HAL Id: hal-00929157https://hal.archives-ouvertes.fr/hal-00929157

Submitted on 1 Jan 1989

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Methods of measurement of milk yield and compositionin nursing mares: a review

M. Doreau, S. Boulot

To cite this version:M. Doreau, S. Boulot. Methods of measurement of milk yield and composition in nursing mares: areview. Le Lait, INRA Editions, 1989, 69 (3), pp.159-171. �hal-00929157�

Lait (1989) 69,159-171© Elsevier/INRA

159

Review article

Methods of measurement of milk yield and compositionin nursing mares: a review

M. Doreau and S. Boulot

INRA, Laboratoire de la Lactation, Theix, 63122 Ceyrat, France

(received 17 March 1988, accepted 23 Decernber 1988)

Summary - This review describes and analyses the most commonly used methods of estimatingmilk yield of nursing mares : milking without or with oxytocin injection, weighing the foal before andafter suckling, dilution of labelled water in foal body water : milking only does not appear toconstitute an accurate method. Weighing before and after suckling requires numerousmeasurements, due to the foal's frequent sucks. Foal body water dilution does not disturb themother-offspring relationships, but can only be used when -milkis the sole water source for the foal.

Conditions of sampling and analysis for milk composition in nursing mares are described;incomplete milking (without oxytocin injection) involves a bias regarding milk fat content becausemilking is unusual. An oxytocin injection allows evaluation of the composition of secreted milk. Theperiod of mare-foal separation should be short and close to the natural interval between sucks.Milk analysis should take into aceount the composition of mare milk fat and protein; infraredspectrophotometry should be calibrated for mare's milk. .

Some of the differences in the Iiterature on milk secretion data can probably be explained by themethods of measurement of milk yield and composition.

mare - milk yield - milk composition - methods of measurement

Résumé - Méthodes de mesure de la production et de la composition du lait chez la jumentallaltante : revue bibliographique. Dans une première partie, cette revue analyse les principalesméthodes d'estimation de la production laitière des juments allaitantes : traite sans ou avecinjection d'ocytocine, pesée du poulain avant et après tétée, dilution d'eau marquée dans l'eaucorporelle du poulain. Seule la traite semble être une méthode imprécise. La pesée avant et aprèstétées nécessite des mesures nombreuses, en raison de la fréquence élevée des tétées dupoulain. La dilution de l'eau corporelle du poulain ne perturbe pas le comportement normal mère-jeune, mais cette technique ne peut être utilisée que lorsque le lait est la seule source d'eau pour lepoulain.

Dans une deuxième partie, les conditions d'échantillonnage et d'analyse de la composition dulait des juments allaitantes sont décrites. Une traite incomplète (sans injection d'ocytocine) omretneun biais sur la détermination du taux butyreux, les juments n'étant pas habituées à la traite.L'injection d'ocytocine permet d'estimer la composition du lait sécrété. La séparation entre la jumentet le poulain doit être de courte durée et proche de l'intervalle naturel entre tétées. L'analyse du laitdoit prendre en compte la composition des matières grasses et des protéines du lait de jument, desorte que les méthodes de spectrophotométrie dans /'infrarouge doivent être calibrées pour le laitde jument. .

more recent method is that of diluting amarker in foal body water. Because of thelack of comparisons between thesemethods in the mare (except for Doreauet al., 1986a), their respective advantagesand shortcomings will be analyzed withreference to observations made on otherspecies, taking into account thephysiological specificities of equids.

Several authors have estimated maremilk from foal growth, more often than notusing relationships established in cattle(Leonhard, 1956). This method will not bedescribed in the present paper as it is tooindirect and provides only roughestimates, as shown in studies in whichmare milk production, measured moreaccuratelv, was compared to foal growth(Bouwman and Van Der Schee, 1978;Doreau et al., 1986b). In particular, therelationship depends on the extent ofmaintenance requirements, thus on foalage.

160 M. DoreauandS. Boulot

En conclu~ion, un~ p.artie des différences de niveau de production ou de composition entreauteurs peut etre expltqueepar le choix des méthodes de mesure.

jument - production laitière - composition du lait - méthodes de mesure

Introduction

The optimal feeding of mares involves theknowledge of their requirements,especially for milk production. These areeasy to measure in dairy females, butmust be estimated in nursing mares,which constitute the main proportion ofherds in most countries, with theexception of the USSR and Mongolia.The main methodological problems arethe determination of the effective yield ofmilk consumed by the foal, and therepresentative sampling of milk. Aliterature survey shows high variations inmilk yield (2-3.5 kg per 100 kg liveweight; Doreau and Boulot, 1989), and inmilk composition (for example, extremereported values of fat content are 0 and7.9%; Neuhaus, 1959). A proportion ofthese high and often unexplainedvariations in mare milk yield and/orcomposition from one trial te another maybe a result of the various measurementtechniques. The purpose of the presentpaper is to critically evaluate the methodsof estimating milk production in nursingmares, in order to assess the reliability ofthe different experimental data,summarized in the reviews of Neseni etal. (1958), Neuhaus (1959) and Doreauand Boulot (1989).

Methods of estimating milk yield

Three types of method are used. The 2most common are milking and foalweighing before and after suckling. A

Mare milking (Table 1)Mare milking is the oldest method ofestimating milk yield in nursing mares. ltwas used as early as 1896 by Camererand Soldner, quoted by Neuhaus (1959).These authors had obtained only lowyields : 4.5 1 as a total of 5 milkings. Thisis not surprising because occasionalmilking of normally nursing females leadsto an underestimation of milk yield, due toincomplete udder emptying. Such datahave beenshown in ewes (Ricordeau andDenamur, 1962) and cows (Le Neindre etei., 1976). The quantity of mare cisternalmilk is low when compared to alveolarmilk : less than half of the total yield

Measurement of milk yield and composition in nursing mares 161

Table 1.Amount of milk obtained in one milking.

.References Experimental conditions

DairymaresAlaguzhin (1964)Cherepanovaand Belokobylenko(1974)

Fedotovand Duisembaev(1974)

Cherepanova (1979)Tarasevich (1979)

Nursing maresBilek et al. (1955)

Neuhaus (1961)Ullrey et al. (1966)

Method of mi/king Amount of mi/k(ml)

1 000 - 2 000

mechanical 375 - 2 000manual 275 - 2 700

mechanical 200 - 1 650

mechanical < 2 200mechanical 600 - 1 800

mechanical 2 300 - 3 000

mechanical 300 - 1 520

manual with oxytocin < 2 000manual 152 - 324

Storch (1969) mechanical 2 000 - 2 500Johnston et al. (1970) manual 220 - 340

Oftedal et al. (1983) manual with oxytocin 000 ± 354

Doreau et al. (1986a) manual with oxytocin 265 - 1 480

(Tumanova, 1966; Dyusembin, 1966; LeDu, 1986). This concept of «cisternalmilk» should be improved in mares sinceBarone and Monnet (1955), by anatomicalobservations, assessed a 60-ml contentfor cisterns of dry mares. In mares, nodifferences in oxytocin release have beenshown between suckling and milking, as

450 - 1 760

405 - 1 405

810 - 1 650

925 - 2 185

at peak of lactation (34 mares)

extreme valuesextreme values

means : comparisonof 3 breeds (47 mares)data on 20 herdsdaily yield : 13--14 kg(20 mares)daily yield : 17-18 kg

4th month of lactation(6 mares)

4 first months of lactation(10 mares)

15 first days of lactation(6 mares)30 min secretion for last suck2 first months of lactation(5 mares)1-2 h secretion for last suck2 days of lactation (10 mares)30 min secretion for last suck9 days of lactation (10 mares)40 min secretion for last suck16 days of lactation (10 mares)50 min secretion for last suck1 month of lactation (10 mares)60 min secretion for last suck2 months of lactation (10 mares)90 min secretion for last suck

indirectly assessed by comparable valuesof blood concentration peaks (Dyusembin,1972a; Sharma, 1974; Kuanyshbekova,1975). However, the measurements werenot carried out on the same breeds : dairybreeds for milked mares, saddle breed fornursing mares. The stimulus providedmay not be as great in occasional

162 M. DoreauandS. Boulot

milkings carried out with mares that areusually nursing. Moreover, the differencesin experimental conditions and methodsof oxytocin analysis do not allow theconclusion to be made that the physio-logical efficiency of milk ejection is thesame for milking and suckling. Whetherthe foal is present or not during milkingalso appears important; Cherepanovaand Belokobylenko (1974) noted that 20%of milked mares withheld their milk whentheir foal was at a distance fram them.

The presence of the foal is notsufficient. Ashcraft and Tyznik (1976)milked one side of the udder while thefoal suckled the other. They obtained lowdaily yields (minimum: 1.5 kg). As milkingtakes longer than suckling, the rapiddecrease in blood oxytocin may not haveallowed the complete emptying of acini atthe end of milking. Neuhaus (1961)observed that when one side was entirelymilked before the other, 35% less milkwas obtained from the second side thanfrom the first side.

Conditioning mares for milking isefficient (Masloboev, 1961; Dyusembin,1968). Several factors which modifyudder emptying remain unknown. Doreauet al. (1986a) observed that when milking10 mares at 5 physiological stages, thevery high variations in milk obtained.during milking when the foal was presentcannat be explained by lactation stage orby the individual effect, as might havebeen assumed. We were not able torelate this quantity ta the apparent calmor stress of the mare during milking.

Many authors since Neuhaus (1961)and Linzell (1972) have recommendedthe use of exogenous oxytocin injections.The problems associated with thismethod have been widely described whenapplied ta bovines or ovines byBoyazoglu (1963), Le Neindre (1973) andDoney et al. (1979). The quantity of milk

consumed by the offspring is generallyoverestimated by taking into account theresidual milk which is high just afterfoaling. The same may be true for equids.Ta our knowledge, the choice of thequantity to be injected has not beenspecifically studied in the mare; 10-25 lUare often injected in cows (cf., Thompsonet al., 1973). Although high amounts donot seem necessary ta these authors, thequantity injected in mares may depend onthe specificity of this species, in which thelack of references is evident, and on themarked individual variations in reactionsto oxytocin injection, as found in ewes(Labussière et al., 1969). In equids,milkings with oxytocin were carried out byNeuhaus (1961) in mares (20 lU,intravenously), Linzell et al. (1972) inponies (0.5 lU, intravenously), Oftedal etal. (1983) and Boulot (1987) in 500-kgmares (20 lU respectively, intramuscularlyand intravenously) and Doreau et al.(1'986a) in 800-kg mares (40 lU,intravenously). These concentrations arehigh and probably non-physiological. Thiscan compensate the higher duration ofmanual milking whsn : compared tomechanical milking or suckling (Le Du,1986). None of the authors performedmilking with oxytocin . when milk yieldassessment was the main purpose. Theobjective was ta obtain maximal milk yieldfor human utilization, to compare withnutrient flow towards the udder, or toobtain a representative sampie foranalysis. The milking routine differedaccording the authors. Oftedal et al.(1983) began milking after oxytocininjection; in this condition an incompleteemptying can occur. Linzell et al. (1972)emptied the udder using oxytocin, thenallowed milk secretion, then emptied theudder. The milk obtained represents thesecretion capacity during a given interval,but slightly differs from the milk sucked bythe foal, which does not completely empty

Measurement of milk yield and composition in nursing mares 163

the udder. Doreau et al. (1986a) usedanother technique : complete emptying bythe foal, then milk secretion during a giveninterval, then milking with oxytocin. Thismethod also slightly overestimates themilk really sucked. The reliability of themethod with oxytocin injection is indicatedwhen it is compared with a body watermarker used as a reference (Doreau etal., 1986a). With 30 measurements, themeans were the same for bath methods,but milking with oxytocin yielded moredispersed and sometimes aberrant datafor the same physiological stage. Onemilking appears insufficient ta estimatedaily milk yield.

Another method, which is probablymore accurate, was used by Nordstrarn(1970). Mares were milked using oxytocinevery hour over a period of 24 h. Thereliability of the method depends on apossible galactopoietic effect of oxytocinwhen repeated injections are made(Dyusembin, 1972a). Further research onthis technique would be of interest.

Weighing before and after sucklingThis technique has been performed veryfrequently in equids since the studiesperformed by Blechschmidt (1933). Thefoal is removed from the mare's box for aperiod of 45 min (Bouwman and Van DerSchee, 1978) ta 8 h (Lukas et al., 1972).The foal is then weighed, allowed ta suck,then weighed again. The differencebetween the 2 weighings gives anestimate of the milk intake. According tathe authors, these measurements arerepeated over a 24- ta 48-h period, or onlyover a periad of 8-12 consecutive hours.Measurements are either diurnal or takenbath day and night. This method hasseveral disadvantages. Ta begin with, itdisturbs the mother-offspring relation-ships and natural nursing behaviour.

When suckings very frequent,immediately atter delivery, measurements

are at wider intervals than naturalsuckings. It is not certain that the youngfoal, which consumes small amounts ofmilk at each sucking, empties the udderwhen it is full. Moreover, if successivesuckings are tao spaced out, the udder'ssecretory ability may decrease during theinterval between suckings. The low uddercapacity, seldom more than 2 1 (Table 1)and its high filling rate, between 1.5-2 h(Alaguzhin, 1964), may result in anunderestimation of milk intake. In arder taimprove the precision of the method at theonset of lactation, Bouwman and Van DerSchee (1978) chose an interval betweenbouts of sucking that takes into accountbehavioural data, and varies between 45min at 1 month and 2 h at 4 months.However, measurements of daily suckingbouts show higher values : 70 atter birth,35 at 1 month and "" 25 at 4 months(Barmincev, 1960; Tyler, 1972; Bouwmanand Van Der Schee, 1978; Martin-Rossetet al., 1978). This method is probably lessreliable than in other species, such asbovines, when suckling frequency is low.ln mares, the high number of weigh-suck-weigh sequences increases therisks of error.

Other problems tend ta decrease thereliability of this method. If measurementsoccur only during daylight, the result maybe biased because foals suck longerduring the day than at night (Duncan etal., 1984). Nevertheless, nocturnal experi-mentation may disturb animal behaviour.Some analytical inaccuracy persists. It isdifficult ta weigh foals without error,particularly atter they are 2 months oldbecause of the limited accuracy of thescales and foal stress. Estimates of theweights of urine and/or feces have ta bemade if defecation or urination occursduring suckling. The later the lactationstage, the better this method proves tabe.

knowledge, this method and weigh-suck-weigh method have not beencompared in equines. However, using amarker can prove to be usefulimmediately after foaling, when suckfrequency is highest, and when theweigh-suck-weigh method is leastfeasible. Moreover, this method does notdisturb maternai nursing and has none ofthe analytical complications of theweigh-suck-weigh technique. Its maindisadvantage is that it cannot be appliedwhen milk is not the only source of water,i.e. after 2 months. In this case, 2 markershave to be used, one in the mother, theother in the offspring (Holleman et al.,1975; Wright and Wolff, 1976). Becauseof its complexity, this method has not yetbeen used in equids.

A completely different method ofmarking milk in the udder has been usedin suckled cows by Nicol and Irvine(1973). One isotope of iodine is given in aknown amount to the offspring; the otheris introduced in a known amount in theudder. The ratio of the 2 isotopes in thecalf blood allows milk intake during onesuck to be calculated. This method can beused at pasture. However, this method isIimited because it disturbs the mother-offspring relationship. Extrapolation of thernilk intake during a suck to daily intake isdifficult, especially as suck frequency ishigh.

164 M. Doreau and S. Boulot

Use of markers

When the offspring receives only milk,intake may be estimated from the waterturnover rate and body volume. A bodywater marker has to be infused intrave-nously into the offspring (deuterium oxideor tritiated water). The decrease in itsconcentration is analyzed. Water intakefrom milk and water excretion (urine,feces, sweat, exhaled steam) would haveto be continuous. If, during the few dayswhen measurements are taken, milkintake and body water volume areconsidered constant, the Iollowinq'relationship can be set up : ..C = ON e -k/v t (C : marker concentration;t : time; a : quantity of marker injected;V : body water volume; k : water intakeper unit of time).

This equation allows water intake bythe offspring to be calculated. If milk drymatter is known, and if metabolic waterproduction from nutrient oxidation isestimated, milk intake can be calculated.This method was first usedby Macfarlaneet al. (1969) in lambs. Improvementswere made by Dove and Freer (1979) inlambs and Doreau and Dussap (1980) infoals by taking into account the Iive-weight gain of the offspring duringmeasurements, and therefore, theincreased body water volume. With this

.modification the method, which has beentested with bottle-fed lambs by these 2authors, was very reliable. Measurementerror was under 5% and there was nosystematic bias. This method has beenused for several years to estimate maremilk production (Doreau and Dussap;1980; Doreau et al., 1982, Oftedal et al.,1983; Doreau et al., 1986a; Boulot, 1987),as in other species (Oftedal, 1984).Although no standardization has been setup in equids, this method is probably asprecise in equines as in ovines becausethe foal sucks frequently. To our

Methods of estimating milk compo-sition

As early as 1937, Linton indicated thepossible bias in milk composition data dueto a bad sampling. Differences inanalytical methods may also occur. Thischapter will successively deal with these 2problems.

Measurement of milk yield and composition in nursing mares 165

Sampling methods of Doreau et al. (1986a) show that theincrease in milk fat during milkingbecomes higher after an oxytocininjection. The possible effect of oxytocinon milk fat has also been reported bySmolders et al. (1989). According toLabussière (1984), the oxytocin release,through compression of acini, should cutthe fat globule aggregates and allow.theircrossing towards the cistern. So, 'everymilking process which involves aninsufficient oxytocin release will give a lowmilk volume of low-fat milk. lt is better toobtain the whole content of one side or, if

As in other species, mare milkcomposition is not constant duringmilking. The only constituent in whichconcentration varies in a significant way isfat. In some individual cases, milk fatcontent at the onset of milking may bebelow 1 g/kg (Doreau et al., 1986a). Thismight explain some zero milk fat contentswhich have been reported in the Iiterature(Neuhaus, 1959). Fat content may be 2 or3 times higher at the end of the milkingthan at the beginning (Table Il). The data

Table Il. Variations in mare milk fat content (%) throughout milking.

Authors No. of Beginning of End of Meanobservations mi/king mi/king

Prowosnik (1931) 0.2 0.85 0.23(N) 0.1 2.23 1.87

0.4 1.45 1.13

Dittrich (1938) 0.1 1.18(N) 1.83 0.76

0.29 5.550.19 2.15

Dyusembin (1972b) 3 0.45 0.82 0.63(0)

Cisternal Alveolar Stripping Residualmilk mi/k mi/k mi/k

Cherepanova andBelokobylenko (1974)(0) 0.15 0.~1.1 1.1-1.4 3.~7.2

Beginning of Last fraction First fraction End ofmi/king before oxytocin after oxytocin mi/king

Doreau et al. (1986a) 50 0.98 1.05 1.29 1.78(N)

N : nursing mares. D : clairy mares.

166

and Labussière and Coindet, 1968, incows). The most accurate solution isprobably to calculate the time ofseparation using the natural intervalbetween 2 suckings. An attempt wasmade by Doreau et al. (1986a) whoapplied the intervals previously found byMartin-Rosset et al. (1978). The longseparation times between mare and foalin the experiments of Gibbs et al. (1982)and Smoczynski and Tomczynski (1982)(respectively 3 and 6 hours) may haveinfluenced milk composition.

M. DoreauandS. Boulot

possible, the two sides of the udder so asto take into account a possible differencein milk composition between the twosides. Prowosnik (1931), Dittrich (1938)and Doreau et al. (1986a) observed at amean no difference in milk compositionbetween the two sides, but with individualdata, milk fat was higher on one sidewhen the obtained milk volume was lowerthan on the other side.

Protein and lactose contents havebeen found to be slightly and non-significantly lower at the end of milkingthan at the beginning (respectively 1.2and 0.5 g/kg; Doreau et al., 1986a). Thisdifference, already observed in otherspecies (Labussière, 1984), can be in part(0.2 glkg) explained by a passivevariation caused by the increase in fatcontent. A different result (increase inprotein and no variation in lactosecontent) was found by Dyusembin(1972b) using a small number of mares.

The use of oxytocin injection, whichpromotes udder emptying, may have ashort-term effect on milk composition.According to Wheelock et al. (1965),when a cow udder is emptied after anoxytocin injection, sodium, chloride andwhey protein rise and lactose falls in thefollowing milking. When several mare milksamples are not obtained successively,oxytocin injection probably has no effecton milk composition. In most cases,literature data come from spot samplings.

ln order to obtain the amount of milkcorresponding exactly to a suckling of theoffspring, numerous authors separatemare and foal before milking (Table III).This time of separation must not beoverestimated because milk stasis in theudder is followed by an increase in fat,protein and minerais and a decrease inlactose content, either for the current orfor the subsequent milking (Petersen andRigor, quoted by Unton, 1937 in mares

Methods of milk analysisFat. Most authors use the Rose-Gottliebreference method or a derived method(Ullrey et al., 1966; Johnston et al., 1970;Oftedal et al., 1983; Pagan and Hintz,1986) or the volumetric method of Gerber(Unton, 1937; Kulisa, 1977; Smoczynskiand Tomczynski, 1982). In these twomethods, the result does not depend onmilk fat composition. Rapid methods suchas turbidimetric analysis (Gibbs et al.,1982) or infrared spectrophotometry(Zimmerman, 1985) should be calibratedfor each animal species. Indeed, the fatrefractive index depends on each species(Grappin and Jeunet, 1979). Conditionsfor a good calibration were defined bySmolders et al. (1989). In mother's milk,an excellent correlation between infraredspectrophotometry and Rose-Gottliebmethod was found by Hundrieser et al.(1984). Dorea et al. (1981), however,found low relationships in mother's milk.The only comparison in mares (Doreau etal., 1985) showed close relationshipsbetween Rose-Gottlieb, Gerber, turbidi-metric and spectrophotometric methods,but resulted in the need for a specificspectrophotometric calibration in mare'smilk.Protein. Most authors have analyzedcrude protein in mare milk by the Kjeldahl

Measurement of milk yield and composition in nursing mares 167

Table III. Methods of mare milk sampling for composition analysis; outline of the Iiterature.

Authors Methods of sampling Amounts of milk obtained(ml)

Linton (1937)

Neseni et al. (1958)

Ullrey et al. (1966)

Johnston et al. (1970)Zimmerman (1985)

Linzell et al. (1972)

Mare--foal separation : 2 hManual milking of the 2 sides

1 sampling by manual milking beforeand after foal suck, bath sides

No mare--foal separation. Manual milkingof one side while the other is sucked 76-162

Mare--foal separation: 30 minManual milking of one side whilethe other is sucked

110-170(Johnston et al.)

Manual milking of one side while theother is sucked, every hour, with anoxytocin injection

Ashcraft and Tyznik (1976) Mare--foal separation: 1 h. Manualmilking of one side while the otheris sucked

Bouwman andVan Der Schee (1978)

Gibbs et al. (1982)

Smoczynski andTomczynski (1982)

Oftedal et al. (1983)

Miraglia et al. (1986)

Doreau et al. (1986a)Boulot (1987)

Mare--foal separation: 30-45 minaccording to lactation stage. Manual milkingof one side while the other is sucked

200

Mare--foal separation : 3 hManual milking of one side whilethe other is sucked

Mare--foal separation : 6 hManual milking while foal is present >300

Mare--foal separation : 1-2 hManual milking of one side while theother is sucked, with an oxytocin injection

SOO± 177

Mare--foal separation : 1 hManual milking of one side whilethe other is sucked

Mare--foal separation : 30-90 min accordingto lactation stage. Manual mil king of thebath sides while foal is present, with anoxytocin injection

964±439

168

protein and ash. This method evidentlyinvolves error and uncertainties regardingthese components.

M. DoreauandS. Boulot

method. Dye-binding protein method(Gibbs et al., 1982) and infrared spectro-photometry (Zimmerman, 1985) haveseldom been used. Regarding fat, theutilization of infrared spectophotometryrequires a specific calibration for maremilk (Doreau et al., 1985). According tothese authors, a dye-binding proteinmethod (amido black calibrated on theKjeldahl method with cow milk) gavehigher results than Kjeldahl (29.6 vs 25.8g/kg) although the milk studied contained7-13% non-protein nitrogen. Thisdifference can be explained by the highercapacity of indicator binding to wheyprotein (Grappin and Jeunet, 1979) whichform '" 50% of mare milk, or to thespecificity of some proteins in mare milk.

Lactose. Lactose is analyzed by verydifferent techniques : Bertrand method(reduction then titration : Kulisa, 1977;Smoczynski and Tomczynski, 1982),phenol-sulphuric acid colorimetriemethod (Oftedal et al., 1983), enzymaticmethod (Pagan and Hintz, 1986) orinfrared spectrophotometry (Zimmerman,1985; Doreau et al., 1986a). With thislatter technique, a specific calibration formare milk does not seem necessary, aslactose is the same chemical entity for ailspecies. The most accurate method isprobably the gas-chromatographictechnique (Kulisa, 1980) which analysesboth lactose and other sugars. Thesedifferent methods were not compared formare milk; it might be that some of thedifferences in lactose content observedfrom one trial to another is explained bythe particular analysis technique utilized.Another source of difference is theexpression of lactose concentration, asanhydrous or monohydrate, which is notalways noticed in the Iiterature.

Numerous authors have given lactoseconcentrations by estimating thedifference between dry matter and fat,

Conclusion

The complexity of the methods forestimating mare milk yield has probablybeen a reason for the small number ofexperiments. It is possible that recentmethod improvements may contributetowards promoting these studies. Despitethe lack of direct comparisons betweenmethods, it is considered that the use ofmarkers is the most accurate method inearly lactation. The weigh-suck-weightechnique, although sometimes lacking inprecision, can be applied in mid-lactation.

Measurements of mare milk compo-sition are more numerous. However,some methodological problems noted inthis review lead us to question thereliability of some results. The mainproblem remains the estimation of fatcontent, which requires very carefulsampling.

References

AlaguzhinA. (1964) Milk productivityof Lokaimares. Konevodstvo Konnyi Sport 34, 13-14(inRussian)AshcraftA. & TyznikW.J. (1976) Effectof dieton volumeand compositionin mare'smilk.J.Anim. Sei.43,248 (abstr.)BarmincevJ.N. (1960) Breedinghorsesfor milkunderconditionsin Kazakhstan.Trudy I-i Kont.Mol. Konevod. Kumysod. Mosk. 20-25 (Anim.Breed. Abstr. (1961) 29, 397) (in Russian)BaroneR. & MonnetC. (1955) Les mamellesdes équidésdomestiques.Bull. Soc. Sei. Vét.Lyon 57,73-92

Bilek J., DusekJ. & LechnerF. (1955) Newfindingson methodsof obtainingmares'milk

Measurement of milk yield and composition in nursing mares 169

and the influence on its secretion. Sborn. Csi.Akad. Zemed. Ved. 28, 859-886 (in Czech)

Blechschmidt H. (1933) Experimentalunter-suchungen und Ermittlungen über dieErnâhrunq des Pferdes auf der Weide. Landw.Jb. 77, 463-560Boulot S. (1987) L'ingestion chez la jument.Etude de quelques facteurs de variation aucours du cycle gestation-lactation. Implicationsnutritionnelles et métaboliques. Thèse Doct.Ing., ENSA Rennes, Univ. Rennes 1Bouwman H. & Van Der Schee W. (1978)Composition and production of milk from Dutchwarmblooded saddle horse mares. Z.Tierphysiol. TIerernahrg. Futtermittelk. 40, 39-53

Boyazoglu J.G. (1963) Aspects quantitatifs dela production laitière des brebis. Ann. Zooteeh.12,237-296

Cherepanova V. (1979) Dairy horses farming inKazakhstan. Konevodstvo Konnyi Sport 8, 5-6(in Russian)

Cherepanova V.P. & Belokobylenko V.T. (1974)Milk ejection characteristics in machine milkingof mares. Proc. 3rd Ali-Union Symp. Physiol.Principles Machine milking, Borovsk (DairySei. Abstr. (1975) 37, 691) (in Russian)

Dittrich H. (1938) Untersuchungen über dieZusammensetzung von Pferdemilch. Mileh-wirtsehaftl. Forseh. 19, 406-412Doney J.M., Peart J.N., Smith W.F. & Louda F.(1979) A consideration of techniques forestimating of milk yield by suckled sheep and acomparison of estimates obtained by twomethods in relation to the effect of breed, levelof production and stage of lactation. J. Agde.Sei. Cambo 92, 123-132Dorea J.G., Horner M.R., Bezerra V.L. & BorgoL.A. (1981) Comparison of two methods fordetermining protein and Iipid in humancolostrum and mature milk. Nutr. Rep. Int. 24,985-991

Doreau M. & Boulot S. (1989) Recentknowledge on mare milk production : a review.Livest. Proe. Sei. 21, in pressDoreau M., Boulot S., Jeunet R. & Trin J.M.(1985) Comparaison de différentes méthodesde dosage des matières grasses et desmatières azotées du lait de jument. Lait 65,149-161

Doreau M., Boulot 5., Martin-Rosset W. &Dubroeucq H. (1986a) Milking lactating maresusing oxytocin : milk volume and composition.Reprod. Nutr. Dev. 26,1-11

Doreau M., Boulot S., Martin-Rosset W. &Robelin J. (1986b) Relationship betweennutrient intake, growth and body composition ofthe nursing foal. Reprod. Nutr. Dev. 26, 683-690

Doreau M. & Dussap G. (1980) Estimation dela production laitière de la jument allaitante parmarquage de l'eau corporelle du poulain.Reprod. Nutr. Dev. 20, 1883-1892

Doreau M., Martin-Rosset W. & Dubroeucq H.(1982) Production laitière de la jument. Liaisonavec la croissance du poulain. C.R. 8e Journéed'Etude du CEREOPA, CEREOPA, Paris,pp. 88-100

Dove M. & Freer M. (1979) The accuracy oftritiated water turnover rate as an estimate ofmilk intake in lambs. Aust. J. Agrie. Res. 30,725-739

Duncan P., Harvey P.H. & Wells S.M. (1984)On lactation and associated behaviour innatural herd of horses. Anim. Behav. 32, 255-263Dyusembin K. (1966) Milk-ejection reflex inmares milked with the modified milkingmachine «Ternp». Trudy Inst. Fiziol. Alma-Ata10,148-152 (in Russian)

Dyusembin K. (1968) Specifie traits of the milkejection reflex in mares. Trudy Inst. Fiziol.Alma-Ata 12,155-159 (in Russian)

Dyusembin K. (1972a) The role of oxytocin inregulating lactation in mares. Trudy Inst. Fiziol.Alma-Ata 17,37-41 (in Russian)

Dyusembin K. (1972b) Uneven distribution ofmajor milk components in consecutive portionsof a milking in farm animais. Trudy Inst. Fiziol.Alma-Ata 17,42-45 (in Russian)

Fedotov PA & Duisembaev K. (1974)Morphological characteristics of mare's udders.Genet. Selekt. Rast. Zhiv. Kazakh. 265-266 (inRussian) (Dairy Sei. Abstr. (1977) 39, 250)Gibbs P.D., Potter G.D., Blake R.W. & McMullan W.C. (1982) Milk production of quarterhorse mares during 150 days of lactation. J.Anim. Sei. 54, 496-499

Grappin R. & Jeunet R. (1979) Méthodes deroutine pour le dosage de la matière grasse etdes protéines du lait de chèvre. Lait 59, 345-360

Holleman D.F., White R.G. & Luick J.R. (1975)New isotope methods for estimating milk intakeand yield. J. Dairy Sei. 58, 1814-1821

Hundrieser K.E., Clark R.M., Jensen R.G. &Ferris A.M. (1984) A comparison of methods

170

Lukas V.K., Albert W.W., Owens EN. & PetersA (1972) Lactation of Shetland mares. J.Anim. Sci. 34, 350 (abstr.)

Macfarlane W.V., Howard B. & Siebert B.D.(1969) Tritiated water in the measurement ofmilk intake and tissue growth of ruminants inthe field. Nature 221, 578-579

Martin-Rosset W., Doreau M. & Cloix J. (1978)Etude des activités d'un troupeau depoulinières de trait et de leurs poulains aupâturage. Ann. Zootech. 27, 33-45

Masloboev AJ. (1961) Characteristics oflactation in mares. Trudy Vses. /nst.Konevodstva. 23, 135-150 (in Russian) (Nutr.Abstr. Rev. (1962) 32, 574-575)

Miraglia N., Quintavalla F., Mariani P. &Catalano AL. (1986) Plasma biochemistrychanges in mares and foals in relation to nutri-tional aspects. Preliminary trials. Proc. 37thAnn. Meet. Europ. Ass. Anim. Prod., pp. 11

Neseni R., Flade E., Heidler G. & Steger H.(1958) Milchleistung und Milchzusammenset-zung von Stuten im Verlaufe der Laktation.Arch. Tierzucht. 1, 91-129

Neuhaus U. (1959) Milch und Milchgewinnungvon Pferdestuten. Z. Tierzucht. 73, 370-392

Neuhaus U. (1961) Obtaining mare's milk forthe feeding of human infants. Vlllth Int. Congr.Anim. Prod. (Hamburg), Il, 184-185

Nicol AM. & Irvine C.H.G. (1973) Measurementof the milk consumption of suckling beef calvesby an isotope dilution method. Proc. N.Z. Soc.Anim. Prad. 33, 176-183

Nordsfrem A. (1970) On the growth of younghorses and on factors affecting the growth.Mast. Thesis, Univ. Helsinki (in Finnish)

Oftedal O.T. (1984) Milk composition, milk yieldand energy ouptput at peak lactation : acomparative review. Symp. Zoo/. Soc. Lond.51,33-85

Oftedal O.T., Hintz H.E & Schryver H.F. (1983)Lactation in the horse : milk composition andintake by foals. J. Nutr. 113, 2196-2206

Pagan J.O. & Hintz H.F. (1986) Composition ofmilk from pony mares fed various levels ofdigestible energy. Comell Veto76, 139-148

Prowosnik L. (1931) Neue Untersuchungenüber die Zusammensetzung der Pferdemilch.Z. Züchtgskde 6, 329-338

Ricordeau G. & Denamur R. (1962) Productionlaitière des brebis Préalpes du sud pendant les

M. Doreau and S. Boulot

for determination of total Iipids in human milk.Nutr. Res. 4, 21-26

Johnston R.H., Kamstra L.D. & Kohler P.H.(1970) Mare's milk composition as related tofoal heat scours. J. Anim. Sei. 31, 549-553

Kuanyshbekova GA (1975) Oxytocic activityof blood in lactating mares. Trudy /nst. Fiziol.A/ma-Ata Kazakhs. SSR 21, 128-131 (inRussian)

Kulisa M. (1977) The composition of mare milkin three horse breeds with reference to N-acetylneuraminic acid. Acta Agrar. Sy/v. Ser.Zooteeh. 17, 25-37 (in Polish)

Kulisa M. (1980) Lactose, free glucose andgalactose levels in the milk of Arab mares.Roez. Nauk. Zooteeh. 7, 31-36 (in Polish)

Labussière J. (1984) Influence des techniquesde traite sur la composition du lait de quelquesespèces domestiques. C.R. Symp. Lacomposition chimique du lait et ses incidencestechnologiques, INRA Rennes, pp. 20

Labussière J. & Coindet J. (1968) Effets de lasuppression de la traite du dimanche soir, chezles bovins de race française frisonne pie noire.Ann. Zootech. 17, 231-244

Labussière J., Martinet J. & Denamur R. (1969)The influence of milk ejection reflex on the flowrate during milking of ewes. J. Dairy Res. 36,191-201

Le Du J. (1986) Mechanical milking of mares.Proc. 37th Ann. Meet. EAAP, Budapest, pp. 12

Le Neindre P. (1973) Observations surl'estimation de la production laitière des vachesallaitantes par la pesée du veau avant et aprèsla tétée. Ann. Zootech. 22, 413-422

Le Neindre P., Petit M. & Muller A (1976)Production laitière des vaches normandes à latraite ou à l'allaitement. Ann. Zootech. 25, 533-542

Leonhard 1. (1956) Preliminary investigationson the production and chemical composition ofmare's milk. Roez. Nauk Ro/n. 70 B3, 333-346(in Polish)

Linton R.G. (1937) The composition of mare'smilk. J. Dairy Sei. 8, 143-172

Linzell J.L. (1972) Milk yield, energy loss inmilk and mammary gland weight in differentspecies. Dairy Sei. Abstr. 34, 351-360

Linzell J.L., Annison F., Bickerstaffe R. &Jeffcott L.B. (1972) Mammary and whole-bodymetabolism of glucose, acetate, palmitate inthe lactating horse. Proe. Nutr. Soc. 31, 72-73

Measurement of mili< yield and composition in nursing mares 171

phases d'allaitement, de sevrage et de traite.Ann. Zootech. 11, 5-38Sharma a.p. (1974) Release of oxytocinelicited by suckling stimulus in mares. J.Reprod. Fert. 37, 421-423Smoczynski S. & Tomczynski R. (1982) A studyof chemical composition of mare milk. 1.Chemical composition of mare milk in first tendays of lactation. Zesz. Nauk. Akad. Roln.-Tech. O/szt. 17, 77-83 (in Polish)Smolders E.A.A. (1989) Composition of horsemilk during the suckling period. Livest. Prod.Sei. (in press)Storch R. (1969) Uber das maschinelle Melkenvon Stuten. Milchwissenschaft 24, 145-146Tarasevich L.F. (1979) The relationship of milkproduction of White-Russian Coach mares withudder shape and measurements. SbornikTrudov. Bie/oruskii Nauchno-/ss/edovate/'skii/nst. Zhivotnovodstva 20, 110-114 (in Russian)(Anim. Breed. Abstr. (1980) 48, 512)Thompson P.O., Paape M.J. & Smith J.w.(1973) Residual milk as affected by dose andtime of injection of oxytocin. J. Dairy Res. 40,221-227

Tumanova E.B. (1966) Certain problems of themorphological physiology of the udder of themare. Zap. Leningr. Se/.'-Khoz. /nst. 102, 130-134 (in Russian) (Dairy Sei. Abstr. (1967) 29,501-502)

Tyler S. (1972) The behaviour and socialorganisation of the New Forest ponies. Anim.Behav. Monogr. 5, 85-196

Ullrey D.E., Struthers R.D., Hendricks D.G. &Brent B.E. (1966) Composition of mare's milk.J. Anim. Sei. 25, 217-221

Wheelock J.V., Rook J.A.E & Oodd EH. (1965)The effect of intravenous injections of oxytocinduring milking and the removal of residual milkon the composition of cow's milk. J. Dairy Res.32,255-262

Wright D.E. & Wolff J.E. (1976) Measuring milkintake of lambs suckling grazing ewes by adouble isotope method. Proc. N.z. Soc. Anim.Prod. 36, 99-102

Zimmerman RA (1985) Effect of ration. oncomposition of mare's milk. Proc. 9th Eq. Nutr.Physiol. Symp. 96-102