Teratognatha, MODESTA OGLOBLIN

Rozen, Jerome G., 2011, Immatures of Exomalopsine Bees with Notes on Nesting Biology and a Tribal Key to Mature Larvae of Noncorbiculate, Nonparasitic Apinae (Hymenoptera: Apidae), American Museum Novitates 2011 (3726), pp. 1-52 : 36-42

TERATOGNATHA MODESTA OGLOBLIN

Table 2

I studied nests of this species at 6 km SW of Pichanal, Salta Province, Argentina, on November 11, 1993, along the border of a large field previously cultivated but in which numerous plants, especially Sphaeralcea and Heliotropium (the food source of Teratognatha modesta ), now grew. Although these plants were widely dispersed throughout this field (perhaps 1 km long and nearly 0.5 km wide), I observed nests only along one edge of the field. Time limitations prevented a broad survey of the field where other nesting aggregations may have existed, though clearly nests were not ubiquitous. The absence of nests from previous generations indicated that this nesting area had not been used by this species previously.

Nest entrances where first identified in barren stretches between clumps of weedy vegetation when females (both with and without pollen) were seen flying close to the ground, stopping, and entering cracks in the surface soil that existed throughout the nesting area. These cracks, caused by drying soil and mostly filled with loose soil, penetrated deeply into the substrate (perhaps as much as 10 cm). Soil on analysis consisted of 22% sand, 40% silt, 38% clay and was classified as clay loam. It was quite homogeneous, containing few stones. Occasionally females entered single surface holes, but on excavation these holes merely led to cracks below surface crust. Consequently, main burrows could not be followed from the surface to cell level, and cells were excavated only by blindly digging areas beneath holes and around cracks into which females had disappeared. Because main nest burrows could not be traced nor their entrances identified, numbers of females to a nest is unknown. However, so many females entered the same crack it seems likely that nests were composite.

NEST ARCHITECTURE: Tunnels of nests, about 2.0 mm in diameter, could be recognized only in the vicinity of cells where tunnels were not soil filled. Their surface was not waterproof and was finally roughened, almost certainly a result of being tamped by the female’s pygidial plate.

The precise number of nests encountered is unknown because we were unable to determine nest boundaries. However, at least four separate clusterings of cells and cell series were found. Centers of these aggregations were at depths of 16, 21, 26, and 28 cm. Most cells were in a straight linear series of three, and certainly in some nests at least two such series were evident. In one case a series of two was found, and some apparent series of three might actually have consisted of two cells in front of which was a celllike chamber without a closure (suggesting an interruption to nest construction). The length of the lateral between the leading cell in a series and the main burrow was 4.0 cm in two cases and 7.0 cm in another. The short laterals were completely soil filled in the first two cases, but a 3 cm gap remained open immediately in front of the cell in the 7.0 cm lateral.

Many cells (and cell series) (dimension in 2) were oriented 45° from horizontal (rear end lower than the front), but more observations on cell orientation are needed. Cells appeared symmetrical around their long axes even after repeated careful examination. Cell walls could not be differentiated from the substrate by color or hardness when dry. They were lined with a transparent, slightly reflective material that was mostly waterproof toward the rear of the cell but permitted water to be absorbed more or less rapidly near the cell closure. The inner surface was darker than the substrate when first excavated but upon drying matched the substrate. The inner surface of the cell was unusually smooth toward the rear, moderately shiny, but in most cells was slightly more irregular toward the cell closure. Septa between cells were very thin, less than 0.5 mm in thickness at the periphery. The front surface of the septum formed the smooth, lined, concave rear of the preceding cell. The opposite surface of the septum was the essentially flat inner face of the cell closure in the form of a spiral of approximately 3–4 four coils to the radius.

PROVISIONING AND DEVELOPMENT: Female Teratognatha modesta transport pollen as moist masses on their hind tibiae and basitarsi. In cells, each female forms the mealy-moist, yellowish provisions into a loaf (figs. 53, 54) that rests its entire length on the cell floor without touching the rear of the cell. These masses varied somewhat in shape and in general had the dimensions given in table 2. They were elongate, rounded at the rear, and had a curved upper surface. As viewed from the side (fig. 53), their highest point was approximately 2 ⁄ 3 toward the rear. From the highest point the top surface curved gently to the nearly flat front. At its base on the cell floor, the front end bore a more or less distinct transverse ridge, questionably corresponding to the “foot” on provisions of Anthophorula chionura ( Rozen and MacNeill, 1957: fig. 2). A reexamination of a preserved food loaf of Exomalopsis bruesi Cockerell (figs. 14, 15) during the current study showed that it too bore this transverse low ridge where the front surface met the cell wall.

The curved white egg rested by its front and rear ends on the gently curved top of the provisions in the sagittal plane of the cell while its midsection curved upward, not touching the surface. The chorion was smooth, shiny, transparent, and lacked reticulations or other sculpturing when viewed stereoscopically. Egg dimensions are given in table 1. Larvae crawled as they fed, leaving a shallow groove on the surface of the provision. The elongate form of the intermediate stage larva appeared characteristic of exomalopsines; the protruding venter of abdominal segment 9 apparently assisted in crawling. The dorsal attachment of abdominal segment 10 to 9 permits the larva to appress its feces against the cell wall, a process that begins before the provisions are completely consumed. Feces are applied to the entire cell wall including the inner surface of the closure.

No cocoons were spun by these larvae, all of which pupated or would have done so if permitted to develop. Interesting is the fact that the salivary lips of the last instar are clearly projecting, though perhaps not as much as those of some other exomalopsines, although the width of the opening is narrow. It is uncertain, therefore, whether a late summer generation might spin cocoons, as suggested by Rozen (1984) and Rozen and Snelling (1986) for certain other exomalopsines. No cocoons from a previous generation occurred at the site, and the absence of empty vacated cells with or without cocoons indicated that the site had not been used before. This species has more than one generation per year; all larvae remained active after feeding and developed pupal tissue internally.

ADULT ACTIVITY: Mating was not observed at the nesting site; hence, it probably occurs at the flowers. Females were most active in the morning, 10 A.M. to 12 noon. Their host plant is Heliotropium , and there can be little doubt that the peculiar modifications of the female’s distal maxillary and labial palpal segments are adaptations for extracting pollen from the flowers of this plant, as was first suggested by Michener and Moure (1957) although the host plant was not known at that time. Other, distantly related bees, such as Calliopsis hesperia (Swenk and Cockerell) , several unnamed South American Calliopsis , and a species each of Callonychium s.l. and Paranychium (all Andrenidae : Panurginae ), harvest pollen from this plant genus and have unusual mouthpart modifications, presumably adaptive for that purpose (New Information).

PARASITISM: No parasites were associated with this site.

DISCUSSION Of NESTING BIOLOGY

This section is intended to provide an overview of the nesting biology of the Exomalopsini . The reader is also directed to an earlier attempt ( Rozen, 1984), which discusses additional information on cocoon spinning, defecation, and voltinism that is not repeated here.

Based on the observations above and literature accounts (see references below), all species are ground nesting, and nest entrances are always without turrets. Many species (hereinafter termed horizontal nesters and identified in table 2 by “H”) tunnel into more or less horizontal surfaces that are exposed to open sunlight, but some species (including several of the subgenus Anthophorisca , at least one species of Chilimalopsis , and Teratognatha modesta ) (termed fissure nesters and identified in table 2 by “F”) utilize cracks in the ground resulting from soil drying and shrinking. Raw (1977) reported that Exomalopsis (E.) pulchella Cresson (as Exo. globosa (F.)) is the only known species to both nest in cracks and on horizontal surfaces. The fissures allow females easy access to moist soil that can be excavated with mandibles, thereby conserving energy and avoiding mandibular wear. Nests of horizontal nesters normally are occupied by more the one female, as was first recognized by Hicks (1936) for Anthophorula (Anthophorula) torticornis (Cockerell) , when he observed four females entering a single nest. So far as currently known, the numbers of females per nest are comparatively few, such as 5–14 for Anthophorula sidae or 12–35 for Exomalopsis solidaginis . However, Hurd and Linsley (1975) claimed that nests of Exo. solani Cockerell may contain colonies of several hundred individuals, and Zucchi (1973) reported 884 females in one nest of Exo. auropilosa Spinola.

Nests of fissure nesters have been impossible to monitor because the entrance to the nest is deep in the crack. By excavating one side of a crack, one immediately upsets the comings and goings of inhabitants. No one has yet removed one side of a crack at night when nest inhabitants are asleep and then collected adults the next day as they emerged from nests on the preserved side. Furthermore, there is a possibility that a crack serves as a counterpart of a main tunnel of horizontal nesters. If that is the case, then nest tunnels opening along the face of a fissure may be the counterparts of the more or less horizontal branches attached to the main tunnel of the horizontal nester. In several cases where nests consisting of only two or three cells ( Anthophorula nitens (Cockerell) and Chilimalopsis parvula ) were found near the fissure, I have proposed that they might have been constructed by a single female, which might be interpreted to support the assumption that the small group of interconnected cells are indeed the counterpart of horizontal side tunnels of a horizontal nest. However, the situation is unresolved because at this time we do not understand the interactions of adults inside nests with multiple females. For example, does a single female in such a nest work a single branch by herself, or does she place her cells wherever she finds herself? Or does she work coopera- tively with other females in building and provision cells as has been suggested by Michener (1966) and others? Such matters are not insoluble and need to be addressed.

Depths of exomalopsine nests as measured by the deepest cells uncovered tend to fall in the first half meter below the surface. The shallowest nests are those of Eremapis parvula , in which cells ranged 5–7 cm deep. By far the deepest exomalopsine nests recorded are those of Exomalopsis auropilosa , two of which were 4.6 and 5.3 m deep ( Zucchi, 1973).

Main burrows tend not to be soil filled, as are side tunnels where they branch from the main tunnel though they seem filled closer to cells. Laterals immediately in front of closed cells are usually filled, except the situation is uncertain for Chilimalopsis parvula , q.v.

Table 2 lists dimensions of nest components and of provisions of those Exomalopsini that have been studied to date. The dimensions in table 2 seem unremarkable in light of the small sizes of many of the bees. Of some interest is that burrow diameters of many species tend to be slightly wider than the burrow opening at the surface (data not supplied); intuitively one would think that nest entrances would tend to wear and widen with passages of adults.

Column titled “Cell tilt” in table 2 gives the approximate range of the tilt of the long axis of the cells from a horizontal position. The anterior end of the cell is always equal to or above the posterior end. Hence, if the long axis is horizontal it would be 0°, and if perpendicular to the horizontal position, it is, of course, at 90°. Because these ranges are based on only a few measurements, they should be considered rough estimates. With most taxa, cell orientation is probably rather variable within a species.

As shown in table 2, many exomalopsine nests have cells in linear series consisting of two cells end-to-end mixed with single cells. However, both in Chilimalopsis parvula and Teratognatha modesta series consist of three cells, and in Anthophorula chionura series can consist of as many as five cells. However, singles and series of two are also encountered in all of them. Only single cells have been reported for A. crenulata (Timberlake) , A. sidae , A. uncicornis , and for all species of Exomalopsis .

A feature that seemed distinctive at first was the presence of a cell wall that was more consolidated than the substrate such as those found with Anthophorula compactula , A. completa , and Exomalopsis solidaginis . On further examination, I realized that the distinction of this feature is obvious only when cells are built into a soft substrate. Those found in consolidated substrates did not have walls that could be differentiated from the matrix. None of the material examined by me appeared to have a wall composed of any material other than that of the surrounding substrate. However, with all species the cell wall has been modified in that it has been manipulated by the female and except for Eremapis parvula (and possibly A. crenulata ) been made hydrophobic, presumably being impregnated with a glandular secretion that apparently also helps consolidate the inner cell surface. The clear, reflective surface of the wall is evidence of this presumed secretion. The cell wall of Er. parvula is dull and water absorptive. Although the cell wall of A. crenulata was reported “polished and shiny,” water droplets were “quickly absorbed” ( Parker, 1984). I have noted that different parts of a cell wall have differential absorption rates as, for example, in A. uncicornis (above), and wonder whether that might explain Parker’s observation.

Cells are normally constructed with one side more elongate and curved than the other side, so that for an inclined cell the upper surface is longer and more curved than the bottom side, the latter straighter and shorter than the upper. This is usually obvious, as illustrated in figure 56. In cells of Eremapis parvula and Chilimalopsis parvula , this asymmetry is difficult to detect because of their small size and because the plane of their closure is closer to being at a right angle to the long axis of the cell in side view. 8 If they are opened in strict lateral view, it becomes apparent (fig. 56). However, I have been unable to detect asymmetry in cells of Teratognatha modesta , and Zucchi (1973: figs. 5b, 5C) illustrated the cells of Exomalopsis auropilosa as symmetrical as well as perpendicular.

With the exception of Eremapis parvula , cell surfaces of exomalopsine bees are lined with a clear material that affords a hydrophobic barrier, presumably safeguarding the occupant from desiccation. However, in the Exomalopsini there is a strong suggestion that this barrier also serves to protect the provisions from coming in contact with the cell wall. For example, in Anthophorula uncicornis this lining is most developed at the very rear of the cell where the provisions are attached. As noted in the case of Chilimalopsis parvula , older larvae cradle food masses so they do not come in contact with cell surfaces, as also shown to be the case for Exomalopsis solidaginis ( Rozen 1984: figs. 23, 24). In Eremapis parvula the peculiar pattern of fecal deposition can be interpreted to be an outgrowth of holding the provision away from cell wall. Although one assumes that this is to prevent wicking of moisture from the provisions by the cell wall, one must also consider safeguarding the provisions from microbial agents.

In all cases cell closures are a spiral on the inside, usually of 3–4 coils to the radius on the inside surface, but for larger species ( Exomalopsis bruesi and Exo. solani ) the number increases to 4–6 coils ( Norden et al., 1994: fig. 2A). In all known cases, the outside surface is concave, smooth, and variably water retardant to a droplet, but the form of the inside surface ranges from flat (or even weakly convex) to deeply concave, with those of Anthophorula sidae , A. uncicornis , and Eremapis parvula being deeply concave and those of A. completa , A. crenulata , A. consobrina (Timberlake) , Chilimalopsis parvula , and Teratognatha modesta being remarkably flat. Others are weakly concave or indeterminate at this time.

Larval provisions of all species are formed into a loaflike mass consisting of mealy-moist pollen with the moisture presumably derived from nectar. In cells that are inclined the loaf rests on the lower, somewhat flatter side of the brood chamber, as in Anthophorula sidae ( Rozen, 1984: fig. 28) or against the rear of the cell, as in A. nitens ( Rozen and Snelling, 1986: fig. 5), Exomalopsis solidaginis ( Rozen, 1984: fig: 20), and Exo. solani ( Norden et al., 1994) . In A. uncicornis the loaf is actually attached to the rear of the cell, and in the perpendicular cell of Exomalopsis auropilosa the loaf is in contact only with the rear of the cell.

In table 2 is the column “Foot?” meaning “is there a pointed projection on the lower front edge of the provision mass?” as original intended by Rozen and MacNeill (1957). Although Zucchi (1973) stated the Exomalopsis auropilosa had such a structure, he identified it as the

8 As noted by Rozen (1984: fig. 17) with respect to Exomalopsis solidaginis the plane of the closure (line e–f) is quite different from the line c–d that is perpendicular to the long axis of the cell, thus accounting for the top surface of the cell being conspicuously longer than the lower surface.

structure at the rear of the provisions where the food mass contacts the rear of the cell rather than at the front end (his fig. 5D). Although the shape of the mass otherwise is convincingly that of an exomalopsine, the structure at the rear of the mass seems to be a novelty because of its position, and accordingly Exo. auropilosa is scored “No.” Although I have scored Eremapis parvula , Teratognatha modesta , and Exomalopsis bruesi as “No,” some of the preserved specimens of each have a suggestion of this projection. Raw stated that provisions of neither Exo. puchella nor Exo. similis bore a “foot.” To be noted: the food mass of Anthophorula uncicornis (fig. 47) is affixed to the cell’s rear wall and still has a foot that never contacts any part of the wall, as also is the case for Exo. solidaginis ( Rozen, 1984: fig. 20).

Eggs of all species that I have studied are deposited on the top surface of the provisions somewhat toward the front end (as in figs. 46, 48, 50, 52, 54) in the sagittal plane of the cell. In all cases the egg contacts the surface of the provisions only with its anterior and posterior ends. Young and intermediate stage larvae slowly crawl over the provisions as they feed, thereby channeling the surface.

Most Exomalopsini spin cocoons including Anthophorula compactula , A. completa , A. crenulata , A. chionura , A. consobrina , A. nitens , A. sidae , Eremapis parvula , Chilimalopsis parvula , Exomalopsis puchella , and Exo. similis . However, of these, A. chionura , A. nitens , and A. sidae have an early generation in the year that transforms to the pupal stage without spinning a cocoon; these species do not spin cocoons until they reach the overwintering generation. On the other hand, cocoons of the following species are unknown: A. uncicornis , Teratognatha modesta , Exo. auropilosa , Exo. bruesi , Exo. solani , and Exo. solidaginis .

It is assumed that cocoon spinning is a plesiomorphic condition among bees, if not among all Hymenoptera . The apparent multiple loss of the ability (or need for) cocoon spinning among various taxa within the small tribe Exomalopsini is worthy of consideration. There is also a strong suggestion in the case of such multivoltine species as Anthophorula chionura and A. consobrina , cocoon spinning can be suspended in summer generations and be initiated again for the overwintering generation. However, the anatomy of the larva of the summer generation is still that of the overwintering one, i.e., salivary lips continue to be projecting. In those exomalopsine taxa that never spin cocoons, the salivary lips are nonprojecting (or project only slightly). Oddly, while strongly projecting lips are no longer present, the salivary opening in all cases is transverse, unlike in larvae of any other non–cocoon spinning bee taxon where the opening tends to be a simple round opening to the salivary duct. The meaning of this is not clear. It could be interpreted that the original assumption that cocoon spinning once lost might indeed evolve again. Or perhaps we do not understand the phylogenetic relationships among the taxa. In any event, it points to the fact that we still have an ongoing puzzle worthy of exploring.

Information presented here does not refer to social interactions of adults bees in communal nests, as has been mentioned by others ( Michener, 1966; Zucchi, 1973; Raw, 1977; Parker, 1984). These are interesting matters that deserve further investigations. This study also does not consider how colonies occupying single nests are established nor does it reveal information about the duration of such assemblages. There is more work to be done.