J. Dairy Sci. 103:2896–2905 https://doi.org/10.3168/jds.2019-16714 © 2020, The Authors. Published by FASS Inc. and Elsevier Inc. on behalf of the American Dairy Science Association®. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). A virtual dairy herd as a tool to teach dairy production and management S. Calsamiglia,1* G. Espinosa,1 G. Vera,2 A. Ferret,1 and L. Castillejos1 1 Animal Nutrition and Welfare Service, Department of Animal and Food Sciences, Universitat Autònoma de Barcelona, Bellaterra 08193, Spain 2 Centre for Research in Agricultural Genomics, Universitat Autònoma de Barcelona, Bellaterra 08193, Spain ABSTRACT semester (2 years later). Results were better in dairy compared with other species questions in the first (9.6 The objective of this project was to develop and test vs. 3.7) and fifth (8.0 vs. 3.8) semesters. The virtual a web-based virtual dairy herd to help students under- dairy herd is an effective tool for teaching introductory stand the structure and functioning of a dairy herd, courses in dairy production. The program can be ac- and to promote active learning. At the beginning of cessed at www​.virtualdairyfarm​.org, and a manual and the course, the instructor defines the profiles of herds videos with instructions for instructors and students to be assigned to students (e.g., herd size, production, are available online. diets, fertility). Each student has a unique herd and Key words: virtual dairy herd, dairy herd management, engages in decision-making for desired management e-learning, teaching tool practices in the herd. Modeled events are based on cow physiology and normal dairy herd management prac- INTRODUCTION tices. Students’ activities and decisions include heat detection, insemination, pregnancy diagnosis, dry-off, One of the main objectives of introductory courses diet specifications, feeding groups, colostrum and milk- in animal production is to expose students to a system replacer feeding, weaning, treatment of diseases, and where knowledge from different disciplines (reproduc- milk withdrawal from the tank if antibiotics are used, tion, nutrition, management, physiology, health, eco- among others. The daily output provides information nomics, and others) is integrated. With traditional on technical indexes, economic performance, counters lectures, integration in introductory courses is difficult of incorrect decisions as feedback for students, and because it requires a good understanding of the system. score. Time in class can be devoted to discussions of Furthermore, teaching animal production is becoming dairy management issues. Additional exercises based more of a challenge due to the increasing numbers of on students’ own herds (e.g., calculating required space students with no rural background in many agricultural for cows, land for forage production, manure manage- and veterinary schools (Kensinger and Muller, 2006). ment) can also be implemented. Students’ performance General consensus exists that active learning improves in the virtual dairy farm was monitored over 3 years. the efficiency of the learning process (Wattiaux and The average score (n = 326) was 87.8 ± 1.1 over 100 Crump, 2006; Freeman et al., 2014). Practice and dis- points, suggesting that self-learning with the virtual cussion strategies are better teaching methods than dairy farm was highly successful. At the end of each are lectures, readings, and demonstrations (Tokuhama- semester, students (n = 277) responded to a survey Espinosa, 2010). To take students into farms where on the experience of working with the virtual dairy they can participate in herd-related daily activities and herd. Most students (>87%) agreed that the virtual decisions would be highly motivating and would im- dairy herd was more effective and motivating than prove the learning process, but it is not always feasible, traditional lectures and helped them understand dairy especially in introductory courses with large enroll- production better. In an unannounced test conducted ment. Case studies are excellent to integrate concepts at least 2 wk before the final exam, students (n = and “learn by doing” but also have problems: (1) they 207) were asked 14 questions on dairy cattle and 14 impose a burden on the instructor to generate a large similar questions on other species taught in the same number of cases and to grade them; (2) students need class through traditional lectures. A similar test on the to have good bases to resolve cases; and (3) they re- same students (n = 142) was conducted in their fifth quire a high degree of motivation from students. With limited time and resources, how can today’s instructors teach introductory dairy science courses with an inte- Received March 31, 2019. Accepted November 11, 2019. grated approach, to an increasingly nonrural popula- *Corresponding author: Sergio.Calsamiglia@​uab​.cat tion of students? The use of new technologies and the 2896 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2897 gamification of the classroom are alternatives that can (every 21 ± 3 d) unless pregnancy occurs. During heat, be of help (Kensinger and Muller, 2006; Medrano et al., heifers show estrous activity the day before and after 2010; Maiga and Bauer, 2013). Computer simulation mounting activity. Artificial insemination should occur models for teaching have been used in many areas, but only if animals have reached sufficient BW (>350 kg) with limited application in animal sciences. However, and height (>127 cm) and there is mounting activ- some successful examples exist in specific areas such as ity. Gestation (278 ± 5 d) will depend on conception genetics (Medrano et al., 2010), reproduction (Johnson rate (60%). Pregnancy can be diagnosed 34 d after in- et al., 1992), and manure management (Albrecht et al., semination, but heat detection on d 21 ± 3 d should be 2006). Although these tools are very useful examples observed, to identify inseminated nonpregnant heifers for teaching in these areas, a model that integrates the and reinseminate them. Daily milk yield is determined different management areas of a dairy herd does not based on Wood’s lactation curve equations (Wood, exist. 1967), taking into account the lactation number and We hypothesized that the use of a virtual dairy herd DIM. After calving, cows return to cyclicity (35 ± 4 would motivate students, help them integrate the dif- DIM), and thereafter reproduction proceeds in a similar ferent management areas of the dairy herd, and im- fashion to that of heifers, except for a lower conception prove the learning process. The objective of this project rate (40%). Cows should be dried off 60 d before the was to develop and test a virtual dairy herd to help expected calving date. students understand the structure and functioning of a Actual cow performance is affected by feeding, dis- dairy herd, and to promote active learning through a eases, and management factors. Intake of DM is de- game-like environment. termined based on NRC (2001) equations. A default feed description and diet composition is provided, MATERIALS AND METHODS but instructors may change it to represent their local conditions. Nutrient requirements and supplies are The project was developed in 2 steps: (1) develop- simplified in terms of energy, CP, and forage NDF ment of the virtual dairy farm; and (2) evaluation of (fNDF). Energy requirements are simplified to 10 Mcal the virtual dairy farm as a learning tool. of NEL/d for maintenance and 0.72 Mcal of NEL/L of milk produced, and protein requirements to 400 g of Step 1: Development of the Virtual Dairy Farm CP/d for maintenance and 90 g of CP/L of milk, based on NRC (2001). Energy and protein supplies are calcu- The software was programmed with JAVA language lated by multiplying DM intake and nutrient content, and runs in an Apache Tomcat server (Apache Software and actual production is determined based on the most Foundation, Wakefield, MA). Herd events are based on limiting nutrient. A minimum level of fNDF (19%) is cow physiology and normal management practices in defined as the threshold for the incidence of acidosis. If small to medium conventional dairies using confine- acidosis occurs, then milk production (−20%) and milk ment facilities. Simulations run on a cow-by-cow and fat content (−15%) are reduced. Students may define day-by-day basis. Calves are born male or female different feeding groups. Each group requires a descrip- (50%), inherit the milk production genetic potential of tion of diet ingredients and cost, and students have to their dam, and are assigned a starting BW (40 kg) and move animals among groups to optimize performance height (75 cm). Newborn calves require actions such as and profitability. ear tagging and navel disinfection during the first day Diseases are described based on occurrence and ef- of life, and colostrum feeding (3 L, 2 times per day) on fects. Occurrence describes incidence, age, month of the the first and second day of life. On d 3, calves should year, DIM at risk interval, and duration of the disease. be moved to the milk-fed group, where they should be Effects describe the influence of disease on intake, milk fed milk replacer (3 L, 2 times per day). Male calves yield and composition, SCC, return to ovarian cyclicity should be sold after 8 d of life. Intake of concentrate after calving, conception rate, abortion, risk of culling, will increase progressively after d 15 of life based on and growth. Treatment may be applied to each disease. NRC (2001) equations. Growth rate is set by default If treatment is applied, the duration of the disease is for BW (0.8 kg/d from birth to first calving) and height reduced, but there is an associated cost. If antibiot- (0.14 cm/d from birth to puberty and 0.07 cm/d from ics are used, milk should be withdrawn from the tank puberty to first calving). Students may define different for the days indicated in the treatment. New diseases heifer feeding groups with different growth rates and can be added as long as the incidence and effects are costs. Weaning is determined based on the amount of described. Sick animals appear in the Sick tab within concentrate consumed (1 kg of DM/d). At puberty (300 the Cow screen, and students are responsible for check- kg BW), the ovarian cycle will start to occur regularly ing sick animal daily, applying appropriate treatment, Journal of Dairy Science Vol. 103 No. 3, 2020 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2898 and withdrawing milk from the tank if antibiotics are in the first semester of the Veterinary School of the used. Default values for the most common diseases Universitat Autònoma de Barcelona, with 120 students (mastitis, laminitis, retained placenta, metritis, ketosis, enrolled per year. The evaluation considered (1) stu- hypocalcemia, and displaced abomasum for cows; diar- dents’ performance using the virtual dairy herd for 150 rhea and respiratory diseases for calves) are provided. d; (2) a survey on student perceptions of the learning The Disease menu can also be used to describe certain tool compared with traditional lectures; and (3) a test management or environmental factors. For example, of knowledge at the end of the semester. heat stress is modeled by default, by assigning a reduc- Students’ Performance Using the Virtual Dairy tion in DM intake (−15%), milk production (−15%), Herd. A scoring system was developed to evaluate milk fat (−5%), milk protein (−5%), heat detection students’ performance while working with the virtual rate (−50%), and conception rate (−50%) for the sum- dairy herd. The score accounted for students’ incorrect mer months. Students will also have to decide to cull decisions in different activities such as heat detection, cows when the number of animals exceeds the maxi- insemination, dry-off, colostrum and milk-replacer mum capacity of the herd. For every day simulated, the feeding, weaning, transfer of animals between groups, fixed and variable costs (such as feed, labor, utilities, treatment of diseases, and milk withdrawal if antibiot- semen costs, and health treatments) and income (sales ics are used. Incorrect decisions are those not done at of milk, calves, and culled cows) are calculated. the appropriate time or done at an inappropriate time. The software is structured with an administrator, in- Bonus points can be obtained by adjusting the diet structors (with their associated courses), and students. to optimize production and economic performance. All The administrator can generate courses with a default students start with 100 points. Points are lost based on herd profile and assigns one or more instructors. Instruc- percentage of incorrect decisions within each activity, tors may modify the default herd to fit the objectives and each activity accounts for a proportional fraction of the course and may also modify the settings at any of the final score. This approach allows the student to time during the development of the course. Students recover lost points if the decision is repeated correctly manage their own unique herd in which cows are gener- because, as number of correct decisions taken increases, ated at random with a milk production potential of the percent of incorrect decisions are diluted. the herd average ± 2,000 L/cow per yr. The instructor Student Perception of the Learning Aid. Stu- also determines the starting date, the number of days dent perception of the usefulness of the methodology to be simulated, and the deadline for completing all to understand dairy production was evaluated by ad- tasks. All changes implemented by the instructors are ministering a survey at the end of the first semester common to all students in the same class. The system that contained 10 items to be scored between 1 (totally allows generation of unlimited instructors, courses by disagree) and 7 (totally agree; Table 1), and an overall instructor, and students within a course. score (1 to 100) for evaluation of the virtual dairy herd as a learning tool. Step 2: Evaluation of the Virtual Dairy Herd Test of Knowledge. We also conducted an unan- as a Learning Tool nounced test at least 2 wk before the final exam. Stu- dents were tested on dairy cattle production concepts Evaluation of the tool was conducted in 3 consecu- learned through the use of the virtual dairy herd and tive years in the course “Bases of Animal Production” on production concepts for other species taught in the Table 1. Survey of student attitudes toward the virtual dairy herd as a learning tool compared with conventional in-class lectures (n = 277) Score frequency1 (%) Item 1 2 3 4 5 6 7 The virtual dairy herd helped you understand dairy cow production cycle 0 1 1 6 22 43 27 The virtual dairy herd helped you understand how a dairy herd works 2 2 3 23 29 28 13 The virtual dairy herd helped you understand other animal production systems 1 0 1 4 19 34 41 Learning with the virtual dairy herd was more effective 1 0 3 9 23 35 29 Working with the virtual dairy herd helped you to remember concepts of dairy production 1 0 1 10 29 32 27 The virtual dairy herd was a more motivating method to learn than traditional lectures 1 1 3 5 22 33 35 The virtual dairy herd increased your interest in other animal production classes 1 1 3 26 22 25 22 (nutrition, reproduction, physiology, etc.) Virtual tools should be used more frequently in other species 0 1 2 5 21 36 35 1 1 = totally disagree; 2 = mostly disagree; 3 = some disagreement; 4 = equal; 5 = some agreement; 6 = mostly agree; 7 = totally agree. Journal of Dairy Science Vol. 103 No. 3, 2020 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2899 Table 2. Questions used on the unannounced test to evaluate whether use of the virtual dairy herd improved short- and long-term retention of knowledge related to dairy cattle production compared with traditional lectures used to teach about other species Questions related to dairy cattle Questions related to other species Length of the ovarian cycle in a cow, d Length of the ovarian cycle in a sheep, d Length of gestation in a cow, d Length of the ovarian cycle in a sow, d When should a cow be dried off? Length of gestation in a sow, d Conditions required to inseminate a heifer Length of gestation in sheep, d Days of colostrum feeding to calves, d When should a goat be dried off? Amount of colostrum fed to calves, L/d Conditions required to inseminate a goat When should a calf be weaned? When should a piglet be weaned, d? Amount of milk replacer fed to calves, L/d When should a lamb be weaned, d? Price of cow’s milk, €/L Slaughter weight of a lamb, kg Price of an 8-mo-pregnant heifer, € Price of goat’s milk, €/L Income from a culled cow, € Price of a pregnant ewe, € Mention 3–5 ingredients of a dairy cow diet Weight at slaughter of a broiler, kg Peak production of a cow, kg/d Age at slaughter of a broiler, d Average production of a cow, kg/d Eggs produced by a layer hen per year, no. same class using traditional lectures. The exam con- about incorrect decisions (point C in Figure 1, showing sisted of 14 questions related to dairy cattle production 3 errors out of 16 heifer inseminations) as immediate and 14 questions of the same type related to poultry, feedback to students. The Economy screen is where swine, or sheep production (Table 2). At the beginning daily income and expenses are calculated and the his- of the fifth semester, 2 years later, the same students torical cash balance reported. As an example, Figure completed the same test without previous notice, to 2 shows the evolution of income and cost following a ascertain whether the method improved long-term re- reduction in cost of feeding due to the introduction of tention of information. high- and low-production groups by a student around d Data were analyzed with SAS (v. 9.4; SAS Institute 16 of simulation (point A in Figure 2), and to increase Inc., Cary, NC) and tested for normal distribution. Dif- income due to an increase in milk price programmed ferences among years in students’ performance in the by the instructor around d 27 of simulation (point B virtual dairy farm exercise and in the survey data were in Figure 2). These are examples of changes that can analyzed using PROC MIXED, where year (1, 2, 3) was be used by the instructor to demonstrate the influence the fixed effect. Scores of the unannounced test were of these and other decisions on herd performance and analyzed using PROC MIXED, where year (1, 2, 3), economics, making students aware of the relevance of semester (first vs. fifth), type of test (dairy vs. other proper management to herd profitability. species), and their interaction were fixed effects, and Most decisions take place in the Cows and Calves the student was the random effect. Differences were screens (Figure 3), where animals may be classified by declared at P ≤ 0.05. any of the main technical indexes, which allows stu- dents to access animals in a specific condition (e.g., RESULTS AND DISCUSSION search for animals around d 21 in the ovarian cycle for heat detection; search for animals 220 d pregnant for The Virtual Dairy Herd drying off). Additional information on each animal can be accessed in the individual cow’s card (+ signs in The challenge of developing the tool was to balance Figure 3). Figure 3 also shows a cow’s individual card animal biology and herd management practices with with the lactation curve (point A in Figure 3), mount- teaching and learning objectives. Furthermore, the soft- ing activity (point B in Figure 3), and specific decisions ware had to be flexible to allow adaptation to different that may be taken on individual cows (inseminate, cull, scenarios, and the interface intuitive and easy to use by or dry off). Cows may also get sick and, when sick, they instructors and students. Interaction of students with should be treated. For example, point A in Figure 3 the virtual dairy farm was achieved through 4 tabs to shows the lactation curve of a cow with a previous clini- gain access to the respective screens: Indicators, Econ- cal metritis event, where milk production was reduced omy, Cows, and Calves. The Indicators screen provides for several days and then recovered. If antibiotics are information on initial and current herd performance used for treatment, milk should be withdrawn from the (point A in Figure 1: milk yield and composition, fertil- milk tank for the required days to avoid being penalized. ity, and so on), overall score (point B in Figure 1, show- In fact, point C in Figure 2 shows the economic effect ing the example score 85 out of 100), and information of not withdrawing milk containing antibiotics, where Journal of Dairy Science Vol. 103 No. 3, 2020 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2900 income for milk sales was zero for that day. Occurrence Students’ Performance Using the Virtual of diseases is designed to help students understand the Dairy Herd. The students’ performance in the virtual consequences of their decisions on herd management, herd exercise (n = 326) was affected by year (P < 0.05; performance, and profitability. All events and decisions 87.7 ± 1.2, 89.8 ± 1.1, and 86.1 ± 1.0, for years 1, 2, are recorded in an individual file for each cow. and 3, respectively), with an overall average score of 87.8 ± 1.1, a median of 90, and first and third quartile Evaluation of the Virtual Dairy Herd Tool of 81 and 98, respectively. Only 2% of students scored less than 60. Results show not only excellent perfor- Most students (typically 85%) in the Veterinary mance but also small variability among students. Part School at the Universitat Autònoma de Barcelona of the success of the tool relies on students’ motivation (Spain) come from urban areas with limited or no rural and the fact that they get immediate feedback on their background, which is becoming a common scenario in management decisions. When an incorrect decisions oc- many veterinary and agricultural schools around the curs, the score is reduced, and feedback is provided to world (Kensinger and Muller, 2006). An introductory the student, pointing to that decision. Because scores course in animal production is required in the first are calculated as a percent of incorrect decisions, in- semester of the veterinary degree. Achieving the ob- creasing the number of correct decisions reduces the jectives of the class with a large enrollment (average percentage of incorrect decisions, allowing improvement of 120 students per semester) is difficult, due to the in the score. This calculation method rewards students complexity of the concepts to be taught, the nonrural who learn from their mistakes. This is a unique feature backgrounds of students, and the limited time avail- that transforms incorrect decisions into learning op- able. portunities. Figure 1. Partial view of the Indicators screen of the virtual dairy herd, describing technical indexes (point A, 29.5 L/d), current score (point B, 85 points over 100), and counter of incorrect decisions (point C, 3 incorrect decisions out of 16 heifer inseminations). Journal of Dairy Science Vol. 103 No. 3, 2020 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2901 Student Perception of the Learning Aid. A sur- tition, reflecting a high degree of motivation, consider- vey conducted at the end of the semester for 3 consecu- ing that most students had limited prior knowledge. tive years asked students (n = 277) their perceptions Test of Knowledge. Results of an unannounced of the usefulness of the virtual dairy farm compared test conducted at least 2 wk before the final exam con- with traditional lectures. Students’ responses were not firmed the perception that the virtual farm was a useful different among the 3 years. When students were asked learning tool. It is a common concern for instructors to score the overall usefulness of the tool between 0 that most students will study class material only in and 100, the average was 83 ± 2.6. The frequency of the weeks before the exam, which results in short-term responses for each specific question is shown in Table memorization of facts. Therefore, the quiz was intended 1. Compared with conventional in-class lectures, 92% of to test the hypothesis that continuous work with the students considered that the virtual dairy herd helped virtual dairy farm allowed students to better gain and them to better understand the dairy cow production retain knowledge. Over the 3 years monitored, the test cycle, 70% consider that it helped them to better was completed by 208 students in the first semester and understand the functioning of a dairy herd, and 94% 122 students 2 years later, in the fifth semester. Year was considered that it helped them to better understand not significant (P > 0.05), but we found an interaction animal production systems. Most students considered of type of exam (dairy vs. other species) by semester the virtual dairy farm more motivating (90%) and more (P < 0.05). Scores were higher in questions related to effective for learning (87%) compared with traditional dairy production compared with equivalent questions lectures. Students (69%) also recognized that the tool on other species taught through traditional lectures in increased their interest in other subjects related to the first (9.6 vs. 3.7 over 14 questions; SE = 0.15; P < animal production compared with traditional lectures. 0.05) and fifth semester (6.9 vs. 3.8 over 14 questions; Most students (92%) think that the virtual dairy farm SE = 0.19; P < 0.05), although the score of the dairy should be used more often and in other species. Less section decreased slightly in the fifth compared with than 7% of students preferred conventional lectures in- the first semester. Results indicate that persistent work stead of the virtual dairy herd. About 42% of students with the virtual dairy herd improved short- and long- would like to continue working with the virtual dairy term learning compared with traditional lectures. farm beyond the required simulation days in the class, Through use of the virtual dairy herd, students were and 67% would participate in an interuniversity compe- able to learn about (a) the biological cycle of cattle from Figure 2. Economy screen in virtual dairy herd. Evolution of income (solid line) and expenses (dashed line), daily balance, and total cash available. The graph shows a reduction in daily costs around d 16 due to the implementation of 2-group feeding (point A), an increase in income due to an increase in milk price at d 27 (point B), and lack of income at d 25 due to the presence of antibiotics in milk (point C). Journal of Dairy Science Vol. 103 No. 3, 2020 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2902 birth to production; (b) reproduction management, simulation conditions for the class. Figure 4A shows a such as puberty detection, heat detection, insemination screenshot of the default reproduction criteria, where at specific criteria, and pregnancy diagnosis; (c) the the instructor can introduce modifications on techni- role of proper nutrition (e.g., modifying the diet to in- cal indexes. Students will manage their own herds, and crease or decrease energy or protein supply, evaluating the main objective is to let students learn the basic milk production and economic benefits, assigning dif- management activities in a dairy herd (Table 3): heat ferent feeding groups); (d) the effects of prices of milk detection, insemination, reproductive cycle, dry-off, and feeds; (e) the relevance of a disease, as through lost prices of feeds and milk, newborn calf management, production, shorter duration if treated, and the need to milk-replacer feeding, weaning, reproduction in heifers, withdraw the milk from the tank if antibiotics are used; basic health issues, and the importance of antibiotic (f) management of heifers and the effects of growth on residues in milk, among others. In a standard class, puberty and proper insemination timing criteria (body students are required to simulate 150 d. Reviewing all weight and height). daily activities in the herd takes about 2 to 3 min per day simulated. If the student’s task is to simulate 150 Class Organization d, then the expected workload to complete the exercise will be 5 to 8 h. The instructor has access to all stu- The virtual dairy farm has been used for 9 years in dents’ herds to follow up the exercise (Figure 4B). The our introductory animal production courses in the Vet- introduction to the virtual dairy herd requires about 1 erinary School at the Universitat Autònoma de Barce- h of lecture to describe the production cycle of dairy lona, after which we can draw some recommendations. cows and 1 h of hands-on computer training to teach We used a standard herd with 100 to 120 adult cows students how to use the software. No additional formal and their respective replacement heifers. The instructor training is done. Students are fully responsible for the may change the default description of the herd or the management of their own virtual dairy herd. Students Figure 3. Cows screen in the virtual dairy herd, showing major technical indicators of all cows. The left-side submenu provides access to cow’s groups, sick cows, and diets. The individual cow card (shown by clicking on the + sign) provides information on lactation curves, where a metritis event reduced milk production from d 8 to 21 postpartum (point A), estrus activity (point B), and decisions that may be taken (insemination, dry-off, or culling of cow). Journal of Dairy Science Vol. 103 No. 3, 2020 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2903 Figure 4. Instructor screens in virtual dairy herd. (A) Instructors have access to a set of screens where default values for a class can be modified (e.g., fertility of cows inseminated on the day of mounting activity is 40%). (B) Instructors have access to all students’ herds to consult about their activities (e.g., student 1529499 has an average production of 30.8 L/d and a final score of 87). Journal of Dairy Science Vol. 103 No. 3, 2020 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2904 Table 3. Physiological events and management decisions that need to be performed by students in the virtual dairy herd Event Decisions/actions Birth 50:50 male:​female Ear tagging (d 1 of life) BW = 40 kg Navel disinfection (d 1 of life) Body height = 75 cm Colostrum feeding (3 L, 2×/d, 2 d; d 1 and 2 of life) Milk-fed calves 0.8 kg ADG, 0.14 cm/d height gain Move from newborns to milk-fed group (d 3 of life) Intake starter Feed 3 L of milk replacer, 2×/d (daily while in milk-fed group) Weaning (when starter consumption = 1 kg) Heifers Heat detection (d 21 ± 3 of estrus cycle) Puberty (300 kg of BW) Insemination (when mounting if BW > 350 kg, height >127 cm) Pregnancy Pregnancy diagnosis (>34 d after insemination) Cows Return to cyclicity Heat detection (d 21 ± 3 of estrus cycle) and insemination (when mounting) Pregnancy Pregnancy diagnosis (>34 d after insemination); dry-off (d 220 to 226 of pregnancy) Feeding Define diet composition and cost Define feeding groups Prevent acidosis (>19% fNDF1) Disease Identify sick animals (check sick cows daily) Treat for disease (when sick) Withdraw milk if antibiotics are used (after treatment) 1 fNDF = forage NDF. may help each other, but each student has to run his ing space, and space for different types of animals in the or her own herd. The instructor has certainty that each herd (i.e., milking and dry cows, calves, and heifers). student has managed his or her herd and does not have Finally, throughout the semester, students visited to grade each student individually, allowing implemen- farms from all species studied as part of the required tation of the system in large classrooms. class activities. A remarkable outcome in the visit to In a typical class, in addition to the daily herd ac- the dairy farm was that students were very motivated tivities, other exercises can be introduced at different to ask questions of farmers and reacted to answers with times. For example, the instructor may introduce one their virtual dairy herd as a benchmark. This higher or more changes after 50 d of simulation, to challenge level of engagement—and, thus, the greater educational students. For example, the instructor might (a) reduce value derived from interacting with the dairy producer the NDF content of forages to cause acidosis, which during the farm visits—was most likely the result of will prompt students to troubleshoot the problem and engaging with the virtual dairy herd. Thus it appears modify the diet accordingly; (b) change the prices of that the virtual dairy herd enhances the value of the milk or feeds to increase awareness of the importance field trips, and, possibly, the reverse was also true. of prices on profitability; or (c) change the incidence The virtual dairy herd is available in Catalan, Span- of a specific disease. These changes can be introduced ish, and English (units in kg and lb), and is acces- during the development of the exercise, and the conse- sible free at www​.virtualdairyfarm​.org. The website quences become the objective of an in-class discussion. contains a manual and short videos with introductions These discussions are very motivating, because any in- for instructors and students on how to use it. To acti- put from the discussion can be applied by each student vate an account as instructor, send a request to info@​ in his or her own virtual dairy herd. Furthermore, in- virtualdairyfarm​.org. class discussions are recognized as an effective teaching method (Chickering and Gamson, 1991; Cameron, 2003; CONCLUSIONS Wattiaux and Crump, 2006). The virtual dairy herd can also be used to ask students to conduct additional The web-based virtual dairy herd is an excellent tool exercises. Some exercises we have used in the classroom to generate an unlimited number of herds to let stu- include (1) calculating the amount of manure produced dents independently control selected daily activities of (total DM and N per year) and total hectarage of land a dairy herd operation. The immediate feedback helps required for its distribution, increasing student aware- students identify incorrect decisions, and automatic ness of environmental issues; (2) calculating the land scoring allows for use in large classrooms without the required for production of different forages or designing burden of individual corrections. Students’ self-learning the sizes of different silos; and (3) drawing at scale the with the tool allows additional time for in-class discus- layout of the facilities, including feedbunk space, drink- sions. The virtual dairy herd can be tailored for use Journal of Dairy Science Vol. 103 No. 3, 2020 Calsamiglia et al.: DAIRY INDUSTRY TODAY 2905 in different environments and with different objectives. Proc. Natl. Acad. Sci. USA 111:8410–8415. https:​/​/​doi​.org/​10​ .1073/​pnas​.1319030111. Results of the exercise, the test, and a survey show that Johnson, P. J., P. A. Oltenacu, and R. W. Blake. 1992. LEARNRE- the virtual dairy herd is an effective learning tool that PRO: A computer-assisted training program for teaching dairy helps students learn and understand dairy production reproductive management. J. Dairy Sci. 75:2288–2293. https:​/​/​doi​ .org/​10​.3168/​jds​.S0022​-0302(92)77990​-9. in a motivating and integrated way. Kensinger, R. S., and L. D. Muller. 2006. Major advances in teaching dairy production. J. Dairy Sci. 89:1155–1162. https:​/​/​doi​.org/​10​ .3168/​jds​.S0022​-0302(06)72184​-1. ACKNOWLEDGMENTS Maiga, H. A., and M. L. Bauer. 2013. Using interactive flash games to enhance students’ learning in animal sciences. NACTA J. 57:6–66. This project was funded by the Generalitat de Cata- Medrano, J. F., A. Ahmadi, and J. Casellas. 2010. Dairy cattle breed- lunya (Barcelona, Spain) under the program Improving ing simulation program. A simulation program to teach animal breeding principles and practices. J. Dairy Sci. 93:2816–2826. of Teaching Quality (Project MQD2010-00170). The https:​/​/​doi​.org/​10​.3168/​jds​.2009​-2640. authors have not stated any conflicts of interest. National Research Council. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed. National Academy Press., Washington, DC. Tokuhama-Espinosa, T. 2010. The New Science of Teaching and REFERENCES Learning. Teachers College Press, New York, NY. Wattiaux, M. A., and P. Crump. 2006. Students’ perception of a dis- Albrecht, G. L., Q. M. Ketterings, K. J. Czymmek, E. M. Van Am- cussion-driven classroom environment in an upper-level ruminant burgh, and D. G. Fox. 2006. Whole farm nutrient management: nutrition course with small enrollment. J. Dairy Sci. 89:343–352. Capstone course on environmental management of dairy farms. https:​/​/​doi​.org/​10​.3168/​jds​.S0022​-0302(06)72100​-2. Nat. Sci. Educ. 35:12–23. https:​/​/​doi​.org/​10​.2134/​jnrlse2006​.0012. Wood, P. D. P. 1967. Algebraic model of the lactation curve in cattle. Cameron, V. L. 2003. Teaching advanced genetics without lectures. Nature 216:164–165. https:​/​/​doi​.org/​10​.1038/​216164a0. Genetics 165:945–950. Chickering, A. W., and Z. F. Gamson. 1991. Applying the seven prin- ciples for good practice in undergraduate education. New Direc- ORCIDS tions for Teaching and Learning. 47:5–7. Freeman, S., S. L. Eddy, M. McDonough, M. K. Smith, N. Okoroafor, S. Calsamiglia https:​/​/​orcid​.org/​0000​-0002​-0978​-3799 H. Jordt, and M. P. Wenderoth. 2014. Active learning increases L. Castillejos https:​/​/​orcid​.org/​0000​-0001​-7762​-192X student performance in science, engineering, and mathematics. Journal of Dairy Science Vol. 103 No. 3, 2020