• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Secondary reinforcement effects of combining food with praise in dogs

ProQuest Dissertations and Theses, 2009
Dissertation
Author: Mary E Travers
Abstract:
Many researchers use food as a positive reinforcer for specific behavior change. Although food is necessary for survival, there are instances where food intake in both animals and humans can be excessive. It is plausible that overeating is the result of learning to associate food with other reinforcers. The purpose of this study was to investigate whether food would function as a conditioned (i.e. secondary) reinforcer in dogs. Subjects were 40 adult small breed (7-251bs) dogs between the ages of one and eight. The study was a 4 x 2 random assignment design consisting of five phases (ABCA'D). Dogs were randomly assigned to one of four reinforcement groups; food and praise, food only, praise only and a non-contingent control group. The mean difference in amount of food consumed from baseline one to return to baseline showed significant differences ( p = 0.038). Group mean differences in average weight change from baseline one to return to baseline also reached statistical significance (p < 0.01). Dogs reinforced with food plus praise were significantly heavier than dogs reinforced with food alone, praise alone or the non-reinforcement control. Dogs reinforced with food plus praise demonstrated fewer trials to criterion and a larger number of responses to food refusal, as hypothesized; however these differences failed to reach statistical significance.

vii Table of Contents Copyright ii Abstract iii Dedication v Acknowledgements vi Table of Contents vii List of Tables x List of Figures xi Chapter 1: Introduction 1 Operant Learning and Reinforcement 2 Principles of Reinforcement 4 Secondary Reinforcement 7 Use of Food in Humans and Animals 10 Possible Causes of Overeating and Overweight 13 Emotional Eating 19 Treatment of Obesity and Overweight 22 Canine Care and Training 24 Training 26 Statement of the Problem 29 Hypotheses 30 Chapter 2: Method 32 Subjects 32 Subject Demographics 32

viii Design 34 Independent Variable 38 Dependent Variables 38 Procedure 38 Recruitment 39 Phase I: Ad lib free feeding 39 Phase II: Restricted feeding 39 Phase III: Group assignment and training 39 Phase IV: Ad lib feeding 41 Phase V: Manipulation check 41 Equipment 42 General Care 44 Chapter 3: Results 45 Descriptive Statistics 47 Statistical Analysis 52 Chapter 4: Discussion 59 Hypothesis One 59 Hypothesis Two 60 Hypothesis Three 61 Implications of Results 64 Strengths of the Study 66 Limitations of the Study 66 Directions for Further Research 67

ix Conclusion 69 References 70

X List of Tables Table Page Table 1. Subject Demographics for Food Only Group 33 Table 2. Subject Demographics for Praise Only Group 35 Table 3. Subject Demographics for Food and Praise Group 36 Table 4. Subject Demographics for Control Group 37 Table 5. Summery of Experimental Procedure 43 Table 6. Mean Food Consumption and Weight by Subject and Group 1 46 Table 7. Mean Food Consumption and Weight by Subject Group 2 47 Table 8. Mean Food Consumption and Weight by Subject Group 3 48 Table 9 . Mean Food Consumption and Weight by Subject Group 4 49 Table 10. Food Consumption Means and Standard Deviations by Group in Ounces 50 Table 11. Weight Means and Standard Deviations by Group in Pounds 51 Table 12. Total Trials to Food Refusal by Subject and Group 53 Table 13. Total Trials to Criterion by Subject and Group 54 Table 14. Means and Standard Deviations for Trials to Criterion and Trials to Food Refusal by Group While Satiated 55

xi List of Figures Figure Page Figure 1. Illustration of Kong1"1 Dog Toy 42 Figure 2. Boxplot of Trials to Food Refusal by Group 58 Figure 3. Boxplot of Trials to Criterion by Group 62

1 Chapter 1 Introduction Although food is necessary for survival and has been demonstrated to be helpful in training situations, there are instances where food intake in humans is excessive, leading to overweight and obesity. It is possible that overeating is the result of feeling good while eating. This feeling of pleasure while eating may be learned through the pairing of the food with some other reinforcer (verbal or non-verbal praise, affection or acceptance), that reinforces eating behavior. For example, when a child is reinforced for paying attention in class with a munchkin, the teacher is attending to the student and will often make a positive statement. Although the teacher may think she rewarded the child with food, she also rewarded the child with attention and verbal praise, thus creating a learned association between the food and another reinforcer. It is important to know if such an association between food and other reinforcers exists for several reasons. One of these reasons is to understand the role that learning may play in overeating. If this association does exist and leads to overeating, it could be concluded that food reinforcement might have some undesirable effects in both humans and other organisms. The purpose of this study is to investigate the possibility that food can become associated with praise in dogs. When this association is learned, it is hypothesized that the dogs are likely to overeat. Such an association could lead to overeating, or eating more than is necessary to sustain life. In other words, the dog would be eating when it is no longer hungry; therefore it would be likely to gain weight.

2 In order to understand how the association between food and other reinforcers may develop, it is necessary to review several relevant areas of psychology. First, basic principles of learning theory, including conditioning, reinforcement and secondary reinforcement will be reviewed. Second, a review of current uses of food reinforcement will be presented. Third an overview of potential contributing factors and treatment of human overeating will be included. Finally, there will be a brief discussion about canine care and training to familiarize the reader with general development, care and training techniques for dogs. Operant Learning and Reinforcement Learning is the process by which organisms acquire new behavior (Thorndike, 1913). Scientists seek to discover the laws of learning. These laws are developed through experimental investigation of associations between stimuli and subsequent behavior. Associative learning, the process by which one event becomes linked with another through individual experience can be generally divided into two mechanisms: classical conditioning and operant conditioning. Classical conditioning, first described by Pavlov involves the pairing of two stimuli (Pavlov, 1927; 1928). With repeated presentations, two formerly unrelated stimuli become associated with each other. Pavlov found that when a new stimulus was associated with an unconditioned stimulus that biologically produced a behavioral response (a reflex), subsequent presentations of the once novel stimulus elicited the behavioral response. During this process, the aforementioned stimulus becomes the conditioned stimulus and the behavioral response becomes the conditioned response (Miller & Wasserman, 1997).

3 In classical conditioning, also known as respondent conditioning, the two stimuli are presented independently of the organism's behavior. Presentation of the unconditioned stimulus and the conditioned stimulus is determined entirely by the experimenter and is not dependent on a specific response of the organism at the onset of the trial (Miller & Wasserman, 1997). Understanding the mechanisms of classical conditioning allowed researchers a glimpse at how organisms learn. However, classical conditioning is generally limited to involuntary behavior (such as blinking in response to a puff of air, salivation in response to food, or knee jerk reactions) (Pavlov, 1927) whereas learning new patterns of behavior involves reinforcement patterns resulting from actions impacting the environment (Skinner, 1953). Named by B. F. Skinner in 1938, operant conditioning is described as outcome mediated, or consequence- controlled behavior. A behavior is consequence- controlled when it is followed by a desirable result and the behavior becomes more likely to happen again. Whereas classical conditioning is applied to reflexive behaviors, operant conditioning describes the learning of responses that increase or decrease in frequency depending on the effect provided by the environment (Staddon & Cerutti, 2003). Previously conceptualized by Thorndike as instrumental learning (Thorndike, 1927), operant conditioning refers to the acquisition of new behavior or the development of habit. In one example, Thorndike studied trial and error learning in cats. He saw that a cat placed in a puzzle box initially found its way out by accident. The behavior immediately preceding release from the box was reinforced; a process Thorndike called "The Law of Effect." Once the cat has learned to open the puzzle box, the cats released themselves from the box in shorter and shorter time intervals in subsequent trials

4 (Skinner, 1953; Thorndike, 1929). Based on the law of effect, new behavior that produces a desirable result is likely to be repeated. Repetition of the behavior, given that it continues to be reinforced in some regular way, leads to increases in the likelihood of that response or what Hull (1943) called habit strength. The stronger the habit, the more frequently the behavior will be emitted. Thus the principles of operant conditioning describe how behavior is acquired. Operant principles can be used to intentionally teach new behavior or modify existing behavior. Principles of operant conditioning are utilized in the implementation of Applied Behavior Analysis (ABA). ABA is a scientific approach to predict and control behavior of humans and other animals. Through the application of ABA, specific target behavior is identified and reinforced as to increase that behavior (Sloane, 1992). Through the principles of operant conditioning, however, one can effect a systematic change of behavior (Malott, Whaley & Malott, 1997). When one intentionally reinforces a pre determined behavior one in effect "teaches" the organism to emit that behavior. For example, ABA can be implemented to increase a behavior, such as the number of words a student can read without prompting. Through reinforcement, the pairing of antecedent, response and consequence is repeated until a connection is firmly established. Principles of Reinforcement The concept of reinforcement is critical in the area of learning and behavior analysis. Skinner (1953) states that by definition, a reinforcer is something that reinforces (increases the likelihood of a behavior) and conversely, a punisher punishes (decreases the likelihood of a behavior). Although this may seem like a circular definition, Skinner explains, "There is nothing circular about classifying events in terms

5 of their effects; the criterion is both empirical and objective." (p. 73). Thus a reinforcer is defined as a consequence which increases the probability of the preceding behavior happening again. Another important concept in the area of reinforcement is the distinction between positive reinforcement and negative reinforcement (Skinner, 1953). Both increase the probability of the preceding behavior occurring again. Negative reinforcement involves the removal of a stimulus whereas positive reinforcement is the addition of a stimulus immediately following a behavior. Negative reinforcement is the removal of an aversive stimulus after a response resulting in increasing the likelihood of that response being repeated (Skinner, 1938). For example, most humans find alarm clocks annoying. The end of the noxious stimulus reinforces the behavior of hitting the snooze button or turning off the alarm. Similarly, a parent may give in to a child's tantrum to make it stop. The end of the tantrum makes it more likely that the parent will be more likely to give in to the tantrum in the future (Patterson, 1966). Both positive and negative reinforcement of a behavior increase the likelihood of that behavior being performed in the future. Many stimuli can serve as positive reinforcers. Yet a reinforcer must be salient, or reinforcing to the organism (Skinner, 1938). As stated above, Skinner uses a circular definition of a reinforcer purposefully to demonstrate that a reinforcer can be anything that increases the likelihood of a behavior. A reinforcer may be reinforcing for one individual but not for another. Therefore a reinforcer is by definition anything that when presented after a behavior increases the likelihood that an individual will emit that behavior again. For example, if children are being reinforced for a behavior with

6 peanuts, the child that likes peanuts (if all else is constant) will learn the behavior whereas the child who does not like peanuts will not learn the behavior under this experimental situation. Similarly, the strength of a reinforcer changes depending on temporary circumstances. Saltine crackers would not be as reinforcing to a dehydrated individual as water, but as the individual becomes rehydrated, water loses strength as a reinforcer. Positive reinforcement is the addition of a reinforcing stimulus after the behavior. Positive reinforcers can be either primary or secondary (conditioned). Reinforcers come in the form of desired objects, tokens, or stimuli. A primary reinforcer is a stimulus that is reinforcing for most members of a given species. It does not depend on a history of conditioning (Kelleher & Gollub, 1962). For example, reinforcers that fulfill biological needs are characterized as primary. They fulfill an immediate need in the organism and most members of a given species find them rewarding. Examples of primary reinforcers are food, water (Kelleher, 1966) and verbal praise (Marschark & Baenninger, 2002). Food and water are common reinforcers in learning experiments (Bitterman & Schoel, 1970). A conditioned reinforcer (referred to as secondary reinforcer by Hull [1934] and second order conditioning by Pavlov [1927]) will be reinforcing to members of a species who have learned to find the stimulus rewarding (Kelleher, 1966). These stimuli are often objects that have a learning history for a specific organism. Examples of object rewards are toys, prizes and merchandise (O'Neill, Horner, Albin, Sprague, Storey & Newton, 1997). These objects acquire their reinforcing qualities through associations. One specific example of this association is reward through tokens. A token is defined as an

7 object that has no value in itself (Pryor, 1995), but can be traded in to obtain goods or services (ex: money, stickers, or stars). Secondary Reinforcement A conditioned, or secondary, reinforcer is defined by Skinner (1953) as, "a stimulus having the effect of a reinforcer because of its relation to a stimulus already having that effect" (p.724). A secondary reinforcer is one that was initially neutral (a sound, light, or motion), yet has been paired so often with an established reinforcer that it becomes reinforcing in and of itself (Pryor, 1995). A classic example of a conditioned reinforcer is the use of a sound, specifically a "click" in the training of various species of animals (Pryor, 1995). Clicker training, as it is commonly known, begins by repeatedly pairing food with the sound of a click until the relationship is well established. After the association is made, the trainer will present a "click" sound to the animal when it displays desired behavior. The click reinforces the behavior and the animal will continue to respond in the correct manner. The work of Breland and Breland (1961) clearly illustrates the development of a secondary reinforcer. While studying food reinforcement and behavior chains in animals, Breland and Breland taught raccoons and pigs to place a wooden coin in a bank for a food reward. They found that with repeated reinforcement of successful attempts to place the coins in banks, they inadvertently created strong secondary reinforcement to the coin itself. The raccoons and pigs became reinforced by presentation of the coins, and ceased to perform the prescribed behavior (putting the coin in the bank) for which they would receive a food reward. Instead, the animals became preoccupied with manipulating the coins. The behavior became more pronounced as the animal became hungrier and it was

8 concluded by the experimenters that the coin represented food in some way to these animals (Myers & Myers, 1965). Despite examples of secondary reinforcement presented by Breland and Breland and Myers and Myers, little is known about the nature of the stimulus representation that maintains secondary reinforcement (Parkinson et al.., 2005). The development of a secondary reinforcer can be explained in two different ways. In order to form a secondary reinforcer, a non-reinforcing stimulus precedes a reinforcing stimulus. Two theories explain the mechanism of association. The first mechanism is classical conditioning. Under this theory, the pairing of the non- reinforcing stimuli and the reinforcer is the same as pairing the conditioned stimulus (CS) and the unconditioned stimulus (US) in classical conditioning. In support of this theory, Bersh (1951) demonstrated that second order reinforcers are most strongly established when presented 0.5 to 1.0 second before the primary reinforcer, which is the same time interval for optimal CS-US pairing. The second theory explains the formation of a conditioned reinforcer according to its discriminative properties (Keller & Schoenfeld, 1950). Keller and Schoenfeld argue that a secondary reinforcer must be a discriminative stimulus for a response in order to become a conditioned reinforcing stimulus. For example, Bugelski (1956) described this as an elicitation effect. He states that when a stimulus (for example a click) has been repeatedly paired with a primary reinforcer (such as food) it becomes a conditioned stimulus. Under this theory, the non-reinforcing stimulus predicts the presentation of the primary reinforcer. In other words, when the animal hears the click it has learned that it will receive food for displaying the behavior at some time in the future. The click

9 predicts the arrival of food. It is argued that establishment of a stimulus as a discriminative stimulus is necessary for a stimulus to become a secondary reinforcer (Keller & Schoenfield, 1950). In addition to discriminative properties, secondary reinforcers may be tokens that can be saved and exchanged for other reinforcers (Catania, 1992). Through association with a primary reinforcer, conditioned reinforcers acquire the ability to reinforce, or increase the likelihood of a behavior occurring (Mazur, 1993). Therefore, such reinforcers are useful in operant conditioning. Secondary reinforcers can be used to strengthen any operant response. Several factors affect how hard the organism will work to obtain a conditioned reinforcer. The number of pairings between primary and conditioned reinforcers, the size and quality of the primary reinforcer and the delay between the conditioned reinforcer and the presentation of the primary reinforcer all affect the strength of the conditioned reinforcer (Mazur, 1993). Secondary reinforcers have several distinct advantages over primary reinforcers. Two of these advantages have to do with ease of use and the ability to reinforce the animal immediately. When one is training a complex series of behaviors, it is disruptive to stop the organism midway through a sequence to deliver a primary reinforcer like food or petting. It is much more effective to simply click and reinforce mid sequence and present the earned reward at the end of the series. Therefore the reinforcement can be presented immediately without disruption to long behavior chains (Pryor, 1984). Secondly, when working with animals such as dolphins it is often not feasible to touch or physically manipulate their behaviors. Shaping via secondary reinforcement is much

10 more practical (Pryor, 1995). Additionally, a secondary reinforcer satiates much more slowly than a primary reinforcer. Satiation is the point where a biological need is satisfied. It is a temporary state where the organism has enough of the primary reinforcer such as food, water, sexual stimulation (Chance, 1999). All primary reinforcers are subject to satiation. For example, a bite of food is reinforcing for a hungry organism but each bite becomes less reinforcing until the organism is satiated. The one main advantage to the use of conditioned reinforcers is their resistance to satiation: "Conditioned reinforcers become immensely powerful. I have seen marine mammals work long past the point of satiety for conditioned reinforcers ... perhaps people who have already earned more money then they can actually spend, have accordingly become addicted to the conditioned reinforcer" (Pryor, 1995, p. 14). Such observations of persistent instrumental responding despite satiation are attributed to the mechanism of conditioned reinforcement (Morgan, 1974). Use of food in humans and animals: Benefits and consequences of food reinforcement Food is defined as a primary reinforcer, or reinforcer that motivates behavior despite learning history. As noted earlier, a primary reinforcer is motivating to most members of species whether or not they have had previous experience with that reinforcer (Epstein & Leddy 2006). Food is considered a primary reinforcer because it fulfills the biological need of hunger. It is only a salient reinforcer so long as the animal is food deprived at the time of the trial. It fulfills the "need" of hunger, which can be a powerful motivating force for performing the requested behavior. Epstein and Leddy (2006) note, "Food is a powerful

11 reinforcer and in some circumstances may be more motivating than drugs of abuse" (p. 23). Food reinforcement is used in many situations and in many settings, from the lab to the classroom. It is possible that in these settings, food is often intentionally or unintentionally paired with neutral signals or other remforcers such as affection or verbal praise to create a secondary reinforcer. Food reinforcement has a long history in the field of psychology. In research, food reinforcement is often utilized in the study of learning principles. Food reinforcement is most often used with animals such as rats (Boakes, Poli, Lockwood, & Goodall, 1978), pigeons (Schaeffer, 1979), monkeys (Kelleher, 1957; Salzbeg, Henton, & Jordan, 1968), rabbits (Rubin & Brown, 1969), cats (Watson, 1954) and dogs (Salzinger & Waller, 1962). Food reinforcement is also used in the teaching and training of both animals and humans. For children and adults, food is often used as a positive reinforcer in the implementation of applied behavior analysis (ABA) with developmental disabilities, autism, behavioral disturbances and learning disabilities (Kazdin, 2001). Similarly, the training of animals from domestic pets to zoo animals often involves reinforcement with food (Pryor, 1995). Current research illustrates the extensive use of food reinforcement in the implementation of ABA with autistic and other developmentally delayed individuals. As stated previously, ABA is an operant approach to the assessment, evaluation and change of behavior (Kazdin, 2001). The range of behavior which has been changed through applied behavior analysis is quite extensive. For example, McManus et al. (2003) used food to reinforce a child diagnosed with Fragile-x syndrome for successful toilet training. To encourage language skills, Butz and Hasazi (1973) reinforced a profoundly retarded

12 14 year-old girl for imitating trainer verbalizations and De Villiers and Naughton (1974) increased language usage in autistic boys with food reinforcement. Food reinforcement has been used to increase imitative behavior in developmentally delayed children (Baer, Paterson & Sherman, 1967). Food reinforcement has also been used to decrease unwanted behavior by increasing alternate or incompatible behavior. Such was the goal of Adelinis, Piazza, and Goh (2001) with their work to decrease the destructive behavior of autistic boys. It has also been used in the treatment of fear and avoidance. For example, Rapp, Vollmer and Hovanetz (2005) had success treating swimming pool avoidance in an autistic girl. In addition, food reinforcement is sometimes combined with other forms of reinforcement. Frankel and Graham (1976) found that food reinforcement and one-on-one teacher to child ratio had a positive effect on the attention and tantrum behavior in autistic and retarded children. As with any reinforcer, a food reinforcer must be presented immediately after the response (Miller & Wasseman, 1997) and therefore the piece of food must be bite-sized and easy to eat. Researchers have reported the use of many different kinds of food rewards including candy such as peanut butter cups or M&Ms (Adalinia, Piazza & Goh 2004) and other food such as pepperoni (McManus, Derby, & McLaughlin, 2003). Food is a desirable reinforcer as it fulfills a biological need and therefore often produces behavior change. It is used in a methodological way, as in ABA yet also in an unplanned way in everyday life. Examples of this can be seen in the classroom when a special education student receives a handful of popcorn for a completed task or in the parents' impromptu trip to the ice cream shop with their child for turning out a good effort at a sporting event. One is planned, researched, tracked and revised. The other is

13 hardly given a second thought. However, reinforcement in both situations could possibly produce behavior change on some level through the association between food and these events. It can be theorized that pairing food with praise, attention or other primary reinforcers will form an association between the food and these other reinforcers. Once this association is created, food, a primary reinforcer, also becomes a secondary reinforcer. Once food becomes a conditioned reinforcer, an organism may eat when it is not experiencing hunger and overeating may result. Overeating is a direct cause of obesity and overweight. Possible Causes of Overeating, Obesity and Overweight The theory that food can become a conditioned reinforcer which would contribute to overeating and overweight has not yet been investigated. In order to explore the possibility that food can become a conditioned reinforcer, current research in the area of obesity must be reviewed. Human obesity is a growing problem, especially in industrialized countries such as the United States. According to the National Health and Nutrition Examination Survey, the percentage of adults who reported being obese increased from 10.5% in the period between 1988 - 1994 to 22.9% in 1994 - 2000 (Lopez, 2004). In the year 2000, 20.1% of the adult population reported being obese and 36.1% reported being overweight. Among children ages two to five, the percentage of who were reported to be obese increased from 7.2% to 10.4% from the years 1999 to 2000 (Lopez, 2004). Being overweight or obese is a stigmatizing condition. This stigma is linked to commonly held attitudes concerning the cause of these problems. For example, it is a popular and common belief that weight is under the control of the person. Inactivity, mood, eating the

14 wrong food, repeated dieting and interpersonal factors are all thought to contribute to obesity (Harvey, Summerbell, Kirk, Hill, & Harvey, 2002). There are actually many factors contributing to being overweight and obese in humans. Factors such as genetics, environment and learning are speculated to play a role in the development of overeating in general and eating disorders in particular (Polivy & Herman, 2002; Southern & Gordon, 2003). Obesity in the United States is thought to be biologically influenced, requiring long-term treatment (Woods, Schwarts, Baskin, & Seely, 2002). Obesity poses a new opportunity for the pharmaceutical companies to invest vast amounts of money in the development of pharmacological therapy for a lucrative market. An example of this market can be seen in the number of weight loss aids available for purchase either over the counter or by prescription from a physician. Because of this, a large body of obesity research has focused on the biological aspects of obesity. Much of this research involves studying rats as well as humans. Experiments in which rats serve as subjects can be more carefully controlled than those with humans. Biologically speaking, there are a number of similarities between humans and rats (Sclafani & Springer, 1976). Using rat studies as a model, researchers defined both internal and external variables that contribute to obesity and overweight. Internal variables such as genetics and metabolism are not under the control of the individual (Krieshok & Karpowitz, 1988). Research strongly suggests a genetic component in the development of obesity. A specific mutant strain of rat, known as the Zucker strain, has a considerably higher baseline weight than rats that have not been genetically altered. After these rats are

Full document contains 93 pages
Abstract: Many researchers use food as a positive reinforcer for specific behavior change. Although food is necessary for survival, there are instances where food intake in both animals and humans can be excessive. It is plausible that overeating is the result of learning to associate food with other reinforcers. The purpose of this study was to investigate whether food would function as a conditioned (i.e. secondary) reinforcer in dogs. Subjects were 40 adult small breed (7-251bs) dogs between the ages of one and eight. The study was a 4 x 2 random assignment design consisting of five phases (ABCA'D). Dogs were randomly assigned to one of four reinforcement groups; food and praise, food only, praise only and a non-contingent control group. The mean difference in amount of food consumed from baseline one to return to baseline showed significant differences ( p = 0.038). Group mean differences in average weight change from baseline one to return to baseline also reached statistical significance (p < 0.01). Dogs reinforced with food plus praise were significantly heavier than dogs reinforced with food alone, praise alone or the non-reinforcement control. Dogs reinforced with food plus praise demonstrated fewer trials to criterion and a larger number of responses to food refusal, as hypothesized; however these differences failed to reach statistical significance.