Introduction
The Diphaglossinae include the largest and most robust Colletidae. They are known for being mostly crepuscular to nocturnal bees that are found in habitats ranging from deserts or near deserts to humid tropical rain forests (Rozen 1984, Almeida et al. 2012). Michener (2007) included nine genera in this subfamily distributed in three tribes, which are restricted to the New World. In contrast, Urban et al. (2012) considered this subfamily as a tribe divided in the subtribes Caupolicanina, Diphaglossina, and Dissoglottina, which include eleven genera. The genus Zikanapis is distributed in warm temperate areas of the continent (Compagnucci 2006) and includes five species from Argentina (Urban and Moure 2001, Michener et al. 2003, Compagnucci 2006, Urban et al. 2012). The genus Ptiloglossa includes 30 or more species ranging from Argentina to Texas and Arizona in USA (Michener 2007). The genus Cadeguala contains two species distributed from the Coquimbo region in northern Chile and Bolivia to Valdivia in southern Chile, and Río Negro province, Argentina (Michener 2007, Montalva et al. 2011). Finally, Diphaglossa, with only one species endemic to Chile, occurs from 30° to 50° of latitude in the continent and in the Chiloé island (Michener 2007, Montalva and Ruz 2010, Montalva et al. 2011).
The behavior and nest architecture of all Diphaglossinae is rather homogeneous and contrasts markedly with that of other colletid subfamilies (Janvier 1955, Roberts 1971, Rozen 1984). Typically, nesting behavior consists of the excavation of a vertical main tunnel by a single female and horizontal laterals ending in a large, vertically oriented cell with fluid provisions. Two of the most interesting features of diphaglossine biology are the construction of curved cells and the retention of a cocoon-spinning behavior and larval morphological features related to this behavior (Rozen 1984, Michener 2007). Most studies about the biology and nest architecture of the subfamily come from North American (Linsley 1962, Linsley and Cazier 1970, Roberts 1971, Rozen 1984, Rozen and Rozen 1986) and Central American species (Otis et al. 1983, Roubik and Michener 1984, Wuellner and Jang 1996). The knowledge of South American taxa corresponds to brief mentions by Schrottky (1906, 1907) about the biology of Ptiloglossa ducalis Smith and Ptiloglossa matutina from Brazil and Paraguay and observations of Cadeguala albopilosa and Diphaglossa gayi from Chile carried out by Claude-Joseph (1926) and Janvier (1933, 1955). Later, Rozen (1984) reinterpreted some aspects of the biology of Cadeguala and Diphaglossa species. Other studies on Cadeguala occidentalis were conducted by Torchio and Burwell (1987) and more recently by Montalva et al. (2011). The biology of some species of the genus Caupolicana was studied by Claude-Joseph (1926), Janvier (1933, 1955), Michener (1966), and Genise et al. (1990). The nesting biology of species of Willinkapis, Cadegualina, Mydrosoma and Mydrosomella is unknown.
Fossil bee cells with a curved shape attributed to Diphaglossinae have been recently recorded from the Cenozoic of Patagonia, Argentina (Sarzetti et al. 2010). Herein is described the nest architecture and some data on the nesting biology of five South American species of Diphaglossinae, to enable in the future a more accurate comparisons with the fossil examples and to extract paleoenvironmental inferences from them.
In this contribution, novel nesting and biological observations are provided for three Caupolicanini species: Ptiloglossa matutina (Schrottky, 1904), Ptiloglossa tarsata (Friese, 1900), and Zikanapis tucumana Moure (1945); and two Diphaglossini species: Cadeguala albopilosa (Spinola, 1851) and Diphaglossa gayi Spinola (1851).
Discussion
The five species studied herein share many ecological preferences, behaviors and features of nest architecture with each other, and with other Diphaglossinae, although some significant differences were also found during this study.
The broad ecological preferences of Diphaglossinae differ greatly among species as shown by its extended geographical distribution (Rozen 1984). Herein are provided values of mean annual temperature (MAT) and mean annual precipitation (MAP), along with vegetation types to understand more precisely this environmental diversity. The southern species, Diphaglossa gayi was found in glades of the hygrophilous evergreen forest (MAT = 11° C and MAP = 2500–3000 mm), whereas Cadeguala albopilosa nested in the xeric Austrocedrus forest (MAT = 8° C and MAP around 1200 mm). The northern species nested besides cultivated fields under warmer conditions (with a MAT between 17° C to 20° C at these localities). Ptiloglossa tarsata in an environment that originally corresponded to the more humid (MAP = 700–800 mm) transition between the drier Chaco and the Yungas and Ptiloglossa matutina habits the humid Paranaense Atlantic forest (MAP = 2000–2300 mm), while Zikanapis tucumana in a more xeric environment (MAP = 250 mm) of the Larrea shrubland.
Diphaglossinae were considered traditionally as dim-light bees (Rozen 1984, Michener 2007). However, among the species studied herein, only Zikanapis tucumana and possibly Ptiloglossa matutina showed dim-light foraging. Wcislo and Tierney (2009) distinguished the following three types of dim-light foraging in bees: (1) matinal, if bees are active before sunrise; (2) vespertine, if bees are active in post-sunset twilight; and (3) crepuscular, if bees are active during both, evening and morning twilight. Females of Zikanapis tucumana showed matinal habits, starting the activity even at night (05:00 am), when the flowers of Solanum sp. began to open before twilight, which occurred at 05:20 am. Probably, during those 20 minutes there were some twilight, imperceptible for the human eye, which allows bees to fly, and the exposition of the flowers of Solanum sp. was synchronized with the start of flights. Females of Ptiloglossa matutina were observed flying after 06:00 pm (Julián Baigorria, pers. comm.), which indicates that this species may be vespertine. However, Schrottky (1907) captured Ptiloglossa matutina flying at 04:00 am in Paraguay. The observations presented herein indicate that Ptiloglossa tarsata is diurnal, starting their activity in the morning and continuing with foraging until 02:00 pm. After 02:00 pm the nests remained open with the female inside. A similar daily activity was recorded for Cadeguala albopilosa, which was active from the morning to 07:00 pm. Diphaglossa gayi was observed active after 06:00 pm with full sunlight. It was impossible to determine the previous daily activity because nests were found after 06:00 pm.
The proposed advantages of colleting pollen during twilight include: 1) reduction of competition for resources, 2) reduction of predators 3) reduction of nest parasites (e.g. Bohart and Youssef 1976, Roubik 1989, Wcislo et al. 2004, Kelber et al. 2005). In addition, dim-light bees are usually active at cooler temperatures during matinal or vespertine hours and this may reduce the exposure to unfavorable thermal conditions, mainly in those species that lives in xeric habits, montane regions, or higher latitudes (Hurd and Linsley 1970, Kelber et al. 2005, Wcislo and Tierney 2009). Among the three species described herein without dim-light foraging activity, Diphaglossa gayi and Cadeguala albopilosa inhabits cold-temperate and humid environments, while Ptiloglossa tarsata is found under warm and humid conditions. These species are active during the day when competition for food is higher. Accordingly, these species of Diphaglossinae seem to be not significantly affected by the factors mentioned previously. The species with dim-light foraging behavior, Zikanapis tucumana and Ptiloglossa matutina, inhabits environments as warm as that of Ptiloglossa tarsata suggesting that 1) this factor is not responsible for its daily activity, or 2) that Zikanapis tucumana and Ptiloglossa matutina may be more sensible to temperature than Ptiloglossa tarsata, or 3) that the influence of the xeric environment favors the matinal behavior of Zikanapis tucumana. Alternatively, the advantage of Zikanapis tucumana and Ptiloglossa matutina to nest during dim-light hours may be to reduce competitors, predators, or nest parasites.
Some aspects of the nest architecture, as the curvature of entrance tunnels and cell necks, were proposed as advantages to face floodings. For example, Roberts (1971) proposed for Ptiloglossa guinnae Roberts nests that the lateral tunnel rose just before the cell neck probably to prevent the entrance of rain water into the cell being provisioned by the female. However, it is difficult to corroborate this hypothesis for all the species described herein. In nests of Zikanapis tucumana, Ptiloglossa tarsata and Ptiloglossa matutina the end of the lateral tunnels raised and then curved downwards (Figs 7, 19 and 21). The same can be inferred from the strongly curved neck of Diphaglossa gayi cells. In contrast, the short lateral tunnels of Cadeguala albopilosa nests, curved downwards without raising anteriorly (Fig. 23). Roberts’ hypothesis (1971) could be corroborated only for Diphaglossa gayi, which nests under MAP around 2500 mm and Ptiloglossa matutina, which nests under MAP around 2300 mm, whereas Cadeguala albopilosa, which nests under MAP of 1200 mm lacks the raising of the entrance tunnels. In contrast, Ptiloglossa tarsata and Zikanapis tucumana, which nest under more xeric conditions, 750 mm and 250 mm respectively, show the raised tunnels. Beyond general xeric environmental conditions of the area, Zikanapis tucumana nested in a soil frequently flooded by irrigation suggesting that particular conditions of nesting sites may be also involved as a selective agent for this behavior.
There is a tendency of Diphaglossinae species to nest in aggregations of few to many bees (Rozen 1984). These aggregations can persist for more than one generation (Rozen 1984, Torchio and Burwell 1987, Montalva et al. 2011). Among the five species studied here only Cadeguala albopilosa and, to a lesser extent, Zikanapis tucumana showed aggregations. The nests of Ptiloglossa tarsata, Ptiloglossa matutina, and Diphaglossa gayi were found isolated.
A tumulus, mostly concentric, surrounded the nest entrances in species studied herein that nested in horizontal surfaces as Zikanapis tucumana, Ptiloglossa tarsata, and Cadeguala albopilosa. Zikanapis tucumana was the only species that constructed a consolidate tumulus, also observed in Ptiloglossa arizonensis Timberlake by Rozen (1984). The nest of Ptiloglossa matutina lacked this typical tumulus probably removed by the abundant precipitations. In Diphaglossa gayi that nested in sloping surfaces, the tumulus was eccentric or absent in vertical sections of soil. Similar conditions were observed in some species of Caupolicana and Cadeguala nesting in sloping surfaces (Rozen 1984).
The nest architecture of the species studied here was mostly similar to other Diphaglossinae described in the literature. The main and lateral tunnels were unlined and lack any particular surface texture in all species studied, with the exception of the main tunnel in nests of Ptiloglossa tarsata, which showed transverse scratches, probably produced by the female mandibles (Fig. 12).
Depending on the species studied herein, from the main tunnel arose 1 to 9 horizontal lateral tunnels connected with one cell as in Zikanapis tucumana or two cells as in Ptiloglossa tarsata, Ptiloglossa matutina, Cadeguala albopilosa, and probably Diphaglossa gayi. The presence of two cells connected to the same lateral is a novel feature for Diphaglossinae. In the five species studied here, the lateral tunnels were filled with soil when connected to closed cells. Janvier (1933) mentioned that the lateral tunnels in the nests of Diphaglossa gayi were unfilled. However, the closed cells of both nests of Diphaglossa gayi recorded herein were not connected by an open lateral to the main tunnel.
The cell earthern closure of Diphaglossinae nests appears to be similar in all species, as in many bees, showing internally a spiral design (Rozen 1984, Almeida 2008). This is consistent with the closure found in Ptiloglossa tarsata showing three coarse coils (Fig. 15). Similar closures were previously recorded also in Ptiloglossa arizonensis, Ptiloglossa fulvopilosa Cameron, Caupolicana gaullei Vachal, Caupolicana albiventris Friese, Cadeguala occidentalis (Haliday) and Cadeguala albopilosa (Rozen, 1984). In contrast, Janvier (1933) indicated that Diphaglossa gayi apparently lacks an earthen closure, and the same is true for Ptiloglossa guinnaeRoberts (1971) and Crawfordapis luctuosa (Smith) (Rozen 1984). Roberts (1971) suggested that in Ptiloglossa guinnae Roberts the lack of a closure could have appeared to facilitate the elimination of CO2 produced by fermentation of provisions. Similarly in the cells of Zikanapis tucumana, Ptiloglossa matutina, Diphaglossa gayi and Cadeguala albopilosa it was impossible to detect an earthern closure. Other character related to the closure was observed in Ptiloglossa tarsata and Ptiloglossa matutina, whose cells show a wad of cotton-like material of unknown origin (Fig. 28). The same structure was observed by Rozen (1984) in cells of Ptiloglossa arizonensis Timberlake. It seems that the presence of this material may be exclusive of Ptiloglossa species.
Rozen (1984) indicated that the most outstanding character of all nests of Diphaglossinae is the shape of their cells. These cells are unique among bees in having a curved neck. The shape of diphaglossine brood cells was discussed by Rozen (1984), who described two types of curved cells. Raised tunnels and more curved cells (90° or more degrees of curvature) were found in Ptiloglossa tarsata, Ptiloglossa matutina, Zikanapis tucumana, and Diphaglossa gayi, as previously described by Rozen (1984) for other species of Ptiloglossa and species of Crawfordapis. In contrast, in Cadeguala albopilosa (Fig. 23) the lateral tunnel is not raised and accordingly the curvature of cell neck is lesser than 90°, as in species of Cadeguala and Caupolicana (Rozen 1984). CT images were useful to confirm the curvature of cell necks of Cadeguala albopilosa observed in the field (Figs 35–37).
The Diphaglossinae are the only group among the Colletidae whose larvae spin cocoons (Rozen 1984, Michener 2007). Rozen (1984) described in great detail the cocoon structure of several species of the genus Ptiloglossa and Crawfordapis luctuosa (Smith), and redescribed that of Cadeguala albopilosa, previously described by Claude-Joseph (1926). Among the five species studied here, only the cocoons of Ptiloglossa tarsata and Cadeguala albopilosa were found. The cocoons of both species were similar to those described by Rozen (1984). Their walls were made of a brown, thin, translucent, and slightly coriaceous material. Cells with cocoons are closed by an operculum made of a disk of silk threads, whose fabric is different among species (Rozen 1984). Below the operculum it was described a structure named filter, composed of a net of silk threads that probably enables gas exchange (Roberts 1971, Rozen 1984, Almeida 2008). Immediately beneath the filter it was described another disk, similar in composition and structure to the operculum, but dome-shaped, which is called the ceiling (Rozen 1984). These structures were recognized only in cocoons of Cadeguala albopilosa. The observations in one cocoon of Ptiloglossa tarsata revealed the presence of one operculum and below it, two other circular disks of similar composition and fabric, which probably had the same function of the filter and the ceiling, but different microstructure.