Professional Documents
Culture Documents
2012 Penstate Nutrition Workshop Proceedings
2012 Penstate Nutrition Workshop Proceedings
2012 PROCEEDINGS
November 13, 2012 7:00 8:00 8:00 12:15 8:25 8:30 8:30 9:15 9:15 10:00 10:00 10:30 10:30 11:15 11:15 12:00 12:00 12:15 12:15 2:00 12:15 1:30 12:00 8:00 2:00 3:15 3:15 3:30 3:30 4:45 5:00 5:00 7:00 7:15 9:30 Registration; coffee and refreshments Novus International Inc. Preconference Symposium Introduction Opportunities and Challenges to Achieving 50,000 Pounds of Milk with Components, Dr. Dale Bauman, Cornell University Using Science to Improve Cow Comfort, Dr. Trevor DeVries, University of Guelph Break Identifying Bottlenecks and Collaborating on Solutions for Change, Mr. Jeff True, True Farms and Mr. Corwin Holtz, Holtz-Nelson Dairy Consultants Unlocking the Bottlenecks to Performance, Dr. Gordie Jones, Central Sands Dairy LLC Questions and Wrap Up Lunch, Sponsored by Novus ARPAS Northeast Chapter annual meeting Exhibit hall open Afternoon workshop - session 1 Break Sponsored by Novus Afternoon workshop - session 2 ARPAS Exam offered Reception in exhibit area Sponsored by Alltech Inc. Evening Session and Dinner Sponsored by Alltech Inc. (RSVP required) Carbon Footprint Feeding Dairy Cows, Dr. Mike Hutjens, University of Illinois Evidence Based Nutrition Meets the Alltech Dairy Advantage, Mr. Jay Johnston, Ritchie Feed and Seed
November 14, 2012 6:30 7:45 Breakfast Sponsored by Multimin USA (RSVP required) The Effect of Injectable Trace Mineral Supplementation on Postpartum Health, Reproduction, and Production Parameters in Holsteins, Dr. Rodrigo Bicalho, Cornell University 6:45 7:45 7:00 - 3:00 8:00 8:45 8:45 9:30 9:30 10:15 10:15 10:45 10:45 11:45 11:45 12:00 12:00 - 1:00 1:00 2:00 2:00 3:00 2:00 Registration; coffee and refreshments Exhibit hall open Strategies for Management of Udder Health and Prevention of Clinical Mastitis, Dr. Ken Leslie, University of Guelph Hot Topics in the World of Silages, Dr. Limin Kung, University of Delaware Evaluating Starch Digestibility for Lactating Cows: An Integrated Approach, Mr. Pat Hoffman, University of Wisconsin-Madison Break Morning workshop - session 1 Break Morning workshop - session 2 Lunch, Sponsored in part by RP Feed Components Afternoon workshop session ARPAS Exam offered
Coffee service and breaks sponsored by Prince Agri-Products, Varied Industries Corporation, and Virtus Nutrition
Post-conference Seminar Sponsored by RP Feed Components (RSVP required) Managing Energy Metabolism in Transition Cows, Dr. Heather Dann, Miner Institute Meeting the Amino Acid Needs of the Transition Cow, Dr. Tom Overton, Cornell University Practical Ration Guidelines for Meeting the Needs of Most Transition Cows, Dr. Patrick French, RP Feed Components
Mark your calendars! Future Dates for the Penn State Dairy Cattle Nutrition Workshop
November 12 - 13, 2013 November 12 - 13, 2014
Grantville, PA
Dairy Products are included in Dietary Recommendations by Public Health Organizations around the World
% of calo ories & nutrients from da airy
Sustainability
meeting the needs of the present without compromising the ability of future generations to meet their own needs s o t r n needs
UN World Commission on Environment and Development, Brundtland Report, 1987.
Sustainability
Economic Social
Environmental
Productivity is Key
Productivity = Productive Efficiency
milk output per resource inputs k t r e inputs milk
Outline
Basis for Animal Differences Limits to Productivity Benefits of Productivity Future Opportunities
Key Questions
What has been the basis for the improvement in milk production? What biological processes are related to th i t the improvement? t?
Cycle of Life
Service functions
y Kidney work Heart work Respiration Nervous functions Liver functions 6 to 7 9 to 11 6 to 7 10 to 15 5 to 10
36 to 50
9 to 12 2 to 4 30 to 40
Total
*Derivation
40 to 56
of these estimates was discussed in detail by Baldwin (1968), Milligan (1971) and Baldwin and Smith (1974).
Amino acids
Glucose
Blood
Acetate OHB
Amino acids
NEAA carbon precursors (from TCA cycle)
Glucose
Glycerol
Fatty acids
Acetate ACC
OHB
Glucose 6 Phosphate
Glycerol 3 Phosphate
Galactose
TAG Synthesis
FAS
MILK PROTEIN
MILK LACTOSE
MILK FAT
Mammary Gland
Nutrient Partitioning
Major source of differences among dairy cows
high producing cows eat more and use nutrients for milk synthesis low producing cows direct less nutrients for milk synthesis; if they do eat more it goes to body fat
Sources of Variation
Efficiency component Digestion and nutrient absorption Nutrient use for maintenance and efficiency of use for productive functions Nutrient partitioning and intake Among-animal variation Low Low High
Outline
Basis for Animal Differences Limits to Productivity Benefits of Productivity Future Opportunities
Key Questions
Is a higher production possible? What are the limits to milk production? Are there concerns about achieving A th b t hi i high production?
Year
10
Protein Fat
Ever-Green-View My 1326 365d 72,170m 2,787F 2,142P Thomas Kestell Family Waldo, WI
11
Protein Fat
Gillette E Smurf Lifetime 478,167 lbs milk (15 yr & 10 lactations) Patenaude Family La Ferme Gillette Embrun, Ontario
12
13
US data (2009-10; 16768 herds) compiled by H.D. Norman and J.R. Wright, USDA-ARS-AIPL (2010)
14
Amish Dairy Farmer response to question Does higher milk production cause your cows to be stressed?
The tightrope walker doesnt fall off the rope because its too high; he falls because he loses his s e e s s balance.
Daniel Miller Dairy Extension Panelist Shipshewana, IN
Outline
Basis for Animal Differences Limits to Dairy Productivity Benefits of Productivity Future Opportunities
15
Key Questions
Why focus on productivity of dairy cows? What is the value to dairy producers?
16
Lactation
350
Maintenance
300
250
80%
200
150
63%
15.3 MJ ME/kg Milk
100
31% 69%
1944
50
37%
2007
20%
50,000 lbs
Year
Goal
17
Benefit: Environment
Environmental impact The production of all food has an environmental impact As productivity increases the s y s e environmental impact of milk production is reduced
Carbon Footprint of a Gallon of Milk Has Been Reduced by 2/3 Since 1944
18
Environmental Impact of Milk Production Has Been Considerably Reduced Since 1944
1944
* Note: per unit of milk Source: Capper et al. (2009) J. Animal Sci. 87:2160
Year
Goal
19
Outline
Basis for Animal Differences Limits to Dairy Productivity Benefits of Productivity Future Opportunities
20
Opportunity
Apply this Understanding t A l thi U d i t di to i Optimize Nutrition and Management so that Animal Well-being is Preserved g d Well being s and the Cow Achieves her Genetic Potential.
21
Thank You!
22
INTRODUCTION Poorly designed and managed facili es cause injuries and increase the risk of health problems including lameness and transi on cow disease, arguably two of the most serious welfare challenges facing the dairy industry (see von Keyserlingk et al. 2009). Producers spend millions of dollars building housing systems for dairy ca le, with the aim of providing a comfortable environment for their animals one that ensures adequate rest, protec on from clima c extremes, and free access to an appropriate, well-balanced diet. Despite these laudable aims, housing systems do not always func on well from the perspec ve of the cow. Poorly designed and maintained facili es can cause injuries, increase the risk of disease, and increase compe on among herd mates for access to feed and lying space. This proceedings chapter will review empirical work focused on the feeding, standing, and lying areas u lized by dairy ca le, and show how these can be be er designed and managed to improve cow comfort, prevent health problems, and improve produc vity. This work has generally evaluated housing systems from the cows perspec ve by asking how the housing aects cow health, what housing the cow prefers, and how the housing aects behavior. The aim of this research is to provide science-based solu ons to prevent problems and improve cow comfort. BETTER LYING AREAS The issue of cow comfort has received considerable interest within the dairy industry, with the bulk of research focused on the design of freestalls and the eect of stall design on stall occupancy and the me spent res ng. Objec ve and accurate records of lying me can be ascertained through use of inexpensive data loggers. To get a reasonable esmate of lying me for a farm, we suggest monitoring a minimum of 30 cows for at least 3 days (Ito et al., 2009). The research we describe below used data loggers and me-lapse video recordings to assess the eects of the surface cows lie down upon and how the stall is congured.
Lying Surface A growing body of research has demonstrated that the surface we provide for cows is one of the most important factors in designing a suitable lying area. Several researchers have measured stall usage, when the animals have no choice between surfaces, to assess how dierent bedding types aect behavior. For example, Haley et al. (2001) used a simple comparison between a space considered high comfort (a large box stall with ma resses) and a stall that represented low comfort (a e stall with concrete ooring). They measured many behaviors including lying, standing, and ea ng mes, the number of mes the cows stood up, and various leg posi ons during lying. Lying mes were 4 h longer and cows were more willing to stand up and change posi ons in the high-comfort housing. Cows also spent more me standing idle in the low-comfort stalls. Collec vely these studies tells us which behavioral measures are likely to change if a cow is uncomfortable, namely, me spent lying and standing, and the number of mes she is willing to stand up. In some of our groups rst work on cow comfort we found that cows on farms with ma resses (and li le bedding) have more severe hock lesions than do cows on farms that using deep-bedded stalls (Weary and Taszkun, 2000). Although similar results have now been found in other research (e.g. Wechsler et al., 2000), and most dairy professionals are aware of the risks of poorly-bedded ma resses, too o en this surface con nues to be used. Cows also clearly prefer lying surfaces with more bedding, and spend more me lying down in well-bedded stalls. In a more recent experiment we examined the eect of the amount of bedding on the me spent lying and standing by cows housed in free stalls (Tucker and Weary, 2004). Each stall was ed with a geotex le ma ress, and bedded with one of three levels of kiln-dried sawdust (0, 1, or 7.5 kg). Cows spent 1.5 h more me lying down in the heavily bedded stalls. In addi on, cows spent less me standing with only the front legs in the stall when the ma resses
23
were heavily bedded. These changes in both standing and lying behavior indicate that cows are hesitant to lie down on poorly bedded ma resses. These dierences in stall comfort may also account for a second important health problem; cows housed on mattresses also have a higher incidence of severe lameness than those housed in deep-bedded sand stalls (Ito et al., 2010). The lying surface can also aect udder health, and many studies have now shown the advantages to cows of using sand or other inorganic bedding as a way of reducing the growth of bacteria associated with environmental mas s (e.g. Zdanowicz et al., 2004). Making the decision to provide a well-bedded surface is just the rst step in achieving a reasonable level of cow comfortthis surface must also be properly maintained. In a series of experiments we have documented how the sand level declines in stalls that are not maintained, and how this decline reduces stall use by cows (Drissler et al., 2005). Sand levels in deep-bedded stalls decrease over a 10-day period, with the deepest part at the center of the stall. Lying me by cows also declines as the stall emp es: every inch decline decreased lying me by about half an hour per day. Contact with concrete while lying down may explain lower lying mes in deep-bedded stalls with less sand, and this concrete also aects leg health. Lesions on the point of the hock are common in deep-bedded stalls (Mowbray et al., 2003), likely due to contact with the concrete curb when stalls are not well maintained. Cows also showed a strong preference for lying on dry bedding during the summer months and when forced to lie down on wet bedding showed a 5-h reduc on in lying me (Fregonesi et al., 2007). Stall Configuration Most indoor housing provides more than just a lying surface for the cows. Typically the space is designed to encourage the cow to lie down in a specic loca on and to use the stall in such a way that feces and urine do not soil the stall. Unfortunately, most a empts to constrain how and where the cow lies down also reduce cow comfort as illustrated by the studies described below. Although some excellent recommenda ons for stall dimensions are now available, too o en new construc ons and renovated barns fail to provide appropriate space. We have conducted several experiments that show how stall size and congura on aect standing and lying mes. For example, in one study we tested the eect of stall width on cow behavior (Tucker et al., 2004), by providing cows
access to free stalls measuring 42, 46, or 50 between par ons. Cows spent an addi onal 42 min/day lying in the widest stalls, likely because they had less contact with the par ons in these larger stalls. Cows also spent more me standing with all four legs in the wider stalls, reducing the me they spent standing par ally (i.e. perching) or fully on the concrete ooring available elsewhere in the barn. STANDING AREA According to popular thinking, when cows are not in the parlor they should be ea ng or lying down. Unfortunately, no one seems to have explained this to the cows. In a number of studies we have found that even when cows have access to well-designed stalls they spend only about half of the day lying down. Cows spend the other 12 h a day on their feet, and we need to take this into account in designing suitable housing. In most barns the surface for standing outside of the stall is wet concretea known risk factor for hoof health (e.g. Borderas et al., 2004). Cows can use the stall as a refuge providing a dry, so er surface for standing. However, this increases the likelihood that cows will urinate and defecate in stall. The common response by barn designers has been to make the stalls more restric ve, forcing cows back into the concrete alley, and explaining in part why lameness is now the most prevalent and costly health problem for cows housed in freestall barns. With our current barn designs we are stuck with two bad choices: use restric ve stalls that keep the stall surface cleaner but force cows back onto the wet concrete or use more open designs and increase frequency of stall maintenance. Of these two op ons we favor the la er, but there may also be a third approachimproving the standing surface elsewhere in the barn. Both the height of the neck rail and its distance from the curb aect standing (Tucker et al., 2005); more restric ve neck-rail placements (lower and closer to the rear of the stall) prevent cows from standing fully in the stall, again increasing the me cows spend on concrete ooring elsewhere in the barn. Gait scores improve when neck rails are moved to a less restric ve posi on so that cows can stand with all four hooves in the stall and worsen when neck rails are more restric ve (Bernardi et al., 2009). Keeping cows out of the stall obviously helps keep the stalls clean. We found that both the narrow free stalls and the more restric ve neck rail placements reduced the amount of fecal ma er that ended up in the stall (Tucker et al., 2005; Bernardi et al., 2009). Although dirty stalls are undesirable, readers should be aware that stall cleanliness alone is a poor measure of stall design. Free stalls that have
24
higher occupancy rates are most likely to contain feces. Thus well-used stalls require more stall maintenance, just like other equipment used on the farm. Alterna ve standing surfaces may also help. A number of studies have shown that access to pasture improves hoof health, likely because under good grazing condi ons the pasture is a more comfortable and poten ally healthy surface to stand upon. For example, Hernandez-Mendo et al. (2007) showed that a rela vely brief period on pasture could help lame cows recover. Non-concrete surfaces can also provide be er trac on and be more comfortable for cows to walk upon. Cows will typically choose to walk upon a rubber surface and avoid concrete if the op on is available, and our research shows that cows slip less frequently and show improved gait when walking on rubber compared to concrete, a dierence that is especially clear for lame cows (Flower et al., 2007). Other work has shown that cows prefer to stand on so er surfaces, and moving the neck rail further from the curb reduces perching behavior and can reduce lameness cases. Bernardi et al. (2009) provided some of the rst experimental evidence that aspects of stall design can reduce the risk of lameness and hoof disease. This study assessed the eect of the posi on of the neck rail and found that over a 5 wk period, although we noted li le change in lying mes, gait scores improved for cows kept in pens without the neck rail compared to pens equipped with the neck barrier. However, these results also illustrate that some changes in design that result in improvements in hoof health come at the expense of cow hygiene and udder health. Although removing the neck rail comes at a hygiene cost (cows standing with all 4 feet in the stall will defecate and urinate more into the stall) there is no clear evidence that it increases the risk of mas s. A high standing me could suggest a decit in the cows environment; for instance, cows housed in pens with insufcient number of lying stalls, low bedding, wet bedding, or restric ve neck rails spend more me standing than those with dry stalls and less restric ve neckrails (Tucker and Weary, 2004; Fregonesi et al., 2007, 2009). Cows that perch with their 2 front feet in the stall during transi on are also at increased risk for lameness (Proudfoot et al., 2010); as stated above, this behavior has been linked with restric ve stall design (Tucker et al., 2005; Fregonesi et al., 2009). Providing cows the opportunity to avoid standing on wet concrete may be especially important during the transi on period. Increased standing me in the pre-partum period
is a key risk factor for hoof health problems later on in lacta on (Proudfoot et al., 2010). BETTER FEEDING AREAS There are several aspects of the feeding environment that aect the cows ability to access feed, including the amount of available feed bunk space per animal and the physical design of the feeding area. Reduced space availability increases compe on in ca le. For example, a recent study by DeVries et al. (2004) showed that doubling feed bunk space from 20 to 40 reduced by half the number of aggressive interac ons while feeding. This reduc on in aggressive behavior allowed cows to increase feeding ac vity by 24% at peak feeding mes, an eect that was strongest for subordinate animals. In addi on to the amount of available feed bunk space, the physical design of the feeding area can also inuence feeding behavior. One of the most obvious features of the feeding area is the physical barrier that separates the cow and the feed, and our research shows that some designs can reduce aggressive interac ons at the feed bunk. For example, Endres et al. (2005) compared the eects of a post-and-rail versus a headlock feed line barrier on the feeding and social behavior of dairy cows. Average daily feeding me (about 4.5 h day) did not dier, but during periods of peak feeding ac vity (90 min a er fresh feed delivery) subordinate cows had lower feeding mes when using the post-and-rail barrier. This dierence in feeding mes was likely due to posi ve eects of the headlock barriers in reducing compe ve interac ons; there were also 21% fewer displacements at the feed bunk with the headlock barrier compared to the post-and-rail barrier. These eects are greatest for the subordinate cow, par cularly at high stocking densi es at the feedbunk. Figh ng for access to feed has also been shown to increase drama cally when cows are temporally feed restricted (e.g. feeding to a slick bunk; Collings et al., 2011). Moreover, combining overstocking with a period of feed restric on may be especially problematic in terms of increasing compe on and reducing feeding behavior for lacta ng dairy cows. Adequate space and me to access feed is essen al to minimize feed bunk compe on in indoor group housing systems. In further research on feed bunk compe on we retested the eects of dierent feed bunk barriers, but did so over a range of stocking densi es (Huzzey et al., 2006). Cows were tested with the barriers described above but using stocking densi es of 0.81, 0.61, 0.41 and 0.21 m/cow (corresponding to 1.33, 1.00, 0.67 and 0.33 headlocks/cow). Daily feeding
25
mes were higher and the dura on of inac ve standing in the feeding area was lower when using a post-and-rail compared to a headlock feed barrier. As well, regardless of barrier type, feeding me decreased and inac ve standing increased as stocking density at the feed bunk increased. Providing adequate feed bunk space during the pre partum period is also essen al as work has shown that overstocking during this period reduces dry ma er intake (Proudfoot et al., 2009) and that cows that consume less are at higher risk for post partum disease (Huzzey et al., 2007). Cows were displaced more o en from the feeding area when the stocking density was increased, and this eect was greater for cows using the post-and-rail feed barrier. Again we found that this eect was greatest for subordinate cows, par cularly at high stocking densi es. Clearly, overstocking the feed bunk decreases me spent at the feed bunk and increases compe on, resul ng in poor feed access. We have recently found very similar eects (less usage and more compe on) when lying stalls are overstocked (Fregonesi et al., 2007). Moreover, when cows are overstocked at the stalls we have observed that cows on average le the feed bunk 30 min earlier when stocked at 150% compared to when they were stocked at 100% (Fregonesi et al., 2007). Other research has shown that providing addi onal par ons (feed stalls) between adjacent cows provides addi onal protec on while feeding and allows for improved access to feed (DeVries and von Keyserlingk, 2006). Providing a feed stall resulted in less aggression and fewer compe ve displacements, eects that were again greatest for subordinate cows. This reduced aggression allowed cows to increase daily feeding me and reduced the me they spent standing in the feeding area while not feeding. Thus, the provision of more bunk space, especially when combined with feed stalls, improves access to feed and reduces compe on at the feed bunk, and this eect is strongest for subordinate cows. These changes in feed bunk design and management could help reduce the between-cow varia on in the composi on of ra on consumed; under conven onal systems subordinate cows can only access the bunk a er dominant cows have sorted the feed (DeVries et al., 2005). The use of a barrier that provides some physical separa on between adjacent cows can reduce compe on at the feed bunk. A less aggressive environment at the feed bunk may also have longer-term health benets; cows engaged in aggressive interac ons at the feed bunk are likely at higher risk for hoof health problems (Leonard et al., 1998).
BARN LAYOUT Cow comfort may also be aected by overall layout of the barn. For example, some work has shown that cows rarely use certain stalls in a pen, while seemingly iden cal stalls are occupied more than 80% of the available me. One study (Gaworski et al., 2003) showed that stalls in the row closest to the feed alley were occupied 41% more frequently than were stalls in more distant rows. In addion, stalls located within the center of each row were used 12% more o en than those stalls located on the periphery of the row (i.e. either near a wall or fence). Natzke et al. (1982) also found that stalls on the periphery were used less than stalls in the interior of the row. These results suggest that certain stalls, par cularly those farther from the feed bunk and on the periphery, are less desirable to dairy ca le perhaps because cows need to walk farther, or because they have to navigate past certain physical (e.g. narrow alleys) or social (e.g. dominant cows) obstacles on their way to the more distant stalls. Indeed, earlier work has indicated that the movements of subordinate animals are prevented by the loca on of dominant cows (Miller and Wood-Gush, 1991). Such factors may partly explain reduced user sa sfac on and lower produc on in those barn designs consis ng of more rows (e.g. 6 and 4 row verses 2 and 3 row barns; Bewley et al., 2001). We also strongly encourage producers to evaluate their facili es on an individual resource basisthe lying, feeding, and standing areas. For example, large dierences in usage can occur even among iden cally congured stalls within the same barn. The fact that stalls within a pen vary in their popularity suggests that stall availability from the cows perspec ve is not the same as from the producers perspec ve. What looks to us as 1:1 cow-to-stall stocking density may seem considerably worse to the cows if some stalls are unacceptable. Another example is providing adequate feed bunk space on a per cow basis; for example, in a 6-row barn the amount of feed bunk space per cow is o en far less than that recommended. A number of lines of evidence now suggest that providing adequate feed bunk space is essen al to maintain dry ma er intake and reduced feed bunk space can have profound eects on rates of illness, par cularly during the transi on period (Huzzey et al., 2007; Goldhawk et al., 2009). CONCLUSIONS Scien c assessments have allowed us to determine which lying, standing, and feeding environments are preferred by dairy ca le, reduce the risk of health problems, and improve comfort. Cows like so er surfaces, for both lying down and for standing upon. Deep-bedded stalls work
26
well for cow comfort, but require maintenance. When it comes to the physical structures used to build freestalls, less is morethe hardware we place in the stall is for our benet and not the cows. The more restric ve we design stalls the less a rac ve they become for the cow. Use of restric ve stall designs can help keep stalls clean, but to avoid problems with hoof health these designs need to be accompanied by be er ooring op ons, such as so er and drier ooring. The design and management of the feeding area is also important. High stocking densi es at the feed bunk increase aggressive compe on and keep subordinate cows away from feed. Physical barriers between cows, including head locks and feed stalls, can help reduce this compe on and increase feed access. ACKNOWLEDGMENTS This paper is based upon an earlier proceedings paper en tled: von Keyserlingk, M. A. G. and D. M. Weary. 2008. Behaviour of Free Stall Housed Dairy Cows. Proceedings of the High Plains Nutri on Conference. pp. 7-16. We gratefully acknowledge our students and our many collaborators in this research. The University of Bri sh Columbias Animal Welfare Program is supported by Canadas Natural Sciences and Engineering Research Council Industrial Research Chair Program with industry contribu ons from the Dairy Farmers of Canada, Westgen Endowment Fund, Pzer Animal Health, BC Ca le Industry Development Fund, BC Milk Producers, BC Dairy Founda on, BC Dairy Educa on and Research Associa on, and Alberta Milk. REFERENCES
Bernardi, F., J. Fregonosi, D. M. Veira, C. Winkler, M. A. G. von Keyserlingk, and D. M. Weary. 2009. The stall design paradox: neck rails increase lameness but improve udder and stall hygiene. J. Dairy Sci. 92:3074-3080. Bewley, J., R. W. Palmer, and D. B. Jackson-Smith. 2001. A comparison of free-stall barns used by modernized Wisconsin dairies. J. Dairy Sci. 84:528-541. Borderas T. F., B. Pawluczuk, A. M. de Passill, and J. Rushen. 2004. Claw hardness of dairy cows: rela onship to water content and claw lesions. J. Dairy Sci. 87:2085-2093 Collings, L. M. K., D. M. Weary, N. Chapinal, and M. A. G. von Keyserlingk. 2011. Temporal feed restric on and overstocking increase compe on for feed by dairy ca le. J. Dairy Sci. 94:54805486 DeVries, T. J., M. A. G. von Keyserlingk, and D. M. Weary. 2004. Eect of feeding space on the inter-cow distance, aggression, and feeding behavior of free-stall housed lacta ng dairy cows. J. Dairy Sci. 87:1432-1438. DeVries, T. J., M. A. G. von Keyserlingk, and K. A. Beauchemin. 2005. Frequency of feed delivery aects the behavior of lacta ng dairy cows. J. Dairy Sci. 88:3553-3562.
DeVries, T. J., and M. A. G. von Keyserlingk. 2006. Feed stalls aect the social and feeding behavior of lacta ng dairy cows. J. Dairy Sci. 89:3522-3531. Drissler M., M. Gaworski, C. B. Tucker, and D. M. Weary. 2005. Freestall maintenance: Eects on lying behavior of dairy ca le. J. Dairy Sci. 88:2381-2387. Endres, M. I., T. J. DeVries, M. A. G. von Keyserlingk, and D. M. Weary. 2005. Eect of feed barrier design on the behavior of loose-housed lacta ng dairy cows. J. Dairy Sci. 88:2377-2380. Flower, F. C., A. M. de Passill, D. M. Weary, D. J. Sanderson, and J. Rushen. 2007. So er, higher-fric on ooring improves gait of cows with and without sole ulcers J. Dairy Sci. 90:1235-1242. Fregonesi, J. A., D. M. Veira, M. A. G. von Keyserlingk, and D. M. Weary. 2007. Eects of bedding quality on lying behavior of dairy cows. J. Dairy Sci. 90:5468-5472. Fregonesi, J. A., M. A. G von Keyserlingk, D. M. Veira, and D. M. Weary. 2009. Cow preference and usage of free stalls versus an open lying area. J. Dairy Sci. 92:5497-5502. Gaworski, M. A., C. B. Tucker, and D. M. Weary. 2003. Eects of two free-stall designs on dairy ca le behavior. Pages 139-146 in Proc. of the Fi h Intl. Dairy Housing Conf., American Society of Agricultural Engineers, St. Joseph, MI. Goldhawk, C., N. Chapinal, D. M. Veira, D. M. Weary, and M. A. G. von Keyserlingk. 2009. Prepartum feeding behavior is an early indicator of subclinical ketosis. J. Dairy Sci. 92:4971-4977 Haley, D. B., A. M. de Passille, and J. Rushen. 2001. Assessing cow comfort: Eects of two oor types and two e stall designs on the behaviour of lacta ng dairy cows. Appl. Anim. Behav. Sci. 71:105117. Hernandez-Mendo, O., M. A. G. von Keyserlingk, D. M. Veira, and D. M. Weary. 2007. Eects of pasture on lameness in dairy cows. J. Dairy Sci. 90:1209-1214. Huzzey, J. M., T. J. DeVries, P. Valois, and M. A. G. von Keyserlingk. 2006. Stocking density and feed barrier design aect the feeding and social behavior of dairy ca le. J. Dairy Sci. 89:126-133. Huzzey, J. M., D. M. Veira, D. M. Weary, and M. A. G. von Keyserlingk. 2007. Behavior and intake measures can iden fy cows at risk for metri s. J. Dairy Sci. 90:3320-3233. Ito, K., D. M. Weary, and M. A. G. von Keyserlingk. 2009. Lying behavior: Assessing within- and between-herd varia on in free-stall housed dairy cows. J. Dairy Sci. 92:4412-4420. Ito, K., M. A. G. von Keyserlingk, S. J. LeBlanc, and D. M. Weary. 2010. Lying behavior as an indicator of lameness in dairy cows. J. Dairy Sci. 93:3553-3560. Leonard, F. C., I. S enezen, and K. J. OFarrell. 1998. Overcrowding at the feeding area and eects on behavior and claw health in Friesian heifers. Pages 40-41 in Proc. 10th Intl. Symp. Lameness in Ruminants, Lucerne, Switzerland. Miller K., and D. G. M. Wood-Gush. 1991. Some eects of housing on the social-behavior of dairy-cows. Anim. Prod. 53:271-278. Mowbray, L., T. Vi e, and D. M. Weary. 2003. Hock lesions and free-stall design: Eects of stall surface. Pages 288-295 in Proc. of the Fi h Intl. Dairy Housing Conf. ASAE, St. Joseph, MI. Natzke, R. P., D. R. Bray, and R. W. Evere . 1982. Cow preference for free stall surface material rubber mats, carpe ng, and a layered mat. J. Dairy Sci. 25:146-153.
27
Proudfoot, K. L., D. M. Veira, D. M. Weary, and M. A. G. von Keyserlingk. 2009. Compe on at the feed bunk during transi on changes the feeding, standing and social behavior of Holstein dairy cows. J. Dairy Sci. 92:3116-3123. Proudfoot, K. L., D. M. Weary, and M. A. G. von Keyserlingk. 2010. Behavior during transi on diers for cows diagnosed with claw horn lesions in mid- lacta on. J. Dairy Sci. 93:3970-3978. Tucker, C. B., and D. M. Weary. 2004. Bedding on geotex le ma resses: How much is needed to improve cow comfort? J. Dairy Sci. 87:28892895. Tucker, C. B., D. M. Weary, and D. Fraser. 2004. Freestall dimensions: Eects of preferences and stall usage. J. Dairy Sci. 87:1208-1216. Tucker, C. B., D. M. Weary, and D. Fraser. 2005. Neck-rail placement: Eect on freestall preference, usage, and cleanliness. J. Dairy Sci. 88:2730-2737. von Keyserlingk, M. A. G., J. Rushen, A. M. B. de Passill, and D.M. Weary. 2009. Invited review: The welfare of dairy ca le Key concepts and the role of science. J. Dairy Sci. 92:4101-4111. Weary, D. M., and I. Taszkun. 2000. Hock lesions and free-stall design. J. Dairy Sci. 83:697-702. Wechsler, B., J. Schaub, K. Friedli, and R. Hauser. 2000. Behaviour and leg injuries in dairy cows kept in cubicle systems with straw bedding or so lying mats. Appl. Anim. Behav. Sci. 69:189-197. Zdanowicz, M., J. A. Shelford, C. B. Tucker, D. M. Weary, and M. A. G. von Keyserlingk. 2004. Sand and sawdust bedding aect bacterial popula ons on teat ends of dairy cows housed in freestalls. J. Dairy Sci. 87:1694-1701.
28
September 2010 scheduled to participate in NOVUS COWS assessment Discussed with Jeff True that herd needed change in cow comfort after seeing multiple sand herds over a 3 day period Challenges production level, foot and leg health, injured cows, involuntary culls and too many deaths from facility (flooring, stuck in stalls, etc.) issues. Scheduled to be off of BST in another 6 months
History of cow comfort at True Farms Cell filled mattresses up to 2005 - 2006 Waterbeds installed 2006 2007 Rubber flooring at feed alleys Rubber strips for travel lanes 2009 Deep bedded sand at satellite farm - 2007 Shallow bedded sand at home farm 2011-2012
By PEN
Pct
AvLS
By PEN
Pct
AvLS
Waterbeds 2-14 Sand 31-34 313 240 1.7 74 78 +3.4 2.2 514 159 2.3 76 83 -2.2 3.4
29
Description
30
73% of Pen 2 scored 3 or greater 8% of Pen 2 scored 4 or greater 6% severe hock damage 40% swollen front knees
31
Landscape Timber
32
33
Laydown Lame Severe Lame Inj Hocks Sev Inj Hocks SCC
THANK YOU
Questions and Discussion?
34
35
Trends in the North American Dairy Industry over the last 25 years can be best described by the term:
CHANGE
PM Fricke, Ph.D.
36
Covenant;
To care for, and keep
Vaccine
First vaccine. Small Pox Cow Pox Latin Vacca = Cow Vaccinate
Cow-inate!!!!
37
Reproduction
Your Interest
BOTTLECKS
38
Bottlenecks
Rate limiting problem Interferes with achieving Dairymans goal Any improvement pays off in better output Before it is completely fixed, something else will become the bottleneck
IN
OUT
Milk Production
Fresh Cow / Dry Cow
Milk Quality
Reproduction
39
Laying
Eating
Drinking
24 Hours
Maternity Pen
40
2 Years
24 Hours
1 Year
2 Years
41
Value of SOPs
Dairy personal are certified in their position! Every person knows and understands their job! Every person that is certified, has passed an exam (oral or written) about their job.
of Cow Comfort
A B Cs
Air quality & ventilation Bunk management
* R ti f Ration formulated, mixed & consumed l t d i d d * Bunk design & space * Feed quality
Cow comfort
Freestalls
4 Reasons Freestalls Fail
Lack of Cushion Neck Rail Placement Lunge Space Limitations Lack of Fresh Air / Vision
(1.7L)
42
As critical for cows entering stall, as for cows leaving the stalls Low neck rails contact cows during entry at the withers Optimal 44-46 (116cm) off of rear curb 44 46 Do not need to be moveable 66 (168cm) from rear of useable curb
Neckrails
168cm
116cm
Bedding
Absorbs Moisture Provides Cushion Prevents Friction
Bedding Types
Sand Mattress Dried manure Combinations
Inorganic--reduces bacterial growth. 40-70 lbs./Stall/day usage rates. Slope 1-2 % to flush effectively. Sand trap to catch sand 0.5% slope for 50ft 0 5% 50ft. Types of sand. Pitch to curb 1/4 per foot (.6cm/30cm).
Sand Bedding
Bedding must be maintained level with the curb for the curb width to be useable
43
Wide stalls
Why are we making them wider? Cows lay diagonally in the stalls If a 28 loop is used with forward lunge, width is 28 lunge not so important. Next slide is a small Jersey springer in a 46 wide stall that is too wide for her, but because of the loop she is lying parallel to the dividers
Once bedding drops below the curb, effective bed length becomes 8-10 inches shorter which is unacceptable to the cow.
Nose to Nose
68 (172cm) for Cows
46-48 68
Why are we making them longer? Some University recommendations are for16-18 apart. Just raises cost of building without increasing milk. Feeding cows do NOT like to be closer than 3 meters Lying cows not so much !!
Feed
Alley
Alley
44
Get it correct!!
And it looks like this!
Get it wrong!!
And it looks like this!
45
Notes to Consultants
Consultants to my Dairy Do not care what you know! Until I know that you care!
Consultants
Is today normal ? Without the usual excuses
Consultants
ID the problem Is it the bottleneck? What are the solutions? Whats the return and how fast? What are the expectations of success?
Consultants
Take ownership of the solution Do NOT make general suggestions
46
INTRODUCTION Generally speaking, dairy producers across North America are proud of their ability to produce high quality milk for daily consump on and for manufacturing of a wide variety of dairy products. High quality milk standards are, in part, maintained through the implementa on of regula ons that focus on the monitoring of milk levels of bacteria, inhibitors, and somatic cells. However, considerable variability exists between various regions on the levels established and enforced through these regula ons. In addi on, many regions use incen ve programs established by local milk coopera ves to encourage par cipa on in udder health management and mas s control programs. In Ontario, Canada, a 6 year reduc on program in bulk tank soma c cell counts (SCC) was implemented in 1989, resul ng in a regulatory framework with an upper limit of 500,000 cells/ml. This program became stable in August 1995, and has been maintained un l July 2012 (OMAFRA, 1990). Dairy markets worldwide have set various regulatory upper limits for bulk tank soma c cell count (BTSCC). For example, the European Union, Australia, New Zealand, Norway and Switzerland are 400,000 cells/ml. The United States has an upper regulatory limit of 750,000 cells/ml, except California, which has a limit of 600,000 cells/ml. Ontario has recently followed the European Communi es, and on August 1 of 2012 the regulatory limit was lowered to 400,000 cells/ml. HISTORICAL PERSPECTIVES OF SCC REDUCTION IN DAIRY HERDS Mastitis control has been an important dairy health management ini a ve for nearly 50 years. Early lessons learned in this area led to the development of a Na onal Mastitis Council (NMC) 5-point mastitis control plan. Subsequently, major extension eorts have been carried out throughout the dairy industry (Rados ts et al., 1994). As a result, the prac ces of post-milking teat disinfec on and blanket applica on of long-ac ng an bio c therapy at drying-o have been widely implemented. In addi on,
milking machine func on, early detec on and treatment of clinical cases, as well as iden ca on and culling of chronically infected cows have been emphasized. Star ng about 35 years ago, the use of SCC, as an indirect indicator of intramammary infec on (IMI), became the standard method of monitoring bulk milk quality (Harmon, 1994). Furthermore, Dairy Herd Improvement (DHI) milk recording agencies began to oer the service of measuring SCC in individual cow composite milk samples on a monthly basis. Between these two sources of SCC data, dairy producers had the tools to accurately monitor their progress in the control of IMI, resul ng in many lessons learned and considerable progress being made. The European Economic Community, as well as other countries such as Canada, developed and implemented SCC penalty programs to force a considerable economic incen ve for implemen ng the 5-point plan to bring herd level SCC under control. The industry dynamics, and the success associated with implementa on of these programs, are well both documented (Sargeant et al., 1998). Several other countries, including the United States, have not emphasized aggressive SCC regulatory limits. In contrast, incen ves for low SCC and other measures of milk quality have been oered at the local milk marke ng level. The vast majority of dairy producers are clearly mo vated by such premium incen ves. Data from DHI centers would suggest that outstanding udder health and milk quality is being achieved by much of the U.S. dairy industry. However, the overall success of these programs appears to be regional, and formidable challenges to achieve excellent mas s control s ll exist on many farms, par cularly in some regions. As the NMC mas s control programs became widely adopted by the late 1980s, herds in many of the major dairying regions of the world experienced changes in the expression of IMI in their cows. New IMI throughout the dry and periparturient periods and clinical mas s in early lacta on became increasingly important (Smith et al., 1985). Valuable lessons have been learned about
47
these changing dynamics of infec on, as well as about the preven on of both clinical and subclinical mas s. These lessons have required the development and use of new nomenclature. Mas s-causing bacteria have been classied into contagious and environmental pathogens, based upon important reservoirs and modes of transmission of each group. Changes in herd size, housing systems, and manure management have markedly increased the importance of environmental pathogens in the North American dairy industry. Also, lessons learned through epidemiological studies have had a major impact on our understanding of the risk factors and biological interacons that result in an increase of environmental IMI. In addi on, studies conducted in several countries have elucidated the incidence and risk factors involved with clinical mas s on commercial dairy farms (Hogan et al., 1989). Nevertheless, signicant challenges for the preven on of new IMI, primarily by environmental pathogens, remain an important focus of the dairy industry. In response to this situa on, NMC has developed, published, and extended a revised Recommended Mas s Control Program, which includes 10 major points of emphasis (Appendix 1). The new aspects of this program are largely focussed on mee ng the challenges presented by environmental pathogens. In other words, considerable emphasis has been placed on dierent aspects of mas s control. For example, the management of milking, housing environment, and the dry period have become extremely important. Pre-milking udder prepara on greatly inuences the preven on of new IMI and maintenance of milk quality (Pankey, 1989). In some regions, the prac ce of premilking teat disinfec on has become a standard prac ce. Interestingly, in other regions of the world such as Scandinavia, in response to public pressure, pre-milking teat disinfec on is illegal. Methods to improve and monitor s mula on of milk letdown, milk-out, and machine removal have also been widely implemented. The design and func on of milking equipment and milking techniques to op mize cleanliness, teat health, and milk ow have all been greatly advanced in last three decades (Spencer, 1989). Appropriate installa on, use, and evalua on of func on of modern milking equipment con nue to be extremely important. There have been important lessons learned in our understanding of the role of the dry period on new IMI, and on subsequent clinical mas s, in early lacta on (Dingwell et al., 2003). Studies conducted in New Zealand, North America, and the United Kingdom have described the importance of management at drying-o, and the occurrence of teat-end closure in the dry period, as important
factors for the preven on of infec on. Long-ac ng dry cow an bio c therapy has received renewed emphasis for the preven on of new IMI, in addi on to its tradi onal role for the elimina on of exis ng infec ons. However, methods to prevent new IMI in the dry period by protec on of the teat-end with external or internal teat sealant products have been developed, evaluated, and implemented (Sanford et al., 2006). The importance of the dry period for mas s control, and for general health, remains as the focus of intense research and development. The design of free-stall housing systems has started to receive great emphasis. The selec on of stall surface and bedding material is known to have a major impact on the rate of new IMI by environmental pathogens (Smith et al., 1985). Considerable evidence from lessons learned points to sand as a preferred bedding material for both cow comfort and udder health. Thus, sand has been widely implemented, notwithstanding the major impact that it has on manure management. On-going challenges from environmental legisla on and public pressure have resulted in the implementa on of numerous sustainable technologies, such as anaerobic manure digesters. With these technologies comes the mo va on to use digested manure solids as bedding materials. However, there is the poten al for many challenges and issues to maintain acceptable rates of clinical mas s and low BTSCC with the use of this bedding material. Therapy of clinical mas s con nues to be a challenging topic of considerable study and debate. No consistent approach to the early iden ca on and treatment of clinical mas s cases has been developed, even though numerous lacta ng intramammary and systemic an bio c preparaons are commercially available, and their ecacies against the common mas s pathogens have been well-documented (Guterbock et al., 1993). The risk and fear of an bio c residues in milk, combined with a reasonably high rate of relapse and recurrent clinical signs, are valid reasons for concern. Thus, it is highly recommended that the management and advisory team for each farm should develop a farm-specic clinical mas s treatment protocol that recognizes the predominant pathogens involved and monitors the success of the protocol. The usefulness of an bio c treatment of subclinical IMI during lacta on remains as a subject of considerable debate. This subject merits further study. Streptococcus agalac ae has been greatly reduced. Conversely, Staphylococcus aureus remains a signicant challenge, with very poor an bio c treatment cure rates during lacta on, and only moderate rates of cure with dry period therapy. However, the use of advanced epidemiolog-
48
ical analy cal tools has elucidated well-dened risk factors for failure of cure, allowing prospec ve decision-making on cows with par cular chronic infec ons (Sol et al., 2000). In addi on, the importance of pre-partum mas s in heifers as an important aspect of mas s control, especially for Staph. aureus, has been well-documented (Oliver et al., 1992; Nickerson et al., 1995). Ra onal use of an bio cs has become a major focus, and appropriate targeted treatment strategies remain extremely important in mas s control programs. Recently, a empts to establish and validate onfarm clinical mas s milk bacteriological culture systems have been documented, and show considerable promise (Lago et al., 2011). On-going challenges to maintain and enhance these programs remain. Throughout the last 30 years, a empts to develop meaningful vaccina on programs to prevent mas s in dairy ca le have been described. In general, the progress in this area is disappoin ng. However, the development and widespread implementa on of core-an gen coliform mas s vaccines has clearly reduced the incidence of severe clinical mas s, and has considerably reduced mortality due to coliform mas s (Gonzalez et al., 1989; Hogan et al., 1992). The development of record systems and monitoring tools has been par cularly impressive, and certainly important for mas s control. Clinical mas s recording systems and automated incorpora on and analysis of individual cow SCC data have allowed dairy producers and their advisors to make decisions and evaluate their success on an ongoing basis. There is a need for greater incorpora on of informa on on the microbiological iden ca on of the bacterial agents involved. The prospects for substan al improvements in this area are encouraging. Recently, new diagnos c techniques such as DNA ngerprin ng of mas s pathogens have helped to be er describe the epidemiology of mas s, factors associated with virulence, and proles determining the source and chronicity of infec ons. These tools have greatly increased our ability to follow infec ons, and in turn have challenged exis ng paradigms about the behavior of environmental and contagious organisms. For example, situa ons where Streptococci and E. coli have become persistent infec ons have been documented. Further studies are needed to con nue to build our knowledge to combat the ever-diversifying nature of organisms that cause mas s. In addi on, we need to pursue a more complex and condi onal understanding of the interac ons of the cow, bacteria, and the environment for mas s control.
DEVELOPING A PROGRAM TO ACHIEVE UDDER HEALTH WITH LOW SCC It is clear that management programs aimed at the reducon and maintenance of low bulk tank soma c cell counts (BTSCC) need to be a major focus for all dairy producers. In this pursuit, it is useful learn from the experiences of other regions that have implemented lower regulatory limits. With the introduc on of the SCC Penalty Program in Ontario, Schukken et al. (1992a) found that udder health management prac ces associated with reducing BTSCC were also closely associated with the improvement of other milk quality indicators, such as microbial inhibitor scores and freezing point devia ons. In a further publicaon, Schukken et al. (1992b) suggest that dairy industries must encourage producers with low BTSCC to maintain those low levels, which is as important as decreasing BTSCC on farms with high BTSCC. Recent studies have shown that herds enrolled in DHI programs, with access to cow-level SCC records are better equipped to manage BTSCC, having lower SCC penalty rates than those herds not enrolled on DHI programs (Hand et al., 2012). In Ontario, par cipa on in DHI has steadily increased, with 75% of dairy herds par cipa ng in 2010 (CanWest DHI, 2010). To assess the benet of DHI par cipa on and thus having access to cow-level SCC, Hand et al. (2012) calculated the odds of a BTSCC (calculated as the weighted average of 4 observa ons per month) exceeding penalty thresholds, as well as the likelihood of incurring a BTSCC penalty (3 out of 4 consecu ve monthly BTSCC exceed penalty thresholds, as per the Ontario Milk Act; OMAFRA, 1990) were calculated for herds enrolled in a DHI program compared with those that were not enrolled. Hand et al. (2012) found that DHI herds were 30% less likely to exceed BTSCC penalty thresholds at a regulatory limit of 500,000 cells/ml and 28% less likely to exceed BTSCC penalty thresholds at a regulatory limit of 400,000 cells/ml. As described in considerable detail earlier in this paper, there are well-established, recommended procedures for mas s control in dairy herds. Each herd should consider implemen ng all of the steps described in the NMC Recommended Mas s Control Program as presented in Appendix 1. Certainly, there are individual items under each of the steps that will not be necessary or useful in a par cular herd. It is suggested that Appendix 1 should be printed and reviewed, and each individual procedure that is not implemented should be recorded and checked for future reference, in order to determine if it should be implemented for a meaningful improvement in the future.
49
MAINTAINING A HIGHLY EFFECTIVE MASTITIS CONTROL PROGRAM A er the herd management team, in conjunc on with udder health and milk quality advisors, have established a meaningful mas s control program, it will be an ongoing challenge to ne-tune and maintain the program over me. As with most aspects of dairy herd management, udder health is a very dynamic and responsive biological system. As environmental factors and management structures change, so will mas s-causing organism populaons and other resistance factors. As such, it is important that the herd management team implement systems to maintain a low bulk milk SCC. To do so, there are seven well-documented habits of a highly eec ve mas s control program. These habits are as follows: 1. Set realis c herd goals - Each herd needs to establish goals for both bulk milk SCC and rates of clinical mass. These goals should include stated alarm points, a list of priori es, investments that need to be made and a commitment to achieving the goals. 2. Iden fy cows with high soma c cell counts - Using DHI SCC data and other informa on sources, cows with (and generally at risk for) high SCC need to be iden ed. Generally speaking, these animals include cows a er calving, a er clinical mas s, a er repeated elevated monthly SCC and those with known chronic infec ons. 3. Analyze your herd data to iden fy the strong and weak points in your herd On a regular basis, at least every 2 to 3 months, herd udder health data should be analyzed to assess the situa on. These data need to include individual cow SCC pa erns, clinical mas s pa erns, and the on-going infec on prole of the herd. 4. Perform a risk assessment of your current mas s control management - It is very useful to use a validated risk assessment protocol for mas s control on a regular basis. This process will help to iden fy and rank known mas s risk factors at work in the herd. 5. Improve the weakest links in your mas s control management - Using the analysis of the herd data and the results of the risk assessment, the herd management team should iden fy and establish a plan to a ack the weakest links in the mas s control program. 6. Develop adequate monitoring procedures U liza on of individual cow SCC data, culture results from clinical mas s cases and regular bulk tank monitoring, as well as rou ne herd data like clinical mas s and reasons for culling, is absolutely paramount in maintaining a highly eec ve mas s control program.
7. Strive for con nuous improvement of milk quality and mas s - All of the above highly eec ve habits need to be implemented con nuously on a rou ne basis. CONCLUSION In conclusion, the establishment and maintenance of a program to achieve superior udder health with an on-going low bulk milk SCC is a rela vely complex and con nuous task. That said, there is ample historical evidence and support that it can be achieved and maintained. In addi on, there are excellent resources readily available from NMC and other sources, as well as well-established highly effec ve habits to maintain success. REFERENCES
CanWest DHI. 2010. DHI par cipa on hits 75% milestone in Ontario [News release]. Guelph, Ontario. Dingwell, R. T., D. F. Kelton, and K. E. Leslie. 2003. Management of the dry cow in control of peripartum disease and mas s. Vet. Clin. North Am. Food Anim. Pract. 19:235265. Dingwell, R. T., K. E. Leslie, Y. H. Schukken, J. M. Sargeant, L. L. Timms, T. F. Dueld, G. P. Keefe, D. F. Kelton, K. D. Lissemore, and J. Conklin. 2004. Associa on of cow and quarter-level factors at drying-o with new intramammary infec ons during the dry period. Prev. Vet. Med. 63:7589. Dohoo, I. R., and K. E. Leslie. 1991. Evalua on of changes in soma c cell counts as indicators of new intramammary infec ons. J. Prev. Vet. Med. 10:225238. Gonzalez, R. N., J. S. Cullor, and D. E. Jasper. 1989. Preven on of clinical coliform mas s in dairy cows by a mutant Escherichia coli vaccine. Can. J. Vet. Res. 53:301305. Guterbock, W. M., A. L. Van Eenennaam, and R. J. Anderson. 1993. Ecacy of intramammary an bio c therapy for treatment of clinical mass caused by environmental pathogens. J. Dairy Sci. 76:34373444. Hand, K. J., A .M. Godkin, and D. F. Kelton. 2012. Bulk milk soma c cell penal es in herds enrolled in Dairy Herd Improvement programs. J. Dairy Sci. 95:240-242. Harmon, R. J. 1994. Physiology of mas s and factors aec ng soma c cell counts. J. Dairy Sci. 77:21032112. Hogan, J. S., K. L. Smith, and K. H. Hoblet. 1989. Field survey of clinical mas s in low soma c cell count herds. J. Dairy Sci. 72:15471556. Hogan, J.S., K. L. Smith, and D. A. Todhunter. 1992. Field trial to determine ecacy of an Escherichia coli J5 mas s vaccine. J. Dairy Sci. 72:15471553. Lago, A., S. M. Godden, R. Bey, P. L. Ruegg, and K. Leslie. 2011. The selecve treatment of clinical mas s based on on-farm culture results: I. Eects on an bio c use, milk withholding me, and short-term clinical and bacteriological outcomes. J. Dairy Sci. 94:44414456. Nickerson, S. C., W. E. Owens, and R. L. Boddie. 1995. Mas s in dairy heifers: Ini al studies on prevalence and control. J. Dairy Sci. 78:16071618. Oliver, S. P., M. J. Lewis, and B. E. Gillespie. 1992. Inuence of prepartum an bio c therapy on intramammary infec ons in primigravid heifers during early lacta on. J. Dairy Sci. 75:406414.
50
OMAFRA. 1990. The Milk Act and Regula ons of Ontario. Ontario Ministry of Agriculture, Food and Rural Aairs, Toronto, Ontario. Pankey, J. W. 1989. Premilking udder hygiene. J. Dairy Sci. 72:13081312. Rados ts, O. M., K. E. Leslie, and J. Fetrow. 1994. Herd Health Food Animal Produc on Medicine. 2nd ed. W. B. Saunders Co., Philadelphia, PA. Sanford, C. J., G. P. Keefe, I. R. Dohoo, K. E. Leslie, R. T. Dingwell, L. DesCteaux, and H. W. Barkema. 2006. Ecacy of using an internal teat sealer to prevent new intramammary infec ons in nonlacta ng dairy ca le. J. Am. Vet. Med. Assoc. 228:1565-1573. Sargeant, J. M., Y. H. Schukken, and K. E. Leslie. 1998. Ontario bulk milk soma c cell count reduc on program: Progress and outlook. J. Dairy Sci. 81:15451554. Schukken, Y. H., K. E. Leslie, A. J. Weersink, and S. W. Mar n. 1992a. Ontario bulk milk soma c cell count reduc on program. 1. Impact on soma c cell counts and milk quality. J. Dairy Sci. 75:3352-3358. Schukken, Y. H., K. E. Leslie, A. J. Weersink, and S. W. Mar n. 1992b. Ontario bulk milk soma c cell count reduc on program. 2. Dynamics of bulk milk soma c cell counts. J. Dairy Sci. 75:3359-3366. Smith, K. L., D. A. Todhunter, and P. S. Schoenberger. 1985. Environmental mas s: Cause, prevalence, and preven on. J. Dairy Sci. 68:15311553. Sol, J., O. C. Sampimon, H. W. Barkema, and Y. H. Schukken. 2000. Factors associated with cure a er therapy of clinical mas s caused by Staphylococcus aureus. J. Dairy Sci. 83:278284. Spencer, S. B. 1989. Recent research and development in machine milking: A review. J. Dairy Sci. 72:19071917.
51
Notes
52
Name / Farm: 1. Establishment of Goals for Udder Health Set realistic targets for average herd somatic cell count (SCC) or linear score and clinical mastitis rate. Review goals on a timely basis, with input from the Herd Udder Health Advisory Team (veterinarian, producer, herd manager, milking personnel and advisors). Prioritize management changes to achieve stated goals. Other: 2. Maintenance of a Clean, Dry, Comfortable Environment Ensure proper stall usage by ensuring adequacy of stall size and design. Maintain clean, dry, and comfortable stalls through appropriate bedding management. Keep cow lots or housing and traffic areas clean and dry. Ensure ventilation system is functioning properly. Ensure proper stocking density in facilities. Control detrimental environmental influences (heat stress, frostbite, stray voltage, insects etc.). Ensure that cows remain standing after milking (provide fresh feed and water). Other : 3. Proper Milking Procedures Examine foremilk to facilitate early detection of clinical mastitis and proper milk letdown. Apply pre-milking teat disinfectant that completely covers the teat skin and allow it to remain on teats for at least 30 seconds. Dry teats using a properly washed and disinfected cloth towel for use on one cow, or a single service paper towel. Wear clean gloves during the milking process to limit spread of contagious pathogens. Attach teat cups squarely and level with the udder within 90 seconds of udder preparation. Adjust cluster during milking to prevent liner slips and squawks. With manual removal, avoid machine stripping and shut off vacuum to the claw before removing cluster. Apply teat disinfectant immediately following teat cup removal, and assure complete coverage of teats.
Date:
Pre- and post-milking teat disinfectants should be selected based on documented efficacy data which can be found on the NMC website (www.nmconline.org) To optimize mastitis control and reduce costs, teat dipping is preferred to spraying as the method of disinfectant application. Milk cows with confirmed contagious intramammary infections last. Other : 4. Proper Maintenance and Use of Milking Equipment Install or update equipment to ISO 5707 (International Organization for Standardization, Milking machine installationsConstruction and performance). Service, maintain, and regularly evaluate equipment function according to manufacturers guidelines, using dynamic evaluation methods and an appropriate record form. Replace inflations and other rubber and plastic parts regularly, according to manufacturers guidelines. Replace broken or cracked inflations and short milk tubes immediately. Sanitize equipment prior to each milking and thoroughly wash and sanitize equipment after each milking. Other: 5. Good Record Keeping For each case of clinical mastitis, record cow identification, date detected, days in milk, quarter(s) affected, number and type of treatments, outcome of treatments (i.e. return to normal milk, time to discard milk) and the causative bacterial pathogen if a sample was cultured on-farm or in a laboratory. Use a computerized or manual record system to manage information, such as individual cow SCC data, on the prevalence and incidence of subclinical mastitis. Other : 6. Appropriate Management of Clinical Mastitis During Lactation Develop and implement a herd clinical mastitis treatment protocol with the Herd Udder Health Advisory team. Carefully consider the economic ramifications of therapy decisions.
53
Collect a pre-treatment milk sample aseptically for microbiological culture so that antimicrobial susceptibility tests can be used when appropriate. Use an appropriate therapeutic regimen; use drugs according to the protocol, or as recommended by the health advisors. Prior to infusion, disinfect the teat with a germicide and scrub the teat-end with an alcohol swab. For infusion of intramammary antibiotics, use a single-dose, regulatory approved product by the partial insertion method. Do not treat chronic non-responsive infections. Observe the correct withdrawal period for the antibiotic used, as stated on the label. If extra-label drug use is necessary, follow regulatory guidelines under the supervision of a veterinarian (i.e. in the systemic treatment of coliform mastitis). Always follow recommended drug storage guidelines and observe expiration dates. Clearly identify all treated cows, and record all treatments in a permanent record. Other : 7. Effective Dry Cow Management Decrease the energy density of the ration during late lactation to reduce milk production before dry-off. Dry cows off abruptly and dry treat each quarter immediately following the last milking of lactation. Disinfect teats and scrub the teat-end with an alcohol swab before infusing. Treat all quarters of all cows with a commercially available approved [long-acting] dry-cow antibiotic and/or an approved internal teat sealant. Use the partial insertion method of dry treatment infusion. Disinfect teats immediately following infusion using any approved post milking disinfectant teat dip. Provide adequate dry cow nutrition to enhance immune system function. Maintain a clean, dry, comfortable environment for dry cows. Dry cow environmental management is important to minimize exposure to pathogens. In situations of high environmental pathogen exposure, use an internal or external teat sealant for dry cows in addition to any antimicrobial product. In herds with coliform mastitis problems, vaccinate with a core antigen endotoxin vaccine following manufacturers directions. Clip flanks and udders to remove excess body hair. Udder singeing may be useful to ensure hair removal. Other :
8. Maintenance of Biosecurity for Contagious Pathogens and Marketing of Chronically Infected Cows Request bulk tank and individual cow SCC data. For suspect animals, further diagnostic efforts may be indicated to identify cases of subclinical mastitis prior to purchasing cows. If possible, obtain aseptically collected milk samples for bacteriological culture from cows prior to purchase. Isolate recently purchased cows, and milk separately, until there is assurance of the absence of intramammary infection. Segregate cows with a persistently high SCC or linear score (i.e. SCC greater than 200,000 or linear score greater than or equal to 4.0 for several months) and observe response to dry treatment or other recommended therapy. Market or permanently segregate cows persistently infected with Staphylococcus aureus or other non-responsive microbial agents (Mycoplasma, Nocardia, Pseudomonas, or Arcanobacterium pyogenes). Consider udder health status of first-calf heifers as this can impinge on herd biosecurity. Other : 9. Regular Monitoring of Udder Health Status Enroll in an individual cow SCC program or use some other monitor of subclinical infections. Use a sensitive cow-side monitor of inflammation in cows suspected of infection and in high risk periods (i.e. early lactation). Monitor distributions of high SCC cows, and rates of change to elevated SCC. Conduct milk bacteriological culture of clinical cases and high SCC cows regularly. Monitor udder health for the herd using reports from the regional regulatory agency or milk marketing organization and DHI. Calculate clinical mastitis rates and distributions regularly, paying particular attention to infections in heifers. Use SCC and clinical mastitis records to evaluate protocols, and to make treatment and marketing decisions. Other : 10. Periodic Review of Mastitis Control Program Obtain objective evaluations from veterinarian, industry field person or extension representative. A step-by-step approach to the review, and a standard evaluation form are useful. Make use of the entire Herd Udder Health Advisory Team: veterinarian, producer, herd manager, milking personnel, and advisors. Other :
54
GOT SHREDLAGE? There is much interest in a new way of chopping and processing corn silage that results in a nal product called shredlage (Ferrare o and Shaver, 2012). Converted choppers have the ability to shred, rather than chop, the corn plant at harvest. Interest in this process is due to the fact that the shredded material has more eec ve ber and has the poten al to be more diges ble. About 25 modied machines were in place this fall. Research on animal performance, packing density, processing scores, and sor ng is underway. ARE YOUR BUGS OKAY? In order for microbial inoculants to have a chance of being eec ve, they must supply the correct number of viable microorganisms when applied to the forage mass. Inoculants are commonly mixed in water for applica on to forage. Mulrooney and Kung (2008) found that some inoculants died rapidly if the water they were in reached high temperatures (>100F). In a recent eld study (Windle et al., 2012) we collected 29 samples of inoculants from inoculant tanks in the eld. Nine of 29 samples had numbers of lac c acid bacteria (LAB) that were lower than expected based on theore cal mixing rates. Of six samples that had water temperatures above 89F, four samples had lower numbers of LAB than expected. Two tanks had water temperatures exceeding 100F (111 and 116F). In those tanks, if applied at the suggested rate, the numbers of nal applied LAB would have been less than 3,000 and 200 cfu/g of fresh forage, respec vely (compared to a theore cal nal applica on rate of 100,000 cfu/g). On a posi ve note, a high percentage of samples had concentra ons of LAB that were higher than expected. Thus, small errors resul ng in under applica on of correct amounts of water probably s ll resulted in adequate numbers of LAB added to the forage. When applying a microbial inoculant, the temperature of the water must be monitored to insure it does not reach levels that would result in death of the bacteria.
EFFECT OF ENSILING TIME ON NUTRITIVE VALUE A growing body of evidence has shown that as corn silage ensiles, the poten al starch diges bility (starch-D) in the rumen increases rela ve to values in fresh forage (Stock et al., 1991; Newbold et al., 2006). This process is probably a result of proteolysis that aects the prolamin matrix interac ng with starch granules (Homan et al., 2011). Because the process is linked to proteolysis, the concentra on of soluble N also increases with me of storage (Der Bedrosian et al., 2012). In order to deal with the poten ally low starch-D in fresh corn plants, preliminary studies from our lab (Young et al., 2012) have shown that the addi on of exogenous proteases to fresh corn plants at ensiling increases in vitro ruminal starch-D faster than that naturally occurring in untreated silage. Research is on-going to determine if this approach is prac cal and cost eec ve. Although starch-D increases with me of ensiling, it does not appear that NDF-D improves with me in the silo, although more corrobora ng evidence is needed. GOT VOLATILES? Vola le organic compounds (VOC) add to global warming. Research has iden ed silages as the primary source of VOC from dairy farms with ethanol being a major contributor to the problem (Malkina et al., 2011). More emphasis is being placed on determining how to control the emissions of vola le organic compounds (VOC) from silages. Rule 4570 is currently in place, and California dairy farms are now required to implement mi ga ons that will lower VOC emissions from silage piles (h p://www.valleyair.org/ farmpermits/applica ons/4570/dairy_caf_phase_2.pdf). Most of the suggested mi ga ons can be characterized as good management prac ces. For example, minimizing the amount of loose silage on the bunker oor between feedings, packing silos ghtly, using a homolac c acid inoculant, and keeping a clean, at silo face are prac ces that will tend to reduce VOC emissions. Could legisla on like this come our way in the future?
55
EFFECTS OF HURRICANE IRENE ON CORN SILAGE FROM 2011 Hurricane Irene caused severe damage to the corn crop of 2011 in the Northeastern U.S. due to ooding and (or) lodging. Samples of damaged and undamaged corn silage were collected from farms and analyzed for microbial and chemical content. The most signicant ndings were higher numbers of yeasts (7.3 log10 cfu/g in damaged vs. 4.9 log10 cfu in undamaged samples) and higher concentra ons of ash in damaged samples (average of 8.93%, with a high of 28.2%) than in undamaged samples (average of 3.85%). In par cular, damaged samples had higher concentraons of iron (average of 2,477 ppm and as high as 12,534 ppm) and aluminum (average of 1,334 ppm and as high as 4,410 ppm). Concentra ons of heavy metals in silages were in general not aected by the hurricane. Overall, indices of silage fermenta on appeared normal, although damaged corn silage had slightly lower concentra ons of fermenta on acids but higher pH than undamaged silage. Deoxynivalenol was detected in 75% (6 out of 8) and 90% (27 out of 30) of undamaged and damaged corn samples, respec vely. Zearalenone was detected in 1 out of 8 (12.5%) control samples, while 7 out of 30 (23.3%) ooded corn samples contained this mycotoxin. Although some complaints about poor animal performance were recorded, it does not appear that there were wide spread nega ve eects from feeding the damaged crop. IN SEARCH OF MARKERS IN SILAGE THAT ARE NEGATIVELY CORRELATED WITH INTAKE AND/OR PRODUCTION Silage researchers have measured the same end products of fermenta on for years. For example, the nal pH, concentra ons of lac c acid, ace c acid, butyric acid, ethanol, and ammonia are the most common chemical components that have been measured. However, it is clear that the accumula on of these compounds do not fully explain the variances observed in intake and animal performance. Several recent publica ons have documented that in fact, silages contain hundreds of vola le compounds (Figueiredo et al., 2009; Malkina et al., 2011), and some may have detrimental eects on rumen microbes or the animal. Some work from Germany (Gerlach et al., 2012) suggests that ethyl lactate may be nega vely associated with DM intake. More work on the poten al of vola le compounds in silages to aect animal performance is on-going. WILD YEASTS IN SILAGES In the U.S. it is generally accepted that lactate-assimila ng yeasts ini ate aerobic spoilage in silage. Measurements
from commercial labs and from research studies have shown that the numbers of total yeasts in spoiling silages can be as high as 109 cfu/g of silage. It is unknown whether the consump on of high numbers of these spoilage yeasts nega vely impacts rumen microbes or the animal directly. Recently, Santos et al. (2011) reported that adding high concentra ons of a spoilage yeast isolated from silage to in vitro ruminal fermenta ons reduced the NDF-D of a TMR in a dose-dependent fashion. The fact that yeasts may have nega ve impacts on rumen fermenta on has also been reported by Chung et al. (2011), who found that the addi on of a yeast from an experimental direct fed microbial formula on resulted in a more acidic rumen when fed to dry cows. More research is needed to determine if consump on of high numbers of yeasts from spoiled silage have nega ve eects on animal performance. ACKNOWLEDGEMENTS We acknowledge the following companies and agencies that provided funding for the analyses on the hurricanedamaged corn: University of Vermont, University of Delaware, Prince Agri Products, BASF, Mycogen Seeds, Lallemand Animal Nutri on, Novus Interna onal, and Renaissance Nutri on. We also thank Ralph Ward and Cumberland Valley Analy cal for their assistance in sample analyses. REFERENCES
Chung, Y. H., N. D. Walker, S. M. McGinn, and K. A. Beauchemin. 2011. Diering eects of 2 ac ve dried yeast (Saccharomyces cerevisiae) strains on ruminal acidosis and methane produc on in nonlacta ng dairy cows. J. Dairy Sci. 94:24312439. Der Bedrosian, M. C., K. E. Nestor, Jr., and L. Kung, Jr. 2012. The eects of hybrid, maturity and length of storage on the composi on and nutri ve value of corn silage. J. Dairy Sci. 95:5115-5126. Ferrare o, L. F., and R. D. Shaver. 2012. Eect of corn shredlage on lacta on performance and total tract starch diges bility by dairy cows. Prof. Anim. Sci. In press. Figueiredo, R., A. I. Rodrigues, and M. do Ceu Costa. 2007. Vola le composi on of red clover (Trifolium pratense L.) forages in Portugal: The inuence of ripening stage and ensilage. Food Chem. 104:14451453. Gerlach, K., K. Weiss, F. Ross, W. Bscher, and K.-H. Sdekum. 2012. Changes in maize silage fermenta on products during aerobic deteriora on and its impact on feed intake by goats. Page 38. Proc. XVI Int. Silage Conf. University of Helsinki. Homan, P. C., N. M. Esser, R. D. Shaver, W. Coblentz, M. P. Sco , A. L. Bodnar, R. Schmidt, and B. Charley. 2011. Inuence of ensiling me and inocula on on altera on of the starch-protein matrix in highmoisture corn. J. Dairy Sci. 94:2465-2474. Malkina, I., A Kumar, P. G. Green, and F. M. Mitloehner. 2011. Iden ca on and quan ta on of vola le organic compounds emi ed from dairy silages and other feedstus. J. Environ. Qual. 40:2836. Mulrooney, C. N., and L. Kung, Jr. 2008. The eect of water temperature on the viability of silage inoculants. J. Dairy Sci. 91:236-240.
56
Newbold, J. R., E. A. Lewis, L. Lavrijssen, H. J. Brand, H. Vedder, and J. Bakker. 2006. Eect of storage me on ruminal starch degradability in corn silage. J. Dairy Sci. 89 (Suppl. 1):190. (Abstr.) Santos, M. C., A. L. Lock, G. D. Mechor, and L. Kung, Jr. 2011. Spoilage yeasts in silage have the poten al to directly impact rumen fermentaon. J. Dairy Sci. 94(Suppl. 1):207. (Abstr.) Stock, R. A., M. H. Sindt, R. M. Cleale, 4th, and R. A. Bri on. 1991. Highmoisture corn u liza on in nishing ca le. J. Anim. Sci. 69:1645-1656. Windle, M. C., C. Wacek-Driver, and L. Kung, Jr. 2012. A survey on the concentra ons of lac c acid bacteria, the pH and temperature of water in tanks for used for microbial inoculants. Study in progress. Young, K. M., J. M. Lim, M. C. Der Bedrosian, and L. Kung, Jr. 2012. Eect of exogenous protease enzymes on the fermenta on and nutri ve value of corn silage. J. Dairy Sci. In Press.
57
Notes
58
INTRODUCTION Grain management prac ces, such as grinding, (Theurer, 1986; Remond et al., 2004), steam aking (Callison et al., 2001), ensiling (Oba and Allen, 2003), or endosperm type selec on (Allen et al., 2008; Lopes et al., 2009), alter physical or chemical characteris cs of feed grains which in turn alter starch diges on and lacta on performance of dairy cows. Despite knowledge (Firkins et al., 2001) of physical and chemical factors that inuence feed grain u liza on by dairy cows, scien sts, nutri on consultants, and feed and forage tes ng laboratories have been challenged to integrate physical and chemical composi on of feed grains into feed grain evalua on systems. Recently, Homan et al. (2012) integrated mean par cle size (MPS) with chemical dierences associated with fermenta on and endosperm type into a common unit, en tled eec ve MPS (eMPS). Eec ve MPS of feed grains, which is conceptually similar to physically eec ve NDF (peNDF), has the poten al to integrate physical and chemical variances within dry and high moisture corn (HMC) into a single unit. Simply, eMPS considers the size of a feed grain par cle plus the degree of bonding (eec veness; strong or weak) within the par cle. The principal components required for determina on of eMPS and the adapta on of eMPS into an integrated feed grain evalua on system are described in this paper. INTEGRATED FEED GRAIN EVALUATION: PRINCIPAL COMPONENTS Mean Particle Size The eect of MPS on starch diges on in ruminants is well known (Firkins et al., 2001). Procedures to determine MPS of dry ground corn are also well dened (Baker and Herrman, 2002; ASAE, 2008). Approximately 150 g of dried, un-ground corn is placed in a series of 14 screens with nominal apertures of 4750, 3350, 2360, 1700, 1180, 850, 600, 425, 300, 212, 150, 106, 75, and 53 um and a pan. The series of screens and pan are placed on an oscilla ng sieve shaker for 15 minutes. Procedures to determine MPS of HMC are nebulous because neither ASAE (2008)
nor Baker and Herrman (2002) dene whether HMC is to be dried prior to sieving. The MPS of dry or HMC is o en determined on farm by dairy producers or nutri on consultants by sieving un-dried samples through a short stack of sieves (Baker and Herrman, 2002). Typically corns are sieved through 6 screens (4000, 2000, 1000, 500 and 250 um) and pan. Recent data (Homan et al., 2012) suggest on-farm determina on of MPS for dry corn is accurate, but on-farm determina on of MPS on un-dried, un-ground HMC causes a large bias, overes ma ng the MPS of HMC by approximately 500 um. Mean Particle SizeCorn Silage Quan fying the MPS of starch in corn silage with or without kernel processing in commercial forage tes ng systems has been challenging. As a commercially viable surrogate, Ferreira and Mertens (2005) recommended determining the starch content of the total sample and the starch content of the DM remaining on screens > 4.75 mm for corn silage. This yields a kernel processing score (KPS) that serves as a starch par cle size index for corn silage. Corn silage is dened as op mally processed if > 70% of starch passes through the 4.75-mm screen. Corn silage can also be dened as adequately (50 to 70%) or inadequately (< 50%) processed by the KPS scoring system. The MPS of corn par cles in corn silage can be es mated (Zwald et al., 2008) from KPS using the equa on (MPS, um = 7,491 + (93.81KPS + 0.37KPS2). Particle Bonding StrengthUnfermented (Dry) Corn Homan et al. (2012) observed that the rela onship between MPS of dry corn and starch degrada on rate could be improved when the MPS of dry corn was adjusted for par cle bonding strength. The best marker of par cle bonding strength in unfermented corns was determined to be prolamin proteins, which encapsulate starch in corn endosperm. Endosperm proteins (prolamin-zein) are not soluble in water (hydrophobic) nor in the innate rumen environment (Lawton, 2002). The nega ve rela onship between prolamin-zein and starch diges on in ruminants
59
has been previously dened (Lopes et al., 2009). The prolamin content of dry corns ranges from 2.5 to 5.5% of dry ma er. Corns > 4.5% prolamin as a % of DM are likely more vitreous, may contain more int genes, are short rela ve maturity hybrids (more int genes) and had adequate N fer lity. Corns with prolamin protein < 3.0% maybe unique opaque-oury hybrids or be grain from N-decient corn. A prolamin protein assay (Larson and Homan, 2008) is available at a number of commercial feed and forage tes ng laboratories. Particle Bonding Weakness Fermented (High Moisture) Corn Similar to dry corn, Hoffman et al. (2012) observed that the rela onship between MPS of HMC and starch degrada on rate was improved when the MPS of HMC was adjusted for par cle bonding weakness. Ammonianitrogen (NH3-N) was observed to be the best marker of par cle weakness in HMC. Ammonia-nitrogen Ammonia-nitrogen is an eec ve marker of par cle weakness in HMC because it is a chemical marker of proteolysis. Ensiling (Homan et al., 2011) has been shown to reduce all , and prolamin-zein subunits of the starch-protein matrix from 10 to 40%. The degrada on of the prolaminzein subunits of the starch-protein matrix of HMC can be more extensive with 60% reduc ons observed. Because prolamin-zeins are primarily responsible for cross-linking starch granules together, the degrada on of zeins in HMC results in starch granule clusters falling apart (weaker) as a result of fermenta on since the cross links holding starch granules together have been degraded. In extensively fermented HMC, NH3-N may represent > 6% of the total nitrogen (or protein). High-moisture corns with < 1% of the total nitrogen (or protein) as NH3-N indicate the degradaon of starch-matrix proteins is minimal. Commercial tests for NH3-N are widely available. Soluble Protein Soluble protein is also a marker of the degrada on of starch matrix proteins in HMC and corn silage. Der Bedrosian et al. (2012) observed weakening of kernel par cles within corn silage to be concomitant with increasing soluble protein, resul ng in greater in vitro starch diges on. Likewise Homan et al. (2011) observed increasing soluble protein content in HMC with advancing ensiling me with concomitant increases in vitro starch degrada on (Homan et al., 2010). Similar results have been previously reported by Benton et al. (2005). Prior to ensiling, about 20% of the protein
in corn is soluble in a buer solu on, but in extensively fermented HMC up to 70% of the protein may be soluble. AN INTEGRATED APPROACHFEEDGRAINV2.0 Concept As a result of current research (Oba and Allen, 2003; Benton et al., 2005; Taylor and Allen, 2005; Allen et al., 2008; Larson and Homan, 2008; Lopes et al., 2009; Homan et al., 2010, 2011, 2012; Der Bedrosian et al., 2012) on grain physico-chemistry and its eects on starch diges bility, a simple model was developed to predict rate of starch diges on, ruminal starch diges on, and total tract starch diges on. The model, FeedGrainV2.0, is publically available at h p://www.uwex.edu/ces/dairynutri on/. FeedGrainV2.0 was developed to provide a simple educa onal pla orm for dairy educators, consultants, and producers to evaluate feed grains using principal components which inuence feed grain diges on and dairy ca le performance. The following text and tables in this proceedings paper dene the general concepts employed in FeedGrainV2.0. Corn Types Redefined One of the major obstacles in developing an integrated feed grain evalua on system is removing empirical-user dened classica ons of corn as either dry or high moisture. There is confusion in the industry whether corns containing low moisture (18.0 to 24.0%) feed like dry corn or HMC. Likewise, there is confusion of whether HMC at harvest or fermented a few days feeds like dry corn or HMC. Data from Homan et al. (2010) and Benton et al. (2005) suggest fresh, unfermented HMC have similar starch degrada on characteris cs as a dry corn contemporary. As a result, dry corn and HMC are redened as unfermented and fermented corns using the concentra on of NH3-N in corn as an unbiased classicatory nutrient. Using NH3-N to dene corn as unfermented or fermented is simple, cost eec ve, unbiased, and accurate for the following reasons. First, dry corn or freshly harvested corn > 15% moisture does not contain any appreciable amount of NH3-N. Therefore, when corn is devoid of NH3-N the corn is not fermented. Second, recent research (Homan et al., 2011; Der Bedrosian et al., 2012) has demonstrated that NH3-N denes both the intensity and dura on of the ensiling process. Because NH3-N concentra ons in HMC increase due to intensity and length of fermenta on, the eects of me and intensity on starch diges bility can be accounted for. Finally, because NH3-N concentra on in HMC is rela vely easy to determine with ammonia probes or via near infrared reectance spectroscopy (NIRS), dening corn types, intensity, and length of fermenta on is rapid and economical.
60
FeedGrainV2.0 Calculations Calcula ons within FeedGrainV2.0 employ a mechanis c model with both physical and nutri onal chemistries of corns inuencing starch diges on parameters. FeedGrainV2.0 does not require measures of in vitro starch diges bility or in vitro gas produc on of feed grains. The following basic steps are used in FeedGrainV2.0 to calculate the outputs described below. A complete view of all calcula ons of FeedGrainV2.0 are detailed within the opera onal program which is available at h p://www. uwex.edu/ces/dairynutri on/. Feed grains are dried and MPS is determined using ASAE (2008) methods. The NH3-N concentra on of feed grains is determined. Because dry and fresh corns are devoid of NH3-N, corns with < 0.50% of NH3-N, % of total N are classied as unfermented. Corns with NH3-N concentra ons > 0.50 are classied as fermented.
High > 85 > 2250 > 73 > 10.0 > 6.0 > 4.4 > 11 > 2.1 > 4.5
Method Chemistry ASAE, 2008 NIRS or Chemistry NIRS or Chemistry NIRS or Chemistry Larson and Homan, 2008 NIRS or Chemistry NIRS or Chemistry NIRS or Chemistry
Or as % of total N.
Unfermented.
Low < 65 < 750 < 65 < 8.0 0.0 < 3.4 < 7.0 < 1.3 < 3.5
The prolamin concentra on of unfermented feed grains is determined. Prolamin concentra on in unfermented feed grains is then used to adjust MPS to eMPS. The MPS of unfermented feed grains with greater prolamin concentraons are considered to be more eec ve at resis ng bacterial diges on of starch, and as a result eMPS may be > MPS. Conversely, unfermented feed grains with lesser prolamin concentra ons are considered to be less eec ve at resis ng bacterial diges on of starch, and as a result eMPS may be < MPS (Figure 1). The NH3-N concentra on in fermented feed grains is used to adjust MPS to eMPS. The MPS of fermented feed grains with lesser NH3-N concentra ons are considered to be equally or marginally less eec ve at resis ng bacterial diges on of starch and as a result eMPS may be MPS. Conversely, fermented feed grains with greater NH3-N concentra ons are considered to be considerably less eec ve at resis ng bacterial diges on of starch and as a result eMPS will be < MPS (Figure 2). A frac onal rate of starch diges on is es mated from eMPS. Post-ruminal ow of starch is es mated assuming feed grain passage rates of 16.0 and 12.0% per hour for unfermented and fermented corns, respec vely. Ruminal starch diges bility is calculated by dierence (100 post-ruminal starch ow).
Figure 1. Rela onship between actual and eec ve mean par cle size (MPS) for a range of prolamin concentra ons. In FeedGrainV2.0 prolamin concentra on of unfermented feed grains is used to adjust MPS to eec ve MPS.
Post-ruminal starch diges bility is es mated assuming surface area of par cles presented post-ruminally is posi vely related to starch hydrolysis poten al.
61
Total tract starch diges bility (TTSD) is es mated as the sum of ruminal starch diges bility plus net post-ruminal diges bility. Es mated TTSD is used as the diges on coecient for starch in a summa ve equa on. Energy values for feed grains are calculated using a summa ve equa on (NRC, 2001) with independent
diges on coecients for starch and non-starch, nonber carbohydrates. A rela ve grain quality index is calculated from TTSD. Validation FeedGrainV2.0 was validated using published research trials involving lacta ng dairy cows as a guide. FeedGrainV2.0 was not evaluated for and is not intended for use for growing beef ca le, cereal grains, or steam aked corns. The valida on of FeedGrainV2.0 required some exibility in research literature interpreta on because of a lack or absence of physical and chemical measurement con nuity in the literature. Placing more rigid trial criteria into the valida on process would have been desirable but would have resulted in a very limited data base to conduct a valida on. In general, trials used to validate FeedGrainV2.0 fed > 80% of starch from grain, reported MPS, made direct comparisons of grain type and reported in vivo TTSD. Most trials did not report prolamin values and a value of 3.9% of DM was used for all trials involving unfermented corn unless otherwise dened. An average NH3-N concentraon of 4.0% was used for all trials involving fermented corn. In vivo TTSD for unfermented corns were adjusted for random study eects (St-Pierre, 2001) but insucient data was available for fermented feeds and TTSD were adjusted for an average MPS slope eect. Finally, in some
Units Low Average High
Figure 2. Rela onship between actual and eec ve mean par cle size (MPS) for a range of ammonia-nitrogen concentraons. In FeedGrainV2.0 ammonia-nitrogen concentra on of fermented feed grains is used to adjust MPS to eec ve MPS.
% as fed eMPS kd RSD TTSD NFC microns % per hour % of starch % of starch % of DM % of DM TDN NEL NEM NEG ME RGQ % of DM Mcals/lb Mcals/lb Mcals/lb Mcals/lb
< 15 < 600 < 13 < 50 < 89 < 77 < 8.0 < 86.5 < 0.86 < 0.93 < 0.63 < 1.35 < 130
> 35 > 2400 > 23 > 70.0 > 95 > 83 > 10 > 89.0 > 0.92 > 0.97 > 0.67 > 1.40 > 170
Starch Fermenta on Rate (As Fed)2 Ruminal Starch Diges bility Starch Diges bility (Total Tract) Non Fiber Carbohydrate Non-starch NFC Total Diges ble Nutrients, 1X Net Energy Lacta on, 3X Net Energy Maintenance Net Energy Gain Metabolizable Energy, 3X Rela ve Grain Quality
1 2
Starch within par cles is es mated to eec vely ferment at this compara ve mean par cle size. Es mated ruminal starch fermenta on rate of the grain in its original form as fed to dairy ca le. Translated from in vitro gas produc on rates of un-dried, un-ground dry and high moisture corns to ruminal passage rates of 16.0 and 12.0% per hour for unfermented and fermented corns, respec vely.
62
Capstone literature used to establish relationships between observed and predicted in vivo ruminal and total tract starch digestibility for FeedGrainV2.01,2
Author(s) Cita on Grain Type Processing Moisture MPS, Prolamin, NH3-N, In vivo Ruminal In vivo TT SD um % DM3 % of N4 TT SD SD (Study adj)5 FeedGrainV2.0 Predicted TT SD Ruminal SD
Ground Ground Ground Rolled Rolled Ground Rolled Ground Ground Ground Fine Grind Coarse Grind Fine Grind Coarse Grind Ground Cracked Ground Rolled Rolled Rolled Rolled Rolled Rolled Rolled Ground Ground Ground Ground Rolled Ground Ground Ground Ground Rolled Ground Rolled Ground Ground Ground Ground Ground Ground
15 15 24.7 24.7 30.0 30.0 30.9 30.9 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 26.9 15 15 15 15 32.0 32.0 15 25.8 30.0 30.0 15 15 15 15 15 36.7 15 15
800 1900 2220 3140 4430 1320 3780 1020 3280 1110 1200 2600 4800 1130 1650 827 3265 1250 1180 2450 1792 1394 1456 3200 3900 682 1292 1017 1540 4330 1660 1550 1600 489 1789 618 1725 730 1800 880 1860 1594 1377
3.9 3.9 . . . . . . 3.9 3.9 3.9 3.9 3.9 3.9 3.9 3.9 3.9 3.9 3.9 3.9 7.5 2.8 1.7 3.9 . 3.9 3.9 3.9 3.9 . . 3.9 . . . 3.9 3.9 5.0 5.0 3.9 . 5.0 2.0
. . 4.0 4.0 4.0 4.0 4.0 4.0 . . . . . . . . . . . . . . . . 4.0 . . . . 4.0 4.0 . 4.0 4.0 4.0 . . . . . 4.0 . .
96.3 94.1 92.4 87.2 85.5 90.2 84.1 91.8 80.4 88.1 98.0 92.2 91.3 96.1 93.6 92.2 85.6 87.3 95.8 87.4 89.6 95.1 96.6 88.2 93.4 92.4 86.4 90.2 85.1 94.2 98.8 93.1 97.4 98.2 95.7 88.9 76.4 91.4 86.0 93.6 94.6 91.7 96.3
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86.8 81.2 60.9 69.2 58.6 49.8 46.4 64.8 34.9 57.0
94.6 91.4 93.0 91.1 88.4 94.9 89.7 95.5 84.3 90.0 93.6 85.8 81.9 94.3 91.4 92.3 87.2 87.7 95.3 86.4 87.5 93.3 94.8 86.0 89.5 94.7 89.6 93.0 88.6 88.6 94.2 92.1 94.3 96.6 93.9 93.3 83.6 93.7 88.6 93.5 93.7 89.9 94.7
93.8 90.0 92.0 91.6 89.2 95.3 90.4 96.0 85.3 92.6 92.3 87.7 80.1 92.6 90.8 93.7 85.3 92.2 92.4 88.2 87.1 92.5 93.3 85.4 92.3 94.2 92.0 92.9 91.1 89.2 94.5 91.1 94.6 97.5 94.2 94.4 90.6 93.6 89.3 93.5 94.1 90.1 93.1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80.0 62.5 59.8 47.4 56.4 44.7 56.0 62.0 46.4 54.7
2000 J. Dairy Sci. 83:28392848 HMC HMC HMC HMC Dry Dry
Callison et al6
Dry Dry
Dry Medium Grind Dhiman et al. Knowlton et al. 2002 J. Dairy Sci. 85:217226 1996 J. Dairy Sci. 79:557-564 Dry Dry Dry Dry Dry Yu et al. Lopes et al. 1998 J. Dairy Sci. 81:777783 2009 J. Dairy Sci. 92:4541-4548 Dry Dry Dry Dry Dry Krause and Combs* 2003 J. Dairy Sci. 86:1382-1397 Krause et al.7 2003 J. Dairy Sci. 86:1341-1353 Dry HMC Dry Dry Dry Dry Ekinci and Broderick 1997 J. Dairy Sci. 80:32983307 HMC HMC Krause et al.* Knowlton et al. 2002 J. Dairy Sci. 85:1936-1946 Dry HMC 1998 J. Dairy Sci. 81:19721984 HMC HMC Dry Dry Remond et al.8 Oba and Allen* Taylor and Allen9*
1 2
2004 J. Dairy Sci. 87:13891399 2003 J. Dairy Sci. 86:184-194 2005 J. Dairy Sci. 88:1413-1424
HMC = high moisture corn, MPS = mean par cle size, TT = total tract, SD = starch diges bility, DM = dry ma er, CP = crude protein. Trial Criteria 1) > 80 % of starch from grain, 2) MPS reported, 3) direct comparsion grain type, 4) in vivo TT SD measured. 3 An average prolamin value of 3.9% of DM was used for all trials involving unfermented corn unless otherwise dened. 4 An average NH3-N value of 4.0% was used for all trials involving fermented corn. 5 In vivo total tract starch diges bili es were adjusted for random study eects (St-Pierre, N.R., JDS:84-741-755). 6 Non structural carbohydrate diges bility was used as a surrogate for TT SD. 7 Pure starch was added to reduce MPS. 8 Inclusion of all treatments in experiments resulted in failure of mixed sta s cal models to converge, sugges ng outlier in vivo data. Prolamin values of 5.0% of DM u lized as the trial indicated semi-int corns were fed. 9 Prolamin values of 5.0 and 2.0 were u lized to represent vitreous (66%) and oury corns (0%) fed in the trial. * Treatments converged.
63
trials non-structural carbohydrate diges bility had to be used as a surrogate for TTSD. CONCLUSIONS The objec ve in developing an integrated feed grain evalua on system (FeedGrainV2.0) is to provide an educa onal pla orm for dairy educators, consultants, and producers to evaluate feed grains using the principal components which inuence feed grain diges on and dairy ca le performance. FeedGrainV2.0 is a public educa onal program and may be used for applica on or advanced study to improve feed grain u liza on in dairy ca le. REFERENCES
Allen, M. S., R. A. Longuski, and Y. Ying. 2008. Endosperm type of dry ground corn grain aects ruminal and total tract diges on of starch in lacta ng dairy cows. J. Dairy Sci. 91(Suppl.1):529. (Abstr.) American Society of Agricultural and Biological Engineers. 2008. Methods of determining and expressing neness of feed materials by sieving. ASAE S319.4. St Joseph, MO. Baker, S., and T. Herrman. 2002. Evalua ng par cle size: MF-2051. Kansas State University Agricultural Experiment Sta on and Coopera ve Extension Service. Manha an, KS. Benton, J. R., T. Klopfenstein, and G. E. Erickson. 2005. Eects of corn moisture and length of ensiling on dry ma er diges bility and rumen degradable protein. Nebraska Beef Ca le Reports: 31-33. Callison, S. L., J. L. Firkins, M. L. Eastridge, and B. L. Hull. 2001. Site of nutrient diges on by dairy cows fed corn of dierent par cle sizes or steam-rolled. J. Dairy Sci. 84:1458-1467. Der Bedrosian, M. C., K. E. Nestor Jr., and L. Kung Jr. 2012. The eects of hybrid, maturity, and length of storage on the compos on and nutri ve value of corn silage. J. Dairy Sci. 95:5115-5126. Dhiman, T. R., M. S. Zamain, I. S. MacQueen, and R. L. Boman. 2002. Inuence of corn processing and frequency of feeding on cow performance. J. Dairy Sci. 85:217226. Eastridge, M. L., A. H. Lefeld, A. M. Eilenfeld, P. N. Go , W. S. Bowen, and J. L. Firkins. 2011. Corn grain and liquid feed as nonber carbohydrate sources in diets for lacta ng dairy cows. J. Dairy Sci. 94: 3045-3053. Ekinci, C., and G. A. Broderick. 1997. Eect of processing high moisture ear corn on ruminal fermenta on and milk yield. J. Dairy Sci. 80:32983307. Ferreira, G., and D. R. Mertens. 2005. Chemical and physical characteris cs of corn silages and their eects on in vitro disappearance. J. Dairy Sci. 88:4414-4425. Firkins, J. L., M. L. Eastridge, N. R. St-Pierre, and S. M. No sger. 2001. Eects of grain variability and processing on starch u liza on by lacta ng dairy ca le. J. Anim. Sci. 79:E218-E238. Homan, P. C., D. R. Mertens, J. Larson, W. K. Coblentz, and R. D. Shaver. 2012. A query for eec ve mean par cle size of dry and high moisture corns. J. Dairy Sci. 95:3467-3477. Homan, P. C., N. M. Esser, R. D. Shaver, W. Coblentz, M. P. Sco , A. L. Bodnar, R. Schmidt, and B. Charley. 2010. Inuence of inocula on and storage me on in vitro gas produc on of high moisture corn. J. Dairy Sci. 93(Suppl. 1):725. (Abstr.)
Homan, P. C., N. M. Esser, R. D. Shaver, W. Coblentz, M. P. Sco , A. L. Bodnar, R. Schmidt, and B. Charley. 2011. Inuence of inocula on and storage me on altera on of the starch-protein matrix in high moisture corn. J. Dairy Sci. 94:2465-2474. Knowlton, K. F., B. P. Glenn, and R. A. Erdman. 1998. Performance, ruminal fermenta on, and site of starch diges on in early lacta on cows fed corn grain harvested and processed dierently. J. Dairy Sci. 81:1972-1984. Knowlton, K. F., M. S. Allen, and P. S. Erickson. 1996. Lasalocid and parcle size of corn grain for dairy cows in early lacta on. 1. Eect on performance, serum metabolites and nutrient diges bility. J. Dairy Sci. 79:557-564. Krause, K. M., and D. K. Combs. 2003. Eects of forage par cle size, forage source, and grain fermentability on performance and ruminal pH in midlacta on cows. J. Dairy Sci. 86:1382-1397. Krause, K. M., D. K. Combs, and K. A. Beauchemin. 2002. Eects of forage par cle size and grain fermentability in midlacta on cows. I. Milk produc on and diet diges bility. J. Dairy Sci. 85:1936-1946. Krause, K. M., D. K. Combs, and K. A. Beauchemin. 2003. Eects of increasing levels of rened cornstarch in the diet of lacta ng dairy cows on performance and ruminal pH. J. Dairy Sci. 86:1341-1353. Larson, J., and P. C. Homan. 2008. Technical Note: A method to quanfy prolamin proteins in corn which are nega vely related to starch diges bility in ruminants. J. Dairy Sci. 91: 4834-4839. Lawton, J. W. 2002. Zein: A history of processing and use. Cereal Chem. 79:1-18. Lopes, J. C., R. D. Shaver, P. C. Homan, M. S. Akins, and S. J. Ber cs. 2009. Type of corn endosperm inuences nutrient diges bility in lacta ng dairy cows. J. Dairy Sci. 92: 4541-4548. Na onal Research Council. 2001. Nutrient Requirements of Dairy Ca le. 7th rev ed. Natl. Acad. Sci., Washington, DC. Oba, M., and M. S. Allen. 2003. Eects of corn grain conserva on method on ruminal diges on kine cs for lacta ng dairy cows at two dietary starch concentra ons. J. Dairy Sci. 86:184-194. Reis, R. B., F. San Emeterio, D. K. Combs, L. D. Sa er, and H. N. Costa. 2001. Eects of corn par cle size and source on performance of lacta ng dairy cows fed direct cut grass legume forage. J. Dairy Sci. 84:429441. Remond, D., J. I. Cabrer-Estrada, M. Chapion, B. Chauveau, R. Coudure, and C. Poncet. 2004. Eect of corn par cle size on site and extent of starch diges on in lacta ng dairy cows. J. Dairy Sci. 87:1389-1399. San Emeterio, F., R. B. Reis, W. F. Campos, and L. D. Sa er. 2000. Eect of coarse or ne grinding on u liza on of dry or ensiled corn by lacta ng dairy cows. J. Dairy Sci. 83:28392848. St-Pierre, N. R. 2001. Integra ng quan ta ve ndings from mul ple studies using mixed model methodology. J. Dairy Sci. 84:741-755. Taylor, C. C., and M. S. Allen. 2005. Corn grain endosperm type and brown midrib 3 corn silage: Site of diges on and ruminal diges on kine cs in lacta ng cows. J. Dairy Sci. 88:1413-1424. Theurer, C. B. 1986. Grain processing eects on starch u liza on by ruminants. J. Anim. Sci. 63:1649-1662. Yu, P., J. T. Huber, F. A. P. Santos, J. M. Simas, and C. B. Theurer. 1998. Eects of ground, steam aked, and steam rolled corn grains on performance of lacta ng cows. J. Dairy Sci. 81:777-783. Zwald A., A. E. Dorshorst, P. C. Homan, L. M. Bauman, M. G. Bertram. 2008. Technical Note: A near infrared reectance spectroscopy technique to predict par cle size of starch within corn silage. J. Dairy Sci. 91: 2071-2076.
64
ABSTRACT Precision Dairy Farming is the use of technologies to measure physiological, behavioral, and produc on indicators on individual animals to improve management strategies and farm performance. Many Precision Dairy Farming technologies, including daily milk yield recording, milk component monitoring, pedometers, automa c temperature recording devices, milk conduc vity indicators, automa c estrus detec on monitors, and daily body weight measurements, are already being u lized by dairy producers. Other theore cal Precision Dairy Farming technologies have been proposed to measure jaw movements, ruminal pH, re cular contrac ons, heart rate, animal posi oning and ac vity, vaginal mucus electrical resistance, feeding behavior, lying behavior, odor, glucose, acous cs, progesterone, individual milk components, color (as an indicator of cleanliness), infrared udder surface temperatures, and respira on rates. The main objec ves of Precision Dairy Farming are maximizing individual animal poten al, early detec on of disease, and minimizing the use of medica on through preven ve health measures. Perceived benets of Precision Dairy Farming technologies include increased efficiency, reduced costs, improved product quality, minimized adverse environmental impacts, and improved animal health and well-being. Real me data used for monitoring animals may be incorporated into decision support systems designed to facilitate decision making for issues that require compila on of mul ple sources of data. Technologies for physiological monitoring of dairy cows have great poten al to supplement the observa onal ac vi es of skilled herdspersons, which is especially cri cal as more cows are managed by fewer skilled workers. Moreover, data provided by these technologies may be incorporated into gene c evalua ons for non-produc on traits aimed at improving animal health, well-being, and longevity. The economic implica ons of technology adopon must be explored further to increase adop on rates of Precision Dairy Farming technologies. Precision Dairy Farming may prove to be the next important technological breakthrough for the dairy industry.
INTRODUCTION Across the globe, the trend toward fewer, larger dairy opera ons con nues. Dairy opera ons today are characterized by narrower prot margins than in the past, largely because of reduced governmental involvement in regula ng agricultural commodity prices. Consequently, small changes in production or efficiency can have a major impact on protability. The resul ng compe on growth has intensied the drive for eciency resul ng in increased emphasis on business and nancial management. Furthermore, the decision making landscape for a dairy manager has changed drama cally with increased emphasis on consumer protec on, con nuous quality assurance, natural foods, pathogen-free food, zoono c disease transmission, reduc on of the use of medical treatments, and increased concern for the care of animals. These changing demographics reect a con nuing change in the way in which dairy opera ons are managed. In large part, many of these changes can be a ributed to tremendous technological progress in all facets of dairy farming, including gene cs, nutri on, reproduc on, disease control, and management. W. Nelson Philpot (2003) captured this change eec vely in describing modern dairy farms as technological marvels. Conceivably, the next technological marvel in the dairy industry may be in Precision Dairy Farming. WHAT IS PRECISION DAIRY FARMING? Precision Dairy Farming is the use of technologies to measure physiological, behavioral, and produc on indicators on individual animals to improve management strategies and farm performance. Many Precision Dairy Farming technologies, including daily milk yield recording, milk component monitoring (e.g. fat, protein, and SCC), pedometers, automa c temperature recording devices, milk conduc vity indicators, automa c estrus detec on monitors, and daily body weight measurements, are already being u lized by dairy producers. Eastwood et al. (2004) dened Precision Dairy Farming as the use of informa on technologies for assessment of ne-scale
65
animal and physical resource variability aimed at improved management strategies for op mizing economic, social, and environmental farm performance. Spilke and Fahr (2003) stated that Precision Dairy Farming, with specic emphasis on technologies for individual animal monitoring, aims for an ecologically and economically sustainable produc on of milk with secured quality, as well as a high degree of consumer and animal protec on. With Precision Dairy Farming, the trend toward group management may be reversed with focus returning to individual cows through the use of technologies (Schulze et al., 2007). Technologies included within Precision Dairy Farming range in complexity from daily milk yield recording to measurement of specic a ributes (e.g. fat content or progesterone) within milk at each milking. The main objec ves of Precision Dairy Farming are maximizing individual animal poten al, early detec on of disease, and minimizing the use of medica on through preven ve health measures. Precision Dairy Farming is inherently an interdisciplinary eld incorpora ng concepts of informa cs, biosta s cs, ethology, economics, animal breeding, animal husbandry, animal nutri on, and engineering (Spilke and Fahr, 2003). POTENTIAL BENEFITS OF PRECISION DAIRY FARMING Perceived benets of Precision Dairy Farming technologies include increased eciency, reduced costs, improved product quality, minimized adverse environmental impacts, and improved animal health and well-being. These technologies are likely to have the greatest impact in the areas of health, reproduc on, and quality control (de Mol, 2000). Realized benets from data summariza on and excep on reporting are an cipated to be higher for larger herds, where individual animal observa on is more challenging and less likely to occur (Lazarus et al., 1990). As dairy operaons con nue to increase in size, Precision Dairy Farming technologies become more feasible because of increased reliance on less skilled labor and the ability to take advantage of economies of size related to technology adop on. A Precision Dairy Farming technology allows dairy producers to make more mely and informed decisions, resul ng in be er produc vity and protability (van Asseldonk et al., 1999b). Real me data can be used for monitoring animals and crea ng excep on reports to iden fy meaningful devia ons. In many cases, dairy management and control ac vi es can be automated (Delorenzo and Thomas, 1996). Alterna vely, output from the system may provide a recommenda on for the manager to interpret (Pietersma et al., 1998). Informa on obtained from Precision Dairy Farming technologies is only useful if it is interpreted and
u lized eec vely in decision making. Integrated, computerized informa on systems are essen al for interpre ng the mass quan es of data obtained from Precision Dairy Farming technologies. This informa on may be incorporated into decision support systems designed to facilitate decision making for issues that require compila on of mul ple sources of data. Historically, dairy producers have used experience and judgment to iden fy outlying animals. While this skill is invaluable and can never be fully replaced with automated technologies, it is inherently awed by limita ons of human percep on of a cows condi on. O en, by the me an animal exhibits clinical signs of stress or illness, it is too late to intervene. These easily observable clinical symptoms are typically preceded by physiological responses evasive to the human eye (e.g. changes in temperature or heart rate). Thus, by iden fying changes in physiological parameters, a dairy manager may be able to intervene sooner. Technologies for physiological monitoring of dairy cows have great poten al to supplement the observa onal ac vi es of skilled herdspersons, which is especially cri cal as more cows are managed by fewer skilled workers (Hamrita et al., 1997). PRECISION DAIRY FARMING EXAMPLES The list of Precision Dairy Farming technologies used for animal status monitoring and management con nues to grow. Because of rapid development of new technologies and suppor ng applica ons, Precision Dairy Farming technologies are becoming more feasible. Many Precision Dairy Farming technologies including daily milk yield recording, milk component monitoring (e.g. fat, protein, and SCC), pedometers, automa c temperature recording devices, milk conduc vity indicators, automa c estrus detec on monitors, and daily body weight measurements are already being u lized by dairy producers. Despite its seemingly simplis c nature, the power of accurate milk weights should not be discounted in monitoring cows, as it is typically the rst factor that changes when a problem develops (Philpot, 2003). Other theore cal Precision Dairy Farming technologies have been proposed to measure jaw movements, ruminal pH, re cular contrac ons, heart rate, animal posi oning and ac vity, vaginal mucus electrical resistance, feeding behavior, lying behavior, odor, glucose, acous cs, progesterone, individual milk components, color (as an indicator of cleanliness), infrared udder surface temperatures, and respira on rates. Unfortunately, the development of technologies tends to be driven by availability of a technology, transferred from other industries in market expansion eorts, rather than by need. Rela ve
66
to some industries, the dairy industry is rela vely small, limi ng corporate willingness to invest extensively in development of technologies exclusive to dairy farms. Many Precision Dairy Farming technologies measure variables that could be measured manually, while others measure variables that could not have been obtained previously. ADOPTION OF PRECISION DAIRY FARMING TECHNOLOGIES Despite widespread availability, adop on of these technologies in the dairy industry has been rela vely slow thus far (Eleveld et al., 1992; Huirne et al., 1997; Gelb et al., 2001). In fact, agricultural adop on of on-farm so ware packages, as a whole, has been much lower than predicted (Rosskopf and Wagner, 2003). The majority of informa on management systems available and used by many dairy producers are underu lized. In prac cality, their use is o en limited to crea ng produc on tables, a en on lists, and working schedules (van Asseldonk, 1999). Perceived economic returns from inves ng in a new technology are likely the main factor inuencing Precision Dairy Farming technology adop on. Addi onal factors impac ng technology adop on include degree of impact on resources used in the produc on process, level of management needed to implement the technology, risk associated with the technology, producer goals and mo va ons, and having an interest in a specic technology (Dijkhuizen et al., 1997, van Asseldonk, 1999). Characteris cs of the primary decision maker that inuence technology adop on include age, level of formal educa on, learning style, goals, farm size, business complexity, percep ons of risk, type of produc on, ownership of a non-farm business, innova veness in produc on, overall expenditures on informa on, and use of the technology by peers and other family members. Eleveld et al. (1992) demonstrated that technology adop on is improved when the technology ts within the normal daily work pa erns of the personnel who will be using it. Farm opera ons with more specializa on of labor are more likely to successfully adopt informa on technology (Eleveld et al., 1992). The most progressive producers will adopt those new technologies that appear to be protable. When a proven technology is not adopted, the opera on observes a lost opportunity cost that may lead to a compe ve disadvantage (Galligan, 1999). INVESTMENT ANALYSIS OF PRECISION DAIRY FARMING TECHNOLOGIES Todays dairy manager is presented with a constant stream of new technologies to consider including new Precision Dairy Farming technologies. Galligan and Groenendaal (2001) suggested that the modern dairy producer can be
viewed as a manager of an investment por olio, where various investment opportuni es (products, management interven ons) must be selected and combined in a manner to provide a prot at a compe ve risk to alterna ve opportuni es. Further, dairy managers must consider both biological and economic considera ons simultaneously in their decisions. Tradi onally, investment decisions have been made using standard recommenda ons, rules of thumb, consultant advice, or intui on. Thus, more objec ve methods of investment analysis are needed (Verstegen et al., 1995). Adop on of sophis cated on-farm decision-making tools has been scant in the dairy industry to this point. Yet, the dairy industry remains a perfect applica on of decision science because: (1) it is characterized by considerable price, weather, and biological varia on and uncertainty, (2) technologies, such as those characteris c of Precision Dairy Farming, designed to collect data for decision making abound, and (3) the primary output, uid milk, is dicult to dieren ate, increasing the need for alterna ve means of business dieren a on. In Compe ng on Analy cs: The New Science of Winning, Davenport and Harris (2007) pose that in industries with similar technologies and products, high performance business processes are one of the only ways that businesses can dieren ate themselves. Investment analyses of informa on systems and technologies are common within the general business literature (Streeter and Hornbaker, 1993; Bannister and Remenyi, 2000; Ryan and Harrison, 2000; Lee and Bose, 2002). However, dairy-specic tools examining investment of Precision Dairy Farming technologies are limited (Carmi, 1992; Gelb, 1996; van Asseldonk, 1999), though investment analyses of other dairy technologies abound (Hyde and Engel, 2002). Empirical comparisons of technology before or a er adopon or between herds that have adopted a technology and control herds that have not adopted are expensive and biased by other, possibly herd-related dierences. As a result, the norma ve approach, using simula on modeling, predominates in decision support models in animal agriculture (Dijkhuizen et al., 1991). Inves ng in new agricultural technologies is all too o en a daun ng and complex task. First, the standard approach using the Net Present Value is o en misleading because it does not adequately account for the underlying uncertain es. Second, the incremental costs and benets of new technologies require complex interac ons of mul ple variables that are o en non-linear and not intui ve. The complexi es surrounding investment in Precision Dairy Farming technologies is one example of this type of complex decision.
67
Ward (1990) listed three benets to investment in technology: 1) subs tu ve, replacing human power with machine power, 2) complementary, improving produc vity and employee eec veness through new ways of accomplishing tasks, and 3) innova ve, obtaining a compe ve edge. In addi on to impacts on produc on, many technologies may also change milk composi on, reproduc ve eciency, and disease incidences (Galligan and Groenendaal, 2001). In an analysis of an investment opportunity at the dairy level, cash ows are generally uncertain because of biological variability or incomplete knowledge of the system (Galligan and Groenendaal, 2001). The impact that a Precision Dairy Farming technology has on produc ve and economic performance is dicult to examine because of the changing nature of the decision environment where investments are o en one- me investments but returns accrue over a longer period of me (Ward, 1990; Verstegen et al., 1995; van Asseldonk, 1999; van Asseldonk et al., 1999a,b). Further, benet streams resul ng from investment in a Precision Dairy Farming technology are highly dependent upon the users ability to understand and u lize the informa on provided by the new technology (Bannister and Remenyi, 2000). An economic analysis of the value of Precision Dairy Farming technologies requires considera on of the eect of adop on on both quality and meliness of decisions (Verstegen et al., 1995). Improvements associated with adop on of new Precision Dairy Farming technologies may increase prots directly through improved u liza on of data provided by the technology or indirectly through recommenda ons of consultants u lizing the new informa on (Tomaszewski et al., 1997). It is dicult, if not impossible to quan fy the economic value of personal welfare associated with a proposed change (e.g. free me or pres ge; O e and Chilonda, 2000). For example, it is nearly impossible to quan fy the sa sfac on of having a healthy herd, reduc on of animal suering, reduced human health risks, and environmental improvements (Huirne et al., 2003). Despite eorts to formalize the ra onal decision making analysis of investment in informa on technologies, many business execu ves ul mately make their investment decision based on gut feel or acts of faith (Silk, 1990; Bannister and Remenyi, 2000; Passam et al., 2003). Ul mately, decision making is and should be dependent upon both ra onal analysis and ins nct (Bannister and Remenyi, 2000). SIMULATION OF DAIRY FARMS Mayer et al. (1998) proposed that with the variety of management issues a dairy manager faces in an ever-changing environment (e.g. environmental, nancial, and biological), best management strategies cannot be veried and
validated with eld experiments. As a result, simula on is the only method of integra ng and es ma ng these eects (Mayer et al., 1998). Simula ons are mathema cal models designed to represent a system, such as a dairy farm, for use in decision-making. Simula on models are useful and cost-eec ve in research that requires complex scenarios involving a large number of variables with large groups of animals over a long period of me under a large range of condi ons (Bethard, 1997; Shalloo et al., 2004). The primary advantages of using mathema cal computer simula on models in evalua ng dairy produc on issues are the ability to control more variables within the model than with a eld trial and the reduced costs associated with this kind of eort (Skidmore, 1990; Shalloo et al., 2004). These economic models can also be useful in evalua ng alternaves where very li le real data is available yet (Dijkhuizen et al., 1995). Simula ng a system is par cularly useful when uncertain, complex feedback loops exist (e.g. disease aects produc on which then impacts other variables further back in the system) (Dijkhuizen et al., 1995). Models that represent system uncertainty, while eec vely using available informa on, provide more realis c insight than models that do not consider a range of responses (Benne , 1992; Passam et al., 2003). Simula on or other systemic methods are preferred to capture the complexity of a dairy system as they can evaluate mul ple biological and economic factors aec ng performance, including management, feeding, breeding, culling, and disease (Skidmore, 1990; Sorensen et al., 1992). Because the dairy system includes environmental, economic, and physical components, accoun ng for interac ons among components and tracing the eects of an interven on through the en re system are essenal (Cabrera et al., 2005). Simula on models are ideal for analyzing investment strategies because they can effec vely examine improvement in biological parameters based on farm-specic data rather than simple industry averages (Jalvingh, 1992; Dijkhuizen et al., 1995; Delorenzo and Thomas, 1996; van Asseldonk et al., 1999b; Gabler et al., 2000). Simula on of a farm can be accomplished by conduc ng two simula ons, one with and one without a proposed change or interven on and then comparing these simula ons to examine the impact on biological or economic parameters of interest (van Asseldonk, 1999). The output of a series of simula ons provides a range of results, more realis cally depic ng biological variability than simple models (Marsh et al., 1987). Risk and uncertainty are major considera ons within a dairy produc on system because of the random nature
68
of milk produc on, biology, disease, weather, input costs, and milk prices (Delorenzo and Thomas, 1996). This risk and uncertainty represents a major por on of the diculty and complexity of managing a dairy opera on (Huirne, 1990). Uncertainty must be considered in decision-making to avoid biased es mates and erroneous decisions (Kristensen and Jorgensen, 1998). Future costs and returns are always uncertain (Lien, 2003). Within precision agriculture, accurate representa on of risk associated with technology adop on is cri cal in the decision making process (Marra et al., 2003). When managers do not have sucient informa on to assess the risk outcomes of decisions, they use subjecve probabili es based on past experiences and their own judgment (Huirne, 1990). In most situa ons, decision makers are primarily concerned with the chances of the realized returns from an investment being less than predicted (Galligan et al., 1987). The ability of a model to reect real world condi ons increases with considera on of more variables (Jalvingh, 1992). Nevertheless, to ensure that the model remains prac cal and reasonable, only variables with the most inuence on the nal desired outcome should be entered into the model as random (Jalvingh, 1992; Lien, 2003). PURDUE/KENTUCKY RESEARCH MODEL Bewley et al. (2010b) developed a simula on model of a dairy farm to evaluate investments in precision dairy farming technologies by examining a series of random processes over a ten-year period. The model was designed to characterize the biological and economical complexi es of a dairy system within a par al budge ng framework by examining the cost and benet streams coinciding with investment in a Precision Dairy Farming technology. Although the model currently exists only in a research form, a secondary aim was to develop the model in a manner conducive to future u lity as a exible, farm-specic decision making tool. The basic model was constructed in Microso Excel 2007 (Microso , Sea le, WA). The @Risk 5.0 (Palisade Corpora on, Ithaca, NY) add-in for Excel was u lized to account for the random nature of key variables in a Monte Carlo simulaon. In Monte Carlo simula on, random drawings are extracted from distribu ons of mul ple random variables over repeated itera ons of a model to represent the impact of dierent combina ons of these variables on nancial or produc on metrics (Kristensen and Jorgensen, 1998). The basic structure of the model is depicted in Figure 1. The underlying behavior of the dairy system was represented using current knowledge of herd and cow management with
rela onships dened from exis ng literature. Historical prices for cri cal sources of revenues and expenses within the system were also incorporated as model inputs. The exibility of this model lies in the ability to change inputs describing the ini al herd characteris cs and the poten al impact of the technology. Individual users may change these inputs to match the condi ons observed on a specic farm. A er inputs are entered into the model, an extensive series of intermediate calcula ons are computed within 13 modules, each exis ng as a separate worksheet within the main Excel spreadsheet. Each module tracks changes over a 10-year period for its respec ve variables. Within these inter-connected modules (Figure 2), the impact of inputs, random variables, and technology-induced improvements are es mated over me using the underlying system behavior within the model. Results of calcula ons within 1
69
module o en aect calcula ons in other modules with mul ple feed-forward and feed-backward interdependencies. Each of these modules eventually results in a calcula on that will inuence the cost and revenue ows necessary for the par al budget analysis. Finally, the costs and revenues are u lized for the project analysis examining the net present value (NPV) and nancial feasibility of the project along with associated sensi vity analyses. Agricultural commodity markets are characterized by tremendous vola lity and, in many countries, this vola lity is increasing with reduced governmental price regula on. As a result, economic condi ons and the protability of investments can vary considerably depending on the prices paid for inputs and the prices received for outputs. Producers are o en cri cal of economic analyses that fail to account for this vola lity, by using a single value for cri cal prices, recognizing that the results of the analysis may be dierent with higher or lower milk prices, for example. In a simula on model, variability in prices can be accounted for by considering the random varia on of these variables. In this model, historical U.S. prices from 1971 to 2006 for milk, replacement heifers, alfalfa, corn, and soybeans were collected from the Understanding Dairy Markets website (Gould, 2007). Historical cull cow prices were dened using the USDA-Na onal Agricultural Sta s cs Service values for beef cows and cull dairy cows sold for slaughter (USDA-NASS, 2007). Base values for future prices (2007 to 2016) of milk, corn, soybeans, alfalfa, and cull cows were set using es mates from the Food and Agricultural Policy Research Ins tutes (FAPRI) U.S. and World Agricultural Outlook Report (FAPRI, 2007). Variaon in prices was considered within the simula on based on historical varia on. In this manner, the vola lity in key prices can be considered within a protability analysis. Although there is probably no direct way to account for the many decisions that ul mately impact the actual protability of an investment in a Precision Dairy Farming technology, this model includes a Best Management Prac ce Adherence Factor (BMPAF) to represent the poten al for observing the maximum benets from adop ng a technology. The BMPAF is a crude scale from 1 to 100% designed to represent the level of the farm management. At a value of 100%, the assump on is that the farm management is capable and likely to u lize the technology to its full poten al. Consequently, they would observe the maximum benet from the technology. On the other end of the spectrum, a value of 0% represents a scenario where farm management installs a technology without changing management to integrate the newly available data in ef-
forts to improve herd performance. In this case, the farm would not recognize any of the benets of the technology. Perhaps most importantly, sensi vity analyses allow the end user to evaluate the decision with knowledge of the role they play in its success. INVESTMENT ANALYSIS OF AUTOMATED BODY CONDITION SCORING To show how it can be used prac cally, this model was used for an investment analysis of automa c body condi on scores on dairy farms (Bewley et al., 2010a). Automated body condi on scoring (BCS) through extrac on of informa on from digital images has been demonstrated to be feasible, and commercial technologies are being developed (Bewley et al., 2008). The primary objec ve of this research was to iden fy the factors that inuence the poten al protability of inves ng in an automated BCS system. An expert opinion survey was conducted to provide es mates for poten al improvements associated with technology adop on. Benets of technology adop on were es mated through assessment of the impact of BCS on the incidence of ketosis, milk fever, and metri s, concep on rate at rst service, and energy eciency. For this research example, industry averages for produc on and nancial parameters, selected to represent condi ons for a U.S. dairy farm milking 1000 cows in 2007 were used. Further details of model inputs and assump ons may be obtained from the author. Net present value (NPV) was the metric used to assess the protability of the investment. The default discount rate of 8% was adjusted to 10% because this technology has not been marketed commercially; thus, the risk for early adopters of the technology is higher. The discount rate par ally accounts for this increased risk by requiring higher returns from the investment. The general rule of thumb is that a decision with a NPV greater than 0 is a go decision and a worthwhile investment for the business. The investment at the beginning of the project includes the purchase costs of the equipment needed to run the system in addi on to purchasing any other setup costs or purchases required to start the system. Recognizing that a simpler model ignores the uncertainty inherent in a dairy system, Monte Carlo simula on was conducted using the @Risk add-in. This type of simula on provides innite opportuni es for sensi vity analyses. Simula ons were run using 1000 itera ons in each simula on. Simulaons were run, using es mates provided by experts, for scenarios with li le to no improvement in the distribu on of BCS and with denite improvement.
70
PROFITABILITY ANALYSIS For the small likelihood of improvement simula on, 13.1% of simula on itera ons resulted in a posi ve NPV whereas this same number was 87.8% for the scenario with a denite improvement. In other words, using the model assump ons for an average 1000 cow U.S. dairy in 2007, inves ng in an automated BCS system was the right decision 13.1% or 87.8% of the me depending on the assump on of what would happen with BCS distribu on a er technology adop on. The individual decision makers level of risk aversion would then determine whether they should make the investment. Although this serves as an example of how this model could be used for an individual decision maker, this protability analysis should not be taken literally. In reality, an individual dairy producer would need to look at this decision using herd-specic variables to assess the investment poten al of the technology. The main take home message was that because results from the investment analysis were highly variable, this technology is certainly not a one size ts all technology that would prove benecial for all dairy producers. SENSITIVITY ANALYSES The primary objec ve of this research was to gain a better understanding of the factors that would inuence the protability of inves ng in an automated BCS system through sensi vity analysis. Sensi vity analysis, designed to evaluate the range of poten al responses, provides further insight into an investment analysis (van Asseldonk et al., 1999b). In sensi vity analyses, tornado diagrams visually portray the eect of either inputs or random variables on an output of interest. In a tornado diagram, the lengths of the bars are representa ve of the sensi vity of the output to each input. The tornado diagram is arranged with the most sensi ve input at the top progressing toward the least sensi ve input at the bo om. In this manner, it is easy to visualize and compare the rela ve importance of inputs to the nal results of the model. Improvements in reproduc ve performance had the largest inuence on revenues followed by energy eciency and then by disease reduc on. Random variables that had the most inuence on NPV were as follows: variable cost increases a er technology adop on; the odds ra os for ketosis and milk fever incidence and concep on rates at rst service associated with varying BCS ranges; uncertainty of the impact of ketosis, milk fever, and metri s on days open, unrealized milk, veterinary costs, labor, and discarded milk; and the change in the percent of cows with BCS at calving 3.25 before and a er technology adop on. Sca er plots of the most sensi ve random variables plo ed against NPV
along with correla on coecients demonstrate how random variables impact protability. In both simula ons, the random variable that had the strongest rela onship with NPV was the variable cost increase. Not surprisingly, as the variable costs per cow increased the NPV decreased in both simula ons (Figure 3). Thus, the value of an automated BCS system was highly dependent on the costs incurred to u lize the informa on provided by the system to alter nutri onal management for improved BCS proles.
Figure 3. Sca er plot of Net Present Value versus annual percentage increase in variable costs (for simula on using all expert opinions provided).
Finally, the results of any simula on model are highly dependent on the assumptions within the model. A one-way sensi vity analysis tornado diagram compares mul ple variables on the same graph. Essen ally, each input is varied (1 at a me) between feasible high and low values and the model is evaluated for the output at those levels holding all other inputs at their default levels. On the tornado diagram, for each input, the lower value is plo ed at the le end of the bar and the higher value at the right end of the bar (Clemen, 1996). Simula ons were run for high and low feasible values for 6 key inputs that may aect NPV. The tornado diagram for the 95th percen le NPV from the simula on with a small likelihood of improvement in BCS distribu on is presented in Figure 4. Herd size had the most inuence on NPV. The NPV was higher for the larger herd because the investment costs and benets were spread among more cows. The next most important variable was the BMPAF. Again, this result was not surprising and reiterates that one of the most important determinants of project success was what the producer actually does to manage the informa-
71
Figure 4. Tornado diagrams for inputs aec ng 95th percentile of Net Present Value for simula ons using the es mates of all survey respondents (BMPAF is the Best Management Prac ce Adherence Factor, RHA milk produc on is rolling herd average milk produc on in lbs).
the costs of the base technologies decrease making them more economically feasible for dairy farms. Because the bulk of research focused on Precision Dairy Farming technologies is conducted in research environments, care must be taken in trying to transfer these results directly to commercial se ngs. Field experiments or simula ons may need to be conducted to alleviate this issue. Because of the gap between the impact of Precision Dairy Farming technologies in research versus commercial se ngs, addi onal eort needs to be directed toward implementa on of management prac ces needed to fully u lize informaon provided by these technologies. To gain a be er understanding of technology adoption shortcomings, addi onal research needs to be undertaken to examine the adop on process for not only successful adop on of technology but also technology adop on failures. Before inves ng in a new technology, a formal investment analysis should be conducted to make sure that the technology is right for your farms needs. Examining decisions with a simula on model accounts for more of the risk and uncertainty characteris c of the dairy system. Given this risk and uncertainty, a stochas c simula on investment analysis will represent that there is uncertainty in the protability of some projects. Ul mately, the dairy managers level of risk aversion will determine whether or not he or she invests in a technology using the results from this type of analysis. Perhaps the most interes ng conclusion from our model case study was that the factors that had the most inuence on the protability investment in an automated BCS system were those related to what happens with the technology a er it has been purchased as indicated by the increase in variable costs needed for management changes and the management capacity of the farm. Decision support tools, such as this one, that are designed to inves gate dairy herd decisions at a systems level may help dairy producers make be er decisions. Precision dairy farming technologies provide tremendous opportuni es for improvements in individual animal management on dairy farms. In the future, Precision Dairy Farming technologies may change the way dairy herds are managed. REFERENCES
Bannister, F. and D. Remenyi. 2000. Acts of faith: ins nct, value, and IT investment decisions. J. Inf. Technol. 15:231-241. Benne , R. M. 1992. The use of economic quan ta ve modeling techniques in livestock health and disease-control decision making: a review. Prev. Vet. Med. 13:63-76. Bethard, G. L. 1997. A microcomputer simula on to evaluate strategies for rearing dairy replacements. Page 161. Vol. PhD Diss. Virginia Polytechnic Ins tute and State University, Blacksburg, VA.
on provided by the technology. There are many nutrional, health, reproduc ve, and environmental decisions made by the dairy producer that have a major impact on changes in body reserves for both individual cows and groups of cows. Management level plays a cri cal role in determining returns from inves ng in a Precision Dairy Farming technology. The level of management in day-today handling of individual cows may also inuence the impact of Precision Dairy Farming technologies. Van Asseldonk (1999) dened management capacity as having the appropriate personal characteris cs and skills to deal with the right problems and opportuni es in the right moment and in the right way. Eec ve use of an informa on system requires an investment in human capital in addi on to investment in the technology (Streeter and Hornbaker, 1993). Then, the level of milk produc on was the next most sensi ve input. As the level of milk producon increased, the benets of reducing disease incidence and calving intervals increased. As would be expected, the NPV increased with an increased base incidence of ketosis because the eects of BCS on ketosis would be exaggerated. The purchase price of the technology had a rela vely small impact on the NPV as did the base culling rate. CONCLUSIONS AND OUTLOOK Though Precision Dairy Farming is in its infancy, new Precision Dairy Farming technologies are introduced to the market each year. As new technologies are developed in other industries, engineers and animal scien sts nd applica ons within the dairy industry. More importantly, as these technologies are widely adopted in larger industries, such as the automobile or personal compu ng industries,
72
Bewley, J. M., M. D. Boehlje, A. W. Gray, H. Hogeveen, S. J. Kenyon, S. D. Eicher, and M. A. S. Russell, M.M. 2010a. Assessing the poten al value for an automated dairy ca le body condi on scoring system through stochas c simula on. Agric. Finance Rev. 70:126-150. Bewley, J. M., M. D. Boehlje, A. W. Gray, H. Hogeveen, S. J. Kenyon, S. D. Eicher, and M. A. S. Russell, M.M. 2010b. Stochas c simula on using @Risk for dairy business investment decisions. Agric. Finance Rev. 70:97-125. Bewley, J. M., A. M. Peacock, O. Lewis, R. E. Boyce, D. J. Roberts, M. P. Coey, S. J. Kenyon, and M. M. Schutz. 2008. Poten al for es ma on of body condi on scores in dairy ca le using digital images. J. Dairy Sci. 91:3439-3453. Cabrera, V. E., N. E. Breuer, P. E. Hildebrand, and D. Letson. 2005. The dynamic North Florida dairy farm model: A user-friendly computerized tool for increasing prots while minimizing N leaching under varying clima c condi ons. Comput. Electron. Agric. 49:286-308. Carmi, S. 1992. The performance of an automated dairy management data-gathering system. Pages 346-352 in Proc. Intl. Symp. Prospects for Automa c Milking. European Associa on for Animal Produc on, Wageningen, The Netherlands. Clemen, R. T. 1996. Making hard decisions: an introduc on to decision analysis. 2nd ed. Duxbury Press, Belmont, CA. Davenport, T. H., and J. G. Harris. 2007. Compe ng on analy cs: the new science of winning. Harvard Business School Press, Boston, MA. de Mol, R. M. 2000. Automated detec on of oestrus and mas s in dairy cows. Page 177. PhD Thesis. Wageningen University, Wageningen, The Netherlands. Delorenzo, M. A., and C. V. Thomas. 1996. Dairy records and models for economic and nancial planning. J. Dairy Sci. 79:337-345. Dijkhuizen, A. A., R. B. M. Huirne, S. B. Harsh, and R. W. Gardner. 1997. Economics of robot applica on. Comput. Electron. Agric. 17:111-121. Dijkhuizen, A. A., R. B. M. Huirne, and A. W. Jalvingh. 1995. Economic analysis of animal diseases and their control. Prev. Vet. Med. 25:135149. Dijkhuizen, A. A., J. A. Renkema, and J. Stelwagen. 1991. Modelling to support animal health control. Agric. Econ. 5:263-277. Eastwood, C., D. Chapman, and M. Paine. 2004. Precision dairy farmingtaking the microscope to dairy farm management. Eleveld, B., R. B. M. Huirne, A. A. Dijkhuizen, and G. Overbeek. 1992. Users in search of farm computer informa on technology: what do farmers want or need? Pages 27-32 in Farm Computer Technology in Search of Users: 4th Intl. Congress for Computer Technology in Agriculture, Paris-Versailles, France. FAPRI. 2007. Food and Agricultural Policy Research Ins tute 2007 U.S. and World Agricultural Outlook. Iowa State University and University of Missouri-Columbia. Ames, IA. Gabler, M. T., P. R. Tozer, and A. J. Heinrichs. 2000. Development of a cost analysis spreadsheet for calcula ng the costs to raise a replacement dairy heifer. J. Dairy Sci. 83:1104-1109. Galligan, D. T. 1999. The economics of op mal health and produc vity in the commercial dairy. Rev. Sci. Tech. O. Int. Epiz. 18:512-519. Galligan, D. T., and H. Groenendaal. 2001. Economic concepts in the valua on of products used in dairy produc on including a real op ons approach. Pages 233-245 in 36th Annual Pacic Northwest Anim. Nutr. Conf., Boise, ID.
Galligan, D. T., W. E. Marsh, and J. Madison. 1987. Economic decision making in veterinary prac ce: Expected value and risk as dual u lity scales. Prev. Vet. Med. 5:79-86. Gelb, E., C. Parker, P. Wagner, and K. Rosskopf. 2001. Why is the ICT adop on rate by farmers s ll so slow? Pages 40-48 in Proc. ICAST, Vol. VI, 2001, Beijing, China. Gelb, E. M. 1996. The economic value of informa on in an informa on system. Pages 142-145 in 6th Intl. Congress for Computer Technology in Agriculture Wageningen, The Netherlands. Gould, B. W. 2007. University of Wisconsin-Madison: Understanding Dairy Markets. Hamrita, T. K., S. K. Hamrita, G. Van Wicklen, M. Czarick, and M. P. Lacy. 1997. Use of biotelemetry in measurement of animal responses to environmental stressors. Huirne, R. 1990. Basic concepts of computerised support for farm management decisions. Euro. R. Agr. Eco. 17:69-84. Huirne, R. B. M., S. B. Harsh, and A. A. Dijkhuizen. 1997. Cri cal success factors and informa on needs on dairy farms: the farmers opinion. Livest. Prod. Sci. 48:229-238. Huirne, R. B. M., H. W. Saatkamp, and R. H. M. Bergevoet. 2003. Economic analysis of farm-level health problems in dairy ca le. Ca le Prac ce 11:227-236. Hyde, J., and P. Engel. 2002. Inves ng in a robo c milking system: a Monte Carlo simula on analysis. J. Dairy Sci. 85:2207-2214. Jalvingh, A. W. 1992. The possible role of exis ng models in on-farm decision support in dairy ca le and swine produc on. Livest. Prod. Sci. 31:351-365. Kristensen, A. R., and E. Jorgensen. 1998. Decision Support Models. Pages 145-163 in Proc. 25th Intl. Dairy Congress, Aarhus, Denmark. Lazarus, W. F., D. Streeter, and E. Jofre-Giraudo. 1990. Management informa on systems: impact on dairy farm protability. North Cent. J. Agric. Econ. 12:267-277. Lee, J., and U. Bose. 2002. Opera onal linkage between diverse dimensions of informa on technology investments and mul faceted aspects of a rms economic performance. J. Inf. Technol. 17:119-131. Lien, G. 2003. Assis ng whole-farm decision-making through stochas c budge ng. Agric. Sys. 76:399-413. Marra, M., D. J. Pannell, and A. Abadi Ghadim. 2003. The economics of risk, uncertainty and learning in the adop on of new agricultural technologies: where are we on the learning curve? Agric. Sys. 75:215-234. Marsh, W. E., A. A. Dijkhuizen, and R. S. Morris. 1987. An economic comparison of four culling decision rules for reproduc ve failure in United States dairy herds using DairyORACLE. J. Dairy Sci. 70:1274-1280. Mayer, D. G., J. A. Belward, and K. Burrage. 1998. Op mizing simula on models of agricultural systems. Ann. Oper. Res. 82:219-231. O e, M. J. and P. Chilonda. 2000. Animal health economics:an introducon. in Fron ers in Bioscience. FAO. Passam, H. C., A. Tocatlidou, B. D. Mahaman, and A. B. Sideridis. 2003. Methods for decision making with insucient knowledge in agriculture. Pages 727-731 in EFITA 2003 Conference, Debrecen, Hungary. Philpot, W. N. 2003. Role of technology in an evolving dairy industry. Pages 6-14 in 2003 Southeast Dairy Herd Management Conf., Macon, GA. Pietersma, D., R. Lacroix, and K. M. Wade. 1998. A framework for the development of computerized management and control systems for
73
use in dairy farming. J. Dairy Sci. 81:2962-2972. Rosskopf, K., and P. Wagner. 2003. Requirements for agricultural so ware and reasons for adop on constraints-results of empirical studies. Pages 651-657 in EFITA 2003 Conf., Debrecen, Hungary. Ryan, S. D., and D. Harrison. 2000. Considering social subsystem costs and benets in informa on technology investment decisions: A view from the eld on an cipated payos. J. Manage. Inf. Sys. 16:11-40. Schulze, C., J. Spilke, and W. Lehner. 2007. Data modeling for Precision Dairy Farming within the compe ve eld of opera onal and analy cal tasks. Comput. Electron. Agric. 59:39-55. Shalloo, L., P. Dillon, M. Rath, and M. Wallace. 2004. Description and valida on of the Moorepark Dairy System Model. J. Dairy Sci. 87:1945-1959. Silk, D. J. 1990. Managing IS benets for the 1990s. J. Inf. Technol. 5:185-193. Skidmore, A. L. 1990. Development of a simula on model to evaluate eec veness of dairy herd management. Page 236. PhD Diss. Cornell University, Ithaca, NY. Sorensen, J. T., E. S. Kristensen, and I. Thysen. 1992. A stochas c model simula ng the dairy herd on a PC. Agric. Sys. 39:177-200. Spilke, J., and R. Fahr. 2003. Decision support under the condi ons of automa c milking systems using mixed linear models as part of a precision dairy farming concept. Pages 780-785 in EFITA 2003 Conf., Debrecen, Hungary. Streeter, D. H., and R. H. Hornbaker. 1993. Value of informa on systems: Alterna ve viewpoints and illustra ons. Pages 283-293 in Farm level informa on systems, Zeist, The Netherlands. Tomaszewski, M. A., A. A. Dijkhuizen, A. G. Hengeveld, and H. Wilmink. 1997. A method to quan fy eects a ributable to management informa on systems in livestock farming. Pages 183-188 in First European Conf. for Informa on Technology in Agriculture, Copenhagen. USDA-NASS. 2007. Agricultural Prices Summary. van Asseldonk, M. A. P. M. 1999. Economic evalua on of informa on technology applica ons on dairy farms. Page 123. PhD Diss. Wageningen Agricultural Univ. van Asseldonk, M. A. P. M., R. B. M. Huirne, A. A. Dijkhuizen, and A. J. M. Beulens. 1999a. Dynamic programming to determine op mum investments in informa on technology on dairy farms. Agric. Sys. 62:17-28. van Asseldonk, M. A. P. M., A. W. Jalvingh, R. B. M. Huirne, and A. A. Dijkhuizen. 1999b. Poten al economic benets from changes in management via informa on technology applica ons on Dutch dairy farms: a simula on study. Livest. Prod. Sci. 60:33-44. Verstegen, J. A. A. M., R. B. M. Huirne, A. A. Dijkhuizen, and J. P. C. Kleijnen. 1995. Economic value of management informa on systems in agriculture: a review of evalua on approaches. Comput. Electron. Agric. 13:273-288. Ward, J. M. 1990. A por olio approach to evalua ng informa on systems investments and se ng priori es. J. Inf. Technol. 5:222-231.
74
INTRODUCTION Promo ng feed intake by lacta ng dairy cows, par cularly those in early lacta on, is cri cal for the improvement and maintenance of milk produc on, health, and body condi on (Grant and Albright, 2001). Gene c selec on prac ces have given rise to dairy ca le that are capable of producing quan es of milk in greater amounts than can be maintained by nutrient intake, par cularly during early lacta on. Past research in dairy ca le nutri on has focused almost exclusively on the nutri ve aspects of the diet, resul ng in many discoveries and improvements in dairy cow health and produc on. Despite many advances this eld we are s ll faced with the challenge of ensuring adequate dry ma er intake (DMI) to maximize produc on and prevent disease, par cularly in dairy cows during the transi on period. This paper will describe the importance of understanding behavior and how knowledge in this area of science can be used to evaluate nutri onal management and housing strategies. It is an cipated that with an improved understanding of feeding behavior, combined with the con nued eorts of nutri onists, dairy producers can manage and design their dairy produc on systems in ways that will allow their cows to fully maximize the poten al of the ra on provided, thereby improving the health, produc on, and welfare of their animals. IMPORTANCE OF FEEDING BEHAVIOR Over the years many researchers have evaluated the many physical, diges ve, and metabolic factors that contribute to the physiological regula on of DMI (see reviews by Allen, 2000; Ingvartsen and Andersen, 2000). The iden ca on of these factors is necessary for the formula on of ra ons that help maximize DMI, par cularly for those early lactaon cows whose energy demands o en cannot be met (Ingvartsen and Andersen, 2000). Since changes in DMI must ul mately be mediated by changes in feeding behavior (Nielsen, 1999), it is important to understand the factors that inuence cow feeding behavior pa erns. To date, the majority of research on dairy nutri on has largely ignored how the diet is consumed. Formula ng diets has tradi on-
ally required li le knowledge about how the diet is consumed; it was enough to simply es mate daily DMI without considering how and what feed was actually consumed. Modern, intensively-housed dairy ca le fed a conserved ra on typically consume their daily DMI in up to 6 hours per day, spread between 7 or more meals per day (DeVries et al., 2003). Management prac ces that cause adult dairy ca le to eat fewer and larger meals more quickly have been associated with an increased incidence of sub-acute ruminal acidosis (Krause and Oetzel, 2006). The reason for this risk is that ruminal pH declines following meals, and the rate of pH decline increases as meal size increases and as dietary effec ve ber concentra on decreases (Allen, 1997). Further, as cows spend less overall me feeding and increase their rate of feed consump on, daily salivary secre on is reduced (Beauchemin et al., 2008), decreasing the buering capacity of the rumen and reducing rumen pH. Alterna vely, when cows slow down their rate of DM consump on and have more frequent, smaller meals throughout the day, rumen buering is maximized, large within-day depressions in pH are avoided, and the risk of sub-acute ruminal acidosis is decreased. Thus, to maximize rumen health, eciency and produc vity, it is important to u lize feeding management strategies that promote the frequent consump on of feed in small meals throughout the day. Total mixed ra ons (TMR) are designed as a homogenous mixture with the goal to help minimize the selec ve consump on of individual feed components by dairy ca le, promote a steady-state condi on conducive to con nuous rumen func on and ingesta ow, and ensure adequate intakes of ber (Coppock et al., 1981). It is not surprising, therefore, that providing feed as a TMR has become the standard on most commercial dairies, par cularly for the lacta ng animals. Unfortunately, even when providing feed as a TMR cows have been shown to preferen ally select (sort) for the grain component of a TMR and discriminate against the longer forage components (Leonardi and Armentano, 2003). Sor ng of the diet can lead to the
75
cows consuming an inconsistent ra on, whereby the raon actually consumed by cows is greater in fermentable carbohydrates than intended and lesser in eec ve ber, and thus can increase the risk of sub-acute ruminal acidosis (DeVries et al., 2008). Likely related to this, in two recent studies it has been observed that such sor ng of a TMR is associated with producing milk with lower fat percentage (milk fat decreases by 0.15% for every 10% refusal of long forage par cles in the ra on; DeVries et al., 2011; Fish et al., 2012). Sor ng of the TMR can also reduce the nutri ve value of the TMR remaining in the feed bunk, par cularly in the later hours past the me of feed delivery (DeVries et al., 2005; Hosseinkhani et al., 2008). For group-fed cows, this may be detrimental for those cows that do not have access to feed at the me when it is delivered, for example when there is high compe on at the feed bunk. In such cases, these cows may not be able to maintain adequate nutrient intake to maintain high levels of milk produc on (Krause and Oetzel, 2006). Thus, feeding management strategies are required to ensure that cows consume the feed as it is formulated for them. It is clear that in addi on to proper formula ng of dairy raons we need to also consider how the ra on is consumed to ensure that the poten al of that ra on is maximized. There is an increasingly growing body of literature in which the knowledge of feeding behavior can be used to iden fy nutri onal management and housing strategies to maximize ra on poten al; including ensuring cows have access to fresh feed throughout the day and minimizing feed bunk compe on. IMPACT OF NUTRITIONAL MANAGEMENT ON FEEDING BEHAVIOR When grazing, ca le o en synchronize their behavior such that many animals in the group feed, ruminate, and rest at the same mes (Miller and Wood-Gush, 1991). Cur s and Houpt (1983) reported that group-housed dairy cows housed indoors also synchronized their behavior, par cularly at feeding. They reported that when cows are fed in groups, the act of one cow moving to the feed bunk s mulates others to feed. It has typically been accepted that dairy ca le exhibit a diurnal feeding pa ern where the majority of feeding ac vity occurs during the day, par cularly around sunrise and sunset (Albright, 1993). However, this observa on is almost exclusively based on the feeding pa erns exhibited by grazing ca le. To gain a be er understanding of how management factors inuence dairy ca le behavior, we have previously examined the normal feeding pa ern of group-housed lacta ng cows fed a TMR ad libitum (DeVries et al., 2003). We found that
the diurnal feeding pa ern was mostly inuenced by the me of feed delivery, feed push-up, and milking. Further, it was clear that the most drama c peaks in feeding ac vity occur around the me of feed delivery and the return from the milking parlor. To follow up on this, we set out in an experiment to determine which of these management prac ces is the primary factor s mula ng dairy ca le to go to the feed bunk (DeVries and von Keyserlingk, 2005). We tested this objec ve by separa ng feed delivery and milking mes by 6 h. When animals were fed 6 h post milking, they increased their total daily feeding me by 12.5%. This change was predominantly driven by a small decrease in feeding me during the rst hour post-milking and a very large increase in feeding me during the rst hour immediately following the delivery of fresh feed. These results indicate that the management prac ce of feed delivery acts as the primary inuence on the daily feeding pa ern of lacta ng dairy cows and not, as previously thought, the me of day. The delivery of fresh feed is clearly an important factor in s mula ng cows to eat. Thus, the frequency of feed delivery should inuence the feeding pa erns of lacta ng dairy cows. To test this predic on, we conducted an experiment to determine whether increasing frequency of feed delivery aects the behavior of group-housed dairy cows (DeVries et al., 2005). This objec ve was tested in two experiments. In the rst experiment, the treatments were: 1) delivery of feed once per day (1x), and 2) delivery of feed twice per day (2x). The treatments for the second experiment were: 1) delivery of feed 2x, and 2) delivery of feed four mes per day (4x). In both experiments, increased frequency of feed provision increased total daily feeding me by 10 and 14 minutes, respec vely, as well as increased the distribuon of feeding me throughout the day. The distribu on of feeding me in both experiments indicated that cows had more equal access to feed throughout the day when provided feed more frequently. Frequency of feed delivery had no eect on the daily lying me of the cows or the total number of aggressive interac ons at the feed bunk. However, we did nd that subordinate cows were not displaced as frequently when fed more o en, indica ng that these cows would have greater access to feed, par cularly fresh feed, when the frequency of feed delivery is high. In addion to these behavioral measures, we also looked at the eects of frequency of feed delivery on feed composi on throughout the day. In both experiments we noted that the neutral detergent ber (NDF) content of the TMR present in the feed bunk increased throughout the day, indica ng that sor ng of the feed had occurred, par cularly on the 1x treatment. Corrobora ng this, it was recently shown in
76
a mul -herd study (Endres and Espejo, 2010) that a greater change in ra on NDF content over the course of the day was associated with lower frequency of feed delivery (i.e. 1x vs. 2x or 3x per day). Increased sor ng of a TMR when fed 1x per day may be par cularly troublesome under some management situa ons. In two recent studies we have inves gated the eects of water addi on to a TMR, fed 1x per day, containing solely haylage and corn silage forage sources (MillerCushon and DeVries, 2009; Felton and DeVries, 2010). In a rst study we reduced TMR dry ma er concentra on from 58 to 48% (Miller-Cushon and DeVries, 2009). In a second study we reduced TMR dry mater concentra on from 56 to 50 to 44% (Felton and DeVries, 2010). In both studies we found that the addi on of water to these higher moisture TMR, containing no dry forage, actually resulted in greater amounts of feed sor ng. In addi on to this, increased amounts of water added to the ra ons also resulted in lower DMI. Reduced DMI with lower dietary DM is likely due, in part, to the increased lling eect of higher moisture ra ons. Interes ngly, in the second study we found that greater amounts of water added to the TMR resulted in greater increases in feed temperature in the hours a er feed was placed in the feed bunk, par cularly with higher ambient temperatures (Felton and DeVries, 2010). This increased feed temperature may be indica ve of feed spoilage and, thus, may be contribu ng to these eects. Thus, when feeding higher moisture TMR during periods of high ambient temperatures, delivering feed 1x per day may not only increase the risk of greater amounts of sor ng, but may also limit DMI. An alterna ve to adding water to TMR may be to add a liquid feed. Liquid feeds, par cularly those molasses-based ones, have good poten al to help bind ra on par cles together and make sor ng more dicult. Oelker et al. (2009) found that using a molasses-based liquid feed reduced sor ng of a corn silage based diet. Similarly, DeVries and Gill (2012), recently found that adding 4.1% of a molassesbased liquid feed to a TMR resulted in less sor ng against the longest ra on par cles. Further, in that study DMI was 1.4 kg/d higher (5.1% increase) and cows produced 3.1 kg/d more 4% fat-corrected milk (7.8% increase) on the liquid feed diet. Interes ngly, Eastridge et al. (2011) recently reported no eect of adding liquid feed on sor ng behavior; however, those researchers used a ra on that was very dry (> 64% DM). One of the most common feeding management prac ces believed to s mulate feeding ac vity is feed push-up.
When fed a TMR, dairy cows have a natural tendency to con nually sort through the feed and toss it forward where it is no longer within reach. This is par cularly problema c when feed is delivered via a feed alley and, thus, producers commonly push the feed closer to the cows in between feedings to ensure that cows have con nuous feed access. In an observa onal study Menzi and Chase (1994) noted that the number of cows feeding increased a er feed push up, however they concluded that feed push ups had minor and brief eects in comparison to milking on the feed bunk a endance. In a more recent study, we tested the s mulatory eect of feed push-up by increasing the number of push ups during the late evening and early morning (DeVries et al., 2003). In that study we found that the addi on of extra feed push ups in the early morning hours did li le to increase feeding ac vity. However, push up does play a vital role in ensuring that feed is accessible when cows want to eat. There is evidence to suggest that the ming of feed delivery is also important for dairy cows. Availability of fresh feed following the return from milking has typically been used to encourage cows to remain standing (while feeding) rather than to lie down. Researchers have shown that the presence of fresh feed in the bunk encourages longer post-milking standing mes (DeVries and von Keyserlingk, 2005). DeVries et al. (2010) recently found that the provision of feed around milking me resulted in the longest post-milking standing mes. Further, this was the rst study to document how post-milking standing me relates to the risk of subclinical udder infec on; cows that lay down, on average, for the rst me 40 to 60 min a er milking tended to have lower odds of a new subclinical udder infec on caused by environmental bacteria compared to cows that lay down within 40 min a er milking. These results suggest that management prac ces that discourage cows from lying down immediately a er milking, such as providing fresh feed frequently through the day (near the me of milking) may help decrease the risk of subclinical mas s. In addi on to the frequency and ming of feed delivery, the amount of feed delivered can also impact the feeding behavior of lacta ng dairy cows. Much of the research in this area has focused on inves ga ng the impact of feed access me. A substan al limita on (8 h) of feed access me has been shown to limit intake and produc on (Marnsson and Burstedt, 1990). Alterna vely, more moderate limita ons (2.5% refusal at 18 h vs. 5% refusal at 23 h) have been shown to have adverse eects on feeding behavior, causing cows to have shorter meals, less feeding me, and faster ea ng rates (French et al., 2005). In a more recent
77
study, Collings et al. (2011) found that temporal restricon of feed access (14 vs 23 h/d access) resulted in lower feeding mes and greater feed bunk compe on. Those researchers also found that at increased stocking densi es this temporal restric on increased feeding rates across the day. Interes ngly, in another recent study, Miller-Cushon and DeVries (2010) found that at lower feeding amounts, feed sor ng was reduced, while DMI was also reduced. IMPACT OF FEEDING ENVIRONMENT ON FEEDING BEHAVIOR One of the specic objec ves of ca le housing is to provide a comfortable environment that will allow cows to meet their behavioral and physiological needs (Phillips, 2001). There are several aspects of the feeding environment that have the poten al to inuence the ability of cows to access feed, including the amount of available feed bunk space per animal and the physical design of the feeding area. Recent observa ons have suggested that at the current industry standard of 0.61 m of feeding space per cow not all animals can access feed at the same me (DeVries et al., 2003). As social animals, ca le tend to synchronize their behavior, including a strong desire to access the feed bunk as a group. Reduced space availability has been shown to result in increased aggressive behavior in ca le (Kondo et al., 1989). When feed bunk space is limited, increases in aggressive behavior are thought to limit the ability of some cows to access feed at mes when feeding mo vaon is high, par cularly a er the delivery of fresh feed (DeVries et al., 2004; Huzzey et al., 2006). Hosseinkhani et al. (2008) recently demonstrated that compe on at the feed bunk drama cally increased the feeding rate at which cows feed throughout the day. These researchers also found that compe vely-fed cows have fewer meals per day, which tend to be larger and longer. In the study by Hosseinkhani et al. (2008) it was also found that compe on changed the distribu on of DMI over the course of the day, resul ng in higher intakes during the later hours a er feed delivery a er much of the feed sor ng had already occurred. Thus, increased compe on promotes feeding behavior that forces subordinate cows to consume more of their feed a er the dominant cows have sorted the TMR. These results suggest that increased compe on at the feed bunk promotes feeding behavior pa erns that will likely increase the between-cow varia on in composion of TMR consumed and the risk of sub-acute ruminal acidosis. Providing more space than the current industry norm has been shown to improve feed bunk access; this increases feeding mes and decreases compe on, with subordinate cows showing the greatest responses (DeVr-
ies et al., 2004; Huzzey et al., 2006). This change will help reduce the varia on in the composi on of feed cows consume as subordinate cows will be able to access the feed prior to it being sorted through by those dominant cows. In addi on to increasing the amount of available feed bunk space, compe on for feed can also be reduced through design of the feeding area. Researchers have shown that a headlock system greatly reduces compe on at the feed bunk compared with a post-and-rail system (Endres et al., 2005; Huzzey et al., 2006). Another op on to reduce compe on is the use of par ons (feed stalls) between the bodies of adjacent cows at the feed bunk. DeVries and von Keyserlingk (2006) demonstrated that feed stalls resulted in increased feeding me and decreased compe on, parcularly for subordinate cows. Their results suggest feed stalls provide addi onal protec on for feeding cows, and improved access to feed beyond that provided by simply increasing the amount of space per animal. It has become increasingly evident from this research that the provision of more feed bunk space (than tradi onally provided), par cularly when combined with a physical barrier separa ng adjacent cows (e.g. feed stalls), will improve access to feed and reduce compe on at the feed bunk, par cularly for subordinate cows. This could help reduce the between-cow varia on in the composi on of ra on consumed by preven ng subordinate cows from being forced to access the bunk only a er dominant cows have sorted the feed. CONCLUSIONS This proceedings chapter summarizes a number of studies that have been undertaken that collec vely provide us with a basic understanding of how knowledge of feeding behavior, par cularly how, when, and what cows eat of the feed provided to them, can be used to maximize the poten al of the ra ons provided to them. In par cular, these can be accomplished through ensuring cows have access to fresh feed throughout the day and minimizing feed bunk compe on. ACKNOWLEDGEMENTS This paper is an updated version of a proceedings paper wri en for, and presented at, the 2011 California Animal Nutri on Conference held in Fresno, CA in May 2011. The paper reviews concepts and data largely generated in collabora on with colleagues, par cularly Marina von Keyserlingk, at The University of Bri sh Columbia Animal
78
Welfare Program. This research was funded in part by the Natural Sciences and Engineering Research Council of Canada, Dairy Farmers of Canada, Westgen Endowment Fund, Investment Agriculture Founda on of Bri sh Columbia, Agriculture and Agri-Food Canada, the Canadian Bovine Mas s Research Network, the Ontario Ministry of Agriculture, Food, and Rural Aairs, the Canadian Foundaon for Innova on, and the Ontario Research Fund, the University of Guelph, and the University of Bri sh Columbia Animal Welfare Program. REFERENCES
Albright, J. L. 1993. Feeding behavior in dairy ca le. J. Dairy Sci. 76:485-498. Allen, M. S. 1997. Rela onship between fermenta on acid produc on in the rumen and the requirement for physically eec ve ber. J. Dairy Sci. 80:1447-1462. Allen, M. S. 2000. Eects of diet on short-term regula on of feed intake by lacta ng dairy ca le. J. Dairy Sci. 83:1598-1624. Beauchemin, K. A., L. Eriksen, P. Nrgaard, and L. M. Rode. 2008. Salivary secre on during meals in lacta ng dairy ca le. J. Dairy Sci. 91:2077-2081. Collings, L.K.M., D.M. Weary, N. Chapinal and M.A.G. von Keyserlingk. 2011. Temporal feed restric on and overstocking increase compe on for feed by dairy ca le. J. Dairy Sci. 94:5480-5486. Coppock, C. E., D. L. Bath, and B. Harris, Jr. 1981. From feeding to feeding systems. J. Dairy Sci. 64:1230-1249. Cur s, S. E., and K. A. Houpt. 1983. Animal ethology: its emergence in animal science. J. Anim. Sci. 57:234-247. DeVries, T. J., F. Dohme, and K. A. Beauchemin. 2008. Repeated ruminal acidosis challenges in lacta ng dairy cows at high and low risk for developing acidosis: Feed sor ng. J. Dairy Sci. 91:3958-3967. DeVries, T. J., L. Holtshausen, M. Oba, and K. A. Beauchemin. 2011. Eect of parity and stage of lacta on on feed sor ng behavior of lacta ng dairy cows. J. Dairy Sci. 94:40394045. DeVries, T. J., M. A. G. von Keyserlingk, and K. A. Beauchemin. 2003. Diurnal feeding pa ern of lacta ng dairy cows. J. Dairy Sci. 86:4079-4082. DeVries, T. J., M. A. G. von Keyserlingk, and D. M. Weary. 2004. Eect of feeding space on the inter-cow distance, aggression, and feeding behavior of free-stall housed lacta ng dairy cows. J. Dairy Sci. 87:1432-1438. DeVries, T. J. and M. A. G. von Keyserlingk. 2005. Time of fresh feed delivery aects the feeding and lying pa erns of dairy cows. J. Dairy Sci.88:625-631. DeVries, T. J., M. A. G. von Keyserlingk, and K. A. Beauchemin. 2005. Frequency of feed delivery aects the behavior of lacta ng dairy cows. J. Dairy Sci. 88:3553-3562. DeVries, T. J. and M. A. G. von Keyserlingk. 2006. Feed stalls aect the social and feeding behavior of lacta ng dairy cows. J. Dairy Sci. 89:3522-3531. DeVries, T. J., S. Dufour, and D. T. Scholl. 2010. Rela onship between feeding strategy, lying behavior pa erns, and incidence of intramammary infec on in dairy cows. J. Dairy Sci. 93:1987-1997.
DeVries, T. J., and R. M. Gill. 2012. Adding liquid feed to a total mixed ra on reduces feed sor ng behavior and improves produc vity of lacta ng dairy cows. J. Dairy Sci. 95:2648-2655. Eastridge, M. L., A. H. Lefeld , A. M. Eilenfeld , P. N. Go , W. S. Bowen , and J. L. Firkins. 2011. Corn grain and liquid feed as nonber carbohydrate sources in diets for lacta ng dairy cows. J. Dairy. Sc. 94:3045-3053. Endres, M. I., T. J. DeVries, M. A. G. von Keyserlingk, and D. M. Weary. 2005. Eect of feed barrier design on the behavior of loose-housed lacta ng dairy cows. J. Dairy Sci. 88:2377-2380. Endres, M. I., and L. A. Espejo. 2010. Feeding management and characteris cs of ra ons for high-producing dairy cows in freestall herds. J. Dairy Sci. 93:822-829. Fish, J. A., and T. J. DeVries. 2012. Varying dietary dry ma er concentraon through water addi on: Eect on nutrient intake and sor ng of dairy cows in late lacta on. J. Dairy Sci. 95:850-855. Felton, C., and T. J. DeVries. 2010. Eect of water addi on to a total mixed ra on on feed temperature, feed intake, sor ng behavior, and milk produc on of dairy cows. J. Dairy Sci. 93:2651-2660. French. P., J. Chamberlain, and J. Warntjes. 2005. Eect of feed refusal amount on feeding behavior and produc on in Holstein cows. J. Dairy Sci. 88(Suppl. 1):175. Grant, R. J., and J. L. Albright. 2001. Eect of animal grouping on feeding behavior and intake of dairy ca le. J. Dairy Sci. 84:E156-E163. Hosseinkhani, A., T. J. DeVries, K. L. Proudfoot, R. Valizadeh, D. M. Veira, and M. A. G. von Keyserlingk. 2008. The eects of feed bunk compe on on the feed sor ng behavior of close-up dry cows. J. Dairy Sci. 91:1115-1121. Huzzey, J. M., T. J. DeVries, P. Valois, and M. A. G. von Keyserlingk. 2006. Stocking density and feed barrier design aect the feeding and social behavior of dairy ca le. J. Dairy Sci. 89:126-133. Ingvartsen. K. L. and J. B. Andersen. 2000. Integra on of metabolism and intake regula on: A review focusing on periparturient animals. J. Dairy Sci. 83:1573-1597. Kondo, S., J. Sekine, M. Okubo, and Y. Asahida. 1989. The eect of group size and space allowance on the agonis c spacing behaviour of ca le. Appl. Anim. Behav. Sci. 24:127-135. Krause, K. M., and G. Oetzel. 2006. Understanding and preven ng subacute ruminal acidosis in dairy herds: a review. Anim. Feed Sci. Technol. 126:215-236. Leonardi, C., and L. E. Armentano. 2003. Eect of quan ty, quality, and length of alfalfa hay on selec ve consump on by dairy cows. J. Dairy Sci. 86:557-564. Mar nsson, K., and Burstedt, E. 1990. Eect of length of access me to feed and allotment of hay on grass silage intake and produc on in lacta ng dairy cows. Swed. J. Agric. Res. 20:169-176. Menzi, W., Jr., and L. E. Chase. 1994. Feeding behavior of cows housed in free stall barns. Page 829-831 in Dairy Systems for the 21st Century. American Society of Agricultural Engineers, St. Joseph, MI. Miller, K., and D. G. M. Wood-Gush. 1991. Some eects of housing on the social behaviour of dairy cows. Anim. Prod. 53:271-278. Miller-Cushon, E. K., and T. J. DeVries. 2009. Eect of dietary dry ma er concentra on on the sor ng behavior of lacta ng dairy cows fed a total mixed ra on. J. Dairy Sci. 92:3292-3298. Miller-Cushon, E. K., and T. J. DeVries. 2010. Feeding amount aects the sor ng behaviour of lacta ng dairy cows. Can J. Anim. Sci. 90:1-7.
79
Nielsen, B. L. 1999. On the interpreta on of feeding behaviour measures and the use of feeding rate as an indicator of social restraint. Appl. Anim. Behav. Sci. 63:79-91. Oelker, E. R., C. Reveneau, and J. L. Firkins. 2009. Interac on of molasses and monensin in alfalfa hay- or corn silage-based diets on rumen fermenta on, total tract diges bility, and milk produc on by Holstein cows. J. Dairy Sci. 92:270-285. Phillips, C. J. C. 2001. Principles of ca le produc on. CABI Publishing, CAB
80
S.E. Nellis, K.A.Weigel and P.C.Hoffman Nellis, Department of Dairy Science University of Wisconsin-Madison Wisconsin-
Body weight, lb w
Week of life
Pollard and Drackley, 2002
Starter consumption
6
5 4 3 2 1 0 1
Batch pasteurizer
145F(63C)/30 min
UV pasteurizer
4 5 Week of life
Labor Efficiency
Milk Production-Intensified Calf Nutrition (Van Amburgh 2011) Trial Foldager and Krohn 1994 Bar-Peled et al., 1998 Foldager et a., 1997 , Ballard et al., 2005 Shamay et al., 2005 Rincker et al., 2006 Drackley et al., 2007 Morrison et al., 2009 Moallem et al., 2010 Treatment Difference (lbs) 3092 998 1143 1543 2162 1100 1841 0 1613
Onfarm ID Net merit (NM$) Breed Performance Index (BPI) Milk Yield (Milk) Fat % Protein % Genomic Individual Inbreeding HH1 HH2 HH3 933 436 1791 1562 -0.05 -0.04 11.4 F F F 941 -96 1186 -677 -0.03 0.02 12.3 F F C
81
Genomics
<>
SNP
AGTCCATGGGGTTATAGAGTCAGACACAGTGGAGTCACACACATACACACG TCACCACGCCGAATTAAGGCGGGGCTGAGACAAGGGCAGGTGAGGCCTCC
genotype haplotype
1
Courtesy of George Wiggans, USDA AIPL
Slope of the line indicates the estimated SNP effect, which is the change in PTA protein per extra copy of a given SNP allele
Genomics 50,000 SNP >$100.00 (Elite Animals) 3000 k SNP (Fall 2010) Discontinued 6000 k SNP Available $40.00 45.00 3 drops of blood or hair Genomic PTAs by 2 3 months of age Reliability = 50.0 70.0 % Integrates Parent Averages Holstein/Jersey Association Pfizer (Clarifide)
82
Fill out the Form - Tedious Detail - Improper Parent/Sire/Dam/Breed ID = 10-15 % Normal
Official ID (Registration #, USDA AIN, Calfhood Vaccination #) 982000133517324 982000133517327 982000133517331 982000133517332 982000133517337 982000133517339 982000133517341 982000133517343 982000133517344 982000133517348 98200013351430 Sample Collector Barcode US4071556 US4071540 US4071529 US4071533 US4071521 US4071517 US4071525 US4071550 US4071513 US4071554 US4071536 Birth Date Breed (yyyy/mm/d (HO/JE/ d) or Birth Dam Registration # or Sire BS) USDA AIN Sex Year Registration # 355HE8261 F HO 11/21/2010 7H7853 Reg. # USDA AIN Vacc. # Other
On farm ID (Herd Management #) 6751 6754 6758 6759 6764 6766 6768 6770 6771 6775 6857
x x x x x x x x x x x
F F F F F F F F F F
HO HO HO HO HO HO HO HO HO HO
11/28/2010 200H3218 11/30/2010 200H6004 11/30/2010 7H10489 12/6/2010 12/8/2010 12/9/2010 14H4929 11H9703 7H9030
35SHE8183 35SHE7962 35SHE7938 35SHE7999 35SLS7488 35SHE5165 35SHE5981 840003003932331 35TDN0336 35SHE5128
+490
83
Some Math: Culling the bottom 15-25 % of heifers has a big effect on genetic improvement
Normal Genetics
Potential Culls ?
Potential Problem : Selecting for Milk Alone Ignores Other Important Traits
GPTA milk +1114 +141 -1041 +490 3 Heifer Average = 581 GPTA milk 4 Heifer Average = 176 GPTA milk (Common AI Improvement)
Old Heifer Management Paradigms Cause Stringing - Average Daily Gain - Average Age at First Calving - Breed when shes 875 by eyeballametrics - Breed when shes 50 tall
Stringing
All of these general historic concepts increase the variance of days on feed
Minimum breeding age 13 mo Minimum breeding BW 875 lbs Average daily gain 1.8 lbs/d Growth variance 0.2 lbs/d Pregnancy rate 50 %
Breed Target Variance and Reproductive Efficiency Influence Age of Calving A Normal Distribution is Not Normal - Stringing
84
AFC = 23.3 months Normal AI Stringing Variance (Days on Feed) = 125 days Non Genetic Weight Variance = 225 lbs
Distribution
Straws 2-4
23
Onfarm ID
HH1
HH2
HH3
6658 6660 6661 6662 6664 6666 6667 6669 6670 6671 6673 6676 6677 6678 6679 6680 6681 6682 6683 6684 6686
F F F F F F F F F F F F F F F F F F F F F
F F F F F C F F F F F F F F F F C F F C F
F F F F F F F F F F F F F F F F F F F F F
924
22
215
-0.01
0.02
-0.26
0.36
0.02
Haplotype Carriers (HH1 HH2 HH3) are Indentified - Do not breed to HH Carrier Bulls - > Chance of Failed Conception, Early Embryonic Loss, or Abortion
AI Mating Services are Adapted for Genomic Data - Production Mating (NM, Cheese Yield, Milk, Fat, or Protein etc) - Inbreeding Defined by Genomics is Considered - Many Type and Traits for Consideration (Limited Example Above)
85
Precision Feeding of Gravid Holstein Heifers: Effect on Growth, Manure Nutrient Excretion, and Subsequent Early Lactation Performance JDS 90 P.C. Hoffman, C.R. Simson, M. Wattiaux
Effect of precision feeding on body size and growth of replacement heifers. Treatment Item Intial Weight, lbs Hip height, in Body condition score Final Weight, lbs Hip height, in Body condition score Growth Average daily gain, lbs/d Feed efficiency, lbs DM/lb gain Excretion DM, lbs/d Hoffman et al. 2005 1.66 13.2 7.7 1.92 10.7 6.9 1.84 11.1 5.8 1220 56.0 3.2 1234 56.3 3.2 1217 56.4 3.2 1036 54.2 3.1 1021 54.6 3.0 1011 54.9 2.9 C-100 L-90 L-80
1
P < 0.01
L-80
12000 11000 10000 9000 8000 7000 6000 C100 L90 L80
NS (P > 0.10)
High Precision Forage Fed Milk, lbs Fat, % Protein, % 20761 3.74 3.05 23041 3.98 2.95
86
L85 29
L80 + I 26
SEM
26
Holstein and Crossbred Heifers 3 22 months of age. Experimental Diets Unsupplemented = 0.28 % P Supplemented = 0.38 % P
0.28 % P Holstein Crossbred 1370.0 57.4 57 4 22.0 65.1 81.1 11.1 16.9 15.8 210.8 22.1 1263.1 54.5 54 5 20.8 64.9 78.9 10.6 16.5 15.8 205.4 21.4 0.38 % P Holstein Crossbred 1406.3 57.3 57 3 22.0 65.1 81.0 11.4 16.6 16.4 214.3 22.3 1293.1 54.8 54 8 21.3 64.8 79.8 10.6 16.8 15.9 210.4 21.5 Effect (P< Diet ns ns ns ns ns ns ns ns ns ns
Item Body weight, lbs Hip height in height, Hip width,in Body length, in Heart girth,in Cannon bone,in Pelvic height,cm Pelvic width,cm Pelvic area,cm2 Pelvic length,in
Holstein and Crossbred Heifers 3 22 months of age. Experimental Diets Unsupplemented = 0.28 % P Supplemented = 0.38 % P
Item Bone density Trabecular bone density,mg/cm3 Cortical bone density, mg/cm3 Total bone density, mg/cm3 Chemical composition P, % Ca, % Ash, %
305 Day Milk ME 305 Milk Peak Milk Total Milk Total Fat Total Protein Fat Corrected Milk Average Log SCC Average MUN Average Milking Speed
ns ns ns 0.08 ns ns
87
Notes
88
SUMMARY Es mates of the amount of varia on (i.e., standard deviaons and ranges) in nutrient composi on of common feeds are readily available from some na onal feed tes ng labs; however, es mates of varia on on feeds within a farm are needed to determine appropriate feed sampling and ra on re-formula on schedules. We conducted a large na onal survey to determine within farm varia on for some common feeds. On average within farm varia on for crude protein (CP) and neutral detergent ber (NDF) was 2 or 3 mes less than varia on in the na onal popula on. Within a farm, substan al day to day varia on in dry ma er, NDF, CP and starch was observed for corn silage and haycrop silage, indica ng that samples should be taken over several days and those results averaged for diet formula on purposes. In some cases, day to day varia on was as great as month to month varia on. The concentra ons of NDF, but not CP, in high moisture corn, dis llers grains, and brewers grains varied almost as much as NDF concentraons in some forages, indica ng that these feeds should be sampled on a regular basis. INTRODUCTION Unques onably the nutrient composi on of all feeds varies, which means that the composi on of the ra ons fed to cows also varies. Although varia on in nutrient composion is a well-accepted fact of life, this varia on has not been extensively quan ed and the eects of varia on on cows have not been studied. A group of researchers and extension specialists at The Ohio State University with the help of collabora ng nutri onists and dairy farmers are working on a large project designed to: 1) Determine the varia on in nutrient composi on of common feeds on commercial dairy farms; 2) Determine how varia on in ingredients aects varia on in nutrient composi on of TMR on commercial dairy farms; 3) Par on sources of varia on (e.g., lab, sampling, and real); 4) Determine covariances among nutrients in common feeds; 5) Determine how varia on in the nutrient composi on of diets aect lacta ng dairy cows; and 6) A empt to put a dollar value
on varia on. This project is ongoing (as of 2012) and it is beyond the scope of this paper to discuss all those objecves. This paper will concentrate on the degree of varia on observed in nutrient composi on of common feeds on dairy farms, some factors inuencing that varia on, and some implica ons of that varia on on ra on formula on. THE IMPORTANCE OF KNOWING NUTRIENT VARIATION The amount of varia on in the nutrient composi on of a feed can aect diet formula on strategies, the economic value of the feed, the sampling and analysis schedule for the feed, and possibly the produc vity and health of cows. An important component of a good ra on formulaon strategy is risk management. Diets are formulated, in part, to minimize the risk of a nutrient deciency that would reduce produc on or impair cow health. The safety factors included in ra on specica ons are dependent on factors such as variability in cow factors within a pen (stage of lacta on, milk yield, parity, etc.), overall quality of nutri onal management on a given farm (monitoring feed bunks, consistency of the total mixed ra on, etc.), and the variability in the nutrient composi on of the ingredients in the diet. A total mixed ra on (TMR) based on ingredients that are highly variable in protein concentra ons will likely be formulated to a greater protein concentra on than a diet based on very consistent ingredients. This overformula on reduces the risk that the diet will be decient in protein because of changes in the composi on of the ingredients. The need for greater safety factors for variable feeds should aect the economic value of the ingredients. For example, a diet might be formulated to contain 16% protein when based on consistent ingredients, but when based on a highly variable ingredient the diet might be formulated to contain 17%. The price of the highly variable ingredient must be discounted enough to cover the cost of feeding the higher protein diet. This also has an environmental cost (i.e., greater N excre on), and in many loca ons that translates directly into an economic cost. The goal of an op mal sampling (and analysis) schedule is
89
to minimize cost. Sampling too frequently increases feed analysis cost, but sampling too infrequently may result in lost produc on or health problems because a change in diet composi on was not iden ed in a mely manner. Feeds that are consistent will have a very dierent op mal sampling schedule than that of highly variable feed. At this me, very li le is known about how varia on in nutrient composi on of the ra on aects produc vity of dairy cows, but if ra on varia on aects the cow then the cost of lost milk (or health problems) will have to be factored into the valua on of the ingredients. VARIATION IN NUTRIENT COMPOSITION OF FEEDS Two excellent sources for nutrient composition data (including varia on) of common feeds are the websites maintained by: DairyOne Coopera ve (Ithaca, NY): www.dairyone.com/Forage/services/default.asp Dairyland Laboratories Inc. (Arcadia WI): www.dairylandlabs.com/ (Note: Other qualied labs are available) For many feeds, these data bases include thousands of samples from a wide geographic area, mul ple years, mul ple hybrids and varie es, diverse growing condi ons, dierent manufacturing systems (e.g., dis lleries or our mills), etc. so that the data from those sources represent a na onal popula on. In other words, if you take a random sample of corn silage or dis llers grain from the U.S. it would likely t the popula on of samples in those two data bases. Although na onal data are valuable, they most likely do not accurately reect varia on in nutrient composi on in feedstus within a farm. As a popula on becomes more specic (e.g., corn silage from across the U.S. vs. corn silage grown on Farm X in 2012), we expect the standard devia on to become smaller because many fewer sources of varia on occur on Farm X than what occurs across the U.S. Knowing varia on in feed composi on at the farm level, rather than at a na onal or global level, will allow us to ne tune ra on safety factors, compare economic value of feeds more accurately, and set up opmal sampling schedules for specic farms. Methodology To determine variability in feed composi on at the farm level, 50 well-managed dairy farms from across the U.S. (20 within Ohio, 30 outside of Ohio) were enrolled in this project. The nutri onists for the farms were given a detailed sampling protocol so that sampling procedures would be consistent across farms, and then all major ingredients
added at the farm to the TMR mixer for the high group were sampled once monthly. Nutri onists were free to add new ingredients or stop feeding ingredients at their discreon. All samples from all farms were sent to a common lab (Cumberland Valley Analy cal Services, Hagerstown, MD) and analyzed via wet chemistry for dry ma er (DM), crude protein (CP), neutral detergent ber (NDF), ash, and major minerals (ash and minerals will not be discussed in this paper). Sta s cs on nutrient composi on were then calculated within each farm and feed. If a farm iden ed dierent popula ons of a feed, sta s cs were calculated for each designated popula on. For example, if a farm sent in samples iden ed as bunker corn silage and bag corn silage, those two feeds were kept separate. If samples were designated simply as corn silage, all samples were considered to be from the same popula on. The nal data set contained 4,700 samples from 49 farms (one farm dropped out of the project) from 10 states (CA, IA, MI, NM, NY, OH, SD, UT, TX, and WI). Data from 49 general ingredients (i.e., all corn silages were considered one ingredient even though farms may have used two dierent corn silages) plus samples of commercial mixes, preblends, and TMR were obtained. For this analysis, ingredients had to be sampled at least 4 mes on a specic farm during the 12-month period and had to be fed by at least 5 farms. Many feeds met these criteria, but for this paper discussion will be limited to corn silage, legume hay, legume silage, small grain silage, straw, dry shelled corn, high moisture corn, dried dis llers grains, soybean meal, and wet brewers grains. Short term (day to day rather than over a year) varia on in haycrop silage and corn silage was evaluated by sampling those silages on 14 consecu ve days on 8 farms near Wooster, OH. On 4 of those farms, independent duplicate samples (mul ple handfuls were taken, mixed, and a subsample was placed in a bag; then that process was repeated) were taken each day. Corn silage was assayed for DM, NDF, and starch, and haycrop silage was assayed for DM, NDF, and CP. All assays were conducted in duplicate at the OARDC Dairy Nutri on Lab using standard wet chemistry methods. Day to Day Variation in Nutrient Composition of Silages Within a farm and over a rela vely short period, day to day varia on in nutrient composi on of silages on many farms was substan al (Figure 1 and Table 1). For corn silage, the day to day varia on in starch was greater than NDF which was greater than DM (coecients of varia on (CV) were 9.5, 6.2 and 5.3, respec vely). The day to day varia on in those nutrients did not follow any discernable pa ern (Figure 1). For the 8 farms, the average range in
90
starch concentra on in corn silage was 12.2 percentage units over a 14 day period. The most consistent corn silage (within a farm) had a range in starch concentra on of 6.3 percentage units and the most variable had a range of 27.7 units (Table 1). If that varia on is real (i.e., not caused by sampling or laboratory error) a devia on of about 14 units from the mean (i.e., half the range) would alter the starch concentra on of the diet by about 3.5 percentage units if silage comprised 25% of the TMR. That degree of change could be enough to cause rumen upset. Day to day varia on in corn silage NDF was also substan al but the varia on was more consistent from farm to farm than the farm to farm varia on in starch. The most consistent farm had a range of 7.3 percentage units in NDF whereas the most variable farm had a range of 11.2 percentage units in corn silage NDF. For the most variable farm, a deviaon of 5.6 units from the mean would change TMR NDF by about 1.4 percentage units (assuming corn silage had a 25% inclusion rate). Although DM was more consistent than the carbohydrates, it s ll ranged within a farm from 5.1 units up to 10.4 units. Because diets are formulated on a DM basis but delivered on an as-fed basis, a deviaon of 5 percentage units in DM could substan ally alter diet composi on.
For hay crop silage, DM was most variable, followed by CP and then NDF (CV = 8.5, 5.8, and 5.0, respec vely). The range in varia on between farms was large (Table 1). The range in DM concentra ons within a farm was more than 5 mes greater for the least consistent hay crop silage compared with the most consistent. The range in NDF and CP varied about 4-fold between the most and least consistent haycrop silages. Contrary to conven onal wisdom, within a farm over a short period of me, haycrop silage was a more consistent source of NDF than was corn silage (5.0 vs. 6.2 for haycrop and corn silage, respec vely). Even though day to day varia on in corn and haycrop silage was substan al, it was much less than what was observed in the na onal popula on. The average standard deviaon within a farm was 2 to 3 mes less than SD for the na onal popula on (Table 1). When developing op mal sampling schedules, if actual variance on a given farm is not known, the average within farm SD shown in Table 1 would be more accurate than using the na onal SD. An obvious ques on is: Why are corn and haycrop silages so variable over a short period of me? One possible reason is laboratory error; however, all assays were conducted in duplicate, and averaging the duplicate values had es-
Corn Silage
50 45 40 35 30 25 20 15 10 0 2 4 6 8 Day 10 12 14 DM NDF Starch 50 45 40 35 30 25 20 15 10 0 2
Alfalfa Silage
DM 4
NDF 6 8 Day 10
CP 12 14
Figure 1. Day to day changes in DM (%) and NDF, CP, and starch (% of DM) of corn silage and haycrop silage on two farms. The farms were chosen because the day to day range in nutrient composi on was approximately the average for the data set (see Table 1).
91
sen ally no eect on SD, CV, or dairy farms in Northeastern Ohio1 ranges when compared to using Corn Silage Hay Crop Silage a single laboratory value for each DM, % NDF, % Starch, % DM, % NDF, % CP, % day (data not shown). Another 2 potential source of variation Averages Mean 38.8 40.4 31.7 43.6 47.6 17.3 is sampling error. Sampling erSD 2.07 2.52 3.02 3.7 2.38 1.01 ror can be dened as varia on among samples from a dened CV 5.3 6.2 9.5 8.5 5.0 5.8 popula on (assuming no laboraRange 7.3 8.8 12.2 11.8 8.5 3.4 tory error). For example, if you Ranges3 had a pile of 2,000 lbs. of corn Mean 31.5 - 45.7 35.4 - 45.0 27.0 - 39.2 32.5 - 55.7 36.1 - 58.2 15.1 - 21.9 silage that was going to be fed SD 1.50 - 3.04 2.16 - 3.27 2.05 - 5.26 1.00 - 6.66 0.92 - 3.64 0.37 - 1.61 to a single pen of cows today Range 5.1 - 10.4 7.3 - 11.2 6.3 - 27.7 3.4 - 19.1 3.2 - 13.6 1.2 - 4.9 and you took 10 samples from 4 the pile, the variation among Na onal Popula on sta s cs those samples is sampling error Mean 33.4 41.7 34.0 40.1 47.0 20.5 (again assuming no analy cal er- SD 6.1 5.4 7.3 10.3 5.7 3.0 ror). To evaluate sampling error, CV 18.3 13.0 21.5 25.7 12.1 14.6 duplicate samples were taken 1 Samples of corn silage and haycrop silage were taken for 14 consecu ve days during a me when the each day on 4 farms. Averaging silages did not knowingly change (i.e., same storage structure, same growing season, same cu ng). the results from the duplicate 2 The mean, standard devia on (SD), coecient of varia on (CV) and range (maximum daily value samples reduced varia on by 13 minimum daily value) were calculated for each farm and then averaged for the 8 farms. to 25% depending on the nutri- 3 These ranges are calculated between farms (e.g., on the most consistent farm corn silage DM ent and type of silage (Figure ranged by 5.1 units but on the most inconsistent farm, it ranged by 10.4 units). 4 Data are from DairyOne Forage Summary (mixed mostly legume silage was used for haycrop silage) 2). Sampling error is clearly an on samples analyzed from May, 2010 to May 2011. important source of varia on, and it is probably greater than 3.5 No Dups our es mates because with the sampling protocol we Duplicates 3 followed, two duplicate samples may not have been adequate to represent all the sampling error. Indeed a sub2.5 stan al amount of the day to day varia on we observed 2 over the 14 day period is most likely sampling error. Even 1.5 though the cow does not experience sampling error, it can 1 have a signicant impact on diet formula on and the cows. High sampling error means that you should not have great 0.5 condence in the results from a single sample. Rather, mul0 ple samples should be taken over a short period of me CS DM CS NDF CS Starch HCS DM HCS NDF HCS CP (days) and the average of those samples should be used for diet formula on. We used Monte Carlo techniques Figure 2. The eect of taking independent duplicate samples to randomly select samples from the day to day study each day on varia on in nutrient composi on of corn silage and determined that the mean of 3 samples consistently (CS) and haycrop silage (HCS) on SD within a farm over 14 days. Each bar is the average of the standard devia ons matched the overall mean (over 14 days) for NDF within calculated for each farm (4 farms). each farm. A single randomly selected sample was within 5% of the mean only 50% of the me. does not increase by a factor of 2, but by a factor of 4 (i.e., 2 squared). Relying on mul ple ingredients, each with a The eect of an ingredient on the varia on observed in the limited inclusion rate, can greatly reduce varia on in the TMR is dependent on varia on in the ingredient and on the TMR. This is illustrated in Figure 3. The NDF concentra ons inclusion rate of the ingredient. The eect an ingredient has of corn silage and hay crop silage from two farms are highly on the varia on of the total diet changes with the square of variable day to day; however, if the TMR contained 25% the inclusion rate. This means that when the inclusion rate corn silage and 25% haycrop silage, the concentra on of of an ingredient is doubled, its eect on total diet variance forage NDF in the diets is less variable. In Farm A (Figure
Standard Deviation Units
Table 1. Day to day varia on in nutrient composi on of corn silage and haycrop silage on 8
92
Farm A
CS 60 55 50 45 40 35 30 25 20 15 0 5 Day 10 15 Haylage Mix 65 60 55 50 45 40 35 30 25 20 0
Farm B
CS Haylage Mix
5 Day
10
15
Figure 3. The eect of blending forages on reducing varia on in the TMR. Forages were sampled daily and assayed for NDF. The line designated as Mix is the concentra on of forage NDF in the TMR assuming corn silage and haycrop silage each comprised 25% of the TMR and no other forages were fed (data are % of DM of silage or total TMR).
Corn Silage
50 45 40 35 30 25 20 15 10 0 2 4 6 8 Month 10 12 DM NDF 70 65 60 55 50 45 40 35 30 25 20 0
Legume Silage
DM NDF
6 8 Month
10
12
Figure 4. Month to month changes in concentra on of DM (%) and NDF (% of DM) in corn silage and legume silage. The farms were chosen because the across months range is approximately equal to the average for the survey (Table 2).
93
3), the CV for corn silage and haycrop silage are 5.3 and 4.3 but the CV for the mix is only 3.7. For farm B, the CV for corn silage is 7.3 and 2.9 for haycrop and for the mix it is 4.3. For both farms, the CV of the mix is much less than the average of the CV for the two forages. Month to Month Variation in Forages Measures of the month to month varia on in nutrient composi on of forages within farms are shown in Tables 2 and 3 and varia on in concentrate feeds are shown in Tables 4 and 5. For comparison purposes, means and SD of the same feeds but for the na onal popula on are shown in Table 6. General observa ons include: 1. As expected, DM concentra ons of silages were much
Table 2. Varia on in nutrient composi on of forages within farms over a 12 month period1 DM, % Corn silage (data from 48 farms) Mean 34.1 SD CV Range Mean SD CV Range Mean SD CV Range Mean SD CV Range Mean SD CV Range
1
more variable than DM of dry forages, and corn silage DM was less variable the haycrop silage DM. 2. Rela ve varia on in CP concentra on (CV) were similar between the forages except for straw, but the high CV for straw is misleading because straw has such a low concentra on of CP. 3. The variation in NDF concentration varied among forages (legume hay was most variable, then legume silage, then small grain and corn silages, and straw was most consistent). 4. Unexpectedly, the rela ve varia on in corn silage NDF over 12 months was essen ally the same as observed over a period of 14 days (Tables 1 and 2 and Figures 1 and 4). However, as would be expected, the average varia on in NDF of haycrop silage over a 12 month period was substan ally greater than what was observed over a 14-day period (CV = 8.7 vs. 5.0). 5. The average SD (within farm) for DM, CP, and NDF were usually one-third to one-half as great as the SD for the
Table 3. Within farm ranges (over a maximum 12 month period) in varia on in nutrient composi on of forages1 DM, % CP, % NDF, % 35.1 - 51.2 1.16 - 6.52 4.1 - 22.1 30.9 - 43.9 0.68 - 7.32 1.6 - 16.6 32.1 - 47.3 1.21 - 6.01 2.6 - 19.0 42.2 - 64.0 1.36 - 6.88 4.0 - 17.2 72.4 - 82.6 1.21 - 4.59 2.8 - 15.5 Corn silage (data from 48 farms) Mean 29.9 - 43.1 6.8 - 11.8 SD Range Mean SD Range Mean SD Range Mean SD Range Mean SD Range
1
CP, % DM 7.9 0.55 7.0 2.0 21.2 2.01 9.4 6.3 21.5 1.64 7.6 5.2 13.0 1.46 11.2 4.6 4.7 1.16 24.7 3.4
NDF, % DM 40.8 2.60 6.4 8.8 37.5 4.20 11.2 13.2 39.9 3.47 8.7 10.9 53.6 3.29 6.1 9.5 79.8 2.41 3.0 7.3
2.67 7.8 9.1 87.8 2.46 2.8 8.2 42.8 6.29 14.7 20.7 37.3 3.27 8.8 10.6 87.8 3.78 4.3 11.7
0.80 - 5.00 2.5 - 17.9 72.3 - 88.5 1.17 - 9.79 2.1 - 20.7 27.2 - 49.9 2.93 - 11.1 8.6 - 41.3 31.4 - 61.7 1.67 - 4.62 4.4 - 17.1 82.3 - 90.8 0.76 - 13.40 1.4 - 42.5
0.24 - 1.27 0.8 - 4.7 18.3 - 24.6 1.18 - 3.62 3.7 - 11.1 17.8 - 23.8 0.33 - 2.64 0.9 - 9.1 10.3 - 17.1 0.59 - 3.76 1.6 - 13.6 3.7 - 5.8 0.27 - 2.73 0.8 - 8.9
Samples were taken once monthly if the forage was fed on the farm. Samples are not necessarily from the same popula on (i.e., corn silage could come from dierent structures, elds, etc; hay crops could be dierent cu ngs, elds, etc.). The sta s cs are calculated within each farm and then averaged across farms.
Samples were taken once monthly if the forage was fed (not all forages were fed for 12 months on all farms). Samples are not necessarily from the same popula on (i.e., corn silage could come from dierent structures, elds, etc; haycrops could be dierent cu ngs, elds, etc. The sta s cs are calculated within each farm and then the minimum and maximum among farms were determined.
94
na onal popula on except for legume hay. The SD for legume hay in the limited popula on (Table 3) may have been similar to the na onal popula on (Table 6) because farms o en purchase hay from na onal markets. More important than average within farm varia on in forage composi on is the range in within farm varia on that was observed (Table 3). Some of the farm to farm dierences in variability could be caused by sampling varia on (dierent people sampled at dierent farms, storage structures diered, etc.) but some is likely real varia on. For the forages on the most consistent farms, the range in CP and NDF concentra ons were so small as to have li le biological (and economic) importance. On the other hand, for the most variable farms, the range in NDF was 16 to 22 percentage units depending on the
Table 4. Varia on in composi on of concentrates within farms over a 12 month period1 DM, % Corn grain (data from 27 farms) Mean 85.3 SD CV Range Mean SD CV Range Mean SD CV Range Mean SD CV Range Mean SD CV Range
1
forage (8 to 11 percentage unit devia on from the mean). At a 25% inclusion rate, this amount of varia on could change forage NDF concentra ons of the TMR by 2 to 3 percentage units. The range in CP concentra ons for the least consistent farms is great enough to also cause substan al change in CP concentra on of the TMR. For the most consistent farms, repeated sampling of the forages would be unnecessary; however, for the more variable farms, frequent sampling would be needed to quickly iden fy when forage composi on changed. Month to Month Variation in Concentrates As expected, the average varia on in DM, CP, and NDF for the concentrates was less than for the forages (Tables 2 and 4). However, varia on in NDF in many of the concentrates was substan al. In most situa ons, the average within farm varia on in CP or NDF concentra ons would have minor eects on TMR nutrient composi on. The average devia on (range divided by 2) in CP concentraon for most of the concentrates would change the CP
Table 5. Within farm ranges (over a maximum 12 month period) in varia on in nutrient composi on of concentrates1 DM, % Corn grain (data from 27 farms) Mean 83.0 - 86.5 SD Range Mean SD Range Mean SD Range 0.39 - 4.50 1.0 - 10.0 57.3 - 77.6 0.69 - 5.76 1.5 - 19.4 86.9 - 91.8 0.58 - 2.22 1.7 - 5.4 87.2 - 91.2 0.29 - 3.25 0.7 - 8.6 21.0 - 24.8 1.15 - 2.58 3.7 - 9.2 CP, % 7.8 - 9.2 0.18 - 1.14 0.5 - 2.7 6.9 - 9.1 0.24 - 0.89 0.8 - 3.2 28.1 - 39.6 0.64 - 1.97 1.5 - 6.2 49.6 - 53.6 0.29 - 3.69 0.6 - 4.4 25.0 - 34.3 1.47 - 3.15 4.7 - 10.0 NDF, % 9.9 - 13.3 0.40 - 2.70 1.2 - 7.7 7.9 - 23.6 0.64 - 5.85 1.9 - 18.9 31.2 - 34.5 1.12 - 5.05 3.1 - 11.9 8.0 - 11.0 0.22 - 3.16 0.5 - 14.9 45.8 - 54.4 1.41 - 3.66 3.0 - 12.0
CP, % 8.4 0.43 5.1 1.2 8.1 0.58 7.2 1.8 30.4 1.02 3.4 2.9 52.4 1.05 2.0 2.3 31.3 2.03 6.5 6.3
NDF, % 11.3 1.30 11.5 3.7 11.6 1.87 16.1 5.9 32.6 2.25 6.9 6.4 8.7 0.95 10.9 3.1 49.1 2.60 5.3 8.1
1.67 2.0 4.6 68.8 3.04 4.4 9.5 89.9 1.31 1.5 3.6 88.3 0.91 1.0 2.5 22.9 1.70 7.4 5.5
Samples were taken monthly if the concentrate was fed (not all feeds were fed for 12 months on all farms). Samples are not necessarily from the same popula on (i.e., dierent lots, mul ple sources, etc.). The stas cs are calculated within each farm and then averaged across farms.
1 Samples were taken monthly if the concentrate was fed (not all feeds were fed for 12 months on all farms). Samples are not necessarily from the same popula on (i.e., dierent lots, mul ple sources, etc.). The sta s cs are calculated within each farm and then the minimum and maximum among farms were determined.
95
concentra on of the TMR by about 0.3 percentage varia on project1 units with an inclusion rate of 20% (a change of 0.6 DM, % CP, % DM NDF, % DM units is possible with wet brewers grains). Assuming a Mean SD Mean SD Mean SD 20% inclusion rate and average within farm varia on Corn silage 33.4 6.16 8.4 1.02 41.7 5.40 for concentrate NDF, diet NDF could change by 0.3 to Legume hay 91.0 1.13 21.3 2.70 38.8 5.23 0.8 percentage units (a change that likely would have Legume silage 41.3 11.00 22.1 2.92 43.7 5.55 li le eect on the cow). The average within farm SD 31.8 11.52 13.8 3.68 56.7 7.83 for the concentrates was much smaller than the SD for Small grain silage Straw 93.4 1.50 5.4 2.31 73.1 7.75 the na onal popula on (Tables 4 and 6) even though many of those concentrates are marketed na on- Corn grain 86.6 2.96 8.3 1.36 10.2 3.14 ally. For the na onal popula on dry corn was more High moisture corn 71.2 6.14 8.6 0.81 10.0 2.53 variable in CP and NDF than high moisture corn, but 1.79 32.0 4.88 34.9 5.78 the opposite was observed for average within farm Dried dis llers grain 88.0 90.9 2.47 51.4 4.79 13.4 6.74 varia on. This discrepancy may be caused by sample Soybean meal Wet brewers grains 24.5 6.34 29.0 4.43 50.6 4.61 mis-iden ca on in the na onal database (e.g., ear corn iden ed as shell corn). In the na onal data base, 1 Data are from DairyOne (Ithaca, NY) for samples analyzed from May, 2010 dried dis llers and wet brewers grains had similar through May 2011. varia on in CP concentra on but the dis llers grains IMPLICATIONS were much more variable in NDF concentra on (CV = 16.1 vs 9.1). For the average within farm varia on, dis llers 1. The amount of within farm varia on for specic feeds differs widely among farms. This means sampling grains were less variable in CP than were wet brewers and schedules should dier widely among farms (increased the two feeds had similar variability in NDF. sampling for farms with greater variability). The range in varia on among farms for the concentrates 2. Day to day variability in corn silage and haycrop silage (Table 5) is more important than the average varia on (Tais very large and o en as great as month to month ble 4). For example, the DM concentra on for high moisture varia on. Single samples should not be relied upon to corn from one farm varied by 19 percentage units over the provide an accurate descrip on of the feed, and subyear, which is enough to substan ally aect the amount of stan al changes in diet formula on should not be done corn DM fed in a diet. The largest month to month range in based on results from a single sample. Results from 2 DM concentra on of wet brewers could also aect TMR. or 3 samples taken within a short period of me (1 or 2 For most of the concentrates (the excep on is wet brewweeks) should be averaged and the average used in diet ers), even for the most inconsistent farm, the varia on in CP formula on. Duplicate samples taken on a single day probably would not cause major issues. The most inconsisreduced day to day varia on but probably not enough tent wet brewers with respect to CP could cause important to jus fy the added costs at least on smaller farms. changes in CP concentra ons of the TMR. On farms that 3. Mul ple sources of nutrients, even highly variable experienced the greatest range in NDF concentra ons for sources, greatly reduce the varia on in the nutrient high moisture corn, dis llers, and wet brewers grains, the composi on of the TMR. NDF concentra on of the TMR could be altered enough to aect cows. The NDF concentra on of soybean meal could 4. For many concentrates, the CP and NDF concentra on (and DM for wet feeds) vary substan ally. The varia on vary tremendously; however, inclusion rate for soybean in NDF was large enough that eects on cows might be meal is usually <10% and some of that varia on is likely observed if the changes in composi on were not used a result of not iden fying low and high protein soybean in diet formula on. meal as separate feeds. Assuming that concentrate feeds within a farm do not vary enough to jus fy sampling and 5. A month to month (day to day) change in nutrient composi on of a feed could substan ally alter its innutrient analysis is clearly wrong at least for NDF. The NDF clusion rates when linear programming is used for diet concentra on of high moisture corn, brewers grains, and formula on. If the change in nutrient composi on was dis llers grains should be monitored. An o en stated comnot real (e.g., sampling varia on), mul ple sampling plaint about dis llers grains is that it is too variable. Based of ingredients could actually increase varia on in TMR. on data in Table 4 and 5, dis llers and brewers grains have Averaging sample results should reduce this eect. similar varia on, and for NDF dis llers grains were actually more consistent than high moisture corn.
96
ACKNOWLEDGEMENTS This project was supported by Na onal Research Ini ave Compe ve Grant no. 2009-55206-05242 from the USDA Na onal Ins tute of Food and Agriculture. We also thank the coopera ng dairy farms and their nutri onal consultants. This paper was originally published in the 2012 Proceedings of the Tri-State Dairy Nutri on Conference, Ft. Wayne, IN.
97
Notes
98
INTRODUCTION Varia on in composi on can be large and dierent across feeds and nutrients. Dietary nutrient varia on can negavely aect feed costs, animal health, and milk produc on (Weiss and St-Pierre, 2009). Inadequate energy supply can reduce milk produc on, impair reproduc on, decrease resistance to disease, and may cause severe weight loss (Weiss, 1997). Strategies such as dietary safety factors (Kohn, 2006), using more ingredients in TMR with lower inclusion rates (Weiss and St-Pierre, 2009), and avoiding feeds that are highly variable in nutrient composi on are all strongly recommended. Safety factors are added to nutrient requirements to reduce the likelihood that a diet will be decient in certain nutrients; however, these safety factors usually increase feed cost but may increase milk produc on and income over feed costs (IOFC; St-Pierre and Thraen, 1999). Determina on of diet safety factors is somewhat guesswork as common nutri on models are incapable of predic ng the varia on in dietary net energy (NEL) or metabolizable protein (MP; Kohn, 2006). Forages and some byproducts are the most variable feeds in nutrient concentra ons; however, these feeds o en are considered bargain feeds based on their mean nutrient concentra ons and cost. Nutrient varia on (expressed as standard devia on, SD) within feed popula ons is provided by the NRC (2001) and several commercial laboratory websites. However, Weiss and St-Pierre (2009) suspect that many of the NRC (2001) es mates may be in error because of small sample size, old data, and presence of outliers. We aim to improve the es mates of varia on in the nutrient composi on of feeds, including NEL and MP, which should allow us to develop be er strategies for managing nutrient varia on and maximizing IOFC. FEED SUMMARIES Composi on data were collected from two major commercial laboratories on a wide range of feeds. Feed classicaon (i.e., feed name or code) provided by the labs was not changed. Feed sta s cal summaries were kept separate
by lab so we would avoid lab eects, dierences in classica on procedures between labs, and other unknown sources of varia on. We developed a stepwise univariate and mul variate sta s cal procedure to iden fy outlier samples and sub-popula ons (Yoder et al., 2012). Outliers in feed popula ons occur because of erroneous labeling of samples on farm (e.g., iden fying a mixed grass-alfalfa hay as alfalfa), taking a poor sample (i.e., the sample does not reect the popula on), improper classica on of sample by the lab, laboratory error, and other unknown factors. Large sets of composi on data for a given feed likely contain mul ple subpopula ons because of manufacturing techniques (e.g. brewery) or geographic loca on. Other subpopula ons could simply be the result of a misclassica on (e.g., labeling corn gluten feed as corn gluten meal). In our procedures we set a limit that if a subpopula on comprised less than 10% of the total number of samples, it was removed because it did not represent the primary popula on but it was not large enough to be classied as a separate feed. We applied this sta s cal procedure to the data provided by the labs and generated several sta s cs of interest (mean, SD, 90% condence intervals, and also the correla ons between major nutrients; e.g., DM, NDF, CP, and ash). The lack of homogeneity in some feed popula ons greatly aected varia on and correla on es mates but had minimal eects on es mates of the mean in most feeds. We discovered that across many feeds, es mates of SD and correla on changed greatly a er outlier samples or subpopula ons were removed (Table 1). For example, the SD es mates of CP concentra on were reduced by 52% for soybean meal and 76% for corn gluten meal a er outliers were removed. Contamina on by roasted soybeans and 44% soybean meal and corn gluten feed, respec vely, likely greatly inated the SD es mates. Across most feeds, the mean nutrient concentra ons were similar to the NRC (2001) and this is shown for six feeds in Table 2. However, some means diered greatly from NRC.
99
Wet brewers grains had a mean fat concentra on 4.6 units higher than NRC (2001) and this is likely because of the number of samples, 260 versus 1. Dried dis ller grains was 2.5 units higher in mean fat concentra on than NRC (2001), and this may be because of dierences in current manufacturing versus approximately 15 years ago. Based on the revised popula ons, SD es mates were much lower for dried dis llers grains and corn gluten feed, but corn silage, legume silage, grass hay, and wet brewers grains were similar or slightly reduced compared to the NRC (2001). Dried dis llers grains contained a small subpopulaon that had approximately a mean fat concentra on of 4.5% (Figure 1) and a er removal of that subpopula on, the distribu on of fat concentra on (Figure 2) in the main popula on was more normal and SD was greatly reduced (Table 2). Three large subpopula ons for wet brewers grains were found based on mul ple nutrient concentraons, DM, CP, NDF, and ash. Figure 3 visually shows at
Table 1. Feed popula on mean and SD nutrient concentraons before and a er sta s cal procedure was applied Before Post Before Post Procedure1 Procedure2 Procedure Procedure CP, % of DM Ash, % of DM Soybean meal 405 n3 Mean 49.4 SD4 7.0 Rye grass silage n 7,849 Mean 15.2 SD 3.6 High moisture corn n 17,788 Mean 8.5 SD 0.9 Corn gluten meal n 127 Mean 46.2 SD 22.9 Canola meal n 642 Mean 40.5 SD 4.6
1
least two popula ons for CP concentra ons. The lack of homogeneity inated the SD of wet brewers grains since these large subpopula ons were combined. Varia on in nutrient composi on (i.e., SD) likely aects the percep on of nutri onal risk when u lizing some feeds. As shown in Table 1, the SD in the NRC (2001) for some feeds may not be appropriate for that feed today. If a feed is suspected to have large subpopula ons, SD es mates should be evaluated with more scru ny, and iden ca on of the actual subpopula on you are using would likely reduce your SD es mate. The SD of a feed also should be considered when evalua ng the results of a single sample and when determining the frequency of sampling. If the SD
346 51.1 3.3 6,719 15.3 3.0 16,772 8.5 0.9 63 68.1 5.4 515 41.1 2.4
405 7.0 4.8 7,849 10.3 2.9 17,788 1.5 0.4 127 5.0 2.5 642 8.0 1.2
346 6.9 0.7 6,719 10.3 2.5 16,772 1.5 0.3 63 3.4 1.9 515 7.9 0.7 Figure 2. Sample frequency distribu on of fat concentra on in the dried dis llers grains dataset a er removal of outliers; n = 2,740, mean = 12.5, SD = 1.3. A er removal of outliers and subpopula ons, the distribu on of fat is more normal.
Subpopulation of low fat samples removed??
Figure 1. Sample frequency distribu on of fat concentraon in the dried dis llers grains data set prior to removal of outliers; n = 3,323, mean = 11.8, SD = 2.5.
Before Procedure = Dataset sta s cs prior to removal of outliers and subpopula ons. 2 Post Procedure = Dataset sta s cs a er removal of outliers and subpopula ons. 3 n = Number of samples in dataset. 4 SD = Standard devia on.
100
in a feed popula on is high, using a single sample result should be avoided and more frequent sampling may be needed. If a nutri onist desires to have a consistent diet,
SD of feeds should be considered when determining their inclusion rates, purchase price, and condence in nutrient formula on inputs for each feed. The correla on matrix of nutrients within a feed popula on refers to how the nutrients are jointly expected to vary from sample to sample. For corn silage, the correla on between starch and NDF is -0.88. Therefore, one should expect corn silage with a low starch concentra on to usually have a high concentra on of NDF and vice versa. Observing the correla on matrices provides some understanding of what mul ple nutrient concentraons are expected to be within a single feed sample. For example, if a corn silage sample contains 45% NDF, 39% starch, and 25% dry ma er, based on the corn silage popula on correla on matrix, one should realize that the sample is highly improbable, although if you examined each nutrient independently, you would conclude that the data looks reasonable. Observing multiple nutrients within a feed sample will provide much be er condence that the sample is representa ve or similar to the general popula on. Knowledge of nutrients and their rela onship also provide support for determining how NEL will vary from sample to sample. Weiss (1997) showed that alfalfa silage with a constant 35% ADF does not result in a constant NEL value, but ranges from 1.28 to 1.44 Mcal/kg due to dierent concentra ons of NDF, lignin, and ash. The range in NEL es mates is because other nutrients (e.g. lignin) are not perfectly correlated to ADF. Observing the correla on matrices of feeds (Table 3) shows that using one nutrient to determine economic or energy value is dicult because some correlaons maybe weaker than expected (e.g. CP and NDF) in some feeds. Current ra on so ware uses the mean concentra on of nutrients in feeds to formulate diets. The variability in nutrient composi on of the formulated diet caused by varia on in feed ingredients is not calculated. Therefore, the condence that the supply of dietary nutrients will be similar day to day is unknown and is not considered when formulating diets. If you set the minimum constraint for
Table 2. Mean and standard devia on es mates for several feed popula ons OSU (2012)1 Corn silage 103,119 n2 Mean SD3 Legume silage n Mean SD Grass hay n Mean SD n Mean SD n Mean SD n Mean SD
1
NRC (2001)
OSU (2012)
NRC (2001)
OSU (2012)
NRC (2001)
CP, % of DM 1033 8.8 1.2 8,576 20.0 3.0 4,702 10.6 3.1 879 29.7 3.3 186 23.8 5.7 1,127 28.4 4
NDF, % of DM 103,119 41.9 4.7 12,597 42.0 5.3 15,146 66.3 5.2 2,740 35.3 4.4 347 36.5 4.6 924 50.0 5.2 1033 45 5.3 8,567 45.7 6.5 4,695 64.4 6.2 493 38.8 7.8 122 35.5 6.8 685 47.1 6.8
Fat, % of DM 102,172 3.1 0.3 12,533 3.4 0.5 14,711 2.8 0.6 2,740 12.5 1.3 139 4.1 1.1 260 9.8 1.2 75 3.2 0.5 1,325 3.1 0.7 542 2.6 0.7 464 10 3.4 68 3.5 1.1 N. A.4 5.2 N. A.
7.9 0.9 12,597 20.7 2.3 15,146 10.7 3.3 2,740 29.6 1.6 347 23.2 2.2 924 29.4 4.8
Datasets obtained from two commercial laboratories and outliers were removed. Corn silage data is from the years 2007 to 2011 and the other feed datasets are from the years 2003 to 2011 (majority 2007 to 2011). 2 n = number of samples in dataset. 3 SD = Standard devia on. 4 N.A. = NRC (2001) does not list n or SD for wet brewers grain.
Table 3. Correla on es mates of several feed datasets n1 Corn silage Alfalfa silage Grass hay Dried dis llers grains Corn gluten feed dry Wet brewers grains
1
101
CP at 16% and a least cost algorithm is used, the diet will be formulated to provide 16% CP using the op mal inclusion of each ingredient that results in the lowest cost. If ingredient X is the cheapest source of CP, the inclusion rate will be the maximum within the users specied constraint. However, if ingredient X is highly variable in CP, it would greatly decrease your condence that CP supply of the diet will be similar day to day. If the SD of dietary nutrients is considered, condence intervals can be constructed, and the formulator can have more informa on on how successful the diet will be at supplying 16% CP. If including ingredient X in the diet results in a condence interval of 14.5 to 17.5% CP, but a diet with ingredient Y results in a condence interval of 15.5% to 16.5% CP, you could feed a lower CP diet with ingredient Y. This might reduce feed costs, and you would have a lower risk of supplying a diet less than 15% CP. So ware (PingPongTM) that calculates the SD of nutrients in diets is being developed at Ohio State. The so ware requires accurate es mates of SD for nutrients and accurate es mates of correla ons among nutrients. Correla on matrices are important for es ma ng NEL and MP since they are calculated from mul ple nutrients (NRC, 2001). Predic ng the varia on of these nutrients via PingPongTM allows for es mates of SD and condence intervals of NEL, MP, and metabolizable amino acids such as methionine and lysine (although es mates of varia on for amino acids are based on very li le data). Observing the SD and condence intervals of nutrients will allow nutri onists to evaluate dierent formula on strategies for minimizing varia on risk. Sensi vity of a models op mal solu on to changes in constraints (e.g. nutrient requirements) or inputs (e.g. feed nutrient concentra ons) should also be evaluated since the SD will be known for nutrient supply. Using PingPongTM, we can also es mate the SD for NEL of individual feeds, which can be used to compare and select individual feeds based on their ability to consistently supply energy. In summary, the preliminary results of this study indicate previous SD es mates of some feeds are highly inated, which may change the percep on of nutri onal risk for a par cular feed. The correla on matrices of some feeds show that some important nutrients (e.g. CP, NDF, and lignin) may have a weaker rela onship than expected and illustrate the importance of using a mul component net energy equa on (Weiss, 1997). The correla on matrices also facilitate the evalua on of mul ple nutrient concentra ons and their rela onships within a feed sample to determine the probability of the sample result occurring, which is more robust than observing individual nutrient
Figure 3. Wet brewers grains, n = 924, mean = 29.4, SD = 4.8. Sample frequency distribu on of CP concentra on in the dataset a er removal of outliers. Three large subpopula ons were indicated by the mul variate sta s cal procedure that included DM, CP, NDF, and ash. The sample frequency distribuon visually shows at least two subpopula ons of CP. Determina on of what the three popula ons precisely represent is unknown, but we speculate that the popula ons may represent dierences in brewery manufacturing.
concentra ons to determine the sample result probability. Accurate SD and correla on matrices enable us to evaluate diet formula on stochas cally, which may possibly improve our ability to formulate an op mal diet. REFERENCES
Kohn, R. A. 2006. How can dairy nutri on models deal with uncertainty? J. Dairy Sci. 89(Suppl. 1):382. (Abstr.). NRC. 2001. Nutrient Requirements of Dairy Ca le. 7th rev. ed. Na onal Academy of Science, Washington D. C. St-Pierre, N. R., and C. S. Thraen. 1999. Animal grouping strategies, sources of varia on, and economic factors aec ng nutrient balance on dairy farms. J. Anim. Sci. 77 (Suppl 2):72-83. Weiss, W. P. 1997. Es ma ng the available energy content of feeds for dairy ca le. J. Dairy Sci. 81:830-839. Weiss, W. P., and N. R. St-Pierre. 2009. Impact and management of variability in feed and diet composi on. Pages 83-96 in Proc. Tri-State Dairy Nutri on Conf. Ft. Wayne, IN. Yoder, P. S., N. R. St-Pierre, and W. P. Weiss. 2012. Iden fying improbable feed samples using a mul variate procedure. J. Dairy Sci. 95(Suppl. 2):737. (Abstr.).
102
INTRODUCTION Forages are the most variable in nutrient composi on of all feeds (Mertens, 2006), usually have the highest dietary inclusion, and o en are an economical source of nutrients. High forage diets that rely on one or two sources of forage can have high varia on in the nutrients provided by the forages. Forage quality, (e.g. NDF and CP), has been shown to be highly variable on farms from day to day and month to month (Stone, 2004; Weiss et al., 2012). Sampling error and highly variable feeds can reduce accuracy of ra on formula on, especially if single samples are relied upon. If sampling error is high (possibly 13 to 25% of varia on; Weiss et al., 2012), reliance on single samples for ra on formula on may result in more frequent and larger changes of diet composi on and therefore increase varia on. Our objec ve was to determine whether forage quality varia on (high sampling error or forage varia on) has an eect on milk produc on. We hypothesized that feeding a high forage diet (25% forage NDF; FNDF) that was highly variable day to day to high producing dairy cows would nega vely aect produc on and intake over me and decrease feed eciency. METHODS Six primiparous (DIM = 71) and 18 mul parous (DIM = 74) were the subjects in a replicated La n square experiment balanced for carryover eects with 21-d periods. Within parity, cows were randomly assigned to groups of 3 and then each cow was randomly assigned to treatment sequence within group. Cows were housed in estalls, fed once daily for ad libitum intake with refusals collected and measured daily, and milked twice daily. One average diet was formulated for the three treatments. Two legume silages and one grass silage were fed and on average had approximately the same dietary inclusion across all three treatments. Dry ma er was tested daily and diets were adjusted immediately if needed. Silage NDF was tested (wet chemistry) two to three mes a week, and diets were adjusted immediately based on
mul ple samples if needed. The control diet (CON) was 24.8% FNDF, and we a empted to maintain 24.8% FNDF concentra on for the en re period (Figure 1). The variable diet (VAR) averaged 24.6% FNDF but varied randomly day to day (Figure 1). Random varia on of FNDF was done by Monte Carlo simula on, and the SD was 2.0% FNDF. The varia on of FNDF in the VAR diet was achieved by changing the ra o of legume silage and grass silage fed but not changing forage to concentrate ra o. The over-reac ng diet (ORR) averaged 24.9% FNDF, and we changed FNDF concentra on four mes during the 21-d period on day 5, day 10, day 15, and day 20 of the period (Figure 1). The varia on of FNDF of the ORR diet was achieved by changing the forage to concentrate ra o but not changing the legume to grass silage ra o. The four changes mimicked extreme sampling error causing diets to be reformulated assuming forage NDF changed when in reality the true concentra on of NDF did not change. To compensate for the false change in NDF, the forage to concentrate ra o was adjusted to maintain the dietary target FNDF. The result was an unnecessary change in forage to concentrate ra o and FNDF with no change in the forage composi on. RESULTS AND DISCUSSION Period DMI was aected by treatment (Table 1), and the ORR treatment cows consumed 1.3 lbs more DM than the control cows (P < 0.05). Average intake of cows on the VAR treatment did not dier from the control cows. The ORR cows appeared to consume similar amounts as the CON cows when the FNDF was more than the CON diet. But when the FNDF was less than the CON diet, the ORR cows consumed more than the CON cows. The lack of cumula ve aect for the VAR treatment across 21 days infers the cows were not nega vely aected by the increased forage quality varia on. Daily DMI tended to be more variable (P < 0.10) with a higher coecient of varia on (CV) for the VAR cows, indica ng that the cows
103
responded to higher or lower FNDF diets as would be expected. However, the VAR cows appeared to compensate for reduced DMI on days of high FNDF by ea ng more on the days of low FNDF diets. For example, on day 16, the VAR cows consumed 60.5 lbs of DM and on day 19 consumed 52.9 lbs of DM. We expected cows to respond to diet FNDF concentra on, and this was observed, but no nega ve cumula ve eects were observed. Therefore, we were unable to accept our hypothesis that uctua ng FNDF varia on nega vely aects cows over me. Milk produc on was not aected by treatment and averaged 100.2, 100.3, and 101.2 lbs/day for the CON,
VAR, and ORR treatment diets, respec vely (Table 1). Daily milk yield appeared to respond to daily changes in dietary FNDF concentra ons as we expected. Milk produc on responded to VAR diet FNDF changes (e.g., 106.5 lbs on day 17 and 93.9 lbs on day 21). The CV was increased and aected by the VAR treatment (P < 0.05) but not aected by the ORR treatment. The change in grass to legume silage ra o (VAR treatment) appeared to have more of an eect on DMI varia on than changes in forage to concentrate ra o (ORR treatment). Based on these preliminary results, intake and milk produc on did not respond to daily uctua ons in forage quality. However, the cows were able to produce more milk during transient increases in forage quality to compensate for losses in milk produc on during mes of decreased forage quality. The cows also appeared to compensate for losses in milk produc on due to the ve day changes in forage to concentrate ra o. This suggests that given a high forage diet with adequate eec ve ber, forage quality varia on does not appear to have a cumula ve nega ve eect on intake or milk produc on over 21 days during daily or ve day varia on periods. REFERENCES
Mertens, D. R. 2006. Quan fying assay varia on in nutrient analysis of feedstus. J. Dairy Sci. 89(Suppl. 1):383. (Abstr.). Stone, W. C. 2004. Nutri onal approaches to minimize subacute ruminal acidosis and lamini s in dairy ca le. J. Dairy Sci. 87:E13-E26. Weiss, W. P., D. Shoemaker, L. R. McBeth, P. S. Yoder and N. R. StPierre. 2012. Within farm varia on in nutrient composi on of feeds. Pages 103-117 in Proc. Tri-State Dairy Nutri on Conf. Ft. Wayne, IN.
Figure 1. Treatment FNDF across 21-d period; control (CON) in top panel, variable (VAR) in middle, and over-reac ng (ORR) at bo om. FNDF concentra ons were determined using daily fed data, daily DM of the silages and ve-day composite NDF samples of the silages.
104
Table 1. Milk Produc on, milk composi on, and intake results CON Milk, lbs/d Milk CV, %2 DMI, lbs/d DMI CV, % FNDF intake, lbs/d FNDF intake CV, % Milk Fat, % Milk Fat, lbs/d Milk Prot, % Milk Prot, lbs/d
1 2
VAR 100.3 6.0 56.7 7.1 13.9 9.8 3.60 3.66 2.76 2.80
ORR 101.2 5.4 58.3 6.6 14.4 10.2 3.58 3.64 2.78 2.82
SE 2.6 0.4 1.5 0.6 0.4 1.2 0.18 0.20 0.05 0.09
CON vs. VAR P <1 0.92 0.02 0.66 0.08 0.29 0.02 0.37 0.23 0.50 0.82
CON vs. ORR P< 0.30 0.36 0.04 0.35 0.02 < 0.01 0.48 0.39 0.79 0.81
100.2 5.0 57.0 6.1 14.1 8.1 3.54 3.59 2.78 2.81
Eect of xed eect of treatment, Period and Cow as random eect. CV: Coecient of varia on was calculated by period average SD divided period mean.
105
Notes
106
INTRODUCTION A more accurate tle for this paper would be Under or Over Supplementa on of Minerals and Vitamins Can be Bad for Cows and Dairy Farm Protability, but that is a bit verbose. Although it is becoming dated, the Dairy NRC (2001) is s ll the best source for es mates of requirements of minerals and vitamins. For good health and high milk yields, cows need to be fed the proper amounts of available minerals and vitamins; however, numerous factors (many of which are unknown) inuence the requirements of minerals and vitamins and their availability. Concentra ons of minerals and vitamins in feeds can be extremely variable and o en they are not assayed. Newer data plus the substan al uncertainty regarding the quan ty of minerals and vitamins consumed, the quan ty absorbed (or available), and the quan ty needed by cows under dierent situa ons mandates appropriate adjustments to NRC requirements. CALCIUM The cow can remove Ca from bone so that a short term deciency of Ca has li le eect on lacta ng cows. A long term deciency can result in weak bones and other skeletal problems. However, because most common feedstus contain some Ca and Ca supplements are inexpensive, Ca deciency is not a real world problem in the U.S. A Ca deciency during the dry period and early lacta on period can cause clinical (i.e., milk fever) and subclinical hypocalcemia; however, Ca deciency is rarely the cause of hypocalcemia. Excess Ca is much more common than Ca deciency, and just modest over feeding of Ca (e.g., 115% of requirement) to dry cows increases the risk of hypocalcemia. With lactating cows, substan al overfeeding (>150% of requirement is needed before problems might be seen. Diets with >1% Ca substan ally reduced selenium absorp on by dry cows. This has not been shown in lacta ng cows but because Se status is o en sub-op mal, feeding lacta ng cows diets with more than 1% Ca should be avoided. Diets with >1.5% Ca may reduce feed intake and milk yield (this is more than twice the requirement). Because Ca is inorganic and contains no energy, high Ca diets tend to be lower in energy.
Bottom Line: For lacta ng cows, no data suggest that modest overfeeding (i.e., 120% of requirement) causes any problem and will ensure diets are not decient. Dietary Ca for dry cows should be fed precisely to requirements. PHOSPHORUS For a lacta ng cow, a diet with approximately 0.35 to 0.4% P is usually adequate, and diets based on typical ingredients without any supplemental P usually will contain between 0.3 and 0.4% P. Diets that provide inadequate P to dry cows (an extremely rare event) can increase the risk of hypocalcemia. Milk yields were reduced when cows were fed diets that provided approximately 85% of P requirement (Wu et al., 2000; Wu, 2005), but no study has reported any benets (produc on, reproduc on, or health) when diets provided more P than NRC requirements. Modestly overfeeding P (~120% of requirement) to dry cows signicantly increases the risk of hypocalcemia. Diets with up to ~0.7% P generally have not adversely affected lacta ng dairy cows, but reduc ons in availability of Ca and Mg are possible. There is no reason to feed diets with 0.7% P, but because many byproducts (e.g., dis llers grains, corn gluten feed, wheat midds) are rich in P and are o en economical, diets for lacta ng cows with 0.5 to 0.55% P are not uncommon. This concentra on of P should not have adverse eects on cows, but you should consider increasing dietary Ca and Mg slightly. Manure excre on of P can be an environmental issue, and it will be substan ally greater when cows are fed diets with 0.5% P compared with diets at requirement (~0.38%). Bottom Line: For lacta ng cows, low P diets have reduced milk yields, indica ng that modest overfeeding should be prac ced. Because of environmental concerns, a safety factor of 105 to 110% of requirement is probably adequate when inorganic P is needed. When byproducts provide extra P, diets with 0.5 to 0.55% P are safe but be aware of en-
107
vironmental implica ons. For dry cows, P should be fed precisely to requirements. POTASSIUM Diets for lacta ng or dry cows rarely require K supplementa on to meet requirements (~0.55% for dry cows and 1.1% for lacta ng cows) because haycrop forages are extremely rich in K. Diets based heavily on corn silage, especially corn silage with a high starch concentra on, may be marginal in K. Erdman et al. (2011) reported increased milk yield and improved feed eciency when a corn silagebased diet was supplemented with K (increased diet K from about 1.1% to 1.4%). The K requirement increases under heat stress, but if diets have some haycrop forage supplemental K probably will not be needed. Dry cow diets will almost never need supplemental K, however, excess K is a substan al problem in prac cal diets. For dry cows, increasing K linearly increases the risk of milk fever (Lean et al., 2006). For lacta ng cows, diets with up to 3% K probably will not aect intake or milk but will increase urine output (more manure) and reduce Mg absorp on substan ally. Dietary Mg should be increased whenever dietary K is greater than about 1%: [1] Increase dietary Mg by 0.08 percentage units above NRC (2001) for every 1 percentage unit dietary K exceeds 1% (Weiss, 2004). [2] Increase dietary Mg by 0.02 percentage units above NRC (2001) for every 1 percentage unit dietary K exceeds 1% (Schonewille et al., 2008). Those two equa ons are markedly dierent, possibly because the data bases used were very dierent. In Schonewille et al. (2008) dietary K ranged from 0.7 to 7.5% (most from grasses), and in Weiss (2004) dietary K ranged from 1 to 2.65% (most from alfalfa). A diet with approximately 0.16% Mg usually meets NRC requirement. If the diet contained 1.7% K, dietary Mg should be 0.22% (equa on 1) or 0.17% Mg (equa on 2). Because Mg deciency is more costly than excess Mg, I would err on the side of poten ally overfeeding and use Equa on 1. Bottom Line: In most situa ons, K deciency will not occur but supplemental K may be benecial for high corn silage diets. High K for dry cows (>0.7 to 1%) increases risk of milk fever and high K (>1%) in lacta ng diets reduces Mg absorp on. Extra Mg is needed when high K diets are fed. MAGNESIUM Most typical diets without supplemental Mg will contain
0.15 to 0.2% Mg which is approximately the NRC requirement; however, typical diets tend to be high in Mg absorpon antagonists (e.g., K and soluble N). Inadequate Mg is a clear risk factor for milk fever. A meta-analysis found that increasing Mg up to 0.4% (NRC requirement is approximately 0.13%) linearly (and substan ally) reduced the risk of milk fever, indica ng that dry cow diets should contain up to 3X NRC recommended concentra ons of Mg. Feeding diets with 0.4% Mg should not cause any problem but will increase supplementa on costs. Benets of high dietary Mg for lacta ng cows are much less clear. Some (but not all) older studies (milk yields averaging about 60 lbs) reported increased milk yields and/or milk fat when diets contained 0.3% Mg compared with control diets (approximately 0.2%). Although diets with up to 0.4% Mg will not cause problems to lacta ng cows, current data do not jus fy that rate of supplementa on. Because of poten al milk yield response and poten al antagonism from high K, diets with 0.25 to 0.3% Mg can be jus ed. Bottom Line: Dry cows must be fed diets with 0.3 to 0.4% Mg to reduce the risk of milk fever. Benets of feeding high concentraons of Mg to lacta ng cows is less clear but balancing the cost of overfeeding (only higher supplementa on costs) to poten al increases in milk and milk fat, a safety factor of 1.4 to 1.6 mes NRC is jus ed. SODIUM AND CHLORIDE All diets will need supplemental Na (i.e., salt), but because salt is inexpensive NaCl deciencies are extremely rare. Over feeding Na via excess salt or in combina on with Na bicarbonate is common. Feeding Na at approximately 2X NRC requirements from a variety of sources (salt + sodium sulfate + sodium sesquicarbonate) over an en re 308-d lacta on had no eect on milk yield (lacta on average milk yield = 70 lbs, 3.5% fat), composi on, or intake (Clark et al., 2009). Some mes feeding excess Na (from buers) can increase milk fat, but this is dependent on starch, ber, forage, etc. Cows can tolerate high Na diets (up to about 1% of diet) as long as non-saline water is readily available. Diets with high Na will increase urine excre on and manure output. Bottom Line: Diets with excess Na (approximately 0.5 to 0.6%, which is about 2X NRC) from buers will not cause any problem for cows if clean water is readily available, but milk yield responses are not consistent.
108
SULFUR Sulfur is essen ally a nutrient for rumen bacteria, not the cow. Diets with inadequate S can reduce ber diges bility, microbial protein synthesis, and feed intake. Diets with approximately 0.2% S are usually adequate to prevent these problems, and in most situa ons dairy cow diets are about 0.2% S without any supplemental S. Because most S in diets is found in amino acids, diets with lower protein also tend to have lower S concentra ons and may need some supplemental S. Diets with supplemental S from inorganic sources during the prefresh period reduce the risk of hypocalcemia. The amount needed depends on concentra ons of Na, K, and Cl. Diets with more than 0.25 to 0.3% S can cause problems when fed for long periods of me (months, not weeks), and diets with more than about 0.5% S can cause problems even when fed for short periods of me. With the excep on of prefresh diets (i.e., low DCAD), there is no reason to increase dietary S above 0.25% with supplemental S; however, because of the increased use of dis llers grain and because some water can be very high in sulfate, cows o en consume diets (or diets equivalent) with more than 0.25% S (water with 350 mg/L sulfate-S is equivalent to increasing dietary S by 0.2 percentage units). A diet with 35% dis llers may have more than 0.4% S. Diets with >0.25% S reduce Cu and Se absorp on, and supplementa on rates for those two minerals should be increased or more bioavailable forms should be fed. Although very rare for dairy ca le, diets with 0.35 to 0.4% S have caused increased mortality in feedlot ca le (mostly via polioencephalomalacia). High starch diets increase the risk of S toxicity; dairy ca le fed moderate starch diets (relave to feedlot diets) probably can handle diets with 0.5% S. Bottom Line: Supplementa on of S for lacta ng cows is usually not required with the possible excep on of lower protein diets. Feeding diets with S concentra ons of 0.3 to 0.4% during prefresh period can reduce hypocalcemia. Excess S is a much greater risk than S deciency. A empt to keep diets (including water) to 0.25% S or less. Increase Cu and Se by at least 1.2X NRC if dietary S is >0.25%. TRACE MINERALS Cobalt Cobalt is required by ruminal bacteria, and the current NRC requirement is 0.11 mg/kg of diet DM. Data published since NRC suggest that increased concentra ons of Co may be benecial (Stangl et al., 2000; Kincaid et al., 2003). Depending on the response criteria, cows, including dry cows, should be fed diets with between 0.2 and 0.9 mg/
kg Co. Older cows may benet more from extra Co than younger cows. The Co eect is likely via vitamin B-12, and factors aec ng responses to B-12 are numerous including supply of methionine, choline, and folic acid. Because of the uncertainty of response, supplementa on rates of Co between 0.2 to 0.5 mg/kg are a good compromise. Ca le can tolerate high concentra ons of Co (at least 20 mg/kg diet). Copper The current NRC requirement for Cu is about 14 mg/kg. Diets that are marginally low in Cu are a risk factor for increased infec ous disease, and feeding diets that are excess in Cu increases the risk for sudden death. Most typical diets contain about 10 mg/kg Cu from basal ingredients, which means about 4 or 5 mg/kg supplemental Cu will be required. Numerous real-world antagonists to Cu absorp on can be present including S (>0.25% of diet), reduced (as opposed to oxidized) Fe, Mo, excess Zn, and pasture consump on (most likely caused by soil inges on). Because of poten al antagonists and because of high variability in Cu concentra on in feeds a safety factor of about 20% can be easily jus ed (i.e., feed diets with about 17 mg/kg Cu). With high S water and substan al inclusion of dis llers grains you should increase another 20 to 40% (20 to 25 mg/kg) and consider using organic source of Cu. If water is also high in Fe (>3 mg/L) addi onal Cu may be needed. Diets with more than 25 mg/kg total Cu are rarely jus ed and long term feeding of excess Cu will cause it to accumulate in the liver. Long term feeding (many months) of diets with 37 mg/kg total Cu resulted in clinical Cu toxicity in Holstein cows (Bradley, 1993). Jersey cows appear to accumulate Cu more eciently than Holstein which puts them at increased risk of Cu toxicity. Iron The NRC iron requirement is approximately 20 to 25 mg/ kg, and most unsupplemented diets are well in excess of that concentra on. However, much of the iron in feeds, especially forages, is poorly absorbed because the Fe is actually from soil contamina on. When evalua ng dietary Fe, the concentra on of Fe in forages should be set at 0. If the remaining basal ingredients provide about 25 mg/kg Fe, the diet is probably adequate. If supplemental Fe is needed, use reduced sources of Fe (e.g., iron sulfate). Because of antagonism of Fe to Cu and Zn and because high Fe increases the requirements for Se and vitamin E, supplemental Fe should be limited to 50 mg/kg. A study with early lactaon cows reported no milk yield response when 30 mg/kg supplemental Fe was fed (Weiss et al., 2010). Iron in water at levels as low as 0.3 mg/L can reduce water consump on but is not adequate to cause other problems.
109
Manganese The NRC requirement for Mn is approximately 18 mg/kg, and data published a er NRC clearly show that the NRC requirement is much too low. In beef heifers, feeding diets with 18 mg/kg Mn resulted in clinical Mn deciency (Hansen et al., 2006). Using diges bility trials to es mate Mn maintenance requirements, we es mated total Mn requirements were between 30 and 50 mg/kg (Weiss and Socha, 2005). Because cows can be fed diets with 1,000 mg/kg of Mn with no eect, feeding diets with 40 or 50 mg/kg of Mn poses no risk and will essen ally eliminate any poten al of deciency. Selenium The benets of feeding diets adequate in Se to the health of dairy cows are unequivocal. All diets fed to dairy animals (calves, heifers, dry cows, lacta ng cows) in the Eastern U.S. should contain 0.3 mg/kg of supplemental Se (this is the maximum allowed by FDA regula ons). Basal ingredients typically contain about 0.1 mg/kg Se so that total diet is about 0.4 mg/kg. Cows have been fed diets with as much as 12 mg/kg of Se (from selenite) for 4 months without any nega ve eects. Because of legal constraints, a safety factor for Se cannot be recommended. If feeding 0.3 mg/kg of supplemental Se from selenite is not adequate, replacing some or all of the supplemental Se with Se-yeast may help. On average true absorp on of Se from inorganic sources is about 50% and about 60% for Se from Se-yeast (calculated from Walker et al., 2010). Based on blood, enzyme, and true absorp on, Se from Se-yeast is about 20% more available than Se from selenite when antagonists are not present. The dierence may be greater for diets with high S. Zinc The NRC requirement for Zn is between about 40 and 55 mg/kg (depending on milk yield), and there is no data showing substan al benets from feeding more than this. Ca le can be fed very high concentra ons of Zn (>500 mg/ kg) without nega ve eects. High Zn (approximately 100 mg/kg), however can reduce Cu absorp on and therefore should be avoided. Diets with a reasonable safety factor (1.2X NRC) to account for uncertainty in Zn concentra ons of basal ingredients are jus able and pose no risk. Bottom Line: The NRC requirements for Co and Mn are too low based on recent studies. Dietary Co should be 0.2 to 0.9 mg/kg (0.4 mg/kg seems a good compromise). Dietary Mn should be between 30 and 50 mg/kg. The Zn and Fe requirements appear adequate. For Zn a small safety factor of 20% is
jus ed based on varia on in Zn of basal ingredients. The Fe contribu on from forages should be ignored when compu ng supply of Fe, but otherwise the NRC Fe requirement is adequate. Because high Fe can be a problem, a safety factor is not recommended. The maximum legal amount of Se should be supplemented. Ini ally this should be from inorganic or a blend of inorganic and Se-yeast (this is based solely on economics). If Se status is s ll not adequate, increase the amount of Se-yeast and reduce or eliminate the inorganic Se. The NRC requirement for Cu is adequate for many situa ons but because of uncertainty a safety factor of 20% is jus ed. With high S diets (e.g., dis llers grains) or water, Cu should be increased another 20 or 40%. Diets with more than 25 ppm total Cu are rarely jus ed and long term feeding (months) may cause problems. FAT SOLUBLE VITAMINS Vitamins were discussed in detail at the 2005 mee ngs. Requirements for vitamins are dicult to establish because measuring vitamins in feeds is some mes very dicult, vitamins can be destroyed or synthesized in the rumen, and responses to changes in vitamin supply is o en very subtle and may take months to observe. Vitamin A The NRC (2001) requirement for vitamin A is about 82,000 IU/day for a dry cow and 72,000 IU/day for a lacta ng cow. Cows fed adequate vitamin A have reduced mas s, metri s, retained fetal membranes, and abor ons compared to cows fed no supplemental vitamin A, but there are no data sugges ng any health benet to feeding more than the current NRC recommenda on. However, cows fed 170,000 IU/day (about 2X NRC) during the dry period and rst 6 wk of lacta on produced more milk (88 vs. 77 lbs/day) than cows fed 50,000 IU/day (0.7X NRC) during that period (Oldham et al., 1991). Conversely, feeding very high vitamin A during the dry period (550,000 vs. 80,000 IU/day) reduced total yield of energy-corrected milk during the rst 100 days of lacta on (Puvogel et al., 2005). Vitamin A is among the least stable vitamins and loss of ac vity can be ~10%/month depending on storage condi ons (Shurson et al., 2011). Vitamin D Because of the interest in vitamin D for human health, vitamin D for dairy cows is also being re-evaluated. Although adequate data are not available to quan ta vely adjust the current NRC vitamin D requirement (approximately 20,000 IU/day), data are available sugges ng poten al benets from increasing vitamin D supplementa on. Stud-
110
ies with humans and limited research with bovine cells have shown that vitamin D has important roles in immune func on and that blood concentra ons of 25-OH vitamin D required for maximal immune response was greater than concentra ons required for op mal Ca metabolism in humans (Lippolis, 2011). Whether this is true for dairy cows will require new studies. Dairy cows housed inside without exposure to sun and fed vitamin D at NRC rates had signicantly lower (about 30%) plasma concentra ons of 25-OH vitamin D than cows fed no supplemental vitamin D but housed outside in the summer with extensive sun exposure (Hymller et al., 2009). The op mal concentraon of plasma 25-OH vitamin D is not known but current supplementa on rates do not provide for maximal concentra ons. Some older studies (see page 165 in NRC) found improved reproduc on measures and increased milk yield when vitamin D was supplemented at about 40,000 IU/day (i.e., 2X NRC). Excess (~2,000 IU/day for months) dietary vitamin D can cause calcica on of so ssues. Vitamin E The NRC requirement for vitamin E is about 500 IU/day for lacta ng cows and 1,000 IU/day for dry cows. Essen ally no new research has been conducted evalua ng vitamin E requirements for lacta ng cows, but several experiments have been conducted with dry cows. Increasing vitamin E supplementa on during the transi on period (2 or 3 wk prepartum un l 1 or 2 wk post partum) has improved measures of immune func on or improved mammary gland health in several, but not all, studies. Supplementaon rates during the transi on period ranged from 2,000 to 4,000 IU/day. No study has shown any nega ve eects of high supplementa on rates in the prefresh period. Therefore the only known cost is the cost of the addi onal vitamin E, but because the supplementa on period is short (a few weeks) and the poten al payo is high (reduced mas s and retained placenta) increasing vitamin E supplementa on during the prefresh period is jus ed. Increasing vitamin E supplementa on during the en re dry period, however is not jus ed. A recent study (Bouwstra et al., 2010) evaluated the eects of feeding 3,000 IU of vitamin E/day (controls were fed approximately 130 IU/ day) during the dry period. On 3 of 5 farms, more cases of mas s occurred when cows were fed high vitamin E, and on 2 farms li le dierence in mas s was observed between treatments. Overall, they reported that feeding high vitamin E increased the risk of mas s by 1.7X compared with the control. Although I have some technical concerns regarding the paper (e.g., methods used to diagnose mas s), the paper clearly shows no benet of increasing amounts of vitamin E and poten ally it might
have nega ve eects. Vitamin E supplementa on during the dry period should be limited to 1,000 IU/day. Bottom Line: Poor stability of vitamin A and poten al for increased milk yield with extra vitamin A jus fy supplemen ng vitamin A at 1.1 to 2X NRC (e.g., 80,000 to 150,000 IU/day). Supplementa on in excess of 150,000 IU/day cannot be jus ed based on available data. The current vitamin E levels are adequate except during the prefresh period when 2,000 to 4,000 IU/day can be jus ied. Much more research is needed with vitamin D, but circumstan al evidence suggests benets from increasing supplementa on up to 2X NRC or ~40,000 IU/day. Because of poten al eects on milk fever and complete lack of data, dry cows should be fed at approximately NRC levels for vitamin D. WATER SOLUBLE VITAMINS Bio n, choline, and niacin are the only water soluble vitamins commonly supplemented to dairy cows. The data suppor ng rou ne bio n supplementa on (approximately 20 mg/day) is strong. A meta-analysis (Lean and Rabiee, 2011) determined that a 2.9 lb/day increase in milk yield is expected with bio n supplementa on, and most studies evalua ng hoof health have reported benets with supplemental bio n. Based on current prices, these responses would be protable. Data suppor ng the response to supplemental rumen-protected choline (RP-Choline) is also good. A meta analysis found an expected increase in milk yield of about 4.4 lbs; however, based on their sta s cal model, a supplementa on rate of about 20 g of RP-choline (actual choline, not product) was required to obtain this response (Sales et al., 2010). Feed intake is o en not stas cally altered, but there is no reason to think choline improves eciency, so an increase in DM intake of 2 to 2.5 lbs/day should be considered in any economic analysis. The milk yield response to RP-choline is substan al but RP-choline is expensive so that its use may not always be protable (especially at the 50 g/day rate). Most experiments with RP-choline use early lacta on cows because they are most likely to be limited in choline. We do not know whether these responses will occur in later lacta on. If pens contain cows at several stages of lacta on, rather than just early lacta on, overall response may be less and this should be considered when deciding on poten al protability. Supplemental RP-choline (only studied at the 50 g/day rate) during the prefresh and fresh period may have some value in reducing fa y liver and type 2 ketosis (Cooke et al., 2007). Ketosis/fa y liver is a very expensive disease, and if RP-choline reduces its incidence it would almost denitely be protable. Longer term approaches
111
to reducing incidence of these diseases are available (e.g., proper management of body condi on and DM intake during the dry period) and should be pursued, but if ketosis/ fa y liver is a problem supplementa on of RP choline at 50 g/day during the prefresh period should be considered. A meta-analysis (Schwab et al., 2005) found li le benet of supplemental niacin at 6 g/day, but at 12 g/day milk protein yields were increased and, depending on milk price, it may be protable at mes. Rumen protected niacin is now available and some data show that it may help cows under heat stress condi ons, but it is expensive and responses in milk yield have generally been quite small. Bottom Line: Cows should be supplemented with bio n (20 mg/day); RP-choline increases produc on, but because of cost the decision to supplement depends on milk price. Current data do not jus fy rou ne supplementa on of niacin. REFERENCES
Bouwstra, R. J., M. Nielen, J. A. Stegeman, P. Dobbelaar, J. R. Newbold, E. H. J. M. Jansen, and T. van Werven. 2010. Vitamin E supplementaon during the dry period in dairy ca le. Part I: Adverse eect on incidence of mas s postpartum in a double-blind randomized eld trial. J. Dairy Sci. 93:5684-5695. Bradley, C. H. 1993. Copper poisoning in dairy herd fed a mineral supplement. Can. Vet. J. 34:287-292. Clark, J. H., R. A. Christensen, H. G. Bateman, and K. R. Cummings. 2009. Eects of sodium sesquicarbonate on dry ma er intake and produc on of milk and milk components by Holstein cows. J. Dairy Sci. 92:3354-3363. Cooke, R. F., N. S. Del Rio, D. Z. Caraviello, S. J. Ber cs, M. H. Ramos, and R. R. Grummer. 2007. Supplemental choline for preven on and allevia on of fa y liver in dairy ca le. J. Dairy Sci. 90:2413-2418. Erdman, R. A., L. S. Piperova, and R. A. Kohn. 2011. Corn silage versus corn silage:Alfalfa hay mixtures for dairy cows: Eects of dietary potassium, calcium, and ca on-anion dierence. J. Dairy Sci. 94:5105-5110. Hansen, S. L., J. W. Spears, K. E. Lloyd, and C. S. Whisnant. 2006. Feeding a low manganese diet to heifers during gesta on impairs fetal growth and development. J. Dairy Sci.89:4305-4311. Hymller, L., S. K. Jensen, H. Lindqvist, B. Johansson, M. O. Nielsen, and E. M. G. Nadeau. 2009. Supplemen ng dairy steers and organically managed dairy cows with synthe c vitamin D3 is unnecessary at pasture during exposure to summer sunlight. J. Dairy Res. 76:372-378. Kincaid, R. L., L. E. Lefebvre, J. D. Cronrath, M. T. Socha, and A. B. Johnson. 2003. Eect of dietary cobalt supplementa on on cobalt metabolism and performance of dairy ca le. J. Dairy Sci. 86:1405-1414. Lean, I. J., P. J. DeGaris, D. M. McNeil, and E. Block. 2006. Hypocalcemia in dairy cows: Meta-analysis and dietary ca on anion dierence theory revisited. J. Dairy Sci. 89:669-684. Lean, I. J., and A. R. Rabiee. 2011. Eect of feeding bio n on milk produc on and hoof health in lacta ng dairy cows: A quan ta ve assessment. J. Dairy Sci. 94:1465-1476.
Lippolis, J. D. 2011. The impact of calcium and vitamin d on the immune systems. Mid-South Ruminant Nutri on Conf., Grapevine, TX:29-34. Na onal Research Council. 2001. Nutrient requirements of dairy ca le. 7th rev. ed. Natl. Acad. Press, Washington DC. Oldham, E. R., R. J. Eberhart, and L. D. Muller. 1991. Eects of supplemental vitamin A and B-carotene during the dry period and early lacta on on udder health. J. Dairy Sci. 74:3775-3781. Puvogel, G., C. R. Baumrucker, H. Sauerwein, R. Ruhl, E. Ontsouka, H. M. Hammon, and J. W. Blum. 2005. Eects of an enhanced vitamin A intake during the dry period on re noids, lactoferrin, IGF system, mammary gland epithelial cell apoptosis, and subsequent lacta on in dairy cows. J. Dairy Sci. 88:1785-1800. Sales, J., P. Homolka, and V. Koukolova. 2010. Eect of dietary rumenprotected choline on milk produc on of dairy cows: A meta-analysis. J. Dairy Sci. 93:3746-3754. Schonewille, J. T., H. Everts, S. Ji akhot, and A. C. Beynen. 2008. Quanta ve predic on of magnesium absorp on in dairy cows. J. Dairy Sci. 91:271-278. Schwab, E. C., D. Z. Caraveillo, and R. D. Shaver. 2005. Review: A metaanalysis of lacta on responses to supplemental dietary niacin in dairy cows. Prof. Anim. Sci. 21:239-247. Shurson, G. C., T. M. Salzer, D. D. Koehler, and M. H. Whitney. 2011. Eect of metal specic amino acid complexes and inorganic trace minerals on vitamin stability in premixes. Anim. Feed Sci. Tech.163:200-206. Stangl, G. I., F. J. Schwarz, H. Muller, and M. Kirchgessner. 2000. Evaluaon of the cobalt requirement of beef ca le based on vitamin B-12, folate, homocysteine and methylmalonic acid. Brit J Nutr. 84:645-653. Walker, G. P., F. R. Dunshea, J. W. Heard, C. R. Stockdale, and P. T. Doyle. 2010. Output of selenium in milk, urine, and feces is propor onal to selenium intake in dairy cows fed a total mixed ra on supplemented with selenium yeast. J. Dairy Sci. 93:4644-4650. Weiss, W. P. 2004. Macromineral diges on by lacta ng dairy cows: Factors aec ng diges bility of magnesium. J. Dairy Sci. 87:2167-2171. Weiss, W. P., J. M. Pinos-Rodrguez, and M. T. Socha. 2010. Eects of feeding supplemental organic iron to late gesta on and early lacta on dairy cows. J. Dairy Sci. 93:2153-2160. Weiss, W. P., and M. T. Socha. 2005. Dietary manganese for dry and lacta ng holstein cows. J. Dairy Sci. 88:2517-2523. Wu, Z. 2005. U liza on of phosphorus in lacta ng cows fed varying amounts of phosphorus and sources of ber. J. Dairy Sci. 88:28502859. Wu, Z., L. D. Sa er, and R. Sojo. 2000. Milk produc on, reproduc ve performance, and fecal excre on of phosphorus by dairy cows fed three amounts of phosphorus. J. Dairy Sci. 83:1028-1041.
Originally published as Minerals and Vitamins for Dairy Cows: Magic Bullets or Just Bullets? in Proc. 2012 Herd Health and Nutri on Conferences, Syracuse, NY and West Lebanon, NH, pages 27-36.
112