Meat Science. Auditing animal welfare at slaughter plants. Temple Grandin. Review. Contents lists available at ScienceDirect

Meat Science 86 (2010) 56 65 Contents lists available at ScienceDirect Meat Science journal homepage: www.elsevier.com/locate/meatsci Review Auditing animal welfare at slaughter plants Temple Grandin Department of Animal Sciences, Colorado State University, Fort Collins CO 80523, USA article info abstract Article history: Received 9 January 2010 Received in revised form 8 April 2010 Accepted 15 April 2010 Keywords: Slaughter welfare Cattle Pigs Stunning The OIE Welfare Standards on slaughter transport, and killing of animals for disease control are basic minimum standards that every country should follow. The OIE, European Union, and many private standards used by commercial industry have an emphasis on animal based outcome standards instead of engineering based standards. Numerical scoring is used by both private industry and some governments to access animal welfare at slaughter plants. Five variables are measured. They are: 1) Percentage of animals effectively stunned on the first attempt, 2) Percentage rendered insensible, 3) Percentage that vocalize (bellow, moo, squeal) during handling and stunning, 4) Percentage that fall during handling, and 5) Percentage moved with an electric goad. Each one of these critical control points measures the outcome of many problems. A good animal welfare auditing system also has standards that prohibit really bad practices such as dragging, dropping, throwing, puntilla, and hoisting live animals before ritual slaughter. On farm and transport problems that can be measured at the slaughter plant are: percentage of lame animals, percentage of thin animals, percentage of dirty animals, percentage with sores, bruises or lesions, death losses, morbidity, and percentage of birds with broken wings and legs. 2010 The American Meat Science Association. Published by Elsevier Ltd. All rights reserved. Contents 1. Introduction............................................................... 57 2. Different types of standards........................................................ 57 2.1. Animal based outcome standards.................................................. 57 2.2. Prohibited practices........................................................ 57 2.3. Input based engineering standards also called resource based.................................... 57 2.4. Documentation standards and paperwork.............................................. 58 3. Vague wording in standards leads to inconsistent enforcement....................................... 58 4. Numerical scoring system for slaughter plants............................................... 58 4.1. Percentage of livestock or poultry stunned effectively on the first attempt............................... 58 4.1.1. Captive bolt........................................................ 58 4.1.2. Electrical stunning..................................................... 58 4.1.3. Controlled atmosphere stunning (CAS)........................................... 58 4.2. Percentage of livestock or poultry with no signs of return to sensibility................................ 59 4.2.1. Evaluating insensibility in cattle, pigs, sheep, and other mammals.............................. 59 4.2.3. Evaluating insensibility in chickens, turkeys, and other poultry................................ 59 4.3. Percentage of cattle, pigs, sheep, and other livestock that fall during handling............................. 59 4.4. Percentage of cattle, pigs, sheep, and other livestock that are moved with an electric goad....................... 59 4.5. Percentage of cattle and pigs that vocalize (bellow or squeal) during handling before stunning..................... 59 5. Slaughter without stunning (kosher and halal)............................................... 60 5.1. Restrain the animal in a comfortable, upright position........................................ 60 5.2. Use a very sharp knife that is twice the width of the neck...................................... 60 5.3. Score the interval from the cut to loss of sensibility......................................... 60 5.4. Cattle, sheep, and goats must be unconscious before removal from the restrainer........................... 60 6.2. Expensive plant renovations not required in most slaughter plants.................................. 60 7. Video auditing to maintain the improvements............................................... 60 8. Clear comments are essential....................................................... 61 E-mail address: cheryl.miller@colostate.edu. 0309-1740/$ see front matter 2010 The American Meat Science Association. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.meatsci.2010.04.022

T. Grandin / Meat Science 86 (2010) 56 65 57 9. On-farm and transport animal welfare problems that can be measured in the slaughter plant in cattle, pigs, sheep, and poultry........ 61 9.1. Body condition score (BCS).................................................... 61 9.2. Lameness measure........................................................ 61 9.3. Foot, leg, and body lesions caused by poor housing or disease.................................... 61 9.4. Broken wings and legs on chickens and turkeys........................................... 61 9.5. Dead animals on arrival at the slaughter plant (DOAs)........................................ 61 9.6. Dirty soiled livestock and poultry.................................................. 62 9.7. Scoring of bruised carcasses.................................................... 62 10. Non-ambulatory downed animals..................................................... 62 11. Conclusions............................................................... 62 References................................................................... 62 1. Introduction Animal welfare is becoming an increasing concern around the world (Seng & Laporte, 2005). Managers, veterinarians, and scientists need to become more knowledgeable on how to assess and audit animal welfare at the slaughter plant. The World Organization for Animal Health (OIE) now has welfare standards for slaughter, transport, and killing animals for disease control (OIE, 2009a,b,c; Shimshony & Chaudry, 2005). Standards for on farm welfare of beef cattle and meat (broiler) chicken have preliminary drafts. The OIE standards are basic minimum standards that both the developed and developing countries have agreed on. In addition to OIE standards, each country has its own animal welfare laws and standards (Defra, 2010; USDA, 2010; MAF, 1996). A third type of standards are private standards that have been created by either large meat buying customers, livestock producer groups, or scientific societies (Barnett & Hemsworth, 2009; Grandin, 2010a; FASS, 2010, National Pork Board (2008), Soil Association (no date). Some of these are stricter than either legislated standards or OIE standards. Legislative standards and private standards should avoid being in direct conflict with OIE by allowing practices that should not be used according to OIE standards. 2. Different types of standards 2.1. Animal based outcome standards Animal based standards measure conditions that are outcomes of either poor management practices, neglect, abuse of animals, or poorly designed equipment. The use of animal based standards is recommended by many researchers (Main, 2009; Hewson, 2003; Wray, Main, Green & Webster, 2003; Wray, Leeb, Main, Green & Webster, 2007, and Webster, 2005). Some examples of welfare problems that can be measured with outcome standards are the percentage of animals that are emaciated, lame, bruised, have lesions, had missed stuns before slaughter or fell down during handling. All of these conditions are outcomes of many different bad practices or poor conditions. For example, lameness in dairy cows can be associated with different factors such as cubicle (freestall) dimensions, poor body condition, type of bedding, or lack of hoof trimming (Dippel, Dolezal, Breninkmeyer, Brinkman, March, Knierim & Winkler, 2009; Barker, Amory, Wright, Browey & Green, 2007) and lesions on chicken carcasses are related to litter quality (Allain, Mirabito, Arnould, Colas, LeBouquin, Lupo & Michel, 2009). An additional example is high numbers of cattle or pigs falling during handling. This can be caused by either slippery floors or causing animals to become agitated by over use of electric goads (Cockram & Corley, 1991; Grandin, 1998a; Gregory, 2007). Stunning methods that fail to produce insensibility can be caused by several factors such as lack of equipment maintenance, agitated animals that make stunner placement difficult, untrained people or poor design of equipment (Grandin, 1998a; Ewbank, Parker & Mason, 1992). Animal based standards are continuous measures that can be numerically scored. For example, an auditor or inspector records the percentage of animals that fall down, the percentage lame, or the percentage where the stunner fails on the first attempt. One of the first animal based scoring systems for evaluating stunning and handling of cattle and pigs at the slaughter plant was developed by Grandin (1997, 1998a). Systems using numerical scoring of animal handling are also described in Maria, Villarrael and Gebresentbet (2004) and Edge and Barnett (2008). The European Union now has a major emphasis on the use of animal based methods for evaluating animal welfare (European Union Welfare Quality, 2009). In these programs, animals are evaluated for body condition, lameness, lesions, abnormal behavior, and many other measures. The OIE slaughter and transport guidelines also have animal based numerical scoring on the percentage of animals falling and electric goad use. In both OIE (2009a,b) and Grandin (2010a), handling practices need improvement if more than 1% of the animals fall during handling. The Food Safety and Inspection Service of the USDA has now adopted the use of the numerical scoring system that was developed by Grandin (1997, 1998a)(FSIS/USDA, 2009). The use of animal based scoring systems has resulted in great improvements in handling and stunning when it was used by restaurant companies to audit slaughter plants (Grandin, 2005, 2006). This system has been in use for over ten years by large meat buying customers in many different countries (Grandin, 2010b). Each animal is scored as either acceptable or not acceptable. For example, the percentage of animals where stunning failed on the first attempt or the percentage of cattle or pigs falling during handling is tabulated. Draft documents for the OIE on the welfare of beef cattle and broiler chickens also have a heavy emphasis on animal based outcome measures. 2.2. Prohibited practices To insure a minimum level of basic animal welfare some really bad practices such as beating or dragging animals are prohibited in both legislated and private standards (Grandin, 2010a; USDA, 2010, National Pork Board, 2008). For example, OIE slaughter standards state that certain practices should never be used such as dragging, dropping, or throwing animals (OIE, 2009b). The puntilla method of stabbing cattle behind the poll to immobilize them before slaughter or cutting tendons is also not permitted by the OIE. Scientific research clearly shows that the puntilla should not be used (Limon, Gultian & Gregory, 2008). Prohibited practices are a discreet measure because the prohibited practice is either present or not present. It is important to specify specific painful or stressful practices that are prohibited to avoid misinterpretation by different people. Both legislation and industry standards contain specific prohibited practices such as prohibiting dragging of conscious disabled livestock (Grandin, 2010a; USDA, 2010). 2.3. Input based engineering standards also called resource based These are standards that specify exactly how to perform a procedure, specify space requirements or specify design of a piece

58 T. Grandin / Meat Science 86 (2010) 56 65 of equipment. Standards that specify stunner or lairage design, should be avoided because they would limit the development of new innovative methods. Most problems with poorly designed equipment can be detected and measured with animal based outcome measures. For example, if a stun box is poorly designed, it is likely that it may cause animals to fall down or lower the percentage of cattle stunned effectively on the first attempt. However, there are a few critical engineering standards that may be required to insure a minimum level of acceptable welfare. Some examples are minimum amperages for electric stunning, bolt velocities for captive bolt and minimum space requirements for lairage pens (OIE, 2009b; Daly & Whitington, 1989; Gregory, 2007). Research clearly shows that severely overloaded trucks or high ammonia levels in an enclosed building are very detrimental to poultry, pig, and cattle welfare (Tarrant, Kenny & Harrington, 1988; Ritter et al., 2006,2007; Kristensen & Wathes, 2000; Jones, Wathes & Webster, 2005; Kristensen, Burgess, Demmers & Wathes, 2000). 2.4. Documentation standards and paperwork The author has audited slaughter plants in eight different countries for restaurant companies such as McDonald's Corporation. Many cases of falsified paperwork were found. Therefore, the author recommends that welfare problems that can be directly observed should be used as the main criteria for passing a welfare audit. 3. Vague wording in standards leads to inconsistent enforcement Vague wording in either regulations or industry standards can lead to inconsistent enforcement. A standard that states to minimize electric goad use or avoid excessive electric goad use will be interpreted differently by different inspectors and auditors. Inconsistent enforcement is a problem. A survey conducted by the GAO (Government Accountability Office) (2010) indicated that enforcement of humane slaughter regulations was very variable between different Federal meat inspectors. Standards and regulations need to be clearly written. An example of a clearly written standard is: All the pigs must have sufficient space to all lie down at the same time without being on top of each other. 4. Numerical scoring system for slaughter plants The scoring system developed by Grandin (1997, 1998a, 2010a) has five numerically scored animal based outcome standards. It is now used as a private standard by major restaurant companies and it has been incorporated into a USDA directive for meat inspectors (Grandin, 2006; FSIS/USDA, 2009). The five animal based measures are: Percentage of livestock or poultry stunned effectively on the first attempt Percentage of livestock that remains insensible after they are hung on the rail must be 100%. For poultry, measure insensibility after stunning, and must be 100% before scaulding. Percentage of livestock animals that fall during handling. Do not use for poultry. Percentage of cattle and pigs that vocalize (squeal, bellow) during handling and stunning. Do not use for sheep or poultry. Percentage of livestock animals moved with an electric goad. DO not use for poultry. It is a practical standard that can be easily implemented in both large and small beef, pork, and sheep slaughter plants (Grandin, 2000, 2003, 2005). In larger plants score a minimum of 100 cattle, pigs or sheep or 500 birds. 4.1. Percentage of livestock or poultry stunned effectively on the first attempt 4.1.1. Captive bolt Baseline data collected before restaurant companies started using numerical scoring indicated that only 30% of the beef plants could stun 95% of the cattle correctly with one shot from a captive bolt (Grandin, 1997, 1998a). The correct position for captive bolt stunning is in the middle of the forehead (Finnie, 1993; Woods, Shearer & Hill, 2010). The most common cause of poor stunning was failure to maintain the stunner or damp cartridges (Grandin, 1998a; Grandin, 2002). After McDonald's Corporation and other restaurant companies started using numerical scoring, the percentage of plants that were able to captive bolt stun 95% or more of the animals on the first shot rose to over 90% (Grandin, 2005, 2006). To pass a welfare audit, the first shot must instantly induce insensibility in95% of the cattle for an acceptable score and 99% for an excellent score (Grandin, 2010a; FSIS/ USDA, 2009). Audit data collected by both the restaurant companies and the author showed that the 95% level is easily attainable (Grandin, 2000, 2005, 2010b). 4.1.2. Electrical stunning Two measures are used. The first measurement is the percentage of pigs, sheep or cattle where the electrodes are placed in the correct locations on the head. The second measurement is the percentage of pigs or cattle that vocalize (squeal or bellow) immediately after the tongs are applied. Correct positioning is essential to induce instantaneous insensibility by passing the electric current through the brain (Croft, 1952; Anil & McKinstry, 1998; Velarde, Ruiz-de-la-Torre, Stub, Diestre & Manteca, 2000a; and Gregory & Wotton, 1984a). Vocalization occurs if the tong is energized before it is in full contact with the head. To pass an audit, the tong must be placed in the correct position on the head on 99% of the animals (Grandin, 2010a). McKeegan, McIntyre, Demmers, Lowe, Wathes, Broek, Coenen, & Gentle 2007; McKeegan, Abeyesinghe, McLeman, Lowe, Demmers, White, Kranen, vanbemmel, Lalnkhaar & Wathes, 2007. Data collected in nine pork plants showed that all the plants achieved this level (Grandin, 2010b). Vocalization due to premature energizing of the tongs must be in 1% or less of the animals. Plants can easily achieve these standards (Grandin, 2001a, 2003). When water bath stunning is used for poultry, the percentage of birds that are rendered insensible is determined. Birds that emerge from the water bath showing no signs of return to sensibility are scored as effectively stunned. 4.1.3. Controlled atmosphere stunning (CAS) There is a great controversy between different researchers on the best gas mixture for stunning poultry (Raj & Gregory, 1990; Raj, 2006; McKeegan et al., 2007a,b; Coenen, Lankhaar, Lowe & McKeegan, 2009). For pigs, most researchers agree that pigs should be rapidly exposed to 90% CO 2 (Hartung, Nowak, Waldmann & Ellerbrock, 2002; Becerril-Herrera et al., 2009). Ninety percent CO 2 for 120 s was more effective for abolishing corneal reflexes than exposure for 90 s (Hartmann, Siegling-Vlitakis, Wolf, Rindermann & Fries, 2009). Corneal reflexes were present in 6% and 15% of the pigs. Unpublished research by the equipment companies in U.S. indicates that behavioral reactions of chickens and turkeys to CO 2 can be greatly reduced by raising the level of CO 2 very gradually from 0% to 55% (Grandin, 2010c). There are definite species and genetic differences on how pigs and poultry react to CO 2 (Grandin, 1988, 2010b). The author proposes that animal based standards should be developed for direct observation of pigs and poultry during the anesthesia induction phase. This could be done through either a window or by a camera. It is the author's opinion that gas mixtures that cause either pigs or poultry to attempt to jump out of the container should not be used. Vigorous struggling or flapping wings before loss of the ability to stand would also not be acceptable. When welfare standards for gas

T. Grandin / Meat Science 86 (2010) 56 65 59 stunning are being developed, the whole system should be evaluated. This includes both anesthesia induction and the handling system used to move the birds into the stunner. Live shackling of poultry is definitely stressful (Kannen, Heath, Wabeck & Mench, 1997; Bedanova et al., 2007). Gas stunning systems eliminate stressful handling because the birds can remain in the travel containers. It also eliminates the opportunity for poorly supervised employees to abuse live birds. A small percentage of birds gasping or head shaking may be a reasonable trade off so that live shackling of poultry can be eliminated. Another method being researched for poultry is low atmospheric pressure stunning (Battula et al., 2008). There is a need for research to determine the stressfulness of this method. 4.2. Percentage of livestock or poultry with no signs of return to sensibility For both livestock and poultry, there is a zero tolerance for skinning, scaulding, limb removal or other invasive dressing procedure on any animal or bird that is exhibiting signs of return to sensibility (Grandin, 2010a). All livestock and birds must be restunned if return to sensibility occurs. 4.2.1. Evaluating insensibility in cattle, pigs, sheep, and other mammals To prevent a failing score on an audit of 100 animals, all animals should be rendered insensible before they are hoisted to the bleeding rail. For poultry, evaluate after the stunner. For livestock, second application of the stunner before hoisting to the bleed rail is scored as a percentage point off the stunning score. For detailed descriptions of signs of return to sensibility. Refer to Grandin ( 2001a, 2002, 2010a,c), Velarde, Gispert, Faucittano, Manteca & Diestre, 2000b; Velarde, Ruizde-la-Torre, et al., 2000a; Gregory (2007, 2008), and OIE (2009b). For all methods of stunning, to insure that an animal is insensible and stunned effectively, the following signs of return to sensibility must all be absent. They are regular rhythmic breathing (Wenzlawowicz & Von Holleben, 2005), response to needle prick on the nose (Limon, GUitian & Gregory, 2010), spontaneous natural eye blinking like live animals in the lairage, righting reflex, and vocalization (moo and squeal)(grandin, 2010a). Kicking limbs are reflexes and should be ignored because they are due to spinal reflexes. A flaccid, soft extended tong is another indicator of effective stunning (Gregory, 2007). There is a need for more research on the relationship between clinical signs that can be observed in a slaughter plant and laboratory measurements of insensibility. 4.2.3. Evaluating insensibility in chickens, turkeys, and other poultry All birds must be completely insensible before they enter the scaulder. Live birds that may have entered the scaulder can be determined by counting the number of dark red scaulded birds that have no throat cut. Birds that are not bled turn dark red. Birds should be monitored for effective stunning and bleeding before the scaulder. A bird is definitely sensible and not properly stunned if it responds to a comb pinch, has spontaneous eye blinking, is vocalizing (cackling) or has full outstretched flapping wings. 4.3. Percentage of cattle, pigs, sheep, and other livestock that fall during handling Falling down during handling in slaughter plants has been documented as a problem (Grandin, 1998a; Cockram & Corley, 1991). Gregory (2007) reported that 40% of the stun boxes in UK abattoirs had slippery floors. Score falling in both the truck unloading area and in the stunning area. Score a fall if an animal's body touches the floor while it is being moved off a truck, through an alley, or through the stunning area. Stun boxes that are deliberately designed to make animals fall down before stunning is an automatic failed audit (Grandin, 2010a). The OIE (2009a,b),(Grandin, 2010a and FSIS/USDA, 2009) both state that handling practices or flooring needs to be improved if more than 1% of the animals fall handling. Restaurant audit data collected in 26 beef plants and 15 pork plants that have been audited for ten years showed that all plants had 1% or less of the animals falling (Grandin, 2010b). 4.4. Percentage of cattle, pigs, sheep, and other livestock that are moved with an electric goad The OIE standards limit electric goad use to battery-operated units and they should not be used on sheep, horses, or small piglets (OIE, 2009a,b). Electric goads are very stressful for both cattle and pigs (Benjamin et al., 2001; Warner, Ferguson, Cottrell & Knee, 2007). Poor handling shortly before slaughter is detrimental to pork quality and raised lactate levels (Hambrecht, Eissen, Newman, Cerstegen & Hartog, 2005; Hambrecht, Eissen, Newman, Smits, dehertog & Verstegan, 2005). Some animal welfare specialists have proposed banning electric goads, but the author has observed in many slaughter plants that an electric goad is preferable to beating an animal or other abusive treatment to make it move. Score the percentage of animals moved with an electric goad. Score each animal as either moved with an electric goad or not moved with one. Handlers should not be allowed to constantly carry electric goads (Grandin, 2006). The electric goad should only be picked up when an animal refuses to move. Restaurant data from plants that had been audited for ten years showed that 26 beef plants and 15 pork plants could achieve the private standard of 25% or less of the animals moved with an electric goad (Grandin, 2010b). Twenty-two out of 26 beef plants used electric prods and 5% or less of the cattle (Grandin, 2010b). 4.5. Percentage of cattle and pigs that vocalize (bellow or squeal) during handling before stunning Vocalizations are correlated with physiological measures of stress in cattle and pigs (Dunn, 1990; Warriss, Brown & Adams, 1994; Weary, Braithwaite & Fraser, 1998; and White et al., 1995). In cattle vocalizations during painful procedures are correlated with the neuropeptide P that is involved in pain perception Coitzee et al., 2008). Each animal is scored as either a vocalizer (squeal, bellow or moo) or a silent animal. All vocalizations that occur in the stun box or restrainer are counted. For cattle, also count vocalizations that occur when the animal is entering the stun box or restrainer. An acceptable score for pigs is 5% or less of the pigs vocalizing (squealing) in the restrainer stun box or stunning pen (Grandin, 2010a). Do not count vocalizations that occur in the lairage or when animals are at rest and not being moved by a person. Vocalizations in cattle can be accurately counted (Welfare Quality, 2009). For cattle, an acceptable score is 5% or less when a head holder is used and 3% or less when no head holder is used (Grandin, 2010a). Welfare audit data from two restaurant companies indicated that all 26 beef plants achieved this standard (Grandin, 2010b). Do not use vocalization scoring for sheep because sheep do not vocalize during aversive stressful handling procedures. In both mammals and poultry, vocalizations that occur after stunning are an indicator of a fully sensible animal. A vocalizing cow or pig hung on the rail is a serious animal welfare problem. Grandin (1998b) reported that 99% of the cattle vocalizations that occurred in the stunning box and race were associated with obviously aversive events such as electric goads, missed captive bolt stuns, slipping in the stun box, gates slammed on animals or excessive pressure from a restraint device. Simple improvements can greatly reduce the percentage of vocalizing cattle or pigs. Reducing pressure applied by a head restraint device reduced the percentage of cattle vocalizing from 23% to 0% (Grandin, 2001b). Adding a light at a dark restrainer entrance to facilitate cattle movement and reduce electric goad use reduced the percentage of cattle vocalizing from 8% to 0%

60 T. Grandin / Meat Science 86 (2010) 56 65 (Grandin, 2001b). Baseline data collected by Grandin (1997) before restaurant audits started indicated that the two worst plants had 32% and 12% of the cattle vocalizing and the four best plants had 7.5%, 2.6%, 6.6%, and 1%. 5. Slaughter without stunning (kosher and halal) Slaughter without stunning is extremely controversial from an animal welfare standpoint. Research indicates that cutting the neck causes pain. Gibson et al. (2009b, 2009a) states that in calves weighing 109 to 170 kg, a cut made with a short 24.5 cm long knife was painful. Another big animal welfare concern is aspiration of blood into the lungs while the animal is still sensible. In cattle slaughtered without stunning, this varied from 36% to 69% of the cattle (Gregory, von Wenzlawowicz & von Holleben, 2008). The OIE (2009b), the EU, and the USA, all permit slaughter without stunning to allow Jews and Muslims to practice their religious beliefs. To improve animal welfare, some religious authorities will accept stunning either immediately before or immediately after the throat cut. Based on observations in over 50 kosher and halal plants where stunning is not allowed, the author recommends the following practices should be used to improve animal welfare during slaughter without stunning. 5.1. Restrain the animal in a comfortable, upright position For design information on upright restraint equipment, refer to Grandin (1992, 2007, 2009). Both Westervelt, Kinsman, Prince and Giger (1976) and Dunn (1990) found that upright restraint was less stressful compared to shackling and hoisting or inverted restraint. Drawings can be obtained on http://www.grandin.com and http:// www.spiritofhumane.com. 5.2. Use a very sharp knife that is twice the width of the neck The author has observed that knives that are too short where the tip gouges into the neck will often cause violent struggling. The author has observed that allowing the wound to close back over the knife during the cut will also cause struggling. Carefully done kosher slaughter done with the special long knife which prevents gouging by the knife tip caused little or no behavioral reaction from most cattle or chickens (Grandin, 1994; Barnett, Cronin & Scott, 2007). 5.3. Score the interval from the cut to loss of sensibility Cattle take longer to lose sensibility after slaughter without stunning compared to sheep (Baldwin, 1971; Blackmore, 1984). Sheep lose sensibility within an average time of 2 to 14 seconds with good cutting technique (Gregory & Wotton, 1984a; Blackmore, 1984). In cattle, some animals may require over a minute to lose sensibility (Blackmore, 1984; Daly, Kallweit & Ellendorf, 1988; Gregory & Wotton, 1984b; Gregory, Fielding, von Wenzlawowicz & von Hollenben, 2010). Careful technique will shorten the time required for cattle to collapse (lose posture and no longer be able to stand) after the cut. Collapse is the initial indicator of onset of insensibility (Blackmore, 1984; Grandin, 1994; Von Holleben et al., 2010). When poor technique was used, only 68% of the cattle collapsed within 30 seconds and good technique increased the percentage of cattle that collapsed within 30 seconds to over 90% (Grandin, 2010c). Gregory et al. (2010) reported similar results where 88% of the cattle collapsed within 30 seconds. It is possible that loss of sensibility may not be complete after the initial collapse (Gregory et al., 2010). Cattle will lose sensibility more quickly if they enter the restraint box calmly and are cut immediately after the head is restrained with a swift knife stroke (Grandin, 1994). Releasing body and head restraints immediately after the cut also facilitates more rapid collapse (http:// www.grandin.com, 2009). Use numerical scoring for evaluating falling, electric goad use and vocalization. The handling variables should be measured the same way as they are measured for conventional slaughter with stunning. A well managed slaughter without stunning plant can easily attain scores similar to a conventional plant. Measurement of collapse times between the cut and loss of posture and eye roll back should also be used to monitor and improve slaughter without stunning methods. 5.4. Cattle, sheep, and goats must be unconscious before removal from the restrainer The animal must be insensible and unconscious before it is removed from the restraint box or restrainer and hung on the rail. An animal showing signs of sensibility when hung on the rail is a failed audit. This is the same as conventional slaughter. 6.2. Expensive plant renovations not required in most slaughter plants In the U.S., the author was hired to implement animal welfare auditing programs at slaughter plants for McDonald's Corporation, Wendy's International, and Burger King (Grandin, 2006). Most plants did not have to install expensive new facilities to achieve numerical scores that were within the values in Grandin (2010a). Out of 50 beef plants, and 24 pork plants, only two had to install totally new restrainer and race systems (Grandin, 2006). The other plants had to make simple changes in equipment and retrain employees. The most common changes to achieve acceptable scores were: 1. Improve maintenance of captive bolt stunners and store cartridges in a dry location (Grandin, 1998a, 2002). 2. Install non-slip flooring in stunning boxes and stunning pens to prevent slipping and falling. Animals will stand calmly and be easier to stun when slipping is prevented. Non-slip flooring is essential (Cockram & Corley, 1991; Grandin, 2010c). 3. Change lighting to improve animal movement. See Grandin (1996, 2010b) for more information. Animals tend to move towards a lighted area (Van Putten & Elshof, 1978; Grandin, 1982, 1996, and Tanida, Miura, Tanaka & Yosimoto, 1996). Installation of indirect lighting over a race entrance will often improve animal movement (Grandin, 1996, 2010c). The light must never shine directly into the eyes of approaching animals. Moving lamps will often eliminate reflections on shiny metal and wet floors that make animals balk and refuse to move. 4. Install shields for people to stand behind and install solid sides on races to prevent approaching animals from seeing people or moving equipment up ahead (Grandin, 1996; Grandin, 1980; Kilgour, 1971). 5. Remove distractions such as dangling chains, coats on fences, loose plastic, hoses on the floor and other objects that make animals balk and refuse to move (Grandin, 1996, 2001b; OIE, 2009a,b). Changes in flooring surface also retard movement (Hutson, 1980; Kilgour, 1971). 6. Reduce noise both from people yelling and equipment. Yelling and intermittent high pitched noise is stressful to cattle and pigs (Talling, Waran, Wathes & Lines, 1998; Waynert, Stookey, Schartzkopf-Genswein & Waltz, 1999). The mechanical equipment for handling animals is extremely noisy in some plants. Research in commercial plants indicates that if may be 80 to 90 db[ ] (Weeks et al., 2009). 7. Better training and supervision of employees. 7. Video auditing to maintain the improvements The author observed that when McDonald's and other companies first started auditing slaughter plants in 1999, many bad practices were done in front of the auditors because managers did not know

T. Grandin / Meat Science 86 (2010) 56 65 61 that they were wrong. In 2010 the author has observed that some plant employees will act good during an audit and then revert to old, bad practices as soon as the auditor leaves. To prevent this from happening, the Cargill Corporation and five other U.S. beef and pork companies have installed cameras that can be monitored by independent third party auditors outside the plant. At any time either the third party auditor or a person at the corporate office can observe stunning, handling in the leadup race, bleed area, and truck unloading. 8. Clear comments are essential When a non-compliance is observed, auditors and inspectors must write clear comments. This will help determine the best corrective action or enforcement penalty. An example of a vague comment would be bad stunning and an example of a clear, well written comment would be 85% of the cattle were prodded with an electric goad and agitated cattle were difficult to stun because they moved their heads. 9. On-farm and transport animal welfare problems that can be measured in the slaughter plant in cattle, pigs, sheep, and poultry Many on-farm animal welfare problems can also be measured in the slaughter plant where it is easy to observe large numbers of animals or birds. Many problems that are detrimental to animal welfare that occur on the farm or during transport and handling can be scored at the plant. 9.1. Body condition score (BCS) The percentage of thin and emaciated animals can be scored (Ritter, Xuc, Dial, Morrison & Marsh, 1999). This is especially important to measure in old cull breeding animals. An animal is scored as either too thin or acceptable. Scoring charts for body condition scoring would have to be made for each country using pictures of the local livestock. Body condition scoring charts for improved dairy cattle breeds such as American Holsteins can be found in Wildman et al. (1982) and Edmonson, Lean, Weaver, Farver and Webster (1989). 9.2. Lameness measure The percentage of lame animals can be counted (Roeber et al., 2001). For poultry, lameness would have to be evaluated on the farm. Lameness is a serious animal welfare issue for intensively raised broiler chickens and Holstein dairy cows (Dawkins, Donnelly & Jones, 2004; Rutherford et al., 2009; Espejo, Endres & Salter, 2006; Knowles et al., 2008). Lameness in dairy cows definitely causes pain (Rushen, Pombourceq & depaisselle, 2006; Flowers, de Passille, Weary, Sanderson & Rushen, 2007). Hardy local breeds of animals often have less lameness. In Holstein dairy cows, lameness ranges from 0 to 13% in the best dairies and 34.9% to 54.4% in the worst 20% (Wray et al., 2007). The best dairies have less than 5% lame cows (Espejo et al., 2006). Thompson, Munkgaard and Toyerson (2008) reported that interobserver reliability for lameness scoring is high when observers are trained. Both limping animals that can keep up with the group when the group is walking, and animals that are not able to keep up should be classified as lame. Broiler chickens should be able to walk ten paces with an even gait (Dawkins et al., 2004). Videos for training people to score lameness can be found at Zinpro/ ASPX_Main/enUS/species/dairy/lameness.aspx for cattle and (Knowles et al., 2008) for chickens. Poultry and pigs can also be scored for leg abnormalities that can cause lameness. Twisted or crooked legs can be easily scored at the slaughter plant. Illustrations of leg abnormalities in pigs can be found in Grandin (2010c) and National Hog Farmer (no date). 9.3. Foot, leg, and body lesions caused by poor housing or disease At the slaughter plant, it is also easy to inspect the legs and feet of both livestock and poultry. Photographic charts for scoring foot pad lesions and feather condition in poultry can be found on the Laywel (2005) and the European welfare quality publications (European Union Welfare Quality, 2009). Poor litter conditions in the poultry barn can cause foot pad lesions in broiler chickens (Dawkins et al., 2004). Additional measures for poultry are foot pad lesions, hock burn and breast blisters. These lesions are caused by poor litter conditions (Allain et al., 2009). In dairy cattle, the percentage of animals with swollen hocks was highly variable. It ranged from 0% on the best dairies to 10.7% on the worst dairy (Grandin, 2010c). Good management of the stalls in freestall (cubicle) barns and reducing overcrowding is a major factor in preventing hock lesions (Rutherford et al., 2008). In the top third dairies with the lowest percentage of severely swollen hocks, the stalls were cleaned and rebedded more than once a week (Fulwider et al., 2007). A hock was scored as severely swollen if it had a swelling larger than a baseball (7.4 cm diameter) on the worst leg. In pigs, the percentage of sows with shoulder lesions can be tabulated (Ritter et al., 1999). Shoulder lesions are usually more common when sows are housed in gestation stalls, but the author has observed them in group housing systems. 9.4. Broken wings and legs on chickens and turkeys In poultry, the percentage of broiler chickens with either broken or dislocated wings is a sensitive indicator of handling practices during catching and loading into the transport containers. Ten years ago the author first started auditing broken wings in chickens, many plant managers thought it was normal to have 5% to 6% of the birds with broken wings. When auditing of broken wings was implemented, the percentage of broken wings dropped to 1% or less in the best plants (Grandin, 2010c). Broken wings should be scored with the feathers on to avoid confusion with wing breakage caused by the feather picking machines. Score a bird as having a broken wing if a wing hangs straight down when the bird is hanging in the shackles. Dislocated wings should also be classified as broken. 9.5. Dead animals on arrival at the slaughter plant (DOAs) The percentage of dead pigs, chickens, or other animals arriving at a slaughter plant dead can vary greatly between different producers and transporters. A survey of 12,333 truckloads of market pigs indicated that the best farms had significantly fewer dead pigs compared to the worst farms (Fitzgerald, Stalder, Matthews, Schultz- Kaster & Johnson, 2009). Many different conditions can increase death losses such as rough handling, overcrowded transport vehicles or crates and metabolic or genetic problems in animals. Ritter, Ellis, Berry, Curtis et al., 2009 contains an excellent review of death loss statistics in pigs from studies all over the world. On a truck, a large 129 kg market pig needs a minimum of 0.462 m 2 per pig or greater (Ritter et al., 2006, 2007). In large intensive pig farms, internal unpublished data showed that when handlers loading trucks became fatigued after about six hours of work, death losses increased. The PSS porcine stress gene can greatly contribute to death losses. Some pork producers bred for the heterozygous condition because it was associated with leaner pork. Murray and Johnson (1998) found that death losses upon arrival at the plant were 9.2% for homozygous pigs, 0.27% for heterozygous pigs, 0.05% for pigs which were free of the stress gene. In poultry, there are also differences in their susceptibility to sudden death losses.

62 T. Grandin / Meat Science 86 (2010) 56 65 9.6. Dirty soiled livestock and poultry Both livestock and poultry can be assessed for cleanliness. Dirty animals are caused by either poor litter conditions in the poultry barn or muddy feedlots. A simple four point scoring system can be used (University of Minnesota Extension; Munoz et al., 2008). 1. Clean animal or bird. Birds must have completely clean feet and cattle may have soil below the knee. 2. Legs are soiled. 3. Legs and belly/breast is soiled. 4. Legs, belly/breast and sides of body are soiled. 9.7. Scoring of bruised carcasses Each animal or bird should be scored to determine the percentage of animals with bruises. Bruises should be classified by both the location on the body and severity with the Australian Carcass Bruise Scoring System (Wythes, Kaus & Newman, 1985). When a plant has a severe bruising problem, the first step that should be taken is to determine if the bruises are occurring inside the plant or outside the plant. If they are occurring inside the plant, they will occur on animals from many different producers. If they are occurring outside the plant, they can usually be traced to a particular producer or trucker. A major cause of bruising is overloaded trucks (Tarrant et al., 1988). Rough handling of both livestock and poultry is another major cause (Grandin, 1981, 2010b). Horned cattle will have more bruises than cattle with no horns (Shaw, Baxter & Ramsey, 1976). Tipping horns will not reduce bruises (Ramsey, Meischke & Anderson, 1976). The cruel practice of cutting horns on adult animals without anesthetics should be banned. Dehorning older cattle is more stressful than disbudding calves (Stafford & Mellor, 2005). Research clearly shows that anesthetics and analgesics should be used (Stewart et al., 2009; McMeekan et al., 1998). Contrary to popular belief, animals can be bruised up until the time of bleeding. Meischke and Horder (1976) found that stunned cattle could be bruised prior to bleeding. Cattle sourced through auctions had more bruises compared to cattle sourced directly from a farm (Hoffman, Spire, Schwenke & Unruh, 1998). 10. Non-ambulatory downed animals Many factors can contribute to causing cattle, pigs or other animals to become non-ambulatory and not able to walk to the stunning area. Loading animals that are not fit for transport onto a truck is a major cause of downers (Roeber et al., 2001). Overloading trucks is another cause of downer animals. When cattle were packed into a vehicle too tightly, an animal that fell down was not able to get back up because the other cattle closed up the opening (Tarrant et al., 1988). A third factor that can cause animals to become non-ambulatory is overuse of ractopamine in pigs. Ractopamine is fed to pigs to increase lean muscle (Fernandez-Duefias et al., 2008; Carr et al., 2005). The label on Paylean (ractopamine) states may increase the number of injured or fatigued pigs during marketing (Elanco, 2009; Espejo et al., 2006). Marchant-Forde, Lay, Pajor, Richert and Schinckel (2003) and Poletto, Rostagno, Richert and Marchant-Forde (2009) found that ractopamine made pigs more difficult to handle and increased the incidence of hoof cracks. The author has observed that pigs that become non-ambulatory due to ractopamine and pigs that became non-ambulatory due to PSS, behave differently. PSS pigs are usually hot when touched, exhibited open mounted breathing, rapid grunts, and quivering. Pigs that became fatigued and non-ambulatory due to an overuse of ractopamine do none of these behaviors. Instead they lie down and remain quiet. If they are allowed to rest, they will recover and may be able to walk to the stunner. They act like they do not have sufficient strength to move. The author observed two large pork slaughter plants in the U.S. that greatly reduced the number of nonambulatory pigs by limiting ractopamine use and replacing a genetic line of pigs with poor leg conformation and a high incidence of lameness with a different genetic line. A chart for measuring leg conformation can be found in Grandin (2010b) and National Hog Farmer (no date). Handling problems with pigs can be further reduced when producers walk the fattening pens to get the pigs accustomed to people walking through them, before they arrive at the plant. Research clearly shows that producers can train their pigs to be easier to handle (Geverink et al., 1998; Abbott, Hunter, Guise & Penny, 1997; Brown, Toth, Stanton, Lawlis & Widowski, 2006). 11. Conclusions The use of numerical scoring and animal based measurements will improve animal welfare during handling and stunning in slaughter plants. These measurements also provide the advantage of greater consistency between different auditors and inspectors. This method for evaluating animal welfare in slaughter plants is practical and easy to implement because there are only five numerically scored measures that assess effectiveness of stunning, insensibility, vocalization, falling during handling, and electric goad use. At the slaughter plant, livestock and poultry can also be assessed for many conditions that are detrimental to animal welfare that are due to problems during transport or poor conditions on the farm. Some of the conditions that can be assessed at the plant are, poor body condition, lameness, broken wings, death losses, animal cleanliness, serious injurious, and obvious neglected health problems. References Abbott, T. A., Hunter, E. J., Guise, J. H., & Penny, R. H. C. (1997). The effect of experiences of handling on pig's willingness to move. Applied Animal Behavior Science, 54, 371 375. Allain, V., Mirabito, L., Arnould, C., Colas, M., LeBouquin, S., Lupo, C., & Michel, V. (2009). Skin lesions in broiler chickens measured at the slaughter house: Relationships between lesions and between their prevalence and rearing factors. British Poultry Science, 50, 407 417. Anil, M. H., & McKinstry, J. L. (1998). Variation in electrical stunning tong placements and relative consequences in slaughter pigs. The Veterinary Journal, 155, 85 90. Baldwin, B. A. (1971). Anatomical and physiological factors involved in slaughter by cantid section. Humane Killing and Slaughterhouse Techniques, University Federation For Animal Welfare (pp. 34 43). Herts, UK: Potters Bar. Barker, Z. E., Amory, J. R., Wright, J. L., Browey, R. W., & Green, L. E. (2007). Management factors associated with impaired locomotion in dairy cows in England and Wales. Journal of Dairy Science, 90, 3270 3277. Barnett, J. L., Cronin, G. M., & Scott, P. C. (2007). Behavioral responses of poultry during kosher slaughter and their implications for the bird's welfare. The Veterinary Record, 160, 45 49. Barnett, J. L., & Hemsworth, P. H. (2009). Welfare monitoring schemes: Using research to safeguard animals on the farm. Journal of Applied Animal Welfare Science, 12, 114 131. Battula, V., Schilling, M. W., Vizzler-Thaxton, Y., Behrends, J. M., Williams, J. B., & Schmidt, T. B. (2008). The effects of low atmosphere stunning and deboning time on broiler meat breast meat quality. Poultry Science, 87, 1202 1210. Becerril-Herrera, M., Alonso-Spilsbury, M., Lemus-Flores, C., Guerrero-Legarreta, I., Hernandez, A., Ramirez-Necocecha, R., & Mota-Rojas, D. (2009). CO 2 stunning may compromise swine welfare compared to electrical stunning. Meat Science, 81, 233 237. Bedanova, I., Vaslarova, E., Chioupek, P., Pisekova, V., Suchuy, P., Blahova, J., Dobsikova, R., & Vecerek, V. (2007). Stress in broilers resulting from shackling. Poultry Science, 80, 1065 1069. Benjamin, M. E., Gonyou, H. W., Ivers, D. L., Richardson, L. F., Jones, D. J., Wagner, J. R., Seneriz, R., & Anderson, D. V. (2001). Effect of handling method on the incidence of stress response in market swine in a model system.supplement 1. Journal of Animal Science, 7, 279 (abstract). Blackmore, D. K. (1984). Differences in behaviour between sheep and cattle during slaughter. Research in Veterinary Science, 37, 223 226. Brown, J. A., Toth, E. L., Stanton, A. L., Lawlis, P., & Widowski, T. (2006). The effects of different frequencies of human interaction on handling responses of market pigs. Suppl. 1. Journal of Animal Science, 84, 301 (Abstract). Carr, S. N., Ivers, D. J., Anderson, D. B., Jones, D. J., Mowrey, D. H., England, M. B., Killerfer, J., Rincker, P. J., & McKeith, F. K. (2005). The effects of ractopamine hydrochloride on lean carcass yield and pork quality characteristics. Journal of Animal Science, 83, 2886 2893. Cockram, S., & Corley, K. T. (1991). Effect of preslaughter handling on the behavior and blood composition of cattle British Veterinary Journal, 147, 444 454.