Secular Trends in the Size and Shape of the Scapula among the Portuguese between the 19th and the 21st Centuries

Simple Summary The human body experiences long-term changes mostly associated with environmental conditions, and generational trends for stature or age at menarche are well-known and studied. That is not the case regarding potential anatomical modifications with time in the human scapula. Accordingly, this study intended to assess size and shape diachronic changes in two Portuguese-reference skeletal collections. Results show that scapular shape and size variation in females is minimal to nonexistent, while in males, an unambiguous decline was detected with time of scapular size. Higher standards of living, including better nutrition and universal healthcare, are associated with an increase in height but also with a slender body—this general trend is possibly related to the scapular size decline with time in Portuguese males. Abstract Potential secular changes in the human scapula are fundamentally unbeknownst, with most of the preceding anatomical studies focusing on long-term changes in the long bones and the skull. As such, the cardinal purpose of this study pertains to the evaluation of secular trends on the shape and size of the scapula in a time period spanning from the 19th to the early 21st centuries. The study sample included 211 individuals (100 males and 111 females) from the Coimbra Identified Skeletal Collection and the 21st Century Identified Skeletal Collection. The size and shape of the scapula were evaluated using geometric morphometrics. Results show secular changes over a relatively short period of time in both the shape and size of the scapula in Portuguese nationals. Shape changes were observed in both sexes but expressed minimally, while a significant negative trend in the size of the scapula was detected in males. Scapular size decrement in males conceivably echoes general trends of the overall anatomy towards a narrower body associated with higher standards of living that include enhanced nutrition and universal healthcare, among other factors.


Introduction
Secular trends refer to biological changes that take place over decades or generations [1]. The best-documented secular trends are the increment in stature and the anticipation of age at menarche [2]. Long-term changes can be precipitated by the restraint of factors that inhibit growth [3]. These are attributed to the improvement of living conditions that started in the late 19th and early 20th centuries, especially in the more industrialized Western countries [4], e.g., dietary improvement-with lower or no periods of caloric deficit-better sanitation, reduction of infant mortality rates due to medical advances and public health policies, as well as rising socio-economic status [5][6][7].
Although environmental factors are acquainted as the precipitating factor for secular changes, their mechanism is not fully understood [5,8]. The primary mechanism cited is  Only complete scapulae were used in this study. Scapulae presenting pathologies or gross taphonomic alterations were excluded. The assigned sex at birth (i.e., biological sex), age at death, and years of birth and death for each individual were obtained from the documentary sources of each collection [45][46][47].

Data Collection: Landmarks and Semilandmarks
Data collection procedures are the same as in Maranho et al. [49]. All bones were positioned with the dorsal surface upwards and photographed with a Canon EOS 70D with a Macro lens (50 mm f/2.5) located at a distance of 50 cm. The camera was fitted in a fixed tripod. The position of the scapula was standardized by placing them on an osteometric board with graph paper, with the inferior part of the glenoid fossa and a point on the lateral border touching the vertical surface of the board. The focus of the camera was on a marked spot in the graph paper.
The images were transferred to a workstation in order to place seven homologous landmarks in each scapula. The landmark set was based on previous research by Taylor and Slice [50] and Scholtz et al. [41]. While ignoring both the acromion and the spine, the landmarks mimic the shape of the scapular body and are clearly recognizable (Figure 1a): Biology 2023, 12, x 4 of with a Macro lens (50 mm f/2.5) located at a distance of 50 cm. The camera was fitted in fixed tripod. The position of the scapula was standardized by placing them on an oste metric board with graph paper, with the inferior part of the glenoid fossa and a point o the lateral border touching the vertical surface of the board. The focus of the camera w on a marked spot in the graph paper. The images were transferred to a workstation in order to place seven homologou landmarks in each scapula. The landmark set was based on previous research by Tayl and Slice [50] and Scholtz et al. [41]. While ignoring both the acromion and the spine, th landmarks mimic the shape of the scapular body and are clearly recognizable (Figure 1a Landmark 4: On the most inferior point of the inferior angle. Landmark 5: Point of intersection of the scapular spine and the medial surface. Th spine was followed until the point at which it would reach the medial border, considerin that sometimes it splits and forms a triangular area.
Landmark 6: On the most superior point of the superior angle. Landmark 7: Point of intersection of the scapular spine and the superior border. Th point of intersection is found by following the superior border until it encounters the sca ular spine. Due to individual variation, the scapular spine sometimes does not interse with the superior border. In those cases, the point was recorded on the basis of the scap lar notch.
All data were assembled with the tps software suite. Landmark digitation was accom plished through tpsDig. A scale was established at 1 cm, and the points were digitize from landmarks 1 through 7 in the same order.
Semilandmarks were also collected through tpsDig, using the function "Draw bac ground curve", allowing the drawing of the contour of each scapula, starting on landma 1 and culminating on landmark 7. Subsequently, the function "Resample curve" was im plemented with 40 points. The scale was also fixed to 1 cm (Figure 1b). Landmark 4: On the most inferior point of the inferior angle. Landmark 5: Point of intersection of the scapular spine and the medial surface. The spine was followed until the point at which it would reach the medial border, considering that sometimes it splits and forms a triangular area.
Landmark 6: On the most superior point of the superior angle. Landmark 7: Point of intersection of the scapular spine and the superior border. The point of intersection is found by following the superior border until it encounters the scapular spine. Due to individual variation, the scapular spine sometimes does not intersect with the superior border. In those cases, the point was recorded on the basis of the scapular notch.
All data were assembled with the tps software suite. Landmark digitation was accomplished through tpsDig. A scale was established at 1 cm, and the points were digitized from landmarks 1 through 7 in the same order.
Semilandmarks were also collected through tpsDig, using the function "Draw background curve", allowing the drawing of the contour of each scapula, starting on landmark 1 and culminating on landmark 7. Subsequently, the function "Resample curve" was implemented with 40 points. The scale was also fixed to 1 cm (Figure 1b). Data analysis was executed with MorphoJ [51] and PAST [52]. All GM procedures begin with a Procrustes Superimposition, or General Procrustes Analyses (GPA), that removes size, location, and orientation data to minimize the sum of squared distances between the homologous landmarks [53][54][55]. The resultant Procrustes shape coordinates only comprise shape information [44,56,57]. After the GPA, a Principal Components Analysis (PCA) was performed. The PCA can explore the key features of shape variation in a sample and ordinate the individuals in morphospace [51], allowing the extraction and evaluation of the main patterns of shape variation [55], simplifying and reducing the complexity of the data [44,54,58]. The comparison of the variation within groups with variation between groups was enacted through a Procrustes ANOVA (a permutation-based ANOVA) using the Procrustes coordinates [59]. Lastly, a regression was performed to evaluate the effects of the years of birth and death on scapular size and shape [27].
Intra-observer and interobserver errors were quantified in a previous study [49], with the results suggesting that errors are negligible.

Landmarks Data
A Procrustes ANOVA was used to evaluate the size and shape biological differences between individuals of the two identified skeletal collections. Analyses were performed within the confines of each biological sex. Concerning the females, only significant differences in shape were found (Table 4). Shape differences are accentuated on the lateral and medial surfaces, with an inferior angle more curved and acute in the CEI/XXI individuals. The pattern of shape variation can be explained by the first four Principal Components, or PCs (PC1-34.33%; PC2-24.19%; PC3-13.31%; PC4-11.07%), which accounted for 82.90% of the total shape variation ( Figure 2). PC1 represents an enlargement of the scapular body, observed on both lateral and medial sides, as the length is slightly diminished. PC2 demonstrates a modest length increase on both superior and inferior surfaces and a small straightening of the upper lateral surface. For PC3, an enlargement of the scapular body in all the lateral borders and the upper part of the medial surface can be observed. PC4 shows an enlargement of almost the entire medial border and a slight length increase in the inferior angle. Nonetheless, substantial overlap in shape variability between the individuals of both collections can be observed. Size does not show a statistically significant influence on shape. The year of birth showed a significant effect on shape, being responsible for 2.24% of total shape variation. The same impact was observed for the year of death, influencing 2.52% of shape. The years of birth or death did not significantly influence scapular size (Table 5). Table 4. Procrustes ANOVA results based on landmark data revealing significant shape differences between females from the CISC and CEI/XXI. The Procrustes ANOVA suggests that the CISC and CEI/XXI males have significant biological differences in both size and shape ( Table 6). The scapulae from the CISC males are generally larger. The medial and lateral borders of CEI/XXI individuals are more curved, and the inferior angle is more acute. The first five PCs described 88.79% of total shape variation ( Figure 3). PC1 (31.16%) shows an enlargement of the body of the scapula for both surfaces and a small shortening of the length for the inferior border. PC2 (24.17%) depicts an increase in the length of the scapular body and a marginal enlargement on both lower and upper lateral and medial borders, as well as a slight decrease on upper and lower lateral and medial surfaces. PC3 (13.19%) displays a width decrease for the lateral and upper medial borders; the remaining medial border shows a minimal enlargement and a slight increase in length can also be observed. PC4 (10.58%) is responsible for a small increase in length. Lastly, PC5 (9.63%) shows a small decrease in length and width, observed on the positioning of the inferior surface and lateral border, respectively. Notwithstanding, males of both collections share a major overlap in shape variability ( Figure 3). Size does not statistically influence the shape. The influence of the year of birth on shape was statistically significant, being accountable for 3.11% of shape variation; likewise, the year of death also showed a statistically significant effect on shape, influencing 3.73% of the total shape variation. Furthermore, the year of birth showed a negative association with size, influencing 8.20% of size variation ( Figure 4). The year of death affected 6.29% of size variation (Table 5).  Table 5. Results of the regressions performed for both sexes regarding landmark and semilandmark data. We tested the impact of size on the shape and the influence of the birth and death periods on both shape and size. P-V represents the p-value; only p < 0.05 are considered statistically significant. %P represents the percentage of the influence on total shape or size variation.

Figure 2.
Graphic representation of the shape changes associated with the PCs in females, the starting shape is depicted as a red outline with hollow dots (at the position of the landmarks), while the target shape is shown as a blue outline with solid dots (at the position of the landmarks). PC1 represents an enlargement of the scapular body. PC2 demonstrates a small increase in length and a small straightening of the upper half of the lateral border. PC3 indicates a modest enlargement observed on the lateral and the upper part of the medial borders. PC4 shows an enlargement on the medial border. Table 5. Results of the regressions performed for both sexes regarding landmark and semilandmark data. We tested the impact of size on the shape and the influence of the birth and death periods on both shape and size. P-V represents the p-value; only p < 0.05 are considered statistically significant. %P represents the percentage of the influence on total shape or size variation. The Procrustes ANOVA suggests that the CISC and CEI/XXI males have significant biological differences in both size and shape ( Table 6). The scapulae from the CISC males are generally larger. The medial and lateral borders of CEI/XXI individuals are more curved, and the inferior angle is more acute. The first five PCs described 88.79% of total shape variation ( Figure 3). PC1 (31.16%) shows an enlargement of the body of the scapula for both surfaces and a small shortening of the length for the inferior border. PC2 (24.17%) depicts an increase in the length of the scapular body and a marginal enlargement on both lower and upper lateral and medial borders, as well as a slight decrease on upper and lower lateral and medial surfaces. PC3 (13.19%) displays a width decrease for the lateral and upper medial borders; the remaining medial border shows a minimal enlargement and a slight increase in length can also be observed. PC4 (10.58%) is responsible for a small increase in length. Lastly, PC5 (9.63%) shows a small decrease in length and width, observed on the positioning of the inferior surface and lateral border, respectively. Notwithstanding, males of both collections share a major overlap in shape variability (Figure 3). Size does not statistically influence the shape. The influence of the year of birth on shape was statistically significant, being accountable for 3.11% of shape variation; likewise, the year of death also showed a statistically significant effect on shape, influencing 3.73% of the total shape variation. Furthermore, the year of birth showed a negative association with size, influencing 8.20% of size variation (Figure 4). The year of death affected 6.29% of size variation (Table 5).

Semilandmarks Data
The Procrustes ANOVA showed that female individuals exhibit significant differences in shape ( Table 7). The scapular shape differences were manifest in the curvature of both the lateral and the upper part of the medial surface, being more accentuated in the CISC females. The superior angle was more prominent in CEI/XXI individuals. The patterns of shape variation can be explained by the first four PCs (Figure 5), which contain 82.39% of total shape variation (PC1-39.45%; PC2-22.04%; PC3-13.62%; PC4-7.28%). PC1 shows a shrinkage on the glenoid fossa area, the lower half of the medial border, and the superior surface. In the lateral surface, an increase in the length was noted until the lower part of the border, where it enlarges. In the remaining medial border, there is a small

Semilandmarks Data
The Procrustes ANOVA showed that female individuals exhibit significant differences in shape ( Table 7). The scapular shape differences were manifest in the curvature of both the lateral and the upper part of the medial surface, being more accentuated in the CISC females. The superior angle was more prominent in CEI/XXI individuals. The patterns of shape variation can be explained by the first four PCs (Figure 5), which contain 82.39% of total shape variation (PC1-39.45%; PC2-22.04%; PC3-13.62%; PC4-7.28%). PC1 shows a shrinkage on the glenoid fossa area, the lower half of the medial border, and the superior surface. In the lateral surface, an increase in the length was noted until the lower part of the border, where it enlarges. In the remaining medial border, there is a small increase in length. PC2 also shows a decrease in width for the glenoid fossa area and both lateral and medial borders. In the superior border, an expansion in length was observed. PC3 reveals the reduction of width for the lateral, medial, and superior surfaces. In the glenoid fossa, the upper part of the lateral border, and the inferior surface, an increase in length was observed. Finally, PC4 shows a slight enlargement in the glenoid fossa and the upper area of the adjacent lateral border. Size does not seem to influence the shape. Both the years of birth and death impacted only 1.54% and 1.82% of the total shape variation, respectively (Table 5). PC3 reveals the reduction of width for the lateral, medial, and superior surfaces. In the glenoid fossa, the upper part of the lateral border, and the inferior surface, an increase in length was observed. Finally, PC4 shows a slight enlargement in the glenoid fossa and the upper area of the adjacent lateral border. Size does not seem to influence the shape. Both the years of birth and death impacted only 1.54% and 1.82% of the total shape variation, respectively (Table 5).   Both size and shape were statistically different between the CISC and CEI/XXI males ( Table 8). The CISC individuals presented with larger scapulae. CEI/XXI males showed more accentuated curves in both medial and lateral, as the inferior and superior angles were more prominent. The results of the PCA showed that the first four PCs (PC1-58.06; PC2-17.94%; PC3-7.51%; PC4-4.51%) were responsible for 88.02% of the total pattern of shape variation ( Figure 6). PC1 shows a slight decrease in width on the glenoid fossa and the lower area of the medial surface. On the lateral border, an increase in length can be observed, and in its lower area, an enlargement occurs; the same can happen in the upper medial surface. PC2 represents a decrease in the width of the scapular body, observed in the lateral and medial borders, and an increase in the length of both the superior and inferior surfaces. PC3 shows another decrease in the width of the lateral and superior borders. Finally, PC4 describes an increase in the width of both the glenoid fossa and the upper area of the medial border. The year of birth did not significantly influence shape variation. In contrast, the year of death was responsible for 4.49% of total shape variation ( Table 5). Table 8. Procrustes ANOVA results based on semilandmarks' data reveal significant differences in both size and shape between males from CISC and CEI/XXI. Both size and shape were statistically different between the CISC and CEI/XXI males ( Table 8). The CISC individuals presented with larger scapulae. CEI/XXI males showed more accentuated curves in both medial and lateral, as the inferior and superior angles were more prominent. The results of the PCA showed that the first four PCs (PC1-58.06; PC2-17.94%; PC3-7.51%; PC4-4.51%) were responsible for 88.02% of the total pattern of shape variation ( Figure 6). PC1 shows a slight decrease in width on the glenoid fossa and the lower area of the medial surface. On the lateral border, an increase in length can be observed, and in its lower area, an enlargement occurs; the same can happen in the upper medial surface. PC2 represents a decrease in the width of the scapular body, observed in the lateral and medial borders, and an increase in the length of both the superior and inferior surfaces. PC3 shows another decrease in the width of the lateral and superior borders. Finally, PC4 describes an increase in the width of both the glenoid fossa and the upper area of the medial border. The year of birth did not significantly influence shape variation. In contrast, the year of death was responsible for 4.49% of total shape variation ( Table 5). Table 8. Procrustes ANOVA results based on semilandmarks' data reveal significant differences in both size and shape between males from CISC and CEI/XXI.

Discussion
This study suggests the presence of secular changes over a relatively short period of time in both the shape and size of the scapula in Portuguese nationals-particularly in males-between the late 19th and early 21st centuries. Secular trends refer to variation among individuals within a population that is justified primarily by differences in birth dates [60]. However, in order to assess secular trends and changing political and socioeconomic circumstances in late-life living conditions, an examination of cohorts based on the year of death can also be of significance [61]. In females, the chronological trend is only detected for shape, with marginal changes associated with both the years of birth and death, suggesting a diachronic stability in the anatomy of the female scapula. In male individuals, a negative trend affecting the size of the scapula is highlighted. The observed scapular size decline in males with time opposes the general worldwide positive trend in stature and length of long bones, such as the radius, femur, and tibia, although these trends are neither universal nor show the same rates in all populations [8,18,20,25,29,[62][63][64][65][66]. In fact, it follows the negative trend observed for measurements in the clavicle, humerus, pelvis, and cranial vault [5,8,20,24,31].
The dwindling of environmental growth-inhibiting factors began only during the 19th century when the living conditions of the population were enhanced through sociopolitical and technological advances, including improvements in sanitation, nutrition, and caloric intake, and the development of public health politics that conducted to the decrease of epidemic diseases while reducing infant mortality [8,18]. In Portugal, the scenario was reasonably different, as the effects of improving living conditions were only observable after the 1970s. The 19th century in Portugal was a tumultuous period characterized by the dearth of medical assistance and social policies, poor nutritional habits, inadequate sanitarian infrastructures, and epidemic diseases, where the base of the economy was largely agriculture-dependent [67,68].
It was only at the end of the century and at the beginning of the 20th century that the first public health measures were successful in decelerating epidemic conditions, slowly diminishing the mortality rate. Nonetheless, unquestionable improvements in living conditions only took place during the 1960s, especially after the creation of the Serviço Nacional de Saúde (National Health Service) in the wake of the Carnation Revolution that established a democratic regime in 1974 [62]. Portugal was then the European country with the highest percentage of young people and the lowest proportion of elderly people, the lowest life expectancy at birth, and the highest rates of fertility and infant mortality [62,63]. Currently, the population growth rate is negligible, and life expectancy has risen to average European values. Infant mortality declined, and it is among the lowest in the world (data available at https://www.pordata.pt/tema/europa/populacao-25, accessed on 24 May 2023). These data reflect the overall upgrading of the living conditions of the Portuguese population, dietary betterment, and the application of several health policies, such as the national vaccination plan and universal healthcare system [62,63,69].
Although the general improvement in the living conditions of the Portuguese population began later than in other European countries, a perceived positive increment in height ensued since the beginning of the 20th century [62,63]. In 1904, the mean height for 18-year-old males was higher in the Santarém district when compared to the Coimbra district [62,63,69]. The tendency to a taller but narrower body, the CEI/XXI males, born between 1910 and 1938, may have been affected by this trend, contrary to the CISC men who were previously born.
The improvement and stabilization of environmental and nutritional factors also impacted the age of menarche, which has been decreasing in the last decades [2,3]. Changes in sexual maturation influence skeletal maturation, leading to changes in adult morphology. Long bones became more linear, narrow, and gracile, and the distal long bones have increased more in length than the proximal long bones [8,20]. The clavicle has also been affected, with a decrease in its maximum length [31]. The pelvic morphology has also been influenced, with contemporary females and males presenting a more gracile pelvis when compared to chronologically ancient individuals [24].
Although there is a positive trend in stature, the human skeleton is veering towards a narrower form [8,24,25,31]. Narrower bodies are mostly connected to the rising standards of living and have been identified, for example, through the secular decrease of the average shoulder bi-deltoid breadth in men [70]. Also, the diachronic length reduction in the humerus was only observed in male individuals [20]. The results observed in the size of the scapulae in males seem to mimic these patterns. Hallgrímsson, Willmore, and Hall [71] theorize that different phenotypes can accumulate genetic variation that remains unexpressed if the environmental context remains stable. Once the population experience changes in the environment, the accumulated genetic variation might start to be expressed. Bipedality exempted the human upper limb from locomotive functions, and the recent mechanization of work has reduced most of their exertion of effort and strain, thus, conceivably increasing the susceptibility to secular changes, a trend that has been, in fact, observed [20,29]. The size of the clavicle has also been declining within a diachronic frame. Langley and Cridlin [31] suggest that the changes in the length of the clavicle are associated with an increase in body mass. The increase in the body mass index can be related to reduced physical activity, earlier skeletal maturation, or the intake of more nutritive food with diminished caloric expenditures [8,31].
The scapula articulates with the head of the humerus and the clavicle, forming the shoulder girdle. The actions performed by the shoulder girdle, such as moving the arm, are possible because most of the sixteen muscles associated are inserted on the scapula [72]. It is possible that the robusticity of the muscles of the shoulder declined in association with the declining size of both the clavicle and humerus. This influence, due to biomechanical forces, might be impacting the size of the scapula. Biomechanical forces are known factors for phenotypic alterations; for example, the robusticity of the masseter, temporal, and pterygoid muscles are decreasing, leading to a more gracile mandible [5,73]. These tendencies might vary in the same population; for example, males are more affected than females in this study, as observed by Jantz and Jantz [20] for the humerus and Spradley, Stull, and Hefner [74] for the cranium.
Different results were observed between the Procrustes ANOVA and regressions. The Procrustes ANOVA compared shape data from the sample classified into two pre-defined groups, an artificial classification, while regressions use data derived from the Procrustes coordinates and centroid size. This group pre-definition can exacerbate or obfuscate the differences related to secular changes observed in shape. In fact, there is some overlap-insignificant but nonetheless present-in the birth period of individuals of both collections. Regarding the techniques, there were also differences between landmarks and semilandmarks. This can be explained by their definition and the type of bone analyzed. Landmarks cannot evaluate surfaces or curves because they are not homologous points between the specimens [43]. Semilandmarks also offer valuable shape information contained on curves and surfaces [38,43,44]. The scapula is a flat triangular bone that has a great surface area with minimal points or protuberances [72,75]; therefore, the application of semilandmarks improved the attained shape information.

Conclusions
Secular changes in stature and age at menarche have concentrated most of the interest in the galaxy of anthropometric studies, with a dearth of analyses focused on the human skeleton, particularly in bones such as the scapula. This study benefited from an ample chronological, biological, and social variation expressed in two Portuguese skeletal identified collections to assess secular trends in the size and shape of the scapula through geometric morphometrics. Also, it was possible to evaluate size and shape variation independently. Both the size and shape of the scapula were affected through a moderately short period of time, with a particular evident tendency of scapular size decrement among males that possibly reflects general trends of the overall anatomy towards a narrower body. In females, the scapula presents, to an extent, with a morphological stasis, with a negligible time-variation of shape and none of size.