Collection of Samples

Two nests of P. paulista were collected in the municipality of Rio Claro (22°23′44.09″S, 47°32′39.98″W), São Paulo state, Brazil. All collected samples were fixed and conserved in 70% ethanol. Voucher specimens of eggs, larvae, and pupae were deposited in the “Adolph Hempel” entomological collection of Centro de Pesquisa e Desenvolvimento de Sanidade Vegetal of Instituto Biológico, São Paulo, Brazil.

Fig. 2.

Immature stages of Polybia paulista: (A) ventral view of a first instar larva; (B) side view of a fifth instar larva; (C) ventral view of a fifth instar larva; (D) side view of prepupa; (E) ventral view of prepupa, showing in the inset the last abdominal somite; (F) side view of pupa. Sizes of scale bars: (A) 0.185 mm; (B) 0.575 mm; (C) 0.650 mm; (D) 0.925 mm; (E) 0.925 mm; (F) 1.200 mm.

Immature Descriptions

The number of larval instars of P. paulista was determined based on maximum head widths following the methods described in Parra & Haddad (1989) based on a sample of 1,378 larvae. Other measures presented were taken from 45 eggs, 707 larvae, and 75 pupae. Terminology used in the description was based on Kojima (1998). Details on sample preparation and the setting of microscopes were described in Solis et al. (2010).

The following abbreviations are employed: (BL) body length; (BW) medial width of body; (d) diam; (Ds) diam of spiracles; (HG) height of galea; (HPa) height of labial palp; (HPm) height of maxillary palp; (HW) head width; (1) length; (LA) length of prementum; (LBH) length of simple setae on body; (LMH-S) length of simple setae on mouthparts; (LMH-B) length of bifid setae on mouthparts; (LMH-R) length of 3-branched setae on mouthparts; (LM) length of mandible; (Lm) length of maxilla; (LR) length of labrum; (n) number of observations; (R) minimum and maximum values of product of the division of the greatest spiracle diam by the diam of other spiracles; (WA) medial width of prementum; (WM) width of mandible base; (WR) medial width of labrum.

Statistical Analysis

All measures are presented with minimum and maximum values. In Table 1, measures are given in millimeter. When compared, measures from different larval instars (see Table 1) were subjected to analysis of variance (One-way ANOVA) followed by Tukey's test.

Determination of the Number of Larval Instars

The obtained frequency distribution of head capsule widths resulted in a multimodal distribution with 5 distinct peaks, suggesting the existence of 5 instars. The first peak includes width values for first-instar larvae found inside the egg chorion and the fifth peak includes width values of prepupae (Fig. 1). Intervals of head widths for each larval instar head width: first instar (0.32–0.49 mm; n = 90), second instar (0.51–0.73 mm; n = 156), third instar (0.78–0.93 mm; n = 181), fourth instar (1.05–1.30 mm; n = 305), and fifth instar (1.40–1.75 mm; n = 642). The obtained number of larval instars yielded a good fit with Dyar's rule (R² = 0.99).

Fig. 3.

Details of the body surface of Polybia paulista larvae: (A) thorax region of a fifth instar; (B) abdominal region of a fifth instar; (C) abdominal region of a third instar; (D) abdominal region of a first instar, displaying a closed spiracle; (E) abdominal spiracle of a fifth instar. Sizes of scale bars: (A) 0.016 mm; (B) 0.028 mm; (C) 0.020 mm; (D) 0.008 mm; (E) 0.010 mm.

The mean growth rate between the larval instars was 1.39, while between the first and the second instars it was 1.34, between the second and third instars it was 1.44, between the third and fourth instars it was 1.40, and between the fourth and fifth instars it was 1.36.

Morphological Description of the Immature Forms

Egg. Elliptical in shape and white, found attached to the nest cell wall. Length and medial width (n = 45) respectively 0.93–1.23 mm and 0.31–0.49 mm, the ratio being 2.73.

Larvae. Obtained bodily measures from all instars are presented in Table 1. Body: 13-segmented, including 3 thoracic and 10 abdominal segments; segmentation upon first- and secondinstar inconspicuous. Body shape short, robust, and slightly curved, with greatest diam upon 4 first abdominal segments, and the following segments gradually reducing in diam (Figs. 2A, 2B, 2C). Anus ventral. Fifth-instar present 2 different body shapes: the one described in the last sentence and an altered shape upon later develop- ment (i.e. prepupa; Figs. 2D, 2E). The prepupa is more elongate, and upon the last abdominal somite presents a round protuberance with 3 lobes (Fig. 2E); during this stage, the nest cell is closed with silk produced by the larva. Body of larvae always grayish white, while prepupae are strongly white. The body tegument of all larval instars is covered with abundant spines and papillae, which increase in density on every subsequent instar (Figs. 3A, 3B, 3C), excepting on the last abdominal segments, where actually they reduce in density. Body setae are few and found only in fifth-instar (Table 2). Vestigial legs (3 pairs) are markedly evident on the ventral thoracic region of fifth-instar (Fig. 2C). There are 10 pairs of spiracles (2 thoracic and 8 abdominal, see Fig. 3E) distributed along a line at mid-width of body as seen from side view, placed near the intersection of segments. Body spiracles are closed in first-instar (Fig. 3D). First pair of spiracles always larger than the others. Cranium: Subelliptical in shape (Fig. 4A), horizontal diam larger, and sitting terminally on the anterior body extremity. Color tends to change from white to a yellowish white tonality over development. Parietal bands evident (Figs. 4A, 4B, 4C). Absence of setae, spines or papillae on the cranium. Antennal orbits circular and concave (fifth-instar, d: 0.12–0.15 mm, n = 10) with 3 basiconic sensilla (Fig. 4D), which may or may not be arranged in line. In fifth-instar, the distance from the lower antennal orbit border to the anterior tentorial pit ranged 0.11–0.13 mm (n = 4), while in prepupae it was 0.22 mm (n = 1). Clypeus poorly delimited. Mouthparts (Fig. 4E): Labrum in anterior view trapezoid-shaped (Fig. 5A), with anterior surface smooth. In ventral view, the labrum presented 2 different shapes depending on larval instar: first- to third-instar and prepupae presented elliptical profiles (Fig. 4E), while fourth- and fifth-instar had strongly emarginate labrum profile (Fig. 5B). Ventral surface of labrum rough, with papillae (only in fifth-instar), 18–22 basiconic sensilla on conical protuberances (all instars) and 7–15 setae (all except first-instar). The palate of fifth-instar is slightly rough, with 6 basiconic sensilla disposed in 2 groups of 3 on the sides. Mandible always triangular; in first and second-instar (Fig. 5C), mandibles are short and robust, wider at base than long, with apex rounded, and there is no lateral tooth. In thirdinstar mandibles are about as wide as long. In fourth and fifth-instar (Fig. 5D), mandibles are more elongate, with base width being less that total length, and there is a lateral tooth near its pointed apex. Anterior face of mandibles presenting some papillae. Maxilla elongate, with setae on the anterior surface (absent in first-instar; at the number of 5–7 setae in second- and third-instar; 11–15 in fourth- and fifth-instar) which are concentrated on the distal half (near the apex). The maxilla of fifth-instar presents spines on the dorsal anterior region (Fig. 5E) covering the first proximal half (near its base). Galea cone-shaped with surface rough, being bifid at apex with each extremity bearing a basiconic sensillum (Fig. 5E). Maxillary palps cone-shaped, with surface smooth and bearing 2 basiconic sensilla and one encapsulated sensillum on top (Fig. 5F). In first- and second-instar, the galea and maxillary palps are simple aggregates of sensilla, with the maxillary palp of first-instar having only 2 basiconic sensilla. Prementum subcircular in shape (Fig. 5G), bearing papillae on the upper half of the anterior surface, above and below the opening of spinneret; on the lower half of prementum, surface is smooth. Presence of setae on the anterior surface: 13–14 setae in third-instar; 19–22 setae in fourth-instar; 24–33 setae in fifth-instar. In first- and second-instar, instead of setae, there are isolated spines on the prementum; first-instar has no papillae. Labial palps cone-like, with surface smooth and topped with 2 basiconic sensilla and one encapsulated sensillum (Fig. 5H). However, the number and type of sensilla varies: 3 basiconic sensilla or 3 basiconic sensilla with one encapsulated sensillum in second-instar; one basiconic sensillum and one encapsulated sensillum in fifth-instar. In first- and second-instar the labial palp is a group of sensilla. First-instar has no visible opening of spinneret, which are quite evident in fifth-instar (1: 0.200–0.236 mm, n = 8), in which they are fringed with an arrangement of fused spines (Fig. 5G). The postmentum of first to third-instar is smooth, while fourth-instar (1: 0.60–0.62 mm, n = 3) have a few spines, and fifth-instar (l: 0.92–1.05 mm, n = 3) have it densely covered with spines, excepting the proximal ventral region on the prementum (Fig. 5G).

Fig. 4.

Fifth instar larva of Polybia paulista, showing: (A) full frontal view of head; (B) dorsal view of head; (C) side view of head; (D) antenna; (E) mouthparts in prepupa. Abbreviations: prementum (la), labrum (lb), mandible (M), maxilla (m), ecdysial sulcus (white arrowhead), parietal band (white arrow). Sizes of scale bars: (A) 0.240 mm; (B) 0.178 mm; (C) 0.143 mm; (D) 0.028 mm; (E) 0.227 mm.

Fig. 5.

Larval mouthparts in Polybia paulista, showing: (A) anterior region of labrum of fifth instar; (B) ventral region of labrum of fifth instar; (C) mandibles of second instar; (D) mandible of fifth instar; (E) maxilla of fifth instar; (F) maxillary palp of fifth instar; (G) prementum of prepupa; (H) labial palp of fifth instar. Abbreviations: opening of spinneret (f), galea (g), maxillary palp (pm), labial palp (pl), sensilla (white arrowheads). Sizes of scale bars: (A) 0.067 mm; (B) 0.060 mm; (C) 0.022 mm; (D) 0.058 mm; (E) 0.055 mm; (F) 0.024 mm; (G) 0.068 mm; (H) 0.028 mm.

Pupa (Fig. 2F). Exarate, white but becoming darker over development, until transforming into black imagoes. Body length and head width of white pupae (n = 75) were respectively: 8.13–11.13 mm and 2.25–2.50 mm.

Determination of the Number of Larval Instars

In the subfamilies Polistinae and Vespinae, the registered number of larval instars is 5, while in Stenogastrinae the number varies between 3 and 5 instars (Kojima 1998). Carvalho & Silva (1975) and Machado (1983) found 5 larval instars in P. paulista, which was confirmed by our present observations. Interestingly, from comparing the mean head widths obtained in this study with the results of Carvalho & Silva (1975), we noted that the size intervals between instars are not the same. Size values obtained by Carvalho & Silva (1975) were larger: first instar (0.360–0.612 mm), second instar (0.648–0.972 mm), third instar (1.008–1.332 mm), fourth instar (1.368–1.836 mm), and fifth instar (1.872–2.340 mm). Consequently, the values of head widths of pupae in their study were also larger than our records. Parra & Haddad (1989) mentioned that the number of larval instars in a given species can vary, being dependent on biotic and abiotic factors such as hereditary traits, rearing methods, ambient temperature, and nutrition. It is very likely that such uncontrolled factors are responsible for the size differences observed, although the estimated final number of larval instars was not affected.

Fig. 6.

Comparison of size ranges of antennal orbit size between Polybia paulista and among different social wasp genera analyzed by Kojima (1998). The ratios between antennal size and head widths were calculated based on images given in Dias Filho (1975) and Kojima (1998). As in Kojima (1998) did not declare any absolute or relative values for determining the character state in character 7, we decided employing the relative ratio to head width. Abbreviations of antennal size as defined by Kojima (1998): (S) for small, and (L) for large; the value given for P. paulista is an average figure, indicated with an arrowhead on the value column of Polybia.

Morphological Description of the Immature Forms

Kojima (1998) based on 28 genera of Polistinae and Vespinae proposed 43 morphological characters of mature larvae for use in phylogenetic analyses. The morphological characters described by Kojima (1998) for Polybia were usually confirmed by our observations with P. paulista. We found however some difficulties when applying a few of these suggested characters with P. paulista. Regarding the size of antenna (character 7), Kojima (1998) proposed 2 character states: small and large, with Polybia being presenting large antennae. The measures obtained in the present study with P. paulista fit in the interval estimated for the genus [0.107–0.208 mm, n = 10; obtained from images presented in Kojima (1998)]. However from analyzing the results with other genera (Fig. 6), we can see that some (e.g. Mischocyttarus and Ropalidia) received the character state of small or large antennae while presenting a broad variation between minimum and maximum size values, suggesting that both character states are present. In one instance, Epipona and Synoeca, 2 genera with the same proportional antennal size received different character states. We think such incongruent results with character 7 could be circumvented by adopting a clear-cut proportion to define relative size of antenna, as was done by Kojima (1998) with character 35, relative size of first spiracle. Based on Figure 6 and analysis of Kojima (1998), we would tentatively suggest that the value of 0.067 [ratio between of antennal orbit /head width; this limit value was set from analyzing 27 species illustrations from Dias Filho (1975) and 116 species from Kojima (1998)] be adopted as the borderline to separate relative antennal size, with the result that some genera (e.g. Polybia) would present both character states. Regarding the size of the postmentum (character 33), 3 character states were proposed by Kojima (1998): small, large, and large ventrally emarginate; in most genera the postmentum was described as small, including Polybia. The measures obtained in the present study with P. paulista fit in the interval estimated for the genus [0.96–1.48 mm, n = 4; values estimated from the illustrations in Kojima (1998)]. As with character 7, the value of 0.69 [ratio between size of postmentum and head; estimated based on 20 species illustrations analyzed in Kojima (1998)] could be used as a limit value between the character states small and large (Fig. 7). In Polybia both character states can be found. Further evaluations with other species might validate the proposed limit values, or even invalidate the use of these 2 characters because of substantial overlap between species.

Fig. 7.

Comparison size ranges of postmentum size between Polybia paulista and among different social wasp genera analyzed by Kojima (1998). The ratios between postmentum size and head widths were calculated based on images given in Kojima (1998). As in Kojima (1998) did not declare any absolute or relative values for determining the character state in character 33, we decided employing the relative ratio to head width. Abbreviations of postmentum size as defined by Kojima (1998): (S) for small, and (L) for large; the value given for P. paulista is an average figure, indicated with an arrowhead beside the Polybia column.

Carpenter et al. (2000) employed 21 larval characters obtained from Kojima (1998) in their cladistics analysis in differentiating between subgenera of Polybia. When analyzing their attributed states to the subgenus Myrapetra, within which P. paulista is included, we confirmed all characters. However, in cases in which more than one state is found in Myrapetra, only one character state was observed in P. paulista, namely: hypostomal ridge nearly straight (character 25), labrum almost as wide as clypeus (character 29), absence of spicules on palate (character 31) and mandibular corium (character 33), apex of maxillary palp flat (character 38), prementum subcircular (character 39), presence of spicules on the side and ventral regions of the postmentum (character 41).

From comparing our present finds with a previous description by Dias Filho (1975), we conclude that the author analyzed fifth instar larvae and we confirm all of his observations regarding: the shape of body, cranium, and mouthparts; and measures of body length and head width. However, few details on these structures were not provided by the author, such as distribution of spinules and setae, and number of sensilla on antennae and palps.

Moreover, we verified most modifications over development in larvae of P. paulista were an increment in size, and with the appearance and development of certain structures upon every subsequent instar, including: increase in the density of body spines and papillae; presence of body setae on fifth-instar larvae; the opening of body spiracles in second instar larvae; 2 types of body shape in fifth-instar larvae; the presence of a lateral tooth in the mandibles of fourth- and fifthinstar larvae; presence of spines on the maxilla of fifth-instar larvae; alterations in the shape of galea and palps upon third-instar larvae from a cluster of sensilla to a conical elevation; and the appearance of spines on the postmentum upon fourth-instar larvae. In other Vespidae species analyzed, besides the increment in general structural size there are other recorded morphological alterations over larval development, such as: appearance of ventral lobes on the first abdominal somite on third-instar larvae of Mischocyttarus cassununga (Ihering) (Giannotti & Silva 1993), and the development of a subapical tooth on second-instar larvae of Polistes lanio (Fabricius) (Giannotti 1995). By employing scanning electron microscopy and by minutely inspecting all development stages, further structural alterations are expected to be recorded.

Finally, the present study confirmed the existence of 5 larval instars in P. paulista, and described all external structural changes occurring over immature development. From comparing our finds with previous published information by Kojima (1998), we were able to confirm most of the proposed characters, yet we could not properly apply some of them: size of antenna and postmentum. We hope that the present description will be followed by other detailed descriptions with larvae of social wasps, emphasizing on alterations that occur in immature development, which are information invaluable to understanding the biology and morphology of these species.