development of predictive models of nutritional condition ... · rachel c. cook,1,2 department of...

Development of Predictive Models of Nutritional Condition for Rocky Mountain ElkAuthor(s): Rachel C. Cook, John G. Cook, Dennis L. Murray, Peter Zager, Bruce K. Johnsonand Michael W. GratsonReviewed work(s):Source: The Journal of Wildlife Management, Vol. 65, No. 4 (Oct., 2001), pp. 973-987Published by: Allen PressStable URL: http://www.jstor.org/stable/3803046 .Accessed: 20/07/2012 14:55

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Allen Press is collaborating with JSTOR to digitize, preserve and extend access to The Journal of WildlifeManagement.

http://www.jstor.org

http://www.jstor.org/action/showPublisher?publisherCode=acg

http://www.jstor.org/stable/3803046?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

DEVELOPMENT OF PREDICTIVE MODELS OF NUTRITIONAL CONDITION FOR ROCKY MOUNTAIN ELK RACHEL C. COOK,1,2 Department of Fish and Wildlife Resources, University of Idaho, Moscow, ID 83844, USA JOHN G. COOK, National Council for Air and Stream Improvement, 1401 Gekeler Lane, La Grande, OR 97850, USA DENNIS L. MURRAY, Department of Fish and Wildlife Resources, University of Idaho, Moscow, ID 83844, USA PETER ZAGER, Idaho Department of Fish and Game, 1540 Warner Avenue, Lewiston, ID 83501, USA BRUCE K. JOHNSON, Oregon Department of Fish and Wildlife, 1401 Gekeler Lane, La Grande, OR 97850, USA MICHAEL W. GRATSON,3 Idaho Department of Fish and Game, 1540 Warner Avenue, Lewiston, ID 83501, USA

Abstract: Despite its preeminence as a game species in North America, little research exists to validate nutritional condition indices for Rocky Mountain elk (Cervus elaphus nelsonii). We developed and calibrated indices of nutritional condition for live and dead Rocky Mountain elk. Live-animal indices included 20 serum and 7 urine chem-

istry variables, a body-condition score (BCS), thickness of subcutaneous fat and selected muscles using ultra-

sonography, bioelectrical impedance analysis (BIA), and body mass. Dead-animal indices included femur and mandible marrow fat, 3 kidney fat indices, and 2 carcass-scoring methods. Forty-three captive-raised cows (1.5 to 7 years old) were randomly divided into 3 seasonal groups (Sep, Dec, and Mar). Within seasonal groups, elk were fed different diets to induce a wide range of condition; all were fed identical diets 7 days prior to sampling to eliminate short-term nutritional effects. Cows were euthanized and homogenized for chemical analysis of fat, protein, water, and ash content. Estimates of fat and gross energy (GE) were compared to each condition indicator using regression, with age and season as covariates. Relations between condition and thyroxine (T4) and insulin-like

growth factor (IGF-1) varied seasonally, and the relation between condition and mandible marrow fat varied

among ages. Subcutaneous fat depth and BCS were most related to condition for live animals (r2 > 0.87, P< 0.001); carcass scores and kidney fat were most related to fat and GE for dead animals (r2 > 0.77, P< 0.001); and IGF-1 and

T4 were the only serum and urine indices at least moderately related to condition (r2 2 0.54, P< 0.001). Nearly all other serum and urine indices, bone marrow indices, and BIA were either poorly correlated with condition or exhibited highly nonlinear relations. These results identify several indices of condition useful for assessing nutritional condition of live or dead elk, and indicate a number of previously used techniques that correlate poorly with total body fat.

JOURNAL OF WILDLIFE MANAGEMENT 65(4):973-987

Key words: body composition, body-condition score, Cervus elaphus, condition, elk, fat index, kidney fat, marrow fat, nutrition, serum chemistry, ultrasound, urine chemistry.

Research on ruminant stock in captive settings has clearly established the role of nutrition on vir-

tually all aspects of individual and herd productivity, but assessment of nutritional effects on

population dynamics of free-ranging ungulates is

rare-particularly for elk outside national parks. Understanding influences of nutrition on herd

demographics is limited by a lack of practical, reliable, and cost-effective techniques for moni-

toring elk condition and nutrition (Cook 2002). Assessing nutritional quality of forage is difficult and expensive, whereas assessing nutritional condition of animals is impractical in the field, inac- curate, or inadequately tested (Robbins 1983, Harder and Kirkpatrick 1994, Saltz et al. 1995).

Moreover, past studies often failed to clarify between indices appropriate for nutritional assessment versus those appropriate for evaluat-

ing nutritional condition. Nutrition is defined as the rate of ingestion of assimilable energy and nutrients, and nutritional condition is the state of

body components (e.g., fat, protein), which, in turn, influences an animal's future fitness (Hard- er and Kirkpatrick 1994).

We evaluated most live- and dead-animal condition indices currently available for large ungulates. We used captive-raised cow elk fed a variety of diets to induce a wide range of body condition to develop the models. For live animals, we assessed serum and urine chemistry, a body-condition scoring system, thickness of subcutaneous

rump fat and select muscles, and bioelectrical

impedance analysis. For dead animals, we assessed femur and mandible fat, 2 carcass-scoring techniques, and several variations of the kidney fat index. We assessed relations between indices

1 Present address: National Council for Air and Stream Improvement, Forestry and Range Sciences Lab, 1401 Gekeler Lane, La Grande, OR 97850, USA.

2 E-mail: [email protected] 3 Deceased.

973

974 DEVELOPMENT OF ELK CONDITION MODELS * Cook et al. J. Wildl. Manage. 65(4):2001

and nutritional condition and developed models using these indices to predict nutritional condition. In a companion paper, Cook et al. (2001) assessed sensitivity, precision, and applicability of models deemed to be most useful and effective.

METHODS Elk and Facilities

Seventeen 1.5-year-old, 19 2.5-year-old, and 7 adult (5- or 7-year-old) elk cows were housed in 4 1-ha pens near Kamela, Oregon (see Cook et al. 1998). All cows were nongravid. The adult elk had been captured as calves from a wild elk herd near La Grande, Oregon, during 1991 and 1993 and bottle-raised (Cook et al. 1996). The sub- adult elk resulted from previous reproduction studies with tame elk and were dam-raised but had daily human contact. Each pen was devoid of vegetation and had a barn containing 9-12 stalls designed for individual feeding and collection of blood, urine, and fecal samples.

Handling and Feeding Within each age group, we randomly assigned

elk to 1 of 3 processing dates, mid-September, late-December, or mid-March, times that man-

agers most often handle wild ungulates. We fed all cows high-quality hay and pellets ad libitum so that they were in good to excellent condition

prior to the study. Beginning 2.5 months before each processing date, we subdivided animals within each age group into 3 nutrition treatments (high, medium, and low) to diverge condition levels. Dietary manipulation involved varying quan- tity and quality of food rations using alfalfa, mixed-grass hays, and pellets. The pellet was formulated from oats, wheat, and alfalfa hay and contained 15.9% crude protein and 3.66 kcal of digestible energy per gram of dry matter (DE/g). The alfalfa hay averaged 17.4% crude protein and 2.66 kcal DE/g; the grass hay consisted of fescue and mixed meadow grasses and averaged 7.0% crude protein and 1.90 kcal DE/g (see Cook et al. 1996 for assay descriptions). The high nutrition diet was designed to maintain high nutritional condition during the 2.5-month feeding trial. We formulated the medium and low diets to induce average body-mass losses of 8-10% and >15%. We

weighed cows biweekly and adjusted feeding levels as necessary to attain desired mass change.

Each morning, we fed elk pelleted food in individual stalls; we provided hay communally in the afternoon in hay mangers located inside each

pen. Mangers were widely spaced to prevent exclusion. Cows were held in the barns for about 2 hours a day, a sufficient time for each to con- sume its designated pellets. This also provided habituation to the barns to reduce handling stress.

We placed all animals on identical diets 7 days prior to data collection to alleviate potential con-

founding effects of short-term nutritional influences on relations between condition indices and nutritional condition. We fed all elk a 35:65 ratio of high-quality pellets:high-quality hay. This provided 50 g of dry matter/(kg body mass)0.75, a maintenance amount for this quality of food. All food was fed individually in stalls to eliminate dominance exclusion.

Data Collection and Animal Processing At the end of the 7-day period, cows were

brought into the barn and fed pellets. We collected urine samples via a galvanized metal pan placed under each stall's floor. Samples were stored frozen at -20 oC until urinalysis (Table 1). The next

day, animals were brought into the barn and anesthetized with 70-100 mg of xylazine hydrochloride, administered intramuscularly. We obtained blood samples within 10 min of the drug injection by jugular venipuncture to avoid confounding drug effects (Cook et al. 1994). We allowed blood to clot for 30-45 minutes in SST vacutainers before cen-

trifuging. Serum was collected and stored frozen at -20 oC until analysis of serum indices (Table 1).

We collected all live-animal measurements (Table 2) while elk were anesthetized. We used a body-condition scoring (BCS) system (modified from Gerhardt et al. 1996) that involved averag- ing 3 separate scores derived from palpation of the ribs, withers, and rump areas (scoring criteria described by Cook 2000). We calculated a body- reserve index (BRI = BCS x body mass; Gerhardt et al. 1996). We followed methods of Farley and Robbins (1994) for bioelectrical impedance analysis. Briefly, cows were positioned sternally recumbent, and we attached electrodes to the gums under each first incisor and on either side of the tailhead with the current-carrying electrodes on the animal's left side. We measured subcutaneous rump fat thickness using ultrasonography (MAXFAT; Stephenson et al. 1998). We also determined scapula muscle (infraspina- tus plus scapular deltoid) thickness taken at the midpoint and 2.5 cm posterior to the scapular spine and longissimus dorsi muscle thickness taken between the 12th and 13th ribs directly beneath the backbone (see Herring et al. 1995).

J. Wildl. Manage. 65(4):2001 DEVELOPMENT OF ELK CONDITION MODELS * Cook et al. 975

Table 1. Serum and urine chemistry indices evaluated for cow elk, 1998-1999. Acronyms (if any) and units of measure are

provided in parentheses.

Serum indices Urine indicesa

Albumin (mg/dl) *Creatinine (C) (mg/dl) *Serum Urea Nitrogen (mg/dl) *Urea Nitrogen (UN) (mg/dl) *Total Protein (g/dl) *Cortisol (gg/dl) *Total Bilitrubin (mg/dl) *Potassium (UK) (meq/I) *Triglycerides (mg/dl) Sodium (UNA) (meq/I) Cholesterol (mg/dl) *Phosphorus (meq/dl) *Creatinine (mg/dl) *Calcium (mg/dl) *Calcium (mg/dl) *Sodium (meq/l) *Potassium (meq/dl) *Inorganic Phosphorus (mg/dl) *Chloride (meq/dl) *Triiodothyronine (T3) (ng/dl) Thyroxine (T4) (gg/dl) *Alkaline Phosphatase (ALP) (lU/I) *Alanine Aminotransferase (ALT) (IU/I) *Asparate Aminotransferase (AST) (IU/I) Gamma Glutamyltransferase (GGT) (IU/I) *Glucose (mg/dl) Insulin-like Growth Factor-1 (IGF-1) (ng/ml)

a Values of urine variables were divided by creatinine values to account for variation in urinary dilution. These ratios were further adjusted to facilitate comparisons to published literature as follows: potassium:creatinine x 100; sodium:creatinine x 100; phosphorus:creatinine x 1,000; calcium:creatinine x

1,000. * Indices with r2 < 0.25 with body condition (%FAT, gross

energy).

We transported the elk the next day by truck to a holding pen and processed them at the Meat Sciences Lab at Oregon State University, Corval- lis. Elk were sedated using xylazine hydrochloride and euthanized via jugular injection of sodium pentobarbital. Cows were hung, eviscerated

(all blood was collected), weighed, and carcass fat, musculature, and visceral fat were visually scored via the Kistner score (Kistner et al. 1980) and what we refer to as the Wyoming Index (Lanka and Emmerich 1996), both of which were

developed for deer (Odocoileus spp.). The Kistner

system requires scoring based on (1) fat in indicator depot sites (cardiac, omental, perirenal, and subcutaneous areas); and (2) the condition of the skeletal muscle mass (Kistner et al. 1980). Each area is scored in increments of 5 from 0 to 15. We found that elk carried more fat around some internal organs (i.e., cardiac and perirenal

areas) than do deer. Thus, we modified the Kist- ner Score by increasing the scoring range from 0 to 20 for these organs, by removing the muscle mass evaluation due to its subjective nature, and

by scoring in increments of 1 point. The

Wyoming Index is based on presence or absence of fat at the rump, stifle, and withers area (Lanka and Emmerich 1996). Because of the difficulty in

using this method on animals with substantial fat reserves, we partitioned the withers score into 3

categories and scored for absence or presence at the crest of the withers, and at the third and fifth vertebra posterior to the crest.

We sawed each carcass (minus the viscera) in half from nose to tail along the vertebrae

(Stephenson et al. 1998). One half of the carcass, along with the hide and hair, was sectioned and stored at -20 oC. This half was later homogenized to determine body composition. The other half was used as a source for various samples. The middle third of the femur and the mandible were collected for bone marrow analysis (ovendry method; Neiland 1970). We also evaluated aver-

age nonfat residue of mandible marrow fat via ether extraction on a subset of 20 mandibles.

We removed and weighed the liver, heart (minus the pericardium and blood), and kidneys with all attached fat. We trimmed perirenal fat

according to Riney (1955), and weighed the kid-

neys, remaining fat, and trimmed fat. The tunica fibrose (the capsule of connective tissue encasing the kidney) was removed with the fat. We calculated kidney fat indices (KFI) based on total fat mass and trimmed fat mass:

KFIfull = (total kidney fat mass x 100)/kidney mass;

KFItrim = (trimmed kidney fat mass x 100)/kidney mass.

For all analyses, we estimated KFIs separately for each kidney and calculated the average (Anderson et al. 1972). We combined the removed organs and blood with the trachea, lar-

ynx, diaphragm, esophagus, and all contents of the pleural and peritoneal cavities, exclusive of the ingesta, weighed, and stored them frozen.

We reweighed half carcasses, visceral masses, and hide samples just before grinding to account for water loss and homogenized them in a whole-

body grinder (Autio 801 B with a Falk 50 hp grinder) at the University of California, Davis. Carcass samples were ground 3 times through a 2.5-cm screen and once with a 6-mm plate open- ing. Viscera and hide samples were ground twice through a 6-mm screen. We collected 2 samples


Table 2. Live and dead animal indices evaluated for cow elk, 1998-1999. Acronyms used in the text are provided.

Live animals Dead animals

index Acronym Index Acronym

Body condition score BCS Modified Kistner score KISTNER Rump score of BCS rumpBCS Kistner score subset

(heart + pericardium + kidneys) KISThpk Maximum rump fat thickness MAXFAT Kidney fat mass KFmass Scapula muscle thickness SCAP Kidney fat index-trimmed KFItrim Loin muscle thickness LOIN Kidney fat index-full

KFIfuI LIVINDEXa LIVINDEX Modified Wyoming index WYI Rump LIVINDEXb rLIVINDEX Femur marrow fat percent FEMUR *Bioelectrical impedance analysis BIA Mandible marrow fat percent MAND *Hind foot length HFL CONINDEXc CONINDEX *Total body length BL Eviscerated mass EVISWT Body mass BM Girth GIRTH Body reserve index BRI

a Combination of MAXFAT and BCS. b Combination of MAXFAT and rump BCS. c Combination of FEMUR and KFI (full or trimmed). * Indices with r2 < 0.25 with body condition (%FAT, gross energy).

of the ground tissues (approx. 1.5 kg of the carcass, approx. 0.75 kg of the viscera, and approx. 0.2 kg of the hide-hair sample) and stored them frozen until chemical analysis.

Body Composition of Wild Elk We included 6 wild cows from northeast Ore-

gon to assess differences in condition indices of wild versus captive elk (Cook et al. 2001). Five of these animals (>2 years old) were included for assessment of mandibular fat. They were captured by baiting in February 1999 and handled as described above except that: (1) animals were held 1-2 days in a holding pen until processing, weighed in a chute system, sedated using xylazine hydrochloride (200 to 300 mg) to collect live-animal data, and then euthanized with sodium pentobarbital; (2) urine was collected directly from the bladder or from the concrete floor when sedated; (3) fetal tissues and fluids were removed and weighed (4 of 6 were pregnant); and (4) cows were aged by tooth wear.

Chemical Analysis We freeze-dried carcass, viscera, and hide sam-

ples to a constant weight and re-homogenized them in a Wiley grinder through a 1-mm screen with dry ice. We determined crude protein by the Kjeldahl procedure (Association of Official

Agricultural Chemists 1980), percent fat by ether extract (Association of Official Agricultural Chemists 1965), and total ash by combustion for >2 hours at 500 oC (Association of Official Agri- cultural Chemists 1960). We combined carcass and viscera compositions according to their rela- tive mass to estimate ingesta-free composition of the whole body. We calculated gross energy (GE in Mcal/kg; Robbins 1993) by:

GE = [(9.11 x %Fat) + (5.65 x %Protein)]/100

We calculated fat-free, ingesta-free lean body

mass by subtracting total fat (kg) from the ingesta-free body mass. Water, protein, and ash were expressed as percentages of lean body mass.

Serum and Urine Chemistry Assays We determined levels of IGF-1 using a double-

antibody radioimmunoassay (RIA; Oklahoma State University, Stillwater). Levels of serum T3 and urinary cortisol were analyzed via RIA (Col- orado State University, Fort Collins). Interpath Laboratory (Pendleton, Oregon) conducted all other serum and urine assays using a Boehringer Mannhein/Hitachi 737 analyzer and reagent sys- tems specified by the manufacturer. We adjusted each urinary value for variations in urine con- centrations by dividing each index by urinary creatinine levels (Table 1).


Statistical Analysis We began by assessing influences of age and

season on relationships between each index and measures of nutritional condition (fat, GE) using ANCOVA (PROC GLM; SAS Institute 1988) with season and age as covariates. This was done to

identify the need for separate equations for each level of the 2 factors. When necessary, data were linear transformed. We excluded all indices with r2 < 0.25 from further analyses. For remaining indices, if age or date effects were significant (P< 0.05), we determined which differed by sequential removal. Where these effects were significant, we conducted a supplementary analysis because in some cases, the ANCOVA apparently indicated significant effects because the range of condition values differed across factor levels. Thus, significant age or season effects may be a

spurious result of little practical relevance. We

developed a predictive equation, with simple linear regression, using only data from factor levels that were not significantly different from each other. We calculated predicted values for data from all 3 factor levels using this equation, and residuals were calculated for each animal. Using 1-way, fixed effects ANOVA (PROC GLM; SAS Institute 1988), residuals from the statistically different level were compared to residuals from the other 2 factors. If significant (P 5 0.05), we con- cluded that the age or season effect was suffi-

ciently large to affect predictive ability, and equations were generated per season or age for that index. If not significant, 1 equation was generated across all ages or seasons.

Next, we generated multivariable models via

multiple regression to increase predictive ability. An initial stepwise regression analysis using Akaike's Information Criterion and Mallows C

(SAS Institute 1988:786) to select regression mod?- els provided results that often were biologically

unrealistic, unstable due to multicollinearity, and

overparametized (25 variables). Instead, we used a more intuitive approach that emphasized sim-

plicity and field practicality. We grouped only indices that would be practical to collect at 1 time under field conditions or at hunter check sta- tions, and considered a variety of logistical limitations. For example, we avoided grouping live- and dead-animal indices, and urine indices were not included with indices requiring capture (urine samples generally are collected from snow). We also limited the number of indices per model to 3 due to our small sample size and added indices only if they were significant in the

model or increased the adjusted r2 ? 0.02 (ra2).

In addition, model variables were screened for mul-

ticollinearity using a variance inflation estimate function; values 210 indicated multicollinearity and such models were rejected (Mendenhall and Sincich 1989). This process reduced potential for spurious models with apparent high predictive ability but little biological justification or

management applicability. Similar procedures were used to generate models for total body com-

ponents (kg of fat and water, and Mcal of gross energy; see Appendices A and B).

Third, we created 2 single-variable indices (LIVINDEX, CONINDEX) from arithmetic com- binations of 2 indices, using indices with different ranges of predictive ability. This procedure mimics that of the CONINDEX (Connolly 1981) for deer in which femur marrow fat and KFI were combined such that (1) when KFI ? 20, CONIN- DEX = (KFI - 20) + femur marrow fat; and (2) when KFI < 20, CONINDEX = femur marrow fat. We combined in a similar manner the variables BCS and MAXFAT so that (1) when MAXFAT

>0.3, LIVINDEX = (MAXFAT - 0.3) + BCS; and

(2) when MAXFAT <0.3, LIVINDEX = BCS (0.3 cm represents the point at which rump fat is

depleted and measurements reflect fascia thickness). We also combined MAXFAT and only the

rump portion of the BCS in the same manner to

produce rLIVINDEX.

RESULTS Total body fat ranged from 1.6 to 19.0%, GE

ranged from 1.39 to 2.77 Mcal/kg, and protein ranged from 16.6 to 24.8% of the ingesta-free body. Live mass ranged from 135 to 245 kg and mass change ranged from +1.0 to -21.5% across the 2.5-month feeding period. Percent body fat was linearly related to GE (Fat% = 10.034x -

12.062, r2 = 0.96, P < 0.001) and percent water

(Fat% = -1.234x + 87.973, r2 = 0.94, P< 0.001) for

captive elk. Water, protein, and ash accounted for 71.7 ? 0.2%, 23.0 ? 0.2%, and 5.3 ? 0.2% (? SE) of the fat-free, ingest-free body mass and these rela-

tionships were used to predict body protein (Appendix C).

Twenty-four indices (Tables 1 and 2) accounted for >25% of the variation in body fat and GE. Most

indices, particularly serum and urine (including urea nitrogen:creatinine, urinary potassium:creatinine, and cortisol:creatinine; Fig. 1), related poorly to nutritional condition (r2 < 0.25). Of these indices, 12 had significant relations with body condition (P < 0.05), whereas the serum indices


20

15 we on

NF 5 .

4 .

. iE-

0 2 4 6 8 10 12 Urea nitrogen:creatinine

20

15 - ONmm NE

Jff 10- Jim

I-,,

5 ? * a U

0 100 200 300 400 500

Urinary potassium:creatinine x 100 20

10

S5 1. . ..5- ons

-0 -------m

0 0.05 0.1 0.15 0.2 0.25 0.3 Urinary cortisol:creatinine

Fig. 1. Relations of 3 urinary indices (urea nitrogen, urinary potassium, and urinary cortisol) with total body fat (%) for 43 captive-raised cow elk. All values are presented as a ratio of creatinine to account for variation in urinary dilution.

urea nitrogen, tryglycerides, creatinine, inorganic phosphorus, and alkaline phosphatase, and the urine indices urea nitrogen:creatinine and cortisol:creatinine did not have significant relations with body condition (P 2 0.06).

Age and Season Effects Cow age significantly influenced relations

between nutritional condition and MAXFAT, femur, and mandibular marrow fat (Table 3), as did season on relations between condition and the modified Kistner score and T4 (Table 3). Of these, only age effects for mandible marrow fat (Fig. 2) and season effects for T4 (Fig. 3) were deemed practically relevant (Table 3) based on ANOVA of residuals. Although IGF-1 did not show a statistically significant seasonal effect, predictive equations differed markedly among seasons (Fig. 3), so we assumed a relevant seasonal effect.

Live Animal Indices For live animals, LIVINDEX (calculated from

either the whole BCS or only the rump portion) was most related to % fat (r2 = 0.90) and GE (r2 =

0.86; Table 4). Both BCS (using the entire score or only the rump portion) and MAXFAT separately were highly related to condition (Table 4), although MAXFAT cannot be used once subcutaneous rump fat is depleted. Body mass alone was poorly related to condition and failed to increase the correlation of BCS when they were combined into a body-reserve index (Table 4). Both bioelectrical impedance analysis and body mass were significant in a multiple regression model (ra2 = 0.86, P < 0.05) used to predict % fat from total body water. Removing bioelectrical impedance analysis, however, failed to lower the ra2, indicat-

Table 3. Significant age and season effects for nutritional condition indices evaluated during the elk body composition study 1998-1999. Index acronyms are presented in Tables 1 and 2.

Index Fat (P) GE (P) Factor-level differencesa ANOVA (p)b

Significant age effects MAXFAT 0.012 0.001 YR # 2Y; AD # 2Y; YR = AD 0.519 FEMUR 0.004 0.007 YR # 2Y; AD # 2Y; YR = AD 0.360 MAND 0.070 0.061 YR # 2Y; AD # 2Y; YR ? AD <0.001

Significant season effects KISTNER 0.004 0.018 SEP # DEC; SEP # MAR; DEC = MAR 0.087

T4 0.035 0.063 SEP # DEC; SEP # MAR; DEC = MAR <0.001

a Three levels were included for the factor Age: yearlings (YR), 2 years old (2Y), and adults >3 years old (AD). Three levels were included for the factor Season: September (SEP), December (DEC), and March (MAR). Differences among factor levels were based on removing 1 level at a time.

b P-values for analysis of variance used to test residuals of significantly different level.


20 2 year olds

10

?

15 ?

20 Y C. 40 45 50 55 60 65 70 75 80

Mandible marrow fat (%)

20

Yearlings

15 Adults n

Do 5o ,. .o .. .o. 8

40 45 50 55 60 65 70 75 80 Mandible marrow fat (%)

Fig. 2. Relation of mandible marrow fat (%) as determined by ovendry method and total body fat (%) for 43 captive-raised cow elk for adults (>3 years old), 2 year olds, and yearlings. Open circles with plus marks represent 5 wild cows from northeast Oregon >3 years old included in adult relation.

ing little value in including it into the model. In addition, frame size and skeletal measurements (chest girth, hind foot length, and total body length) were unrelated to condition either separately or combined with other indices. Combin-

ing either BCS or LIVINDEX with muscle measurements (scapula or loin) increased correlation to condition (Table 5).

Thyroxine (T4) and IGF-1 were the only single serum or urine indices useful in predicting condition. Even so, this relation was restricted to early and late winter (Dec and Mar; Table 4). Models with multiple serum and urine indices only weak-

ly increased the correlation. The combination of

T4, In(IGF-1), and alkaline phosphatase gave the highest ra2 (0.90) using serum and urine variables, but only in December (Table 5).

Dead Animal Indices For dead elk, the modified Kistner score and

the Kistner subset score (heart, pericardium, and

kidney) were most related to condition (Table 4). The Wyoming Index was moderately related to

20 O

5 2 20 o-

20,)

15

0

0 20 40 60 80 Insulin-like growth factor (nglml)

Sep Dec Mar . .--O-- p--

Fig. 3. Seasonal trends in the relationship of thyroxine (T4) and insulin-like growth factor-1 (IGF-1) to total body fat (%) for 43 captive-raised cow elk.

condition but only when subcutaneous fat was present. Kidney fat mass (KFmass) alone was superior to KFIfull, and KFIfull was superior to the traditional method of trimming (KFItrim;

Table 4). Although CONINDEX worked well (r2 =

0.70), it was linear only at low to moderate levels of condition (<12.5% fat) but had no predictive ability at higher levels of condition. Using KFItrim (r2 = 0.72) or KFIfull (r2 = 0.70) to devel-

op the CONINDEX produced the same linear

relationship, but using KFmass produced a nonlinear relation. Combining the Wyoming index with KFmass produced a model that was superior to both single-index models (Table 5).

Femur marrow fat produced an r2 of 0.89 using a transformation of the dependent variable, but was highly curvilinear and useful only below 6% body fat (Fig. 4). Mandible marrow fat was less curvilinear, but due to the confounding effect of

age, our estimate of the relationship for adults is tentative due to small sample size (n = 12, includ-


Table 4. Single index regression equations for ingesta-free body fat (%) and gross energy (Mcal/kg) for Rocky Mountain cow elka. All regression equations are significant (P 5 0.001). Index acronyms are presented in Tables 1 and 2.

Ingesta-free body fat Ingesta-free gross energy x ,0 ,1? r2 0 f,8 r2

LIVINDEX -0.055 2.316 0.90 1.287 0.183 0.86 rLIVINDEX 0.259 2.237 0.89 1.314 0.176 0.85 MAXFATb 5.630 3.550 0.87 1.734 0.281 0.81 BCS -5.526 4.776 0.87 0.842 0.381 0.85 Rump BCS -4.618 4.478 0.86 0.924 0.355 0.83 BRI -0.263 0.016 0.79 1.271 0.001 0.76 In(IGF-1)c -0.802 3.165 0.54 1.156 0.266 0.54 In(lGF-1)d -0.932 4.213 0.73 1.164 0.349 0.74 In(IGF-1)e -7.915 5.504 0.82 0.737 0.421 0.81

T4d -6.229 2.997 0.72 0.707 0.252 0.74

T4e -7.151 2.693 0.65 0.839 0.199 0.59 BM -9.670 0.104 0.44 0.536 0.008 0.42 LOIN -20.302 5.701 0.44 -0.491 0.484 0.48 KISTNER -1.300 0.177 0.92 1.193 0.014 0.87

KISThpk -4.469 0.405 0.90 0.922 0.035 0.89

In(KFmass) -13.050 4.573 0.86 0.275 0.359 0.81

In(KFIfull) -9.630 4.586 0.77 0.544 0.359 0.72

In(KFItrim) -13.528 5.981 0.74 0.237 0.469 0.70 CONINDEXf 1.730 0.050 0.70 1.438 0.004 0.65 WYI 1.515 0.454 0.69 1.394 0.037 0.70 MAND9 -0.861 0.048 0.47 -0.300 0.016 0.52 MANDh -3.986 0.094 0.80 -0.950 0.026 0.64 FEMUR' -0.777 0.007 0.89 -0.863 0.004 0.52

a Linear equation is of the form: y = f,1x + ,Po b Only measurements 20.3 cm were used to determine parameter estimates. C Regression equation intended for use in September. d Regression equation intended for use in December. e Regression equation intended for use in March.

f Combination of FEMUR and KFifull. 9 Regression equation for use in yearlings in the form In(y) = f,1x + ,0.

h Regression equation for use in adults (>3.5 years old) in the form In(y) = px + o0. iLinear equation is of the form: -1/y = ,1x + f,o

ing our wild elk data; Fig. 2). The nonfat residue component of mandible marrow was 7.85 + 0.42%. This residue was unrelated to age (P= 0.66) or animal condition (P = 0.77) using ANCOVA.

DISCUSSION Our approach to end the nutrition treatments

and place all elk on identical maintenance diets 7 days prior to data collection is infrequently used, and it strongly tempers our results and conclusions. By choosing to feed identical diets, we opted to provide inferences only of the ability of indices to predict nutritional condition. Our data provide no insights of relations between indices and nutrition. We emphasize this because it is our

experience that many biologists are unaware of this key distinction between nutrition and nutritional condition and their associated indices.

Live Animal Indices Ultrasonography of rump fat and muscle thick-

ness has been used in the livestock industry (Bul- lock et al. 1991, Domecq et al. 1995) and in wildlife studies (Stephenson et al. 1998). Bright- ness mode real-time ultrasonography generates grayscale 2-dimensional images that readily dis- tinguish tissue layers (skin, fat, and muscle) that are measurable with electronic calipers (? 1 mm; Stephenson et al. 1998). Although highly correlated to body fat and gross energy, ultrasound fat


Table 5. Multiple index models for ingesta-free body fat (%) and gross energy (Mcal/kg) for Rocky Mountain cow elka. All regression equations are significant (P ? 0.001). Index acronyms are presented in Tables 1 and 2.

y xl x2 x3 Pl P2 P3 Po r a2

%FAT LIVINDEX LOIN SCAP 2.882 -2.525 0.057 10.939 0.94 %FAT BCS LOIN SCAP 6.228 -2.586 -0.081 4.033 0.92 %FAT BCS SCAP 5.549 -2.055 -1.594 .0.89 %FAT BCS LOIN 5.631 -1.814 1.318 0.88 %FATb ALP T4 In(IGF-1) -0.011 1.479 2.650 -3.623 0.87 %FATb T4 In(IGF-1) 1.741 2.572 -6.305 0.84 %FAT WYI In(KFmass) 0.016 3.467 -10.477 0.89 GE LIVINDEX LOIN SCAP 0.222 -0.141 -0.045 2.031 0.93 GE BCS LOIN SCAP 0.488 -0.152 -0.069 1.542 0.92 GE BCS SCAP 0.448 -0.184 1.212 0.91 GEb ALP T4 In(IGF-1) -0.001 0.127 0.215 0.936 0.90 GEb T4 In(IGF-1) 0.150 0.208 0.701 0.86 GE WYI In(KFmass) 0.016 0.252 0.523 0.86

a Multiple regression equation is of the form: y = o + 31fx, + P322 + P33x3 b Regression equations intended for use in December.

measurements in elk become meaningless after subcutaneous rump fat is depleted (<5.5% body fat). At this point, measurements <0.3 cm reflect

only fascia thickness.

Body-condition scoring (BCS) has been used to evaluate condition and reproductive status and to

predict calving difficulty and pregnancy rates in livestock (Jeffries 1961, Wildman et al. 1982, Edmonson et al. 1989). Despite its potential for

subjectivity, studies on livestock (Wright and Rus- sell 1984) and wild ungulates (Gerhardt et al. 1996) reported high correlation between BCS and

body fat content, which is consistent with our

findings. Moreover, BCS was linearly related to fat and GE over the entire range of condition in our

study, and did not display seasonal or age effects. Gerhardt et al. (1996) reported predictability

was enhanced by combining BCS and body mass in a body-reserve index (BRI). Our BRI showed a weaker relationship to key body components than did BCS, but it did predict total amount of body components (kg) as accurately as BCS (Appendix A). Variation in ingesta weight, level of dehydra- tion, pregnancy status, and animal frame size

probably weakened the relation between BRI and nutritional condition in our study.

An arithmetic combination of BCS and rump fat thickness (LIVINDEX) was the most correlated to condition of any live-animal index. Com-

bining these 2 indices reduces potential subjectivity over moderate and high levels of condition where rump fat is more effective, and relies sole-

ly on BCS only on the low end of condition (the range where BCS appears to be least subjective; Cook, unpublished data). In addition, using LIVINDEX combined with muscle measurements (loin and scapula) adds additional objective measurements at the low end of condition, expands the utility of ultrasonography, and may track

body protein reserves once fat deposits are large- ly depleted (Cook 2000). Although we present the LIVINDEX relation as linear for simplicity, a

polynomial fit may be more biologically correct

(Fig. 5), especially for the extremes of body condition. This assertion is based on experiments (Cook, unpublished data) that included severely

20 OO

15

10

, ...............i.,..,..., 0 20 40 60 80 100 Femur marrow fat (%)

Fig. 4. Relation of femur marrow fat as determined by ovendry method to total body fat (%) for 43 captive-raised cow elk. Open symbols represent the range where femur fat loses sensitivity.


25 %fat = -9.8863584 + 9.1871285x -

201.3831754x2 + 0.083951218x3

10 -.

0

1 2 3 4 5 6 7 8 9 LIVINDEX

3.25

30 GE = 0.37206184 + 0.82280587x - . 0.12853573x2 + 0.007762878x3

S2.75

S2.50 .

0 2.00

1.25

1 2 3 4 5 6 7 8 9 LIVINDEX

Fig. 5. Relations of LIVINDEX (a combination of rump fat thickness and body condition score) with total body fat and gross energy presented as both a linear and polynomial fit. The polynomial equation is presented for each graph. Equations using rLIVINDEX (a combination of rump fat thickness and rump BCS) are as follows: Fat% = -7.1527185 + 7.323081x - 0.98980456x2 + 0.057445567x3; GE = 0.47782556 + 0.77267613x - 0.12202139x2 + 0.0075108143x3.

malnourished elk and other elk with consider-

ably more fat than those in the present study. Bioelectrical impedance analysis (BIA) has

been shown to work well on nonruminants such as bears (Ursus americanus; Farley and Robbins

1994, Hilderbrand et al. 1998). However, we found no relation of BIA to total body water, and thus body fat, probably due to (1) the difficulty in standardizing body position of elk under field conditions (i.e., inconsistent positioning changes signal path geometry, thereby introducing bias; Hall et al. 1989, Farley and Robbins 1994); (2) the influence of rumen water on current flow; and (3) the need for a nonconductive surface, which is difficult to achieve in field conditions.

Serum Indices We found that most serum indices were poorly

related to condition (see also Franzmann 1985, Harder and Kirkpatrick 1994). However, insulin-

like growth factor-1 and thyroxine were significantly related to condition. Insulin-like growth factors (IGF-1 and IGF-II) are a family of polypeptides, related structurally and evolutionarily to proinsulin (Gluckman et al. 1987). They are involved in many biological processes including postnatal growth, lactation, reproduction, and immune function (McGuire et al. 1992). Nutrition and IGF-1 are

directly correlated (McGuire et al. 1992, Adam et al. 1995, Webster et al. 1996), and seasonal effects also have been reported (Ringberg et al. 1978, Webster et al. 1996, Quinlan-Murphy 1998). Levels of IGF-1 are correlated with body-mass change as well (Suttie et al. 1989, Quinlan-Murphy 1998).

Two thyroid hormones, triiodothyronine (T3) and thyroxine (T4), are important in metabolism, growth, and thermogenesis. In general, increas-

ing levels of T3 are associated with increasing metabolic activity (Newsholme and Leech 1983). Thyroxine may be more diagnostic of nutritional condition, whereas T3 may be more diagnostic of short-term nutrition (Bahnak et al. 1981, Franz- mann 1985). But studies investigating effects of nutrition and condition on T3 and T4 offer mixed results. Most report a definite response of T3 to energy in the diet (e.g., Bahnak et al. 1981, Hamr and Bubenik 1990, Cook et al. 1994, Brown et al. 1995), but response of T4 was inconsistent (see Ryg and Jacobsen 1982, Hamr and Bubenik 1990). In most cases, however, condition of animals was not evaluated. Other studies that evaluated condition (chemical composition or mass loss) found T3 to be more correlated than was T4 (Watkins et al. 1991, Quinlin-Murphy 1998), the opposite of our findings. However, the interaction between T4 and season suggests that despite relationships in winter, T4 probably reflects metabolic status more so than condition. During December and March, T4 may vary more in response to animal condition than in September (i.e., perhaps there is an interaction between photoperiod and condition on T4).

We found other serum indices to be of little value for assessing nutritional condition. We suspect many of these are rate variables, or more reflec- tive of short-term nutritional status. For example, 2 of the most studied ones, serum urea nitrogen and glucose, have found to be strongly influenced by short-term nutritional status (Ganong 1979, Cook et al. 1994, Harder and Kirkpatrick 1994).

Urinary Indices

Urinary indices were poor predictors of condition and were of little value for identifying even


our most emaciated elk. For example, 7 elk had <5% body fat (a level in which increased muscle catabolism should be occurring; Parker et al. 1994); only 2 of these 7 had elevated urinary nitrogen:creatinine ratios (>3.5; DelGiudice et al. 1991), 3 had elevated urinary potassium:creatinine ratios (>200), and 2 had elevated cortisol:creatinine ratios (>0.1 tig/mg; Saltz and White 1991; Fig. 1). These results should be ex-

pected, however, particularly given our decision to place all elk on equal, maintenance diets a week before data collections. This is because these urinary indices generally reflect the rate at which lean body tissue is consumed. Nutritional condition is a state variable and cannot be measured in terms of rates (Saltz et al. 1995). Al-

though urinary indices have received consider- able attention and are easy to collect from snow, our data (and see Parker et al. 1994, Saltz et al. 1995) indicate these indices should never be used to assess condition.

Dead Animal Indices Because marrow fat appears to be 1 of the last

stores of fat to be used, low values of marrow fat tend to indicate acute nutritional deprivation (Harder and Kirkpatrick 1994). Bubenik (1982) suggested that marrow fat percentages in red deer (Cervus elaphus) >80% prior to winter and >60% in late winter indicated excellent condition, but our results supported those of Mech and DelGiudice (1985), who found that any loss of this fat reserve indicated poor condition. Aver-

age femur marrow fat values for animals with >6% body fat were 91.8 ? 0.56% (?SE; n = 33), thus values >90% only indicate >6% body fat, not excellent or even moderate condition. Values <90% indicate poor condition (<6% fat), and values <12% indicate acute starvation (Ratcliffe 1980, Depperschmidt et al. 1987).

Mandible fat was considered an easier-to-collect alternative to femur fat because it shares a linear relation with femur fat in moose (Alces alces; Bal- lard et al. 1981) and caribou (Rangifer tarandus; Davis et al. 1987). But our findings suggest caution in adult elk because we found a more com-

plex, nonlinear relation (and see Snider 1980, Cederlund et al. 1986, Okarma 1989). Differ- ences in studies may be due to variation among species. Whatever the cause, mandible fat in our elk ranged from 50.0% to 72.1% across a femur fat range of 14.6% to 95.3%, a difference apparently not explained by a higher mandible nonfat residue component (Davis et al. 1987, Marquez

and Coblentz 1987). Thus, elk that had depleted large amounts of femur fat depleted little of their mandible reserves, suggesting rapid depletion of mandible fat at very low levels of condition (see also Okarma 1989). In addition, mandible fat levels in our elk never exceeded 75%, as reported for red deer by Okarma (1989; but see Ballard et al. 1981 for moose and Davis et al. 1987 for caribou), suggesting that, unlike femur fat, some uti- lization of mandibular fat may be occurring at

high levels of condition.

Finally, differences in depletion of mandible fat

among age classes further limited the value of this index. Data from yearlings were linear with a hint of curvilinearity, whereas adult data

(including wild elk) followed an exponential relation. Data from 2-year-old cows, however, showed no relation between mandible fat and total body fat. This lack of relationship is difficult to explain. Differential fat deposition and depletion patterns in post-pubescent, pre-adult animals similar to those found in teenage human females is a possibility (Tanner 1968, Sinclair and

Dangerfield 1998). We also wonder if eruption of adult molars and premolars, growth processes that come to completion in 2 year olds, might influence mandible fat levels as well. Whatever the explanation, we caution use of mandible fat as a condition indicator.

Kidney fat indices (KFItrim, KFIfull, KFmass) pro-

duced a logarithmic relation with condition, limit-

ing its use for cows in very poor or very good condition (see also Finger et al. 1981, Depperschmidt et al. 1987). Combining kidney fat and femur marrow fat indices (CONINDEX) extended the linear range to a lower level of condition but did not extend the upper limit. Among KFI evaluated, perirenal fat mass (KFmass) alone was superior to either method using kidney mass. Moreover, al-

though we did not find a significant seasonal effect on kidney mass, other studies have reported such effects (Dauphind 1975, Anderson et al. 1990). Either way, our data support conclusions of Harder and Kirkpatrick (1994) that dividing kid-

ney fat by kidney mass probably should be avoided, particularly when comparing across seasons.

Carcass/musculature scores have produced strong correlations to condition, particularly the Kistner score (Kistner et al. 1980, Watkins et al. 1991). We also found a tight linear relationship between the modified Kistner score and body components. Also, using only the heart, pericardium, and kidney scores (the most easily attainable and identifiable parts of the Kistner


score), we were able to predict body composition almost as well as the whole score across the entire range of condition. The Wyoming index represents an even simpler and more practical tech-

nique. It works relatively well as long as subcutaneous fat is present, but combining with other indices improves its value. As an example, we combined it with KFmass and produced a superior model to either single-variable models.

MANAGEMENT IMPLICATIONS The large number of nutritional condition

indices present biologists with daunting choices. This is particularly true for elk biologists because no study has provided a definitive assessment of these indices for elk (Cook 2002). This study, plus our companion study (Cook et al. 2001) were intended to fill this gap for Rocky Mountain elk, with explicit focus on nutritional condition. Biol-

ogists face 2 levels of choices: (1) monitoring nutrition versus nutritional condition; and (2) deciding which technique provides affordable, yet reliable data. Nutrition data usually are con- founded by changes in weather, forage phenology and availability, and even the stress of capture; thus, useful insights from nutritional indices

require relatively long-term, sequential sampling (e.g., DelGiudice et al. 1991, Garrott et al. 1996). In contrast, nutritional condition results from cumulative energy balance, so a single estimate of nutritional condition provides useful insights cov- ering a substantial period of time. This funda- mental difference should be considered in light of objectives of research and monitoring efforts. Our findings are that traditional serum and urine indices provide little potential for assessing condition (see also Saltz et al. 1995), but there are valuable nutritional condition indices such as the Kistner full or subset score for dead animals and the LIVINDEX or body-condition score for captured, live animals (Cook et al. 2001).

ACKNOWLEDGMENTS Financial support for this study was provided by

Idaho Department of Fish and Game and Rocky Mountain Elk Foundation. Support of the elk herd was provided by the Northwest Forestry Association, Boise Cascade Corporation, Nation- al Council for Air and Stream Improvement, U.S. Forest Service Pacific Northwest Forest and Range Experiment Station, Oregon Department of Fish and Wildlife, and Rocky Mountain Elk Founda- tion. This support was facilitated by L. L. Irwin, L. D. Bryant, R. A. Riggs, and J. G. Kie. We

acknowledge the foresight and work of J. W. Thomas that established the infrastructure that supported this study. We thank the University of Idaho for logistical support, the U.S. Forest Ser- vice for supplying the ultrasonograph, B. L. Dick, J. C. Nothwang, R. O. Kennedy of the U.S. Forest Service, and D. Borum of Oregon Department of Fish and Wildlife for transporting the elk, R. L. Greene of the Oregon Department of Fish and Wildlife for logistical aid, B. Dixon of Oregon State University for allowing the use of the meat processing lab, and R. D. Sainz and T. Mori of the University of California (Davis) for use of the whole-body grinder and their slaughter facilities. T. R. Stephenson of the Alaska Department of Fish and Game provided ultrasonograph training, D. Vales and the Muckleshoot Indian Tribe, and L. C. Bender and R. Spencer of the Washington Department of Fish and Wildlife allowed use of Green River wild elk condition data. C. T. Rob- bins provided the BIA machine and valuable guid- ance on processing procedures. B. B. Davitt provided analysis of tissue composition and elk food. L. C. Bender reviewed the manuscript. We also thank T. E. George, N. Suverly, M. Alldredge,J. H. Noyes, and S. E. George for their assistance with handling and processing animals. This research was conducted in accordance with approved animal welfare protocol (Wisdom et al. 1993).

LITERATURE CITED ADAM, C. L., C. E. KYLE, P. YOUNG, AND T. ATKINSON.

1995. Effect of nutritional growth restriction on tim- ing of reproductive development and plasma con- centrations of insulin-like growth factor-i and growth hormone in male red deer (Cervus elaphus) reared in constant photoperiod. Animal Science 61:155-160.

ANDERSON, A. E., D. C. BOWDEN, AND D. E. MEDIN. 1990.

Indexing the annual fat cycle in a mule deer population. Journal of Wildlife Management 54:550-556.

-, D. E. MEDIN, AND D. C. BOWDEN. 1972. Indices of carcass fat in a Colorado mule deer population. Journal of Wildlife Management 36:579-594.

ASSOCIATION OF OFFICIAL AGRICULTURAL CHEMISTS. 1960. Official methods of analysis. Ninth edition. Associa- tion of Official Agricultural Chemists, Washington, D.C., USA.

. 1965. Official methods of analysis. Tenth edition. Association of Official Agricultural Chemists, Washington, D.C., USA.

. 1980. Official methods of analysis. Thirteenth edition. Association of Official Agricultural Chemists, Washington, D.C., USA.

BAHNAK, B. R., J. C. HOLLAND, L. J. VERME, AND J. J. OZOGA. 1981. Seasonal and nutritional influences on growth hormone and thyroid activity in white-tailed deer. Journal of Wildlife Management 45:140-147.

BALLARD, W. B., C. L. GARDNER, J. H. WESTLUND, AND S. M. MILLER. 1981. Use of mandible versus long-


bone to evaluate percent marrow fat in moose and caribou. Alces 17:147-164.

BROWN, R. D., E. C. HELLGREN, M. ABBOTr, D. C. RUTH- VEN, III, AND R. L. BINGHAM. 1995. Effects of dietary energy and protein restriction on nutritional indices of female white-tailed deer. Journal of Wildlife Man- agement 59:595-609.

BUBENIK, A. B. 1982. Physiology. Pages 125-179 inJ. W. Thomas and D. E. Toweill, editors. Elk of North America: ecology and management. Stackpole Books, Harrisburg, Pennsylvania, USA.

BULLOCK, K. D., J. K. BERTRAND, L. L. BENYSHEK, S. E. WILLIAMS, AND D. G. LUST. 1991. Comparison of real- time ultrasound and other live measures to carcass measures as predictors of beef cow energy stores. Journal of Animal Science 69:3908-3916.

CEDERLUND, G. N., R.J. BERSTROM, F. V. STALFELT, AND K. DANELL. 1986. Variability in mandible marrow fat in 3 moose populations in Sweden. Journal of Wildlife Management 50:719-726.

CONNOLLY, G. E. 1981. Assessing populations. Pages 287-345 in O. C. Wallmo, editor. Mule and black- tailed deer of North America. University of Nebraska Press, Lincoln, USA.

COOK, J. G. 2002. Nutrition and food habits. In D. E. Toweill and J. W. Thomas, editors. North American elk: ecology and management. A Wildlife Manage- ment Institute book. Smithsonian Institution Press, Washington, D.C., USA. In Press.

, L. L. IRWIN, L. D. BRYANT, R. A. RIGGS, D. A. HEN-

GEL, AND J. W. THOMAS. 1994. Studies of elk biology in northeast Oregon. 1993 Progress Report. Nation- al Council of the Paper Industry for Air and Stream Improvement, Corvallis, Oregon, USA.

AND J. W. THOMAS. 1998. Relations of forest cover and condition of elk: a test of the thermal cover hypothesis in summer and winter. Wildlife Monographs 141.

- , L. J. QUINLAN, L. L. IRWIN, L. D. BRYANT, R. A. RIGGS, AND J. W. THOMAS. 1996. Nutrition-growth relations of elk calves during late summer and fall. Journal of Wildlife Management 60:528-541.

COOK, R. C. 2000. Studies of body condition and reproductive physiology in Rocky Mountain elk. Thesis, University of Idaho, Moscow, USA.

-- , J. G. COOK, D. L. MURRAY, P. ZAGER, B. K. JOHN- SON, AND M. W. GRATSON. 2001. Nutritional condition models for elk: which are the most sensitive, accu- rate, and precise? Journal of Wildlife Management 65:988-997.

DAUPHINt, T. C., JR. 1975. Kidney weight fluctuations

affecting the kidney fat index in caribou. Journal of Wildlife Management 39:379-386.

DAvis, J. D., P. VALKENBURG, AND D.J. REED. 1987. Corre- lation and depletion patterns of marrow fat in caribou bones. Journal of Wildlife Management 51:365-371.

DELGIUDICE, G. D., F. J. SINGER, AND U. S. SEAL. 1991.

Physiological assessment of winter nutritional deprivation of elk of Yellowstone National Park. Journal of Wildlife Management 55:653-664.

DEPPERSCHMIDT, J. D., S. C. TORBIT, A. W. ALLDREDGE, AND R. D. DEBLINGER. 1987. Body condition indices for starved pronghorn. Journal of Wildlife Manage- ment 51:675-678.

DOMECQ, J. J., A. L. SKIDMORE, J. W. LLOYD, AND J. B. KANEENE. 1995. Validation of body condition scores

with ultrasound measurements of subcutaneous fat of dairy cows. Journal of Dairy Science 78:2308-2313.

EDMONSON, A.J., I.J. LEAN, L. D WEAVER, T. FARVER, AND G. WEBSTER. 1989. A body condition scoring chart for Holstein dairy cows. Journal of Dairy Science 72:68-78.

FARLEY, S. D., AND C. T. ROBBINS. 1994. Development of two methods to estimate body composition of bears. Canadian Journal of Zoology 72:220-226.

FINGER, S. E., L. BRISBIN, JR., AND M. H. SMITH. 1981. Kid-

ney fat as a predictor of body condition in white-tailed deer. Journal of Wildlife Management 45:964-968.

FRANZMANN, A. W. 1985. Assessment of nutritional status. Pages 239-260 in R. J. Hudson and R. G. White, editors. Bioenergetics of wild herbivores. CRC Press, Boca Raton, Florida, USA.

GANONG, W. F. 1979. Review of medical physiology. Ninth edition. LANGE Medical Publications, Los Altos, California, USA.

GARROTT, R. A., P. J. WHITE, D. B. VAGNONI, AND D. M. HEISEY. 1996. Purine derivatives in snow-urine as a dietary index for free-ranging elk. Journal of Wildlife Management 60:735-743.

GERHARDT, K. L., R. G. WHITE, R. D. CAMERON, AND D. E. RUSSELL. 1996. Estimating fat content of caribou from body condition scores. Journal of Wildlife Man- agement 60:713-718.

GLUCKMAN, P. D., B. H. BREIER, AND S. R. DAVIS. 1987.

Physiology of the somatotropic axis with particular reference to the ruminant. Journal of Dairy Science 70:442-466.

HALL, C. B., H. C. LUKASKI, AND M.J. MARCHELLO. 1989. Estimation of rat body composition using tetrapolar bioelectrical impedance analysis. Nutrition Reports International 39:627-633.

HAMR, J., AND G. A. BUBENIK. 1990. Seasonal thyroid hormone levels of free-ranging white-tailed deer (Odocoileus virginianus) in Ontario. CanadianJournal of Zoology 68:2174-2180.

HARDER, J. D., AND R. L. KIRKPATRICK. 1994. Physiologi- cal methods in wildlife research. Pages 275-306 in T. A. Bookhout, editor. Research and management techniques for wildlife and habitats. Fifth edition. The Wildlife Society, Bethesda, Maryland, USA.

HERRING, W. O., J. K. BERTRAND, AND L. L. BENYSHEK. 1995. Importance of real-time ultrasound use for live animal evaluation. Pages 102-103 in 1995 Depart- ment of animal and dairy science annual report. Uni- versity of Georgia College of Agricultural and Envi- ronmental Sciences, Athens, USA.

HILDERBRAND, G. V., S. D. FARLEY, AND C. T. ROBBINS. 1998. Predicting body condition of bears via two field methods. Journal of Wildlife Management 62:406-409.

JEFFRIES, B. C. 1961. Body condition scoring and its use in management. Tasmanian Journal of Agriculture 32:19-21.

KISTNER, T. P., C. E. TRAINER, AND N. A. HARTMANN. 1980. A field technique for evaluating physical condition of deer. Wildlife Society Bulletin 8:11-17.

LANKA, R., AND J. EMMERICH. 1996. Fat deposition measurements as an index to mule deer body condition and spring/summer habitat conditions. Wyoming Game and Fish Department, Cheyenne, USA.

MARQUEZ, M., AND B. E. COBLENTZ. 1987. Metatarsal and mandibular marrow fat in black-tailed deer. Journal of Wildlife Management 51:38-40.

McGUIRE, M. A., J. L. VICINI, D. E. BAUMAN, AND J. J.


VEENHUIZEN. 1992. Insulin-like growth factors and binding proteins in ruminants and their nutritional regulation. Journal of Animal Science 70:2901-2910.

MECH, L. D., AND G. D. DELGIUDICE. 1985. Limitations of the marrow-fat technique as an indicator of body condition. Wildlife Society Bulletin 13:204-206.

MENDENHALL, W., AND T. SINCIcH. 1989. A second course in business statistics: regression analysis. Third edition. Dellen Publishing, San Francisco, California, USA.

NEILAND, K. A. 1970. Weight of dried marrow as indicator of fat in caribou femurs. Journal of Wildlife Management 34:904-907.

NEWSHOLME, E. A., AND A. R. LEECH. 1983. Biochemistry for the medical sciences. John Wiley & Sons, New York, USA.

OKARMA, K. 1989. Mandible versus long bone marrow fat in red deer. Acta Theriologica 34:537-544.

PARKER, K. L., G. D. DELGIUDICE, AND M. P. GILLINGHAM. 1994. Do urinary nitrogen and cortisol ratios of creatinine reflect body-fat reserves in black-tailed deer? Canadian Journal of Zoology 71:1841-1848.

QUINLAN-MURPHY, L. J. 1998. Influences of age, condition, nutrition and season on serum and urine chemistry of Rocky Mountain elk. Thesis, Oregon State University, Corvallis, USA.

RATCLIFFE, P. R. 1980. Bone marrow fat as an indicator of condition in roe deer. Acta Theriologica 25:333-340.

RINEY, T. 1955. Evaluating condition of free ranging red deer (Cervus elaphus), with special reference to New Zealand. New Zealand Journal of Science and Tech- nology 36:429-463.

RINGBERG, T., E. JACOBSON, M. RYG, AND J. KROG. 1978. Seasonal changes in levels of growth hormone, somatomedin and thyroxine in free-ranging, semi- domesticated Norwegian reindeer (Rangifer tarandus tarandus [L]). Comparative Biochemistry and Physi- ology 60A:123-126.

ROBBINS, C. T. 1983. Wildlife feeding and nutrition. First edition. Academic Press, Orlando, Florida, USA.

. 1993. Wildlife feeding and nutrition. Second edition. Academic Press, San Diego, California, USA.

RYG, M., AND E. JACOBSEN. 1982. Seasonal changes in

growth rate, feed intake, growth hormone, and thyroid hormones in young male reindeer (Rangifer tarandus tarandus). Canadian Journal of Zoology 60:15-23.

SALTZ, D., AND G. C. WHITE. 1991. Urinary cortisol and urea nitrogen responses to winter stress in mule deer. Journal of Wildlife Management 55:1-16.

- , - , AND R. M. BARTMANN. 1995. Assessing animal condition, nutrition, and stress from urine in the snow: a critical review. Wildlife Society Bulletin 23:694-704.

SAS INSTITUTE. 1988. SAS/STAT? user's guide. Edi- tion 6.03. SAS Institute, Cary, North Carolina, USA.

SINCLAIR, D., AND P. DANGERFIELD. 1998. Human growth after birth. Sixth edition. Oxford University Press, New York, USA.

SNIDER, J. B. 1980. An evaluation of mandibular fat as an indicator of condition in moose. Proceedings of the North American Moose Conference Workshop 16:37-50.

STEPHENSON, T. R., K. J. HUNDERTMARK, C. C. SCHWARTZ, AND V. VAN BALLENBERGHE. 1998. Predicting body fat and body mass in moose with ultrasonography. Cana- dian Journal of Zoology 76:717-722.

SUTTIE,J. M., P. F. FENNESSY, I. D. CORSON, F.J. LAAS, S. F.

CROSBIE,J. H. BUTLER, AND P. D. GLUCKMAN. 1989. Pul- satile growth hormone, insulin-like growth factors and antler development in red deer (Cervus elaphus scoti- cus) stage. Journal of Endocrinology 121:351-360.

TANNER,J. M. 1968. Growth of bone, muscle and fat during childhood and adolescence. Pages 3-18 in G. A. Lodge and G. E. Lamming, editors. Growth and development of mammals. Butterworth, London, United Kingdom.

WATKINS, B. E.,J. H. WITHAN, D. E. ULLREY, D.J. WATKINS, AND J. M. JONES. 1991. Body composition and condi-

Appendix A. Single index predictor equations for ingesta-free body fat (kg) and gross energy (Mcal) for Rocky Mountain cow elka. Index acronyms are presented in Tables 1 and 2.

Ingesta-free body fat Ingesta-free gross energy x 80o , r2 P0o , r2

LIVINDEX -5.130 5.126 0.94 102.639 55.916 0.91 rLIVINDEX -4.424 4.947 0.93 111.949 53.592 0.89 MAXFAT 4.779 9.071 0.93 212.362 97.727 0.88 BCS -16.896 10.463 0.89 -30.135 115.517 0.89 Rump BCS -14.887 9.803 0.88 -2.722 106.612 0.85 BRI -6.510 0.037 0.89 79.047 0.422 0.93 BM -32.012 0.260 0.58 -243.890 3.121 0.69 KISTNER -6.458 0.369 0.85 92.130 3.963 0.80 KISThpk -13.041 0.844 0.83 17.262 9.184 0.81 ln(KFmass) -30.389 9.422 0.78 -159.509 100.111 0.72 In(KFIfuII) -21.933 9.115 0.65 -58.674 94.265 0.57 WYI -0.721 0.954 0.65 149.290 10.484 0.64 EVISWT -292.430 4.791 0.84

a Linear equation is of the form: y = flx + p,


tion evaluation of white-tailed deer fawns. Journal of Wildlife Management 55:39-51.

WEBSTER, J. R., I. D. CORSON, R. P. LITTLEJOHN, S. K. STU- ART, AND J. M. SUTTIE. 1996. Effects of season and nutrition on growth hormone and insulin-like growth factor-1 in male red deer. Endocrinology 137:689-704.

WILDMAN, E. E., G. M. JONES, D. E. WAGNER, R. L. BOMAN, H. F. TROUTT, JR., AND T. N. LESCH. 1982. A dairy cow BCS system and its relationship to selected production characteristics. Journal of Dairy Science 65:495-501.

WISDOM, M. J., J. G. COOK, M. W. ROWLAND, AND J. H.

NOYES. 1993. Protocols for care and handling of deer and elk at the Starkey Experimental Forest and Range. U.S. Forest Service General Technical Report PNW-GTR-311.

WRIGHT, I. A., AND A. J. F. RUSSELL. 1984. Partition of fat, body composition and body condition score in mature cows. Animal Production 38:23-32.

Received 1July 2000. Accepted 1 June 2001. Associate Editor: Krausman.

Appendix B. Multiple index models for ingesta-free body fat (kg), gross energy (Mcal), and water (kg) for Rocky Mountain cow elka. Index acronyms are presented in Tables 1 and 2.

y xl x2 Pi P2 P0 ra2

Fat (kg) WYI In(KFmass) 0.477 0.035 0.685 0.82 GE (Mcal) EVISWT KISThpk 2.895 4.895 -213.057 0.93 GE (Mcal) EVISWT KISTNER 2.089 2.919 -176.245 0.93 GE (Mcal) EVISWT In(KFmass) 3.299 47.804 -334.384 0.92 GE (Mcal) GIRTH KISThpk 4.690 6.585 -531.975 0.90 GE (Mcal) EVISWT WYI 3.665 4.228 -215.792 0.89 GE (Mcal) GIRTH In(KFmass) 5.583 69.121 -766.772 0.89 Water (kg) EVISWT 0.593 23.117 0.91 Water (kg) EVISWT In(KFmass) 0.719 -4.050 26.671 0.95 Water (kg) EVISWT KISThpk 0.751 -0.404 16.509 0.95 Water (kg) EVISWT KISTNER 0.748 -0.162 14.224 0.95 Water (kg) EVISWT WYI 0.714 -0.439 15.158 0.95 Water (kg) BM 0.415 23.758 0.86

a Multiple regression equation is of the form: y = po + f31x + 2X2 + fl3x3

Appendix C. Calculation of ingesta-free body mass, body water, body protein, and ash for cow elk.

Dependent variable Predictive equation1 Predictive equation2

IFBMa -677.305 + 106.632 x In(BM)b

Water% -0.759 x Fat% + 70.755 LBM (kg)c Water(kg)d + 0.717 LBM% Water% + 0.717 100 - Fat% Protein (kg) 0.3123 x Water(kg) LBM (kg) x 0.23 Protein% LBM% x 0.23 Ash% LBM% x 0.053 Ash (kg) LBM (kg) x 0.053

a Ingesta-free body mass. We caution that this equation was generated using captive elk that were fed hay and pellets. Thus, application may be limited for other settings.

b Live-animal body mass. C Fat-free, ingesta-free lean body mass. d See Appendix B for predictive equations.

development of predictive models of nutritional condition ... · rachel c. cook,1,2 department of...

Documents