A solitary mud�daubing wasp, Brachymenes dyscherus (Hymenoptera: Vespidae) f:ron BrazH with evidence of a life-cyde polyphenism

Abstrad: Nests (n=60 ) of a solitary mud-daubing wasp, Brachymenes dyscherus were conected at Fazenda Santa Carlota, Cajuru, Sao Paulo, Brazil ín 1995 and 1996 . The multi-celled mud nests were constructed on the interior walls of abandoned houses in dry and shaded places. Nests were composed of 1 to 4 layers with cens constructed in two paralle! series. The number of cens per nest ranged from 3 to 62. Cells containing prey (larvae of Lepidoptera) numbered fram 4 to 29. B. dyscherus is a uniyoltíne species. Adults emerge from September to December, and nesting talces place at the beginning of the following year; however, the developmental period for some immatures with proloFlged diapause l3.Sted up to 596 days. The total duration of immature stages �n nests collected in 1996 was less for males than for females. The sex ratio for the total populatíon was 1 .5 males: 1 female with males emerging earher than females. A life table was constructed, and details of lhe life cycle of the wasps and parasitoids are presented. The mos! common mOl1ality factors were either endogenous or due 10 Melittobia sp. (Eulophidae) .

divides Vespidae into the subfamilies Euparagiioae, Masarinae, Eumeninae (solitary or presocia!wasps), Stcnogastrinae, Polistinae and Vespinae.Most species in the Vespidae belong to the subfamily Eumeninae, with more than 180 genera and 3.000 species (Carpenter 1986) occurring throughout the world.The eumenine female preP it: es a brood cel!, lays an egg, then she qUlc!fy makes several foragmg tnps, stonng the gathered food with her egg.When enough food has been placed in the cel! far the complete growth of the larva, she seals the cel! and begins work on another.The females of a few solitary vespids may oviposit after they have stored the larval food and sorne are progressive provisioning during the early larval period (eg, see Cowan 1991).
In the eumenines, the genus Brachymenes contains two species.B. dyscherus is divided in two subspecies: dyscherus (Saussure 1852), restricted from Pan ama to Argentina, and catarinae Giordani Soika 1990, collected only in Santa Catarina State, Brazil.The second species is B. wagnerianus (Saussure 1875), occurríng in Panama, Venezuela, Colombia, Equador and Pem.In Brazil, B. dyscherus dyscherus occurs from the State of Pará to Río Grande do Su1 (Giordani Soika 1990).This species, according to Bertoni (19 18, in Giordani Soika 1990), builds its nest with mud on the interior waUs of man-made structures, natural rock walls, or dead trees, Nest dimensions are 40 x 80 mm, and nests exhibit a rough and irregular surface.In addition to specíes that construct their nests with mud or masticated plant material, there are species of Eumeninae that excavate the soil or occupy and modify preexisting cavities (Iwata 1976).
The purpose of this paper is present biological data on B. dyscherus, a common, yet poorly understood species.

MATERIAL AND METHODS
Study site: The study was carried out at F azenda Santa Carlota, Cajuru, SP (210 17'-18' S, 470 12'-18' W), Brazil.The farm, with an area of 6.300 ha and altitudes ranging from 540 to 944 m, is drained by several streams.Approximately 2.200 ha of the total area are covered with isolated patches of native vegetation (semideciduous forests, riparian forests, cerrados (s.l.) and cerradoes).In addition there are areas of secondary growth in various stages of ecological succession.Rhodic Ferralsol and Orthic Ferralsol soils predominate on the farm (Oliveira et al. 1983).The study area contains sorne abandoned buildings surrounded by abandoned orchards, pastures and cultivated land (mainly sugar cane).
Two seasons, distinguished by temperature and precipitation, occurs in the study area.From May to August there is a cool, dry season, with mean monthly temperatures ranging from 18 -22.1°Cand precipitation from O -81.7 mm.From September to April there is a hot, wet season with mean monthly temperatures ranging from 19.9 -26.1°C and precipitation from 35.3 -422.6 mm.
Nests: A total of 60 nests of B. dyscherus ( 8 on 7/7/95, 28 on 7/13/95,8 on 8/15/95, 8 on 4/18/96 and 8 on 8/2196) were collected using a palette knife and taken to the laboratory.In sorne nests, irnmatures from the first layer of brood cells (always in contact with the substrate) were exposed, in which case the immatures were collected and placed in labeled individual vials and identified.In the laboratory all nests were placed in plastic boxes (10 x 10 x 3 cm gerbox-type) covered with same material and together with the vials were kept at room temperature until the adults emerged.After emergence the nests were opened to analyze their cells and structure.The prey from cells in which the irnmatures did not develop were collected and counted.In sorne nests the following measurements were taken: length, width and height of the nests and of the cells of the frrst layer.The thickness of the walls between the cells and rows of cells of the frrst layer were measured in addition to the distance from the rows to the nest covering (Fig. lA).The diameter of adult emergence holes and of the orifice left by female for provisioning the cell was also measured (Fig. lB).
The greatest width of the head (GWH) determined the width of produced individuals.A life table was constructed (for nests collected in 1995 and 1996) showing the various causes of mortality at different stages in the life cycIe and the relative number of immature individuals dying from each known cause.Voucher specimens were deposited in the Entomology Collection of the Department of Biology, FFCLRP-USP.

RESULTS
Nest Construction and Architecture: The nests are constructed with mud of one or more colors on the walls inside abandoned houses, located at least 1 m from the floor in dry locations with little light.Of 60 collected nests, five had a cell in construction, and 15 had finished cells, either empty or in the initial stages of provisioning.
The occurrence of abandoned balls of cIay next to or even away from the nests suggests that before starting nest constructiolJ, the wasps may sometimes initiate nests and then abandon the site.The nest is constructed with the longer axis perpendicular to the ground with all the cells joined to one another.The three first cells are constructed one beside the other (first layer), and then one or two over the first ones (second layer).After this, two parallel rows of cells are constructed from bottom to top, always starting from the substrate surface (Fig. 1 C).The second layer is constructed over a cell row in the first layer, and then a third or a fourth as may be the case (Figs.ID, E, F). (Aquí fig. 1) After this, the female retums to the second row of the fmt e layer and repeats the process.Of the total number of nests analyzed, 3 had only one layer of cells; 29 had two; 20 had three, and 8 had four layers.The distribution of the brood cells in superimposed layers (cross-section of nest) and in the two parallel series (ventral view of nest) can be observed in Figures 2A and B, respectively.
Nests collected on 4/18/96 were observed in construction on 3/27/96.On trus day, 1 never observed more than one female per nest.Sorne of the females were collecting mud nearby for nest construction.This behavior consisted of two distinct procequres.Initially, the female leaves the nest toward a small puddle of water, flying soon after to a place with loose earth (approximately 2 m away).The second procedure entails the female regurgitating the water, mixing it with the earth and later retuming to the nest with a ball of mudo Evidently females can construct more than one celI per day.A nest with five cells that was observed on 3/27/96 being constructed by one female, had 41 closed cells when it was observed a second time on 4/18/96.Females fulIy construct and provision a cell before working on another.
Once completed, the nests are totally covered with mud, resulting in a distinctive  shape.The surface is rough, and in sorne cases it is possible to distinguish the balls of mudo The width of this covering, 0.96 cm, was measured in one nest.GeneralIy, this covering made it impossible to distinguish individual cells (Fig. 2C).DUling nest construction, a group of closed cells may or may not be covered, because sorne of the nests collected with cells that were abandoned, partially constructed, or those being provisioned were almost completely covered.Fifty-four nests had varying dimensions depending on the number of cel! layers present.
The maxímum mean nest length increased with the number of layers.The same occurred with respect to maximum mean nest height and number of layers (Table 1).
The number of cells per nest varied from 3 to 62, with most nests having from 3 to 25 (Fig. 3); in nests with 4 layers, the maximum number of cells counted was 23 for the first ¡ayer, 20 for the second, 16 for the third, and 5 for the fourth (Table 2).With respect to the first cell layer (the one in contact with the substrate), the thickness of the dividing walls between two contiguous cel\s (lateral walls) varied from 1.1 to 8.3 mm (x -= 2.6 ± 1.2 mm, n=146); the thickness of the walls between the celIs and the nest surface (external walls) varied from 1.6 to 13.8 mm (x= 6.7 ± 204 mm, n=148); the thickness of the walls between the two rows of cells (internal walls) varied from 1.5 to 9.1 mm (x= 3.5 ± lA mm, n=83).The cells had a maximum length varying from 13.0 to 23.8 mm (x= 17.8 ± 2.2 mm, n=121), maximum height from 6.3 to 11.5 mm (x= 8.9

Number of cells
Cells were provisioned with paralyzed caterpillars of Lepidoptera, and the number of these per cel! varied from 4 to 29 (5(= 9.6 ± 4.8, n=93).Most cells contained from 4 to 13 prey (Fig. 4).The cocoons, milkish white and more or less transparent, were affixed to the cell walls (Fig. 2 B).Whole prey or parts of prey, together with larval feces, were found adhered to the external surface of the cocoons.days (x= 233.3 ± 44.1 days, n=50), and for females it was 167 to 320 days (x= 263.9 ± 56.9 days, n=51), again significantly different (Mann Whitney, Z=-4.41; p<0.05).
B. dyscherus is a univoltine species.Most larvae develop into adults during the same year of nest construction, with the greatest number of emergencies occurring from September to December(Fig.5).However, in one nest collected on 7/7/95, and in five nests collected on 7/13/95, sorne individuals had much longer development periods, evidencing polyphenism.The length of this period, ca1culated between the nest collection date and the emergence of the adults, varied fram 343 to 596 days (n=17).In addition to these cases, two others should be pointed out: a prepupa fram one nest collected on 7/13/95 died in July/98, and a prepupa from another nest collected on the same day pupated in January/98.Soon after, this pupa died.
Moreover, in all of these nests there were emergencies with normal immature stage durations.If we take into consideration that the nests may have been completed months before the collection (in 1996 the nests were completed before the cold season, or in other words, April), the duration of these immature stages may be even longer.The emergence of males began before females (Fig. 6).
Adults size, sex ratio and mortality: Although there was sorne overlap in size distribution, the females with the GWH ranging fram 3.01 to 4.29 mm (n=30) were significantly larger than males whose GWH ranged from 2.72 to 3.94 mm (n=30) (Mann Whitney, Z=5.82; p<0.05) .
The sex ratio determined for the total population produced from the nests was 1.5 males to 1 female which is significantly different from 1:1 (X 2 = 31.19;df= l; p<0.05) (Tab.3).For the 44 nests collected in 1995 and the 16 in 1996, the sex ratios were 1.7 males to 1 female and 1.3 males to 1 female, respectively, which are all significantly different from 1: 1 (X 2 = 28.02 and X 2 = 4.32, respectively; df=1 ;p<0.05).Mortality from developmental failure and parasitoids was observed in 40.7% of the provisioned cells being more accentuated in the first layer.Lost samples, cells accidentally damaged and empty cells co r responded to 4.4% of the cells (Table 3).

Life table and mortality factors:
Life tables were constructed for B. dyscherus on the basis of 1177 cells from the nests (750 in 1995 (Table 4) and 427 in 1996 (Table 5).Unidentified parasitoids (11 cells) and lost samples (27 cells) were not included in the total.When the exact cause of death was indeterminate, it was classified as endogenous mortality.The endogenous causes are probably the result

DISCUSSION
The major behavioral change of many aculeate Hymenoptera relative to their parasitoid origin was the evolution of nesting behavior is thought to play an important role in reducing immature morta l ity (Evans 1977).Iwata (1976) c1assified solitary Vespidae into three categories with respect to nesting habits: (1) burrowers -excavate nests in soil, (2) renters -occupy and modify preexisting cavities and (3) builders -construct their entire nest from materials such as mud or masticated plant material.The construction material and form of nests are influenced by the availability of nest sites and construction materials, as well as the ability of particular designs to thwart nest parasites and predators (Cowan 1991).
As observed here in B. dyscherus, most mud-daubing vespids obtain their mortar for nest construction by drinking water at one place and then flying to a site of dry, c1ayey soil where they bite at the ground, regurgitate water, and mix the resulting mud to the proper consistency (lwata 1938, 1939).They then fly to the nest with the globule of mud and incorporate it into the nest structure (Cowan 1991).Because the exposed m u d nests of sorne solitary vespids are quite durable and do not wash away in rain, it may be that saliva is added to the water and soil to strengthen the nest (Isely 1913).Often nests are located near a source of water (Isely 1913, Markin and Gittins 1967, Freeman and Jayasingh 1975), and Gess (1981) has argued that sorne wasps may be limited geographically and seasonally by lack of water for mudo Thus, mud nests are commonly found around human habitations because these are as generalIy provide an abundance of suitable nesting and mud collecting sites (Smith 1979).
Like B. dyscherus, other eumenine species apply additional material to the outside surface of the nest after nest construction and provisioning (Williams 1919, Jayakar andSpurway 1965).In the case of the sphecid wasp Sceliphron laetum (F.Smith 1856) (similar to B. dyscherus), the entire nest is plastered over with a thick layer of mud and raised ridges up to 8 mm thick, once a certain number of cells have been completed (Srnith 1979).
According to Brooke (1981) the females of eumenine provide cells with many prey items when prey are small, but females provide only a few when the prey are large.Th is may explains the great variability in the number of prey collected by B. dyscherus in the present study.
In solitary vespids, the pupal stage can last 9-22 d ays, depending on the species and sexo Females require more time (Krombein 1967).I also observed this variation in males (16-25 days) and females (19-28 days).The occurrence of prepupal diapause (leading to a very prolonged totál development period for immatures) observed in nests of Monobia angulosa Saussure 1852 (Camillo et al. 1997) collected in the same area as the current study, was also observed in B. dyscherus.These authors showed that those species have two different strategies in response to adverse climatic conditions.One of the strategies is to remain for a long period of time in the prepupa stage, and the other, in the adult stage.This can be proven because new nesting was only observed in the following year, although emergencies had occurred after the month of August.Jayakar and Spurway (1965) observed that sorne wasps tolerate weather adverse seasonal conditions in the tropics with a prepupal diapause similar to that of temperate-zone species, but individuals of other species simply seem to disappear during such unfavorable times.Maybe they migrate to more favorable habitats or seek shelter as adults and remain inactive (Cowan 1991).
B. dyscherus is a univoltine species, but in other tropical wasp species that enter diapause during the dry season (Jayakar and Spurway 1965), the numerically ma1e-biased sex ratio in the diapausing generations (Jayakar and Spurway 1966) is a hint of the existence of partial bivoltinism (Cowan 1991).
Parasitoids of the families Chrysididae, Eulophidae, Ichneumonidae and Tachinidae attack various irnmature stages of B. dyscherus.Species of Chrysididae attack solitary wasps and lay eggs in the stocked cells of his host just before the cells are closed off.• Melittobia (Eulophidae) have been commonly reared from the nests eumenines, vespids and sphecids, bori ng its way into closed mud cells and laying large numbers of eggs on fully-grown prepupae inside.Ichneumonidae are parasitoids that consume their host from the outside, or they may develop within the body of the immature stages of other arthropods (Clausen 1940, in Gauld 1987).Tachinidae are parasites of the larvae of Lepidoptera and Hymenoptera, among others.
They lay their eggs over the body of the host and the larvae perforate the body wall and feed on the body interior (Borror and Delong 1969).

Fig
Fig .l .A -Ventral view of nest with 12 cells in the first layer , iIIustrating sorne of the rneasure rnents taken : 1thickness of thelateral wal ls bet ween the cells; 2 -bet ween the rows of cells and 3 -bet ween the cells and the nest surface; B -nest with a cell ready for prov isioning.The arrow indicates the orifice left by the female; C -Nest with a cell from the first ¡ayer , from one of the series , in construction; D -Nest with a cell from the second ¡ayer , from one of the series , ready for prov isioning; E -Nest with a cell from the third layer , from one of the series , in construction; F -Nest with a cell from the trhir layer , from one of the series , ready fro prov isioning..The Iines correspond to 1.0 cm.

Fig. 2 .
Fig. 2. A -Nest corss-section showing the position of the celllayers; B -Ventral vi ew of a nest with 14 cornleted cells and one ready fro provisioning (abandoned) .In addition , it is possible to distinguish sorne prepupae and remnants ofi :ocoons; C -Dorsal vi ew of nests.Note the balls of mud placed on the cov ering .The lines correspond to 1.0 cm.

Fig . 3 .
Fig .3 .Frequency distribution of the number of cells per nest of Brachymenes dyscherus.

Fig . 4 .
Fig .4. Frequency distribution of the number of pre y per eell of Brachymenes dyscherus.Development period: Irr espective of the collection date of the nests, the duration of the pupalstage varied from 16 to 25 days for males (x= 20.2 ± 4.0 days, n=107) and from 19 to 28 (x= 21.5 ± 3.5 days, n=121) days for females, a statistically significant difference (Mann Whitney, Z=4.05; p<0.05).The total duration of the immature stages was estirnated from the time lapsed between the nest collection date (recently completed nests collected in April/96) and the emergence of the individuals.This time varied depending on sexo For males it was 158 to 309 Fig .6. Number of emerged males and females of Brachymenes dyscherus obtained monthly .

TABLE 1
Dime/!sions (mm) of /les!s ofBrachymenes dyscherus in re/atio/!to the number of layers

TABLE 3
Adults produced, parasitoids and dead imatures, lost samples and cel/s accidentally demaged, and empty cells in each layer in nests ofBrachymenes dyscherusLije table for Brachymenes dyscherus(nests collected in 1995)showing the percentage mortality and its causes at different stages of the lije cycle

TABLE 5 Lije
Endogenous mortality was very high at all stages, accounting for 221 of the 366 deaths, or 60.4%, in 1995 and 98 of the 156 deaths, or 62.8%, in 1996.Death due to parasitic wasps, Melittobia sp.(Eulophidae), was the next most common mortality factor, accounting for 110 deaths, or 30.0%, in 1995 and 47, or j O.l %, in 1996.(Here tables 4 and 5).
table for Brachymenes dyscherus(nests collected in 1996)showing the percentage mortality and its causes at dif f erent stages of the lije cycle