Urban Habitat Mosaics Structure Invasive Termite Gene Flow Without Genetic Erosion

A replicated field study in Guadeloupe found that urban and forest habitats differed in the genetic composition of the invasive termite Heterotermes tenuis, even though colonies in the two environments retained broadly similar levels of neutral genetic diversity. Across three paired urban–forest landscapes, habitat was associated with stronger genetic differentiation than geographic distance, while colony breeding organization remained similar. The findings indicate that urban transformation can reorganize genetic structure without producing the genetic erosion commonly associated with habitat fragmentation.

The researchers sampled 50 nests during April–June 2023: five forest and five urban nests in the Baie-Mahault and Petit-Bourg pairs, and 25 forest and five urban nests in the Sainte-Rose pair. Paired sites were separated by 4.7–11 kilometres, and forest sampling points were at least 200 metres from forest edges. Urban sites were small woodlots or wastelands embedded in matrices dominated by impervious surfaces. They were warmer, more strongly illuminated at night and surrounded by more people than the corresponding forests, which were cooler, higher and more heavily covered by trees.

To characterize genetic variation, the team developed species-specific microsatellite markers for H. tenuis and genotyped 414 workers from the 50 nests. Sequencing produced 2,481,770 raw paired-end reads; after filtering and supplementary processing, the researchers identified thousands of candidate loci, validated 95 by amplification and then selected 20 for multiplex assays. Two markers were later excluded because of poor amplification and high rates of missing data, leaving 18 loci. Quality-control tests detected no evidence of null alleles, and correction for them had little effect on estimated differentiation, with global FCT changing from 0.144 to 0.137.

All 18 retained loci were polymorphic, and nests contained 30–51 alleles each. Sixty-five alleles occurred in both habitat types, whereas 14 were confined to urban nests and 21 to forest nests. Mean allelic richness was 2.35 ± 0.26 in urban habitats and 2.11 ± 0.26 in forests. Mean observed heterozygosity was 0.56 ± 0.06 in urban nests and 0.47 ± 0.10 in forests, while expected heterozygosity averaged 0.46 ± 0.05 and 0.38 ± 0.07, respectively. These descriptive differences, however, did not translate into a general habitat effect in mixed-effects models that accounted for locus, nest, site and site pair.

After the hierarchical analysis, habitat explained little variation in any genetic-diversity measure. Its estimated effects were 0.098 for allelic richness, 0.039 for expected heterozygosity, 0.055 for observed heterozygosity and 0.025 for the inbreeding coefficient, with none statistically significant. Pair-specific results showed that the apparent urban advantage was concentrated at Sainte-Rose: urban nests there had significantly greater allelic richness, expected heterozygosity and observed heterozygosity than forest nests. Neither the Baie-Mahault nor Petit-Bourg pair showed a detectable habitat difference in diversity. The overall conclusion was therefore preservation, rather than loss, of neutral genetic variation in urban colonies.

Genetic differentiation among individual nests was substantial in both habitats. Mean FCT was 0.255, with a 95% confidence interval of 0.224–0.292, in forests and 0.243, with an interval of 0.213–0.275, in urban sites. Principal-coordinate analyses revealed habitat-associated separation in each of the three pairs, while hierarchical clustering showed the clearest division between habitats at Sainte-Rose, intermixed habitats at Petit-Bourg and an intermediate pattern at Baie-Mahault. Permutational multivariate analysis of variance attributed 14.4%, 17.0% and 21.5% of genetic variation to habitat in the three pairs, respectively.

The habitat pattern persisted in a global analysis of all 50 nests. In marginal tests, habitat explained 13.8% of total genetic variation and site pair explained 15.9%, while their interaction was not significant. Multivariate dispersion did not differ between habitats or site pairs, making unequal within-habitat spread an unlikely explanation. The researchers also repeated the analysis across 999 random subsamples containing five nests per site; habitat remained significant in every subsample, with a median R2 of 0.198 and a 95% percentile range of 0.179–0.218. This sensitivity analysis addressed the unbalanced number of forest nests collected at Sainte-Rose.

Habitat-related differentiation was more pronounced than isolation by distance. A positive relationship between geographic and genetic distance occurred at only one of the six sites, the Baie-Mahault forest site, where the Mantel correlation was 0.600 and p was 0.033. No significant relationship appeared in the other sites. A maximum-likelihood model incorporating repeated nest pairs found no geographic-distance effect, with a coefficient of 0.001 ± 0.009 standard error and p = 0.878, whereas comparisons between habitats were more strongly differentiated than comparisons within either habitat, with a coefficient of 0.128 ± 0.042 and p = 0.002.

Colony breeding structure did not appear to change markedly with urbanization. No mixed-family colonies were detected. Thirty-seven of 50 nests, or 74%, were classified as extended families containing multiple related neotenic reproductives, while 13 were simple families headed by a monogamous pair. Extended families accounted for 28 of 35 forest nests and nine of 15 urban nests, and their frequencies did not differ significantly between habitats (Fisher’s exact test, p = 0.170). Because the prevalence of multiple reproductives can affect effective dispersal, relatedness and within-colony genetic variation, this stability indicates that landscape context and colony biology may operate through distinct components of genetic structure.

The results do not establish that urban habitat alone caused the differentiation. Forest and urban sites also differed in elevation, soil temperature, artificial light, human density and tree cover, and the study design could not separate these correlated variables. Neutral microsatellites likewise cannot distinguish demography from adaptive differentiation. Historical introductions, habitat-dependent colony establishment and human-mediated transport remain plausible explanations: H. tenuis nests in food resources, so movement of infested wood or plants could generate long-distance connections while local habitats retain distinct genetic composition. That transport mechanism was not directly tested, and the study covered only three site pairs in an introduced range without native-range comparisons, limiting generalization to other tropical cities and invasion histories.

Nevertheless, the study shows that genetic diversity and genetic structure need not move together. Urban colonies retained levels of allelic richness and heterozygosity comparable to those in forests, yet urban and forest assemblages consistently differed in genetic composition. At the spatial scale investigated, habitat context was more informative than simple distance between colonies, and breeding organization remained stable across the two environments. Resolving whether the pattern reflects contemporary dispersal, differential establishment, environmental selection or partially admixed introduction lineages will require native-range sampling, genome-wide analyses, continuous measures of urbanization and explicit modeling of transport corridors.

Subject of Research: Urban landscape genetics and colony breeding structure in the invasive termite Heterotermes tenuis in Guadeloupe

Article Title: Urban Habitat Mosaics Structure Gene Flow but Preserve Genetic Diversity in an Invasive Termite

Article References: Gheyle, E., Duquesne, E., Roisin, Y., & Fournier, D. (2026). Urban Habitat Mosaics Structure Gene Flow but Preserve Genetic Diversity in an Invasive Termite. Ecology and Evolution, 16(10), Article e74439. https://doi.org/10.1002/ece3.74439

Image Credits: AI Generated

DOI: 10.1002/ece3.74439

Keywords: urban landscape genetics, invasive termites, Heterotermes tenuis, microsatellite markers, gene flow, genetic diversity, population structure, colony breeding systems, Guadeloupe, urbanization, Urban, Habitat