INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE STANDARD METHOD OF QUANTIFYING energy expenditure is to measure oxygen consumption by indirect calorimetry and convert it to energy using a conversion factor for the substrate being oxidized (3, 8, 40). The principal advantages of this method are its accuracy and repeatability. However, the resultant measurement only reflects an animal's metabolism under the conditions of the chamber in which it is measured. When several animals are housed together within a chamber (for example, a lactating female and offspring), it provides a summed value of the oxygen consumption of all the individuals. Values of individual metabolism measured for separated offspring may be different from the metabolic rates of individuals when they are within the litter for a number of reasons. For example, they may have elevated metabolic rates due to an increased surface area by being prevented from huddling with conspecifics (5, 14, 39), the activity of isolated offspring might increase due to search behavior for warmth and/or conspecifics, because socially mediated (19) or CO₂-mediated (30, 31) metabolic suppression might not occur. Alternatively, separated offspring might be less active due to lack of stimulus from conspecifics (28). Indirect calorimetry experiments that are designed to assess the energy expenditures of individual suckling offspring may involve an extended period of separation from the mother that could be "unphysiological," particularly for species that suckle frequently (6).

The doubly labeled water (DLW) technique (16, 21, 32, 35) provides an alternative method of measuring CO₂ production and thus energy expenditure. It potentially overcomes some of the problems associated with indirect calorimetry and causes relatively minor disruption of behavior, food intake or energy expenditure (38). In the context of lactation, a major advantage of this technique is that it provides information on individual metabolism without the need to separate individuals from their social environment.

A fundamental assumption of the DLW technique, however, is that all substances entering the animal are labeled at background levels and there is no recycling or reentry of labeled or unlabeled CO₂ (17, 21). Isotope-based estimates of water turnover and milk intake might also be compromised by reentry of water (1, 23, 25). Difficulties occur if the isotope eliminated from one individual contaminates the route of uptake of another. The normal sources of exchangeable hydrogen entering the body are from free water (drunk), from preformed water in food, and water formed by oxidation of foodstuffs. Sources of oxygen include water by all the above routes and inhaled atmospheric oxygen. The most likely sources of additional exchanges of water and CO₂ are skin surface exchanges, including the respiratory tract (20, 21).

Elevated unlabeled ambient CO₂ levels might increase the rate at which oxygen enters and leaves the body (the exchange rate of CO₂ being dependent on the external concentration of CO₂), thus increasing the loss of ¹⁸O from the body and leading to subsequent overestimation of CO₂ production (21). This might pose a problem, for example, in fossorial animals, in which a limited mixing of air in the living space might result in a high ambient CO₂ concentration (10, 12, 15). In lactating common shrews Sorex araneus, a higher energy expenditure was measured by DLW than by indirect calorimetry (26). The authors concluded that this increase may have been due to elevated levels of unlabeled CO₂ around the mother caused by the respiration of her litter. Quantification of such effects are rare, but high levels of unlabeled ambient CO₂ led to an overestimation of CO₂ production by up to 80% in kangaroo rats Dipodomys merriami (21).

Elevated ambient levels of labeled CO₂ or water vapor might cause different effects depending on their enrichment in relation to the enrichment of the body of the animal (20). For instance, high levels of ambient CO₂ labeled at enrichments greater than in the body may decrease the apparent ¹⁸O turnover rate and thus reduce the estimate of CO₂ production. Alternatively, high levels of ambient ²H-labeled water would tend to decrease the apparent ²H elimination rate in a labeled animal without affecting the ¹⁸O elimination rate. This would potentially increase the estimate of CO₂ production (20). These effects may be particularly relevant when using DLW to measure CO₂ production in a lactating mother and her offspring or when estimating milk production using isotope transfer because of the long periods the mother and offspring spend in close contact. Many individuals in a small area might increase the local ambient concentrations of CO₂ and/or water vapor, either labeled or unlabeled, which could affect isotope elimination rates (21, 35).

In addition to gaseous routes, ¹⁸O and ²H might be recycled between individuals by other means. Lactating carnivores routinely ingest the urine and feces of their young, which serves both to maintain nest hygiene and minimize maternal water losses (2, 9, 33, 34). For example, tritium accumulated in the body of a lactating dingo Canis familiaris dingo when her young were injected with isotope (1). Similarly, when milk intake was measured by deuterium dilution in dog Canis familiaris puppies (24) and black bear Ursus americanus cubs (25) and one pup or cub from each litter was left unlabeled to serve as a control, significant increases in deuterium enrichment were found in these controls. Deuterium accumulated in control puppies by up to 11.5% and 10.6% of the levels in dosed puppies (at 15-16 and 22-23 days of lactation, respectively) (24) and, if not accounted for, this has been suggested to produce underestimates of milk production by up to 13.7% (25).

When one-half of the nursing pups in rat Rattus norvegicus litters were dosed with tritiated water (11), mother rats reclaimed roughly two-thirds of the water transferred to pups in milk. Tritium also accumulated in the bodies of the unlabeled control pups. In this system, the authors showed that at least one-half of the tritiated water in the initially unlabeled pups came not from milk, but directly from the labeled pups and suggested the transfer was a result of gaseous exchange of enriched water vapor. Thus the mother's consumption of her young's urine and the sharing of evaporated water by huddled siblings were important in the pattern of water and isotope recycling.

We performed a series of experiments to quantify the extent of isotope transfer between lactating domestic dogs Canis familiaris and their dependent offspring to establish their potential effects on DLW estimates of CO₂ production and isotope estimates of water turnover. Previous studies have left a single offspring unlabeled to control for recycling (24, 25). If differences in uptake of isotope are significant between offspring, then leaving a single pup unlabeled might not be a representative sample to assess the extent of recycling. Therefore, when labeling offspring we left one-half of the pups in each litter unlabeled to examine the consistency of the response in the uptake curves of isotope enrichment between the unlabeled individuals. Our data provide an indication of the likely routes by which recycling occurs during lactation and the magnitude of such effects in this species.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals and housing. Fifteen lactating domestic dogs aged 2-6 yr with litters aged 24-30 days were used in this study. Three were Miniature Schnauzers weighing on average 6.0 kg (SD = 0.9), and 12 were Labrador Retrievers, weighing 29.6 kg (SD = 2.6). At 24 days old, Labrador pups had a mean weight of 2.2 kg (n = 18, SD = 0.4) and Schnauzer pups 0.78 kg (n = 6, SD = 0.1). Animals were bred and reared at the Waltham Centre for Pet Nutrition (WCPN), Melton Mowbray, UK. Mothers were housed with their litters in a climate-controlled building at 20-25°C in individual pens with access to outside runs. Details of the design of the housing conditions and the pens have been described previously (18). Pens were 2 × 2.6 × 3.2 m high with concrete floors. Synthetic bedding and a heated area of the floor of each pen were provided for the offspring to sleep in. Each pen was illuminated naturally through a window and supplemented by electric lighting (12:12 light to dark). Adults were fed a dry commercial diet (Waltham Formula Expert Growth) from the onset of gestation until the offspring were eventually weaned. The diet consisted of 26.9% crude protein, 14.4% crude fat, 44.8% nitrogen-free extract, 7.4% water, and 6.5% ash and had a metabolizable energy content of 15.5 kJ/g (WCPN, unpublished data). Food and water were supplied in deep stainless steel bowls that were too high for the pups to eat or drink from. Two Labrador litters (mean = 2.35 kg, SD = 0.21, n = 16 pups) were weaned early to examine pup-to-pup isotope transfer in the absence of the mother. Pups in these two litters were introduced to solids (Pedigree Chum puppy food) at 14 days old and by 24 days were fully weaned.

DLW protocol. On day 24 of lactation, blood samples of all animals were taken to estimate background enrichments of isotopes. Experimental animals were weighed (Sartorius F150 balance, ±0.1 g, Gottingen, Germany), then dosed intravenously with DLW [2.3 parts 90% enriched ¹⁸O water (Enritech Rehovot, Israel) and one part 99.9% enriched ²H water (MSD Isotopes Pointe-Claire, Quebec, Canada) made isotonic with sodium chloride]. Doses were 7.0 ml for adults and 1.0 ml for pups, ~0.3 ml/kg body weight. Syringes were weighed before and after administration of isotope (Sartorius 4-figure balance ± 0.0001 g). Blood samples were taken 5 h postdose in adults and 2.5 h postdose in pups to estimate initial isotope enrichments. This was sufficient time for isotopes to equilibrate within the body water of the animals (M. Scantlebury and J. R. Speakman, unpublished data). Thereafter, blood samples were taken at 24-h intervals for 4 days (experiment 1) and 7 days (experiments 2 and 3) to estimate isotope elimination rates (36). Blood samples were collected into 1.5-ml heparinized serological tubes from which 3 × 100 µl glass capillaries were immediately filled and heat sealed. The remaining blood was frozen at 80°C as a backup. To estimate the enrichment of the original injectate, a known weight of the DLW injectate (~0.2 g) was diluted with a known weight of tap water, the enrichment of which had been determined previously. The enrichment (of the resulting mixture) was used to calculate the enrichment of the injectate (27, 35). The enrichment of the original injectate was calculated from the mean of 10 such dilutions.

Samples of blood (from the 100-µl sealed capillaries), feces, urine, and milk were vacuum distilled into Pasteur pipettes (22), and the resultant distillate was used for the determination of ²H and ¹⁸O content (42). Hydrogen gas was prepared using the zinc reduction technique for the determination of ²H:¹H enrichment (42) and carbon dioxide gas by small sample equilibration (37) for determination of ¹⁸O:¹⁶O enrichment. During analysis some of the reaction vessels for determination of ²H enrichment were lost due to an analytical accident. Therefore, there was a larger sample size of ¹⁸O excreta results than ²H excreta results. ¹⁸O and ²H isotope ratios were analyzed using a gas source isotope ratio mass spectrometer (Optima, Micromass IRMS, Manchester, UK), using isotopically characterized gasses of CO₂ and H₂ (CP grade gases BOC, Aberdeen, UK) in the reference channel. Cylinder gasses were characterized relative to standard mean ocean water (SMOW) and standard light arctic precipitate (7, 41), supplied by the International Atomic Agency (Vienna, Austria). We ran duplicate analyses of three laboratory standards with each batch of samples that allowed elimination of any effect of daily variation in the performance of the mass spectrometer. All isotope enrichments were measured in delta per mil relative to the working standards, after which they were normalized to delta relative to SMOW (35) and then converted to parts per million (ppm) using the established ratios for the reference materials. In all cases the log-converted isotope enrichments above background in labeled individuals decreased linearly over time after injections. The r² values averaged 99.5% for both isotopes in both breeds.

Experiment 1: transfer of ¹⁸O and ²H between pups in the absence of the mother. Two fully weaned 24-day-old litters of 10 and six pups were used in this experiment. Pups of each litter were housed together as intact groups. All individuals were blood sampled to estimate initial isotope enrichments, after which one-half the pups in each litter (5 pups and 3 pups) were labeled with DLW. All individuals were blood sampled for the next 4 days to determine isotope elimination rates in labeled individuals and possible isotope uptake in unlabeled individuals.

Experiment 2: transfer of ¹⁸O and ²H between pups and from pups to mother. Six Labrador and three Miniature Schnauzer bitches with 24-day-old unweaned litters (6 Labrador litters, mean = 6.5, SD = 0.84 pups and 3 Schnauzer litters, mean = 4.7, SD = 1.2 pups) were used in this experiment. All animals were blood sampled to obtain background levels of isotopes, after which one-half the pups in each litter were injected with DLW. All animals were blood sampled for the next 7 days to determine isotope elimination in labeled individuals and possible isotope uptake in unlabeled individuals.

Experiment 3: transfer of ¹⁸O and ²H from mother to pups. Six Labradors with their litters (mean = 6.2, SD = 0.98 pups) were used in this experiment. All individuals were blood sampled on day 24 of lactation to estimate background isotope enrichments. Adults were then dosed with DLW, and all individuals were blood sampled for the next 7 days to determine isotope elimination in mothers and uptake in offspring.

Collection of pup feces, pup urine, and milk. In addition to blood samples, approximate daily samples of pup excreta (feces and urine) were collected immediately after production to establish a potential route of transfer of isotope from offspring to mother. Excreta were collected within 10 min of production, and stored at 80°C in rubber-sealed glass bottles until analysis. Pups were allowed to leave their maternal pens twice each day to enable us to obtain samples of excreta without the attentions of the mother. Samples of excreta were collected from pups during this period because the mothers otherwise immediately ate all excreta that was produced. Milk samples were collected daily from each mother by manual expression into 20-ml glass bottles. Pups were removed from their mother 40 min prior to milking. All nipples were milked, and all glands were emptied as completely as possible (24). Oxytocin (10 IU/ml im dose, 0.5 ml for Miniature Schnauzers, 1.0 ml for Labrador Retrievers) was administered prior to milking on alternate days in experiments 2 and 3. Milk samples were stored (at 80°C) in rubber-stopped glass bottles until analysis. Milking took ~20 min, after which the pups were returned to their mother.

Statistical methods. Least squares linear regression was used to analyze changes in isotope enrichment over time. Reduced major axis regression was used to analyze correlations between excreta and milk enrichment and blood isotope enrichment. Analysis of covariance (ANCOVA) was used to examine differences in gradients of enrichment increases between individual unlabeled pups in experiment 1 and between unlabeled pups and unlabeled mothers in experiment 2. In both cases, litter size was entered as a covariate. One pup per litter was selected at random for analysis in experiments 2 and 3 to ensure independence of measurements. Statistical analyses were performed using Minitab 5.2 software (29).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiment 1: one-half the pups in a litter labeled in the absence of the mother. There was a significant decrease in the enrichments of both isotopes over time in water distilled from the blood of the labeled individuals of both litters (Table 1). There was no change in the enrichment of isotopes in the blood of the unlabeled pups of either litter over time (Table 1). The differences in the gradients of the enrichment curves between the unlabeled pups of each litter were not significant for either isotope [litter 1 ANCOVA: F(3,11) = 1.23, P = 0.346 for ¹⁸O, and F(3,11) = 1.45, P = 0.280 for ²H; litter 2 ANCOVA: F(2,8) = 0.620, P = 0.562 for ¹⁸O and F(2,8) = 0.670, P = 0.539 for ²H]. Absence of an overall trend in isotope enrichment was thus not a result of an increase in some individuals at the same time as a decrease in others. The data for these two litters indicate that there was negligible transfer of ¹⁸O or ²H from labeled pups to unlabeled pups in the absence of the mother.

Experiment 2: one-half the pups of a litter labeled in the presence of the mother. The log-converted enrichments above background in water distilled from the blood of the labeled pups decreased linearly over the experimental period (Table 2). From each litter, one labeled pup and one unlabeled pup were selected at random for inclusion in the analysis to ensure independence of measurements. In the unlabeled pups and mothers, enrichment levels of both isotopes increased significantly throughout the experimental period (data pooled across all individuals; Table 2). The gradient of the ¹⁸O enrichment curve was steeper for the unlabeled pups than for unlabeled mothers [ANCOVA: F(1,92) = 4.95, P = 0.028; Fig. 1A]. Pups increased in ¹⁸O by an average of 12.5 ppm over the 7 days (SD = 5.4, n = 9), which was greater than the increase in unlabeled mothers (mean = 4.6 ppm, SD = 4.2, n = 9; t = 2.62, P = 0.021) over the same period. In contrast, there was no significant difference in the gradients of the enrichment curves of unlabeled pups and unlabeled mothers for the ²H isotope. However, the enrichment curve of the pups was significantly elevated above that of the mothers [ANCOVA: F(1,92) = 4.52, P = 0.036; Fig. 1B]. Unlabeled pups increased in ²H by 6.6 ppm (SD = 2.5, n = 9) and unlabeled mothers by 4.0 ppm (SD = 2.2, n = 9; t = 2.35, P = 0.036). Within each of the nine litters, there were no significant differences in the gradients of the enrichment curves between the unlabeled pups (Table 3).

View larger version (1615K): [in this window] [in a new window]   Fig. 1.   Enrichment increases [parts/million (ppm)] above background of 18O (A) and 2H (B) in unlabeled offspring and unlabeled mothers when one-half of members of a litter were labeled with 2H218O. Dashed lines and open circles represent enrichment increases in unlabeled offspring. Filled circles and continuous lines represent enrichment increases in mothers. Data pooled across all litters (n = 9). High variability reflects mostly variation between individuals.

The enrichment of ¹⁸O and ²H in water distilled from pup excreta was positively correlated with the enrichment of ¹⁸O and ²H respectively in water distilled from simultaneously collected pup blood. The reduced major axis regression equation for ¹⁸O was ¹⁸O (excreta) = 0.833 ¹⁸O (blood) + 340 (r² = 0.86, P < 0.001, n = 37, data pooled across all pups, Fig. 2A). Similarly, the reduced major axis regression equation for ²H was ²H (excreta) = 0.990 ²H (blood)  0.505 (r² = 0.975, P < 0.001, n = 27, data pooled across all pups, Fig. 2B). Fewer samples were available for ²H enrichment determination due to loss of some samples during analysis.

View larger version (1918K): [in this window] [in a new window]   Fig. 2.   Relationship between excreta isotope enrichment and blood isotope enrichment in pups. A: 18O enrichment; B: 2H enrichment. Correlations for both isotopes were highly significant. Dotted lines are lines of equality. Open and solid circles represent urine and feces, respectively. On average, urine was higher in enrichment than feces because samples were taken at different times in experiment.

There was no significant difference between the measured ¹⁸O enrichment of the pup excreta and the ¹⁸O enrichment of the simultaneously collected pup blood (paired  test, P = 0.79). There was, however, a significant difference between the deuterium enrichment of pup excreta and that simultaneously measured in pup blood (paired  test, P = 0.0015) with pup blood being on average 7.2 ppm (SD = 10.5, n = 27) more enriched than pup excreta.

Experiment 3: transfer of ¹⁸O and ²H from mother to pups. When DLW was administered to lactating mothers, there was a significant decrease over time in the log-converted enrichments above background of both isotopes in each mother (Table 4) (Fig. 3, A and B). Labeling lactating mothers with DLW produced significant increases in enrichment of ¹⁸O and ²H in the blood of all the pups in their litters (Table 4) (Fig. 3, A and B). Pups increased in ¹⁸O enrichment by an average 31.2 ppm (SD = 8.3, n = 6 litters) and in ²H by an average 22.2 ppm (SD = 4.4, n = 6 litters).

View larger version (1313K): [in this window] [in a new window]   Fig. 3.   Isotope elimination curves (ppm) of mother 1 and uptake curves in her litter. A: 18O enrichment; B: 2H enrichment. Error bars denote standard deviations of pup enrichments.

There was a significant positive relationship between the enrichment of ¹⁸O in the blood of labeled mothers and the enrichment of ¹⁸O in simultaneously collected milk. The reduced major axis regression equation was ¹⁸O (milk) = 0.956 ¹⁸O (blood) + 89.2 (r² = 0.993, P < 0.001, n = 20, data pooled across all 6 mothers, Fig. 4A). Blood was on average 3 ppm more enriched in ¹⁸O than milk (mean = 2,091 ppm, SD = 77 for milk; mean = 2,094 ppm, SD = 76 for blood; paired t = 2.26, P = 0.036). ²H milk enrichments were also positively correlated with simultaneously collected ²H blood enrichments. There was no difference in the enrichment of ²H between blood and milk (mean = 203 ppm, SD = 39 for milk; mean = 205 ppm, SD = 39 for blood; paired t = 1.72, P = 0.10). The reduced major axis regression equation was ²H (milk) = 0.992 ²H (blood) + 1.65 (r² = 0.992, P < 0.001, n = 19, data pooled across all 6 mothers, Fig. 4B).

View larger version (1714K): [in this window] [in a new window]   Fig. 4.   Relationship between milk isotope enrichment and blood isotope enrichment. Data were pooled across all 6 mothers. A: 18O enrichment; B: 2H enrichment. Correlations for both isotopes were highly significant.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiment 1. There were no increases in enrichment of either isotope in the unlabeled pups of either litter, despite the likelihood that transfer of isotope from labeled to unlabeled pups could occur by a variety of mechanisms. This suggests firstly, that the ambient levels of CO₂ and water vapor, labeled or otherwise, surrounding the pups were insufficient to cause significant transfer of isotopes between individuals. Secondly, the fecal-oral route was unlikely to have been a source of isotope contamination between these pups. Pups from all three experiments were monitored with time-lapse video recorders for 24-h periods. The pups spent much of their time huddled together, and there were no obvious behavioral differences between groups. Failure to detect isotope uptake between weaned pups was therefore not because they spent most of their time separated. Recycling between pups would be unlikely to be a significant source of error in calculations of water turnover or DLW estimates of CO₂ production in groups of pups in the absence of the mother.

Experiment 2. In theory, the enrichment of both isotopes might be expected to be higher in the excreta than in the blood because of the time lag in their equilibration with blood. However, the ²H enrichment of the excreta was lower, and there was no significant difference for ¹⁸O. This may indicate the excreta remain in exchange equilibrium until their elimination. The ²H enrichment of the excreta may be reduced by interactions with exchangeable hydrogen in the solid component of the waste following excretion.

In this experiment, pups could interact with their mother as well as each other. A potential route of isotope transfer was, therefore, from labeled pup to mother to unlabeled pup. This might occur, for example, as a result of the mother ingesting excreta from labeled offspring and subsequently transferring isotope to unlabeled offspring via milk, as has been suggested to occur in suckling beagles (24) and black bears Ursus americanus (25). In nursing laboratory rats Rattus norvegicus, gaseous routes of isotope transfer might play an important role in isotope recycling (11). In the current experiment, unlabeled pups did increase in the enrichment of both ¹⁸O and ²H. This probably did not occur as a result of pup-to-pup interactions, for example by rebreathing labeled exhaled CO₂ or water vapor, as no increase in enrichment of unlabeled individuals was observed in experiment one. This may have indicated that any increase in isotope enrichment of the unlabeled pups was likely to have involved the mother. However, the enrichment increases in all nine mothers were significantly less than the increases observed in their unlabeled pups (Table 2, Fig. 1, A and B). This precludes offspring excreta ingestion by the mother and subsequent isotope transfer via milk as the sole recycling mechanism and implies an additional route or routes.

One possibility is that unlabeled nursing pups may have obtained recycled isotope by the exchange of saliva on nipples. With the use of the mean differences in enrichment between unlabeled and labeled offspring over the time period (~300 ppm difference in ¹⁸O and 150 ppm difference in ²H), a 2-kg unlabeled pup would have to consume at least 50 ml of enriched saliva over 7 days to achieve the 7- to 8-ppm ¹⁸O and 2- to 3-ppm ²H differences in enrichment observed between unlabeled pups and unlabeled mothers (~7.5 ml of enriched saliva each day). If a nursing pup took 0.5 ml of saliva from a mammary gland, which had been left there by the previous individual, each time it started to suckle, then it would need to start suckling at a gland contaminated with enriched saliva 15 times in the course of a day to take in sufficient saliva to account for the observed increases in isotope enrichment. The likelihood of an unlabeled pup coming into contact with a mammary gland contaminated with enriched saliva depends on the ratio of labeled to unlabeled pups. With 50% of the pups labeled, as in the current experiment, the total number of times a pup would need to start suckling at a new gland over 24 h would be 30. Our impression was that this level of suckling activity at the glands does occur over 24 h, although we did not collect quantitative data on this aspect of the suckling behavior.

There is another possibility that might have occurred in addition to saliva ingestion. Although there was no transfer of isotope between weaned pups 24-28 days of age (experiment 1), this does not necessarily mean that there was no direct transfer of isotope between labeled and unlabeled pups of the same age that had not been weaned (experiment 2). This may occur because the composition of feces from unweaned offspring is different from weaned offspring; the former being primarily of endogenous origin, whereas the latter includes undigested food residues. Hence, the palatability of feces to pups may have differed between experiments 1 and 2. It is possible, therefore, that the pups in experiment 2 ingested the feces of the labeled pups, but the weaned pups of a similar age (experiment 1) did not. During observations, the mother always ate pup excreta, and, at no time in either experiment was pup excreta seen to be eaten by another pup. However, because most observations were performed during the day, it was possible that some isotope transfer via excreta ingestion between pups occurred during the night when the pups were huddled together. We cannot eliminate this as a potential additional route of transfer.

The increase in enrichment of unlabeled offspring was consistent within litters. Because the current data show that there were no significant differences between the isotope uptake rates of the unlabeled pups within any given litter, these experiments suggest that leaving a single pup rather than several unlabeled offspring is an adequate experimental design to control for the effects of recycling.

One of the impacts of recycling might be to depress the measured elimination rates of both isotopes. Therefore, removal of the effects of recycling would increase the apparent elimination rates of both isotopes. In the current experiment, the effect on the ¹⁸O was greater than on the ²H. When this effect was removed, both elimination curves increased in gradient, but the effect was greater for ¹⁸O. Hence, one might expect the energy expenditure to differ when calculated after removal of the effects of recycling. We modeled the effect of these changes for measurements of CO₂ production by DLW in a typical puppy. The calculated energy expenditure of a 2-kg animal with a respiratory quotient of 0.85 and elimination rates of 0.175/day of ¹⁸O and 0.132/day of ²H without the effects of recycling was 917 kJ/day (parameters based on our own unpublished data). We assumed that the amount of isotope recycled to the labeled puppies was the same as that recycled to the unlabeled puppies. Hence, the decrease in enrichment of the labeled puppies over the 7 days would have been 12.5 and 6.6 ppm for ¹⁸O and ²H, respectively, greater than the measured values if there were no recycling. With the effects of recycling taken into account, the recalculated values of the elimination rates became 0.194/day for ¹⁸O, an increase of 11.1%, and 0.146/day for ²H, an increase of 10.9%. Accordingly, because deuterium turnover is directly proportional to water turnover (35), deuterium-based calculations of water turnover will be 10.9% greater when the effects of recycling are accounted for. The recalculated value of energy expenditure was 980 kJ/day, 7% greater than the previous value without the effects of recycling.

Experiment 3. This experiment examined the transfer of ¹⁸O and ²H from lactating mothers to offspring when the mothers were dosed with DLW. Milk would be expected to be an important route of isotope transfer (4, 13), but potential additional routes might be via ingestion of mother's excreta by the pups, or rebreathing of exhaled labeled CO₂ by the mother or the pups (21). ¹⁸O and ²H administered to lactating mothers rapidly appeared in their milk and consequently in their nursing offspring. Over a 7-day period, pups increased in enrichment of ¹⁸O by 31.2 ppm and ²H by 22.2 ppm. Enriched offspring could potentially recontaminate the mother by the processes outlined in experiment 2. However, the relatively modest enrichment increases (above background) observed in experiment 3 in pups are unlikely to have a significant effect on the enrichment of their mother. The 2- to 300-ppm enrichments above background in pups observed in experiment 2 only produced increases in enrichment in the mother of both isotopes by ~4 ppm. Recycling may, therefore, have profound effects on water turnover and energy expenditure estimates in suckling offspring, but less so in lactating mothers.