Post by brobear on Apr 20, 2020 5:34:41 GMT -5
Continued: Excavation site and specimen description
The specimen of the cave bear (JST4), the third phalanx, was excavated in Stajnia Cave located in the Kraków-Częstochowa Upland in Poland (50° 36′ 58″ N, 19° 29′ 04″ E, Fig. Fig.1).1). The specimen under study shows a morphometry typical of speleoid bear forms (Fig. (Fig.2).2). It had the greatest length = 39.3 mm and the proximal height = 27.1 mm. The measurements better correspond to the cave than brown bear. Mean and standard deviation of these parameters are respectively 38.3 ± 3.7 and 25.9 ± 2.3 mm for U. spelaeus from Buse di Bernardo (Italy) (Santi et al. 2011) and 36.6 ± 5.5 and 25.1 ± 3.7 mm for U. ingressus from Gamssulzenhöhle (Austria) (Alscher 2013), whereas subfossil Ursus arctos from Austria and France is characterized by smaller dimensions, i.e. 33.8 ± 5.8 and 20.9 ± 6.0 mm, respectively (Alscher 2013). Therefore, the measurements for the specimen JST4 are closer to the mean values of the cave bear and are much higher than those for the brown bear. Moreover, the phalanx is more massive and not as slender as in the brown bear. It has also a well-developed articular surface with subcircular contour (Bonifay 1966; Torres Pérez-Hidalgo 1989).
The Stajnia Cave is famous for the discovery of the first remains of Neanderthals in Poland (Nowaczewska et al. 2013; Urbanowski et al. 2010). Six stratigraphical complexes named from A to G were distinguished in this cave, spanning the period of more than 100,000 years (Żarski et al. 2017). The youngest complexes A and B represent MIS 1 sediments. The complex C consists of several layers, marked from the top as C6, C7 and C18 corresponding to MIS 2 as well as C19 from the younger MIS 3. The deposits consist of poorly sorted sandy loams containing sharp-edged limestone rubble, dated to the LGM. The cave bear specimen under study was found at the bottom of layer C18 deposited during the Leszno (Brandenburg) Phase. The complex D of about 60 cm average thickness includes three units D1, D2 and D3, dated to the older part of MIS 3. They yielded the remains of Homo neanderthalensis and very rich Late Middle Palaeolithic flint artefacts. The archaeologically sterile complex E is most probably of MIS 4 age, whereas the oldest complexes F and G are dated to MIS 5. Remains of the Late Pleistocene cave bear (U. spelaeus sensu lato) were found in large numbers throughout the whole profile from layers G to C6. A tooth of cave bear from layer D1 was dated to >49,000 years BP. In total, 323 teeth and bones of the bear were found in this cave.
Laboratory analysis of specimen
The bone was cut into two parts and one of them was submitted to radiocarbon dating in Poznań Radiocarbon Laboratory, whereas the other one was used for genetic analyses. DNA extraction was performed in a laboratory dedicated to ancient DNA analyses observing strict contamination precautions. Samples were thoroughly cleaned with sterile toothbrush and bleach, rinsed with ddH2O and pulverized in cryogenic mill (Spex). A portion of the obtained bone powder was used for DNA extraction following the established protocol (Baca et al. 2012). A 309-bp-long fragment of mitochondrial DNA control region was amplified in singleplex PCR reactions with five different primer pairs (Baca et al. 2012). PCR products were pooled and converted into Illumina sequencing library following the protocol proposed by Meyer and Kircher (2010). Two uniquely indexed libraries based on independently amplified PCR products were produced, pooled with other libraries and sequenced on MiSeq platform. Adaptor and quality trimming were performed with Adapter Removal (Lindgreen 2012). PCR primer sequences were trimmed in Mothur (Schloss et al. 2009). The readings were assembled in SeqMan Pro (DNASTAR). Consensus sequences from two replicates were called according to the guidelines proposed by Stiller et al. (2009). To further confirm AMS date obtained in Poznań Radiocarbon Laboratory, the remaining bone powder (ca. 300 mg) was sent to GADAM Centre in Gliwice, Poland, for independent dating.
Analysis of DNA sequence
To verify taxonomic and phylogenetic position of the analysed specimen, we compared it with all 141 sequences of cave bears (U. spealeus, U. ingressus, U. rossicus and U. kudarensis) available in GenBank as well as 490 sequences of brown bear (U. arctos). The number of base differences per site (p distance) between the sequence under study and the others was calculated in MEGA6 (Tamura et al. 2013). The median-joining network (Bandelt et al. 1999) was constructed using the Network 4.6.1 software (fluxus-engineering.com). The MP algorithm was used to resolve reticulations in the final network (Polzin and Daneschmand 2003).
Phylogenetic trees were created by Bayesian method in MrBayes 3.2.3 (Ronquist et al. 2012) and maximum likelihood in morePhyML 1.14 (Criscuolo 2011) based on PhyML 3.0 (Guindon et al. 2010). In the MrBayes analyses, we adopted a mixed model to sample appropriate substitution models across the larger space in the Bayesian MCMC analysis itself, avoiding the need for a priori model testing (Huelsenbeck et al. 2004). Additionally, we applied gamma-distributed rate variation across sites with five discrete rate categories as suggested by jModeltest 2.1 based on Bayesian Information Criterion (BIC) and decision theory (DT) criterion (Darriba et al. 2012; Guindon et al. 2010). Two independent runs starting from random trees, each using eight Markov chains, were applied. The trees were sampled every 100 generations for 20,000,000 generations. In the final analysis, we selected trees from the last 5,500,000 generations that reached the stationary phase and convergence (i.e. the standard deviation of split frequencies stabilized and was lower than the proposed threshold of 0.01). The tree inferred in morePhyML was based on the best-fit substitution model TPM2uf+Γ found in jModeltest 2.1 among 1624 candidate models according to BIC and DT criteria. The best heuristic search algorithm, nearest neighbour interchanges (NNI) and subtree pruning and regrafting (SPR), in morePhyML was applied. The non-parametric bootstrap analysis in PhyML was carried out applying 1000 replicates and assuming the approximate likelihood ratio test (aLRT) based on a Shimodaira-Hasegawa-like procedure in morePhyML.
Estimation of extinction time
To determine the extinction time of the cave bear, we collected 207 dates of its remains. The dates were carefully selected from the set of 513 dates reported in various references, including an excellent and comprehensive review of dating the cave bear remains done by Pacher and Stuart (2009). We discarded the dates based on molecular, uranium series, uranium/thorium, stratigraphy context and strata dating, dates of U. kudarensis remains, without dating error, infinitive dates and dates out of range 47,500 ± 3000 BP after calibration, as well as the dates of remains with unclear affiliation to the cave bear. All the dates were calibrated to the years BP in OxCal v4.2.4 (Bronk Ramsey et al. 2013) using intCal13 atmospheric curve (Reimer et al. 2013). In the assessment of the extinction time, calibrated mean values and standard deviations were used.
The best-fitted distribution to the set of dates was selected adopting the Akaike information criterion (AIC) and Schwarz Bayesian criterion (BIC) based on the maximum likelihood method applying fitdist from library fitdistrplus in R package (R_Core_Team 2015). Besides R package, statistical analyses were also performed in Statistica (StatSoft_Inc. 2011). To define the extinction time, we performed a procedure based on five methods devised by Strauss and Sadler (1989), Solow (1993), Roberts and Solow (2003), Solow and Roberts (2003) and McInerny et al. (2006) and implemented by Rivadeneira et al. (2009). In addition, we applied the newly developed inverse-weighted McInerney Gaussian-resampled (GRIWM) (Bradshaw et al. 2012) and bootstrap-resampled (BRIWM) methods (Saltré et al. 2015). In the last two approaches, we assumed 10,000 iterations and α level 0.05.
The specimen of the cave bear (JST4), the third phalanx, was excavated in Stajnia Cave located in the Kraków-Częstochowa Upland in Poland (50° 36′ 58″ N, 19° 29′ 04″ E, Fig. Fig.1).1). The specimen under study shows a morphometry typical of speleoid bear forms (Fig. (Fig.2).2). It had the greatest length = 39.3 mm and the proximal height = 27.1 mm. The measurements better correspond to the cave than brown bear. Mean and standard deviation of these parameters are respectively 38.3 ± 3.7 and 25.9 ± 2.3 mm for U. spelaeus from Buse di Bernardo (Italy) (Santi et al. 2011) and 36.6 ± 5.5 and 25.1 ± 3.7 mm for U. ingressus from Gamssulzenhöhle (Austria) (Alscher 2013), whereas subfossil Ursus arctos from Austria and France is characterized by smaller dimensions, i.e. 33.8 ± 5.8 and 20.9 ± 6.0 mm, respectively (Alscher 2013). Therefore, the measurements for the specimen JST4 are closer to the mean values of the cave bear and are much higher than those for the brown bear. Moreover, the phalanx is more massive and not as slender as in the brown bear. It has also a well-developed articular surface with subcircular contour (Bonifay 1966; Torres Pérez-Hidalgo 1989).
The Stajnia Cave is famous for the discovery of the first remains of Neanderthals in Poland (Nowaczewska et al. 2013; Urbanowski et al. 2010). Six stratigraphical complexes named from A to G were distinguished in this cave, spanning the period of more than 100,000 years (Żarski et al. 2017). The youngest complexes A and B represent MIS 1 sediments. The complex C consists of several layers, marked from the top as C6, C7 and C18 corresponding to MIS 2 as well as C19 from the younger MIS 3. The deposits consist of poorly sorted sandy loams containing sharp-edged limestone rubble, dated to the LGM. The cave bear specimen under study was found at the bottom of layer C18 deposited during the Leszno (Brandenburg) Phase. The complex D of about 60 cm average thickness includes three units D1, D2 and D3, dated to the older part of MIS 3. They yielded the remains of Homo neanderthalensis and very rich Late Middle Palaeolithic flint artefacts. The archaeologically sterile complex E is most probably of MIS 4 age, whereas the oldest complexes F and G are dated to MIS 5. Remains of the Late Pleistocene cave bear (U. spelaeus sensu lato) were found in large numbers throughout the whole profile from layers G to C6. A tooth of cave bear from layer D1 was dated to >49,000 years BP. In total, 323 teeth and bones of the bear were found in this cave.
Laboratory analysis of specimen
The bone was cut into two parts and one of them was submitted to radiocarbon dating in Poznań Radiocarbon Laboratory, whereas the other one was used for genetic analyses. DNA extraction was performed in a laboratory dedicated to ancient DNA analyses observing strict contamination precautions. Samples were thoroughly cleaned with sterile toothbrush and bleach, rinsed with ddH2O and pulverized in cryogenic mill (Spex). A portion of the obtained bone powder was used for DNA extraction following the established protocol (Baca et al. 2012). A 309-bp-long fragment of mitochondrial DNA control region was amplified in singleplex PCR reactions with five different primer pairs (Baca et al. 2012). PCR products were pooled and converted into Illumina sequencing library following the protocol proposed by Meyer and Kircher (2010). Two uniquely indexed libraries based on independently amplified PCR products were produced, pooled with other libraries and sequenced on MiSeq platform. Adaptor and quality trimming were performed with Adapter Removal (Lindgreen 2012). PCR primer sequences were trimmed in Mothur (Schloss et al. 2009). The readings were assembled in SeqMan Pro (DNASTAR). Consensus sequences from two replicates were called according to the guidelines proposed by Stiller et al. (2009). To further confirm AMS date obtained in Poznań Radiocarbon Laboratory, the remaining bone powder (ca. 300 mg) was sent to GADAM Centre in Gliwice, Poland, for independent dating.
Analysis of DNA sequence
To verify taxonomic and phylogenetic position of the analysed specimen, we compared it with all 141 sequences of cave bears (U. spealeus, U. ingressus, U. rossicus and U. kudarensis) available in GenBank as well as 490 sequences of brown bear (U. arctos). The number of base differences per site (p distance) between the sequence under study and the others was calculated in MEGA6 (Tamura et al. 2013). The median-joining network (Bandelt et al. 1999) was constructed using the Network 4.6.1 software (fluxus-engineering.com). The MP algorithm was used to resolve reticulations in the final network (Polzin and Daneschmand 2003).
Phylogenetic trees were created by Bayesian method in MrBayes 3.2.3 (Ronquist et al. 2012) and maximum likelihood in morePhyML 1.14 (Criscuolo 2011) based on PhyML 3.0 (Guindon et al. 2010). In the MrBayes analyses, we adopted a mixed model to sample appropriate substitution models across the larger space in the Bayesian MCMC analysis itself, avoiding the need for a priori model testing (Huelsenbeck et al. 2004). Additionally, we applied gamma-distributed rate variation across sites with five discrete rate categories as suggested by jModeltest 2.1 based on Bayesian Information Criterion (BIC) and decision theory (DT) criterion (Darriba et al. 2012; Guindon et al. 2010). Two independent runs starting from random trees, each using eight Markov chains, were applied. The trees were sampled every 100 generations for 20,000,000 generations. In the final analysis, we selected trees from the last 5,500,000 generations that reached the stationary phase and convergence (i.e. the standard deviation of split frequencies stabilized and was lower than the proposed threshold of 0.01). The tree inferred in morePhyML was based on the best-fit substitution model TPM2uf+Γ found in jModeltest 2.1 among 1624 candidate models according to BIC and DT criteria. The best heuristic search algorithm, nearest neighbour interchanges (NNI) and subtree pruning and regrafting (SPR), in morePhyML was applied. The non-parametric bootstrap analysis in PhyML was carried out applying 1000 replicates and assuming the approximate likelihood ratio test (aLRT) based on a Shimodaira-Hasegawa-like procedure in morePhyML.
Estimation of extinction time
To determine the extinction time of the cave bear, we collected 207 dates of its remains. The dates were carefully selected from the set of 513 dates reported in various references, including an excellent and comprehensive review of dating the cave bear remains done by Pacher and Stuart (2009). We discarded the dates based on molecular, uranium series, uranium/thorium, stratigraphy context and strata dating, dates of U. kudarensis remains, without dating error, infinitive dates and dates out of range 47,500 ± 3000 BP after calibration, as well as the dates of remains with unclear affiliation to the cave bear. All the dates were calibrated to the years BP in OxCal v4.2.4 (Bronk Ramsey et al. 2013) using intCal13 atmospheric curve (Reimer et al. 2013). In the assessment of the extinction time, calibrated mean values and standard deviations were used.
The best-fitted distribution to the set of dates was selected adopting the Akaike information criterion (AIC) and Schwarz Bayesian criterion (BIC) based on the maximum likelihood method applying fitdist from library fitdistrplus in R package (R_Core_Team 2015). Besides R package, statistical analyses were also performed in Statistica (StatSoft_Inc. 2011). To define the extinction time, we performed a procedure based on five methods devised by Strauss and Sadler (1989), Solow (1993), Roberts and Solow (2003), Solow and Roberts (2003) and McInerny et al. (2006) and implemented by Rivadeneira et al. (2009). In addition, we applied the newly developed inverse-weighted McInerney Gaussian-resampled (GRIWM) (Bradshaw et al. 2012) and bootstrap-resampled (BRIWM) methods (Saltré et al. 2015). In the last two approaches, we assumed 10,000 iterations and α level 0.05.
