Another study conducted in the Chianti area showed that, followin

Another study conducted in the Chianti area showed that, following the expansion of cultivations AZD6738 in longitudinal rows, versus continued maintenance of terraces, erosion increased by 900% during the period 1954–1976, and the annual erosion in the longitudinal vineyards was approximately 230 t/ha (Zanchi and Zanchi, 2006). As a typical example, we chose the area of Lamole, situated in the municipality of Greve in Chianti, in the province of Florence. The area is privately

owned. The geological substrate is characterized by quartzose turbidites (42%), feldspathic (27%) sandstones, with calcite (7%), phyllosilicates (24%) and silty schists, while in the south there are friable yellow and grey marls of Oligocene origin (Agnoletti et al., 2011). For this specific area, where the terracing stone

wall practice has been documented since the nineteenth century (see the detail of Fig. 7, where the year “1868” is carved in the stone), some authors have underlined a loss of approximately 40% of the terracing over the last 50 years due to less regular maintenance of the dry-stone walls (Agnoletti et al., 2011). As of today, 10% of the remaining terraces are affected by secondary successions following the abandonment of farming activities. Beginning in 2003, the restoring of the terraces and the planting of new vineyards follows an avant-garde project that aims at reaching an optimal level of mechanization as well as leaving the typical landscape elements undisturbed. However, a few months after the restoration, Ibrutinib the terraces displayed deformations and slumps that became a critical issue for the Lamole vineyards. Recently, several field surveys have been carried out using a differential GPS (DGPS) with the purpose of mapping all the terrace failure signatures that have occurred since

terraces restoration in 2003, and to better analyze the triggering mechanisms and failures through hydrologic and geotechnical instrumentation analysis. Fig. 8a Ketotifen shows an example of terrace failure surveyed in the Lamole area during the spring 2013. In addition to these evident wall slumps, several minor but significant signatures of likely instabilities and before failure wall deformations have been observed (Fig. 8b and c). The Fig. 8b shows a crack failure signature behind the stone wall, while Fig. 8c shows an evident terrace wall deformation. The research is ongoing, anyway it seems that the main problem is related both to a lack of a suitable drainage system within terraces and to the 2003 incorrect restoration of the walls that reduced the drainage capability of the traditional building technique (a more detailed description and illustrations about this problem are given in Section 3.2).

Third fire generation anomalies also regard a potential shift of

Third fire generation anomalies also regard a potential shift of the lightning-caused fire regime season, generally concentrated in summer, to the spring season. During spring 2012, an extraordinary lightning fire ran over an area of 300 ha in the south-eastern Alps (“Tramonti

fire”, Friuli, 29th March–10th April). Similarly, recent large summer fires ignited by lightning have attracted public attention because of their extent, as for Pictilisib nmr example the “Monte Jovet Fire” in 2013 (Friuli), which lasted almost one month and spread over an area of 1000 ha, with crown fire phases and flames up to 50 m in height ( Table 1). The listed hot-spots and anomalies may indicate the shift towards a new generation of large natural fires as yet undocumented ( Conedera et al., 2006 and Pezzatti et al., 2009). The short historical overview on fire epochs and generations of large fires in the Alps makes it very clear how disturbance by fire has been and still is a prominent agent in shaping Alpine landscapes and habitats, producing a selective

pressure on species life-history traits and related distribution (Ravazzi et al., 2005), particularly since the last Ice Age (Tinner buy Everolimus et al., 2000, Vannière et al., 2011 and Colombaroli et al., 2013). In the subalpine belt, late glacial forest vegetation consisted of mixed stands of Pinus cembra, Betula spp., Pinus sylvestris, Pinus mugo and Larix decidua ( Vescovi et al., 2007). Periods when natural fire events were low in frequency (early Holocene) favoured why P. cembra dominance ( Gobet et al., 2003), while increases in fire activity (fire intervals of 200–300 yrs) favoured P. sylvestris, Picea abies, P. mugo, L. decidua, and Betula spp. ( Ali et al., 2005 and Stähli et al., 2006). However, during the second fire epoch the increased anthropogenic use of fire for land management resulted in a reduction of the tree component and an opening of the landscape. Some signs at landscape scale of the second fire epoch are still visible in several subalpine rangelands, where the timberline is artificially lowered and the combination

of pastoral fires and recurrent grazing maintain a savannah-like open forest structure (Conedera et al., 2007 and Conedera and Krebs, 2010). Relevant examples of cultural landscapes still maintained by periodic burning and grazing are the open wide-standing larch forests (Fig. 6, left) (Gobet et al., 2003, Ali et al., 2005, Schulze et al., 2007, Genries et al., 2009 and Garbarino et al., 2013), as well as the lowland Calluna vulgaris dominated heathlands ( Fig. 6, right) with sparse birches and oaks ( Borghesio, 2009, Ascoli and Bovio, 2010 and Vacchiano et al., 2014b). The third fire epoch has also been contributing to shape Alpine landscapes. Fire use bans and fire suppression have successfully reduced the overall area burnt in several Alpine regions, e.g., Pezzatti et al.

, 2008, Deli et al , 2005 and Tóth et al , 2011) The key feature

, 2008, Deli et al., 2005 and Tóth et al., 2011). The key features of the adult BBB result from a sequence of cell:cell CAL-101 chemical structure interactions during development between the ingrowing vessel sprouts and the associated cells of the NVU (Liebner et al., 2011). When brain microvessels are isolated from adult mammalian brain and brain endothelial cells are cultured from these vessel fragments, they retain many key features of the BBB phe-notype. In 1969, Siakotos and colleagues described for the first

time a method to successfully isolate bovine and human brain endothelial cells (Siakotos et al., 1969). Nearly a decade later, Panula et al. demonstrated the migration of rat brain endothelial cells from isolated capillaries. These cells were able to grow in culture and had strong alkaline phosphatase activity (Panula et al., 1978). Tontsch and Bauer (1989) simplified the culture methods for isolating murine and porcine brain endothelial cells (e.g. avoiding sieving steps, gradient centrifugations) and optimised the culture medium to increase cell yield. They also found that when proliferative factors such as endothelial cell growth supplement (ECGS) and heparin were removed from culture medium, the morphology of cells changed from spindle-shape to cobblestone phenotype. Through a series of experiments, DeBault and Cancilla gave evidence for the influence of

astrocytic factors on BBB phenotype of brain endothelial cells (DeBault and Cancilla, 1980a, DeBault and Cancilla, 1980b and DeBault, 1981). These studies led to the development of co-culture models of the BBB (Joó, Apoptosis inhibitor 1985). We chose to develop a porcine BBB model for several reasons: (1) A single pig brain gives a high yield of cells compared to that from rat or mouse. (2) Porcine brains are relatively easy to obtain as they are

a by-product of the meat industry; there is no need to have animal breeding facilities buy Paclitaxel on site to maintain a continuous supply of brain tissue. (3) Porcine brain endothelial cells (PBECs) generally retain many key features of the BBB following isolation, and the rate of loss of BBB phenotype in culture is less than for rodent or bovine BBB models (Deli et al., 2005), therefore co-culture with astrocytes is not essential to induce functional expression of tight junctions (i.e. high TEER) (Patabendige et al., this issue). (4) The porcine genome, anatomy, physiology and disease progression reflect human biology more closely than many established laboratory animals (Walters et al., 2011). (5) The availability of miniature pigs and novel porcine transgenic disease models make the pig the most suitable animal model to study human disease (Bendixen et al., 2010 and Lunney, 2007). The miniature pig is now a well established ‘large’ mammalian model for pharmacokinetics/toxicology studies (Bode et al., 2010) and is also used for surgical studies to generate organs for xenotransplantation (Vodicka et al., 2005).

5 the average concentration of paired samples from 10 m into the

5 the average concentration of paired samples from 10 m into the wetland (14.8 ± 4.5 mg kg−1). These values were not statistically different (p = 0.31; t = 1.02; df = 90). The average concentrations for PAHs were 2.7 times higher than samples from 1 m into the wetland, compared with 10 m inland (3427 ± 2,072 vs. 1168 ± 305 μg kg−1), and were not statistically significantly different from each other (p = 0.28; t = 1.08; df = 90). The variability from 1 to 10 m is such that there might be a 10-fold difference, either higher or lower, in the concentration

of PAHs and a lesser amount for target alkanes ( Fig. 3). These results are similar to those of Culbertson Doxorubicin nmr et al. (2007a) who demonstrated high spatial heterogeneity over as little as 5 m in the concentrations of oil remaining in salt marshes 20 years after the West compound screening assay Falmouth, MA, oil spill. The concentrations of target alkanes and PAHs measured in June 2011 and September 2011 along a 90 or 100 m transect (Fig. 4) illustrate how this small reduction from 1 to 10 m continues further into the wetland with only a slight attenuation in concentration, if any. It is not surprising that oil would be carried 100 m into the wetland

in light of the multiple high water events between 2010 and the end of 2012 (Fig. 5). The tidal range is nominally around 30–50 cm throughout these estuaries, whereas the three tropical storms and two hurricanes inundated the wetland between 50 and 100 cm water depth. The turbulence of the storms and water height also came with currents that would have mobilized oil in the wetlands, Dichloromethane dehalogenase bays, or even offshore, and brought it into and over the wetland and perhaps out again. There was not a good correlation between the five SCAT categories of shoreline oiling observed

during the active portions of the spill and the quantity of oil at the same sites that were measured in our study (Fig. 6). There was no difference in the average concentration of aromatics at all SCAT sites that were oiled or not. Further, there was no difference in the amount of oil amongst the SCAT categories for the specific site where the sample came from. These results agree with the conclusion of Michel et al. (2013; p. 4) that “these descriptors are not adequate by themselves to develop cleanup strategies and goals for each habitat type or shoreline segment.” The SCAT team assessments are a necessary first-order assessment for many purposes, including near real-time response operations, but these assessments may not be useful for quantifying relationships between dose and response, changes with time, or spatial distribution horizontally and vertically. Field observations comparing oil exposures along the marsh shoreline should consider taking their own site-specific measurements of oil concentration rather than rely on these surveys to define the relative exposure at the plot level (e.g., 1–10 m). The average concentration of target alkanes and PAHs for each sampling trip are in Fig.

The prevalence of clinical symptoms of TMD in an American populat

The prevalence of clinical symptoms of TMD in an American population was about 6 – 12% [9]. However, there is a peak occurrence between 20 and 40 years of age [10]. One part of TMD is the articular disorders (internal derangement) which is a noninflammatory arthropathy and equates changes in the disc-condyle relationship

[11] and [12]. A recent study among 6-8 year old children showed that 35% of these children had at least one clinical sign of TMD. [13] The TMJ also plays a role in posture and body biostatics [14]. T1 mapping of cartilage after http://www.selleckchem.com/erk.html delayed gadolinium diethylenetriaminepentaacetate acid ion (Gd-DTPA)2- enhancement, called delayed Gadolinium-Enhanced Magnetic Resonance Imaging of Cartilage (dGEMRIC), has emerged as a promising biochemical Magnetic Resonance Imaging (MRI) technique for the quantitative evaluation of articular cartilage [15]. The dGEMRIC has been validated as

a clinically useful tool for the relative glycosaminoglycan content of repair issue after various types of chondrocyte transplantation [16]. Furthermore, in combination with T2 mapping a dGEMRIC provided complementary information on a biochemical properties of a cartilage repair tissue [17]. The dGEMRIC index, i.e., the T1 relaxation time following (Gd-DTPA)2- administration (T1(Gd)), is an indirect measure of the glycosaminoglycan (GAG) concentration of cartilage tissue [18], [19] and [20]. At field-strengths of 3 T, the biochemical MRI measurement of smaller joint cartilage, such as the ankle joint or lumbar facets, becomes possible in find more satisfactory image resolution and clinically reasonable measurement time [21], [22] and [23]. Recently, these biochemical techniques were adapted to fibrocartilaginous tissues, such as the menisci [24] and [25], where, similar to the fibrocartilage structure of the TMJ disc, GAGs are less abundant compared to hyaline cartilage [2] and [26]. Recent results showed that T2 mapping

technique enables ultrastructural analysis of the composition of the TMJ disc and is feasible in vivo [24]. Developed heptaminol for hyaline cartilage, dGEMRIC imaging is an important step towards noninvasive compositional cartilage imaging, because it can show the biochemical ultrastructure of healthy and diseased cartilage. Different studies have demonstrated the ability of dGEMRIC to detect changes in cartilage degeneration before morphological changes occur, in early-stage osteoarthritis (OA) [27] and [28]. The dGEMRIC method can also be used for the monitoring of the maturation of repair tissue after different cartilage repair surgeries [25] and [29] and for longitudinal cohort evaluation of cartilage regeneration [30]. To our best knowledge, no dGEMRIC feasibility studies have been done yet on the disc of the TMJ.

It is interesting to note that the length of follow-up trended to

It is interesting to note that the length of follow-up trended toward significance with close/positive-margin Selleckchem ATM/ATR inhibitor patients having longer follow-up than negative-margin patients (63.1 vs. 58.5 months, p = 0.06). This may represent surgeons increasingly attempting to achieve wider surgical margins in patients selected for APBI or a change in patient selection. Despite

these limitations, this analysis represents the largest collection of close/positive-margin APBI patients evaluated to date and supports the recommendation to obtain margins of 2 mm or greater before the adjuvant application of APBI. Good clinical outcomes were seen in patients undergoing APBI regardless of margin status. However, nonsignificant increases in the rates of IBTR were noted in patients with close or positive margins similar to

what is observed with WBI. Statistically significant increases in IBTR were noted for DCIS patients with close margins. Further prospective studies are required to validate these results and define the appropriate margin status for patients undergoing APBI. “
“Penile carcinoma accounts for 0.4–0.6% of all malignant neoplasms among men in Europe [1] and [2]. Its incidence may reach 20% in some Asian, African, and South American countries. Penile cancer is a disease of elderly men MK-2206 in Europe and North America, with a peak incidence in the sixth decade of life (3), although it may affect a younger age group

in developing countries. Most tumors of the penis are squamous cell carcinomas and occur most commonly on the glans, prepuce, and the coronal sulcus. For small lesions, treatment enabling the penis body to be preserved, notably penis brachytherapy (PB) (4), is recommended to improve the quality of life. Surprisingly, sexuality, which is nevertheless an important component of the quality of life in men with cancer, has not been well studied after conservative treatment of penile cancer. By analyzing a previous series of 51 patients treated between 1971 and 1989, we obtained information about the Edoxaban persistence of sexuality and penile erections of patients (5), but we did not have access to information on the impact of PB on all sexual functions and sexual behavior. To answer these questions, we established a database in the Catalan and Occitan Oncology Group, which includes two cancer centers each in France and Spain. We analyzed the oncologic outcome of penile cancer, and conducted a survey by questionnaire on the sexual functions and behavior after PB treatment, in the two French centers.

45 μm Acrodisc

45 μm Acrodisc LGK-974 molecular weight (Pall) filters prior to chromatographic analysis. HPAEC-PAD was performed with a Dionex ICS3000 equipped with a CarboPac PA10 column (4 × 250 mm) coupled with PA10 guard column (4 × 50 mm). Separation of the sugars was performed with a flow rate of 1 ml/min eluting in a gradient from 10 mM NaOH to 18 mM NaOH over 20 min. After washing for 20 min with 100 mM AcONa in 28 mM NaOH, the column was re-equilibrated with 10 mM NaOH for 20 min. The effluent was monitored using an electrochemical detector in the pulse amperometric mode with a gold working electrode and an Ag/AgCl reference electrode. The Dionex standard quadruple-potential waveform for carbohydrates was used. The resulting chromatographic

data were processed Y 27632 using Chromeleon software 6.8. Calibration curves were built for each sugar monomer (0.5–10 μg/ml). The standards were hydrolysed and analysed in the same way as the samples. For GlcNAc, glucosamine (GlcN) was the species detected by HPAEC-PAD after hydrolysis. 1H NMR analysis was performed to estimate the O-acetylation level.

It was also used to confirm the identity of the OAg samples (typical signals of the OAg chain can be detected, confirming the presence of the characteristic sugars) and in particular for calculating the molar ratio of Rha to abequose (Abe) by comparing the integrals of the two peaks corresponding to Rha-H6 and Abe-H6. The dried OAg sample was subsequently solubilized

in deuterium oxide (D2O) and transferred to a 5 mm NMR tube. A first spectrum was collected in D2O and a second one Farnesyltransferase after de-O-acetylation achieved by adding sodium deuteroxide (NaOD) to a final 200 mM concentration and heat treatment (37 °C for 2 h for complete de-O-acetylation). The first 1H NMR spectrum was recorded to ensure the absence of impurities at the same chemical shift of the acetate anion released after de-O-acetylation of the sample that would interfere with the quantification of the O-acetyl content. O-acetylation level was quantified by comparing acetate (released after treatment with NaOD) and Rha-H6 peaks, and expressed as molar % of O-acetyl with respect to OAg chain repeating units (based on Rha present only in the OAg chain at one sugar per repeating unit). NMR experiments were recorded at 25 °C on Varian VNMRS-500 spectrometer, equipped with a Pentaprobe. Acquisition time of 5 s, relaxation delay of 15 s and number of scans of 64 was set for the acquisition of the spectra. For data acquisition and processing VNMRJ ver. 2.2 rev. C and Mestrenova 6.1 (Mestrelab Research) were used respectively. 1-D proton NMR spectra were collected using a standard one-pulse experiment. Chemical shifts were referenced to hydrogen deuterium oxide (HDO) at 4.79 ppm (1 H). OAg samples were analysed by HPLC-SEC after derivatisation with semicarbazide to quantify α-ketoacid present at the terminus KDO.

The measures presented above do not satisfy this condition and th

The measures presented above do not satisfy this condition and thus must be modified. The percentage-measures defining a percentage of still-at-sea were changed to percentage of no-longer-at-sea. The time-measures were changed to one over the measure. To avoid

to large values, a lower limit of 1 hour was used in this study. The measures were not normalized because normalization Erastin solubility dmso would not affect the routes. However, the actual value of the measure is important for weighting when the measure is not the only term in the target function. The route was found by a very simplistic approach. The route progresses from grid point to grid point of the model grid with a step of at most two. Only eight directions would be possible if only the neighboring grid points were allowed. The number of directions is increased to sixteen by also considering the neighboring grid points one step further. The cost of stepping from one grid point to any of its 16 possible neighboring grid points was calculated in the following way. The direct path between the two grid points passes two or more grid squares. The distance of the

path in each grid square is calculated and multiplied with the value of the measure in the grid square and finally Talazoparib mouse summed up. Dijkstra’s algorithm (Dijkstra, 1959) was used to find a globally optimal path. The mean surface currents for the simulation period are depicted in Fig. 3. In Fig. 4, the investigated measures of this study and the distance to the nearest coast

are shown. The color scale is chosen such that the area of each color is the same in each picture (see figure caption for details) to facilitate comparisons between measures. In the following, higher/lower values refer to higher/lower according to this normalized color scale. The measures form two groups that share many common features. The results of the first group are depicted in Fig. 4a–c, which are the percentage-measures containing information about ID-8 the situation after 30 days. The results of the second group are shown in Fig. 4d–e, which are the time-measures containing information about time scales for reaching the 90% level. In the narrow passage between Öland and Gotland, the mean current is directed toward the open area south of Gotland. The percentage-measures are higher than the time-measures. These time-measures are more similar to the distance to the nearest coast. In the more open area northwest of Gotland, the mean current is directed toward Gotland and the south. In this area, the percentage-measures are lower than the time-measures, which in turn are lower than the distance to the nearest coast. In the western Arkona basin, the mean currents are directed toward The west. Here, all measures are lower than the distance to the nearest coast. There are results that divide the measures in different ways, e.g., in the Bornholm Channel, where the mean current is directed toward the Arkona basin.

These methods were optimized as previously

described with

These methods were optimized as previously

described with some modification [27]. For both methods, each mass spectrum was obtained from the sum of 10 scans of 150 laser shots each and using 512 K data points. Typically, the target plate offset was 100 V with the deflector plate set at 180 V. The ion funnels operated at 100 V and 6.0 V, respectively, with the skimmers at 15 and 5 V. The analyzer entrance was maintained at −7 V, and side kick technology was used to further optimize peak shape and signal intensity. The two acquisition settings differentiate for the trapping potentials (LM, 0.6 and 0.55 V; GSK126 chemical structure HM, 0.95 and 0.80 V), the required excitation power (LM, 25%; HM, 28%) and pulse time (LM, 10 μs; HM, 20 μs), the time of flight to the ICR cell (LM, 1.350 ms; HM, 2.700 ms) and the quadrupole filter mass (LM, m/z 1300; HM, m/z 2500). For each spotted sample, two duplicate spots were measured using the LM and the other two using the HM. Approximately 4.5 h were needed to measure 384 MALDI spots (i.e. originating from 96 different serum samples). DataAnalysis Software 4.0 SP 5 Ceritinib chemical structure (Bruker Daltonics) was used for the visualization and the calibration of the spectra. Prior to the measurement of each MALDI plate the FTICR system was externally

calibrated using a commercially available peptide mix and a protein mix (Bruker Daltonics). The spectra obtained using the LM were internally calibrated only when used for identification purposes. The m/z-values used for the internal calibration of the LM and the HM are reported in Table S1 in the Supplementary Material. Peaks were

determined using the FTMS algorithm with a signal-to-noise threshold of 3 and using the centroid for peak position with a percentage height of 80. Protein and/or peptide signals in RPC18 profiles were quantified as follows. First, based on visual inspection of the profiles, 457 and 670 peaks were selected for the LM and HM spectra, respectively, for further analysis. To this end, a so-called reference file was compiled for both types of profiles in such a way that for each selected peak the m/z-value, Liothyronine Sodium a peak number and an m/z-window were reported. In the LM profiles, this m/z-window ranged from 0.015 to 0.166 Da while in the HM it ranged from 0.05 to 0.31 Da reflecting the peak width along the spectra. Then, the in-house developed Xtractor tool was used to determine the intensity of each user-defined peak. This open source tool generates uniform data (peak) arrays regardless of spectral content (http://www.msutils.org/Xtractor). MALDI-FTICR profiles were exported as XY (.xy) files, all containing m/z values with corresponding intensities. Although peptide and proteins were measured up to 10,000 Da using the HM method, the peak selection was limited to 9043.3 Da. The analysis of the spectra in the m/z-range from 9043.3 to 10,000 is on-going and the results will be presented in a separate study.

Alternatively, Silva et al (2010) found different moisture (45 5

Alternatively, Silva et al. (2010) found different moisture (45.5–51.5 g/100 g), protein (26.9–59.6 g/100 g), and fat contents (36.6–48.2 g/100 g) and pH (5.99–7.13) values in Coalho cheese made from cow’s milk marketed in the Brazilian Northeast. Micelle structures of goat milk differ from cow milk in average Palbociclib diameter, hydration and mineralization. Average mineralization levels of micelles in goat are higher than in cow milk. There is an inverse relationship between the mineralization

of the micelle and its hydration, which also means that goat milk is less hydrated than cow milk (Park, Juárez, Ramos, & Haenlein, 2007) which explains the tenderness of cow cheese. The protein content of CCM and the pH values for CCM, CCGM and CGM significantly differed (P < 0.05) between the 1st and 28th day of storage. The pH values presented no significant differences (P > 0.05) among the different cheeses. Sheehan et al. (2009) observed a decrease in the pH values of semi-hard cheeses manufactured from a mixture of caprine and bovine milk during 150 days of cold storage. According to Sheehan et al. (2009), cow’s

milk presents pH values higher than those of goat’s milk after pasteurization and before the inoculation of the starter culture during the cheeses manufacture, results also observed for our study. The pH values of cheeses made from goat’s milk tend to decrease during the first thirty days of ripening, followed by an increase after this time, while the pH values of cheeses made from cow’s milk tend to decrease during Akt inhibitors in clinical trials the first sixty days of ripening, with a slight increase after this time (Mallatou, Pappas, & Voutsinas, 1994). Goat’s milk also presented a more pronounced alkalinity and buffering capacity in comparison to cow’s Calpain milk, which is mainly related to the associated casein and phosphate systems (Galina, Osnaya,

Cuchillo, & Haenlein, 2007). Low pH values make calcium phosphate micelles more soluble increasing the loss of soluble calcium of whey during the draining of curdled milk (Park, 2006). Pappa et al. (2006) found a decrease in the protein content of ripened cheeses during storage regardless of the kind of milk (goat’s, ewe’s and cow’s) used in their production. Changes in the protein content of cheeses during storage have been related to protein hydrolysis and the production of water-soluble nitrogen compounds, which are released in the brine (Pintado et al., 2008). The moisture, salt and pH values of cheeses are related to the time of ripening because ripened cheeses present lower moisture, greater hardness, higher acidity and higher salt content than unripened cheeses (Freitas & Malcata, 2000). However since our cheeses were only slightly ripened few significant variations of such parameters were observed throughout storage time.