Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Companson of New Protein Systems for Ruminants INTRODUCTION Several new theoretical protein systems have been proposed that have potential application to feeding ru- minants. These new systems require several additional concepts that the current National Research Council (NRC) systems, such as that for dairy cattle (1978), do not. Dietary intake crude protein (Ilk is either degraded (DIP) in the rumen, with partial or total conversion to bacterial and protozoa! crude protein (BCP), or passed from the rumen as undegradec! intake protein (UIP). Microbial growth in the rumen requires either DIP, which may include either dietary nonprotein nitrogen (NPN), or a net ruminal influx of endogenous urea as crude protein (RIP) from either saliva or across the ru- men wall. Production of BCP associated with microbial growth is related to energy fermented in the rumen and is expressed most commonly as a function of apparently fermented organic matter (FOM). Excess DIP increases the concentration of ruminal ammonia and increases the ruminal efflux of ammonia as crude protein (REP) by absorption and passage. Production of BCP repre- sents both a protein requirement and a subsequent source of protein for the tissue needs of the cow. Effi- ciency of ruminal utilization of protein is 1.0 when DIP exactly meets the BCP need. When DIP equals BCP need, RIP must equal REP. The theoretical efficiency of tissue utilization of protein is maximum when BCP and UIP exactly meet the cow's tissue need. The theoretical efficiency of producing milk is maximum when both ef- ficiencies in rumen and tissue utilization are maximum, i.e., neither DIP nor UIP is excessive. The new concepts require that ruminal undegrada- bility (UIPIP) must be specified in addition to a total tissue protein requirement. As higher milk production requires more total protein and available DIP exceeds that converted to BCP, more undegradable protein 7 sources increase the efficiency of N use. A similar situa- tion prevails in the rapidly growing animal. Objectives of this paper are (1) to present a compara- ble tabulation of factors used in five U. S. and five Euro- pean factorial systems that are static or partly static sys- tems; (2) to calculate, as an example, minimal dietary protein and optimal UIPIP based on factors for a 600-kg cow producing from 10 to 40 kg of milk per day; and (3) to compare the expected flow of N into the small intes- tine and into the sinks of milk, urine, and feces. Papers previously published (Verite et al., 1979; Waldo, 1979; Chalupa, 1980a; Verite, 1980, Waldo and Glenn, 1982) have compared factors of some systems. Yerite et al. (1979) compared the protein concentration in dry mat- ter (DM) required for milk production from 15 to 35 kg, and Geay (1980) compared protein required for growth based on several systems. Waldo and Glenn (1982) com- pared the distribution of dietary protein and N to milk, urine, and feces in five European systems. FACTORS IN AVAILABILITY OF ABSORBED PROTEIN Factors from 10 systems are compared in Tables 1 and 2. The current NRC dairy system (Swanson, 1977, 1982; NRC, 1978) is included as a reference. Four new U.S. systems have been proposed. These systems will be called Burroughs (Burroughs et al., 1971, 1974, 1975a, b; Trenkle, 1982), Satter (Roffler and Satter, 1975a,b; Satter and Roffler, 1975; Satter, 1982), Chalupa (1975b, 1980a), and Cornell (Fox et al., 1982; Van Soest et al., 1982~. Two new European systems the ARC system in Great Britain (Roy et al., 1977, ARC, 1980) and the PDI grele system in France (Verite et al., 19793 are official proposals within each country. Kaufmann (1977b, 1979) has proposed a system in Ger
~ Ruminant Nitrogen Usage many, and Landis (1979) has proposed a system in Switzerland. Haselbach (1980) and Schurch (1980) also presented discussions relative to the proposal of Landis. Danfaer (1979) has outlined many factors in a model of protein utilization from Denmark. Danfaer et al. (1980), Madsen et al. (1977), and Molter and Thomsen (1977) also presented data from Denmark that will be used for some values not specified in the model of Dan- faer. Danfaer does not propose a system but gives some factors in protein utilization. The Cornell system intro- duces dynamic factors. The new systems require specification of several new factors to describe availability of protein at the intes- tine. The division of IP into DIP and UIP fractions must be specified. The proportional production of BCP from DIP (BCPDIP) must be specified. Production of BCP must be related to dietary energy, which frequently is expressed as either FOM or apparently cligested organic matter (DOM). The division of BCP into nucleic acid N as crude protein equivalent (NCP) and bacterial and protozoa! true protein (BTP) must be specified. If theo- retical urinary and fecal N excretion are to be calcu- lated, the digestible nucleic acid N as crude protein equivalent (DNP) must be specified. Intake Protein per Unit of Dry Matter The required IP concentration in the dietary DM (IPDM) in the newer systems is still variable and directly related to milk production as in the NRC system based on total protein or crude protein (CP) (Table 1~. How- ever, in the newer systems, high milk production can be sustained with less IP if UIPIP also is increased, and dry cows in low production can be fed less IP if UIPIP is decreased. Undegraded and Degraded Intake Protein per Unit of Intake Protein All of the new systems consider the dietary IP to be divided into undegraded (UIP) and degraded (DIP) fractions, with the proportional division represented by UIPIP and DIPIP. This division is most generally con- sidered continuously variable (Table 1~. The Chalupa and ARC systems place all proteins into four classes hav- ingUIPIPofO.20 + 0.10,0.40 + 0.10,0.60 + 0.10, and > 0.70. The Cornell system defines an indigestible intake protein (IIP) fraction in IP. The Cornell system also further subdivides DIP into soluble and potentially degradable subfractions and includes a dynamic degra- dation of the potentially degradable subfraction. The proper division of IP into UIP and DIP is the ma- jor new input required for these new systems to be effec tive in practice. The derivations or sources of these data are not always specified. Burroughs et al. (1975b) have an extensive tabulation of feedstuff degradabilities for their system. Satter is collecting in vivo data for com- mon dairy feeds used in the north central United States. Chalupa (1980a) has accepted the ARC tabulation of feeds into four classes. The PDI system (Demarquilly et al., 1978) uses an extensive tabulation of solubility and in vitro fermentability. Verite and Sauvant (1981) pro- posed equations for calculating digestible protein reach- ing the intestine from IP and protein of concentrates sol- uble in salt solution. The other systems propose neither a source of undegradability data nor an analytical method for obtaining the data. Crude Bacterial Protein per Unit of Degraded Intake Protein Six systems assume no loss or gain of protein in the production of BCP from DIP (BCPDIP), or BCPDIP = 1.00, but the ARC and Chalupa systems assume BCP- DIP of dietary urea to be 0.80 (Table 1~. The Satter sys- tem assumes an RIP equivalent to 12 percent of dietary IP and 90 percent utilization of ruminal ammonia, or bacterial and protozoa! crude protein/ruminally avail- able nitrogen as protein (BCPRAP); if degradability = 0.7,then the netinfluxisO.12 - 0.7~1.0 - 0.9) = 0.05 for BCPDIP = 1.05. The Danfaer model assumes BCPDIP = 0.90. The Cornell system proposes a range of BCPDIP from 0.5 to 0.9. Such low efficiencies in- crease implied protein requirement by increasing esti- mates of ammonia absorption and urinary excretion. It seems unrealistic to assume that degradation of pro- tein can be optimized for high milk production so that conversion of DIP to BCP fully attains 1.00, even though this is the goal of any ideal protein system. Wa- ter passage from the rumen will elute some ammonia that must be replaced. The dynamic model of Baldwin et al. (1977a) indicates that one-fourth of ammonia leaves the rumen by passage and three-fourths by ab- sorption in a 40-kg sheep fed a 22.5 percent CP alfalfa hay at 37.9 g of DM/h. Kaufmann (1977a) found duode- nal N (g/100 g of feed N) = 34.2 + 1032.711P (percent of dietary DM) in 45 observations on lactating dairy cows; this equation implies a gain of total N in the ru- men below 15.7 percent dietary IP and a Toss of total N above 15.7 percent. Hogan (1975) described protein reaching the intestines (gig IP) = 0.33 + 0.18 DOM intake with r = 0.96 using sheep; assuming that DOM = 0.67 DM, this equation implies a gain of protein in the rumen below 14.2 percent IP and ~ loss of protein above 14.2 percent IP. Oyaert and Bouckaert (1960) de- scribed the percentage of protein N intake absorbed in
- ~ - ~ o c-' ~ - ;> cc ~ in So .= ~ ~ ~ - ~ - O tiC CD D 1) it, ~ 0 m o ~00 Z~ o ._ - ._ Ct O 'ct ._ sol o Cal Ct Cal - o Ct - - x ~ ~ O ; ~OC A., . V it. C,= 11 ~ OO Cal O Cal 0 C 0 CCG) ~ ~ fir <73 A,, · - ~ . . ~ ~ ~ > _ _ $ O ~ ~ >` _ .- C=~ ~ _ _ O .~ 0= . O~ 0_ _ it. ~. _ SO ~ a c s D~-;~ Q m E ~ a .= O ~. ~ C ~ O _ ~C o ~ D .~ c ~c~ E ~ ou ' ~ 2 ' o ~-=-~ ~ D :: L~e~ C~00 L~ C`1CO - . ~ CC ~ U: ~ C5 - 00 _ . - a 11 C`1 - O P" ., c~ L~ C~OO C~ O O C~ _OC ~ ~V: ~ V: :> OCO CC C= ~O ~_ ~ a ~> - a ~11 ~ L - ) O) O L~ =0 C~ o ~ Z =. . ~· ~ · · _ · C~ O O ~ O - OC 00 ~ ~ ,^ U ~. U: C C - - - Ilc~ O ~ O O ~ O ;> ~ o - +1 o C~ . o - . +1 . . o .6 - ._ Ct Z C5 P~ C: & & ~ ~ Z ~ o OC & z ~ ~c az ~a~ . ~,^ ;~ - a 11 c~ ~ z O r ~ ~4 ~. v, V3 c ~ ~: . - ~c ~ c " ~ c5 o =~Ic5 a4 ~ cts C) ~ 3 t4 C, Ct ~ C) ~ D c; c, C C `~, 3 _ _ _ ~ _ _ ~ ·_ O ~ ~ o4= C) ~ e~ ~o C) ~ D a, 5, oo _ = ~_ C w - O D ~ ~ , ~ W ~ :,$ ~ 05 D W D 'a; o D D W E C ~C ~ b0 C) ~ O C) O ~ COC ~ .= =,~£ ~ _ 0~ bC ~ C) 9 , ~ . ~v (D _ c - ~2 c ~.e E .`A E E C O ° 3- 3 ~ O ~ 0m ~ ~C) _ .' _ ~C 3 ~ ~ _ ~: ~C ~ ~- ~- ~Ct ~ ._ ~ ~C -=c E o , ~ ~ ~Z 'o ~ ~=^ ~ ~c~ ~:}) 3~ o . - ~ ~^ C o ~o O ~ ~ D ~ ~ _ =~ ~o -3 ~ C) O E ~ ~ + o W D ~ - S cz ~ _ _ ~ ' - c:SO C) >~ o o ~c, -= e ,~ ~ - ' &~-= ~ ~ ~ 3 = ~, 3 c E c ~ ° e & ~ c~ ' ~ ~ ~ ~ - , E ° ~ _ D ~ ~ ~ E ~ ~ Z 0 =~ w ~ - - .= ~ _ c ._ o - o vc 3 V, Ct C .0 U, ~ C tV: ,_ ~ ~ =5 V C Ct ~Le: ~ ~ ~; ~_ ~ 00 D ~11 ~a v~ O v ~ - E 0 3 _= L U W a c u L. ~O ~ ~ ° i' ~ !; ~ ~ ~ ~ 51!
~a ~ ~ a ; ~_ _ ~ l A !1 ) c~ ° ; z 0 z _ ~ z ~OF 0 . 9 0 y _ . 3 ~ ~ ~ ~ ~Z ~ Z ~Co ~To ~ ,Yz _ YE = · Z ~ O ' ~ :) ~ ~ ~) | - ~ C A _ ~C I o _ 11 L ~~ - 1 1 1 II ~o l ~ =~ C _ . t = c' Polo Y ~ ' 0 8 0 ~ ~ ~ ~Z ~ ~° Y ~ O ~- - ~ y 0 ~ . ~ · C _ z ~ 8 8 - _ _ E ~ x 3 ~ Cal ~: cat Cal ~ ~ ~ ~ ~ c _ E ~_ ~; , ~ ; Y - ~c ~ ~_ ~i2 E ~ z z e ~Z E z Z _ 8 ~' _ _ u ~c E ° ~- 1; - 8 ~= c ~ ~j IO
the rumen = - 20.8 ~ 1.62 NH3 - N (mg/100 ml of rumen fluid) with r = 0.91 for n = 10 diets fed to sheep; this equation implies a gain of protein in the rumen be- low 13 mg NH3 - N/100 ml and a loss above 13 mg NH3 - N/100 ml. Crude Bacterial Protein per Unit of Fermented Organic Matter The most common expression of BCP to microbial en- ergy fermentation is as a function of FOM (BCPF'OM). Five systems assume proportional BCPFOM that range from 0.15 to 0.25 (Table 1~. The Satter, PDI, and Lan- dis systems relate BCP to DOM (BCPDOM) and specify no proportions for BCPFOM. The Cornell system is dy- namic for this factor. Thomas (1973), in his summary, calculated a mean of 0.20 for 27 diets fed to sheep and cattle; Kaufmann (1977a) calculated a mean of 0.20 for 11 diets fed to dairy cows. Diet does affect this factor. McMeniman et al. (1976) calculated means of 0.1375 for diets with high concentrates and 0.1962 for diets with fresh or dried forages. Chamberlain and Thomas (1980a) present a range from 0.0625 to 0.1375 for diets consisting of only hay-crop silages. Fermented Organic Matter per Unit of Digested Organic Matter When BCP is based on FOM, it is necessary to assume a proportional FOM per unit of DOM (FOMDOM). These proportions range from 0.65 to 0.68 except in the Burroughs system, for which 0.52 is used (Table 1~. The Cornell system is dynamic for this factor. Crude Microbial Protein per Unit of Digestible Organic Matter The eight fully static systems assume proportional BCP per unit of DOM (BCPDOM) that range from 0.0975 to 0.153 (Table 1~; excluding the low 0.0975 for the Chalupa system and the high 0.153 for the Danfaer model narrows the range from 0.12 to 0.135. Digested Organic Matter per Unit of Dry Matter Because BCP is a function of DOM and IF usually is specified as a proportion of dietary DM, some specifica- tion of the proportional DOM per unit of DM (DOMDM) must be made. No system describes this rela- tionship well. For subsequent calculations in the latter sections of this chapter when no value is specified, a value of 0.67 will be used; taken from the analyses of Tyrrell and Moe (197S) on the data of Wagner and Comparison of New Protein Systems for Ruminants I} Loosli (1967~. Diets containing from 25 to 75 percent concentrate and being consumed from 2.82 to 4.05 times maintenance had total digestible nutrients (TDN) from 66 to 68 percent. Increasing the percentage of con- centrate increased intake, but digestibility depression offset the expected increase of digestibility. True Bacterial Protein per Unit of Crude Microbial Protein All systems, except the Landis system, split the BCP into BTP and NCP components with the proportional division represented by BTPBCP and NCPBCP. All sys- tems specify 0.80 for BTPBCP and 0.20 for NCPBCP except the Danfaer model, which specifies 0.85 ant! 0.15, respectively (Table 1) . Digestible Bacterial Protein per Unit of True Bacterial Protein All systems, except the Landis system, specify digest- ~ble bacterial true protein (DBP) per unit of BTP (DBPBTP). These proportions range from 0.70 to 0.90 with 0.80 most frequently used (Table 1~. Digestible Bacterial Protein per Unit of Crude Bacterial Protein All new systems specify a proportional DBP per unit of BCP (DBPBCP) with the range from 0.56 to 0.72 (Ta- ble 1) . All of the new proportions are below the 0.75 that is assumed in the present NRC system. Digestible Nucleic Acid Nitrogen per Unit of Crude Nucleic Add Nitrogen Only three systems specify any rate for NCP. The Danfaer model assumes a proportional 0.85 for digest- ible nucleic acid N as protein equivalent (DNP) per unit of NCP (DNPNCP). The Chalupa system assumes DNPNCP = 1.00. Kaufmann (1977b) describes his sys- tem as having a NCPBCP of 0.10 that has a digestibility of 0.85 and another NCPBCP of 0.10 that has a digest- ibility of 0. For a greater equivalence to the Danfaer model, the Kaufmann system is redescribed in Table 1 as having the full NCPBCP of 0.20 that has a digestibil- ity of 0.425. This description gives the same distribution of nucleic acid N as the original Kaufmann system. Specification of the cligestibility of nucleic acid N is nec- essary to compare the theoretical urinary and fecal ex- cretion of N with ire vivo results. It is also unreasonable to expect that the nucleic acid N will be more digestible than the bacteria that contain it.
12 Ruminant Nitrogen Usage Digestible Undegraded Protein per Unit of Undegraded Intake Protein All systems specify proportional digestible unde- graded intake protein (DUP) per unit of UIP (DUPUIP) with the range from 0.70 to 0.90 (Table 1~. The PDI system assumes variable digestibility as specified in their equation 5 (Verite et al., 1979~. Substitution of their equation 4 into their equation 5 gives a true digestibility of undegradedN = t(0.65 - 0.143) x insolubleNin- take]/0.65 x insoluble N intake = 0.78 for a mean; this mean is also equal to the mean of their variable range from 0.60 to 0.95. The difference between this mean and their variable digestibilities is that their variable di- gestibilities include all of the residual error of a particu- lar feed associated with the derivation of the constants describing fecal N output. Fecal Metabolic Protein Fecal metabolic protein is the most variable factor in the systems (Table 1~. Fecal metabolic protein either is not specified as a separate factor or is specified as a sepa- rate factor that is a function of DM intake (DMI), indi- gestible dry matter intake (IDMI), or indigestible or- ganic matter intake (IOMI). The units are either as absorbed (FPA) or net (FPN) protein. The ARC and Sat- ter systems do not specify any FPA or FPN as a separate factor. The Cornell system mentions it as a separate fac- tor but does not specify an equation. The PDI system specifies FPA as 0.057 per unit of IOMI. The NRC sys- tem specifies FPA as 0.068 per unit of IDMI or as 0.030 per unit of DMI with slightly less accuracy. The Cha- lupa system specifies FPN as 0.03 per unit of DMI. How- ever, it considers two-thirds of this to be undigested bac- terial cells that are accounted for as indigestible bacterial protein and indigestible nucleic acid protein equivalent in this publication. All other systems specify FPA as a function of DMI with factors ranging from 0.012 to 0.026. The Burroughs system specifies the FPA as a reduction of absorbed protein (AP) from feed. All systems specifying FPA may use that FPA as a reduction of AP from feed. In viva data on lactating cows fed forage-concentrate diets may be used for comparing with the fecal meta- bolic protein specifications of these systems. Boekholt (1976) found that digestible protein (DP) as a percent- age of dietary dry matter or DP (percent of DM) = 0.833 x IP(percentofDM) - 3.31,withr2 = 0.95,sy.x = 0.469, and n = 362 for dairy cattle whose mean milk production wasl8.9 + 5.2kg/day~asastandarddevia- tion) and IP (percent of DM) was 16 + 2.3 percent. This equation implies a fecal metabolic protein fraction of 0.033 gig of dietary DM. Waldo and Glenn (1982) per formed a similar analysis on the data of Conrad et al. (1960) and found DP (percent of DM) - 0.861 x IP (percent of DM) - 2.86, with r2 - 0.92, sy.x - 0~743, and n = 177 for dairy cattIe whose milk production was 11.8 + 3.1 kg/day and IP (percent of DM) was 15.3 + 2.8. This equation implies a fecal metabolic protein of 0.029 g/g of dietary OM. That these proportions could round to 0.030 is support for the slightly less accurate estimate of the NRC system, but as FPA rather than FPN. At 15 percent dietary IP, these equations imply that 57 percent of the fecal N arises from this source and that 19 to 22 percent of dietary IP is required to meet this need. It seems realistic to include a factor for fecal meta- bolic protein, since it has so much quantitative impor- tance. FACTORS IN THE REQUIREMENT FOR ABSORBED PROTEIN No factors other than those already used in the NRC system are required in the new protein systems. The to- tal protein requirement includes that for fecal metabolic protein, maintenance, and production. The mainte- nance requirement may include fecal ~netabolic pro- tein, urinary endogenous protein, and surface protein. The production requirement is the sum of one or more of four factors lactation, conceptus, weight change, and growth (including surface material). Maintenance FECAL METABOLIC PROTEIN Fecal metabolic protein is considered differently in the systems (Table 2~. The ARC and Satter systems do not specify any FPA or FPN as a separate factor. The Cornell system cloes not specify how the fecal metabolic protein fraction is considered. The Burroughs system has specified F PA as a reduction from AP available from feed. The Kaufmann and Landis systems specify F PA as a component of total requirement independent of main- tenance; the model of Danfaer also seems to in.clude FPA as a component of total requirements. The inclu- sion of F PA either as a feed reduction factor or fully con- sidered as a component of total requirement supplies the full requirement for fecal N. The NRC system includes a part of the fecal metabolic protein in the maintenance requirement an`1 the remainder in the production re- quirement. The Chalupa system specifies one-third of total fecal metabolic protein in maintenance. The PDI system includes only part of total fecal metabolic pro- tein as a component of the requirement; the remaining requirement for fecal N excretion must be met by reduc- ing assumed urinary N excretion.
URINARY ENDOGENOUS PROTEIN Neither the PDI nor Landis systems specify this factor (Table 2~. All other systems specify an endogenous uri- nary protein equivalent either in units of absorbed (UPA) or net (UPN) protein. This requirement is usually calculated as a power function of body weight near 0.75. SURFACE PROTEIN No specification is made in the Landis system and a zero specification is made in the Burroughs, Satter, Cor- nell, Danfaer, and Kaufmann systems (Table 2~. All other systems specify a surface protein requirement in units of absorbed (SPA) or net (SPN) protein. The SPN is less than 5 percent of the total maintenance for the Cha- lupa and NRC systems. The SPN represents about 20 percent of the maintenance requirement of the ARC sys- tem. Because the PDI system uses the SPA required to give N retention equal to hair and scurf loss as its total maintenance requirement, its SPA is relatively higher than all others. TOTAL The total maintenance requirement as AP is in Table 2 for comparative purposes. The total maintenance re- quirements are extremely variable, ranging from 100 to 395 g of AP. Unfortunately, these are not equivalent be- cause some include fecal metabolic protein and some do not. The relatively low requirement of the Burroughs system can be accounted for partially by the deduction of fecal metabolic protein from available AP. Similarly, the relative low requirements of the Danfaer, Kauf- mann, and Landis systems are accounted] for by their consideration of fecal metabolic protein as a separate component of total requirement. No equivalent factors can account for the lowest requirement in the ARC sys- tem; however, the failure to include a fecal metabolic protein factor probably contributes to its smallness. Production LACTATION The lactation protein requirement as absorbed (LPA) units is the assumed protein concentration or lactation net protein requirement as net (LPN) units divided by the assumed efficiency (LPNLPA) (Table 2~. The most commonly assumed efficiency is 0.70. The efficiency of Burroughs is highest at 0.95 and of Danfaer is lowest at 0.56 (Table 2) . The LPA requirement for 30 kg of milk is in Table 2. These requirements range from 990 to 1,920 Comparison of New Protein Systems for Ruminants 13 g of LEA, with the major cause of differences being dif- ferences in efficiency. The ARC system has the second lowest requirement for milk along with the lowest re- quirement for maintenance and no reduction of avail- able AP by fecal metabolic protein. CONCEPTUS The conceptus protein requirement as absorbed (YPA) units for the last 2 months of pregnancy varies from 107 to 205 g (Table 2~. A frequent requirement is about 160 g. Five systems have not described a require- ment for the conceptus. WEIGHT CHANGES IN LACTATION Six systems rlo not specify a factor for weight change. When weight change is specified as retained protein in net (RPN) units in the NRC, ARC, Oanfaer, and Satter systems, the proportions range from 0.112 to 0.225. The validity of the NRC and ARC systems assuming a differ- ent proportion for gain and loss when the units are de- fined as RPN seems questionable. The efficiency for weight change is the same as for lactation in the ARC and Satter systems. GROWTH Six systems have proposed the proportional gain as protein (Table 2~. Burroughs et al. (1974) assume the proportional retained protein as net (RPN) units in live- weight gain (G) declined from 0.150 at 150 kg of live- weight to 0.110 at 500 kg for finishing steers and heifers of early maturing breeds. Chalupa (1975b) adopted these same data. The ARC (1980) assume proportional RPN in empty body weight gain (EBWG) declined from 0.181 at 50 kg of empty body weight (EBW) to 0.140 at 500 kg for steers of an average size with 0.6 kg EBWG/ day. Proportional RPN is changed by a factor of 0.90 for smaller breeds and 1.10 for larger breeds. Proportional RPN is changed by a second factor of 0.90 for heifers and 1.10 for bulls. Proportional RPN is changed by a third factor of 0.013 subtracted from 1.0 for each 0.1 kg of EBWG greater than 0.6 and 0.013 addec] to 1.0 for each 0.1 kg of EBWG less than 0.6. The PDI (Verite et al., 1979) system assumes proportional RPN in G to decrease from 0.186 to O. 135 with maturity; the proportion var- ies with liveweight, G. breed, and sex. Robelin and Daenicke (1980) extend the PDI system by giving a set of equations for describing the proportional RPN as con- tinuous functions of liveweight and G within very early maturing steers, early maturing bulls, and late matur- ing bulls. Fox et al. (1982) describe the retention of RPN as a function of EBW for steers of medium frame size.
14 Ruminant Nitrogen Usage This steer of medium-frame size is considered a refer- ence animal with an equivalent weight equal to its actual weight. Eight other frame sizes and two other sexes are specified that require adjustment factors for converting their actual weight to equivalent weight; these equivalent weights, theoretically, have the same body composition. Adjustment factors for steers are 1.25 for smallest frame, 1.00 for medium frame, and 0.83 for largest frame. Adjustment factors for heifers are 1.56 for smaldest frame, 1.25 for medium frame, and 1.04 for largest frame. Adjustment factors for bulls are 1.04 for smallest frame, 0.83 for medium frame, and 0.69 for largest frame. The NRC (1978) requirement for dairy cattle is in Table 2. The NRC (1984) requirement for beef cattle calculates RPN as a function of the energy concentration of gain for steers. Composition of gain of medium-frame heifers is assumed equivalent to medium-frame steers weighing 15 percent more. Com- position of gain of bulls and large-frame steers was as- sumed equivalent to medium-frame steers weighing 15 percent less. Liveweight, daily gain, breed or frame size, and sex are the four most important factors affect- ing the protein energy ratio in growing and fattening cattle. The functional change in protein and fat deposi- tions with increasing energy deposition remains some- what controversial. The proposals range from linear changes in energy deposited as fat and protein (Tyrrell et al., 1974; Geay, 1984) to an asymptotic maximal de- position of protein (Byers, 1982b) and to a maximal pro- tein (reposition followed by a decrease (Anrique, 19764. The efficiencies of converting retained protein as ab- sorbed (RPA) units to RPN, or RPNRPA, range from 0.45 (NRC, 1978) to 0.75 (ARC, 1980) in Table 2. The NRC (1984) requirements for beef cattle assume an effi- ciency of 0.66. Data on sheep are not specific in the various models. As a consequence they are not covered in this discussion. Comments on gain and concepts applying to it would be appropriate for sheep in lieu of more definite data. DYNAMIC MODELS Dynamic models have been proposed that describe protein utilization for the entire animal. Other dynamic models describe ruminant digestion of dietary crude protein and carbohydrates, while others describe nitro- gen metabolism in the ruminant without any reference to energy. Some of these models are considered prelimi- nary. Generally, the models are not published in full detail so that direct communication with the authors is required for enough detail to use, compare, or challenge them. Two dynamic models for specifying protein require- ments for ruminants are being developed in the United States. At Michigan, Fox et al. (1976) introduced a net protein system. Bergen et al. (1979) calculated the up- per limit of ruminal microbial protein synthesis. Bergen et al. (1982) describe the efficiency of microbial protein synthesis in relation to specific growth rate, growth yield, and maintenance in rumen bacteria. They dem- onstrate an increase in ribonucleic acid/protein ratio as specific growth rate increases in anaerobic bacteria. Such a large difference in this ratio must raise questions about the constancy of the ratio of nonammonia nitro- gen and amino nitrogen entering the small intestine or apparently absorbed in the small intestine. Johnson and Bergen (1982) describe the effects of diet on the fraction of organic matter digestion occurring in the rumen and efficiency of microbial protein production. Wailer et al. (1982) describe their progress toward a dynamic model of protein requirement for the ruminant that considers economics as well as nutrition and emphasizes the alge- bra and linear programming necessary to consider the uncertainty of feed composition and least cost formula- tion. At Cornell, Van Soest et al. (1982) propose a rumen submodel for nitrogen utilization that describes the out- put of protein by using the following inputs: soluble pro- tein; three true protein subfractions based on the degra- dability rates (B~, rapid; Be, intermediate; and Be, slow); bound protein; nonstructural carbohydrate, po- tentially digestible organic matter; rates of digestion for each protein, nonstructural carbohydrate and poten- tially digestible organic matter subfractions; and rates of passage for liquids and solids. Fox et al. (1982) com- plete the total model by describing the factors in the cal- culation of requirements for growth (Table 2~. The nitrogen flux within the rumen as REP loss of ammonia by absorption and passage and as RIP gain of urea from saliva or blood are very important. Nolan et al. (1976) describe the nitrogen dynamics on a three pool model of rumen ammonia, plasma urea, and cecal am- monia in a sheep eating about 22 g air dry feed/in that contained 18.7 percent CP. When mean dietary N in- take was 16.3 g/d, mean rumen ammonia N was 20.9 mg/100 ml, mean plasma urea N was 18.1 mg/100 ml, and total flux of rumen ammonia was 15.0 go/. Of this total flux 28.7 percent was recycled, and 71.3 percent was irreversible loss via influx and efflux. The influx sources, as a percentage of total, were: dietary ant] en- dogenous sources, 61.9; blood urea, 6.9; and from cecal ammonia but not via blood urea, 2.5. The efflux losses, as a percentage of total, were: absorbed, 44.4, micro- bial protein synthesized into tissue, 20.6; and cecal am- monia from rumen microbes, 6.3. Only 40 percent of rumen bacterial N came from ammonia N. and only 20 percent of urea degraded in the intestinal tract was de- gracled in the rumen. Mazanov and Nolan (1976) de
Comparison of New Protein Systems for Ruminants 15 scribe the nitrogen dynamics in a nine-part model using the above data combined with data from lower N in- takes. When mean dietary N intake was 14. I6 g/d, total flux of ammonia N was 9.11 g/d. Of this total flux, 19.2 percent was recycled. The influx sources, as a percent- age of total, were: dietary amino N. 67.0; urea, 13.2; and dietary ammonia N. 0. 6. The efflux losses, as a per- centage of total, were: microbial N not recycled, 31.3; and ammonia N absorbed or passed, 49.5. Baldwin et al. (1977a, proposed a model of ruminant digestion that uses 12 chemical inputs: lignin, cellulose, hemicellulose, pectin, starch, soluble carbohydrate, or- ganic acids, lipids, ash, insoluble protein, soluble pro- tein, and NPN. The model uses one physical input: frac- tion retained on 1-mm sieve. This model emphasizes the importance of ammonia passage as a loss of N from the rumen. Baldwin and Denham (1979) present another model of N metabolism in the rumen that emphasizes the difference in affinity of the two major enzymes for ammonia. Glutamic dehydrogenase is constitutive and has a low affinity for ammonia (Km = 5 mM); glu- tamine synthetase is induced at low ammonia concen- trations and has a high affinity for ammonia (Km = 0.2 mM). Such a difference may explain why microbial growth is not limited until concentrations fall below 3 to 5 mg/100 ml, but microbial fermentation of the carbo- hydrates in some diets is limited at concentrations below 20 to 25 ma/ 100 ml, a critical point in comparing in vitro and in vivo results. The dietary differences in ruminal methylamine concentration (Hill and Mangan, 1964) may affect the competitive uptake of ruminal ammonia by bacteria. Black et al. (1980-1981) describe a model of rumen function that uses these chemical inputs: beta-hexose (lignin, cellulose, and hemicellulose); alpha-hexose (pectin and starch), soluble carbohydrate (including glycerol), total fatty acids, inorganic sulfur; ash; protein (true protein and free amino acids); NPN (including nu- cleic acids); potential degradability of beta-hexose; and potential degradability of protein. The mode] has one physical input: modulus of fineness of diet. Other mod- eling inputs are: feed intake; time feeding; time rumi- nating; and reduction in maximum rate or degradation of beta-hexose, alpha-hexose, and protein due to diet. Endogenous inputs are: true protein, NPN, and inor- ganic sulfur. Beever et al. (1980, 1981) did a sensitivity analysis for 22 variables that could not be set with confi- dence. Six variables with a high sensitivity, i.e., a change in protein flow greater than 40 percent from the possible range in input variable, were: potential degra- dability of protein, fractional outflow rate of water, fractional outflow rate of microbes, energy required for microbial maintenance, salivary flow, and proportion of rumen ammonia available for microbial growth. Faichney et al. (1980) found predictions of this model to be closer to observations in one data set than predictions from the ARC and PDI systems. COMPARISON AND CHALLENGE OF SYSTEMS WITH IN VI VO DATA A comparison and challenge of implications of the systems with in vivo data from lactating cattle is inform- ative after their assumptions and calculations are under- stood. For these comparisons and challenges we have calculated IF as a percentage of DM, optimum UIPIP, either the sum of BTP and UIP or the sum of BCP and UIP reaching the small intestine, fecal N as a percentage of dietary N. urinary N as a percentage of dietary N. and milk N as a percentage of dietary N for a 600-kg cow producing 10, 20, 30, and 40 kg milk/day with degrada- bility optimal. These predicted data then are compared with expected in viva data on protein reaching the small intestine (Tamminga and van Hellemond, 1977; Journet and Verite, 1979; Bohr et al., 1979) and in vivo data on the distribution of N in feces, urine, and milk (Conrad et al., 1960; Boekholt, 1976~. Some additional data and assumptions are required because BCP pro- duction is a function of FOM and fecal metabolic pro- tein is a function of DMI, IDMI, or IOMI. No attempt was made to compare and challenge the Cornell system because it contains several dynamic relationships. Energy Standards Except for the ARC and PDI systems, the new protein systems are published without any specific statement of or reference to an energy standard. Production of BCP is related to energy fermented in the rumen, which is more frequently FOM. Fecal metabolic protein is related to some dietary component, most frequently dietary DM. These or other required energy variables must be speci- fied for a complete system. The energy requirements and the DM intakes used in these comparisons and chal- lenges of protein feeding systems and their sources are in Table 3. Chalupa (1980a) used metabolizable energy as the energy unit, but energy requirements were not fully elaborated, so TDN is used as the energy requirement for the Burroughs, Chalupa, NRC, and Satter systems. Assumptions about concentrations of energy in dietary DM vary as well as the assumptions about absolute amounts of either. In going from 5 to 40 kg of milk, en- ergy concentrations increase 85 percent in the PDI sys- tem, 33 percent in the Kaufmann system, and 65 per- cent in the Landis system. When energy concentration is not specified, we have assumed it to be constant (based on the proposed maximum of Tyrrell and Moe, 1975) with DOMDM = 0.67 as discussed earlier for TDN.
16 Ruminant Nitrogen Usage TABLE 3 Dry Matter Intakes and Energy Standards When Energy Concentration Varied as Used in Comparison and Challenge of Protein Systems PDI ~KaufmannC Landisf NRC° ARCb DanfaerC Dry Dry Dry Dry Dry Dry Milk Matter Matter Matter Matter Matter NEL Matter (kg) (kg) (kg) (kg) UFL (kg) SE (kg) (MJ) (kg) 5 8.6 7.2 7.5 7.1 12.3 4375 9 51.2 11.5 10 10.9 9.4 9.8 9.3 13.7 5750 11 66.9 13.0 15 13.2 11.7 12.1 11.5 15.1 7125 13 82.6 14.5 20 15.4 14.1 14.4 13.7 16.5 8500 15 98.3 16.0 25 17.7 16.4 16.7 16.1 17.9 9875 17 114.0 17.5 30 20.0 18.8 19.0 18.6 19.2 11250 19 129.7 19.0 35 22.2 21.3 21.3 21.0 20.6 12625 20 145.4 20.5 40 24.5 23.6 23.6 23.5 21.9 14000 21 161.1 22.0 a Calculated from total digestible nutrients for maintenance of the mature, lactating, 600-kg cow and production of milk with 3.5 percent fat (NRC, 1978) by dividing by .67. Used for the NRC, Burroughs, Chalupa, and Satter systems. Calculated from megajoules of metabolizable energy for maintenance of a 600-kg cow with 0 live- weight change and production of milk with 3.68 percent fat while being fed a diet with metabolizab~lity or q = .60 (ARC, 1980) by dividing by 11. Calculated as Scandinavian feed energy (FE) units from Madsen et al. (1977) (Table 3) starting from N in microbial net protein divided by 16 g N per FE divided by digestibility or .60 divided by efficiency or .71; then 16.S FE = 19 kg dry matter from Danfaer et al. (1980, p. 12~. From Verite et al. (1978, Table 12.3~. UFL = the French net energy unit and is the total requirement for a 600-kg cow consuming good quality forage and producing milk with 4.0 percent fat. 'From Kaufmann (1977b, 1979). SE = starch equivalent unit. Data arelinearly interpolated and extrapolated from the data in Table 2 (Kaufmann, 1979~. fFrom Landis (1979~. NEL = net energy for lactation. Data are linearly interpolated and extrapolated from data in Table 2. Such different energy assumptions contribute to differ- ences among the protein systems. While the use of TDN is questioned by many, the available data for alternatives are not as numerous. Of even more importance is the fact that in use many of the alternative energy terms are derived from TDN or an estimate of TDN. Thus, we do not feel that TDN is, in fact, an improper base. Additional Assumptions Several additional assumptions that were required in one or more of the systems and their bases are in Table 4. Where the disposition of NCP is not described, its digest- ibility was assumed to be 0.85, and the excretion of di- gested fraction was assumed to be via urine. Minimum Dietary Intake Protein Percentage Dietary IP percentages required in nine systems, when undegradability is optimal, are compared (Figure 3) . Differences among the systems are smaller (from 9 to 13.2 percent IP) at 10 kg of milk but become larger (11 to 17.4 percent IP) at 40 kg of milk. The Danfaer model requires the highest IP percentage at every milk yield. Presumably, this higher requirement is primarily a result of a lower assumed efficiency of milk production. The Burroughs, ARC, and Satter systems require a much lower IP percentage than other systems at higher milk production. The probable causes of their low re- quirements are the highest efficiency for converting AP to milk assumed in the Burroughs system; the second lowest AP requirement for lactation plus a low AP re- quirement for maintenance with no separate fecal metabolic protein requirement in the ARC system; and no fecal metabolic protein requirement as a function of DM intake either alone or as a component of mainte- nance in the Satter system. Optimum Undegradability Optimum IP undegradabilities required when IP per- centage is minimum are compared (Figure 4~. Differ- ences among the systems arelarge (7 to 41 percent at 10 kg of milk) and remain large (20 to 55 percent at 40 kg of milk). The undegradability of many common diets for dairy cows is considered to be near 0.30 (Satter and Rof- fler, 1975~. The ARC and Burroughs systems do not re- quire an undegradability as high as 0.30 at 40 kg of milk per day. These low undegradability requirements are
Comparison of New Protein Systems for Ruminants 17 TABLE 4 Additional Assumptions of Protein i and Energy Relationships Assumption Systemsa Proteins' CBPDIP= 1.00 N DNPCNP = 85c A, P. B. S _ LNPLMP= .60 L E 14 Energyd at DM = TDN/.67e N. C, S LIZ, 11 MJ ME/kg DMf A O 19 kg DM = 16.5 FE. D AM = DM x .9 p DE - ME/.82'' A DONI = DM x .67 D '0 DOM = UFL x .732i P DOM = NEL/9.31i L 1 kgDOM = 900 SEE K 19 MJ ME/kg DOM/ A IOM = 0M - DOM P aA, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kaufmann; L, Landis; N. NRC; P. PDI; and S. Satter. b CBPDIP, crude bacterial protein/degraded intake protein; DNPCNP, digestible nucleic acid bacterial protein/crude nucleic acid bacterial protein; LNPLMP, lactation net protein/lactation metabo- lizable protein. 'From Danfaer (1979). ~DM, dry matter; TDN, total digestible nutrients; ME, metaboliz- able energy; FE, Scandinavian feed energy unit; OM, organic matter; DE, digestible energy; DOM, apparently digested organic matter; UFL, French net energy unit; NEL, Swiss net energy unit; SE, starch equivalent; IOM, indigestible organic matter. From analyses of Tyrrell and Moe (1975) on data of Wagner and Loosli (1967). fFrom ARC (1980, see p. 112). "From Danfaer et al. (1980, see p. 12). From ARC (1980, see p. 136). iFrom INRA (1978, see p. 589); and ARC (1980, see Table 4.7). ME = 2.73 UFL and DOM = ME/3.72 so DOM = UFL x (2.73/3.72) = UFL x .732. Calculated from Landis (1979~. .135 CBPDOM/.0145 CBPNEL = 9.31. From Kaufmann (1977b). From ARC (1980, see p. 136~. related to the low protein percentages, and their causes as discussed earlier. The Chalupa system always requires an undegradability greater than 0.30. This high unclegradability results primarily from the low BCPDOM. A plot of the UIPIP as a function of concentration of IF (Figure 5) indicates a great diversity among the sys- tems. The differences are largely caused by assumptions about changes of energy concentration and dry matter intake for meeting the additional energy needs for high milk procluction. If increasing energy requirements are met by increasing energy concentration more than DM intake, as in the Kaufmann, Landis, and PDI systems, then protein concentration varies more than UIPIP. If ~6 o / No/ / A 1 0 20 30 M I LK (kg/day) FIGURE 3 Intake protein percentage in dry matter as a function of milk production. A, ARC; B. Burroughs; C, Cha- lupa; D, Danfaer; K, Kaufmann; L, Landis; N. NRC; P. PDI; and S. Satter. increasing energy requirements are met by increasing DM intake more than energy concentration, as in the ARC, Burroughs, Chalupa, Danfaer, and Satter sys- tems, then UIPIP varies more than protein concentra- tion. Protein Reaching the Small Intestine IN VIVO REFERENCE DATA Three data sets (Tamminga and van Hellemond, 1977, Journet and Ferrite, 1979; Bohr et al., 1979) are available that describe protein flowing into the duode- num of the lactating cow. Tamminga and van Helle- mond (1977) observed amino acid N (g/day) = 32.3 DOM (kg/day) - 8.63, with r2 = 0.90. Their organic matter intakes ranged from 4.7 to 14.6 kg/day, N intake ranged from 140 to 430 g/day, and digestible IF ranged from 11.2 to 23.1 percent of DOM in 49 observations. Rohr et al. (1979) observed amino acid N (g/day) = 31.42 DOM (kg/day) - 40.56, with r2 = 0.8S. Their organic matter intakes ranged from 8.88 to 15.14 kg/ day, N intake ranged from 205 to 413 g/day, and crude protein ranged from 12.9 to 15.6 percent of dietary DM in 21 observations. These two equations indicate that DOM is a primary determinant of protein entering the small intestine. Journet and Verite (1979) observed non- ammonia N (g/day) = 23.85 DOM (kg/day) + 0.60 in vitro nondegradable N (g/day) + 8. 6, with R2 = 0.886
IS Ruminant Nitrogen Usage 60: 50 Be 40 - a, 6 30 cr LL of 10: o 1 B 0 10 20 MILK (kg/day) 30 40 FIGURE 4 Undegradability of dietary intake protein as a function of milk production. A, ARC; B. Burroughs; C, Cha- lupa; D, Danfaer; K, Kaufmann; L, Landis; P. PDI; and S. Satter. in equation 2 with lactating cows. Their DOM intakes ranged from 4.3 to 12.2 kg/clay, and nondegradable N intakes ranged from 40 to 266 g/day in 42 observations. The equation of Tamminga and van Hellemond always gives a greater expectation than the equation of Bohr et al. (1979) because Tamminga and van Hellemond sam- pled posterior but Rohr et al. (1979) sampled anterior to the pancreatic and bile ducts. SYSTEM COMPARISONS First, one type of predicted flow into the small intes- tine of true protein (STP) was calculated as the sum of BTP plus UIP without endogenous protein for each sys- tem. Another type of predicted flow into the small intes- tine of crude protein (SCP) was calculated as the sum of BCP plus UIP without endogenous protein for each sys- tem. Second, three expecter] protein flows into the small intestine were calculated for each system based on DOM intake and UIP intake, if required, in the three equa- tions just discussed. These two types of estimates of pro- tein flow will be called predicted for the two former and expected for the three latter. Comparisons of the predicted protein flow, as STP, and expected protein flow, as amino nitrogen, are in Figures 6 and 7, compar- ison of predicted protein flow, as SCP, and expected protein flow, as nonammonia N. are In Figure 8. Pre- dicted flows into the small intestine from the ARC, Bur roughs, and Satter systems were less than expected flows in all three comparisons. This difference probably results from their low AP requirement and their low IF concentration in the DM. The predicted flow in the Landis system was always less than the expected flow, and an explanation for this is not clear. The predicted flow from the NRC was highest and generally greater than expected in the two comparisons with expected flow based on DOM. This high predicted flow for the NRC system probably results from no subtraction of NCP. The predicted flow in the Danfaer model was next highest. This probably resulted from having the highest AP requirement and the highest IF concentration in the Dot. The predicted flows of the Kaufmann and PDI sys- tems were similar to the expected flows. The predicted flow for the Chalupa system was similar to the expected flow based on DOM but decreased relative to expected flow when undegradable protein became a partial basis of expectation. This difference probably resulted from the high undegradability and the low microbial protein production per unit of DOM. Fecal Crude Protein Equivalent Excretion Relative to Crude Protein Percentage Fecal crude protein (F P) equivalent was calculated as the sum of indigestible bacterial protein (IBP), indigest- ible nucleic acid crude protein (INP) equivalent, indi 60 50 _ an 40 6 30 tr: C' LU He TIC 20 _ 10 _ K ' Lo So AD B O _I I 8 10 PROTEIN (%dm) D / 1 1 1 1 12 14 16 18 FIGURE 5 Undegradability of dietary intake protein as a function of intake protein percentage in dry matter. A, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kaufmann; L, Landis; P. PDI; and S. Satter.
Comparison of New Protein Systems for Ruminants 19 gestible undegraded dietary protein (IUP), and fecal metabolic protein as absorbed (FPA) or net (FPN) units. The FP excretion, as a percentage of dietary IP, is ex- pressed as a function of IP as a percentage of dietary DM (Figure 9~. Reference curves are plotted from the analy- sis (Waldo and Glenn, 1982) of the data of Conrad et al. (1960) and Boekholt (1976~. The ARC, Burroughs, Cha- lupa, and Satter systems and the Danfaer model predict fecal excretions lower than expected from the data of Conrad et al. (1960) or Boekholt (1976~. The probable causes of these low excretions are the use of zero fecal metabolic protein in the ARC and Satter systems as a function of dietary DM and a relatively low fecal meta- bolic protein in the Chalupa, Danfaer, and Burroughs systems. The use of zero fecal metabolic protein pro- duces a relatively constant percentage output of N in- take in the feces, and use of a low fecal metabolic protein produces a curve with less slope than expected. The PDI system predicts a fecal output in the general range of that expected from the data of Conrad et al. (1960) and Boekholt (1976), but it declines more rapidly as concen- tration increases; the more rapid decline occurs because fecal metabolic protein actually decreases due to lower indigestible OM as milk production and protein concen- tration increase. The NRC system predicts a fecal excre- tion essentially equal to that expected from Boekholt 4 _ 3 - ~2 ~ c, a: - 1 A N EXPECTED (kg/day) FIGURE 6 Protein flow into small intestine predicted from the system versus that expected based on the digestible organic matter of the system and the equation of Tamminga and van Hellemond (1977~. A, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kaufmann; L, Landis; N. NRC; P. PDI; and S. Satter. N 3 _ - - y - ~2 LU 1 - ~: 0 1 2 3 4 EXPECTED (kg/day) FIGURE 7 Protein flow into small intestine predicted from the system versus that expected based on the digestible organic matter of the system and the equation of Bohr et al. (1979~. A, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kaufmann; L, Landis; N. NRC; P. PDI; and S. Satter. (1976) and slightly higher than expected from the data of Conrad et al. (1960~. The Kaufmann and Landis sys- tems predict fecal excretions most similar to those which occur because their assumptions for fecal metabolic pro- tein and digestibility are similar to those in the data of Conrad et al. (1960) and Boekholt (1976~. Fecal Crude Protein Equivalent Excretion Relative to Milk Production The FP excretion, as a percentage of dietary IP, is ex- pressed as a function of milk production in Figure 10. Two reference points for these data are 37.5 percent from Boekholt (1976) and 33 percent from the analysis (Waldo and Glenn, 1982) of the data of Conrad et al. (1960~. Basically, the same comments apply to Figure 10 as were made for Figure 9. The Satter system and the ARC system, to a lesser degree, predict low outputs that are nearly constant because they assume zero fecal metabolic protein per unit of feed DM. The Chalupa, Danfaer, and Burroughs systems predict low outputs that decrease gradually with increasing milk production because they assume a minimal fecal metabolic protein. The PDI system predicts an output in the expected range, but its predicted output declines rapidly because this fecal metabolic protein output actually declines with increasing milk production. The NRC system pre
20 Ruminant Nitrogen Usage / //// o, ~ 1 _ 1 0 1 2 3 4 EXPECTED (kg/day) FIGURE 8 Protein flow into small intestine predicted from the system versus that expected based on the digestible organic matter plus undegraded protein intake of the system and the equation of Journet and Verite (1979) . A, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kaufmann; L, Landis; P. Pl)I; and S. Satter. diets more fecal output than expecter] because it assumes lower digestibility. The Kaufmann and Landis systems predict fecal output that follow the expected curve (Fig- ure 9) but are slightly higher than expected. Urinary Crude Protein Equivalent Excretion Relative to Milk Production Urinary crude protein (UP) equivalent was calculated as the algebraic sum of rumen efflux of crude protein (REP) equivalent or a rumen influx of crude protein (RIP) equivalent; digestible nucleic acid crude protein (DNP); maintenance protein as absorbed (MPA3 units that is free of any fecal metabolic N. if possible; and the protein difference of LPA minus LPN. If necessary, fe- cal metabolic N was subtracted to balance the system. Possibly, the tissue utilization of nucleic acids should be considered based on the finding of a 47 percent retention of activity in the tissues of the ruminating lamb by Raz- zaqueetal. (1981~. The UP excretion as a percentage of dietary IP is ex- pressed as a function of milk production (Figure 11~. Two reference points for these data are 35.7 percent from Boekholt (1976) and 38.6 percent from the data of Conrad et al. (1960~. The Satter system predicts a uri- nary excretion greater than expected primarily because it assumes a low efficiency of milk production. The Dan faer system predicts a urinary excretion greater than ex- pected because it assumes a low efficiency for milk pro- duction and assumes an efflux of N as ammonia from the rumen to the blood. The ARC and Burroughs systems did not predict high urinary excretions as might be ex- pected in order to balance low fecal excretions. Their predicted urinary excretions were similar to those of Boekholt and Conrad; their low dietary IP were ac- counted for by their high efficiencies of producing milk from AP. The PDI system is the only one that predicted an increasing UP excretion as milk production in- creased, and these excretions were generally lower than those of Boekholt and Conrad. This increasing urinary N excretion seems to result from the decreasing fecal metabolic N excretion at high milk production. The Chalupa system predicts UP excretions that are consis- tent with those of Boekholt and Conrad. The NRC sys- tem predicts low UP excretions that result from an as- sumption of zero DNP. The Landis system predicts a relatively low UP excretion because no DNP fraction is included. The Kaufmann system predicts a low UP ex- cretion because the digestibility of nucleic protein equivalent is only one-half of the more common assump- tion. Milk Nitrogen Output Relative to Milk Production Output of milk protein in net (LPN) units as a per- centage of dietary IP was expressed as a function of milk 60 50: ~ 30 o A: UJ 20 10 X it_ PROTEIN (% DM) FIGURE 9 Fecal protein as a function of intake protein per- centage in dry matter. A, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kaufmann; L, Landis; N. NRC; P. POI; S. Sat- ter; X, Boekholt (1976); and Y. Conrad et al. (1960).
Comparison of New Protein Systems for Ruminants 21 production (Figure 12~. Two reference points for these data are 24.9 percent from Boekholt (1976) and 21.7 percent from the data of Conrad et al. (1960~. In all systems milk protein is assumed equal to requirement as LPN units. The high fractional output for the ARC and Burroughs systems is primarily a function of the low protein intake. All of these systems predict a higher out- put of dietary IP in milk than either the mean of 24.9 percent from Boekholt (1976) when mean milk produc- tion was 18.9 kg/day or the mean of 21.7 percent from the data of Conrad et al. (1960) when the mean milk production was 11.8 kg/day. Increasing milk produc- tion to the average (25 kg/day) assumed here and opti- mizing degradability both will increase the fractional output of dietary IP into milk. It seems overly optimistic to assume that outputs greater than 40 percent can be obtained easily. CRITICAL COMMENTS ON OMISSIONS OF SOME SYSTEMS Comparison and analysis of these systems as described earlier emphasize three important points that fre- quently are overlooked but should receive more empha- sis. First, a fecal metabolic protein fraction is needed for FP excretion to correspond to in vivo data. Second, this fecal metabolic protein fraction should be considered ei- ther a separate component of total requirement or a feed reduction component and not be included in mainte 60- 50t 40 ' L K ~_ 10 _ O O JO 1 1 1 20 30 40 M l ~ K (kg/day) FIGURE 10 Fecal protein as a function of milk production. A, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kauf- mann; L, Landis; N. NRC; P. PDI; S. Satter; X, Boekholt (1976~; and Y. Conrad et al. (1960~. so_ so - - ~ 40 _ ._ - z ~ 30 _ o CL ~ 20 _ CC 10 _ O D. 0 10 20 30 40 M I LK (kg/day) FIGURE 11 Urinary protein as a function of milk produc- tion. A, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kaufmann; L, Landis; N. NRC; P. PDI; S. Satter; X, Boekholt (1976); and Y. Conrad et al. (1960). nance per se for simplicity as has been done in some cases (or ignored in others). Third, specification of the DM intake and DOM, or other energy components, are an integral part of any complete protein system. Fecal metabolic protein is an important component of the protein requirement of the ruminant. Fecal meta- bolic protein represents about 57 percent of the total EP and about 20 percent of the IP requirement at 15 per- cent dietary IP for the negative intercept from either the equation of Boekholt (1976) or the equation (Waldo and Glenn, 1982) based on the data of Conrad et al. (1960) as its estimate. The failure to include a fecal metabolic protein factor in a protein feeding system will result in underestimation of requirements for IP percentage in dietary DM and for undegraclability of dietary protein. The fraction of dietary nitrogen excreted in the feces will be underestimated, and the fractional FP excretion as a function of IP concentration will not have the char- acteristic in viva hyperbolic curvature. Fecal metabolic protein is most commonly related to DMI except for the PDI systems where it is related to IOMI and the NRC system where it is related to IDMI. The PDI equation (Verite et al., 1979) is based on sheep and has an R2 = 0.74; the equations of Boekholt (1976) and the data of Conrad et al. (1960) are based on lactat- ing cattle and have r2 = 0.95 and 0.92, respectively. Fecal metabolic protein is more highly correlated with DM than IOMI. Three g of FMP/100 g of DMI is ~ good simple interim proportion. Fecal metabolic protein is a
22 Ruminant Nitrogen Usage 60 50 ~ 40 He o CL ye 30 20 10 o ON Lip X P y 1 1 0 10 20 30 40 MILK (kg/day) FIGURE 12 Milk protein as a function of milk production. A, ARC; B. Burroughs; C, Chalupa; D, Danfaer; K, Kauf- mann; L, Landis; N. NRC; P. PDI; S. Satter; X, Boekholt (1976~; and Y. Conrad et al. (1960~. function of DM intake, maintenance protein is a func- tion of body weight, and production protein is a func- tion of milk output. Fecal metabolic protein is consid- ered most simply either as a separate component of the total requirement along with maintenance and produc- tion as used by Danfaer (1979), Kaufmann (1979), and Landis (1979) or as a feed reduction component as used by Burroughs et al. (1975b). When the units are consid- ered as SPA, there is little conceptual difference be- tween these methods of accounting. Operationally, this is much simpler than either having the fecal metabolic protein for a part of dietary DM included in mainte- nance and the remainder accounted for in the produc- tion requirement as in the NRC (1978) system or having one-third of the total fecal metabolic protein require- ment per unit of DM in maintenance (Chalupa, 1980a). No AP system is complete until all of the integral en- ergy components required in the system are described. The BCP requirement per se and its contribution to the animal's need for AP are a function of the energy fer- mentecl in the rumen; generally, this component is ap- parently FOM. The fecal metabolic protein require- ment is related to another feed component; generally, this component is dietary DM. The relationship be- tween these two components or the digestible energy concentration in the diet thus is needed. The assumption made here is that digestible organic matter, TDN, or energy digestibility must be asymptotic at about 67 per- cent based on the analyses of Tyrrell and Moe (1975) of the data of Wagner and Loosli (1967~. This assumption of a constant energy concentration is different from the PDI assumption where energy concentration is 85 per- cent greater for high milk production than for low milk production. The relative changes of concentration ver- sus undegradability of protein are affected largely by the relative changes of digestibility and intake of en- ergy, respectively, in requirements for higher produc- tion.
