An ecological investigation of decaying coarse woody debris in montane forests has demonstrated that an eight-class log decomposition scale originally proposed by McCullough in 1948 effectively captures distinct successional dynamics across cryptogamic and vascular plant communities. Reporting in Annals of Forest Science, researchers Damian Chmura, Monika Staniaszek-Kik, and Jan Żarnowiec evaluated 630 colonized fallen logs across four distinct forest types in southern Poland to resolve long-standing questions regarding how species assemblages change as wood deteriorates. Rather than following a uniform trajectory, the succession of lichens, liverworts, mosses, and vascular plants revealed pronounced taxonomic divergence along the wood decay continuum. The study established that species turnover in the most species-rich forest stands closely mirrors the classic relay floristics model, wherein sequential species groups replace one another over time, rather than an initial floristic composition pathway where pioneer colonizers simply persist alongside newly arriving organisms.

Coarse woody debris represents an essential yet temporary habitat within forest ecosystems, serving as an ephemeral substrate for saproxylic invertebrates, fungi, and epixylic vegetation. As fallen logs decompose, gradual modifications in physical structure, wood density, chemical composition, and water retention continuously reshape the microenvironmental conditions available to colonizing organisms. Historically, researchers have categorized this physical progression using decay classifications ranging from four to eight stages, beginning with early visual assessments such as Hackiewicz-Dubowska’s five-class system in 1936 and McCullough’s eight-class system formulated in 1948. McCullough’s scale, inspired by trunk decay classifications previously applied in entomological studies, tracked log deterioration through visible physical criteria, notably log shape changes, wood softening, moisture accumulation, and the progressive sloughing of bark and branches. Subsequent ecological research has frequently debated whether an eight-stage classification provides genuine ecological resolution or introduces redundant stages, with competing contemporary frameworks advocating for simplified four-, five-, or six-degree systems.

The primary theoretical aim of the research was to test whether the vegetative colonization of deadwood conforms to Frank Egler’s relay floristics model or his alternative initial floristic composition scenario. Under the relay floristics framework, inspired by early Clementsian succession concepts, colonizing species progressively modify their environment and are sequentially replaced by subsequent waves of organisms, culminating in high taxonomic turnover across advancing decomposition stages. Conversely, the initial floristic composition hypothesis posits that most species establish relatively early in succession, with subsequent compositional shifts reflecting changes in dominance and competitive suppression rather than complete species replacement. Because deadwood availability diminishes over time while its physical matrix softens and accumulates moisture, understanding whether epixylic cryptogams and vascular plants undergo continuous replacement or gradual accumulation is critical for characterizing forest microhabitat dynamics. Furthermore, modern analytical advances in beta-diversity partition compositional change into nestedness and turnover components, providing statistical means to assess these historical ecological theories along complex environmental gradients.

Field investigations were conducted within the Karkonosze National Park and its surrounding buffer zone in the Giant Mountains of southwestern Poland, a high-relief subrange of the Sudetes characterized by Carboniferous granite and metamorphic bedrock. To capture diverse montane conditions, the investigators deployed a stratified sampling design across 180 established ten-meter by ten-meter plots alongside 164 additional logs situated outside plot boundaries, examining a total of 650 coarse woody debris items. Among these, 630 logs supported living flora and lichens, while 20 logs remained completely uninhabited. Tree species identifications revealed that 447 colonized logs were Norway spruce (Picea abies), 156 were European beech (Fagus sylvatica), 20 were silver birch (Betula pendula), three were rowan (Sorbus aucuparia), two were European larch (Larix decidua), and two could not be taxonomically determined. The logs were distributed across four distinct forest communities: lower montane beech forests, lower montane fir-spruce stands, lower montane planted spruce forests, and upper montane spruce forests belonging to the Calamagrostio villosae-Piceetum association.

At each sampling location, the researchers recorded extensive geographic, topographic, dendrometric, and microclimatic variables to characterize the physical decay environment. Absolute elevation was documented using geographic positioning systems, terrain slope was measured using a clinometer, and aspect was converted to a continuous degree scale. Overhead canopy cover was visually categorized into five shade classes ranging from under thirty percent to over eighty percent illumination reduction. For every log, investigators measured length, base diameter, and top diameter with measuring tapes, subsequently calculating total surface area based on the geometric formula for a truncated cone. Ground adherence, defined as the percentage of log length in direct physical contact with the forest floor, was recorded alongside tactile moisture assessments scored on a three-point ordinal scale. Degree of decomposition was assigned using McCullough’s original eight-stage classification, while vegetative cover for all inhabiting lichens, bryophytes, and vascular plants was estimated using modified Londo percentage intervals.

To evaluate species richness without distorting count data characterized by excessive zeros, the authors formulated zero-inflated generalized linear mixed models using a Poisson distribution and log link function within the glmmTMB statistical package. The statistical structure incorporated an interaction term between the four taxonomic groups and the four forest communities, treating McCullough’s eight decay stages as a continuous, mean-centered environmental covariate while designating sampling sites as random intercept effects. Simulation-based residual diagnostics confirmed appropriate model dispersion, uniformity, and zero inflation handling. Community-level composition was evaluated using Bray-Curtis dissimilarity matrices, multivariate dispersion testing, and principal coordinates analysis. Indicator species analysis was executed across decay stages using 999 permutation iterations to determine whether specific taxa were strictly confined to particular decomposition classes. To identify which external environmental factors most strongly directed community organization, the authors performed multidimensional fuzzy set ordination based on Horn distance dissimilarity alongside classical principal component analysis.

Across the 630 surveyed logs, the botanical inventory identified 195 distinct species, comprising 68 mosses, 53 lichens, 45 vascular plants, and 29 liverworts. Spruce logs hosted the highest overall diversity with 165 recorded species, averaging roughly eleven species per log, followed by beech logs with 110 total species, birch logs with 46, rowan with 27, and larch with 27. Likelihood ratio testing demonstrated a highly significant interaction between taxonomic groups and forest types, confirming that richness responses along the decomposition gradient differed fundamentally among organism types. Mosses consistently exhibited the highest estimated marginal mean richness across all forest types. In upper montane spruce forests, liverworts, mosses, and vascular plants exhibited distinct humped-back richness curves that peaked at intermediate decomposition stages, whereas lichen richness declined steadily as logs decayed. Conversely, in fir-spruce forests, taxonomic richness generally climbed across successive decay classes, with lichens showing only a slight intermediate rise before dropping.

Multivariate testing of beta-diversity confirmed statistically significant compositional differences across the eight decay stages when all forest types and taxonomic groups were analyzed simultaneously. Pairwise comparisons revealed that every subsequent decomposition stage differed significantly from the initial decay class, demonstrating that vegetative communities undergo progressive restructuring as wood breaks down. However, compositional differences between the sixth, seventh, and eighth decay classes were not statistically significant, reflecting community stabilization as logs approached advanced humification. When examining pairwise differences from the first decay stage across individual taxonomic groups, lichens exhibited nineteen significant stage differences, liverworts exhibited fifteen, mosses showed thirteen, and vascular plants showed only nine. Crucially, indicator species analysis revealed that multiple taxa were statistically confined exclusively to the eighth decomposition class, including liverworts such as Calypogeia integristipula, mosses such as Polytrichum formosum, and vascular plants like Homogyne alpina and Calamagrostis villosa, verifying that the eighth stage is ecologically non-redundant.

Multidimensional fuzzy set ordination and principal component analysis demonstrated that wood decomposition stage and elevation were the two primary drivers governing epixylic community assembly, jointly yielding a cumulative correlation coefficient of 0.65 with vegetative dissimilarity. Log decomposition correlated positively with forest floor adherence and tactile moisture, but correlated negatively with bark cover and overall log length. Stepwise ordination revealed that remaining bark cover, log surface area, and upper log diameter also exerted statistically significant influences on species composition. The researchers observed that the replacement of pioneer lichen communities by bryophytes and vascular plants reflects shifting substrate conditions; decortication strips away epiphytic relics, while exposed, softening wood provides moisture and nutrients favorable to true epixylics and late-successional generalists. Because lichens favor drier, well-aerated bark surfaces, they experience pronounced competitive displacement as water-retaining mosses expand across humifying logs, explaining why lichens displayed the highest rate of species turnover.

The authors concluded that McCullough’s eight-class decay scale provides necessary ecological resolution for tracking deadwood vegetation, with the highest species turnover occurring in upper montane spruce forests where trajectories strongly align with relay floristics. In contrast, vascular plants behaved as late-stage colonizers with low turnover, gradually invading advanced decay stages from the adjacent forest floor where rotting wood approximates humified soil. Nonetheless, the authors acknowledged important methodological limitations inherent to space-for-time chronosequence sampling, noting that variation in log dimensions, microtopography, and uneven decay class distributions introduce spatial variability that cross-sectional surveys cannot entirely disentangle. They emphasized that confirming absolute temporal succession dynamics will ultimately require multi-decadal permanent plot monitoring from the moment of tree fall until complete organic disintegration. Even so, the empirical findings demonstrate that eight-stage decomposition scales capture subtle, ecologically meaningful floristic transitions that coarser classification systems inevitably obscure.

Subject of Research: Vegetative succession of lichens and plants across eight stages of deadwood decomposition in montane forests

Article Title: Lichen and plant succession on decaying logs: McCullough’s (1948) decomposition scale revisited

Article References: Chmura, D., Staniaszek-Kik, M., & Żarnowiec, J. (2026). Lichen and plant succession on decaying logs: McCullough’s (1948) decomposition scale revisited. Annals of Forest Science, 83(1), Article 28. https://doi.org/10.1186/s13595-026-01343-9

Image Credits: AI Generated

DOI: 10.1186/s13595-026-01343-9

Keywords: coarse woody debris, deadwood decomposition, McCullough decay scale, relay floristics, initial floristic composition, epixylic flora, lichens, bryophytes, liverworts, mosses, vascular plants, beta-diversity, species turnover, Karkonosze National Park, Giant Mountains, multidimensional fuzzy set ordination, forest ecology, montane spruce forest