- Email: [email protected]
Milk intake and production curves and allosuckling in captive Iberian red deer, Cervus elaphus hispanicus
ANIMAL BEHAVIOUR, 2000, 60, 679–687 doi:10.1006/anbe.2000.1515, available online at http://www.idealibrary.com on
Milk intake and production curves and allosuckling in captive Iberian red deer, Cervus elaphus hispanicus T. LANDETE-CASTILLEJOS*, A. GARCIA u †, J. GARDE†‡ & L. GALLEGO†
*Seccio´n de Recursos Cinege´ticos, IDR and †Departamento de Ciencia y Tecnologi´a Agroforestal, ETSIA, Universidad de Castilla-La Mancha ‡Instituto de Investigacio´n en Recursos Cinege´ticos (IREC), CSIC-UCLM, Spain (Received 15 October 1999; initial acceptance 30 November 1999; final acceptance 2 June 2000; MS. number: 6384R)
In two experiments, we compared milk intake (assessed by weighing calves before and after suckling) and milk production (by hand milking hinds) in Iberian red deer both when calves sucked their mothers together (group-suckling experiment) and when mother and offspring were isolated (isolation-suckling experiment). In both experiments, the general lactation curve for calves increased to a peak and then decreased (type I, standard lactation curve in mammals), whereas the curve for hinds decreased from the start (type II). However, in the experiment on group suckling, calves ingested 17.2% more milk than that produced by their mothers from weeks 6 to 20. In both isolation- and group-suckling experiments, hinds showed an overproduction of milk decreasing from weeks 1 to 5. This decreasing overproduction coincides with a similar trend in calf mortality reported in the literature and might thus be aimed at ensuring calves have sufficient nutrients when mortality is highest. In addition, allosuckling observations in the group-suckling experiment showed an inverse relationship between milk production and percentage of allosucking attempts. Allosucking attempts were also more frequent after the milk overproduction period. Both findings suggest that allosuckling is a response to compensate for a reduced maternal milk supply.
of recorded data and the fitting of such data to mathematical models (reviewed by Serchand et al. 1995) produce a lactation curve. In the standard lactation curve in mammals (type I, Arman et al. 1974; Loudon et al. 1983; Robbins et al. 1987) the amount of milk increases with the increasing energetic requirements of the offspring, peaks when the mother can no longer provide all the energy required by the calf and then starts a slow decrease as milk is substituted by solid food (Lee et al. 1991). In addition, curves decreasing continuously from the start of lactation have been found in red deer (termed type II, Arman et al. 1974; Sadleir 1980; Loudon et al. 1983). These have been attributed to a poorer diet (Loudon et al. 1983, 1984); however, both type I and II lactation curves have been found in red deer fed the same diet ad libitum (Garci´a et al. 1999). Some of the studies based on the frequency and time of suckling have reported allosuckling or suckling between a female and an offspring different to hers. Allosuckling has been reported in 74 species in the majority of mammalian taxa (reviewed by Packer et al. 1992) and has been considered either a maladaptive behaviour arising from crowded conditions in captivity or adaptive behaviour in natural conditions.
Observations on the frequency and length of suckling have been commonly used as estimates of nutrient transfer from mother to offspring during lactation in behavioural research (Clutton-Brock et al. 1982; Gauthier & Barrette 1985; Birgersson et al. 1991; Ekvall 1998). However, the literature shows little evidence for a relationship between length of suckling bouts and milk transferred (reviewed by Cameron 1998). In some instances, for example fallow deer, Dama dama, no relationship has been found between time spent suckling and fawn growth (Birgersson & Ekvall 1994), and isotope labelling has shown no relationship between time or frequency of suckling and milk transfer in horses, Equus caballus (Cameron et al. 1999). The most common methods used to estimate milk transfer directly are milking the mothers (which estimates the ceiling of maternal production; Robbins et al. 1987; Beal et al. 1990; Landete-Castillejos et al., in press) and weighing calves before and after suckling (termed double weighing, an intake estimate; Loudon et al. 1983, 1984; Garci´a et al. 1999). Both the plotting Correspondence: T. Landete-Castillejos, Seccio´n de Recursos Cinege´ticos, IDR, Universidad de Castilla-La Mancha, 02071 Albacete, Spain (email: [email protected]). 0003–3472/00/110679+09 $35.00/0
2000 The Association for the Study of Animal Behaviour
679
2000 The Association for the Study of Animal Behaviour
680
ANIMAL BEHAVIOUR, 60, 5
Crowding may result in misdirected behaviour by the mother (Boness 1990; Tiplady 1990), or opportunistic thefts by offspring (Murphey et al. 1995) or prevent the mother repelling allosuckling attempts (Le Boeuf & Briggs 1977). Under natural conditions, allosuckling has been explained as an altruistic behaviour, either reciprocal (Pusey & Packer 1994), directed to kin (Ekvall 1998) or to defend offspring communally from infanticide or predation (Pusey & Packer 1994). In Scottish red deer, Clutton-Brock et al. (1982) found that allosuckling frequency was very low or nonexistent, whereas Cowie et al. (1985) found allosuckling in the majority of hinds whose calves had been removed. In contrast to red deer, and despite allosuckling being attributed to crowding in captivity in fallow deer, the frequency of allosuckling in free-ranging fallow deer (Ekvall 1998) is similar to that in captivity (Birgersson et al. 1991; Birgersson & Ekvall 1994). In addition, Ekvall (1998) found that the proportion of allosuckling bouts increased with the stage of lactation. Although milk transfer in allosuckling has been confirmed in the field by radioisotope labelling (Hoogland et al. 1989), no study has estimated the amount of milk transferred. In this study we compared the milk production of red deer hinds with the milk intake of their calves both when calves sucked freely in their social group and when they sucked their mothers in isolation. Thus, we aimed to assess milk over- or underproduction relative to calf intake from both their mother (isolation-suckling experiment) and including intake from other hinds (group-suckling experiment). We also investigated the relationship between these variables and allosuckling observations during lactation. METHODS Twenty-two Iberian red deer hinds and their 21 calves were housed in an outdoor enclosure (5000 m2) under the conditions described in Garci´a et al. (1999). Both during gestation and throughout lactation, hinds were fed diets based on suggestions by Brelerut et al. (1990): barley straw and hay from barley, alfalfa, oat and sweet beetroot (16% protein). Deer had ad libitum access to food. Fifteen days after birth, calves had free access to a creep feeder containing an 18% CP concentrate diet. We used 14 hinds and 13 calves in the first experiment comparing the milking estimates with those of weighing the calf before and after suckling (henceforth termed double weighing) when all calves were released together for suckling (henceforth termed group-suckling experiment). We used eight hinds and their calves in the second experiment comparing milking and double-weighing estimates carried out by isolating mother–calf dyads (henceforth termed isolation-suckling experiment).
Group-suckling Experiment Two hinds were 2 years old and the other 12 were 5 years old. One of the hinds lost her calf at 7 weeks but continued producing milk so her milking data were used
in the analysis. Calving took place between 4 June and 12 August 1998. Mean hind weight in the week after calvingSD was 94.13.34 kg.
Milking Beginning 2 weeks after calving, we milked hinds every 4 weeks up to week 34. Weaning was enforced simultaneously in all hinds to keep social conditions constant throughout the experiment. Thus, every hind remained in the group for at least 34 weeks. Before milking, we separated hinds from their calves for 6 h (0800–1400 hours) in a deer-handling facility. Individuals were hand milked under anaesthesia until the four quarters of the udder were totally emptied, which took 10–60 min. Daily milk production was considered to be four times the amount collected in each trial. To reduce health risks from anaesthesia, a low-dose combination of xylazine (0.5 mg/kg live weight) and ketamine (1 mg/kg) was injected into the jugular vein. After inducing anaesthesia, we injected 10 IU of oxytocin into the right jugular vein 1 min before milking was started. After milking, we reversed anaesthesia with a yohimbine injection (0.25 mg/kg live weight). Hinds were weighed weekly on an electronic balance to within 50 g.
Weighing the calf before and after suckling Double-weighing estimates were made weekly after a 6-h isolation period, in a similar way to that described in the literature (Arman et al. 1974; Loudon et al. 1983; Garci´a et al. 1999). Calves remained together indoors until the start of each double-weighing procedure to reduce handling stress, whereas hinds remained outdoors. Prior to and after suckling, calves were weighed in a restraining box placed on an electronic balance to within 5 g. All calves were released simultaneously outdoors for 15 min. Hinds and their calves were fitted with matching colour collars for identification. Two observers recorded the identity of calves and hinds performing successful suckling bouts (contact with the udder for longer than 1 s). Suckling occurred very quickly and took only 1–2 min, with a few suckling bouts during the rest of the suckling period. As a result, we focused observations on successful attempts and did not record rejections in sufficient detail to used them in statistical analysis. Because calves often suck hinds in a lactation stage different to the age of the alien calf, we used the term ‘allosucking bout’ as the suckling bout timed according to the age of the calf, and the term ‘allonursing bout’ as that timed according to the lactation stage of the hind (i.e. the number of allosucking and allonursing bouts coincide, but may show a different temporal pattern). To reduce stress in the calves only one daily double-weighing measure was made (see Ethical note). We obtained daily milk production in both cases by multiplying the one daily record by 4 and fitting the resulting data to Wood’s (1967) gamma function as described by Garci´a et al. (1999). In the double-weighing procedure, differences of less than 50 g were recorded as blanks. Data were analysed for the first 19 weeks (no suckling was observed in any calf after this date).
LANDETE-CASTILLEJOS ET AL.: MILK CURVES IN DEER
Isolation-suckling Experiment Eight hinds, four 6 years and four 3 years old, and their calves were used in this experiment. Mean hind weight after calvingSD was 97.13.2 kg. Births took place between 6 and 31 of May 1999. Because two calves never sucked their mothers during this experiment, we used data from only six hinds. We carried out milking similarly to that of the groupsuckling experiment. However, because hinds had shown a good tolerance to the milking under anaesthesia the previous year, and to increase the definition of the fit for the first few weeks (when milk production changes more quickly), we milked them weekly for the first 6 weeks and then once every 4 weeks until week 18. Double weighing was done as in the previous year, but mother–calf dyads were isolated together for 15 min inside the handling premises. As in the previous year, and to reduce stress, calves were housed together. Double weighing (as milking) was done for 18 weeks.
Statistical Analysis We computed milk intake and production curves by fitting the logarithmically transformed data to the regression line derived from the gamma function (Wood 1967): Yt =atb e −ct; where Yt is the yield at week t. The parameters a, b and c thus estimated were used to obtain both the lactation curves and the estimated daily production. For curves of type I b was positive (Garci´a et al. 1999), whereas for those of type II b was negative or not statistically significant. We estimated the general model for milk intake (double weighing) and yield (milking records) curves by pooling data to compute an averaged curve of each type of method. Because R2 does not discriminate between lack of fit and pure error in regressions involving repeated measures, in the general models, R2/R2max was computed to show the percentage of variation explained that was not due to pure error (Draper & Smith 1998). We compared intake and production curves both for the general model and in sections in which differences were expected (i.e. periods of overproduction or overintake). The first comparison did not require a statistical test because it involved a different model (see Table 1). For intake and yield curves on sections showing overintake or overproduction we fitted the logarithmically transformed data to the logarithm of week, as the function Yt =atb accurately describes both the rising section of a lactation curve and its decreasing phase separately. Despite hind overproduction lasting up to week 5 in the group-suckling experiment, we compared weeks 1–6 because the first two milking trials had been conducted on weeks 2 and 6. If the test comparing regression lines showed significant differences between them, we computed the percentage of milk overproduced or that ingested above the mother’s production from the general model. Similar comparisons were made between pairs of mother–offspring curves that were marginally significant (P<0.10, always in intake curves). We used the
Statgraphics package for these analyses (Statgraphics 1997). Management was designed to keep calf mortality in our premises as low as possible. Thus it was unlikely to resemble mortality in a free-ranging population, and the sample size was too small to be used for statistical comparisons. In consequence, we compared published research on natural rates of calf mortality during lactation (Clutton-Brock et al. 1982) with the ratio milk yield:milk ingested for the period of milk overproduction found in both experiments (5 weeks). The test was conducted with a Pearson correlation. We used one-way ANOVAs to assess differences in the total milk production of the hinds during lactation (i.e. assessed by milking), the mean weight of the hinds during this period and their weight loss (difference between weight at the end of lactation and that after calving), and the weight of the calf within 24 h of birth between individuals classified with regard to the type of curve (classified in two groups I or II) both for milking and double-weighing curves, or the existence of a period when intake exceeded production. For frequency of suckling bouts, both those initiated by the calf (number of bouts observed at every age of calves) and those received by the mother (bouts observed on each lactation week of hinds), we used Pearson correlations. In contrast, we compared the percentage of allosucking and allonursing bouts (number of suckling bouts from alien hinds or calves versus number of total bouts observed per trial) with a test comparing the regression line fitting each variable (Statgraphics 1997). Differences in the percentage of allosucking or allonursing bouts between the period of milk overproduction and the rest of lactation were assessed by one-way nonparametric ANOVAs (Meddis 1984). Data shown are records observed, not daily or weekly estimates.
Ethical Note We followed Spanish and European guidelines and laws in the use of animals in research.
Separation of calves from their mothers The period of mother–calf separation was based on published research (Arman et al. 1974; Loudon et al. 1983), which in turn was based on suckling frequency of this species (Kelly & Drew 1976; Loudon et al. 1983). Calves were always kept with other calves for the separation period (even in the isolation-suckling experiment) to reduce stress. They were kept indoors rather than outdoors because they showed signs of stress (including weight loss) when chased to the handling premises for the first weighing. These symptoms were alleviated when they were driven together with their mothers after suckling for the second weighing. Calves gave distress calls, particularly at the beginning of the experiment when only a few calves were born. Later on, experience of older calves and a larger number of individuals resulted in fewer such calls. Noise was kept to a minimum to reduce stress. Hinds were also kept together outdoors to minimize stress. Nevertheless, hinds were more habituated
681
682
ANIMAL BEHAVIOUR, 60, 5
than calves to handling and showed very few signs of stress. Also, they did not show apparent signs of rejection of their calves, and most rejections observed appeared in older calves after they had already sucked the hinds once or more. No calf stopped growing during the experiment.
Handling during double weighing Every effort was made to minimize stress during handling of calves at weighing. This was easy when calves were small, which produced very reliable measures. As they grew, calves became more restless, occasionally jumping to avoid being caught, and we often had to take them in our arms to reduce stress when driving them to the balance.
Anaesthesia We chose the combined action of xylazine and ketamine as anaesthetic because the synergism between them allowed a much lower dose to be effective in inducing insensitivity and also because it allowed a quicker recovery. This mixture has been widely used for immobilization in a wide range of species, including pregnant white-tailed does, Odocoileus virginianus, that were anaesthetized 10 times in a 6-month period with ‘no effect on fawning date or survival’ (DelGiudice et al. 1986, page 247). Lactating hinds have been subjected to anaesthesia more often than in our study: up to 17 times in 300 days, or once in 2 weeks in Arman et al. (1974) without apparently affecting calves. Anaesthetized individuals lay down after 50 s. After the injection of yohimbine they got up within 2 min. The majority could walk, but some lay down again. They appeared to be disoriented at first, and tongue control (the tongue of the animals was stretched out of the mouth to drain saliva and prevent choking when lying down) was not evident for 1–5 min after they stood upright. A few minutes (5–10 min) after getting up, hinds showed standard signs of ruminating. We left them inside the handling premises to avoid other hinds attacking them. After the recovery period, hinds appeared to behave normally. Hinds were anaesthetized for 15–60 min (mean ca. 30 min). They needed a longer time under anaesthesia at the beginning of lactation because milk production was greater and also because habituation to milking and frequent handling during lactation reduced resistance to anaesthesia.
Possibility of the anaesthetics passing to the milk Ketamine does not pass into the milk and is eliminated in the kidney. Xylazine could pass into the milk, although it has not been detected at 5 and 21 h postanaesthesia in dairy cows (Plumb 1999). However, this possibility appears to be very small because hinds were milked dry and it would take them some hours to fill the udder. Moreover, as mentioned above, reports of anaesthesia of pregnant hinds did not find any effect on calves at such a sensitive stage in their life.
Standardization of the filling state of the udder Double-weighing techniques in cattle require a standardized filling state of the udder (i.e. the weighing is
done twice so that the second and valid record starts with an empty udder; Beal et al. 1990). However, despite this, and despite research in cattle showing that several consecutive measures are needed for the double weighing to reach milking precision we weighed calves only once a week to reduce isolation stress (Garci´a et al. 1999). This resulted in more variable double-weighing estimates than milking ones (see Results) and so we had to use marginal significance (P<0.1) values to compare mother–offspring dyads of curves. Similarly, the most advisable protocol for milking from a technical point would be either to milk the udder empty or to allow the calves to suck their mothers at the start of the isolation period. However, that would mean anaesthetizing the hinds twice, or separating the calves from their mothers overnight in addition to the following 6-h experimental period, both of which were ethically unacceptable. To disrupt lactation as little as possible, milking and double-weighing estimates in the same week were done 3–4 days apart. RESULTS Milk production estimated by double weighing was similar to that obtained by milking when individuals sucked together (193.1 versus 193.6 litres). However, double weighing underestimated milking when calves sucked their mothers in isolation (143.9 versus 200.6 litres). The first method was more variable than milking and so, only three out of 13 curves from individuals reached marginal significance (P<0.10), whereas 13 out of 14 milking curves reached significance (P<0.05). In the isolationsuckling experiment only one of six double-weighing curves of individual calves was significant, whereas four of six milking curves were. Such variability makes comparisons and assessment of effects of any factor more difficult. The percentage of variability explained by the model also reflects this pattern: thus, for calves in the group-suckling experiment, R2/R2max of double weighing was 80.88%, and the mean R2 of individual curves was 32.684.51%; whereas for milking curves these values were R2/R2max =99.23% and that for individual curves 87.493.49%. In a similar way, the R2/R2max for the general model in the isolation-suckling experiment was 77.53% for the double-weighing curve and 98.24% for the milking one. The curve type of the general model estimated by double weighing was also different to that estimated by milking hinds in both treatments (Table 1): the calf intake curve was a standard mammal lactation curve (type I), whereas hinds had a curve of type II, both in the group- and isolation-suckling experiments (Fig. 1; general model for double weighing: group suckling: Yt =942.6 t0.5682 e 0.1021, P=0.007; isolation suckling: Yt = 1026.6 t0.6062 e 0.1205; P=0.016; general model for milking curves: group suckling: Yt =1994.0e 0.0612; P<0.001; isolation suckling: Yt =2334.8 e 0.0431; P<0.001; see Table 1 for model details). This pattern was also found in the frequency of each type of individual curve: for the groupsuckling experiment 61.5% of calves had curves of type I, whereas 64.3% of hinds had curves of type II; in the isolation-suckling experiment all the calves had intake
LANDETE-CASTILLEJOS ET AL.: MILK CURVES IN DEER
Table 1. Parameters and their confidence interval at 95% probability level for lactation curves estimated by milking and double weighing both when calves suckled their mothers in a group, and when calves suckled their mother in isolation
Curve
Group suckling Milking*
R2/R2max (model) (%)
P (model)
F
99.23
0.001
129.47
Parameter
ln a b c
Double weighing
Isolation suckling Milking*
80.88
98.24
0.007
0.001
5.23
56.78
77.53
0.016
4.45
−0.0144 −0.0612
ln a
6.8486
b
0.5682
c
−0.1021
ln a b c
Double weighing
7.5979
7.7569 −0.0897 −0.0431
ln a
6.9340
b
0.6062
c
−0.1205
95% confidence interval
7.3785 7.8173 — −0.0718 −0.0506 6.5190 7.1783 0.0181 0.9553 −0.1649 −0.0394 7.6577 7.8616 — −0.0546 −0.0316 6.5238 7.3442 0.0899 1.1225 −0.2055 −0.0355
P
0.001 >0.1 0.001 0.001 0.004 0.002 0.001 >0.1 0.001 0.001 0.022 0.006
The model fitted corresponds to Wood’s (1967) gamma function: yield Y at week t, Y(t)=a tb e −ct. Because the function is fitted as a regression line on logarithmically transformed data and the antilog transformation introduces a bias in the standard error and confidence intervals, a natural logarithm of parameter ‘a’ along with its confidence intervals are given instead of their back-transformed value. R2/R2max indicates, for regression involving repeated measures, the percentage of variability explained by the model from that ‘explainable’ (ie. not due to pure error of repeating measures). *Values of parameters correspond to final model, which did not include ‘b’ as P>0.1.
curves of type I whereas 50% of the hinds had curves of type II. However, curves of mother–offspring dyads showed a weaker relationship: 38.5% of dyads in the group-suckling experiment had intake curves of type I and production curves of type II, 23.1% had both curves of type I or II, and 15.4% had intake curves of type II and production curves of type I; in the isolation-suckling experiment 50% of dyads had only type I curves and 50% had intake curves of type I and production curves of type II. The type of intake and production curves in the groupsuckling experiment did not seem to depend on the total amount of milk produced during lactation, the weight or weight loss of the hind, or calf weight at birth. Thus, there were no differences between hinds with curves of type I and those with curves of type II in milk production (t12 = 0.473, P>0.1), mean weight during lactation (t12 = 0.325, P>0.1), or percentage of body weight lost (t12 =0.687, P>0.1). Calf birth weight did not vary with the type of lactation curve of the mother (t12 =0.510, P>0.1), or of the calf (t11 =0.50, P>0.1). The type of curve of the calves did not affect the mean weight of their mothers (t11 = 0.150, P>0.1), or their percentage of body weight lost during lactation (t11 =1.434, P>0.1). Our data
are not adequate to assess the effect of calf sex on every variable. In the group-suckling experiment intake apparently exceeded hind production from week 6 onwards (Fig. 1; the slight overproduction after week 27 is meaningless because intake data were not collected at this time), whereas production exceeded intake in weeks 1–5. The production curve was significantly different from the intake one for weeks 1–6 (overproduction period: intercepts: F1,3 =6.39, P=0.014; slopes: F1,3 =5.54, P=0.021) and from week 6 onwards (overintake period: intercepts: F1,3 =4.96, P=0.027; slopes: F1,3 =2.32, P>0.1). Similar results were obtained in the isolation-suckling experiment, where, again, regression lines for intake and production curves differed from weeks 1 to 5 (overproduction period: intercepts: F1,3 =17.18, P<0.001; slopes: F1,3 =5.67, P=0.022), whereas they differed only in intercepts but not slopes from week 6 onwards (intercepts: F1,3 =5.08, P=0.028; slopes: F1,3 =0.35, P>0.1). However, after the peak in intake in this case, intake underestimated milking records. The results in the group-suckling experiment might have been an artefact of unbalanced double-weighing data. However, this seems unlikely because intake and
683
ANIMAL BEHAVIOUR, 60, 5
3000 (a ) 2500 2000 1500 1000 500 Milk (ml)
684
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33
3000 (b) 2500 2000 1500 1000 500 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Week
Figure 1. General Wood’s (1967) model (fitted to pooled data from all individuals, which are also shown in the figure) for intake curves estimated by weighing red deer calves before and after suckling ( , ——) and those of milk production estimated by hand milking of hinds ( , – – –). (a) Calves suckled their mothers together; (b) each calf was isolated with its mother.
production curves in mother–offspring pairs also differed from each other in both individuals where both curves were significant and intake exceeded production in the midde of lactation (calf 2, second born: intake curve: R2 =44.38%, F2,10 =3.99, P=0.053; model: Yt =1285.9 t0.8180 e 0.1456 t, all coefficients significant at P<0.05; production curve: R2 =84.54%, F1,7 =38.28, P<0.001; model: Yt =2359.7 e 0.0516 t, both coefficients significant at P<0.001, but coefficient b= 0.1631 discarded at P>0.1; calf 7, seventh born: intake curve: R2 =58.45%, F2,8 =5.63, P=0.03; model: Yt =775.2 t1.3779 e 0.2320 t, all coefficients significant at P<0.05; production curve: R2 =88.85%, F1,7 =47.81, P<0.001; model: Yt =3242.8 e 0.1089 t, both coefficients significant at P<0.001, but coefficient b=0.4821 discarded at P>0.1). The comparison between intake and production curves in periods of overintake or overproduction did not yield any significant result, probably because of the small sample size (intercept of both curves in calf 2 occurred at week 2–3, which included only one milking record; the rest of the segments could be compared). Total milk production estimated from the general model in the group-suckling experiment was 58.3 litres for the 5
weeks where production exceeded intake. For this period the estimated intake was 42.5 litres. Thus, hinds produced 44.0620.1% (XSE; N=5) more milk than their calves ingested. From week 6 to 28 (overintake period judged from the intersection of the curves), hinds produced 123.0 litres, whereas calves ingested 139.1 litres. Thus, calves ingested 12.441.31% (N=22) more than their mothers were producing every day. For weeks 6–20 (where we have data for both curves), calf intake was 17.520.5% (N=14) greater than hind production. In the isolation-suckling experiment hinds produced 72.0 litres during weeks 1–5, whereas calves sucked 45.4 litres. Thus, hinds produced 65.9321.65% more than their calves ingested. Published data on the percentage of calf deaths for the first 5 weeks in free-ranging deer showed a significant correlation with the overproduction both in the groupsuckling experiment (r3 =0.938, P=0.019) and in the isolation-suckling one (r3 =0.950, P=0.013). If all individual curves from the group-suckling experiment are considered (significant and nonsignificant), six calves (42.86%) showed periods when they sucked more milk than their mothers produced (two were males). Of these, four had a curve of type I, whereas three of their mothers had a curve of type II. The hinds whose calves sucked other hinds (assessed by comparing lactation curves) did not differ from the others in their total milk production (t11 = 0.025, P>0.1), weight (t11 =0.191, P>0.1), percentage of body weight lost during lactation (t11 =1.298, P>0.1), or birth weights of their calves (t11 =0.903, P>0.1). The type of curve did not affect the frequency of allosucking bouts observed (21 =0.04, P>0.1): 29.4% occurred between dyads with curves of type I, 37.3% between calves with curves of type I allosuckling hinds of type II, 15.7% between calves of type II and hinds of type I, and 17.6% between dyads with curves of type II.
Allosuckling Observations We observed 37.80% allosucking bouts (number of allosucking bouts:total number of suckling ones). These occurred both when the offspring of the hind was already sucking and when the alien calf was sucking alone. The number of suckling bouts decreased with the stage of lactation (suckling bouts: r17 = 0.724, P<0.001; nursing bouts: r14 = 0.568, P=0.022; Fig. 2). In contrast, allosucking and allonursing bouts constituted an increasing percentage of suckling bouts as lactation proceeded (allosucking bouts: r17 =0.810, P<0.001; allonursing bouts: r12 =0.786, P=0.001; Fig. 3). Calves allosucked more often after the 5-week milk overproduction period than before (XSE of the percentage of daily allosuckling during the first 5 weeks=22.453.67%, that for weeks 5–20, 61.517.46%; H1 =8.871, P<0.01). This difference was also significant, although smaller, for hinds (allonursing bouts: 28.166.13 versus 44.099.15%; H1 =4.312, P<0.05). Although this result suggests a difference between trends of allosuckling and allonursing, the regression lines were not statistically significant (intercepts: F1,3 =1.61, P>0.1; slopes: F1,3 =1.39, P>0.1).
LANDETE-CASTILLEJOS ET AL.: MILK CURVES IN DEER
DISCUSSION
20 18
Suckling bouts
16 14 12 10 8 6 4 2 0
5
10 Week
15
20
Figure 2. Daily frequency of total suckling bouts according to the stage of lactation of the calf (x) and the hind (C). The types of suckling bout do not coincide because allosucking is timed differently according to whether the age of the alien calf or the lactation stage of the hind is considered.
100
Allosuckling/allonursing bouts (%)
90 80 70 60 50 40 30 20 10 0
5
10 Week
15
20
Figure 3. Percentage of allosucking bouts (x) and allonursing bouts (C) during lactation. Fitted lines correspond to the following models: allosucking: Y=8.2888+4.2944t; allonursing: Y=13.6267+2.8126t. Lines did not differ significantly in slopes or intercepts.
The percentage of allosucking bouts recorded on each trial correlated inversely with the estimated milk production of the hind for that week (r17 = 0.790, P<0.001). In contrast, the percentage of allosucking bouts by each calf did not correlate with its mother’s total milk production (r9 =0.288, P>0.1), her weight (r9 =0.389, P>0.1) or the percentage of body weight she lost during lactation (r9 =0.321, P>0.1), although there was a positive correlation between allosucking bouts and calf birth weight (r9 =0.692, P=0.018). Conversely, the percentage of allonursing bouts increased with increasing milk production by the nursing hind (r9 =0.619, P=0.042), but there was no correlation with her body weight (r9 =0.549, P=0.08) or the percentage of weight she lost during lactation (r9 =0.212, P>0.1).
Our results show that double weighing and milking estimate processes in lactation that differ from each other. Reports in the literature, however, consider them as alternative methods for estimating milk production. Indeed, double weighing has been used more often than milking because it is easier to do, although it is more variable (Beal et al. 1990; Garci´a et al. 1999; this study). If, as considered in the literature, both methods estimate the same process they should produce the same type of curve. However, both in the general model and in more than half of the mother–offspring dyads (52.6%) the type of intake curve was different to that for production. It seems unlikely that the pattern of lactation curves in the general models was due to the particular conditions in which the experiment was conducted: this pattern was repeated the following year with different individuals differing in age; intake curves were of type I both when individuals sucked their mothers in isolation and when they did so together with other calves; and hinds had a curve of type II both when trials were conducted once a month (which might have given insufficient detail to detect the peak characteristic of curves of type I) and when trials were carried out weekly for the first 6 weeks to increase curve definition. In the literature, in contrast, the standard lactation curve for all mammals is considered a single one (type I) and intake and production curves are not distinguished (Lee et al. 1991). It is no surprise that calves had a general curve of type I, as their energetic requirements rise with increasing body weight until the slow process of weaning is enforced by the mother. It is more striking, though, that hinds had a general curve of type II and overproduced milk during the first 5 weeks, particularly considering the high cost of lactation in red deer (e.g. lactating hinds suffer a higher mortality rate than barren ones, Clutton-Brock et al. 1982). The milk overproduction might ensure a sufficient flow of energy for the calf during the first few weeks of life, when both growth rate and mortality are greatest. This is more strongly suggested by the significant correlation between increasing percentage of summer deaths reported in this period and increasing milk overproduction. However, these results should be taken with great care, as the test compares different sets of animals (summer deaths recorded by Clutton-Brock et al. 1982 under free-ranging conditions) and at a different time. Other species such as fallow deer also show milk production greater than intake during their first few weeks of life, as calves terminate 40% of the suckling bouts during their first week of life (suggesting they feed to satiation), but their mother terminates all the observed suckling bouts from the second week onwards (Birgersson & Ekvall 1994). Unfortunately, as our deer are managed to keep them in good condition with low mortality, our data are inadequate to assess whether calf mortality is higher for mothers producing the smallest amount of milk during the first few weeks. Nevertheless, indirect results suggest that this might be so under natural conditions, as factors increasing the mortality of calves under natural conditions, such as an increasing calving delay (Clutton-Brock
685
686
ANIMAL BEHAVIOUR, 60, 5
et al., 1982), also reduce the amount of milk produced (Landete-Castillejos et al., in press). However, our results are not conclusive and further experiments are needed to discriminate between these and alternative/ complementary hypotheses. Our results do not allow us to assess which factor caused the hinds or calves to have one type of curve or the other, as neither the weight, the percentage of body weight lost by the hinds, nor the calf birth weight differed between individuals with either type of curve. It is particularly striking that hinds did not differ in the amount of milk produced as a function of the type of curve displayed, in contrast to Loudon et al.’s (1983) findings. The origin of each type of curve might be due to an unexamined factor, such as dominance rank among hinds or calves. The comparison of intake and production curves also showed a period in which intake was greater than production. This is likely to differ both in length and milk surplus for each individual case (in fact, many dyads of curves did not even intersect at any stage), and it may vary with ecological conditions. However, it can provide a rough estimate of the amount of milk that is sucked from other hinds assuming that the mother’s milk is not, in turn, sucked by other calves. Although such an overintake period could arise from an imbalance in the data, such that calves with greatest intakes contributed more to the general model in the overintake period and those with smallest intakes contributed more in the tails, this does not seem very likely. This is because there was a milk overintake in two mother–offspring dyads where the fit was significant. In addition, sustained allosuckling is suggested by the fact that the hind that lost her calf in week 7 maintained her milk production and did not become dry, presumably because other calves often sucked her. Such incidence of allosuckling has often been considered a consequence of crowded conditions in the experimental set-up. Although it is likely that conditions under which double weighing was carried out might have influenced the amount of milk sucked from other hinds compared with what might happen under natural conditions, it seems unrealistic to assess the amount of milk allosucked in the field with direct methods such as those used here. Alternatively, the amount of milk allosucked might be estimated from observations, but the utility of these is controversial (Cameron 1998), and when compared with the more reliable direct methods, as has been done in horses (Cameron et al. 1999), they have not shown any relationship. Finally, although the incidence of allosuckling we found differs from that reported in other studies of red deer (which in turn, also varies: none in Clutton-Brock et al. 1982; only when the mother is removed in Cowie et al. 1985), studies in other cervid species such as fallow deer have shown that the frequency of allosuckling is similar in captivity (73%, Pelabon et al. 1998; 23–57%, Birgersson et al. 1991; Birgersson & Ekvall 1994) and free-ranging individuals (43%, Ekvall 1998). Moreover, the frequency of allosuckling reported here for Iberian red deer (38%) falls within the range of that found in captivity and in free-ranging fallow deer.
The comparison between hind milk production and observed allosucking frequency shows that the percentage of allosucking bouts increases as the milk production of the mothers decreases. This suggests that calves may allosuck as an attempt to maintain the flow of milk intake despite the reduction enforced by their mothers. The comparison of allosucking and allonursing percentages before and after the milk overproduction period shows a smaller difference for hinds than for calves. This might suggest that the timing of allosucking differs from that of allonursing and, therefore, that hinds may usually receive a high proportion of sucking attempts by calves older than their own, although the comparison of regression lines belonging to allosucking and allonursing bouts did not achieve statistical significance. This point deserves further study because such allosucking attempts by older calves of mothers in their early lactation stages, if proven, might be caused by the milk overproduction during the first few weeks of lactation. Pelabon et al. (1998) also found that fallow deer does during the first 2 weeks of lactation were allosucked by older calves, which was considered to be a result of the longer suckling bouts (ad libitum of the calf) at this stage of lactation. Finally, the fact that some of the allosuckling attempts involve the diversion of milk destined for young calves whose only source of energy is milk to older calves that can take solid food might suggest that allosuckling in red deer could be a form of milk theft detrimental to the offspring of the allosucked hind. This is consistent with the general pattern found by Packer et al. (1992) in a wide range of mammal species, where allosuckling is considered to be theft in monotocous species (single offspring per litter). However, further research is needed, particularly assessing the body condition and immune system status of calves suffering different rates of milk theft from their mothers. Acknowledgments This research has been partly funded by European project FEDER no 1FD97-0540. We thank Bernardo Albin ˜ ana, Fulgencio Cebria´ n, Francisco Villar Go´ mez, Tanya Goettinger and Mai Nun ˜ o for their help in data collection, and two anonymous referees for their comments. References Arman, P., Kay, R. N. B., Goodall, E. D. & Sharman, G. A. M. 1974. The composition and yield of milk from captive red deer (Cervus elaphus L.). Journal of Reproduction and Fertility, 37, 67–84. Beal, W. E., Notter, D. R. & Akers, R. M. 1990. Techniques for estimation of milk yield in beef cows and relationships of milk yield to calf weight gain and postpartum reproduction. Journal of Animal Science, 68, 937–943. Birgersson, B. & Ekvall, K. 1994. Suckling time and fawn growth in fallow deer (Dama dama). Journal of Zoology, 232, 641–650. Birgersson, B., Ekvall, K. & Temrin, H. 1991. Allosuckling in fallow deer, Dama dama. Animal Behaviour, 42, 326–327. Boness, D. J. 1990. Fostering behavior in Hawaiian monk seal: is there a reproductive cost? Behavioral Ecology and Sociobiology, 27, 113–122.
LANDETE-CASTILLEJOS ET AL.: MILK CURVES IN DEER
Brelerut, A., Pingard, A. & The´ riez, M. 1990. Le Cerf et son uElevage. Paris: INRA. Cameron, E. Z. 1998. Is suckling behaviour a useful predictor of milk intake? A review. Animal Behaviour, 56, 521–532. Cameron, E. Z., Stafford, K. J., Linklater, W. L. & Veltman, C. J. 1999. Suckling behaviour does not measure milk intake in horses, Equus caballus. Animal Behaviour, 57, 673–678. Clutton-Brock, T. H., Guinness, F. E. & Albon, S. D. 1982. Red Deer: Behaviour and Ecology of Two Sexes. Edinburgh: Edinburgh University Press. Cowie, G. M., Moore, G. H., Fisher, M. W. & Taylor, M. J. 1985. Calving behaviour of farmed red deer. In: Deer Branch Course no 2 (Ed. by N. S. Beatson & P. R. Wilson), pp. 143–154. Ashburton: New Zealand Veterinary Association. DelGiudice, G. D., Mech, D. L., Paul, W. J. & Karns, P. K. 1986. Effects on fawn survival of multiple immobilizations of captive pregnant white-tailed deer. Journal of Wildlife Diseases, 22, 245–248. Draper, N. R. & Smith, H. 1998. Applied Regression Analysis. 3rd edn. New York: J. Wiley. Ekvall, K. 1998. Effects of social organization, age and aggressive behaviour on allosuckling in wild fallow deer. Animal Behaviour, 56, 695–703. Garci´a, A., Landete-Castillejos, T., Molina, A., Albin ˜ ana, B., Ferna´ ndez, C., Garde, J. & Gallego, L. 1999. Lactation curves in captive Iberian red-deer (Cervus elaphus hispanicus). Journal of Animal Science, 77, 3150–3155. Gauthier, D. & Barrette, C. 1985. Suckling and weaning in captive white-tailed and fallow deer. Behaviour, 94, 128–149. Hoogland, J. L., Tamarin, R. H. & Levy, C. K. 1989. Comunal nursing in prairie dogs. Behavioral Ecology and Sociobiology, 24, 91–95. Kelly, R. W. & Drew, K. R. 1976. Shelter seeking and suckling behaviour of the red deer calf (Cervus elaphus) in a farmed situation. Applied Animal Ethology, 2, 101–111. Landete-Castillejos, T., Garcia, A., Molina, P., Vergara, H., Garde, J. & Gallego, L. In press. Milk production and composition in captive Iberian red deer (Cervus elaphus hispanicus): effect of birth date. Journal of Animal Science. Le Boeuf, B. J. & Briggs, K. T. 1977. The cost of living in a seal harem. Mammalia, 1, 167–195.
Lee, P. C., Majluf, P. & Gordon, I. J. 1991. Growth, weaning and maternal investment from a comparative perspective. Journal of Zoology, 225, 99–114. Loudon, A. S. I., McNeilly, A. S. & Milne, J. A. 1983. Nutrition and lactational control of fertility in red deer. Nature, 302, 145–147. Loudon, A. S. I., Darroch, A. D. & Milne, J. A. 1984. The lactation performance of red deer on hill and improved species pastures. Journal of Agricultural Science, Cambridge, 102, 149–158. Meddis, R. 1984. Statistics Using Ranks. Oxford: Blackwell. Murphey, R. M., Paranhos da Costa, M. J. R., Gomes da Silva, R. & de Souza, R. C. 1995. Allonursing in river buffalo, Bubalus bubalis: nepotism, incompetence or thievery? Animal Behaviour, 49, 1611–1616. Packer, C., Lewis, S. & Pusey, A. 1992. A comparative analysis of non-offspring nursing. Animal Behaviour, 43, 265–282. Pelabon, C., Yoccoz, N. G., Ropert-Coudert, Y., Caron, M. & Peirera, V. 1998. Suckling and allosuckling in captive fallow deer (Dama dama, Cervidae). Ethology, 104, 75–86. Plumb, D. C. 1999. Veterinary Drug Handbook. 3rd edn. (Pocket edn). Ames, Iowa: Iowa State University Press. Pusey, A. E. & Packer, C. 1994. Non-offspring nursing in social carnivores: minimizing the costs. Behavioral Ecology, 5, 362–374. Robbins, C. T., Oftedal, O. T. & O’rourke, K. I. 1987. Lactation, early nutrition, and hand-rearing of wild ungulates, with special reference to deer. In: Biology and Management of the Cervidae (Ed. by C. M. Wemmer), pp. 429–442. Washington, DC: Smithsonian Institution Press. Sadleir, R. M. F. S. 1980. Milk yield of black-tailed deer. Journal of Wildlife Management, 44, 472–478. Serchand, L., McNew, R. W., Kellogg, D. W. & Johnson, Z. B. 1995. Selection of a mathematical model to generate lactation curves using daily milk yields of Holstein cows. Journal of Dairy Science, 78, 2507–2513. Statgraphics. 1997. Statgraphics Plus, Version 3.1, User Manual. Rockville, Maryland: Manugistics. Tiplady, B. A. 1990. Multiple nursing in free-living muskoxen (Ovibos moschatus). Canadian Field Naturalist, 104, 450–454. Wood, P. D. P. 1967. Algebraic model of the lactation curve in cattle. Nature, 216, 164–165.
687
Milk intake and production curves and allosuckling in captive Iberian red deer, Cervus elaphus hispanicus T. LANDETE-CASTILLEJOS*, A. GARCIA u †, J. GARDE†‡ & L. GALLEGO†
*Seccio´n de Recursos Cinege´ticos, IDR and †Departamento de Ciencia y Tecnologi´a Agroforestal, ETSIA, Universidad de Castilla-La Mancha ‡Instituto de Investigacio´n en Recursos Cinege´ticos (IREC), CSIC-UCLM, Spain (Received 15 October 1999; initial acceptance 30 November 1999; final acceptance 2 June 2000; MS. number: 6384R)
In two experiments, we compared milk intake (assessed by weighing calves before and after suckling) and milk production (by hand milking hinds) in Iberian red deer both when calves sucked their mothers together (group-suckling experiment) and when mother and offspring were isolated (isolation-suckling experiment). In both experiments, the general lactation curve for calves increased to a peak and then decreased (type I, standard lactation curve in mammals), whereas the curve for hinds decreased from the start (type II). However, in the experiment on group suckling, calves ingested 17.2% more milk than that produced by their mothers from weeks 6 to 20. In both isolation- and group-suckling experiments, hinds showed an overproduction of milk decreasing from weeks 1 to 5. This decreasing overproduction coincides with a similar trend in calf mortality reported in the literature and might thus be aimed at ensuring calves have sufficient nutrients when mortality is highest. In addition, allosuckling observations in the group-suckling experiment showed an inverse relationship between milk production and percentage of allosucking attempts. Allosucking attempts were also more frequent after the milk overproduction period. Both findings suggest that allosuckling is a response to compensate for a reduced maternal milk supply.
of recorded data and the fitting of such data to mathematical models (reviewed by Serchand et al. 1995) produce a lactation curve. In the standard lactation curve in mammals (type I, Arman et al. 1974; Loudon et al. 1983; Robbins et al. 1987) the amount of milk increases with the increasing energetic requirements of the offspring, peaks when the mother can no longer provide all the energy required by the calf and then starts a slow decrease as milk is substituted by solid food (Lee et al. 1991). In addition, curves decreasing continuously from the start of lactation have been found in red deer (termed type II, Arman et al. 1974; Sadleir 1980; Loudon et al. 1983). These have been attributed to a poorer diet (Loudon et al. 1983, 1984); however, both type I and II lactation curves have been found in red deer fed the same diet ad libitum (Garci´a et al. 1999). Some of the studies based on the frequency and time of suckling have reported allosuckling or suckling between a female and an offspring different to hers. Allosuckling has been reported in 74 species in the majority of mammalian taxa (reviewed by Packer et al. 1992) and has been considered either a maladaptive behaviour arising from crowded conditions in captivity or adaptive behaviour in natural conditions.
Observations on the frequency and length of suckling have been commonly used as estimates of nutrient transfer from mother to offspring during lactation in behavioural research (Clutton-Brock et al. 1982; Gauthier & Barrette 1985; Birgersson et al. 1991; Ekvall 1998). However, the literature shows little evidence for a relationship between length of suckling bouts and milk transferred (reviewed by Cameron 1998). In some instances, for example fallow deer, Dama dama, no relationship has been found between time spent suckling and fawn growth (Birgersson & Ekvall 1994), and isotope labelling has shown no relationship between time or frequency of suckling and milk transfer in horses, Equus caballus (Cameron et al. 1999). The most common methods used to estimate milk transfer directly are milking the mothers (which estimates the ceiling of maternal production; Robbins et al. 1987; Beal et al. 1990; Landete-Castillejos et al., in press) and weighing calves before and after suckling (termed double weighing, an intake estimate; Loudon et al. 1983, 1984; Garci´a et al. 1999). Both the plotting Correspondence: T. Landete-Castillejos, Seccio´n de Recursos Cinege´ticos, IDR, Universidad de Castilla-La Mancha, 02071 Albacete, Spain (email: [email protected]). 0003–3472/00/110679+09 $35.00/0
2000 The Association for the Study of Animal Behaviour
679
2000 The Association for the Study of Animal Behaviour
680
ANIMAL BEHAVIOUR, 60, 5
Crowding may result in misdirected behaviour by the mother (Boness 1990; Tiplady 1990), or opportunistic thefts by offspring (Murphey et al. 1995) or prevent the mother repelling allosuckling attempts (Le Boeuf & Briggs 1977). Under natural conditions, allosuckling has been explained as an altruistic behaviour, either reciprocal (Pusey & Packer 1994), directed to kin (Ekvall 1998) or to defend offspring communally from infanticide or predation (Pusey & Packer 1994). In Scottish red deer, Clutton-Brock et al. (1982) found that allosuckling frequency was very low or nonexistent, whereas Cowie et al. (1985) found allosuckling in the majority of hinds whose calves had been removed. In contrast to red deer, and despite allosuckling being attributed to crowding in captivity in fallow deer, the frequency of allosuckling in free-ranging fallow deer (Ekvall 1998) is similar to that in captivity (Birgersson et al. 1991; Birgersson & Ekvall 1994). In addition, Ekvall (1998) found that the proportion of allosuckling bouts increased with the stage of lactation. Although milk transfer in allosuckling has been confirmed in the field by radioisotope labelling (Hoogland et al. 1989), no study has estimated the amount of milk transferred. In this study we compared the milk production of red deer hinds with the milk intake of their calves both when calves sucked freely in their social group and when they sucked their mothers in isolation. Thus, we aimed to assess milk over- or underproduction relative to calf intake from both their mother (isolation-suckling experiment) and including intake from other hinds (group-suckling experiment). We also investigated the relationship between these variables and allosuckling observations during lactation. METHODS Twenty-two Iberian red deer hinds and their 21 calves were housed in an outdoor enclosure (5000 m2) under the conditions described in Garci´a et al. (1999). Both during gestation and throughout lactation, hinds were fed diets based on suggestions by Brelerut et al. (1990): barley straw and hay from barley, alfalfa, oat and sweet beetroot (16% protein). Deer had ad libitum access to food. Fifteen days after birth, calves had free access to a creep feeder containing an 18% CP concentrate diet. We used 14 hinds and 13 calves in the first experiment comparing the milking estimates with those of weighing the calf before and after suckling (henceforth termed double weighing) when all calves were released together for suckling (henceforth termed group-suckling experiment). We used eight hinds and their calves in the second experiment comparing milking and double-weighing estimates carried out by isolating mother–calf dyads (henceforth termed isolation-suckling experiment).
Group-suckling Experiment Two hinds were 2 years old and the other 12 were 5 years old. One of the hinds lost her calf at 7 weeks but continued producing milk so her milking data were used
in the analysis. Calving took place between 4 June and 12 August 1998. Mean hind weight in the week after calvingSD was 94.13.34 kg.
Milking Beginning 2 weeks after calving, we milked hinds every 4 weeks up to week 34. Weaning was enforced simultaneously in all hinds to keep social conditions constant throughout the experiment. Thus, every hind remained in the group for at least 34 weeks. Before milking, we separated hinds from their calves for 6 h (0800–1400 hours) in a deer-handling facility. Individuals were hand milked under anaesthesia until the four quarters of the udder were totally emptied, which took 10–60 min. Daily milk production was considered to be four times the amount collected in each trial. To reduce health risks from anaesthesia, a low-dose combination of xylazine (0.5 mg/kg live weight) and ketamine (1 mg/kg) was injected into the jugular vein. After inducing anaesthesia, we injected 10 IU of oxytocin into the right jugular vein 1 min before milking was started. After milking, we reversed anaesthesia with a yohimbine injection (0.25 mg/kg live weight). Hinds were weighed weekly on an electronic balance to within 50 g.
Weighing the calf before and after suckling Double-weighing estimates were made weekly after a 6-h isolation period, in a similar way to that described in the literature (Arman et al. 1974; Loudon et al. 1983; Garci´a et al. 1999). Calves remained together indoors until the start of each double-weighing procedure to reduce handling stress, whereas hinds remained outdoors. Prior to and after suckling, calves were weighed in a restraining box placed on an electronic balance to within 5 g. All calves were released simultaneously outdoors for 15 min. Hinds and their calves were fitted with matching colour collars for identification. Two observers recorded the identity of calves and hinds performing successful suckling bouts (contact with the udder for longer than 1 s). Suckling occurred very quickly and took only 1–2 min, with a few suckling bouts during the rest of the suckling period. As a result, we focused observations on successful attempts and did not record rejections in sufficient detail to used them in statistical analysis. Because calves often suck hinds in a lactation stage different to the age of the alien calf, we used the term ‘allosucking bout’ as the suckling bout timed according to the age of the calf, and the term ‘allonursing bout’ as that timed according to the lactation stage of the hind (i.e. the number of allosucking and allonursing bouts coincide, but may show a different temporal pattern). To reduce stress in the calves only one daily double-weighing measure was made (see Ethical note). We obtained daily milk production in both cases by multiplying the one daily record by 4 and fitting the resulting data to Wood’s (1967) gamma function as described by Garci´a et al. (1999). In the double-weighing procedure, differences of less than 50 g were recorded as blanks. Data were analysed for the first 19 weeks (no suckling was observed in any calf after this date).
LANDETE-CASTILLEJOS ET AL.: MILK CURVES IN DEER
Isolation-suckling Experiment Eight hinds, four 6 years and four 3 years old, and their calves were used in this experiment. Mean hind weight after calvingSD was 97.13.2 kg. Births took place between 6 and 31 of May 1999. Because two calves never sucked their mothers during this experiment, we used data from only six hinds. We carried out milking similarly to that of the groupsuckling experiment. However, because hinds had shown a good tolerance to the milking under anaesthesia the previous year, and to increase the definition of the fit for the first few weeks (when milk production changes more quickly), we milked them weekly for the first 6 weeks and then once every 4 weeks until week 18. Double weighing was done as in the previous year, but mother–calf dyads were isolated together for 15 min inside the handling premises. As in the previous year, and to reduce stress, calves were housed together. Double weighing (as milking) was done for 18 weeks.
Statistical Analysis We computed milk intake and production curves by fitting the logarithmically transformed data to the regression line derived from the gamma function (Wood 1967): Yt =atb e −ct; where Yt is the yield at week t. The parameters a, b and c thus estimated were used to obtain both the lactation curves and the estimated daily production. For curves of type I b was positive (Garci´a et al. 1999), whereas for those of type II b was negative or not statistically significant. We estimated the general model for milk intake (double weighing) and yield (milking records) curves by pooling data to compute an averaged curve of each type of method. Because R2 does not discriminate between lack of fit and pure error in regressions involving repeated measures, in the general models, R2/R2max was computed to show the percentage of variation explained that was not due to pure error (Draper & Smith 1998). We compared intake and production curves both for the general model and in sections in which differences were expected (i.e. periods of overproduction or overintake). The first comparison did not require a statistical test because it involved a different model (see Table 1). For intake and yield curves on sections showing overintake or overproduction we fitted the logarithmically transformed data to the logarithm of week, as the function Yt =atb accurately describes both the rising section of a lactation curve and its decreasing phase separately. Despite hind overproduction lasting up to week 5 in the group-suckling experiment, we compared weeks 1–6 because the first two milking trials had been conducted on weeks 2 and 6. If the test comparing regression lines showed significant differences between them, we computed the percentage of milk overproduced or that ingested above the mother’s production from the general model. Similar comparisons were made between pairs of mother–offspring curves that were marginally significant (P<0.10, always in intake curves). We used the
Statgraphics package for these analyses (Statgraphics 1997). Management was designed to keep calf mortality in our premises as low as possible. Thus it was unlikely to resemble mortality in a free-ranging population, and the sample size was too small to be used for statistical comparisons. In consequence, we compared published research on natural rates of calf mortality during lactation (Clutton-Brock et al. 1982) with the ratio milk yield:milk ingested for the period of milk overproduction found in both experiments (5 weeks). The test was conducted with a Pearson correlation. We used one-way ANOVAs to assess differences in the total milk production of the hinds during lactation (i.e. assessed by milking), the mean weight of the hinds during this period and their weight loss (difference between weight at the end of lactation and that after calving), and the weight of the calf within 24 h of birth between individuals classified with regard to the type of curve (classified in two groups I or II) both for milking and double-weighing curves, or the existence of a period when intake exceeded production. For frequency of suckling bouts, both those initiated by the calf (number of bouts observed at every age of calves) and those received by the mother (bouts observed on each lactation week of hinds), we used Pearson correlations. In contrast, we compared the percentage of allosucking and allonursing bouts (number of suckling bouts from alien hinds or calves versus number of total bouts observed per trial) with a test comparing the regression line fitting each variable (Statgraphics 1997). Differences in the percentage of allosucking or allonursing bouts between the period of milk overproduction and the rest of lactation were assessed by one-way nonparametric ANOVAs (Meddis 1984). Data shown are records observed, not daily or weekly estimates.
Ethical Note We followed Spanish and European guidelines and laws in the use of animals in research.
Separation of calves from their mothers The period of mother–calf separation was based on published research (Arman et al. 1974; Loudon et al. 1983), which in turn was based on suckling frequency of this species (Kelly & Drew 1976; Loudon et al. 1983). Calves were always kept with other calves for the separation period (even in the isolation-suckling experiment) to reduce stress. They were kept indoors rather than outdoors because they showed signs of stress (including weight loss) when chased to the handling premises for the first weighing. These symptoms were alleviated when they were driven together with their mothers after suckling for the second weighing. Calves gave distress calls, particularly at the beginning of the experiment when only a few calves were born. Later on, experience of older calves and a larger number of individuals resulted in fewer such calls. Noise was kept to a minimum to reduce stress. Hinds were also kept together outdoors to minimize stress. Nevertheless, hinds were more habituated
681
682
ANIMAL BEHAVIOUR, 60, 5
than calves to handling and showed very few signs of stress. Also, they did not show apparent signs of rejection of their calves, and most rejections observed appeared in older calves after they had already sucked the hinds once or more. No calf stopped growing during the experiment.
Handling during double weighing Every effort was made to minimize stress during handling of calves at weighing. This was easy when calves were small, which produced very reliable measures. As they grew, calves became more restless, occasionally jumping to avoid being caught, and we often had to take them in our arms to reduce stress when driving them to the balance.
Anaesthesia We chose the combined action of xylazine and ketamine as anaesthetic because the synergism between them allowed a much lower dose to be effective in inducing insensitivity and also because it allowed a quicker recovery. This mixture has been widely used for immobilization in a wide range of species, including pregnant white-tailed does, Odocoileus virginianus, that were anaesthetized 10 times in a 6-month period with ‘no effect on fawning date or survival’ (DelGiudice et al. 1986, page 247). Lactating hinds have been subjected to anaesthesia more often than in our study: up to 17 times in 300 days, or once in 2 weeks in Arman et al. (1974) without apparently affecting calves. Anaesthetized individuals lay down after 50 s. After the injection of yohimbine they got up within 2 min. The majority could walk, but some lay down again. They appeared to be disoriented at first, and tongue control (the tongue of the animals was stretched out of the mouth to drain saliva and prevent choking when lying down) was not evident for 1–5 min after they stood upright. A few minutes (5–10 min) after getting up, hinds showed standard signs of ruminating. We left them inside the handling premises to avoid other hinds attacking them. After the recovery period, hinds appeared to behave normally. Hinds were anaesthetized for 15–60 min (mean ca. 30 min). They needed a longer time under anaesthesia at the beginning of lactation because milk production was greater and also because habituation to milking and frequent handling during lactation reduced resistance to anaesthesia.
Possibility of the anaesthetics passing to the milk Ketamine does not pass into the milk and is eliminated in the kidney. Xylazine could pass into the milk, although it has not been detected at 5 and 21 h postanaesthesia in dairy cows (Plumb 1999). However, this possibility appears to be very small because hinds were milked dry and it would take them some hours to fill the udder. Moreover, as mentioned above, reports of anaesthesia of pregnant hinds did not find any effect on calves at such a sensitive stage in their life.
Standardization of the filling state of the udder Double-weighing techniques in cattle require a standardized filling state of the udder (i.e. the weighing is
done twice so that the second and valid record starts with an empty udder; Beal et al. 1990). However, despite this, and despite research in cattle showing that several consecutive measures are needed for the double weighing to reach milking precision we weighed calves only once a week to reduce isolation stress (Garci´a et al. 1999). This resulted in more variable double-weighing estimates than milking ones (see Results) and so we had to use marginal significance (P<0.1) values to compare mother–offspring dyads of curves. Similarly, the most advisable protocol for milking from a technical point would be either to milk the udder empty or to allow the calves to suck their mothers at the start of the isolation period. However, that would mean anaesthetizing the hinds twice, or separating the calves from their mothers overnight in addition to the following 6-h experimental period, both of which were ethically unacceptable. To disrupt lactation as little as possible, milking and double-weighing estimates in the same week were done 3–4 days apart. RESULTS Milk production estimated by double weighing was similar to that obtained by milking when individuals sucked together (193.1 versus 193.6 litres). However, double weighing underestimated milking when calves sucked their mothers in isolation (143.9 versus 200.6 litres). The first method was more variable than milking and so, only three out of 13 curves from individuals reached marginal significance (P<0.10), whereas 13 out of 14 milking curves reached significance (P<0.05). In the isolationsuckling experiment only one of six double-weighing curves of individual calves was significant, whereas four of six milking curves were. Such variability makes comparisons and assessment of effects of any factor more difficult. The percentage of variability explained by the model also reflects this pattern: thus, for calves in the group-suckling experiment, R2/R2max of double weighing was 80.88%, and the mean R2 of individual curves was 32.684.51%; whereas for milking curves these values were R2/R2max =99.23% and that for individual curves 87.493.49%. In a similar way, the R2/R2max for the general model in the isolation-suckling experiment was 77.53% for the double-weighing curve and 98.24% for the milking one. The curve type of the general model estimated by double weighing was also different to that estimated by milking hinds in both treatments (Table 1): the calf intake curve was a standard mammal lactation curve (type I), whereas hinds had a curve of type II, both in the group- and isolation-suckling experiments (Fig. 1; general model for double weighing: group suckling: Yt =942.6 t0.5682 e 0.1021, P=0.007; isolation suckling: Yt = 1026.6 t0.6062 e 0.1205; P=0.016; general model for milking curves: group suckling: Yt =1994.0e 0.0612; P<0.001; isolation suckling: Yt =2334.8 e 0.0431; P<0.001; see Table 1 for model details). This pattern was also found in the frequency of each type of individual curve: for the groupsuckling experiment 61.5% of calves had curves of type I, whereas 64.3% of hinds had curves of type II; in the isolation-suckling experiment all the calves had intake
LANDETE-CASTILLEJOS ET AL.: MILK CURVES IN DEER
Table 1. Parameters and their confidence interval at 95% probability level for lactation curves estimated by milking and double weighing both when calves suckled their mothers in a group, and when calves suckled their mother in isolation
Curve
Group suckling Milking*
R2/R2max (model) (%)
P (model)
F
99.23
0.001
129.47
Parameter
ln a b c
Double weighing
Isolation suckling Milking*
80.88
98.24
0.007
0.001
5.23
56.78
77.53
0.016
4.45
−0.0144 −0.0612
ln a
6.8486
b
0.5682
c
−0.1021
ln a b c
Double weighing
7.5979
7.7569 −0.0897 −0.0431
ln a
6.9340
b
0.6062
c
−0.1205
95% confidence interval
7.3785 7.8173 — −0.0718 −0.0506 6.5190 7.1783 0.0181 0.9553 −0.1649 −0.0394 7.6577 7.8616 — −0.0546 −0.0316 6.5238 7.3442 0.0899 1.1225 −0.2055 −0.0355
P
0.001 >0.1 0.001 0.001 0.004 0.002 0.001 >0.1 0.001 0.001 0.022 0.006
The model fitted corresponds to Wood’s (1967) gamma function: yield Y at week t, Y(t)=a tb e −ct. Because the function is fitted as a regression line on logarithmically transformed data and the antilog transformation introduces a bias in the standard error and confidence intervals, a natural logarithm of parameter ‘a’ along with its confidence intervals are given instead of their back-transformed value. R2/R2max indicates, for regression involving repeated measures, the percentage of variability explained by the model from that ‘explainable’ (ie. not due to pure error of repeating measures). *Values of parameters correspond to final model, which did not include ‘b’ as P>0.1.
curves of type I whereas 50% of the hinds had curves of type II. However, curves of mother–offspring dyads showed a weaker relationship: 38.5% of dyads in the group-suckling experiment had intake curves of type I and production curves of type II, 23.1% had both curves of type I or II, and 15.4% had intake curves of type II and production curves of type I; in the isolation-suckling experiment 50% of dyads had only type I curves and 50% had intake curves of type I and production curves of type II. The type of intake and production curves in the groupsuckling experiment did not seem to depend on the total amount of milk produced during lactation, the weight or weight loss of the hind, or calf weight at birth. Thus, there were no differences between hinds with curves of type I and those with curves of type II in milk production (t12 = 0.473, P>0.1), mean weight during lactation (t12 = 0.325, P>0.1), or percentage of body weight lost (t12 =0.687, P>0.1). Calf birth weight did not vary with the type of lactation curve of the mother (t12 =0.510, P>0.1), or of the calf (t11 =0.50, P>0.1). The type of curve of the calves did not affect the mean weight of their mothers (t11 = 0.150, P>0.1), or their percentage of body weight lost during lactation (t11 =1.434, P>0.1). Our data
are not adequate to assess the effect of calf sex on every variable. In the group-suckling experiment intake apparently exceeded hind production from week 6 onwards (Fig. 1; the slight overproduction after week 27 is meaningless because intake data were not collected at this time), whereas production exceeded intake in weeks 1–5. The production curve was significantly different from the intake one for weeks 1–6 (overproduction period: intercepts: F1,3 =6.39, P=0.014; slopes: F1,3 =5.54, P=0.021) and from week 6 onwards (overintake period: intercepts: F1,3 =4.96, P=0.027; slopes: F1,3 =2.32, P>0.1). Similar results were obtained in the isolation-suckling experiment, where, again, regression lines for intake and production curves differed from weeks 1 to 5 (overproduction period: intercepts: F1,3 =17.18, P<0.001; slopes: F1,3 =5.67, P=0.022), whereas they differed only in intercepts but not slopes from week 6 onwards (intercepts: F1,3 =5.08, P=0.028; slopes: F1,3 =0.35, P>0.1). However, after the peak in intake in this case, intake underestimated milking records. The results in the group-suckling experiment might have been an artefact of unbalanced double-weighing data. However, this seems unlikely because intake and
683
ANIMAL BEHAVIOUR, 60, 5
3000 (a ) 2500 2000 1500 1000 500 Milk (ml)
684
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33
3000 (b) 2500 2000 1500 1000 500 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Week
Figure 1. General Wood’s (1967) model (fitted to pooled data from all individuals, which are also shown in the figure) for intake curves estimated by weighing red deer calves before and after suckling ( , ——) and those of milk production estimated by hand milking of hinds ( , – – –). (a) Calves suckled their mothers together; (b) each calf was isolated with its mother.
production curves in mother–offspring pairs also differed from each other in both individuals where both curves were significant and intake exceeded production in the midde of lactation (calf 2, second born: intake curve: R2 =44.38%, F2,10 =3.99, P=0.053; model: Yt =1285.9 t0.8180 e 0.1456 t, all coefficients significant at P<0.05; production curve: R2 =84.54%, F1,7 =38.28, P<0.001; model: Yt =2359.7 e 0.0516 t, both coefficients significant at P<0.001, but coefficient b= 0.1631 discarded at P>0.1; calf 7, seventh born: intake curve: R2 =58.45%, F2,8 =5.63, P=0.03; model: Yt =775.2 t1.3779 e 0.2320 t, all coefficients significant at P<0.05; production curve: R2 =88.85%, F1,7 =47.81, P<0.001; model: Yt =3242.8 e 0.1089 t, both coefficients significant at P<0.001, but coefficient b=0.4821 discarded at P>0.1). The comparison between intake and production curves in periods of overintake or overproduction did not yield any significant result, probably because of the small sample size (intercept of both curves in calf 2 occurred at week 2–3, which included only one milking record; the rest of the segments could be compared). Total milk production estimated from the general model in the group-suckling experiment was 58.3 litres for the 5
weeks where production exceeded intake. For this period the estimated intake was 42.5 litres. Thus, hinds produced 44.0620.1% (XSE; N=5) more milk than their calves ingested. From week 6 to 28 (overintake period judged from the intersection of the curves), hinds produced 123.0 litres, whereas calves ingested 139.1 litres. Thus, calves ingested 12.441.31% (N=22) more than their mothers were producing every day. For weeks 6–20 (where we have data for both curves), calf intake was 17.520.5% (N=14) greater than hind production. In the isolation-suckling experiment hinds produced 72.0 litres during weeks 1–5, whereas calves sucked 45.4 litres. Thus, hinds produced 65.9321.65% more than their calves ingested. Published data on the percentage of calf deaths for the first 5 weeks in free-ranging deer showed a significant correlation with the overproduction both in the groupsuckling experiment (r3 =0.938, P=0.019) and in the isolation-suckling one (r3 =0.950, P=0.013). If all individual curves from the group-suckling experiment are considered (significant and nonsignificant), six calves (42.86%) showed periods when they sucked more milk than their mothers produced (two were males). Of these, four had a curve of type I, whereas three of their mothers had a curve of type II. The hinds whose calves sucked other hinds (assessed by comparing lactation curves) did not differ from the others in their total milk production (t11 = 0.025, P>0.1), weight (t11 =0.191, P>0.1), percentage of body weight lost during lactation (t11 =1.298, P>0.1), or birth weights of their calves (t11 =0.903, P>0.1). The type of curve did not affect the frequency of allosucking bouts observed (21 =0.04, P>0.1): 29.4% occurred between dyads with curves of type I, 37.3% between calves with curves of type I allosuckling hinds of type II, 15.7% between calves of type II and hinds of type I, and 17.6% between dyads with curves of type II.
Allosuckling Observations We observed 37.80% allosucking bouts (number of allosucking bouts:total number of suckling ones). These occurred both when the offspring of the hind was already sucking and when the alien calf was sucking alone. The number of suckling bouts decreased with the stage of lactation (suckling bouts: r17 = 0.724, P<0.001; nursing bouts: r14 = 0.568, P=0.022; Fig. 2). In contrast, allosucking and allonursing bouts constituted an increasing percentage of suckling bouts as lactation proceeded (allosucking bouts: r17 =0.810, P<0.001; allonursing bouts: r12 =0.786, P=0.001; Fig. 3). Calves allosucked more often after the 5-week milk overproduction period than before (XSE of the percentage of daily allosuckling during the first 5 weeks=22.453.67%, that for weeks 5–20, 61.517.46%; H1 =8.871, P<0.01). This difference was also significant, although smaller, for hinds (allonursing bouts: 28.166.13 versus 44.099.15%; H1 =4.312, P<0.05). Although this result suggests a difference between trends of allosuckling and allonursing, the regression lines were not statistically significant (intercepts: F1,3 =1.61, P>0.1; slopes: F1,3 =1.39, P>0.1).
LANDETE-CASTILLEJOS ET AL.: MILK CURVES IN DEER
DISCUSSION
20 18
Suckling bouts
16 14 12 10 8 6 4 2 0
5
10 Week
15
20
Figure 2. Daily frequency of total suckling bouts according to the stage of lactation of the calf (x) and the hind (C). The types of suckling bout do not coincide because allosucking is timed differently according to whether the age of the alien calf or the lactation stage of the hind is considered.
100
Allosuckling/allonursing bouts (%)
90 80 70 60 50 40 30 20 10 0
5
10 Week
15
20
Figure 3. Percentage of allosucking bouts (x) and allonursing bouts (C) during lactation. Fitted lines correspond to the following models: allosucking: Y=8.2888+4.2944t; allonursing: Y=13.6267+2.8126t. Lines did not differ significantly in slopes or intercepts.
The percentage of allosucking bouts recorded on each trial correlated inversely with the estimated milk production of the hind for that week (r17 = 0.790, P<0.001). In contrast, the percentage of allosucking bouts by each calf did not correlate with its mother’s total milk production (r9 =0.288, P>0.1), her weight (r9 =0.389, P>0.1) or the percentage of body weight she lost during lactation (r9 =0.321, P>0.1), although there was a positive correlation between allosucking bouts and calf birth weight (r9 =0.692, P=0.018). Conversely, the percentage of allonursing bouts increased with increasing milk production by the nursing hind (r9 =0.619, P=0.042), but there was no correlation with her body weight (r9 =0.549, P=0.08) or the percentage of weight she lost during lactation (r9 =0.212, P>0.1).
Our results show that double weighing and milking estimate processes in lactation that differ from each other. Reports in the literature, however, consider them as alternative methods for estimating milk production. Indeed, double weighing has been used more often than milking because it is easier to do, although it is more variable (Beal et al. 1990; Garci´a et al. 1999; this study). If, as considered in the literature, both methods estimate the same process they should produce the same type of curve. However, both in the general model and in more than half of the mother–offspring dyads (52.6%) the type of intake curve was different to that for production. It seems unlikely that the pattern of lactation curves in the general models was due to the particular conditions in which the experiment was conducted: this pattern was repeated the following year with different individuals differing in age; intake curves were of type I both when individuals sucked their mothers in isolation and when they did so together with other calves; and hinds had a curve of type II both when trials were conducted once a month (which might have given insufficient detail to detect the peak characteristic of curves of type I) and when trials were carried out weekly for the first 6 weeks to increase curve definition. In the literature, in contrast, the standard lactation curve for all mammals is considered a single one (type I) and intake and production curves are not distinguished (Lee et al. 1991). It is no surprise that calves had a general curve of type I, as their energetic requirements rise with increasing body weight until the slow process of weaning is enforced by the mother. It is more striking, though, that hinds had a general curve of type II and overproduced milk during the first 5 weeks, particularly considering the high cost of lactation in red deer (e.g. lactating hinds suffer a higher mortality rate than barren ones, Clutton-Brock et al. 1982). The milk overproduction might ensure a sufficient flow of energy for the calf during the first few weeks of life, when both growth rate and mortality are greatest. This is more strongly suggested by the significant correlation between increasing percentage of summer deaths reported in this period and increasing milk overproduction. However, these results should be taken with great care, as the test compares different sets of animals (summer deaths recorded by Clutton-Brock et al. 1982 under free-ranging conditions) and at a different time. Other species such as fallow deer also show milk production greater than intake during their first few weeks of life, as calves terminate 40% of the suckling bouts during their first week of life (suggesting they feed to satiation), but their mother terminates all the observed suckling bouts from the second week onwards (Birgersson & Ekvall 1994). Unfortunately, as our deer are managed to keep them in good condition with low mortality, our data are inadequate to assess whether calf mortality is higher for mothers producing the smallest amount of milk during the first few weeks. Nevertheless, indirect results suggest that this might be so under natural conditions, as factors increasing the mortality of calves under natural conditions, such as an increasing calving delay (Clutton-Brock
685
686
ANIMAL BEHAVIOUR, 60, 5
et al., 1982), also reduce the amount of milk produced (Landete-Castillejos et al., in press). However, our results are not conclusive and further experiments are needed to discriminate between these and alternative/ complementary hypotheses. Our results do not allow us to assess which factor caused the hinds or calves to have one type of curve or the other, as neither the weight, the percentage of body weight lost by the hinds, nor the calf birth weight differed between individuals with either type of curve. It is particularly striking that hinds did not differ in the amount of milk produced as a function of the type of curve displayed, in contrast to Loudon et al.’s (1983) findings. The origin of each type of curve might be due to an unexamined factor, such as dominance rank among hinds or calves. The comparison of intake and production curves also showed a period in which intake was greater than production. This is likely to differ both in length and milk surplus for each individual case (in fact, many dyads of curves did not even intersect at any stage), and it may vary with ecological conditions. However, it can provide a rough estimate of the amount of milk that is sucked from other hinds assuming that the mother’s milk is not, in turn, sucked by other calves. Although such an overintake period could arise from an imbalance in the data, such that calves with greatest intakes contributed more to the general model in the overintake period and those with smallest intakes contributed more in the tails, this does not seem very likely. This is because there was a milk overintake in two mother–offspring dyads where the fit was significant. In addition, sustained allosuckling is suggested by the fact that the hind that lost her calf in week 7 maintained her milk production and did not become dry, presumably because other calves often sucked her. Such incidence of allosuckling has often been considered a consequence of crowded conditions in the experimental set-up. Although it is likely that conditions under which double weighing was carried out might have influenced the amount of milk sucked from other hinds compared with what might happen under natural conditions, it seems unrealistic to assess the amount of milk allosucked in the field with direct methods such as those used here. Alternatively, the amount of milk allosucked might be estimated from observations, but the utility of these is controversial (Cameron 1998), and when compared with the more reliable direct methods, as has been done in horses (Cameron et al. 1999), they have not shown any relationship. Finally, although the incidence of allosuckling we found differs from that reported in other studies of red deer (which in turn, also varies: none in Clutton-Brock et al. 1982; only when the mother is removed in Cowie et al. 1985), studies in other cervid species such as fallow deer have shown that the frequency of allosuckling is similar in captivity (73%, Pelabon et al. 1998; 23–57%, Birgersson et al. 1991; Birgersson & Ekvall 1994) and free-ranging individuals (43%, Ekvall 1998). Moreover, the frequency of allosuckling reported here for Iberian red deer (38%) falls within the range of that found in captivity and in free-ranging fallow deer.
The comparison between hind milk production and observed allosucking frequency shows that the percentage of allosucking bouts increases as the milk production of the mothers decreases. This suggests that calves may allosuck as an attempt to maintain the flow of milk intake despite the reduction enforced by their mothers. The comparison of allosucking and allonursing percentages before and after the milk overproduction period shows a smaller difference for hinds than for calves. This might suggest that the timing of allosucking differs from that of allonursing and, therefore, that hinds may usually receive a high proportion of sucking attempts by calves older than their own, although the comparison of regression lines belonging to allosucking and allonursing bouts did not achieve statistical significance. This point deserves further study because such allosucking attempts by older calves of mothers in their early lactation stages, if proven, might be caused by the milk overproduction during the first few weeks of lactation. Pelabon et al. (1998) also found that fallow deer does during the first 2 weeks of lactation were allosucked by older calves, which was considered to be a result of the longer suckling bouts (ad libitum of the calf) at this stage of lactation. Finally, the fact that some of the allosuckling attempts involve the diversion of milk destined for young calves whose only source of energy is milk to older calves that can take solid food might suggest that allosuckling in red deer could be a form of milk theft detrimental to the offspring of the allosucked hind. This is consistent with the general pattern found by Packer et al. (1992) in a wide range of mammal species, where allosuckling is considered to be theft in monotocous species (single offspring per litter). However, further research is needed, particularly assessing the body condition and immune system status of calves suffering different rates of milk theft from their mothers. Acknowledgments This research has been partly funded by European project FEDER no 1FD97-0540. We thank Bernardo Albin ˜ ana, Fulgencio Cebria´ n, Francisco Villar Go´ mez, Tanya Goettinger and Mai Nun ˜ o for their help in data collection, and two anonymous referees for their comments. References Arman, P., Kay, R. N. B., Goodall, E. D. & Sharman, G. A. M. 1974. The composition and yield of milk from captive red deer (Cervus elaphus L.). Journal of Reproduction and Fertility, 37, 67–84. Beal, W. E., Notter, D. R. & Akers, R. M. 1990. Techniques for estimation of milk yield in beef cows and relationships of milk yield to calf weight gain and postpartum reproduction. Journal of Animal Science, 68, 937–943. Birgersson, B. & Ekvall, K. 1994. Suckling time and fawn growth in fallow deer (Dama dama). Journal of Zoology, 232, 641–650. Birgersson, B., Ekvall, K. & Temrin, H. 1991. Allosuckling in fallow deer, Dama dama. Animal Behaviour, 42, 326–327. Boness, D. J. 1990. Fostering behavior in Hawaiian monk seal: is there a reproductive cost? Behavioral Ecology and Sociobiology, 27, 113–122.
LANDETE-CASTILLEJOS ET AL.: MILK CURVES IN DEER
Brelerut, A., Pingard, A. & The´ riez, M. 1990. Le Cerf et son uElevage. Paris: INRA. Cameron, E. Z. 1998. Is suckling behaviour a useful predictor of milk intake? A review. Animal Behaviour, 56, 521–532. Cameron, E. Z., Stafford, K. J., Linklater, W. L. & Veltman, C. J. 1999. Suckling behaviour does not measure milk intake in horses, Equus caballus. Animal Behaviour, 57, 673–678. Clutton-Brock, T. H., Guinness, F. E. & Albon, S. D. 1982. Red Deer: Behaviour and Ecology of Two Sexes. Edinburgh: Edinburgh University Press. Cowie, G. M., Moore, G. H., Fisher, M. W. & Taylor, M. J. 1985. Calving behaviour of farmed red deer. In: Deer Branch Course no 2 (Ed. by N. S. Beatson & P. R. Wilson), pp. 143–154. Ashburton: New Zealand Veterinary Association. DelGiudice, G. D., Mech, D. L., Paul, W. J. & Karns, P. K. 1986. Effects on fawn survival of multiple immobilizations of captive pregnant white-tailed deer. Journal of Wildlife Diseases, 22, 245–248. Draper, N. R. & Smith, H. 1998. Applied Regression Analysis. 3rd edn. New York: J. Wiley. Ekvall, K. 1998. Effects of social organization, age and aggressive behaviour on allosuckling in wild fallow deer. Animal Behaviour, 56, 695–703. Garci´a, A., Landete-Castillejos, T., Molina, A., Albin ˜ ana, B., Ferna´ ndez, C., Garde, J. & Gallego, L. 1999. Lactation curves in captive Iberian red-deer (Cervus elaphus hispanicus). Journal of Animal Science, 77, 3150–3155. Gauthier, D. & Barrette, C. 1985. Suckling and weaning in captive white-tailed and fallow deer. Behaviour, 94, 128–149. Hoogland, J. L., Tamarin, R. H. & Levy, C. K. 1989. Comunal nursing in prairie dogs. Behavioral Ecology and Sociobiology, 24, 91–95. Kelly, R. W. & Drew, K. R. 1976. Shelter seeking and suckling behaviour of the red deer calf (Cervus elaphus) in a farmed situation. Applied Animal Ethology, 2, 101–111. Landete-Castillejos, T., Garcia, A., Molina, P., Vergara, H., Garde, J. & Gallego, L. In press. Milk production and composition in captive Iberian red deer (Cervus elaphus hispanicus): effect of birth date. Journal of Animal Science. Le Boeuf, B. J. & Briggs, K. T. 1977. The cost of living in a seal harem. Mammalia, 1, 167–195.
Lee, P. C., Majluf, P. & Gordon, I. J. 1991. Growth, weaning and maternal investment from a comparative perspective. Journal of Zoology, 225, 99–114. Loudon, A. S. I., McNeilly, A. S. & Milne, J. A. 1983. Nutrition and lactational control of fertility in red deer. Nature, 302, 145–147. Loudon, A. S. I., Darroch, A. D. & Milne, J. A. 1984. The lactation performance of red deer on hill and improved species pastures. Journal of Agricultural Science, Cambridge, 102, 149–158. Meddis, R. 1984. Statistics Using Ranks. Oxford: Blackwell. Murphey, R. M., Paranhos da Costa, M. J. R., Gomes da Silva, R. & de Souza, R. C. 1995. Allonursing in river buffalo, Bubalus bubalis: nepotism, incompetence or thievery? Animal Behaviour, 49, 1611–1616. Packer, C., Lewis, S. & Pusey, A. 1992. A comparative analysis of non-offspring nursing. Animal Behaviour, 43, 265–282. Pelabon, C., Yoccoz, N. G., Ropert-Coudert, Y., Caron, M. & Peirera, V. 1998. Suckling and allosuckling in captive fallow deer (Dama dama, Cervidae). Ethology, 104, 75–86. Plumb, D. C. 1999. Veterinary Drug Handbook. 3rd edn. (Pocket edn). Ames, Iowa: Iowa State University Press. Pusey, A. E. & Packer, C. 1994. Non-offspring nursing in social carnivores: minimizing the costs. Behavioral Ecology, 5, 362–374. Robbins, C. T., Oftedal, O. T. & O’rourke, K. I. 1987. Lactation, early nutrition, and hand-rearing of wild ungulates, with special reference to deer. In: Biology and Management of the Cervidae (Ed. by C. M. Wemmer), pp. 429–442. Washington, DC: Smithsonian Institution Press. Sadleir, R. M. F. S. 1980. Milk yield of black-tailed deer. Journal of Wildlife Management, 44, 472–478. Serchand, L., McNew, R. W., Kellogg, D. W. & Johnson, Z. B. 1995. Selection of a mathematical model to generate lactation curves using daily milk yields of Holstein cows. Journal of Dairy Science, 78, 2507–2513. Statgraphics. 1997. Statgraphics Plus, Version 3.1, User Manual. Rockville, Maryland: Manugistics. Tiplady, B. A. 1990. Multiple nursing in free-living muskoxen (Ovibos moschatus). Canadian Field Naturalist, 104, 450–454. Wood, P. D. P. 1967. Algebraic model of the lactation curve in cattle. Nature, 216, 164–165.
687