Milk Composition and Production in Free-living Allied ... - Helene Marsh

Australian Journal of Zoology, 1996,44,659-74

Milk Composition and Production in Free-living Allied Rock-wallabies, Petrogale assimilis

J. C. ~ e r c h a n t * ,H. MarshB, P. Spence@ and G. ~ e ' a t h ~ *CSIRO Division of Wildlife and Ecology, PO Box 84, Lyneham, ACT 2602, Australia. BDepartmentof Tropical Environment Studies and Geography, James Cook University, Qld 48 11, Australia. C~epartmentof Biomedical and Tropical Veterinary Science, James Cook University, Qld 481 1, Australia.

Abstract Milk composition and the rates of milk consumption by pouch young were examined in free-living allied rock-wallabies, Petrogale assimilis. Milk solids concentrations were approximately 16% (wlw) at 70 days post-partum and increased to about 22% by 170 days when young first left the pouch. By permanent pouch emergence (about 200 days), concentrations had declined and stabilised at about 19%. Milk carbohydrate peaked at 12% (wlv) at 150 days; lipid concentrations averaged 8% (wlw) at 200 days. The subsequent decline in carbohydrates was the main cause of the fall in milk solids. Protein concentrations increased gradually from about 3% (wlv) at 70 days to plateau at 5.5% at about 200 days. Milk consumption rates were measured from 72 to 159 days post-partum with 2 2 ~turnover. a Milk consumption, about 3 mL day-' initially, increased to an average of about 15 mL day-' by 150 days. The mass gained by a pouch young between 72 and 159 days for each millilitre of milk consumed was not correlated with lactational stage and averaged 0.21 + 0.014 (s.e.) g mL-'.

Introduction In marsupials it is important that milk samples are obtained at known stages of the lactational cycle as previous studies have shown that their milk compositions change considerably during lactation (Green 1984; Green and Merchant 1988; Merchant 1989; Munks et al. 1991; Merchant et al. 1994). This requirement makes it difficult to undertake useful studies of milk composition in free-ranging marsupials and consequently most data have been obtained from captive animals (Green and Merchant 1988; Merchant 1989). Sequential milk samples from individual animals have been obtained in the field only from the ringtail possum, Pseudocheirus peregrinus (Munks et al. 1991) and the koala, Phascolarctos cinereus (Krockenberger 1996), while in wild red kangaroos, Macropus rufus, Muths (1996) collected single milk samples from shot or stunned females with young of various ages. Munks et al. (1991) reported that the composition of milk in captive and wild ringtail possums, P. peregrinus, differed and suggested that this may be a consequence of diet. These are the only studies of marsupial milk so far attempted in free-living animals. Detailed studies of milk consumption by macropodid pouch young are limited to captive studies of the tammar wallaby, Macropus eugenii (Green et al. 1988; Dove and Cork 1989) and the brush-tailed bettong, Bettongiapenicillata (Merchant et al. 1994). Petrogale assimilis, the allied rock-wallaby, is a medium-sized macropodid (adults 3.54.5 kg) that lives in the wet-dry tropics of northern Queensland where the climate is characterised by irregular summer rainfall and long dry seasons (Delaney 1993). As do other rock-wallabies, P. assimilis shelters in rock piles during the day and emerges to feed in the surrounding countryside at night. It is an opportunistic herbivore that generally prefers forbs and browse to grass (Horsup and Marsh 1992). Births may occur throughout the year and pouch young first leave the pouch temporarily at 170 days of age and permanently at about 200 days (Delaney 1993). Most pouch young in the present study were weaned by about 270 days.

J. C. Merchant et al.

This study of lactation in free-living allied rock-wallabies, P. assirnilis,was part of a project to determine their seasonal water and energy requirements. This is the first time that both milk composition a n d consumption b y p o u c h y o u n g (= maternal m i l k production) h a v e been determined in a free-living marsupial.

Materials and Methods Animals This research is part of a longitudinal study of a colony of rock-wallabies at 'Black Rock', a sandstone outcrop about 250 km west of Townsville, Australia, conducted by Marsh and her co-workers since 1986. The climate is typical of the wet-dry tropics and the vegetation is predominantly open woodland (Horsup and Marsh 1992). Annual rainfall is extremely variable and averages 731 mm, most of which (70%) falls from December to March. Rock-wallabies were caught in cage traps baited with a mixture of peanut butter, rolled oats and honey. Animals were removed from traps between 2200 and 2300 hours after which the traps were rebaited and rechecked at 0600 hours. All trapped animals were restrained in hessian bags for weighing, examination of pouches and collection of milk samples. Pouch young were removed from their mothers after trapping was completed in the morning. The pouch young were weighed (+ 0.1 g), measured and then placed on damp paper towelling in ventilated containers on a wire mesh stand in a polystyrene insulated box. The space beneath the mesh was filled with warm water and the temperature maintained at 37OC by regular renewal of the warm water. The ages of pouch young were estimated from measurements taken when first found in the pouch, by using growth curves from known-age animals (Delaney and De'ath 1990). Milk Consumption Milk consumption estimations were made during five field trips in July 1990 (mid-dry season), October 1990 (late-dry), April 1991 (late-wet), July 1991 (mid-dry) and October 1991 (late-dry). Milk consumption rates were determined in 18 pouch young by the 2 2 ~turnover a technique of Green and Newgrain (1979). On removal from the pouch, pouch young were injected with 50-100 pL of 2 2 ~ a (200 kBq rnL-I), depending on size. Pouch young were separated from their mothers for 5 h, which allowed accumulation of milk in the mammary glands of the mother and equilibration of 2 2 ~with a the exchangeable sodium pool (ES) of the pouch young. Two blood samples of 2 pL and 1C-20 pL were then taken from the pouch young via the lateral tail vein and collected in Microcap capillary tubes. The pouch young were then returned to the pouch. The upper age limit for estimation of milk consumption by this technique is determined by the age at which pouch young begin to eat solid food and therefore ingest sodium from a source other than milk. In this study the age range over which estimations were made was 72-159 days. The 2-pL blood sample was diluted with 1 mL water in a l&mL Eppendorf tube, sealed and frozen. This dilution allowed estimation of blood sodium concentrations by atomic absorption spectrophotometry (Varian 1000). The second blood sample of 10-20 pL was placed in approximately 200 pL of water in a 5-mL scintillation tube, decolourised with H202, sealed and frozen. These samples were dried in the laboratory and 3 mL of scintillation cocktail (PCS Amersham) added to each sample. Measurement of the specific activity of 2 2 ~was a carried out with a Beckrnan liquid scintillation spectrometer. A final set of blood samples was obtained at the end of the measurement period when the mother was recaptured. Estimations were carried out over periods of 14 days (July 1990 only) or 21 days. Changes in body mass and ES were assumed to be linear during each experimental period and sodium influx was derived from change in specific activity of 2 2 ~ina the blood during the experimentalperiod, as follows: Sodium influx (mmol) = {ln[(Nal x ES,)/(Na2 x ES2)](ES2- ESl)Iln(ES21ESl)]+ (ES, - ES1) where Nal and Na2 are the initial and final specific activities of 2 2 ~ and a ESl and ES, are the initial and final ES pools respectively.

Milk Composition To facilitate milking, some mothers were anaesthetised with sodium methohexital (Brietal Sodium, Eli Lilley and Co. Sydney; 15 mg per kg body mass i.v.) and injected (i.m.) with oxytocin (Syntocinon; Sandoz, Australia. 1 i.u, per kg body mass). Volumes of milk up to about 0.5 mL were obtained dependent on the age of the young. Anaesthetising the mother also allowed the pouch young to be more readily assisted back on to the teat. Most pouch young reattached to the teat without assistance.

Milk Composition and Production in Rock-wallabies

Initial dilutions of milk (5 FL in 1 mL of water) were made before fresh milk samples were sealed and frozen. These dilutions were later used to determine electrolyte concentrations (Na, K, Ca and Mg) with an atomic absorption spectrophotometer (Varian 1000). Total milk solids and carbohydrate (total hexose) were determined by the methods of Green et al. (1980) and Messer and Green (1979) respectively. Total protein content was determined by the Coomassie Blue dye-binding method (Bradford 1976). Nine milk samples provided sufficient milk for analysis by the Kjeldahl procedure and total nitrogen was estimated by a microdiffusion technique (Conway 1962). The dye-binding method produced values that were 112.7 2 20.7 (s.d.)% of those derived by the Kjeldahl method. All protein values presented here have been corrected downwards to account for this discrepancy. A conversion factor of 6.38 was used to convert total nitrogen to estimated protein. After allowing an ash content of 5% (wlw) of milk solids, fat content was estimated by difference because of insufficient volumes of milk in samples. For the same reason an experimental value for milk ash content was not derived. The value of 5% (wlw) for the ash content of the solids fraction of the milk is based on data from Ben Shaul (1962) and a review by Jenness and Sloan (1970). Since marsupial milk composition, and possibly ash content, varies with lactational stage we have chosen to use an intermediate value for ash. The energy content of milk was estimated from its proximate composition by assuming energy values of 39, 24 and 17 kJ g-' for fat, protein and carbohydrate respectively (Kleiber 1975). Statistical Analyses Due to the strong non-linearities in the data, the changes in the composition of rock-wallaby milk with the age of pouch young were modelled by locally weighted regression (loess smoothers: Cleveland 1979; Chambers and Hastie 1992). This technique attempts to predict the value of the response by simple regression on the data in the neighbourhood of each value to be predicted. As the neighbourhood moves through the data, the points near the boundary to one side drop out and are replaced by points on the other side. The low weighting of the boundary points ensures a smooth curve. The span (width of the neighbourhood) of each smoother was chosen by the Akaike Information Criteria and varied from 25 to 75% of the data. The local fits were linear and in all cases were superior to local quadratic fits. Approximate pointwise standard errors were calculated for the smoothed curves (see Figs 1, 2, 3). Sixteen females provided 70 samples of milk for analysis of composition while suckling a total of 24 pouch young. Because individual females provided 1-10 mi& samples, correlations between successive observations on the same animal are possible and should be accounted for in any inferential procedures. However, in this case, the correlations were generally found to be weak and hence observations were assumed to be independent for the fitting of the loess smoothers. The validity of the estimated standard errors was assessed by resampling procedures for two of the dependent variables with very different smoothed curves (protein and carbohydrate). Single observations were randomly selected for each mother, thereby ensuring independence, and a smoother was fitted in the same manner as for the full data set. The procedure was repeated. The envelope of these fits agreed well with the single fits to the full data sets when adjusted for the reduced sample sizes. The relationships between milk intake and pouch young age were predominantly linear or exponential and parametric linear regression models were used. For the exponential models with pouch young age as the dependent variable, log2 transformations were used, the choice of base simplifying estimation of doubling times for the independent variables. Differences between macropodid species in the relationship between milk consumption and pouch young age or body mass were examined by nested linear regression models. All data analyses used the S-Plus package (S-Plus for WINDOWS 1994).

Results Milk Solids T h e solids content of the milk averaged 16% (wlw) at 70 days post-parturition when pouch young were about 20 g (Fig. l a ) . Concentrations increased to a n average of 22% (wlw) by 170 days w h e n pouch young started spending increasing periods outside the pouch. During this period of temporary pouch emergence (170-200 days) the concentrations of milk solids were highly variable (Fig. l a ) . The concentrations declined and stabilised at about 19% by 200 days coincident with permanent vacation of the pouch (Delaney 1993). Most young were weaned by about 270 days although milk was obtained from one mother until her pouch young was 3 4 9 days old.


Protein The protein content of milk was about 3% (g per 100 mL) at 70 days and after a short plateau phase increased to concentrations of about 5.5% at about 200 days and continued to increase gradually to 6 3 % at 300 days (Fig. lb). Carbohydrate Carbohydrate concentrations increased slowly from about 8.5% (g per 100 mL) at 70 days to concentrations averaging 10% at 170 days (Fig. Ic). During the period of temporary pouch emergence carbohydrate concentrations decreased markedly to about 3% and continued to decline to 2% after 200 days. Lipid Lipid levels averaged about 3% (wlw) at 70 days post-parturition (Fig. Id). Concentrations then increased to maximum values during the period of temporary pouch emergence, averaging about 8% at 200 days. There was considerable individual variation during this period. Samples from the mothers of older pouch young suggest a continuing gradual increase in lipid concentrations during the period of weaning. Protein :Carbohydrate: Lipid Ratios The major part of the total solids in the milk during the period from 70 days until temporary pouch emergence at 170 days was carbohydrate, which decreased gradually from 54 to 41%. Protein was relatively stable at 21 to 18% over the same period while lipid increased from 20 to 36%. Between 170 days and permanent emergence at 200 days the carbohydrate contribution decreased markedly and was only 13% of the solids fraction at 215 days. Protein began to increase at temporary pouch emergence and was contributing about 37% by 215 days. Lipid continued to increase and by 215 days had reached 45% of the solids. Data are sparse after this time but suggest a continued gradual downward trend for carbohydrate and an increase for lipid. Milk Energy The change in total milk energy content (kJ per 100 g) during lactation is shown in Fig. 2a. Energy content was about 350 kJ per 100 g at 70 days after which it increased to a plateau of about 530 kJ per 100 g during the time of temporary pouch emergence. Energy content then increased gradually until weaning. The energy contribution from protein increased gradually from about 75 kJ per 100 g at 70 days to about 150 M per 100 g at pouch emergence (Fig. 2b). Carbohydrate and lipid were the main contributors to milk energy until temporary emergence of the pouch young at about 170 days when the contribution from carbohydrate decreased and lipid became the major contributor to milk energy (Fig. 2c, 6). Electrolytes At 7 0 days post-parturition Na concentrations were 33 mM; K concentrations were about 24 mM (Fig. 3a, b). Na concentrations then decreased to minimum values of 15 mM at 170 days while K concentrations declined only slightly to 20 mM. As a result Na and K concentrations crossed at about 140 days. Na concentrations then increased to 28 mM at 270 days and once more crossed K concentrations which continued at about 20-25 mM. For later lactation, data are sparse and highly variable. The pattern for Ca was different from those for Na and K. Ca concentrations were 38 mM at 70 days post-partum increasing to a maximum concentration of 57 mM at 150 days followed by a marginal decline thereafter. Mg concentrations gradually increased from 5 mM at 7 0 days to maximum values about 15 mM at 270 days but were increasingly variable in the later stages of lactation.

Pouch young age (days)

Fig. 1. Changes in the concentrations of (a) milk solids, (b) protein, (c) carbohydrate and (d) lipid in the milk of Petrogale assirnilis as a function of pouch young age. The solid line for each relationship represents a loess smoothed regression and the dotted lines approximate 90% confidence bands.



0

6 C .d a

8%

au e' 2

$

.-C

ri

Ld

.:$

5 2%

a

a X d f S

S g

3g*

- I B cg . $ J d j ! Z b8 sa 3M a ,'I08 a-T $ . 2 e ~ ~ ~ g uo 3a c8 c E 8 g g . 2 pa o.;; es 0 . &gn a a " E r . 5 0q.s 2 * ee-5 0 % 0 bg Qq =g i;; gsezz.g Q

gz

h -Q)


Milk Intake The daily rates of milk intake of 18 pouch young, ranging in average age from 72 to 159 days, increased from about 3 mL day-' to about 13 to 18.8 mL day-' (Table 1). Regressions of daily milk intake (I, mL day-') were calculated against both the age (A, in days) and body mass (M, in g) of the pouch young (Fig. 4a, b) and are described by Equations 1 and 2:

The relationship between body mass and pouch young age was exponential between the ages of 72 and 159 days. Body mass doubled every 35.2 days (95% CI = 31.3, 40.3 days). Milk intake also increased exponentially with pouch young age during this period (Fig. 4a) but at a slightly slower rate than the increase in body mass. Milk intake doubled every 42 days (95% CI = 34.6, 55.2 days). Milk intake was proportional to body mass to the power 0.801, this power being marginally less than 1 (t16 = 2.14, P = 0.048). Thus, the rate of increase of milk intake is less than the rate of increase of body mass. For example, the estimated milk intake per day for a 30-g pouch young was 3.76 mL day-l whereas for a 120-g animal it was 11.39 mL day-', 76% of the intake of the expected value if milk intake had increased proportionally with body mass. Since milk intake increased at a slower rate than body mass it would be expected that milk intake per body mass should decrease with pouch young age. The direct evidence for such a decrease was equivocal (P = 0.107) when the log2-transformed ratio of milk intake to body weight was regressed on pouch young age (Fig. 4c). The mass gained by pouch young over the period of measurement allowed estimation of crude growth efficiency (CGE): the efficiency with which the pouch young converted each rnillilitre of milk into body tissue (Table 1). The CGE ranged from 0.103 to 0.274 g mL-I [mean

Table 1.

Animal No.

Milk consumption and crude growth efficiencies (CGE) in pouch young of Petrogale assirnilis Age (days)

Mass (g)

Mass gain (g day-')

Milk intake (mLday-')

CGE (g m~-')

Energy intake (kJ day-')

-

2

4

8

120

160

Pouch young age (days)

80

2

64

128

Pouch young mass (g)

Y .2

C .-

C

([I

E

-2 J

C

.-C Y .-

([I

5.

I

7

n

$

0)

2

U

32

0

a E

([I

Y

a,

E

J

-

U

(d

2.

I

n

.-

2 h U

V) V)

0.062

0.088

0.125

0.177

120

160 Pouch young age (days)

80

Fig. 4. Milk intake by pouch young of P. assirnilis as a function of (a) pouch young age and (b) pouch young mass. (c) Milk intake per gram body mass shown as a function of pouch young age. Least squares regressions (solid lines) with 95% confidence bands (dotted l i e s ) indicate that milk intake increases exponentially with pouch young age and linearly with pouch young mass.

2

C .Y .-

C

Y ([I

a,

J

U

([I

I 2.

r

16

32


= 0.21 + 0.014 (s.e.) g mL-'1. As expected from the lack of a significant relationship between milk intake per unit body weight and pouch young age, CGE was not correlated with age of pouch young (? = -0.3343, n = 18, P = 0.175). No young were available for estimations of body composition, so their efficiency in using their daily energy intake in milk could not be assessed. Discussion Milk Composition This study presents the first sequential data from free-living macropodids on both milk composition and milk production. Sequential milk samples are generally difficult to obtain from free-living mammals but are important in studies of marsupial milk composition because of the major changes that occur during pouch life (reviewed by Green 1984; Green and Merchant 1988; Merchant 1989). With the exception of the work by Munks et al. (1991), Krockenberger (1996) and Muths (1996), studies of milk composition have used captive animals. The pattern of change in composition of milk from P. assirnilis during the pouch life of the young is similar to that observed for marsupials in captive conditions suggesting that, in this respect, captivity and, presumably, a different diet does not alter it greatly. Since no comparisons with milk samples from captive rock-wallabies were made we cannot comment on the possibility of differences in composition due to diet, which has been suggested by Munks et al. (1991) for the milk of the ringtail possum. Despite the overall similarity to the pattern of change in other marsupials the milk of P. assirnilis, over the period of temporary pouch emergence, decreased in milk solids content from maximal values about 22% (wlw) and higher to stabilise at about 19%. This is a feature shared only by the possums Pseudocheirus peregrinus (25 to 13%, Munks et al. 1991) and Trichosurus vulpecula (28 to 18%, Cowan 1989) and the koala, Phascolarctos cinereus (34 to 28%, Krockenberger 1996). Sparse data from the red kangaroo, Macropus rufus (= Megaleia rufa) (Lemon and Barker 1967), suggested that a decrease in milk solids content may occur also in this species but subsequent data presented by Muths (1996) provided no further evidence. The rock-wallaby, therefore, is the only macropodid studied so far that shows a decrease in milk solids at pouch emergence. On the basis of this decrease and other compositional characteristics of koala milk shared by the possums P. peregrinus (Munks et al. 1991) and T. vulpecula (Cowan 1989), Krockenberger (1996) suggests that these features may reflect their arboreal and folivorous lifestyle. The present observation on P. assirnilis does not necessarily contradict this hypothesis since P. assirnilis is described by Horsup and Marsh (1992) as a flexible feeder that eats mainly dicotyledonous plants and even climbs into trees for food. A sharp decrease in the concentration of milk solids also occurs in the bandicoot, Isoodon macrourus, at the onset of weaning (Merchant and Libke 1988) but in this species the young are completely independent within a few days so that the precipitous decline represents the termination of lactation. Cowan (1989) and Munks et al. (1991) relate the fall in milk solids concentrations to the increasing intake of solid food by young possums. Munks et al. (1991) suggest that with the increase in nutrients from another source there is less need for milk to satisfy all the nutritive and physiological requirements of the young. While most of this research has been carried out in captivity, Munks et al. (1991) include supporting data from a wild population. The concentrations of solids in the milk of P. assirnilis were highly variable during the period of temporary pouch emergence (Fig. la) and may reflect individual variation in the time spent by pouch young outside the pouch or sucking milk from the teat. Our data provide further evidence from a field situation that a decline in milk solids may be correlated with pouch emergence in some marsupials and the increasing consumption of solid food by the young. Eisenberg (1981) suggests that selection for milk of a high nutritive content may be relaxed in favour of both a lower energy drain on the mother and a prolonged mother-young bond. A long mother-young bond may be advantageous to P. peregrinus to provide young with the opportunity to learn complex foraging behaviours and arboreal agility (Munks et al. 1991). The


lifestyle of a young rock-wallaby is no less complicated: many food plant species must be identified (Horsup and Marsh 1992) and the trails that allow negotiation of precipitous rock faces at high speed must be learned. A long association with its mother would provide a young P. assirnilis with the opportunity to learn these skills. As in other macropodids such as Macropus eugenii, M. rufogriseus and M. rufus (Green et al. 1980; Merchant et al. 1989; Muths 1996), the protein content of the milk of P. assirnilis is low ( 3 4 % ) during the first 100 days of pouch life. In contrast, in the macropodoid brush-tailed bettong, B. penicillata (Merchant et al. 1994), the protein content of the milk increases from about 6.5 to 9% during this period. Carbohydrate concentrations in the milk of P. assirnilis are little different from those recorded for M. eugenii, M. rufogriseus and B. penicillata (Green et al. 1980; Merchant et al. 1989, 1994) and reach similar maximum values. The highest average value of 6.2% recorded by Muths (1996) for M. rufus is approaching half that of the other macropodids. The decline in carbohydrate concentration when young emerge from the pouch is characteristic of most marsupials so far studied, particularly the Macropodoidae (Merchant 1989). When P. assirnilis pouch young are 110 days old, lipid concentrations in the milk are similar to those in the milk of M. eugenii and M. rufogriseus with pouch young of the same age (4% cf. with 2-3% wlw), but by first pouch emergence at 170 days when emerging young have an increased energy requirement for locomotor activity and thermoregulation they have increased to about 8% (wlw). The pouch young of M. eugenii and M. rufogriseus do not leave the pouch temporarily until 210 and 230 days respectively and the increase in milk lipid concentrations is consequently delayed relative to P. assirnilis, although they reach 23% in M. eugenii and 12% in M. rufogriseus about the time their young leave the pouch (Green et al. 1983; Merchant et al. 1989). The recent study of M. rufus (Muths 1996) showed that milk lipid concentrations at about the time of pouch emergence, 235 days, were also about 10%. The limited data available suggest that the peak lipid values for M. eugenii are exceptionally high; most macropodids studied so far have maximum milk lipid concentrations comparable to those of P. assirnilis. The maximum energy content of the milk of P. assirnilis is about 530 kJ per 100 g, a value somewhat lower than that in M. rufogriseus, 800 kJ per 100 g (Merchant et al. 1989) and half that recorded for M. eugenii, 1150 kJ per 100 g, a value that reflects the higher lipid content of the milk of M. eugenii (Green et al. 1988). The smaller potoroids, Potorous tridactylus and B. penicillata, have more-concentrated milks with peak milk solids concentrations about 40% (wlw) and energy contents of about 900-1 100 kT per 100 g (Crowley et al. 1988; Smolenski and Rose 1988; Merchant et al. 1994). Specific differences within the Macropodoidae are reflected in the electrolyte (Na + K) concentrations of their milks at the time of peak carbohydrate concentration. Linzell and Peaker (1971) point out that the isosmolarity of milk is maintained predominantly by its carbohydrate and electrolyte content. The Na + K concentration at peak carbohydrate concentrations in the milk of P. assirnilis (38 mM) is within the range observed for three of the macropodid species for which there are data (M. eugenii, 37 mM, Messer and Green 1979; Green et al. 1980; M. rufogriseus, 47 mM, Merchant et al. 1989; Setonix brachyurus, 40 mM, Bentley and Shield 1962). In contrast, for M. rufus, this value is about 80 rnM (Muths 1996), which is therefore more similar to that measured in the milk of the two macropodoid species, P. tridactylus (80 mM, Crowley et al. 1988) and B. penicillata (70 mM, Merchant et al. 1994), which have considerably higher concentrations of Na + K in their milk at this time. The significance of such differences is obscure but has been associated with the molecular size of the milk oligosaccharides; small oligosaccharides exert greater osmotic pressures than large oligosaccharides and therefore lower concentrations of electrolytes are required to maintain isosmolarity (Messer and Green 1979). The Ca concentration of the milk of P. assirnilis in early lactation (40 mM; 1.6 g L-') is similar to values recorded for most other marsupial species (1.7-2.1 g L-', Green 1984; Green et al. 1991; Munks et al. 1991; Muths 1996) other than the wombat, Vornbatus ursinus (4.2 g L-l, Green


1984). In late lactation, milk Ca concentrations in P. assirnilis are about 55 mM (2.2 g L-I), which is at the low end of the range for other marsupials (2.3-53 g L-l) including other macropodids [e.g. M. rufus (72.6 mM) 2.9 g L-' (Muths 1996), M. eugenii 4 g L-l and M. rufogriseus 5.5 g L-' (Green 1984)l. In this respect, the milk of P. assirnilis is more similar to the late-lactation milk of the glider possum, Petaurus breviceps (2.3 g L-', Green 1984) and ringtail possum, P. peregrinus (1.7 g L-l) although there are no clear trends throughout lactation in the latter species (Munks et al. 1991). There are no data for V. ursinus on the Ca content of the milk during late lactation. As in other marsupials (Green 1984), Mg concentrations in the milk of P. assirnilis were much lower than those of Ca (cf. Fig. 3c and d). Mg concentrations increased from 5 mM (0.12 g L-l) at 70 days to 15 mM (0.34 g L-') at 270 days. Concentrations of Mg in the early milk of P. assirnilis are similar to those recorded for M, eugenii (0.1 g L-l), M. rufogriseus (0.1 g L-I), Dasyurus viverrinus (0.17 g L-') (Green et al. 1987; Green and Merchant 1988), M. rufus (0.16 g L-I; Muths 1996), V. ursinus (0.17 g L-'; Green 1984), P. peregrinus (0.14 g L-'; Munks et al. 1991), Monodelphis dornestica (0.21 g L-'; Green et al. 1991) and Didelphis virginiana (0.24 g L-l; Green et al. 1996). With the exception of P. peregrinus (Munks et al. 1991), M. dornestica (Green et al. 1991) and D. virginiana (Green et al. 1996) in which there is little change, Mg concentrations in the milks of the other species are higher in late lactation than earlier and range from 0.28 to 0-5 g L-'. There are no data for V. ursinus. Milk Consumption The milk consumption rates of the pouch young of P. assirnilis are the first published for a marsupial in the field. The only other data are for captive M. eugenii (Cork and Dove 1986; Green et al. 1988), B. penicillata (Merchant et al. 1994), 1. rnacrourus (Merchant et al. 1996) and P. peregrinus (Munks 1990). Milk consumption rates in P. assirnilis increased as pouch young increased in mass from about 30 to 184 g over the period 72-159 days (Table 1, Fig. 4a, b). However, there was no correlation between CGE and lactational stage in this age range. In M. eugenii (Green et al. 1988), CGE ranged from 0.21 to 0.25 over the first 154 days of lactation but had increased to 0.32 by 196 days and to 0.37 by 224 days. The increase in CGE was associated with an increase in milk solids, particularly lipid after 182 days (Green et al. 1988). Up to 159 days, when our measurements of milk intake stopped, the milk solids concentrations of P. assirnilis were very similar to those of M. eugenii (Green et al. 1980). It is unlikely that P. assirnilis would show the same increase in CGE as lactation progresses because after pouch emergence the milk solids content decreases and the energy requirements of the young for locomotion and thermoregulation presumably increases. Furthermore, in M. eugenii the increase in CGE occurred while the pouch young was still permanently in the pouch and entirely reliant on milk. Green et al. (1980) suggest that the increasing variability in the various parameters of milk composition in M.eugenii as the pouch young gets older may correlate more closely with pouch young body mass than age and consider that even the duration of lactation may be correlated with differential growth rates of young. Green et al. (1988) show that in M. eugenii there is also an increasing variability in milk intake rates with increasing pouch young age. They examined data for a group of young aged 110 days and showed that the variability was primarily due to variable growth rates of young influenced by maternal mass. Variances were greatly reduced when milk intake was regressed against pouch young body mass. Because of insufficient numbers of pouch young, we could not analyse the data for possible differential rates of milk intake in P. assirnilis against body mass at a specific age. However, in future, it may be preferable to use mass-specific estimates of milk intake by pouch young in assessing the milk output of mothers in P. assimilis. The milk composition and consumption data now available for P. assirnilis allow crude comparisons of the growth of their young with those of M. eugenii. Interspecific comparisons present obvious problems and the milk of a species is generally considered to be best suited to


the needs of its own young. M. eugenii and P. assirnilis are of similar adult size (about 4-5 kg, Green et al. 1988) but the time to permanent exit is 22% longer in M.eugenii (245 days) than in P. assirnilis (200 days) (Poole 1987; Delaney 1993). Restricting comparisons to the period 72-159 days of lactation for which we have data for P. assirnilis, their pouch young increase in mass from 30 g to about 180 g (Table 1) while the pouch young of M. eugenii increase from 30 to 210 g. The exponential growth curves for body mass for the two species were significantly different (Fig. 5a, F2,23 = 6.2, P = 0.007). M. eugenii had a higher growth rate than P. assirnilis, doubling its mass every 28 days compared with 35.2 days. There were no specific differences in the relationship between milk intake and body mass = 0.06, P = 0.943). The difference between species in between 72 and 159 days (Fig. 5b, F2,23 the variation of milk intake per unit body mass with age approached significance (Fig. 5c, F2,23 = 3.38, P = 0.056). The common regression showed a significant decrease with age (t2,23= 5.52, P < 0.001), suggesting that with more data the relationship in Fig. 4c would also be significant. It is, however, more appropriate to compare the two species on a relative scale based on the duration of pouch life and so reflect more accurately comparable developmental stages. The range for comparison is now 36-79% of pouch life duration, 72-159 days in P. assirnilis and 88-194 days in M.eugenii. Over this period daily milk intake increases from about 3 mL day-' to about 14 mL day-' in P. assirnilis and from 3.06 to 42.25 mL day-' in M. eugenii (Green et al. 1988). Dove and Cork (1989) report that milk intake in M. eugenii is highest at pouch emergence (86 mL day-') when pouch young weigh 1110 g. When the young of P. assirnilis leave the pouch permanently they are about 310 g (Delaney 1993). If P. assirnilis also reaches peak milk production at about the time of pouch emergence, we could expect its milk production a technique did not allow by 200 days to be higher than our estimates. The 2 2 ~ turnover estimation of milk intake beyond 170 days in this study, as pouch young may be consuming sodium from sources other than milk by this time. However, extrapolated estimates of milk intake at 200 days and 310-g body mass, from the age-based and mass-based equations (Equations 1 and 2 above), predict milk intakes of 25.3 and 26.5 mL day-' respectively. These values are well below those reported for M.eugenii. The greater milk intake of pouch young of M.eugenii is probably sufficient to account for most of the observed difference in body mass. Over the experimental period the efficiency of the pouch young of P. assirnilis in using milk averaged 0.21 + 0.014 (s.e.) g mL-' (n = 18) (Table 1) and was similar to that of M. eugenii (0.23 g mL-' from Green et al. 1988). By about 200 days, the pouch young of M. eugenii have increased their efficiency in using milk to 0.32 g mL-' (Green et al. 1988). It is apparent, therefore, that M.eugenii young benefit not only from a greater intake of milk but in late pouch life are more efficient than P. assirnilis in converting it into body tissue. The reason for the apparently marked differences in milk production by two species of similar adult size may reflect the considerable differences in the lifestyles of the two species. M. eugenii inhabits a temperate region where seasonal changes are highly predictable and it has evolved a well-defined reproductive pattern resulting in peak lactation and pouch young emergence coincidental with a good food supply. In contrast, the habitat of P. assirnilis is characterised by climatic extremes: high rainfall during the wet season, followed by dry seasons of varying severity. P. assirnilis breeds continuously (Delaney 1993) and therefore peak lactation and pouch young emergence may occur throughout the year and not necessarily at times conducive to successful lactation. A lower rate of milk production (= milk intake) and consequent lower growth rate of young may allow a mother to cope more successfully with times of severe food and water restriction. This study of lactation in P. assirnilis has shown that the patterns of change in milk composition determined in captive marsupials reflect those of a free-living species in its natural environment. In the field, up to the time of their emergence from the pouch, the young of P. assirnilis consume milk of quality similar to that of a similar-sized macropodid, the tammar wallaby, M.eugenii, in captivity, but consume much less. The influence of seasonal changes on a mother's ability to produce milk of adequate quality or quantity has not been examined. This



area of study may provide explanations for the marked seasonal differences in mortality of the pouch young of P. assirnilis observed by Delaney (1993) w h o found that pouch young born during the late dry season were approximately twice as likely to survive compared with those born in the remainder of the year.

Acknowledgments This research was supported by grants from the Australian Research Council, CSIRO and James C o o k University. W e thank the management of 'Lyndhurst' station for permission to work o n their property and Drs S. Cork, B. Green and L. Hinds and K. Newgrain and S. Delean for helpful criticism of the paper.

References Ben Shaul, D. M. (1962). The composition of the milk of wild animals, lnternutional Zoo Yearbook 4, 33342. Bentley, P. J., and Shield, J. W. (1962). Metabolism and kidney function in the pouch young of the macropodid marsupial, Setonix brachyurus. Journal of Physiology 164, 127-37. Bradford, M. M. (1976). A rapid and sensitive method for the quantification of microgram quantities of protein utilising the principle of protein-dye binding. Analytical Biochemistry 72, 248-54. Chambers, J. M., and Hastie, T. J. (1992). Statistical Models. In 'Statistical Models in S'. (Eds J. M. Chambers and T. J. Hastie.) pp. 13-44. (Wadsworth and Brooks: California.) Cleveland, W. S. (1979). Robust locally weighted regression and smoothing scatterplots. Journal of the American Statistical Association 74, 829-36. Conway, E. J. (1962). 'Microdiffusion Analysis and Volumetric Error.' (Crosby Lockwood & Son: London.) Cork, S . J., and Dove, H. (1986). Milk consumption in late lactation in a marsupial, the tammar wallaby (Macropus eugenii). Proceedings of the Nutrition Society of Australia 11,93. Cowan, P. E. (1989). Changes in milk composition during lactation in the common brushtail possum, Trichosurus vulpecula (Marsupialia: Phalangeridae). Reproduction, Fertility and Development 1, 325-35. Crowley, H., Rose R., and Woodward D. (1988). Changes in milk composition during lactation in the potoroo, Potorous tridactylus (Marsupia1ia:Potoroinae). Australian Journal of Biological Science 41, 289-96. Delaney, R. (1993). Life history and reproductive ecology of a tropical rock-wallaby, Petrogale assirnilis. Ph.D. Thesis, James Cook University of North Queensland. Delaney, R., and De'ath, G. (1990). Age estimation and growth rates of captive and wild pouch young of Petrogale assirnilis. Australian Wildlife Research 17,491-9. Dove, H., and Cork, S. J. (1989). Lactation in the tammar wallaby (Macropus eugenii). I. Milk consumption and the algebraic description of the lactation curve. Journal of Zoology (London)219, 385-97. Eisenberg, J. F. (1981). 'The Mammalian Radiations. An Analysis of Trends in Evolution, Adaptation, and Behaviour.' pp. 307-8. (The University of Chicago Press: Chicago.) Green, B. (1984). Composition of milk and energetics of growth in marsupials. In 'Physiological Strategies in Lactation'. (Eds M. Peaker, R. G. Vernon and C. H. Knight.) pp. 369-87. (Academic Press: London.) Green, B., and Merchant, J. C. (1988). The composition of marsupial milk. In 'The Developing Marsupial. Models for Biomedical Research'. (Eds C. H. Tyndale-Biscoe and P. A. Janssens.) pp. 41-54. (Springer-Verlag: Berlin.) Green, B., and Newgrain, K. (1979). Estimation of the milk intake of sucklings by means of 22Na. Journal of Mammalogy 60,556-9. Green, B., Newgrain, K., and Merchant, J. C. (1980). Changes in milk composition during lactation in the tammar wallaby (Macropus eugenii). Australian Journal of Biological Science 33, 35-42. Green, B., Griffiths, M., and Leckie, R. M. C. (1983). Qualitative and quantitative changes in milk fat during lactation in the tammar wallaby (Macropus eugenii). Australian Journal of Biological Sciences 36,455-61. Green, B., Merchant, J. C., and Newgrain, K. (1987). Milk composition in the eastern quoll, Dasyurus viverrinus (Marsupialia:Dasyuridae). Australian Journal ofBiologica1 Science 40, 379-87.


Green, B., Merchant, J. C., and Newgrain, K. (1988). Milk consumption and energetics of growth in pouch young of the tammar wallaby Macropus eugenii. Australian Jountal of Zoology 36, 217-27. Green, B., Van de Berg, J. L., and Newgrain, K. (1991). Milk composition in an American marsupial (Monodelphis domestica). Comparative Biochemistry and Physiology 99B, 663-5. Green, B., Krause, W. J., and Newgrain, K. (1996). Milk composition in the north American opossum (Didelphis virginianu). Comparative Biochemistry and Physiology 113B. 619-23. Horsup, A., and Marsh, H. (1992). The diet of the allied rock-wallaby, Petrogale assimilis, in the wet-dry tropics. Wildlife Research 19, 17-33. Jenness, R., and Sloan, R. E. (1970). The composition of milks of various species: a review. Dairy Science Abstract 32,599-612. Kleiber, M. (1975). 'The Fire of Life.' (Robert E. Krieger Publishing Co.: Huntington, NY.) Krockenberger, A. K. (1996). Composition of the milk of the koala, Phascolarctos cinereus, an arboreal folivore. Physiological Zoology 69, 701-18. Lemon, M., and Barker, S. (1967). Changes in milk composition of the red kangaroo, Megaleia rufa (Desmarest) during lactation. Australian Journal of Experimental, Biological and Medical Science 45, 213-19. Linzell, J. L., and Peaker, M. (1971). Mechanism of milk secretion. Physiological Review 51,564-97. Merchant, J. C. (1989). Lactation in macropodoid marsupials. In 'Kangaroos, Wallabies and Ratkangaroos'. (Eds G. Grigg, P. Jarman and I. Hume.) pp. 355-66. (Surrey Beatty & Sons: Sydney.) Merchant, J. C., and Libke, J. A. (1988). Milk composition in the northern brown bandicoot, Isoodon macrourus (Peramelidae, Marsupialia). Australian Journal of Biological Science 41,495-505. Merchant, J C., Green, B., Messer, M., and Newgrain, K. (1989). Milk composition in the red-necked wallaby, Macropus rufogriseus banhianus (Marsupialia). Comparative Biochemistry and Physiology 93A, 483-8. Merchant, J. C., Libke, J. A., and Smith, M. J. (1994). Lactation and energetics of growth in the brush-tailed bettong, Bettongia penicillata, (Marsupialia: Potoroidae) in captivity: ~ust&lianJournal of Zoology 41,267-77. Merchant, J. C., Libke, J. A., and Newgrain, K. W. (1996). Milk consumption and estimates of growth energetics in pouch young of the northern brown bandicoot, Isoodon macrourus (Peramelidae, Marsupialia), in captivity. Journal of Zoology (London) 238,483-94. Messer, M., and Green, B. (1979). Milk carbohydrates of marsupials 11. Quantitative and qualitative changes in milk carbohydrates during lactation in the tammar wallaby (Macropus eugenii). Australian Journal of Biological Science 32, 519-63 1. Munks, S. (1990). Ecological energetics and reproduction in the common ringtail possum, Pseudocheirus peregrinus. Ph.D. Thesis, University of Tasmania, Hobart. Munks, S., Green, B., Newgrain, K., and Messer, M. (1991). Milk composition in the common ringtail possum, Pseudocheirus peregrinus (Petauridae:Marsupialia). Australian Journal of Zoology 39, 403-16. Muths, E. (1996). Milk composition in a field population of red kangaroos, Macropus rufus (Desmarest) (Macropodidae : Marsupialia). Australian Journal of Zoology 44, 165-75. Poole, W. E. (1987). Australian marsupials. In 'The UFAW Handbook on the Care and Management of Laboratory Animals'. (Ed. T. B. Poole.) pp. 207-28. (Bath Press: Avon.) Smolenski, A. J., and Rose, R. W. (1988). Comparative lactation in two species of rat-kangaroo (Marsupialia). Comparative Biochemistry and Physiology 90A, 459-63.

Manuscript received 17 September 1996; revised and accepted 20 November 1996