Temporal variation in climatic factors influences phenotypic diversity of Trochulus land snails

Temporal differentiation of wild populations of T. hispidus and climatic parameters

Comparison of morphometric features of T. hispidus shells collected in different years in two geographic regions, i.e., Wrocław and Lubawka, showed significant differences depending on the year of collection. The largest number of differences was revealed in shells from Wrocław (Figs. 1 and 2A; Additional file 2: Table S1). Out of 210 comparisons (15 pairs of collection years × 14 features), 84 were statistically significant (Additional file 2: Table S2), e.g., shell diameter (D) was significantly different in 11 cases, shell height (H) and shell width (W) in 10 cases, body whorl height (bwH), the number of whorls (whl), umbilicus major (U) and minor (u) diameters in 9 cases and aperture height/width ratio (h/w) in 7 cases. Nine features obtained more than 10% difference between shells in at least one comparison of mean values, e.g., U 24%, u 19%, H 16% and D 15% (Additional file 2: Table S2). Umbilicus major (U) and minor (u) diameters showed the largest average percentage difference, i.e., 12% and 10%, respectively, in comparisons of all years.

For snails from Lubawka, out of 84 comparisons (6 pairs of collection years × 14 features) only 8 were statistically significant (Additional file 2: Table S3). The shells differed significantly in their aperture height (h) and width (w) in 3 comparisons. The h feature showed the percentage difference up to 9% (Additional file 2: Table S3) and the largest average difference was 4.5%.

Besides the phenotypic variation, climatic parameters also showed high fluctuations in the studied period (Fig. 2B,C, Additional file 2: Table S4). The maximum difference reported between temperature parameters in some years prior to sample collection in Wrocław was up to 3.7 °C for the maximum winter temperature, while the maximum difference in the relative humidity was up to 11% for autumn. The maximum temperature difference in Jelenia Góra close to Lubawka was up to 3.5 °C for the minimum winter temperature, while the relative humidity differed at most by up to 8% in summer.

Differences in shell morphometry under various climatic conditions

The distinction between shells collected in individual years and changes in climatic parameters along the same period suggest that these differences can be associated with the climate. Therefore, we calculated the average value of a given climatic parameter for each season and studied region and next divided the collected shell data into two groups according to this value. The first group included the shells that developed in conditions above this average and the second below this average (Additional file 2: Table S5). The differences between these groups were statistically significant for 15 out of 16 considered climatic parameters for at least two shell features (Fig. 3). Similarly, each of 14 features significantly separated the groups based on at least two climatic conditions. The results demonstrated that the mean winter temperature substantially influenced nine morphometric shell features, whereas eight characters were changed due to the maximum winter temperature as well as the mean and minimum temperatures in spring, summer and autumn. Umbilicus major (U) and minor (u) diameters as well as umbilicus relative diameter (U/D) were significantly different in 14 pairs of groups characterized by various climatic parameters. In 11 pairs, the height/width ratio (H/W) was significantly different and shell height (H) in 10 pairs.

The umbilicus diameters (u and U) as well as umbilicus relative diameter (U/D) clustered together in the dendrogram based on the mean percentage difference, which indicates that they similarly responded to climatic conditions (Fig. 3). The features u and U revealed the strongest average increase of all features, from 4.1 to 10.5% in shells developed in higher temperatures in all seasons. The largest percentage difference exceeding 10% was recorded for groups separated according to the mean summer and autumn temperatures as well as the maximum summer and minimum autumn temperatures. The U/D ratio was also significantly greater with the mean percentage difference of 2.8–7.6% in shells grown under high temperatures for all seasons and almost all temperature types. On the other hand, the u and U diameters as well as the U/D ratio were on average by 3.7–6.0% significantly smaller in shells developed under higher humidity in summer and winter.

The height/width shell ratio (H/W) was grouped with H and bwH features in the dendrogram and was on average by up to 3.6% significantly smaller in shells grown under higher temperatures in all seasons for almost all types of parameters. The maximum winter temperature caused a significant increase, on average by ca. 3%, in shell height (H) and body whorl height (bwH), whereas higher temperatures in other seasons led to their decrease by up to 3.4% (Fig. 3).

The shells that were grown in autumn with a relatively high maximum temperature were characterized by ca. 3% significantly smaller aperture height (h) and aperture height/width ratio (h/w), which were clustered together in the dendrogram (Fig. 3).

Other four features, shell diameter (D), number of whorls (whl) as well as shell (W) and aperture width (w), formed an additional cluster in the dendrogram (Fig. 3). All of them were on average significantly greater in shells collected one year after winter that was characterized by relatively higher mean and maximum temperatures. The percentage difference was greater, with 3.6–3.9% for W and D.

In the dendrogram, the climatic parameters were clustered in several groups indicating their similar influence on the morphometric features of shells (Fig. 3). There are separate clusters for temperature and humidity parameters with the exception of the autumn maximum temperature and autumn humidity, which are grouped together. The other temperature parameters for warmer seasons are separated from those for winter, which indicates that they differently influenced the shell morphometry.

Correlations between morphometric shell features and climatic parameters

The influence of climatic conditions on the shells collected in individual years was also assessed using Spearman’s correlation coefficient between the morphometric features and climatic parameters (Fig. 4). Of 224 potential relationships 113 were statistically significant. The spring mean temperature was significantly correlated with 10 morphometric features. Summer humidity and six temperature parameters, i.e., the minimum temperatures as well as the spring and winter maximum temperatures, significantly correlated with eight shell features. Minor umbilicus diameter (u) and umbilicus relative diameter (U/D) were significantly correlated with almost all climatic parameters, i.e., 15, umbilicus major diameter (U) and height/width ratio (H/W) with 13 and the ratio of umbilicus minor to its major diameter (u/U) with 11.

As in the case of percentage difference, we can also recognize groups of morphometric features that were similarly correlated with climatic parameters (Fig. 4). Features U/D, U and u were significantly positively correlated with all or almost all temperature parameters for four seasons with the coefficients up to 0.34, 0.30 and 0.36, respectively. On the other hand, the significant correlation coefficients between these features and the humidity in spring, summer and winter were negative and reached − 0.34.

Another group of features included shell height/width ratio (H/W), shell height (H) and body whorl height (bwH) (Fig. 4). All of them showed significant negative correlations with all temperature parameters for spring and summer as well as the minimum autumn temperature, and H/W also with the mean and maximum autumn temperatures as well as the mean and minimum winter temperatures. The correlation coefficients reached − 0.28, − 0.27 and − 0.28, respectively. These three features significantly correlated with summer and spring humidity, at up to 0.23.

The number of whorls (whl), shell width (W), shell diameter (D), demonstrated a similar correlation with climatic parameters (Fig. 4). They showed the largest and significant correlation coefficients with winter temperatures: up to 0.24, 0.22 and 0.22, respectively. The ratio of umbilicus minor to its major diameter (u/U) showed significant positive correlation up to 0.22 with temperature of warmer seasons.

The climatic parameters were grouped into several clusters indicating their similar relationships with morphometric features (Fig. 4). Humidity parameters of warmer seasons formed a separate cluster and temperature parameters were grouped according to seasons. The winter parameters were connected with autumn humidity and separated from temperatures for warmer seasons.

Modelling relationships between morphometric shell features and climatic parameters

The joint influence of many climatic parameters on morphometry of shells collected in individual years was studied using a linear mixed-effects (LME) model after exclusion of correlated parameters and a linear ridge regression (LRR) model including all climatic parameters. The latter allows for the inclusion of correlated variables. We separately investigated the seasonal maximum, mean and minimum temperature parameters in combination with seasonal humidity parameters (Additional file 2: Table S6) because they are obviously correlated.

Umbilicus minor (u) and major (U) diameters as well as umbilicus relative diameter (U/D) turned out best explained by the climatic parameters (with R² > 0.15) in two models (Additional file 2: Table S6). Moreover, u, U and U/D were described in LME models by the largest number of significant climatic parameters, i.e., 15. The features u and U had also the largest number of significant parameters in LRR models, i.e., 18 out of 24 possibilities. The largest average values of temperature coefficients for the LRR models were 0.66 for D, 0.58 for W, 0.32 for H, 0.26 for U and 0.22 for u. Thus, all the above-mentioned features were under the strongest influence of the climatic conditions.

In the case of LRR models, the coefficients at the winter mean temperature were most often selected as significant, in 12 out of 14 possibilities (Additional file 2: Table S6). The humidity coefficients for autumn were significant in 30 cases of 42 possibilities. The highest average absolute values of coefficients in climatic variables were those for the summer (0.63), spring (0.31) and autumn (0.24) minimum temperatures as well as the summer mean temperature (0.31). Thus, the temperatures of warmer seasons were more important for developing shell morphology. Seasonal humidity coefficients showed similar values compared to each other.

Comparison of shell morphometry of T. hispidus and T. sericeus kept under various conditions

In order to verify the influence of different climatic parameters on Trochulus shell morphometry in selected conditions, we compared shells from three groups of T. hispidus, which represented several subsequent generations: (1) parental snails collected in the wild in Wrocław-Jarnołtów, (2) their offspring bred in the laboratory for two generations and (3) offspring of the second laboratory-bred generation transplanted again into a garden in Wrocław (Fig. 5A–C). The comparison of the group 2 and 1 was to verify if laboratory conditions with controlled temperature and humidity can influence the shell morphometry within only one generation, whereas including the group 3 in the comparison, we wanted to check if snails raised in wild garden conditions can recover the original phenotype. Furthermore, we transplanted into the same garden conditions T. sericeus, which was collected in the wild in Muszkowice (Fig. 5D,E). In this case, we verified if two originally different ecophenotypes T. hispidus and T. sericeus, develop the same shell morphometry under the same conditions.

Conditions in which these snails developed were different. According to WorldClim, the wild environment of T. hispidus in Wrocław was generally warmer than that of T. sericeus in Muszkowice (Additional file 2: Table S7). The largest difference was 1.4 °C for the maximum summer temperature. Relative humidity was lower in Wrocław by up to 2% for warmer seasons but was higher in winter by 1.6%. The difference between the wild and garden localities in Wrocław was much smaller and did not exceed 0.41 °C. The garden conditions were less humid, by up to 2%. However, data from WorldClim are generalized over a longer period and wider regions, so may not well reflect local conditions in the studied places. Actually, the Wrocław site was an open habitat covered with a nettle community like a garden patch, while the Muszkowice site was overgrown by a beech forest, which most likely maintained a higher humidity and a more stable temperature.

Laboratory temperatures were substantially different from those in the field, especially for winter (by 18–19.7 °C) as well as for spring and autumn (by 8.2–12 °C). Laboratory humidity was by up to 4.5% lower compared to winter and 5.9–9.9% higher than in spring and summer.

A discriminant function analysis (DFA) for the defined groups of snails provided their interesting grouping and separation (Fig. 6). The analysis identified three significant discriminant functions (p < 00002). The first two functions explained 63% and 27% of variance, respectively. The first function was best positively correlated with umbilicus relative diameter (U/D, r = 0.82) as well as umbilicus major and minor diameters (U and u, r = 0.63). The second function was best positively correlated with relative height of body whorl (bwH/H, r = 0.43) and negatively with shell height (H, r = − 0.68), shell diameter (D, r = − 0.51), shell width (W, r = − 0.50), aperture width (w, r = − 0.48), body whorl height (bwH, r = − 0.47) and aperture height (h, r = − 0.43).

In the DFA plot (Fig. 6), the wild T. sericeus created a distinct group separate from others. The laboratory-bred T. hispidus also formed a well-defined cluster, which only partially overlapped the other sets. Many laboratory-kept specimens were located far from others in the plot, which indicates their disparate shell morphometry. However, wild T. hispidus as well as garden T. sericeus and T. hispidus were grouped together.

In agreement with DFA, statistical tests showed that the laboratory-bred T. hispidus differed significantly in 9 and 10 morphometric features from the wild parents of T. hispidus and its garden-bred offspring, respectively (Additional file 2: Tables S8 and S9). The average percentage differences calculated from the absolute values of all individual features were in these cases 7.3% and 9.8%, respectively. The wild and garden T. hispidus appeared to be different in 6 features with an average percentage difference of 6.4%. Aperture width (w) and height/width ratio (H/W) increased significantly by 10% and 8% in the laboratory snails in comparison to their wild parents (Fig. 7). Then, after their offspring were raised in the garden, these features decreased significantly by 17% and 9% to the values that were not statistically significantly different from those of the wild parents. The same trend was followed by height of body whorl (bwH) as well as shell height (H), width (W) and diameter (D). The initial increase in these features was 14–5% and the subsequent decrease was 23–13%. The garden-raised T. hispidus had the shell statistically different from the wild forms, but the laboratory snails were still much more distinct in these features among the compared groups. On the other hand, umbilicus relative diameter (U/D) and relative height of body whorl (bwH/H) decreased by 16% and 5% in the laboratory-bred T. hispidus in comparison to their wild parents, and then became larger by 12% and 4% in the garden snails, which were not statistically different form the wild T. hispidus in these features (Fig. 7).

The wild T. sericeus was significantly different from the wild T. hispidus in nine features with an average percentage difference of 28% and differed in six features from its descendants raised in the garden on average by 26% (Additional file 2: Tables S8 and S9). The latter differed only in 4% in four shell features from T. hispidus which lived in the same conditions. U/D, U and u were the most distinctive differences between T. sericeus and T. hispidus from the wild environment (Fig. 7). These features were smaller in the former by 115–107% and were subjected to a substantial average increase by 110–101% in T. sericeus raised in the garden. After this drastic change, the offspring became statistically indistinguishable from T. hispidus living in the same conditions. The same trend was observed in the number of whorls (whl), which was significantly smaller in the wild T. sericeus than in the wild T. hispidus by 6%. After the garden experiment, whl increased significantly by 4% and the shell became similar to that of the wild T. hispidus. In turn, height/width ratio (H/W) was larger in the wild T. sericeus than in T. hispidus by 12%, and dropped by 11% in the garden snails to the value observed in the wild T. hispidus (Fig. 7).

Correlation between climatic parameters and shell morphometry of T. hispidus and T. sericeus kept under various conditions

We noticed significant correlations between the selected morphometric features and the climatic parameters. For example, the maximum summer temperature was positively correlated with umbilicus relative diameter (U/D) with the Spearman’s correlation coefficient of 0.64, while negatively correlated with height/width ratio (H/W), body whorl height (bwH), shell height (H) and aperture width (w). The correlation coefficients were − 0.56, − 0.55, − 0.53 and − 0.47, respectively. The relative humidity in spring and summer showed positive relationships with shell height (H, ρ = 0.65), body whorl height (bwH, ρ = 0.56), aperture width (w, ρ = 0.54), shell height/width ratio (H/W, ρ = 0.47), and aperture height (h, ρ = 0.47). Shell height (H) was also positively correlated with the minimum autumn temperature (H, ρ = 0.47) and negatively with the winter humidity (H, ρ = − 0.47). Relative height of body whorl (bwH/H) demonstrated the highest positive correlations with autumn (bwH/H, ρ = 0.49) and winter humidity (bwH/H, ρ = 0.48), and negative (bwH/H, ρ = − 0.48 and − 0.49) with almost all temperature parameters except the maximum summer temperature.

Source: Ecology - nature.com