According to Blackwelder (1982) the range of this large beetle (19.8 ± 1.5 mm, n=6) (Figs 1a–c) extends from Mexico to Nicaragua. However, specimens collected later (L.D. Gomez near San Vito, Costa Rica; H. Stockwell, Cerro Campana, Panama Prv., 30 V 70 and 18 VII 76; D.M.W., Los Santos Prv., Cerro Canajagua, 25 V 92 and Colon Prv., Cerro Galera, 1 V 02; M. Cuignet, Colon Prv. Sta. Rita Ridge Rd. km 2, 2 XI 02; S. Van Bael, Bocas del Toro Prv., Chiriqui Gde., 17 I 04; M. Membache, 1 VI 92, Colon Prv., Gamboa; S. Lankowski, Panama Prv., Parque Metropolitano, 15 IV 07) indicate the species extends to at least 10 km east of the Panama Canal. These records plus observations of a Doryphora paykulli adult following a tightly arranged group of larvae moving between leaves on their food plant near Chiriquí Grande, Bocas del Toro Province (S. Van Bael, pers. comm.) documents the presence of the species in the Caribbean as well as in the Pacific lowlands of Panama and provides the first unequivocal record of subsocial habits for the species.

Subsequently, individual Doryphora paykulli adults were observed (D.W., S.L.) during late April and early May of 2005 moving slowly and feeding among low vegetation. Pairs of individuals were observed interacting aggressively on the small leaf fragments remaining on the host plant, Prestonia seemannii Miers (Apocynaceae) under late dry season conditions along trails in the Parque Metropolitano (elevation 30 m, 8°59.24'N, 79°32.797'W), Panama City. Whether these were contests over resources or precopulatory courtship is unclear, however, analysis of video taken of one of these interactions shows the use of the mesosternal horn in dislodging a competitor, much as described by Eberhard (1981) for the closely related beetle, Doryphora sp. near punctatissima. Adult females were discovered during mid-May of both 2005 and 2007 in close proximity to eggs attached to the underside of cupped, newly expanded Prestonia seemannii leaves (Fig. 1a) on the western slope of Cerro Pelado, Gamboa (elevation 95 m, 9°7.29'N, 79°41.78'W). Eggs measured 3.5 × 0.9 mm and at first were an opaque, cherry-red, with the chorion becoming transparent and larvae distinguishable as development progressed. Clutches initially contained ten eggs with roughly ten new eggs added each day over the course of 3 to 5 days. The female became noticeably more defensive the second day of oviposition, straddling the eggs and jerking from side to side apparently in response to movement by the observer.

Larval emergence began on days 6 and 7 following first oviposition with clutches (n=3) at that time containing from 40–50 eggs. Within 24 hours after larvae began to emerge many of the original eggs were missing, apparently consumed by early-emerging larvae such that clutches were reduced to 8, 15 and 20 surviving larvae accompanied by some intact, opaque, less-developed eggs and opened eggs with red-colored residues of once-developing larvae visible under close inspection (Fig. 1). One day later an adult was observed (S.L.) returning to a natal group after having fallen from the plant on a detached leaf, then deposit a single egg which was immediately consumed by a nearby larva. First instar larvae (1.5–2.0 mm in length) had small black heads at the time of emergence, these easily distinguished them from second instar larvae appearing, 1 to 2 days later with red head capsules and a larger and more rotund appearance. Larvae expanded rapidly in size following their first meals but did not feed on leaf tissue until after the first molt, 2 to 3 days after emerging from the egg. Mothers at times tightly straddled their aggregated first and second instar larvae on the natal leaf, preventing their advance down the leaf petiole (Fig. 1b). While guarding, mothers reacted aggressively by charging to the edge of the leaf when a thin stick was introduced to the area by an observer. Charges, stamping and shaking continued for at least two minutes after the stimulus was presented and removed. The strongest reaction was given to a camera held approximately 10 cm under and to the side of the natal leaf. The mother seemed to be reacting to the camera lens–suggesting that a mirror held near guarding mothers might provide a non-invasive means of assessing defensiveness. On one occasion an Ectatomma tuberculatum (Olivier) worker was observed to pass by the base of the petiole, eliciting aggressive shaking of the natal leaf and short charges, after which the ant reversed course and departed that portion of the plant. As larvae became larger and began moving between leaves the intensity of the mothers’ reactions to foreign stimuli appeared to subside. Increasingly, mothers were seen feeding on leaves and leaf petioles, rather than guarding, as larval development proceeded.

Once the natal leaf was consumed, larvae began moving down the petiole to the stem, where they then moved either up or down in smaller groups to other leaves, sometimes moving as solitary individuals. Mothers, often fed from the pedicel of the leaf just consumed, occasionally accompanied by one or two larvae. Mothers actively trampled upon the backs of larvae still located on the pedicel, in effect pushing them away from the leaf and toward the stem. Mothers on other occasions stepped on and over larvae, rapidly tapping larvae with antennae and tarsi until they reversed direction. After leaving the natal leaf, mothers resumed guarding one of the several larval groups that reassembled. However, some groups continued to split into ever smaller units and moved to adjoining leaves and stems, leaving mothers guarding smaller sets of offspring and spending more time travelling among groups in what seemed to the observer as an effort to herd offspring back together (Fig. 1c). Mothers also increasingly divided their time between guarding and feeding on the cortex of the stem, girdling the vine over distances from a few cm to nearly 1 m. Seven days following eclosion larvae were arranged largely in doublets, girdling stems as intact leaves had disappeared in substantial sections of the plant. Larvae then descended in unguarded pairs to pupation sites by backing down the stem from which all cortex was stripped (Fig. 1d), effectively killing that section of the plant. By day eleven, most larvae had descended the host plant and moved along small above-ground roots into the leaf litter. One mother was last seen guarding two slow-developing larvae high (2 m) on the plant thirteen days after oviposition. Development from first oviposition to larvae wandering on the ground took approximately 20 days. Several larvae collected and placed in a plastic container with moist leaves, molted at day 4 and eclosed as teneral adults on days 18 and 19, a metamorphosis period slightly shorter than the 24 days estimated for Doryphora sp. near punctatissima by Eberhard (1981). Development from oviposition to the eclosion of adults in mid-June required 35 days. Monthly visits to the study area throughout the rest of the year were successful in finding solitary, feeding adults on nearby host plants but not in finding signs of additional reproductive activity. Thus, this species seems to have but a single generation per year timed to the period of accelerated leaf growth by its host plant. The nearly synchronous May onset of reproduction in both Doryphora paykulli in Panama and Doryphora sp. near punctatissima in Colombia (Eberhard 1981) is likely the product of similar climate regimes in the two species’ ranges and subsequent effects on host phenology. Finally, the high morphological similarity of these neighboring species, the presence of larval cannibalism in both, but the presence of maternal care in only one, raises intriguing questions regarding the lability of defensive behaviors and underscores the importance of reconstructing phylogenetic relationships for as many Doryphora species as possible.

Recent observations by F.F. reveal clearly that maternal care is expressed by Doryphora (Megistomela) reticulata (Fabr.) in the cerrado of south-central Brazil (Fig. 5) (see also photo in Chaboo 2011). Photographs of this species in the Boqueirão Biological Reserve, Minas Gerais of Brazil (elevation 1200 m; 21°20.76'S, 44°59.49'W) in 2005 clearly show behaviors strikingly similar to that observed in Doryphora paykulli in Panama. Females oviposit on the underside of partially-expanded, apical leaves of Prestonia tomentosa (Apocynaceae) (Fig. 1e, f). Larvae emerge and are tightly straddle-guarded by the female (Fig. 1g), but unlike Doryphora paykulli and Doryphora sp. near punctatissima, no larval cannibalism of eggs was observed. Larvae guarded by the mother continued to feed on leaves and strip cortex, eventually descending to the ground tended by the mother prior to pupating nearby in the soil (Fig. 1h). Indeed, of the many Chrysomelinae species associated with Solanaceae and other plant families studied at Serra do Japi and other sites near Campinas in Central Brazil (Table 1), Doryphora reticulata is the only species in which mothers are known to actively guard their larval brood. Inferences regarding subsocial habits in Platyphora conviva (Reid et al. 2009) are incorrect according to J. Vasconcellos-Neto (personal communication, 2013). Further, ongoing studies in the eastern lowlands of Bolivia by one of the authors (D.W.) have found no evidence of maternal care occurring in any of 16 species of Doryphorini.

Within New World Chrysomelinae, reports of subsociality until recently were limited to a single species studied at the La Selva Field station in the Atlantic lowlands of Costa Rica (Choe 1989). However a misidentification of that species (not by the author) lead to erroneous attribution of subsocial behavior to Labidomera suturalis, rather than to an unidentified species of Platyphora. The species was subsequently described and named Platyphora selva by Daccordi without comments on Choe’s behavioral observations (Daccordi 1993). As noted by Reid et al. (2009), this first record of subsocial behavior in Neotropical chrysomelines led to a number of reports citing the original paper and repeating the taxonomic error (e.g. Windsor and Choe 1994, Kudô and Hasegawa 2003, Costa 2006).

Choe (1989) observed 18 guarding Platyphora selva females in two different years, all feeding on Lycianthes (Witheringia) heteroclita Sendtm. (Solanaceae) in the Atlantic lowlands of Costa Rica. His observations were remarkable in first describing how females of this species tightly guarded offspring by straddling. By removing mothers from roughly half of the families, he was able to demonstrate that guarding was highly effective in preventing predation by the gigantic ponerine ant, Paraponera clavata Fab. The importance of maternal defenses in reducing losses to parasitoids, however, was not investigated. Further, it was noted that mothers always guarded groups of four or fewer larvae; but eggs of the beetle were never observed during the study. From observations of related taxa (see below) we now suspect that Platyphora selva is not oviparous, but instead deposits temporally isolated clutches of four larvae. This inference remains to be documented and is based on the habits of the morphologically similar species, Platyphora microspina, which occurs widely (but rarely) in neighboring Panama. Regrettably, sequence data are not yet available for Platyphora selva.

Platyphora microspina was initially observed on Cerro Campana (Parque Nacional Altos de Campana), along a ridgeline approximately 50 m west and up-slope from the Podocarpus trail in July 1999 (elevation 900 m; 8°41.07'N, 79°55.82'W). Large numbers of adult and immature beetles were observed feeding on Markea megalandra (Dunal), a woody hemiepiphyte which grows within the canopy of forests at elevations of 1000–2000 m in Western and Central Panama (Correa et al. 2004). Larvae and adults of Platyphora microspina (Fig. 2a) were largely associated with quick-growing sprouts coming from a portion of the plant damaged earlier by limb fall. A small number of Platyphora microspina including one female tending three small and partially-sclerotized larvae were moved to a terrarium in an air conditioned laboratory containing host plant cuttings to facilitate observations. Larvae remained physically in contact with one another, often beneath one or more legs of the mother during the first three days. However, as larvae grew in size and spent more time feeding, the mother moved to the side of the group for the remaining 10–12 days of development and feeding (Fig. 2b). Single larvae occasionally left the aggregation, apparently to find new leaves, and through alternative bouts of substrate tapping with the tip of the abdomen—approximately one to two taps per second for two to three minutes—isolated larvae appeared able to call or stimulate their siblings and mother to visit new feeding sites (Fig. 2c). The mother was also observed physically nudging inactive larvae. As the first cohort neared the end of its feeding period the mother deposited another cohort of 4 larvae, briefly leaving 7 larvae of two distinct cohorts and size classes together under the mothers care (Fig. 2d). One day later larvae in the first cohort fell to the base of the terrarium, became inert and later pupated. An additional two cohorts produced by the same mother, each containing 4 larvae, were subsequently observed in the lab. Larvipositions occurred over a span of 28 days, 10 days between cohort 2 and 3, 17 days between cohorts 3 and 4. The larval feeding period for the second cohort lasted 21 days. The non-feeding prepupal period lasted 8 days and the pupal period 7 days. Thus, the interval between larviposition and adult emergence takes approximately 5 weeks in this species. Observations were terminated after the fourth cohort, so reproduction possibly continues for an even longer period in this relatively non-seasonal, premontane forest. The species is dependent upon the continued presence of its hemiepiphytic host plant on Cerro Campana and similar small refuges along the cordillera passing through Panama. Approximately eight years after these observations were made a single adult specimen resembling Platyphora microspina was collected at the Cana field station near the Colombian border. One year later at the same site, a group of recently emerged adults of the same species were observed on a woody shrub in the family Solanaceae, 3–5 m to the side of the entrance of to Cana gold mine. Preliminary analysis of its COI gene sequence shows that it is nearly identical to that of Platyphora microspina on Cerro Campana. Continuing observations at this and similar remote sites, coupled with molecular sequencing, should add considerably to our knowledge of this species and its relationship to Platyphora selva and similar species in South America.

Female Proseicela vittata (Fig. 3a) were observed by D.W. tending offspring during each of three visits to Montagne de Kaw, French Guiana (elevation 150 m; 4°32.686'N, 52°09.151'W), 11–18 April 2010, 25–31 January 2011 and 19 June–2 July 2012. All individuals and family groups were found on a single species of host plant, Solanum morii S. Knapp, a 1-3 m shrub with glabrous leaves and pendulous green and white fruit and a colonist of disturbed areas (Plate 121a, Mori et al. 2002). We searched host plants for this species mainly along logging roads. While numerous individuals and family groups were found on each trip, most groups contained older larvae. Only four females were found tending recently deposited larvae, 11 to 18 in number, which had not begun to feed. The brood tended by one female contained a single large larva feeding and resting beside 17 freshly deposited larvae (Fig. 3b). Within 2 days the single large larva descended alone to pupate, a sign that while broods may overlap in this species (as in Platyphora microspina) the period of overlap is brief. Normally, individual larvae within cohorts were remarkably similar in size (Fig. 3c). The only exception came if they were observed on day 2 or 3 while molting was in progress. Despite three observation periods per day of approximately 15 min per family, possible predators and parasitoids were rarely observed. And while a Pachycondyla ant or a carabid beetle may have been responsible for the abrupt loss of 15 of 17 larvae from one female over night, ongoing predation has yet to be observed in this species.

The 17 larvae belonging to another female were observed to take approximately 30 hours to consume the entire lamina of the natal leaf. While the last of the leaf was being consumed some larvae began to molt while still on the remnants of the natal leaf. The mother maintained a tight grip on the leaf petiole (blocking behavior), but eventually larvae pushed by and began traversing nearby stem and petioles solitarily or in small groups. Commonly families split into two or more separated feeding groups at this stage, with the mother usually remaining with a larger group. Groups often reunited but others remained separated until pupation. The transition from the natal to second leaf appears to be a crucial and dynamic time for larvae and events proceeded differently for most groups. During this period mothers moved actively among different leaves and branches in what seemed to be attempts to herd and reconstitute a single larval group. While we observed what we interpret as herding behavior in most species in this report, its possible importance to group safety and success remains open and in need of experimental study.

Strong indications of maternal care in Proseicela bicruciata (Fig. 3d) were gathered by G.D. while walking trails in the montane cloud forest of the Yanayacu Biological Station and Center for Creative Studies, Napo Province, Ecuador (2150 m elevation, 0°36.27'S, 77°53.25'W) during the first week of July 2011. A total of five females were found, each tending small groups of uniform larvae (Fig. 3e), on the undersides of large leaves of Solanum abitaguense S. Knapp growing in streamside habitats (Fig. 3f). The group containing the smallest larvae was composed of nine individuals, while groups with larger larvae contained five and six individuals. The group with five larvae had two individuals separated on leaves 20–30 cm in different directions from the central three larvae arranged in a small rosette with heads to the inside and the mother to the side.

Observations of Proseicela spectabilis were taken by G.D. while walking the main trail leading to the Cascadas de San Rafael, Reventador, Napo province, Ecuador (1300 m elevation; 0°6.07'S, 77°35.18'W) on July 17, 2011. A single female (Fig. 3g), was perched half above its larvae and half on the branch of its food plant. Six of the seven larvae in the formation were tightly aggregated around the thin stem of the plant, while the seventh larva was on a leaf a few centimeters away. A tachinid fly was present on the dorsal surface of a larva located on the side opposite the mother (Fig. 3h). The fly departed when the observer approached but promptly returned to land on the larva opposite the mother. The host plant (Fig. 3i) was later identified as a nightshade, Solanum sp. section Dulcamara.