Carcass Quality, Meat Quality and Sensory Properties of the Dual-Purpose Chicken Lohmann Dual

Over 40 million day-old layer line cockerels are culled in Germany each year, due to economic reasons, leading to a recently instigated research focus on the potential of dual-purpose breeds as an alternative to conventional poultry husbandry, especially the practice of culling. This paper aims to explore and assess the dual-purpose chicken breed “Lohmann Dual” (LD) performance (n = 30) and sensory characteristics (n = 48). Carcass and meat quality traits are evaluated, and descriptive sensory analysis of breast muscles is conducted. To define the scope of characteristics, a market sample of “Ross” Line (n = 35) is adducted. LD carcasses are characterized by higher leg than breast yield; carcass, breast and leg weights are higher in Ross. LD meat has a lower pH, differs in color, has higher drip and thawing losses, but lower cooking loss. LD breast muscles are firmer as indicated by shear force measurements, which is confirmed through the sensory analysis. Appearance, odor and flavor differ between the lines. Overall, distinguishable differences are found between both breeds. Further research should focus on the marketing aspect of the dual-purpose line, as some characteristics could draw consumers to this product. Animal welfare and ethical concerns should further be considered when considering dual-purpose breeds as a feasible alternative to culling.


Introduction
One-day-old male layers are usually culled directly after hatching by homogenization, as rearing these animals would be economically detrimental, due to their inefficient growth rate [1,2]. Investigations into the rearing of male layers showed that their performance is less attractive for meat production [3][4][5]. According to the animal welfare regulations of the European Union and Germany it is forbidden to cause pain, suffering or harm to animals without a sensible reason. As the culling of millions of chicks, due to economic reasons is not clearly justifiable as a "sensible reason", this subject has become the topic of much criticism by animal welfare organizations, the public, and policy-makers. Nonetheless, attempts to forbid the killing of chicks have not yet been successful, mainly because suitable alternatives are not yet available or economically relevant. Unfortunately,
Chicken carcasses from LD (n = 78, age: 64 days,) were randomly selected after commercial slaughter. The LD birds were electrically stunned in a water bath (9 s, 100-150 mA), exsanguinated by neck cut, scalded (58 • C, 150 to 210 s), eviscerated, and chilled at 4 • C for at least two hours. Whole carcasses of Ross 308 birds (Aviagen UK Ltd., Midlothian, UK) with a mean age of 42 days were acquired from a commercial abattoir. These animals were stunned with CO 2 in two phases (1. phase: 23% O 2 , 27% CO 2 for 1 min; 2. phase: 70% CO 2 , 0% O 2 for 2 min). They were exsanguinated after neck-cutting, scalded (54 • C, 210 s), eviscerated, pre-chilled for one hour at 0 • C and subsequently chilled for further two hours at 0.5 • C.
After chilling, all carcasses were transported at 4 • C for ca. 4 h to the laboratory facilities of the Department of Animal Sciences of the Georg-August-University. The following procedure was equally conducted for both breeds: Twenty-four hours after slaughter (24 h postmortem [p.m.]) the carcasses were weighed and manually dissected. 30 Musculi pectorales superficiales (MPS) were used for analysis of the meat quality traits. Due to different sizes of the breast muscles a total of 48 MPS of LD and 24 MPS of Ross were prepared for sensory evaluation, and any remaining breast muscles were appointed for training panelists (definition of sensory attributes and the evaluation scheme, training of scale usage). The samples for sensory analysis were sealed in an air-evacuated polyethylene bags and frozen at −20 • C until training or evaluation.

Carcass-Characteristics and Meat Quality Traits
MPS (boneless, skinless) and both legs (with bones, skinless) were weighed. The MPS and leg yields in percent were calculated in relation to the carcass weight.
The pH values were determined 24 h p.m. with a pH-meter by insertion of the pH-electrode and a thermometer (Knick, Portamess 913, Berlin, Germany) into the center of the MPS. Beforehand, the pH-meter was calibrated using standardized buffers (pH 4.0, 7.0).
Lightness (L*), redness (a*) and yellowness (b*) values were determined with a chromameter CR-400 (KONICA MINOLTA, Langenhagen, Germany) in triplicate on the ventral non-defected (no discolorations or petechial bruises) surface of the MPS. The aperture size was 8 mm, the illuminant D65 and standard observer angle was 10 • .
To determine drip loss (DL), the intact MPS were hung in separate boxes equipped with a lid and stored at 4 • C between 24  × 100) [18]. All drip loss samples were then vacuum packaged in plastic bags and frozen at −20 • C until thawing/cooking loss and shear force analysis was conducted.
Three months after slaughter the packaged drip loss samples were thawed overnight at 4 • C, which was followed by an acclimation at room temperature for about one hour. The MPS were unpacked, carefully dried with paper and reweighed to determine the percentage thawing loss (TL) [18]. The thawed MPS were packaged again and cooked in a hot water bath (3043, Köttermann, Uetze/Hänigsen, Germany) at 77 • C for approximately 30 min (LD) and 60 min (Ross) to reach a core-temperature between 74 and 75 • C. The core-temperature of each sample was determined by insertion of a temperature electrode into the center of the MPS (testo 926, Lenzkirch, Germany). After cooking (Figure 1), the samples were, unpacked, carefully dried with paper, stored at room temperature and reweighed for the analysis of the percentage of cooking loss (CL) [18]. Three months after slaughter the packaged drip loss samples were thawed overnight at 4 °C, which was followed by an acclimation at room temperature for about one hour. The MPS were unpacked, carefully dried with paper and reweighed to determine the percentage thawing loss (TL) [18]. The thawed MPS were packaged again and cooked in a hot water bath (3043, Köttermann, Uetze/Hänigsen, Germany) at 77 °C for approximately 30 min (LD) and 60 min (Ross) to reach a coretemperature between 74 and 75 °C. The core-temperature of each sample was determined by insertion of a temperature electrode into the center of the MPS (testo 926, Lenzkirch, Germany). After cooking (Figure 1), the samples were, unpacked, carefully dried with paper, stored at room temperature and reweighed for the analysis of the percentage of cooking loss (CL) [18]. For the determination of shear force (SF) and shear energy (SE) values, the samples were measured using a TA.XTplus texture analyser (Stable Micro Systems Ltd., Surrey, UK), with a 5 kg load cell and a Meullenet-Owens Razor Shear Blade (MORS-Blade). The conditions for the shear analysis according to Xiong et al. [19], were as follows: Pre-test-speed 100 mm/min, test speed 600 mm/min, post-test-speed 300 mm/min, trigger type 10 g (contact force), blade penetration depth 20 mm and data acquisition rate 200 PPS. Each intact MPS-sample was penetrated six times perpendicular to muscle fibre orientation. The following results were measured: SF (in N) was fixed as the first turning point in the deformation curve representing the maximum force to cut the first muscle fibres. SE (in N × mm) represented the whole area under the curve from the start to the turning point of the SF-measurement.

Descriptive Sensory Analysis
Descriptive analysis was selected for objective sensory evaluation of the samples [20,21]. Evaluation was performed by ten trained panelists, who were experienced in descriptive sensory profiling of pork and poultry products. The sensory panel was selected according to ISO 8586-1 [22]. Training with chicken samples lasted for 7 weeks (2 days/week, duration: 1.5 to 2 h per session) according to Lawless [21]. The final sensory evaluation was performed over two weeks (2 days/week, duration: 1.5 to 2 h per session). All panelists evaluated samples of both breeds in triplicate in one session. Each sample was evaluated for up to 12 min, followed by a break to neutralize the senses of smell and taste with water and white bread. Sensory training and evaluation were carried out in the Laboratory for Sensory Analysis and Consumer Research at the University of Goettingen equippedaccording to ISO 8589:2010 [23]-with ten individual cubicles, air exchange and standardized lighting. During the sensory training and evaluation phases the panelists were in the cubicles and evaluated each sensory attribute on a 10 cm unstructured line scale with scale endpoints as given in Table 1, but mainly ranging from 0 (weakest perception) to 100 (strongest perception) [24]. The software EyeQuestion (Logic8 BV, Elst, The Netherlands) was used for data acquisition. For the determination of shear force (SF) and shear energy (SE) values, the samples were measured using a TA.XTplus texture analyser (Stable Micro Systems Ltd., Surrey, UK), with a 5 kg load cell and a Meullenet-Owens Razor Shear Blade (MORS-Blade). The conditions for the shear analysis according to Xiong et al. [19], were as follows: Pre-test-speed 100 mm/min, test speed 600 mm/min, post-test-speed 300 mm/min, trigger type 10 g (contact force), blade penetration depth 20 mm and data acquisition rate 200 PPS. Each intact MPS-sample was penetrated six times perpendicular to muscle fibre orientation. The following results were measured: SF (in N) was fixed as the first turning point in the deformation curve representing the maximum force to cut the first muscle fibres. SE (in N × mm) represented the whole area under the curve from the start to the turning point of the SF-measurement.

Descriptive Sensory Analysis
Descriptive analysis was selected for objective sensory evaluation of the samples [20,21]. Evaluation was performed by ten trained panelists, who were experienced in descriptive sensory profiling of pork and poultry products. The sensory panel was selected according to ISO 8586-1 [22]. Training with chicken samples lasted for 7 weeks (2 days/week, duration: 1.5 to 2 h per session) according to Lawless [21]. The final sensory evaluation was performed over two weeks (2 days/week, duration: 1.5 to 2 h per session). All panelists evaluated samples of both breeds in triplicate in one session. Each sample was evaluated for up to 12 min, followed by a break to neutralize the senses of smell and taste with water and white bread. Sensory training and evaluation were carried out in the Laboratory for Sensory Analysis and Consumer Research at the University of Goettingen equipped-according to ISO 8589:2010 [23]-with ten individual cubicles, air exchange and standardized lighting. During the sensory training and evaluation phases the panelists were in the cubicles and evaluated each sensory attribute on a 10 cm unstructured line scale with scale endpoints as given in Table 1, but mainly ranging from 0 (weakest perception) to 100 (strongest perception) [24]. The software EyeQuestion (Logic8 BV, Elst, The Netherlands) was used for data acquisition. The samples for training and objective sensory evaluation were prepared as presented above for cooking loss. Due to organization of training and evaluation sessions, the samples have been stored for 6 weeks to 3 months. Therefore, determination of SF and SE was performed parallel to final sensory evaluation after 3 months. The muscles were cut into pieces (10 mm × 10 mm × 10 mm, Figure 2) and were stored for up to 2 min on a heated plate until evaluation. For sensory analysis the samples were allocated in a randomized order, labeled with a 3-digit random code, and given to each single panelist for evaluation. Due to the clear weight differences of the LD and Ross MPS, either 2-3 (LD) or 5 (Ross) samples were obtained from each muscle. Each panelist evaluated two pieces per sample. The results of 120 assessments per poultry genetic were used for further statistical analysis. Detailed information about the sensory attributes evaluated during the sessions is presented in The samples for training and objective sensory evaluation were prepared as presented above for cooking loss. Due to organization of training and evaluation sessions, the samples have been stored for 6 weeks to 3 months. Therefore, determination of SF and SE was performed parallel to final sensory evaluation after 3 months. The muscles were cut into pieces (10 mm × 10 mm × 10 mm, Figure  2) and were stored for up to 2 min on a heated plate until evaluation. For sensory analysis the samples were allocated in a randomized order, labeled with a 3-digit random code, and given to each single panelist for evaluation. Due to the clear weight differences of the LD and Ross MPS, either 2-3 (LD) or 5 (Ross) samples were obtained from each muscle. Each panelist evaluated two pieces per sample. The results of 120 assessments per poultry genetic were used for further statistical analysis. Detailed information about the sensory attributes evaluated during the sessions is presented in Table 1. The attribute list was permanently available for each panelist during each evaluation session.

Data Analysis
Data analysis was performed with Statistical Analysis System (SAS, version 9.3., 2011, SAS Institute Inc., Cary, NC, USA). After analysis of the carcass-characteristics and meat quality values for normality (Shapiro-Wilks-test) parameters showing normal distribution were analyzed using a 2sample-t-test. Pooled values were used for variance equality and Satterthwaite approximation for unequal variances. The other (non-parametrical) data were analyzed with the Wilcoxon-two-sample test. ANOVA (mixed model) was applied for statistical analysis of the results from the sensory evaluation using the GLM procedure of SAS. The statistical model contains genetic of the animals (Ross; LD) as a fixed effect and additionally the panelist and day of evaluation as random effects. An alpha-level of 0.05 is considered significant in all statistical tests.
For reasons of clarity the mean and standard deviation values of the data (parametrical and nonparametrical) are presented in the tables and the non-parametrical data are marked with a * in the appropriate tables.

Data Analysis
Data analysis was performed with Statistical Analysis System (SAS, version 9.3., 2011, SAS Institute Inc., Cary, NC, USA). After analysis of the carcass-characteristics and meat quality values for normality (Shapiro-Wilks-test) parameters showing normal distribution were analyzed using a 2-sample-t-test. Pooled values were used for variance equality and Satterthwaite approximation for unequal variances. The other (non-parametrical) data were analyzed with the Wilcoxon-two-sample test. ANOVA (mixed model) was applied for statistical analysis of the results from the sensory evaluation using the GLM procedure of SAS. The statistical model contains genetic of the animals (Ross; LD) as a fixed effect and additionally the panelist and day of evaluation as random effects. An alpha-level of 0.05 is considered significant in all statistical tests.
For reasons of clarity the mean and standard deviation values of the data (parametrical and non-parametrical) are presented in the tables and the non-parametrical data are marked with a * in the appropriate tables.

Carcass-Characteristics and Meat Quality Traits
In the present study, the carcass, MPS, and leg weights, as well as the MPS yields of the LD animals were lower (p < 0.0001 for all) compared to the Ross broiler. Only the leg yields of the LD birds were higher (p < 0.0001) than the results of the Ross birds ( Table 2). Performance of LD in this study was similar to recently published data of 63 day-old LD with regard to carcass and leg weights [25]. However, considering a dressing percentage of 67% [26], carcass weights in the present study were clearly higher compared to weight results of 63 day-old LD presented by Habig et al. [27]. The carcass, breast and leg weights, as well as breast and leg yield values of the Ross broiler were mainly in line with other investigations [28,29]. Although some comparisons of meat quality traits for fast-and slow-growing broiler exist, comparable studies are rare, due to differences in chicken breed, age and handling procedure. However, in accordance to the present results, lower carcass and breast weights in combination with higher leg yields were also found for LD birds by Mueller et al. [26] and in general for slow-growing birds in other publications [25,[30][31][32][33]. Murawska and Bochno [5] found increasing leg yields with increasing age in (slower-growing) layer lines, but decreasing leg yields in Ross after three weeks of age. Therefore, a suggestion would be to market animals of the LD genetic as a whole bird. There might be a chance for slow-growing genotypes in Europe, as the Label Rouge program in France markets poultry products [25,33], which are mainly consumed as whole carcasses [34]. On the contrary, Castellini et al. [14] considered industries demand for processed poultry products as a limiting factor for the marketing of whole carcasses, which have mainly been sourced from free-range or organic production systems. With regard to the meat quality traits (Table 3), the MPS of the Ross genetic had higher pH 24 h p.m. values (p = 0.0079) than the LD animals. These results are in agreement with other studies that present lower pH values in the MPS of LD and other slow-growing birds [26,[35][36][37][38]. Possible reasons for these pH differences are higher stress sensitivity accompanied with a higher activity during transport and slaughter (e.g., wing-flapping, struggling) of slow-growing birds [36,38,39].
The Ross L* 24 h p.m. and b* 24 h p.m. values were higher, i.e., lighter and more yellow (p < 0.0001 for both), and the a* 24 h p.m. values lower (p < 0.0001), i.e., less red in comparison to the LD animals (Table 2). Hemoglobin, as well as myoglobin content, and the different redox forms of the myoglobin, are the main determining factors of meat color [40] depending on factors like meat species, age or muscle type [17]. Investigation by Berri et al. [41] also showed higher L* and lower a* values in selected birds with higher muscle mass, as is the case in this study, where Ross animals have larger MPS compared to the LD animals. This result was also supported by the investigations of Zhao et al. [42]. However, the age difference between Ross and LD should be considered, as the lower L* and higher a* might also be related to the higher age, due to higher myoglobin content in the MPS of the LD animals. Again, this assumption was supported by several reports [41][42][43] and in agreement with color values for the older (LD) compared to younger (Ross) birds. DL-and TL-results of the MPS were higher (p < 0.0001) in LD, but cooking losses of this genetic were lower (p = 0.0046) compared to the Ross birds (Table 2). These findings agree with other studies that also showed reduced water holding capacities in LD, as well as in other slow-growing birds [26,44,45]. The DL results might be related to the pH values, as meat with a lower pH 24 h p.m. is accompanied by a higher DL [25,36,43,45]. This assumption was supported by Berri et al. [46] who found a negative correlation between ultimate pH and DL (r = −0.41; p ≤ 0.001). Additionally, the higher DL of slow-growing birds might also be caused by the larger surface area-to-volume ratio of the smaller breast fillets [36,45,47] or the age difference between LD and Ross birds [48]. The higher CL in Ross birds is likely related to the longer cooking time, but could also be a compensation of the increased DL and TL values, i.e., the higher DL and TL may have reduced final CL in LD birds.
Values for SF and SE were higher (p < 0.0001) in the LD MPS, which is concordant to findings of Mueller et al. [26], comparing LD and Ross and other slow-growing breeds. Reasons might be the differing liquid losses during storage, thawing and cooking (31.0% total loss for LD; 27.9% total loss for Ross,) and possibly structural differences of the MPS, due to the different slaughter ages. However, data about the impact of the slaughter age on tenderness, respectively shear force values, are inconsistent throughout the literature. Nakamura et al. [49] and Janisch et al. [28] found higher shear values for older compared to younger broiler, whereas Sonaiya et al. [50] and Krischek et al. [51] could not detect age-related differences. Connective tissue mainly influences the evaluation of texture attributes, which were performed by compression [52], like SF in this study. Collagen content varies between age and genetic breed, as higher collagen content was found in broiler chicken until three weeks of age and in layer birds up to five weeks, followed by decreasing contents until 10 and 15 weeks [53]. As LD is a cross between meat and layer lines [11], the influence of age and breed has to be considered. Higher collagen-content, higher cross linkage of the connective tissue and smaller fiber diameter were reported for older and slower growing birds [40,54,55]. In consequence, "more dense" connective tissue with more cross-linked collagen content might by responsible for higher shear force values in dual-purpose breed. However, further investigation is necessary to clarify the differences of the shear force values.

Objective Sensory Characteristics
Six of the 19 sensory attributes analyzed differed significantly (p ≤ 0.05) between the LD and Ross genetics (Table 4). For the flavor, the most distinct difference were the sour taste, which was significantly higher in LD (p = 0.0003) and the sweet taste that was more pronounced in Ross (p = 0.0053) (Figure 3a). Secondly, there were big differences in the perceived mouthfeel and texture, LD samples had firmer (p = 0.0005) and less tender (p < 0.0001) textures and tended to be juicier (p = 0.0501) than the Ross samples (Figure 3b). With regard to the appearance less visible fibers on the surface were detected for LD (p = 0.0032). Regarding the odor profile, small, but significant differences were found: Breast muscles of LD exhibited a stronger stable-like smell (p = 0.0349) and tendentiously a higher intensity of smell (p = 0.054) (Figure 3a). Appearance, texture, juiciness and flavor are important sensory quality attributes of poultry products [40,56,57] and discrimination of taste attributes is more perceptible with increasing age of birds [58]. Values representing ratings on an unstructured 100-point-scale from 0 (the weakest perception) to 100 (the strongest perception) without units.
In general, genetic variation is usually only accountable for minor differences in taste attributes [59] and the discrimination by aroma and flavor can only be performed by a subgroup of panelists, as was stated by Jahan et al. [57]. In addition to the significantly higher 'stable-like' smell, overall smell intensity was objectively rated tendentiously stronger in the slow-growing birds LD. These findings are appropriate to higher perceived intensities of meat odor in birds with increasing age [24,39].
The sourer perception in LD is likely related to the lower pH values [60]. The reason for the 'sweeter' taste of the Ross samples might be also related to this difference in pH, as less sour perception leads to a more pleasant 'sweet' perception. Another reason for the differing sour and sweet taste perception in the present study could be linked to differences in the muscle composition, like peptides and fat [61], as significantly differences of the fat or protein contents in the MPS of Ross and LD birds are likely present [26]. However, this would need to be further evaluated.
The higher fibrousness, meaning more visible fibers on the surface, in the Ross birds may have been caused by larger fiber swelling during cooking [62], as well as higher fiber diameters in this genetic.
The latter assumption is supported by Berri et al. [48], who found larger fiber diameter in fast-growing birds compared to older intermediate and slow-growing chicken and by own unpublished results.
important sensory quality attributes of poultry products [40,56,57] and discrimination of taste attributes is more perceptible with increasing age of birds [58].  In this study, the LD were perceived as firmer and less tender, which was supported by numerous findings and could be also related to the significantly higher shear force and energy results in this genetic in comparison to the Ross. Lower tenderness ratings were stated for older birds [25,39,42]; as more mature tissue, due to higher collagen content, crosslinking [15,40] and thicker perimysium [42] is linked to firmer and less tender ratings. Insoluble cross-linked collagen shrinks with heating and is supposed to cause firmness and toughness, because of a smaller fiber size and greater moisture loss [15]. However; as there have been many factors, such as genotype, diet, age and deboning time, that can influence meat texture [24,63], an adjusted production line for alternative genetics, like LD, may improve texture quality characteristics. In concordance, the present results were furthermore supported by Horsted et al. [16], who found a positive correlation between tenderness and sweet taste (as was the case for the Ross samples).
Allen et al. [64] claimed negative influences of selection for higher ultimate pH and greater processing yields on sensory quality, especially juiciness. Similarly, Ross tended to be less juicy and higher juiciness was perceived in the older birds [50] of the LD genetic.
Given the results of this study, future research in sensory analysis should primarily concentrate on determining the market potential of dual-purpose chicken, primarily at the consumer level. It would be valuable to find out, how the varying sensory properties of different genetic lines agree with consumers preferences, i.e., does a target group for more firmer meat with a distinct odor and flavor profile exist? For instance, firmer meat with a more intense flavor was preferred in Thailand [35]. Furthermore, the influence of information on consumers' acceptance, such as animal welfare throughout the production system, should not be underestimated and could influence consumer preferences.

Conclusions
The dual-purpose breed Lohmann Dual is less efficient in meat production, as lower weight properties were obtained over a longer rearing period. The sensory quality differs from the commercial Ross line; future studies should focus on the market potential of chicken meat with such flavor and texture characteristics. Eventually, sensory attributes of LD are linked to a more "traditional" eating-experience and could fit to consumers demand.
One benefit of the dual-purpose breed is the utilization of all animals either for meat or egg production, with only a few restrictions compared to highly specialized breeds. Large strides can be made in animal welfare with the production of these animals. However, each chicken breed needs an optimized husbandry system, feeding and processing requirements and further adjustments are still necessary, before dual-purpose production constitutes a reliable alternative to the culling of day-old male cockerels.

Conflicts of Interest:
The dual-purpose chicken Lohmann Dual was kindly provided by Lohmann Tierzucht, Cuxhaven. Arrangement of housing by Big Dutchman. Boehringer Ingelheim Vetmedica GmbH assisted with the arrangement of vaccination. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results. There is no other conflict of interest declared by all other authors.