table 1).In all other censused web complexes, web units tended to be 2 to 4 web unit diameters apart; and aggregated M. niveus webs tended to be 2 to 4 web diameters apart also.Some of the larger communal webs of M. gregalis in the laboratory had surface areas and volumes estimated to be as large as 1300 cm2 and 1500 cm3.If the densities in these large webs were comparable to those in smaller webs, they can be estimated to have contained 300 to 700 spiders.Based on the descriptions of Diguet (1909a, b) and Burgess (1976), 4 m2 will be used as an approximate estimate for the surface area of larger communal webs on trees in Mexico.If densities in nature are comparable to those in the laboratory then these very large webs may have contained as many as 20,000 spiders.

WEB OCCUPANTS
Solitary species, including M. niveus, usually occurred one spider per web (Jackson, 1978a), and this was true of all censused webs in the aggregations of M. niveus.However, each occupied web unit of the three communal, territorial species usually contained a set of several spiders.These could be almost any combination of females, males, and immatures of varied sizes (Table 3), except that web units were never shared by two adults of the same sex.In the only instances in which male-male or female-female pairs were seen together in the same web units, aggressive interactions took place, after which one individual departed.In contrast, large numbers of M. gregalis belonging to all sex/age classes shared the same communal webs without aggressive behavior or cannibalism.

DISTRIBUTION OF WEB OCCUPANTS
Large proportions of each sex/age class of M. trivittatus occupied the nests (Tables 3 and 4); but adult females were especially prone to be inside nests, and there was a preponderance of immatures in the interstitial web.These data were consistent with more casual observations of M. trivittatus in other locations and of the other communal, territorial species.During the day, M. niveus and other solitary species were most often, but not always, inside nests.All M. niveus in censuses were inside nests.
Many hours of casual observation in the laboratory gave the impression that adult males and immatures of M. gregalis were more prone to be on the outer surfaces of webs, and females were more prone to be in the interior.(Data related to this will be presented below.)Î).The average adulftfffifratk» was approximately^7.For the 33 transect c$&susei (Method No. 2), popula^fefeiMpPsitiOn was 12.3 ^^S ^tfe adu^y w a les.3.0 ± 3.85% adult males, SÉûQjlg.12.É^ÿlaree immatures, and 46,2 ± 13.1 MpBMrali immatures (adult sex ratio, ca. with a tôttï^f 1^^^B.-14spiders occurring $|g each census.The three sections of f$ef|$e,lh were comparable.

Sex/Age
The occurrence in certain areas of very large populations of some species, especially M. g^e^^^^nd M. trivittatus, and the great densities of spiders within these populations contrast markedly with the majority of %nd Mallos species.Most species are solitary, with individuals living in relatively widely spaced indi vidual webs on stems and leaves of vegetation (Jackson, 1978a); and although density censuses were not carried out for these, our impression is that an area of 50 to 100 m2 in a population of a solitary species would usually contain fewer than a dozen indi viduals.In contrast, the largest web complex of .trivittatus occupied approximately this much space and contained more than 10,000 individuals.

A related comparison can be made by looking at the amount of web area or volume per individual spider for different species.
There is a trend toward decreasing web surface per spider as one goes from solitary to communal species.Individual webs of solitary species tend to have a surface area of 35 cm2 (Jackson, 1978a).Since each web unit in web complexes of communal, territorial species tends to contain 2 or 3 spiders (Table 3), and since web units tend to have a surface area of 4 to 21 cm2 (Jackson, 1978a), there tends to be 1 to 7 cm2 of surface area per spider in these species.In the laboratory, in communal webs of M. gregalis, there tends to be a surface area of 2.1 cm2 and a volume of 6.4 cm3 of web per spider.(Volume was not calculated for other species since their webs tended to be largely 2dimensional; Jackson, 1978a.)The influence of prey availability on these trends deserves atten tion.Diptera of the same approximate size relative to the spider seem to predominate in the diet of species of Dictyna and M allos   in Chapala, ^Ifeierging fWm nearby Lake Çhapala in thè 'k?fee afternoon and early ëvening.Diptera Iftfe also numerous in the viciâky of web com plexes of àt San Anton Falls.In general, wherever iwlrfr 'werC*freb completes, flê r ^% e ^a lso numerous Diptera.
However, the question of why'#ome-species live in large, dense populations, while Others do not, cannot be answered simply on the basif ©f prey densities in the habitats o L ^fer e n t species.Riechert and Tracy (1975) l ^e shown relationships between density within Spider populations and §É §f availability.However, they found prey f illa b ilit y to depend not or® .on the absoute abundance of prey but also oil factors that influence how the spider experiences prey abundance, especially its th^çpal relations with its environment.
Another consideration is that species of and Mallos with differing aggregating tendencies may occur; side-by-side in the same habitats (Ja^poitj 1978a).For example, M. niveus were found in individual webs on the same trees with web complexes of M. trivittatus in the Çhiracahua Mountains.It would seem that dictynid species that routinely occur in aggregated states (communal, terri torial and communal, non-territorial) and those that generally live more widely dispersed (solitary) are somehow adapted to exploit different sets of resources, but we have no clear insights at this time concerning what these different resources might be.
The adaptive advantages and disadvantages for animals related to living in groups have been subjects of considerable interest in recent years (see Wilson, 1975).For a review of ideas concerning spiders.see Buskirk (1975) and Lubin (1974).With the dictynids, we need to compare the varied species with respect to the importance of each potential advantage and disadvantageflgiven the nature of the resources each species exploits.

PHENOLOGY
Most Dictyna and Mallos probably have an annual life cycle in nature, with adult females and males present during spring, summer and/or fall (Chamberlin and Gertsch, 1958).The mating season for Mallos species generally seems to be later than that of Dictyna species, although the season for some species may last many months.Since both adult males and adult females were found, evidently each census was carried out during the mating season of the species involved, although not necessarily during the peak of the season for all species.No doubt, if censuses had been undertaken at different times of the year, different ratios of each sex/age class would have been found for each species.
Seasonal changes in the ratios of the different sex/age classes in laboratory populations of M. gregalis apparently occur, although data have not been collected.Each sex/age class was found throughout the year, but males were less numerous in winter than in other seasons.The differences in adult sex ratios in the two types of censuses of M. gregalis may be a reflection of the tendency, noted during casual observation, for females to predominate in the interior of webs, with males predominating on the exterior.How-ever, another factor that must be considered is that the two types of censuses were made at different times of the year, and differences may be influenced by phenology.

causes © D istribution patterns w|^h^I;'*|veis
In all species, eggs tend to be oviposited'the vicinity of the nests, and the teîÂàpsey for females to be found in nests is probably at least partly related to this.hav^^^advertising routine, as part of their daily activity pattern, in whiefe they perform behavior referred to as "pluck-walking."Females sdlM. to indicate their receptivity to pluck-walking males by failing to run away (Jackson, 1978b).We do not know the factors that determine female receptivity, but perhaps receptive females are more likely to be on the outer surfaces of webs.For example, females might be unreceptive near the time of oviposition, and oviposition takes place in the interior of the web.Although the difficulty of observing behavior of spiders in the interior of webs should be kept in mind, it is of interest that the majority of observed instances of pluck-walking involved males on the surfaces of webs.
We have no information at this time concerning how new communal webs are formed by M.
One possibility is that new webs are founded by single adult females or by groups of several adult females.In the laboratory web censuses (method 1), the webs with fewer spiders (predom inant adult females) may have been ones that were newly founded^,with more spiders and a greater proportion of smaller immatures may have been older webs in a period of rapid growth,.Further stw S y^^ecially in the natural habitats of these spiders^» tuéüessary.Information is also lacking on the formation tifweb com p lies by calcarata.

CAUSES OF SKEWER :i f p RATIOS?
The causes of the ^^^fé-rarity of males in all species investigated are unknown.We dJpbt know the sex-ratios atM iifeing.If maternal investment in progeny» the.tvfe'sexes is equal, we Spould expect a 1:1 ratio (F is h e r/1 ^;lp iit s^ffeïpilfilWv ' 1967;Trivets and Willard, 1973).There are a number o fftf^r s that might s&w the adult ratio in favor of females éÿfén if the rak-io at hatching is 1:1.Ë.arli#t |t-lÿdiel (Jackson, 1978b) suggested that adult mates, of solitary and com munal, territorial species are relativgly ftomadiife, expending con-siderabl^,>iy^ and energy wandering about stafehklg for female^ As a result, a sizeable proportion of the males in population^ Of these species might not be found in censuses of webs.Also, the nomadic character of males might subject them to earlier mortality from predation, starvation and other factors.Shorter male longev ity would skew the adult se l rati© in favor of fdrfiales.However, mortality factors such as predation would seem less important for M. gregalis populations in the laboratory;, gpfc the sex ratio was skewed in favor of females here also,, suggesting that mortality factors of a more intrinsic nature might be involved.Females of each investigated species oviposited several batches of eggs over a period of time.In contrast to females, males may be adapted to a lifestyle that emphasizes courtship, mating, and searching for females, in conjunction with greater vagility and smaller size, at the expense of maintenance functions that would serve to prolong survival.
from complete disassembly of 19 M. gregalis webs or subunits of webs (census Method No. 1, see text).Area and volume are estimates, since borders of webs were irregular.Small immature: ca. 1 mm in body length (color: uniformly brown).Large immature: body length ca. 2 mm (gray markings present on body).Adult females: ca. 3 mm.Adult males: ca. 3 mm.Web class A: 3-D and lacking evidence of deserted areas.B: 3-D and containing conspicuous deserted areas.C: 2-D and lacking conspicuous deserted areas.D: 2-D (but not as flat as C) and containing conspicuous deserted areas.E: 3-D, small discrete web on irregular substrate.A-D: census areas were subunits of large webs, except for some of C that were discrete webs.3-D: web occupies relatively much space in each of 3 dimen sions.2-D: web confined predominately to 2 dimensions ("flat")."Deserted" areas: no longer occupied by M. gregalis, not substantially adhesive, and covered by fly carcasses and dust.Note: as population size increases, composition of population shifts toward decreasing percentages of females and increasing percentages of im-£ matures.Kendall rank correlation (Sokal and Rohlf, 1969) for increasing percent ages of small immatures with increasing population size (P < 0.001).
vicinity of M. gregalis webs in Mexico during the rainy season.Studies are needed to determine whether Diptera populations are larger and/or more predictable in habitats con taining communal webs of M. gregalis compared to neighboring areas without communal webs.Great numbers of Diptera were noted in the metal cuIVtrt in Arizona that contained the enormous web complex of M. trfrittatus.Another Consideration in this case is that flies entering tljt cuN^rt may be especially vulnerable to capture, sine# inside the thsif: were almost completely surrounded by web: Great masses of eewatocerous flies were active in the vicinity ©Ï wSw p3n|plexes nèsts p a y be optimal resting sites with *3ggpect to p ro te c tic ^^o m predators and parasites y ackson, 1978*-).If space wti&in a nest Is ^èeOm^anied $gg aggTesbehavior, which' f^K !lik e lv , tentorial 'fpedilp then Êçfpâles may l i f ^a n advantage related to tfifir larger Females in e ||lp species have rl|e largest body size of any seÿg^ge class.C^»pared to males, females pay be more sedentary^ and rf Wronger tend^fe^lo 'rtcsts may be r^te^ÉP^B^This might also apply to ^^^lparison of females wit.K'lmmat^Mflf the im m a tu ^^e the f e ^^^^S w e A the life historié o f spe cies.Also, by virtue of tf^r smaller immatures may be , safer from predation on^fej& r ^S w j^S r st-rtial web l$iln females, H$nce they under strands of sIÉfi particles of debrt e etc.. are tord small td be* eifji^^w#rfem ales.As a reauBjËbbfté-.might be.reMér.^ef^^fa^ip^^pMagainst immatures fâ$at remain o u t a d id ^^A i& ^redÿlJ larger femateS '-fl^feremain outside rtesfe^i females IlÉfém §&be be in theitiletiiot of webs of M. r0mgMis, we might txpeet ^a ^r 'iWpepd'ConsideraMe time search^ ing for and courting females iùf,f^|nterirpr of webs.Instead, more male^^Éfe; to be on: :^^xtefi®r surfaces of the webs.In gregalis, unlike M. tfiykfatus?and D. c&îéÊmta, it$ §'.ftt&gi3£e of females and ^ïk'ipun by females atf^y#T-iliëaiM#fefecourMbfp behavior.Instead, malefe.

Based on the spiders in the two censuses in the culvert at E. Turkey Creek, this large web complex was estimated to have had a population consisting of 2 111 adult females» adult males, and 64% immatures. The adult sex, ratio (Male: Female) was approxi mately i;3. Based on ..the censuses at Chapala, populations of D. calcarata in web complexes Q» buildings were composed of 14% adult females, 5 ^ .adult males, and 8 lp|irnmatures; and the adult sex ratio was 1:3, jts for ^ t^A ta tu s No ma consistes of M. r t i v é u s; . Genera|EM É âmâll numbers of males of the jgplitary species were found',.were predominantly ones in Pvebs with females (Jackson, (^7 7 ^; Based on the 19 ^to u ses of fyt. ere^Wm bv Method No. 1 . (disassembly), population composition was 29.8 ||j|25.49% adult H m ales
, 4 S y E O t ^ adult males, 17.6 ± 12.17%

Table 3 . Percentages of web units (in web complexes) occupied by different sex/ age class combinations. Data based on nests alone
(M. trivittatus)