Evaluation of fruit and leaf morphological variability in bearberry (Arctostaphylos uva-ursi (L.) Spreng.) Dinaric Alps populations

The Dinaric Alps (Dinarids) are a mountain range of southern and south-central Europe, usually recognized as one of the most important centres of European plant biodiversity and endemism (Médail and Diadema 2009, Médail et al. 2019, Liber et al. 2020). One of the largest mountains in this area is Velebit. Situated on the Adriatic coast of Croatia, it stretches about 145 km from the Vratnik pass above Senj in the northwest to the bend of the Zrmanja river in the southeast (Forenbacher 1990). The entire mountain area has been protected as a nature park, encompassing two national parks and one strict nature reserve. Velebit is home to about 2000 plant species, subspecies and varieties with a large number of rare, endangered, and protected plants (Topić et al. 2010), such as e.g. Sibiraea laevigata (L.) Maxim. and Degenia velebitica Hayek.

In higher altitudes and on Velebit summits, particularly prominent are Arctic-alpine (northern alpine) plants, persisted in this area since the Last Glacial Period (Birks and Willis 2008), such as e.g. Arctostaphylos uva ursi (L.) Spreng., Dryas octopetala L. and Vaccinium vitis-idaea L. Presently, the main part of those plants’ distribution area is in the cold Arctic tundras, whereas in the south of Europe they only grow in high mountainous areas, where the climate is somewhat similar to the Arctic. Nowadays those plants in the southern mountainous areas are isolated because they are surrounded by a large area of what is for them an unfavourable habitat (Ronikier et al. 2023). In this isolation each group of plants lives independently, which often results in large differences between populations (Ægisdóttir et al. 2009, Bradburd et al. 2013). Today these high-mountain and Arctic plant species are considered to be particularly sensitive to climate change because of their specialized adaptation to the cold environment (Lesica and McCune 2004, Kougioumoutzis et al. 2021, Watts et al. 2022, Walas et al. 2023).

In the group of Arctic-alpine plants, because they are endangered and rare in the Dinaric area, we were particularly interested in the species Arctostaphylos uva-ursi, commonly known as bearberry. A. uva-ursi is a small procumbent woody groundcover shrub growing 5–15 cm high (Herman 1971, Forenbacher 1990). Its leaves are evergreen, leathery and shiny, oblanceolate to obovate in shape, sometimes narrowly elliptic, 1–3 long and 0.5–1 cm wide (Schütt 2008). Besides the leaves, this species is also characterized by bisexual, bell-shaped, white, pink, or pink-tipped flowers, and 0.6–0.8 cm large globose and red fruits (Idžojtić 2019). Flowering takes place in April and May, and the fruits ripen in August and September (Krüssmann 1962). The fruits persist on the plants into early winter, and are consumed by bears and many other animals. This species is widely distributed across circumboreal regions of the subarctic Northern Hemisphere (Schütt 2008), including North America, Europe and Asia. In the Dinaric area (Forenbacher 1990), there are isolated populations on karst mountainous pastures and glades, and in dwarf pine communities at altitudes above 1000 m. Due to excessive harvesting in nature, it has been placed on the Red List of Threatened Plants of Croatia, in the Vulnerable category (Marković et al. 2005), and it is strictly protected by law.

This species is best known for its leaves, which are the main natural source of arbutin (Asensio et al. 2020), used for centuries to treat urinary tract infections and other renal diseases. In addition, bearberry also has edible fruits (Schütt 2008). Raw and unprocessed fruits have an astringent and acidic taste, while cooking and drying makes them sweeter and more appetizing. Given the significance of this plant’s leaves in folk medicine, there are numerous papers dealing with their phytochemical variability (Asensio et al. 2020, Shamilov et al. 2021, Sugier et al. 2021, Kravchenko et al. 2022). However, prior to the present study, little was known about the morphological variability of bearberry. In their paper, Remphrey et al. (1983) analysed the morphology and shrub architecture of bearberry in north Canada. Due to the variability of branch architecture and hairiness, as well as leaf hairiness, as many as 14 subspecies and varieties have been recognized, although nowadays they are considered synonyms by World Flora Online (WFO 2024).

In this study, we investigated the fruit and leaf diversity of A. uva-ursi populations in the northern area of the Dinaric Alps. Specifically, we asked three questions: (1) What is the morphological diversity of bearberry fruits and leaves in the studied area?; (2) Do the morphometric traits of fruits and leaves correlate with each other?; and (3) Are there any differences between the populations and shrubs within the populations, and is there any morphological structure within the species in the studied area?

Descriptive statistics, i.e. means, standard deviations, minimal and maximal values and coefficient of variations of each morphological trait, are given in Table 2 for populations, and for the overall populations’ sample. The coefficient of variations for the overall sample ranged from 4.2% to 24.5%. In terms of the traits pertaining only to fruits, only one has normal variability, namely fruit mass (m), while the other three variables (FH, FW and FH/FW) are characterized by low variability. Furthermore, the leaves were found to have nine out of ten studied traits with variability ranging from 11.5% to 24.4%, i.e. normal variability. The PMLW/LL variable exhibited variability under 10%.

Average fruit mass (m) for populations overall was 0.24 g, fruit height was 6.7 mm, and fruit width was 7.7 mm. The fruit length-to-width ratio was 0.87, ranging from

0.79 to 0.99. The highest average fruit mass (m) of 0.37 g was measured on a shrub from population Zečjak, and the lowest average of 0.15 g on a shrub from population Zavižan.

Average leaf blade length (LL) and width (MLW) were 1.5 cm and 0.7 cm, respectively. Leaf blade length (LL) and width (MLW) ranged from 1.06 to 1.95 cm, and from 0.57 to 0.95 cm, respectively. Leaf blade width at 90% of leaf blade length (LW90) was significantly lower than maximum leaf blade width (MLW), amounting to 0.45 cm. The leaf blade length/width ratio (MLW/LL) ranged from 0.34 to 0.59, and the ratio of length and distance from leaf blade base to the point of maximum leaf width (PMLW/LL) ranged from 0.55 to 0.66. The petiole (PL) was 0.43 cm long on average.

The results of the correlation analysis of the studied leaf and fruit morphological traits showed 57 significant correlations, out of 91 possible pairs (Table 3). The majority of those were positive, with only eight negative correlations found. Furthermore, out of the total number of correlations, only 17 demonstrated r values larger than 0.7.

When leaf traits are considered, 32 significant correlations were detected, among which 12 were strong (r<0.7). The largest number of correlations was observed for PMLW (nine) and LL (eight), and the lowest for LA10 (five) and PMLW/LL (three). Observing the fruit morphological traits, significant correlations were noted between all tested pairs except between fruit mass (m) and FH/FW. On the other hand, fruit mass (m) was strongly correlated with fruit width (FW) and fruit height (FH).

Table 2 The results of the descriptive statistical analysis for the studied populations and morphometric traits. Fruit morphometric traits analysed: m–fruit mass (g); FH–fruit height (mm); FW–fruit width (mm); FH/FW–fruit height/fruit width. Leaf morphometric traits analysed: LA–leaf area (cm²); P–leaf perimeter (cm); LL–leaf blade length (cm); MLW–maximum leaf width (cm); PMLW–leaf blade length measured from the leaf base to the point of maximum leaf width (cm); LW90–leaf blade width at 90% of the leaf blade length (cm); LA10–angle closed by the main leaf vein and the line defined by the leaf blade base and the point on the leaf margin, at 10% (˚); PL–petiole length (cm); MLW/LL–maximum leaf width/leaf blade length; PMLW/LL–leaf blade length measured from the leaf base to the point of maximum leaf width/leaf blade length. Descriptive parameters: M–arithmetic mean, SD–standard deviation; Min-Max–range; and CV–coefficient of variation (%).

Table 3 The results of correlation analysis between studied leaf and fruit traits. Morphometric traits’ acronyms as in Table 2. *** significant at

p < 0.001, ** significant at 0.001 < p < 0.01, * significant at 0.01 < p < 0.05, ns depicts non-significant values (p > 0.05).

The analysis of leaf and fruit traits showed 20 significant correlations, all positive. No significant correlations were detected between fruit traits and traits related to leaf shape, or between leaf traits and FW/FH, i.e. the trait describing the shape of the fruit.

The first three principal components had eigenvalues above 1 and accounted for 84.7% of total variability (Table 4). The first principal component (PC1) was highly positively correlated with nine leaf and fruit traits (P, LL, LA, PMLW, FH, m, MLW, LW90 and FW), all referring to the size of the leaf and fruit. The second principal component (PC2) was highly positively correlated with two leaf traits (MLW/LL and LA10), and highly negatively with one leaf trait (PMLW(LL), whereas the third principal component was highly positively correlated with one fruit trait (FH/FW). The traits highly positively or negatively correlated with the second and third principal component relate to the leaf and fruit shape. Furthermore, there was no clear grouping of individuals by population within the studied data.

At the level of individual traits, the populations differed in only one (MLW/LL) of the 14 studied morphological traits (Table 5). AMOVA demonstrated that the differences between the studied individuals within the populations account for 100% of the total variability. Furthermore, to determine the diversity within each of the studied populations, multivariate diversity index (MDI) values were calculated. The MDI values in the studied populations (Table 1), based on fruit and leaf morphological traits, ranged from 3.542 (P3–Kiza) to 4.164 (P3–Zavižan).

Table 4 Pearson’s correlation coefficients between morphometric traits and scores of the first three principal components. Morphometric traits’ acronyms as in Table 2.

The leaf dimensions of A. uva-ursi obtained in this research were consistent with the data provided in the botanical literature (Krüssmann 1962, Schütt 2008). Specifically, the average leaf in Croatian bearberry populations measured 1.5 cm in length and 0.7 cm in width, which falls within the respective ranges of 1–3 cm and 0.5–1 cm listed by the mentioned authors. The same applies to the dimensions of the fruits, which, with an average height of 6.7 mm, fall within the range of 6–8 mm specified by Schütt (2008) and Idžojtić (2019). Furthermore, bearberry demonstrated moderate coefficients of variability, ranging from 7.1 to 24.5% in fruit morphology, and from 4.2 to 24.4% in leaf morphology. This represents approximately half the variability of another Arctic-alpine species, Dryas octopetala, whose variability of comparable traits ranged from 27.6 to 56.2% (Marcysiak 2014).

Table 5 The results of the hierarchical analysis of variance. Morphometric traits’ acronyms as in Table 2.

Figure 2 Biplot of the principal component analysis (PCA) based on 14 fruit and leaf morphometric traits in the studied bearberry populations. Each individual shrub is indicated by a small sign, while the population barycenters are represented by larger ones. Morphometric traits’ acronyms as in Table 2.

Significant correlations were found between both fruit and leaf morphometric traits. These findings align with general expectations, as it is well-known that when one dimension of a fruit or leaf increases, other related dimensions also tend to increase, indicating the linked nature of their development. Furthermore, significant correlations were also found between bearberries’ leaf and fruit morphological traits, particularly those related to leaf and fruit size. In other words, shrubs with smaller leaves had smaller fruits, and vice versa. Such correlation analyses are rarely conducted and often result in no significant correlations or only weak ones (Poljak et al. 2024b). Nonetheless, our results align with the findings of Fishler et al. (1983), who highlight leaf area as a crucial factor influencing fruit growth. A larger photosynthetically active leaf area could contribute more to better fruit development, ultimately affecting fruit size.

The finding that bearberry populations on Velebit do not differ significantly is indeed unexpected, given the typical genetic differentiation observed in alpine landscapes with high mountain ridges and steep valleys, which often limit gene flow among plant populations (Cain et al. 2000, Theurillat and Guisan 2001). This usually results in stronger genetic differentiation among populations compared to more homogeneous landscapes (TillBottraud and Gaudeul 2002). Alternatively, heterogeneity across small spatial scales in these habitats could also favour adaptive phenotypic plasticity (Alpert and Simms 2002, Hamann et al. 2016). However, in our study, 100% of the morphological variability was attributed to within-population variability. This suggests effective and strong gene flow between populations, which in the case of the barberry populations can be explained by the movement of the wildlife, such as deer, bears, and various small mammals and bird species, which feed on bearberry fruits (Crane 1991), and the historical and ongoing pastoral movements on the Velebit Mountain (Alegro and Šegota 2019). This human-mediated seed dispersal, along the wildlife gene flow, could counteract the population isolation typically caused by the rugged alpine terrain. In general, such effective gene flow ensures a healthy and resilient ecosystem by maintaining genetic diversity within plant populations, which is essential for their long-term survival and adaptability (Ellstrand and Elam 1993, Frankham 1996).

Furthermore, another possible explanation for the lack of differentiation between the studied populations might stem from their evolutionary history and recolonization processes after the last glaciation. Considering their geographical proximity, it is safe to assume they originated from the same refugial area. Comparable patterns of limited population differentiation, or more precisely, the absence of significant differentiation, have also been observed in other woody species, such as Juniperus deltoides R.P.Adams (Brus et al. 2010) and Salix reticulata L. (Marcysiak 2012).

Contrary to the assumption that altitude affects plant morphology (Körner et al. 1989, Cruz-Maldonado et al. 2021), our research shows that the bearberry populations we studied did not show any variations in their leaf and fruit morphology along the altitudinal gradient. This lack of phenotypic plasticity could be due to several reasons. Firstly, such a small, procumbent shrub with evergreen and small leaves may exhibit less morphological plasticity compared to physiological traits due to the inherent stability of structural characteristics in response to environmental changes (Arnold et al. 2022). Additionally, we cannot exclude the possibility that some morpho-anatomical traits, such as leaf thickness and stomatal density, in bearberry populations that we studied change with altitude, which has been previously reported in alpine plants (Bertel et al. 2022). Furthermore, similar soil conditions in mountainous pastures where these populations are located may have favoured the same phenotypes and the absence of phenotypic plasticity. Such consistent environmental conditions across different altitudes could lead to uniform selection pressure, resulting in similar leaf and fruit morphological traits despite the variation in altitude. Finally, studies on alpine plants, such as Arabis alpina L. (de Villemereuil et al. 2018), have indicated that populations from higher altitudes often exhibit reduced phenotypic plasticity compared to those from lower altitudes. This reduced plasticity could be explained by the extreme and stable environmental conditions at higher altitudes, which favour more specialized and less plastic traits. Nevertheless, although we did not find variations in leaf and fruit morphology along the altitudinal gradient, we did observe that population diversity changes with altitude. Specifically, the population from the lowest elevation exhibited the least variability, while the population from the highest elevation showed the greatest diversity.

Along with effective gene flow and the size of bearberry populations, an additional factor that positively affects the maintenance of high diversity within populations is the fact that these populations are located within national parks and nature parks. Protected areas are known to play a key role in biodiversity conservation and in reducing exploitation of the species of interest, leading to greater species richness and its abundance (Gray et al. 2016, Cavalcante de Souza and Prevedello 2020). The legal protection helps to ensure the conservation of bearberry’s genetic diversity, which is crucial for its longterm survival. Protecting such species in their natural habitats is essential for maintaining overall biodiversity and ecosystem stability (Hanjalić Kurtović et al. 2025). In general, the protection of mountain plants involves preserving their natural habitats, preventing habitat fragmentation, and mitigating the impacts of climate change. As the results of our study indicate, conservation efforts should also focus on maintaining traditional land-use practices that support biodiversity, such as sustainable grazing. Legal protection, public awareness, and scientific research including long-term monitoring are essential components in the effective conservation of these species.

The study revealed several important conclusions regarding the morphological variability of bearberry populations in the Dinaric Alps. Firstly, the observed morphological variability in fruit and leaf traits was generally moderate. This moderate variability aligns with previous studies on similar species in different regions, suggesting a stable pattern of morphological variability within this species. Significant correlations were identified among both fruit and leaf traits, as well as between them, which supports the notion that leaf area significantly influences fruit development, highlighting the connection of vegetative growth and reproductive success. No significant differences were found among populations, as all variability was being attributed to within-population differentiation. This unexpected pattern of morphological variability could be explained by several factors: the effective gene flow facilitated by the movement of wildlife and humanmediated seed dispersal through traditional pastoral practices; the evolutionary history and recolonization processes after the last glaciation; and the absence of phenotypic plasticity in structural traits. All in all, our study emphasizes the importance of conservation efforts in maintaining species diversity and ecosystem stability. The legal protection of bearberry populations within the Northern Velebit National Park and Velebit Nature Park appears to play a crucial role in preserving the diversity of bearberry populations.