The investigation took place on Barro Colorado Island in Lake Gatun, within the Panama Canal. Erin Krichilsky collected nests from April through June 2018 by walking through the forest between 09:00 and 15:00 local time, when the nocturnal bees were expected to be inside their nests. Potential nesting pieces were identified in the understory by a circular entrance collar. Approximately 250 nests were collected during the broader field season for other research, but the anatomical study included only occupied pieces that were not too rotten or damaged. The analyzed sample comprised 30 solitary and 11 social nests of M. genalis, together with 15 solitary and three social nests of M. amoena.

Researchers divided the 59 nest-bearing pieces into 14 wood morphospecies and examined one representative from each group. Disks about 4 cm long were cut from the samples, and thin transverse, radial longitudinal and tangential longitudinal sections measuring 18–30 micrometres were prepared where the material was firm enough to section. Other samples were photographed with a scanning electron microscope because decay prevented cutting. Investigators measured 25 anatomical features under magnifications from 4 to 1,000 times and compared the material with wood-identification databases and published anatomical descriptions. The measurements included vessel density, mean tangential vessel diameter, vessel-element length, fibre length, ray height, ray frequency, perforation plates, tracheids, pits, axial parenchyma and vessel dimorphism.

The nest substrates represented eight plant families and 11 genera: seven lianas and four trees. Uncaria tomentosa accounted for 22 samples, Hylenaea praecelsa for 12, Guettarda sp. for five, Faramea sp. for four, Combretum sp. for three, Theobroma sp. and Terminalia sp. for two each, and Hippocratea volubilis, Serjania sp., Davilla sp. and Leretia cordata for one each. Three additional morphospecies could not be identified because their tissues were too deteriorated. Four plants—U. tomentosa, H. praecelsa, Guettarda sp. and Terminalia sp.—contained nests made by both bee species. The samples did not show a difference in substrate use between solitary and social colonies, although M. amoena used stems and branches of smaller diameter than M. genalis, consistent with an earlier observation.

Microscopic comparison revealed consistent anatomical differences between the liana and tree substrates. Lianas generally had wider vessels, longer vessel elements and longer fibres. Mean tangential vessel diameter ranged from less than 50 micrometres to more than 200 micrometres across the lianas, while vessel density ranged from 5–20 vessels per square millimetre to 40–100 vessels per square millimetre. Mean vessel-element length extended from less than 350 to about 800 micrometres. All of the lianas showed vessels in two distinct size classes, and most had simple perforation plates. Other recurring features included abundant axial parenchyma, broad rays and, in some species, vasicentric tracheids. The tree substrates generally had narrower vessels, more fibres, higher wood density and lower vessel density per square millimetre.

These traits fit the structural and hydraulic demands associated with climbing plants. Lianas rely on host trees for support and include species with stems extending more than 10 metres, large leaf area and substantial water demand. Wide vessels and simple perforation plates can support efficient water conduction, while narrow or vasicentric tracheids occur alongside the larger conduits. The mixture of soft and stiff tissues, relatively abundant axial parenchyma and broad rays is also associated with flexibility, storage and injury repair. In the literature, this combination of anatomical characteristics has been described as the “lianescent vascular syndrome.” Because lianas depend on host trees and must remain mechanically flexible, their wood differs from that of self-supporting trees and shrubs in several measurable ways.

The nesting wood was described as moderately hard to soft and, in many instances, nearly crumbling. The authors had expected Megalopta bees to use material soft enough to chew with their robust mandibles but strong enough to protect the nest, although mechanical strength was not measured in this study. Some analyzed pieces nevertheless remained firm, showing that the bees can excavate in relatively resistant material. The observations therefore link the sampled substrates with their anatomy, but do not quantify the relationship between wood hardness and nest-site selection. The study also did not test whether the bees selected particular plants because of nesting suitability, their association with forage plants, or a combination of those factors. One tree substrate, Faramea sp., has also served as a pollen source for both bee species and has in some years been abundant in their brood cells, but the present data cannot establish the reason for its use as nesting material.

Plant abundance limits what can be inferred from the number of nests. A survey of rooted lianas larger than 1 cm in diameter within the island’s 50-hectare Forest Dynamics Plot recorded 911 Hippocratea volubilis individuals, 659 Davilla individuals and 206 U. tomentosa individuals; totals varied substantially among Combretum and Serjania. The survey covered the plot rather than the entire island, and its figures were not combined with a statistical comparison of substrate availability and nest frequency. Although U. tomentosa was less abundant in that plot than H. volubilis, it produced the largest number of nest samples. However, cat’s claw has brown bark that peels in sheets and distinctive terminal branches with square cross-sections and paired thorns, making it readily recognizable to people without specialist botanical training. Researchers familiar with local nesting patterns may therefore have checked cat’s claw more often, biasing its representation in the collections.

The results provide a detailed anatomical baseline for future tests of how wood-nesting bees use forest substrates. Suspended or elevated dead wood can persist above the forest floor under conditions that slow decomposition, while many lianas have narrow stems and anatomically distinctive tissues. Those traits may place suitable nesting material in the understory, where the bees construct brood cells within stems and branches. Availability nevertheless changes as dead material decays, loses mass through water percolation, accumulates salt or falls to the ground, where decomposition may be faster. By documenting the plants and cellular structures used by two species of Megalopta, the study shows that nest substrates vary widely and that lianas supplied most of the observed specimens. Its central limitation is equally clear: the observations identify what the bees used in the sampled area, but do not prove a preference for lianas, cat’s claw or any other plant after accounting for substrate abundance, detectability and decay.

Subject of Research: Wood-anatomical traits and nest substrates of two nocturnal Megalopta bee species in Panama

Article Title: Plant species used as wood substrata for nesting by Megalopta genalis Meade-Waldo, 1916 and M. amoena Spinola, 1853 ( Hymenoptera , Halictidae )

Article References: Harwood, J. A., & Wcislo, W. T. (2025). Plant species used as wood substrata for nesting by Megalopta genalis Meade-Waldo, 1916 and M. amoena Spinola, 1853 (Hymenoptera, Halictidae). Journal of Hymenoptera Research, 98, 293-309. https://doi.org/10.3897/jhr.98.108041

Image Credits: AI Generated

DOI: 10.3897/jhr.98.108041

Keywords: Megalopta bees, nest substrates, wood anatomy, lianas, Barro Colorado Island, Panama, nocturnal bees, social bees, plant species, Plant, species, used