EFFECTS OF FREQUENT MILKING ON MILK SECRETION DURING LACTATION IN THE GOAT: RELATION TO FACTORS WHICH LIMIT THE RATE OF SECRETION

Quarterly Journal of Experimental Physiology (1982) 67, 33-31 Printed in Great Britain EFFECTS OF FREQUENT MILKING ON MILK SECRETION DURING LACTATION IN THE GOAT: RELATION TO FACTORS WHICH LIMIT THE RATE OF SECRETION D R BLATCHFORD AND M PEAKER Hannah Research Institute, Ayr, Scotland (RECEIVED FOR PUBLICATION 1 AUGUST 1981) SUMMARY Goats were milked hourly with the aid of oxytocin at different stages oflactation Udder volume and milk yield were also measured The marked variation between goats in the time after parturition at which peak milk yield is attained and in the rate of decline after peak is illustrated Hourly milking had a stimulatory effect on the rate of milk secretion in early lactation (before peak) and in declining lactation (after peak), in both cases at previous milk yields of 1 1-1 48 g/ml udder volume d There was no stimulatory effect of hourly milking on milk yield at or near peak lactation (yield before the experiment > 1-48 g/ml volume d) or in late lactation (<11 g/mld) The responses of milk yield to hourly milking are discussed in relation to the factors which limit the rate of secretion In particular, it is concluded that a stimulatory response indicates that before the experiment the rate of secretion could not have been limited directly by the arterial supply of one or more substrates for milk synthesis It is stressed that the results were obtained under one dietary regime only INTRODUCTION The factors which control and limit the rate of milk secretion are unknown, other than in gross endocrinological and nutritional terms Dietary intake is of undoubted importance and it has been suggested that availability of certain substrates for milk secretion (eg some of the amino acids, see for example, Spires, Clark, Derrig & Davis, 1975; Mepham, 1976; Ranawana & Kellaway, 1977; rskov, Grubb & Kay, 1977) may be rate-limiting under 'normal' conditions or with certain diets in goats and cows; it has also often been implied that the supply of such substrates may limit secretion directly at the mammary level In goats, as well as in other animals (see references in Linzell & Peaker, 1971), frequent milking, with or without the aid of exogenous oxytocin to facilitate milk ejection, often leads to a temporary increase in the rate of milk secretion By frequent milking of one gland only Linzell & Peaker (1971) showed that the increase is caused by a local intramammary mechanism and that it depends on actual milk removal since local massage (akin to milking and which, like milking, sometimes leads to a small temporary increase in mammary blood flow ofsimilar magnitude - Annison & Linzell (1964), Reynolds, Linzell & Rasmussen (1967), the late J L Linzell & M Peaker (unpublished observations)) had no stimulatory effect Moreover, frequent milking had no effect on the arterial concentrations of many actual or potential substrates for milk secretion (Annison & Linzell, 1964; D R Blatchford & M Peaker, unpublished observations) It was suggested that there is present in milk a chemical factor which acts to partially inhibit milk secretion by a local negative feed-back While the identity of this factor(s) is being pursued (Maule Walker & Peaker, 1982), it was

34 D R BLATCHFORD AND M PEAKER realized that the response to frequent milking could throw light on the factors acting to limit the rate of secretion If the rate of secretion does increase, then it indicates that substrate supply per se could not have been rate-limiting at the mammary level at the time the experiment was done since obviously no local mechanism could overcome such a limitation imposed by arterial concentration or by blood flow Since it was clear from the studies of Linzell & Peaker (1971) and from the earlier work of Linzell (1967) that while on average milk secretion increased in response to frequent milking, there was considerable variation in that some goats showed a clear stimulation while others showed little or no change Therefore, responses to frequent milking have been studied throughout lactation in order to determine whether responsiveness can be related to stage of lactation and to the changes which occur in milk yield METHODS Animals Goats in their second to fourth lactation were used They were fed 1-2 kg/d hay and 15-1-7 kg/d concentrates (BOCM 'Red Label'; manufacturer's analysis, 4 5% oil, 16% protein and 9% fibre), half the concentrate ration being given after each milking This dietary intake is equivalent to 21-9-23 7 MJ/d metabolizable energy All were mated at the first or second oestrus of the breeding season and all were pregnant by day 2 of lactation (one by day 154; day of parturition taken as day zero of lactation) The animals were milked twice-daily at approximately 83 and 16 h; the milk yields (by weight) and the time of milking were recorded Experimental procedure At each stage of lactation, animals for study were taken from a pool of eleven goats On the day of the experiment, the animals were milked, starting at approximately 93 h, every hour for 8 h, 1 mu oxytocin being given iv each time (Linzell & Peaker, 1971); milk yields and times of milking were recorded After the afternoon milking on the day following hourly-milking, udder volume was determined after the injection of 1 mu oxytocin and removal of residual milk (empty udder volume) (Linzell, 1966) Analytical methods Milk Na, K, Cl and lactose were determined in a Technicon auto-analyser II (Neville & Peaker, 1981) Derived data Previous yield was taken as the mean yield in ml/h over the 3d immediately before each experiment The yield on hourly milking was taken as the total weight of milk removed during the 8 h of hourly milking plus the yield at the next morning milking This was done in order to obtain a true daily rate of secretion, since the amount of milk remaining in the udder was found in preliminary experiments to be virtually identical after the two morning milkings (day of the experiment and the day after (Linzell & Peaker, 1971 and unpublished); this yield was then expressed as ml/h Previous yields and yields on hourly milking were also calculated in relation to udder volume (ie yield (g)/udder volume (ml d) From plots of weekly yield against time, the week of peak yield was identified and the time taken as the time (d) from parturition to the middle of that week In animals which reached peak very early (see Results), daily records were examined At the time of each experiment, previous yield was also expressed as a percentage of the maximum (peak) yield for that goat in that lactation Previous milk composition was taken as the mean over 3 d before the experiment (mean of two glands) RESULTS Changes in milk yield and udder volume during lactation In goats, milk yield varies widely between animals Furthermore, the shape of the 'lactation curve' (ie yield per unit time plotted against time) is also highly variable although of a general pattern - a rise to peak yield followed by a more gradual decline Some animals reach peak sooner than others; in some yield declines more rapidly than in others This

RESPONSES TO FREQUENT MILKING 35 1 -, * E 6 8 * 8 - e * 4 2 5 1 15 2 25 3 Time from parturition (d) Fig 1 Milk yield (expressed as a percentage of the maximum yield of the animal during the lactation) in relation to stage of lactation All points obtained at > 2 d plus the point indicated by the arrow were from pregnant animals Thirty-nine experiments in eleven goats variation complicates the interpretation of experiments done at different times after parturition since animals at one stage may be in different physiological and metabolic states The variation in, and general pattern of, the lactation curve is illustrated for the goats used in the present experiments in Figs 1 and 2 In these Figures yield, at the time the experiments were done, expressed as a percentage of the maximum yield of the goat in that lactation, is plotted against the time from parturition (Fig 1) or the time from peak (Fig 2) The variation in the time to reach peak is evident; linear regression analysis indicates that peak yield was attained at a mean of 4 d post partum with 95 confidence limits of 1-81 d Similarly, the variation in the rate of decline in yield is evident both in terms of the time after parturition and the time after peak yield This marked variation was particularly evident after about 2 d of lactation (or 13 d after peak) by which time all the animals were pregnant In Fig 3 yield, expressed per unit of udder volume (g milk/ml tissue d), is plotted in relation to the time from peak yield At or near peak the yield was relatively high in relation to udder volume (> 1[5 g/ml d) Before peak, the yield was in the range 1 1-1 5 g/ml d, and from about 5 d after peak, yield declined more steeply Responses of milk yield to hourly milking Responses to hourly milking were classified as positive (yield on hourly milking 1/previous yield > I 1%) or negative (< 11%) Although this limit is somewhat arbitrary it does have some statistical foundation When the distribution of responses to hourly milking was plotted for all experiments, distinct positive skewness was apparent High values were successively eliminated until a normal distribution curve was achieved

36 1 D R BLATCHFORD AND M PEAKER so * 8 - -1 x 6 F 4 S C- 4 F 2 - I I I -5 +5 1 Time from peak yield (d) l 15 2 25 Fig 2 Milk yield (expressed as a percentage of the maximum yield of the animal during the lactation) in relation to time from peak yield All points obtained at > 13 d plus the point indicated by the arrow were from pregnant animals Thirty-nine experiments in eleven animals 25 I -n L r :- 1 5 - " 1- I F ) > 5 - * -- -- Fig 3 O - 5 5 1 Time from peak yield (d) 15 2 25 Milk yield expressed as g/ml udder volume d in relation to time from peak yield The dashed lines show the 'zone of responsiveness' (see Fig 5 and text) Thirty-nine experiments in eleven animals Then, using the SD for the remaining values, the standardized normal deviate was calculated for various values above the mean (using probability for one-sided deviation) (Campbell, 1967) At an hourly/previous yield of 11% the chances of a 'negative' observation falling above this value are approximately 2% Responses of milk yield to hourly milking have been assessed in relation to: time from parturition, time from peak yield and previous yield

16 RESPONSES TO FREQUENT MILKING 37 e14 8 12 1 * 1-8 L l 5 1 15 2 25 3 Time from parturition (d) Fig 4 Response of milk yield to hourly milking in relation to stage of lactation All points obtained at > 2 d plus the point indicated by the arrow were from pregnant animals Thirty-nine experiments in eleven animals The response of milk yield to hourly milking is expressed as a percentage of the previous rate as described in the text As shown in Fig 4, some goats showed positive responses (yield on hourly milking 1/previous yield of > 1 1%) in early lactation Of the eight studied at this stage, five showed a response of > 11% The three which did not respond were all ones in which peak yield had already been reached Later in lactation, the response appeared variable; however, no goat showed a positive response at > 2 d of lactation A similar pattern was evident in relation to time from peak yield (data not shown) A much clearer relationship emerged when response to hourly milking was plotted against previous yield in g/ml udder volume d (Fig 5) At previous yields above 1 5 g/ml d and below -1 g/ml d, no animals showed a positive response to hourly milking It was between these two values that responses to hourly milking were obtained Of fifteen experiments in which yield fell within this range, eleven goats showed a response of > 11% As can be seen from Fig 5, the four which did not respond were only just in this range A X2-test using columns of 4 g/ml d above and below 1[5 g/ml d and responses to hourly milking abo'e 11% (positive) and below 11% (negative) showed a highly significant inequality of distribution (P < 1) Therefore, a 'zone of responsiveness' to hourly milking was identified in these experiments and defined in terms of yield per unit of udder volume For further discussion, the zone has been set to include all those experiments in which a response of > 1 1% was obtained: 1 1-1 48 This zone is drawn on Fig 5 as well as on Fig 3 The yield obtained in response to hourly milking in terms of g/ml udder volume day was calculated and is shown in Fig 5 It can be seen that in nine of the eleven experiments in which the response was > 11%, yield was raised to values above the upper limit of the zone of responsiveness (ie to > 1 48 g/ml d), in other words to values found in peak lactation (see above) The over-all pattern of responsiveness to hourly milking can be explained by reference to Fig 3 It can be seen that animals may pass through the zone of responsiveness twice

38 16 D R BLATCHFORD AND M PEAKER 1l4 126-1 8 5 1 15 2 2-5 Yield (g/ml tissue d) Fig 5 Response to hourly milking in relation to milk yield expressed as g/ml tissue d before hourly milking (see text for details of calculation) The arrows indicate, for animals showing a response to hourly milking of > 1 IO, the yield per unit weight of tissue during the response The dotted lines indicate the 'zone of responsiveness' (see text) Stages of lactation: A, 3 12;, 6-1;, 14-181; U, 2-24; E], 25-29 Thirty-nine experiments in eleven animals in a lactation: first, during early lactation when yield is ascending to peak; second, during declining lactation Therefore, the lactation can be divided into four periods: (i) ascending - lactation (yield 1 1-1I -5 g/ml d) - positive response to hourly milking, (ii) peak lactation (yield > I1-5 g/ml d) - not responsive to hourly milking, (iii) declining lactation (yield -1-l5 g/ml d) - responsive to hourly milking, (iv) late lactation (yield < [1 g/ml d) - not responsive to hourly milking However, as explained above, the lengths of these phases varies markedly between animals Examination of the data showed that the ability to respond to hourly milking is not a property of particular animals regardless of the stage of lactation In four animals which were studied at all stages of lactation positive responses were obtained only when the milk yield fell within the zone of responsiveness (ie before peak and during declining lactation) and not at other times Response of milk composition to hourly milking In goats milked hourly with the aid of oxytocin, milk sodium and chloride concentrations increase while those of potassium and lactose decrease (Linzell & Peaker, 197 1) In these experiments milk composition at the third hour of hourly milking (the time when the change in composition is maximal or near maximal: Linzell & Peaker, 1971) are considered There was no obvious pattern to the responses in milk composition to hourly milking in relation to time from parturition or previous yield and there was no significant correlation between the change in sodium concentration and the response of milk yield to hourly milking (data not shown) There was also no significant difference in the change in lactose concentration between

RESPONSES TO FREQUENT MILKING 39 those which showed a response in milk yield of > 11% (fall of 2 mmol/l) and those which did not respond (fall of 4-6 mmol/l in the third hour) DISCUSSION From the evidence of Linzell & Peaker (1971) that the increase in milk yield on hourly milking is mediated by a local intramammary chemical mechanism in the milked gland only, it is clear that in the two periods of responsiveness to frequent milking (ascending lactation and declining lactation when milk yield is in the range 1 1-1 -48 g/ml udder volume d) that have been identified, the supply of one or more substrates for milk synthesis could not have been rate-limiting at the mammary level Therefore, in these periods the rate of milk secretion must have been under the rate-limiting control of other, probably hormonal, factors It is tempting to speculate that the lack of response to hourly milking at or near peak and in late lactation indicates that substrate supply is directly rate-limiting Even in late lactation when yield is low this argument could hold since available nutrients are diverted to body stores (see Vernon, 198) However, there are other possible explanations which at present cannot be excluded These include; (i) the local inhibitor is not produced, (ii) the mammary cells are not responsive to removal of the inhibitor, (iii) the cells are already secreting at their maximal intrinsic rate (at peak lactation) even with the local inhibitor present It must be stressed that the present results were obtained under one dietary regime Different diets at the various stages of lactation may alter the response to hourly milking For example, in preliminary experiments we have found that when dietary intake is restricted during declining lactation (and when from previous milk yield and udder volume measurements a stimulatory effect would have been expected), milk yield falls and there is then no stimulation by hourly milking These preliminary studies might indicate that substrate supply can become directly rate-limiting Dr P C Thomas has kindly estimated the metabolizable energy (me) intake of the animals on the diet described at about 22 MJ/d In lactating goats from this herd, Thomson (1979) found the metabolic rate to be approximately 11 MJ/d, leaving about 11 MJ/d for export in milk, for depostion in body stores or, in the case of pregnant animals, for fetal requirements From milk composition before each experiment and previous yield, energy output in milk was calculated In those studies done at or near peak lactation, the estimated basic requirements, plus the energy secreted in milk was greater than estimated dietaty intake by from 2 to 7 5 MJ/d In other words at or near peak the goats were in negative energy balance and must have been drawing on body reserves At lower yield before and after peak, the goats were in 'positive energy balance' The identification of rate-limiting factors in milk secretion is of major importance both in human lactation (particularly in areas of the world where the diet is poor) and in the productivity of dairy animals The use of frequent milking provides one tool for exploring the physiological nature of limitation at least in dairy animals Moreover, it can clearly be used to identify periods or conditions when it is worth investigating by other means whether or not the supply of one or more substrates for milk synthesis could be directly limiting the rate of secretion For example, Linzell & Mepham (1974) and Mepham & Linzell (1974) infused mixtures of essential amino acids or of non-essential amino acids or glucose into the arterial supply of autotransplanted mammary glands in order to determine whether supply of these substances normally directly limits the rate of milk or milk protein synthesis

31 D R BLATCHFORD AND M PEAKER Only in one goat was an indication obtained that milk and milk protein yield increased Clearly, such experiments should be done at times when there is reason to suspect, from negative responses to hourly milking, that the fact that substrate supply appears not to be directly rate-limiting at certain stages of lactation does not preclude substrate supply to the whole body from influencing milk secretion at these times by indirect, presumably hormonal, mechanisms We are most grateful to Mrs J M White for assistance, to Mr R M Mabon, Dr P L Clarke and Miss E Y Brechany for the chemical analyses, and to Mr T F Hutchison and Mrs S M Paton for care of the animals REFERENCES ANNISON, E F & LINZELL, J L (1964) The oxidation and utilization of glucose and acetate by the mammary gland of the goat in relation to their over-all metabolism and to milk formation Journal of Physiology 175, 372-385 CAMPBELL, R C (1967) Statistics for Biologists Cambridge: Cambridge University Press LINZELL, J L (1966) Measurement of udder volume in live goats as an index of mammary growth and function Journal of Dairy Science 49, 37-311 LINZELL, J L (1967) The effect of very frequent milking and of oxytocin on the yield and composition of milk in fed and fasted goats Journal of Physiology 19, 333-346 LINZELL, J L & MEPHAM, T B (1974) Effects of intramammary arterial infusion of essential amino acids in the lactating goat Journal of Dairy Research 41, 11-19 LINZELL, J L &-PEAKER, M (1971) The effects of oxytocin and milk removal on milk secretion in the goat Journal of Physiology 216, 717-734 MAULE WALKER, F M & PEAKER, M (1982) Prostaglandins and lactation Acta veterinaria scandinavica (in the Press) MEPHAM, T B (1976) Amino acid supply as a limiting factor in milk and muscle synthesis In Principles of Cattle Production, ed SWAN, H & BROSTER, W H, pp 21-219 London: Butterworths MEPHAM, T B & LINZELL, J L (1974) Effects of intrammary arterial infusion of non-essential amino acids and glucose in the lactating goat Journal of Dairy Research 41, 111-121 NEVILLE, M C & PEAKER, M (1981) Ionized calcium in milk and the integrity of the mammary epithelium in the goat Journal of Physiology 313, 561-57 RSKOV, E R, GRUBB, D A & KAY, R N B (1977) Effect of postruminal glucose or protein supplementation on milk yield and composition in Friesian cows in early lactation and negative energy balance British Journal of Nutrition 38, 397-45 RANAWANA, S S E & KELLAWAY, R C (1977) Responses to postruminal infusion of graded levels of casein in lactating goats British Journal of Nutrition 37, 67-79 REYNOLDS, M, LINZELL, J L & RASMUSSEN, F (1967) Comparison of four methods for measuring mammary blood flow in conscious goats American Journal of Physiology 214, 1415-1424 SPIRES, H R, CLARK, J H, DERRIG, R G & DAVIS, C L (1975) Milk production and nitrogen utilization in response to postruminal infusion of sodium caseinate in lactating cows Journal of Nutrition 15, 1111-1121 THOMSON, E M (1979) The effect of cold exposure on mammary metabolism in the ruminant PhD Thesis, University of Glasgow VERNON, R G (198) Lipid metabolism in the adipose tissue of ruminant animals Progress in Lipid Research 19, 23-16